403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 06_jes_296_11_16_2020 journal of engineering science 11(2), 2020, 61-73 doi: https://doi.org/10.3329/jes.v11i2.50898 predictability of orographic rainfall features over sylhet using nwp model md. omar faruq1, 2*, m. a. k. mallik1, m. a. m. chowdhury2, m. a. e. akhter3 and m. arif hossain1 1bangladesh meteorological department, agargaon, sher-e-bangla nagar, dhaka 1207, bangladesh 2department of physics, jahangirnagar university, savar, dhaka, bangladesh 3department of physics, khulna university of engineering & technology, khulna 9203, bangladesh received: 18 october 2020 accepted: 16 november 2020 abstract topography and orography are two physical factors which produce high impact rainfall over the north-eastern part of bangladesh. to predict the orographic rainfall of 29 march 2017 over sylhet, bangladesh an attempt has been performed using weather research and forecasting (wrf) model. the model has been run in a single domain of 10 km horizontal resolution for 48-h and 72-h using six hourly global final datasets from 0000 utc of each initial day of the event as initial and lateral boundary conditions with nssl 2-moment microphysics scheme, kain–fritsch cumulous scheme and yonsei university planetary boundary layer (pbl) scheme. the model outputs such as sea level pressure, wind flow, vorticity, wind shear, humidity, convective available potential energy (cape), convective inhibition, lifted index, k-index, total total index and rainfall have been analyzed. the model predicted weather parameters were visualized by grid analysis and display system (grads) and geographic information system (gis) software and validated with observed data of bangladesh meteorological department (bmd), tropical rainfall measuring mission (trmm) and european centre for medium-range weather forecasts (ecmrwf) data. the analysis determines that the cape of magnitude 8001000 jkg-1, positive vorticity of (6-10)×10-5s-1 and relative humidity of 80-100% up to 500-400 hpa levels are accountable for the happening of the orographic extreme rainfall and other parameters are compatible with the observed or theoretical values. this study indicates that the model with an appropriate model set up is capable to predict the orographic precipitation realistically well and can be used for upcoming events. keywords: wrf-arw; vorticity; cape; cin; li; tt. 1. introduction tropical bangladesh is situated in the north-eastern part of south asia. the great himalayas stand at some distance to the north, while in the south lays the bay of bengal (bob). west bengal borders on the west and in the east lies the hilly and forested regions of india and myanmar. these pretty topographical settings of low lying plain of about 1,47,570 square kilometers has been crisscrossed by many rivers and streams. the elevation of delta area is not more than 150 m above sea level and most of it belongs to 1-2 m above sea level. flood water covers most of the land surface during the rainy season and damages crops and decrease national economy. the orographic rainfall (or) mainly occurs in the northeastern hilly regions of bangladesh (faruq et al., 2019). when moist air is lifted and moves over a mountain range, orographic effect is produced. as the air rises and cools, orographic convective clouds form and serve as the source of rainfall, most of which falls upwind of the mountain ridge. some also fall a short distance downwind of the ridge and sometimes called spillover. on the lee side of the mountain range, rainfall is usually low, and the area is assumed to be in a rain sleuth. the influence of mountains upon rain is often profound, generating some of the earth's dampest places. orographic effects on rainfall are also responsible for some of the planet's sharpest climatic transitions. the classic example is the so called 'rain shadow'; for a mountain range oriented perpendicular to the prevailing winds, rainfall is greatly enhanced on the windward side and suppressed in the lee. however, the full range of orographic influences is much broader than this. rainfall can be heightened in the lee, over the crest, or well upwind of a mountain. almost all orographic influences on rainfall occur due to rising and descending atmospheric motions forced by topography. these motions can be forced mechanically, as air affecting on a mountain is lifted over it, or thermally, as heated mountain slopes prompt buoyancy-driven rotations. rising motion causes the air to expand and cool, which is significant since the amount of water that may exist as vapor in air is an approximately exponential function of temperature. thus, if cooling is adequate, air saturates and the water vapor condenses into cloud droplets or forms cloud ice-crystals. these droplets and crystals develop by various processes until they become large enough to fall as rain drop. it is important to emphasize that moist ascent over topography alone is typically inadequate to make rainfall. these orographic effects generally modify rainfall during pre-existing storms (browning et al., 1974; smith et al., 2006). globally, the advanced research wrf (arw) model is being usually used for the mockup of a variability of weather events, such as heavy rainfall (niyogi et al., 2006; routray et al., 2010; dodla and ratna, 2010; osuri et al., 2012) and tcs (osuri et al., 2012; pattanaik et al., 2009; davis et al., 2008). * corresponding author: mallikak76@yahoo.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 62 md. omar faruq et al. predictability of orographic rainfall features ……. so, an effort has been made to forecast the high impact rainfall of 119 mm recorded over sylhet, bangladesh on 29 march 2017 by wrf model and an endeavor is also taken to find out the meteorological features of that event. 2. experimental setup, data used and methodology the advanced research wrf model, developed at the national center for atmospheric research (ncar), is one of the two different dynamical hubs of the wrf model. the other core version, the non-hydrostatic mesoscale model (nmm), was developed at the environmental modeling center of the national centers for environmental prediction (ncep). the numerical weather prediction (nwp) model used in this study is the arw model of version 3.9.1. the wrf model provides a flexible and portable open-source community model for both atmospheric research and operational forecasting (skamarock et al., 2008). it is a limited-area, non-hydrostatic basic equation model with multiple options for numerous parameterization schemes for different physical methods. figure 1: domain for the nwp study figure 2: kalpana-1-tir imageat (a) 1415 utc and (b) 2115 utc of 29 march 2017. table 1: wrf model and domain configurations dynamics non-hydrostatic wrf version 3.9.1 number of domains 1 central points of the domain 230 n and 900e horizontal grid distance 10 km number of grid points 251×251 run time 72 and 48 hours time step 25 map projection mercator vertical levels 38 horizontal grid distribution arakawa c-grid time integration 3rd order runge-kutta spatial differencing scheme 6th order centered differencing initial conditions final (fnl: 1° × 1°) lateral boundary condition specified options for real-data top boundary condition gravity wave absorbing bottom boundary condition physical or free-slip diffusion and damping simple diffusion microphysics nssl 2-moment scheme cumulus physics kain-fritsch (kf) scheme land surface parameterization 5 layer thermal diffusion scheme pbl parameterization yonsei university (ysu) scheme radiation scheme rrtm for long wave and dudhia for short wave surface layer monin-obukhov similarity theory scheme journal of engineering science 11(2), 2020, 61-73 63 the 1⁰ resolution fnl data covering the entire globe every 6-h were taken as the initial and lateral boundary conditions. 30 arc sec united states geological survey (usgs) data gtopo30 were used as topography and 25 categories usgs data were taken as vegetation/land use (modis and hi-def lakes) coverage. the observed rainfall data of bmd were used to compare or authenticate the model simulated rainfall. the simulation was done on a single domain of 10 km horizontal resolution and the domain (251×251 grid points with 38 unequally spaced sigma levels) of the nwp study is shown in figure 1. the details of the model and domain configuration are listed in table 1. 3. result and discussions at first the multi-cell thunderstorm was not developed over sylhet and adjoining area at 1415 utc of 29 march 2017 and is not captured by kalpana-1-tir shown in figure 2(a). at 2115 utc of 29 march 2017, the multi-cell thunderstorm is developed over sylhet and adjoining area which is merged afterwards and captured by kalpana1-tir satellite and shown in figure 2(b). different meteorological parameters are analyzed to describe orographic rainfall over sylhet and adjoining area. (hpa) figure 3: observed mslp on 0000, 0600, 1200 & 1800 utc of (a-d) 29 march 2017; predicted mslp on 0000, 0600, 1200 and 1800 utc of 29 march based on the initial condition of 0000 utc of (e-h) 27 march and (i-l) 28 march 2017, respectively. 3.1 sea level pressure the analysis of mslp on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 compared with observed mslp on 0000, 0600, 1200 and 1800 utc of 29 march 2017 which is presented in figures 3(a-l). the model has simulated 3-4 isobaric lines with two hpa differences penetrated towards bangladesh from bob and a trough of low extends up to ne part of bangladesh from west bengal and adjoining area is shown in figures 3(e-l). the convergence zone lies along the trough. the (d) (e) (f) (g) (h) (i) (j) (k) (l) (b) (a) (c) 64 md. omar faruq et al. predictability of orographic rainfall features ……. model has also simulated the high-pressure area over meghalaya and a pressure gradient force in the southeastern part of bangladesh which is consistent to the observation. so, this high probability of moisture incursion towards ne part of bangladesh from the bob is the source of energy to accelerate the buoyancy processes in wind side of the hills over sylhet and neighborhood and it is the supportive condition for the formation of orographic clouds and afterwards convective rainfall over sylhet and adjoining area. 3.2 wind at 10m height, 850 and 500 hpa levels the model simulated wind flow at 10-meter height on 0000, 0600, 1200 and 1800 utc of 29 march based on 0000 utc of 27 and 28 march is compared with bmd’s observed wind flow at 10-meter height on 0000, 0600, 1200 and 1800 utc of 29 march 2017 which is shown in figures 4(a-l). the model simulated wind flow at 850 and 500 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 shown in figures 5(a-h) and figures 6(a-h), respectively. figure 4: observed wind flow at 10-meter height on 0000, 0600, 1200 & 1800 utc of (a-d) 29 march 2017; predicted wind flow at 10-meter height on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (e-h) 27 march and (i-l) 28 march 2017, respectively. from the 10-meter height wind analysis, it is found that south-southwesterly wind is blowing from the bob towards sylhet through central part of bangladesh and changes its direction and started to blow from east to west direction which is matched to the observed easterly wind flow of sylhet and same scenario is found at 850 hpa wind flow. this wind brings high amount of moisture towards bangladesh and due to the orographic effect in the ne part of bangladesh, this high moisture uplifted and enhanced shallow or deep convection. this uplifted wind mixes with the model predicted wind at 500 hpa level blowing from the west/northwest and it is very dry and (e) (f) (g) (h) (i) (j) (k) (l) (b) (a) (d) (c) journal of engineering science 11(2), 2020, 61-73 65 cold. owing to the mixing of this cold wind and the uplifting moist wind, cloud formation occurs over sylhet and adjoining area and is responsible for orographic rainfall. figure 5: predicted wind flow at 850 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. figure 6: predicted wind flow at 500 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. 3.3 relative vorticity at 850 and 500 hpa levels the model simulated relative vorticity at 850 and 500 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 is displayed in figures 7(a-h) and figures 8(a-h) respectively. vorticity is the first critical factor in the determination of thunderstorm type and potential storm severity. positive vorticity is related to updraft and negative vorticity is correlated to downdraft. (e) (f) (g) (h) (b) (a) (d) (c) (e) (f) (g) (h) (b) (a) (d) (c) 66 md. omar faruq et al. predictability of orographic rainfall features ……. the model simulated vorticity at 850 hpa level over sylhet and adjoining area (marked by the circle) is found positive of magnitude (6-10)×10-5 s-1 and negative of magnitude (8-10) ×10-5 s-1 on 0000, 0600, 1200 and 1800 utc of 29 march 2017, respectively. this positive and negative vorticity is supportive for occurring of deep convective clouds and afterwards heavy rainfall. on the other hand, the vorticity at 500 hpa level over sylhet and adjoining area is dominated by positive vorticity which is the indication of priority of further updrafts and supportive for deep convection. (s-1) figure 7: predicted relative vorticity at 850 hpa on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. (s-1) figure 8: predicted relative vorticity at 500 hpa on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. ×10-5 (e) (f) (g) (h) (b) (a) (d) (c) ×10-5 (e) (f) (g) (h) (b) (a) (d) (c) journal of engineering science 11(2), 2020, 61-73 67 3.4 vertical wind shear vertical wind shear is the second critical factor in the determination of thunderstorm type and potential storm severity. wind shear plays a fundamental role in determining the internal dynamics of a thunderstorm, along with the organization of a group of thunderstorms. both speed and directional wind shear are important when considering thunderstorms (atmos.uiuc.edu, 2010). the model regenerated wind shear between 500 and 850 hpa level (u500 – u850) on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 is depicted in figures 9(ah). higher positive value of vertical wind shear is related to thunderstorm intensification and lower wind shear is supportive for tropical cyclone strengthening. the model simulated wind shear of the order of (0-20) knots is related to updraft and lower value of vertical wind shear of the order of (0-5) knots governs downdraft over sylhet and adjoining area (marked by the circle). (knots) figure 9: predicted vertical wind shear between 500 and 850 hpa on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. 3.5 relative humidity at 2m height and its vertical cross-section the model simulated rh at 2m height on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 is compared with observed rh at 2m height on 0000, 0600, 1200 and 1800 utc of 29 march 2017 shown in figures 10(a-l). the model generated rh is about 75-95% over sylhet and adjoining area which is consistent to the observation. the abundant moisture percentage is responsible for buoyancy of air and finally cloud formation. it is also mentionable that rh is more than 95% at the wind side of hilly region of sylhet and adjoining area which trigger precipitation processes over those regions. the heavy rainfall is shown in figures 16(a-e) occurred at the right side of the dry line drawn in the picture (border of dry and cold air with moist and warm air) and the prerequisite is moisture abundance. the model simulated vertical cross-section of rh on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 for latitude 24.9on is shown in figures 11(a-h) and for longitude 91.9oe is shown in figures 12(a-h) respectively. the model simulated latitudinal cross-section of rh along 24.9on indicates that 60-80% of moisture is extended up to 600 hpa level along 87-90 oe and 60-100% of moisture is extended up to 400 hpa level along 91-93 oe, whereas the longitudinal cross-section of rh along 91.9oe indicates that 60-80% of moisture is extended up to 400 hpa level. (e) (f) (g) (h) (b) (a) (d) (c) 68 md. omar faruq et al. predictability of orographic rainfall features ……. (%) figure 10: observed rh at 2m height on 0000, 0600, 1200 & 1800 utc of (a-d) 29 march 2017; predicted rh at 2m height on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (e-h) 27 march and (i-l) 28 march 2017, respectively. (%) figure 11: predicted latitudinal (24.9on) cross-section at sylhet of rh on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. (e) (f) (g) (h) (i) (j) (k) (l) (b) (a) (d) (c) (h) dry line dry line dry line dry line dry line (e) (f) (g) (h) (a) (b) (d) (c) journal of engineering science 11(2), 2020, 61-73 69 % figure 12: predicted longitudinal (91.9oe) cross-section of rh on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. 3.6 cape at 850 hpa level the model simulated cape at 850 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of 27 and 28 march 2017 is shown in figures 13(a-h). the model simulated cape is found 400 – 800 jkg-1 over sylhet and adjoining area at 0000 and 0600 utc at developing stage of the system. afterwards the cape is increasing to 1000 jkg-1 or more over the study region. the value of cape is 1000 jkg1 or more is liable for moderate unstable [uk ag weather center] condition of the atmosphere which enhanced heavy to very heavy rainfall. the amount of 24h rainfall over sylhet was recorded, 119 mm (bmd) which is consistent to the model simulated cape value. (jkg-1) figure 13: predicted cape at 850 hpa on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively (e) (f) (g) (h) (b) (a) (d) (c) (a) (e) (b) (d) (c) (f) (g) (h) 70 md. omar faruq et al. predictability of orographic rainfall features ……. (jkg-1) figure 14: predicted cin at 850 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on the initial condition of 0000 utc of (a-d) 27 march and (e-h) 28 march 2017, respectively. 3.7 cin at 850 hpa level the model simulated cin at 850 hpa level on 0000, 0600, 1200 and 1800 utc of 29 march 2017 based on 0000 utc of 27 and 28 march 2017 is shown in figures 14(a-h). the model simulated cin is found 0–80 jkg-1 over sylhet and adjoining area from 0000 to 1800 utc at developing and dissipating stage of the system. the value of cin is supportive for potentially unstable (jeff haby) atmospheric condition and it is the favorable condition for the convection and afterwards occurring of very heavy rainfall over sylhet and adjoining area. 3.8 different indices of thermodynamic diagram the rawinsonde observation over sylhet was not available, so, the nearby sylhet station is chosen for thermodynamic diagram and validation for different indices. the model simulated thermodynamic parameters as well as thermodynamic diagram at sylhet, bangladesh on 0000 utc of 29 march 2017 based on 0000 utc of 27 and 28 march 2017 is compared with bmd’s thermodynamic parameters and diagram on 0000 utc of 29 march 2017 which is shown in figures 15(a-c). the model predicted sounding diagram of t and td lines are very close to the observation and it is the indication of unstable atmosphere. different thermodynamic indices are described as follows: the model simulated li is found negative (-1.2⁰c for 48h and -2˚c for 72h model run) and the observed li is found also negative (-1.52˚c) which is the marginal unstable to large instability of the atmospheric condition (uk ag weather center). the model predicted k-index is found 43⁰c for 48h and 45˚c for 72h model run and the observed k-index is found 32.6˚c which is the moderate convective potential of the atmospheric condition (jeff haby). this thermodynamic index is predicted by the model good enough and it is consistent to the observation. so that the wrf model can be used for predicting this parameter even 72 h before of the event occurrence. the model generated tt is found 44˚c for both 48h and 72h model run and the observed tt is found 39.8˚c which is supported for likely thunderstorm formation (jeff haby). the model predicted cape at sylhet is found 1522 jkg-1 for 48h and 1632 jkg-1 for 72h model run and the observed cape is found 1521.65 jkg-1 which is liable for marginal to moderate unstable [uk ag weather center] condition of the atmosphere. the model predicted cin at sylhet is found 103 jkg-1 for 48h and 64 jkg-1 for 72h model run and the observed cin is found 56 jkg-1 at 0000 utc which is supportive for moderate to potentially unstable (uk agriculture weather center) condition of the atmosphere. these thermodynamic indices indicate to favorable condition for occurring heavy to very heavy rainfall. (e) (f) (g) (h) (b) (a) (d) (c) journal of engineering science 11(2), 2020, 61-73 71 ki = 32.6˚c tt = 44.8˚c li = -1.44˚c cape = 1521.65 jkg-1 cin = 56 jkg-1 ki = 31˚c; tt = 43˚c; li = -1.20c cape = 1522jkg-1; cin = 103jkg-1 ki = 35˚c; tt index = 45˚c; li = -2˚c cape = 1632jkg-1; cin = 64jkg-1 figure 15: observed thermodynamic diagram and thermodynamic parameters at sylhet, bangladesh on 0000 utc of (a) 29 march 2017; predicted thermodynamic diagram and thermodynamic parameters on 0000 utc of 29 march 2017 of 0000 utc of (c) 27 march and (b) 28 march 2017, respectively. (mm) figure 16: predicted 24 hour accumulated rainfall of 29 march 2017 based on 0000 utc of (a) 27 march 2017 (b) 28 march 2017; (c) observed, (d) trmm and (e) ecmwf 24 hour accumulated rainfall of 29 march 2017 respectively. (e) (d) (a) (b) (c) (a) (b) (c) 72 md. omar faruq et al. predictability of orographic rainfall features ……. 3.9 rainfall analysis the model predicted accumulated 24-hour rainfall of 29 march 2017 based on 0000 utc of 27 and 28 march 2017 is compared with observed, trmm and ecmwf accumulated 24-hour rainfall of 29 march 2017 which is shown in figures 16(a-e). the signature of the spatial distribution of wrf model is well-matched to the observed rainfall than that of trmm and ecmwf. in both cases very heavy rainfall is predicted by the wrf model equitably well over the wind side of orographic region, sylhet. the wrf model simulated rainfall is overemphasized by trmm and underestimated by ecmwf. the model predicted 24-hour rainfall of 29 march 2017 based on 0000 utc of 27 and 28 march 2017 is compared with 24-hour observed, trmm and ecmwf rainfall of 29 march 2017 at sylhet which is shown in figures 17(ae). the computational analysis represents that the model simulated 24-hour rainfall by 48-hour advanced run is closer than that of trmm, ecmwf and 72-hour prediction. the model performance is good enough and gives the better results with the minimization of the lead time to predict the rainfall over sylhet and adjoining area. figure 17: predicted 24-hour rainfall of 29 march 2017 at sylhet based on the initial condition of 0000 utc of (a) 27 march 2017 and (b) 28 march 2017, (c) observed, (d) trmm and (e) ecmwf rainfall of 29 march 2017 respectively at sylhet, bangladesh. 4. conclusions on the basis of the present study, the following conclusions can be drawn: i. the nssl 2-moment microphysics coupling to the kain–fritsch cumulus scheme and ysu pbl scheme options of wrf model produces representative results in both spatial and quantitative evaluations. therefore, these schemes have been considered as the prediction of thunderstorm which passes over ne part of bangladesh. ii. the model predicted lowest mslp of the thunderstorm is about 1006-1008 hpa for 48-h, and 72-h model run. the model captured the south-westerly wind flow at 10-m height and 850 hpa level which transports of moisture from the enormous area of the bob towards the ne part of bangladesh and neighborhood, this south-westerly wind changes to easterly by the confrontation of hills. one of its components is uplifted and conjugate with the west-northwesterly dry and cold wind at 500 hpa which is very close to the observation. iii. the model predicted vorticity over ne part of bangladesh at 850 hpa level is positive of magnitude (0610)×10-5 s-1 and negative of magnitude (6-10)×10-5 s-1 for 48-h and 72-h model run which directs updraft and downdraft. at 500 hpa level, positive vorticity is dominant, is the indication of further updrafts of the system. iv. the model predicted positive vertical wind shear of the order of (0-20) knots is governing and it is the indication of further updraft of the system. v. the rh is found 75-100% over sylhet and adjoining area which is very close to the observation and 60100% moisture is extended up to 600-350 hpa levels. vi. the model simulated cape is found 400 – 800 jkg-1 at developing stage and 800-1000 jkg-1 or more at mature stage over sylhet and adjoining area which is moderately liable for unstable condition of the atmosphere. the model simulated li is -2 to -1.2˚c, k-index lies between 31 to 35˚c and tt is found 43 to 45˚c which are close to the observation and responsible for the marginal unstable to large instability of the atmospheric condition. vii. the model captured the rainfall event reasonably well enough though some spatial and computational error exits. the bias correction may increase the competency of the model for prediction of orographic rainfall and may be forecasted more precisely and accurately. 0 20 40 60 80 100 120 140 (a) (b) (c) (d) (e) r ai nf al l ( m m ) journal of engineering science 11(2), 2020, 61-73 73 acknowledgement the authors are thankful to ncar, ncep, usgs and bmd for providing model source code and support, topography and land use, initial and lateral boundary conditions and rain gauge observed data. we are grateful to director of bmd for his constant support, inspiration and encouragement throughout the research work. references browning, k. a., hill f. f., and pardoe c. w., 1974. structure and mechanism of precipitation and the effect of orography in a wintertime warm sector, q. j. r. meteorol. soc., 100, 309–30. davis, c., wang w., chens. s., chen y., corbosiero k., demaria m., dudhia j., holland g., klemp j., reeves j. m. h., rotunno r., snyder c., xiao q., 2008. prediction of landfalling hurricanes with the advanced hurricane wrf model, mon. wea. rev., 136, 1990–2005. dodla, v. b. r., and ratna s. b., 2010. mesoscale characteristics and prediction of an unusual extreme heavy precipitation event over india using a high resolution mesoscale model, atmos. res., 95, 255–269. faruq, m. o., chowdhury m. a. m., akhter m. a. e., mallik m. a. k., hassan s. m. q., and huque s. m. m., 2019. simulation of a heavy rainfall event of 17 may, 2016 and its thermodynamic features over sylhet, bangladesh using wrf model, the atmosphere, 8(1) 81–90. haby, j., http://www.theweatherprediction.com/habyhints. niyogi, d., holt t., zhong s., pyle p. c., and basara j., 2006. urban and land surface effects on the 30 july 2003 mesoscale convective system event observed in the southern great plains, j. geophys. res., 111, d19107, doi: 10.1029/2005jd006746. osuri, k. k., mohanty u. c., routray a., makarand a. k., and mohapatra m., 2012. sensitivity of physical parameterization schemes of wrf model for the simulation of indian seas tropical cyclones, nat. hazards, 63, 1337–1359. pattanaik, d. r., and rao y. v. r., 2009. track prediction of very severe cyclone ‘nargis’ using high resolution weather research forecasting (wrf) model, j. earth syst. sci., 118, 309–329. routray, a., mohanty u. c., rizvi s. r. h., niyogi d., osuri k. k., and pradhan d., 2010. impact of doppler weather radar data on simulation of indian monsoon depressions, quart. j. roy. meteor. soc., 136, 1836– 1850. skamarock, w.c., klemp j.b., dudhia j., gill d.o., barker d.m., duda m.g., huang x.y., wang w., powers w.g., 2008. a description of the advanced research wrf, version 3. ncar technical note. boulder. smith, r. b., 2006. progress on the theory of orographic precipitation, special paper 398, tectonics, climate, and landscape evolution, s. d. willett et al., eds., geological society of america, 1–16. uk ag weather center cape: http://weather.uky.edu/cape.html. ww2010.atmos.uiuc.edu/(gh)/guides/mtr/svr/comp/wind/home.rxml © 2020 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. microsoft word 10_jes_293_12_03_2020 journal of engineering science 11(2), 2020, 107-115 doi: https://doi.org/10.3329/jes.v11i2.50902 effect of elevated temperature on residual strength of self compacted concrete snahashish paul1, muhammad harunur rashid 1, and md. anisur rahman2 1department of civil engineering, khulna university of engineering & technology, khulna 9203 2deputy project director, dhaka khulna (n-8) project, bangladesh received: 20 september 2020 accepted: 03 december 2020 abstract self compacted concrete (scc) is a material used in the construction industry to ensure proper compaction of concrete without providing any external energy. in case of exposure of scc to accidental fire, an assessment of its residual capacity is needed. this study covers the observation of residual compressive strength, tensile strength and modulus of elasticity of self compacted concrete under elevated temperatures (150, 300, 450, 600 and 800⁰c) and cooling conditions (air cooling and water quenching). the compressive strength increased at 150⁰c and decreased continuously after this temperature. however, tensile strength and modulus of elasticity decreased at elevated temperatures compared with ambient temperature. the compressive strength at ambient temperature (30⁰c) was 27.0 mpa, and it raised to 28.7 mpa at 150⁰c for air cooling and 27.8 mpa for water quenching. keywords: cooling condition; elevated temperature; mechanical properties; rheological properties; self compacted concrete. 1. introduction self-compacted concrete is characterized as a high-performance concrete that has excellent fresh-state deformability and high segregation resistance and can be put and compacted under its own weight without any external energy being applied. it was first introduced in japan in 1988 and was named as “high-performance concrete”, and further development was carried out by different researchers (bartos & grauers, 1999; okamura & ouchi, 2003). self-compacted concrete is a better option than normal concrete in areas where compaction and placement of concrete are difficult (mathew & paul, 2012). scc occupies the shape of a formwork in plastic state and when hardened it creates a dense and homogenous concrete with better engineering properties and is more durable than traditional vibrated concrete (lenka & panda, 2017; salhi et al., 2017). however, while the use of scc is technologically, socially and overall beneficial, its production may cost between 2-3 times higher than normal concrete. hence for the reduction of the cost of scc, mineral admixtures could be used to increase the workability of the concrete mix (heikal et al., 2013; pathak & siddique, 2012). the composition of scc can fluctuate between plants and countries alike, even if the fresh and hardened properties are similar, in terms of flowing and passing capability. concrete interfaces zone properties probably improved due to the presence of higher amount of fine and extra-fine particles (ding et al., 2010; 2011). scc has the ability to flow without any induced compaction through any congested sections between the reinforcement and fills entirely within the formworks without leaving any voids (badogiannis et al., 2015; calado et al., 2015; kanadasan et al., 2015; rama seshu & pratusha, 2013). now-a-days fire accident has become one of the most frequently occurring hazards all over the worlds. several fire accidents due to the various reasons and with different magnitudes are frequently happening. so, there is a growing concern about the structural safety before construction of an infrastructure and reconstruction after a building has experienced a fire hazard. in cases of an accidental explosion, concrete properties change after burning. hence, the change in concrete properties due to extreme temperature exposures is necessary to understand (vasusmitha & rao, 2012). a large number of studies were performed on normal strength concrete and high strength concrete on residual strength after facing high temperature (anand et al., 2014; castillo, 1990; phan & carino, 1998; rama seshu & pratusha, 2013). spalling and reduction in strength due to high temperature lead the concrete to failure. the type of powder used in the production of scc will significantly impact the spalling and strength characteristics (bakhtiyari et al., 2011). the study on the properties of scc exposed to elevated temperatures is a growing concern now-a-days (noumowé et al., 2006). some investigations were performed on residual compressive strength of scc exposed to high temperature (annerel et al., 2007; fares et al., 2009, 2010; heikal et al., 2013; persson, 2004; tufail et al., 2017). the strength of concrete decreased while heated to elevated temperatures of 650c and thereafter (demirel & keleştemur, 2010; tufail et al., 2017). it was observed that the strength loss between 100-200c (chang et al., 2006), 100-400c is slightly lower compared to the strength loss after these temperatures (anand & arulraj, 2014; tao et al., 2010). some researchers investigated on the performance of elevated temperature of concrete and found that concrete temperature increases up-to 200c (heikal, 2000), 300c (fares et al., 2009; rama seshu & pratusha, 2013) *corresponding author: snahashish@ce.kuet.ac.bd https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 108 snahashish paul et al. effect of elevated temperature on residual …… and after these temperatures, concrete strength started to decrease. however, this variation of strength loss is different for different grade of concrete. high strength concrete is more sensitive to temperature than normal strength concrete. higher grade of concrete shows a significant reduction in strength and stiffness than normal strength concrete (anand & arulraj, 2014; tao et al., 2010). in case of normal strength concrete, scc is more sensitive to temperature than normally vibrated concrete (rama seshu & pratusha, 2013). some experimental research on the performance of scc was carried out using different mineral additives and elevated to high temperature of up to 600°c (rukavina et al., 2015). it was found that mineral additives have great influence on the residual compressive strength between ambient temperature and 400°c. at the time of exposure of reinforced concrete structural elements to high temperature due to good fire resistance properties of concrete, it is often possible for the structure to withstand. but at the same time, the chemical composition, physical structure and water content of concrete are modified. cooling conditions play an important role in these modifications and greatly produce an impact on the strength characteristics of concrete. so, it is necessary to observe the mechanical properties of concrete after a fire accident when extinguishing fire by water or goes out slowly (rama seshu & pratusha, 2013). in this study, investigations on the residual capacity of concrete such as residual compression capacity, residual tension capacity modulus of elasticity will be done due to exposure of self-compacting concrete to elevated temperature and two different cooling conditions for evaluating the structural performance. 2. materials and methods 2.1 materials in this experiment, for preparing self compacted concrete locally available materials were used. ordinary portland cement (cem-i) was used as a binding agent with coarse sand and crushed stone. to fill the void of the concrete stone dust was used as a filler material and for getting higher slump value with moderate w/c ratio, superplasticizer was used. the properties of the materials and the gradation curve for coarse aggregate used in this research are represented in table 1 & figure 1, respectively. table 1: properties of materials used to prepare scc materials properties unit value test standard binder (portland cement) specific gravity 3.15 astm c128 normal consistency % 26.5 astm c187 initial setting time minutes 155 astm c191 final setting time minutes 275 astm c191 fine aggregate specific gravity (ssd) 2.55 astm c128 absorption % 4.55 astm c128 fineness modulus 2.70 astm c136 unit weight kg/m3 1573 astm c29 filler (stone dust) specific gravity 2.78 astm c128 coarse aggregate (maximum size of aggregate = ½ inch) specific gravity (ssd) 2.78 astm c127 absorption % 1.93 astm c127 void % 42.7 astm c29 unit weight kg/m3 1587 astm c29 figure 1: gradation curve for coarse aggregate journal of engineering science 11(2), 2020, 107-115 109 2.2 mix design for self compacted concrete the first aim of this mix design is to find out the optimum proportions of aggregate that can produce a highly compacted concrete when filling a casing by the self-weight of that concrete. the calculation was done by using the physical properties of the aggregate used to achieve theoretically almost zero void condition. theoretical zero void condition was achieved by determining the voids between the aggregates and filling that voids by cement paste as well as filler paste. to ensure proper flowing ability, passing ability and resistance to segregation ability of the concrete several trials were done and a final mix proportion was obtained that satisfy all the requirements of self-compacted concrete. the trials were done by adjusting the water cement ratio and the dose of superplasticizer. table 2 represents the final mix proportions of self-compacted concrete. table 2: final mix proportions of scc materials amount cement (cem-i) 485 kg/m3 coarse aggregate (stone chips) 713.35 kg/m3 fine aggregate (coarse sand) 865.94 kg/m3 water 237 kg/m3 filler material (stone dust) 17.85 kg/m3 super plasticizer (con-lub) 12 ml/kg of cement water cement ratio, w/c 0.474 water filer ratio, w/f 0.40 2.3 test on fresh properties of self-compacted concrete the essential properties on which concrete is claimed as self-compacted concrete (scc) can be found by some tests on the fresh state of that concrete. slump flow test was done to ensure the flowing ability, v funnel flow tests at 10s, and 5 min (t5min) was done to check the flowing ability as well as resistance to segregation ability and l box test was conducted to assure the passing ability of concrete through congested openings like beamcolumn joint. all properties were determined in 30 minutes following mixing in order to maintain their fresh quality and to reduce the impact of workability loss on the test results. figure 2: tests on fresh properties of scc (a) slump flow test (b) v funnel test (c) l box test (d) standard dimension for l box 2.3.1 slump flow test the slump flow test for scc is similar to the conventional slump value test established for normal concrete, but the only difference is, this concrete does not require any external energy to fill the slump cone. according to efnarc, concrete with a slump flow range of 600 mm to 800 mm confirms self compaction ability. a non(a) (b) (c) (d) 110 snahashish paul et al. effect of elevated temperature on residual …… absorbent board marked with a 700 mm diameter circle was used (figure 2a). the slump cone was placed at the centre of this circle and filled with concrete without any temping. the top surface of the slump cone was levelled by using a trowel. the cone was removed vertically, and the concrete was allowed to spread. the final diameter of the spread concrete was recorded. 2.3.2 v funnel flow test in the v funnel flow test, the time required to flow through the v funnel indicates the viscosity of concrete which is a measure of segregation resistance ability.approximately 13 liters of concrete were prepared to perform this test. a v funnel with standard dimension (figure 2b) is placed vertically on flat ground and the inner surface of the funnel was moistened. a bucket was placed beneath the funnel and the trap door was closed. the funnel was completely poured with concrete in such a way that there is no application of external energy. the top of the funnel was smoothed by a trowel and the door was opened within 10 seconds. the time required to complete the discharge of concrete through funnel was recorded as the flow value for 10 seconds retention (t10s). to determine the t5min flow value, the funnel was filled again and waited for 5 minutes before opening the door. the time required to pass the concrete through the outlet of v funnel after 5 minutes retention period in the funnel was recorded. 2.3.3 l box test l shaped mold was used in the l box test (figure 2c), which consists of a vertical section and a horizontal section having a movable door in their connection point (figure 2d). in the joining point of these two sections, there was some fixed reinforcing bar to create an obstruction in passing. the movable door was opened after filling the vertical section and waited for the concrete to pass through the obstacles to reach the other end of the horizontal section. the height of concrete at both ends of the horizontal section was measured as h1 & h2. 2.4 sample preparation and curing total 84 cylindrical specimens with a dimension of 100 mm diameter and 200 mm height were cast with a mix ratio obtained from the trial used to fulfil the requirements of fresh state properties of self-compacted concrete. proper curing was confirmed by keeping the specimen into a curing chamber for the ages of 3 days, 7 days, 28 days, and 90 days. 2.5 preheating, heating and cooling conditions a preliminary drying process was carried out for at least two hours at a temperature of 1052˚c in an electric oven before heating the samples to elevated temperature in the furnace. this preheating was done to reduce the internal moisture content of the specimen as loss of free water from concrete generally occurs at about 100˚c. figure 3: astm e119 standard time temperature curve vs heating rate in the research figure 4: electric furnace with controlling system a laboratory furnace with an internal dimension of 240 mm length, 240 mm width, and 275 mm depth is used for heating the specimens in different elevated temperatures (150, 300, 450, 600 and 800˚c). it has a control unit on its left side, as shown in figure 4. the temperature of the furnace can be raised about 1200˚c. it has a cover on its top which has a diameter of 178 mm. the heating process was performed according to the rate of heating of the astm e119 standard time-temperature curve. this curve represents the rate of change in temperature during a fire hazard. in this research up to 700-degree celsius, the heating rate has followed the astm e119 time-temperature curve for fire and a slight deviation from the fire curve was observed during an elevated journal of engineering science 11(2), 2020, 107-115 111 temperature of 800˚c. the samples were heated for one hour at the desired maximum elevated temperature and cooled by two different cooling conditions. one is cooling naturally in the air for 24 hours, and another is force cooling into the water for 1 hour and kept in air for 23 hours. 2.6 tests on mechanical properties of scc tests on mechanical properties were performed after 24 hours from starting of cooling. the compressive strength, splitting tensile strength and elastic modulus of different elevated temperature and ambient temperature was determined using a digital compression testing machine. 3. results 3.1 properties of self-compacted concrete the results obtained from several trials on fresh state characteristics of produced concrete by slump flow, v shaped funnel flow and l box passing test are shown in figure 5. the selected trial mix has a slump flow of about 657 mm, which represents a good flowing capacity of concrete by its own weight through any congested section by eliminating frictional resistance. table 3 represents the fresh state rheological properties of selected trial mix proportions. figure 5: rheological properties of scc for different trial mix proportion the time required to pass through the outlet of the v-shaped funnel after 10 seconds of placement of concrete gives a value of 11.93 seconds indicates moderate viscosity. european federation of national associations representing for concrete (efnarc) recommended a range of about 8 to 12 seconds for possessing good flowing property where the limiting value of 8 seconds represents a concrete of very low viscosity. the variation between t5min flow time and t10s flow time is 2.92 seconds confirmed the segregation resistivity of the concrete. table 3: fresh state rheological properties of selected trial mix proportion test name parameter result slump flow test slump value (mm) 657 v shaped funnel flow test t10s flow time (sec) 11.93 t5min flow time (sec) 14.85 l box passing test (h2/h1) 0.82 3.2 compressive strength the compressive strength development of self-compacted concrete for different age of curing at air temperature is shown in figure 6a. the maximum compressive strength attained after 90 days of curing was 27.0 mpa at room temperature of 30˚c. figure 6b represents the compressive strength at different elevated temperatures. while heated to about 150˚c has some increase in compressive strength and a continuous decrease in strength was observed from 300˚c and thereafter heating. above 110c, the chemically bound water from calcium silicate hydrate (c-s-h) was started to release, and the intermolecular stress was increased due to thermal expansion of aggregate (bingöl & gül, 2009; seleem et al., 2011). when subjected to elevated temperature the strength of concrete structure started to decrease from 300c due to the evaporation of chemically bound water and dehydration of ca(oh)2 into free lime leading to the formation of micro-cracks inside the concrete (demirel & keleştemur, 2010; tufail et al., 2017). the strength of concrete is greatly affected due to the expansion of lime during the cooling period (demirel & keleştemur, 2010). the strength may be increased if 112 snahashish paul et al. effect of elevated temperature on residual …… the adverse effect of lime can be minimized by adding some mineral additives. mineral containing silicon dioxide from stone dust reacts with calcium hydroxide from cement to decrease the amount of lime from the system. as a result, an increment of strength was observed at 150c. this increment of strength also occurs when the free water from a concrete body was released, enhancing the frictional resistance between failure plane. elevated temperature up to 300c also enhance the hydration process of anhydrous cementitious component resulting in the proper bonding between aggregates and hence increase the strength (fares et al., 2009; rama seshu & pratusha, 2013; seleem et al., 2011). figure 6: (a) compressive strength against different age of curing (b) compressive strength against different elevated temperatures and cooling conditions in this test, it was observed that the concrete strength for water quenching & cooling is slightly lower than that for air cooling after heated to elevated temperatures. after a fire accident, when extinguished by sudden quenching in water produce thermal shock due to the sudden drop of temperature within a few minutes (botte & caspeele, 2017). as a result, natural cooling in air shows better result than sudden quenching in water after fire exposure. after 150c, in both cooling conditions, the strength of the concrete was decreased because of the expansion of lime that affects the volume occupied by other cementitious components (heikal, 2000; seleem et al., 2011). the concrete strength beyond 400c, decreased rapidly due to the dehydration of c-s-h (heikal, 2000; tufail et al., 2017). effect of high temperature and cooling conditions on the residual strength of scc can be observed from figure 7. figure 7: comparison of residual compressive strength for different elevated temperatures and cooling conditions the residual compressive strength was determined as a percentage of the strength of reference specimens (29⁰c) which was available after considering the heating and cooling process. the minimum compressive strength was observed at 800⁰c. the available compressive strength of scc after heated to 800˚c and cooled at air is 57.1 %, and when cooled at the water for 1 hour, the residual strength is 54.1% compared to 30⁰c. 3.3 tensile strength the splitting tensile strength values from the tests on the specimens at ambient temperature and after heated to high temperature are represented in figure 8. the tensile strength of concrete significantly decreased after (b) (a) journal of engineering science 11(2), 2020, 107-115 113 elevated to high temperatures. the maximum tensile strength was noticed at natural temperature, and a gradual reduction was found on heating. the effect of cooling conditions was found similar to that of compressive strength. suddenly cooling in water for one hour shows a significant reduction in tensile strength compared to cooling in the air for 24 hours. the residual tensile strength of scc is shown in figure 9. the residual tensile strength was above 92 % for cooling in the air up to 300⁰c, whereas sudden cooling in water shows an additional 6-7% drop in tensile strength. the tensile strength of self-compacted concrete was dropped to about 52% of its original capacity when heated to 800⁰c and cooled in water for one hour. for the same temperature, cooling through the air for 24 hours gives 15-16% more strength. figure 8: tensile strength for different elevated temperatures and cooling conditions figure 9: comparison of residual tensile strength for different elevated temperatures and cooling conditions 3.4 modulus of elasticity the modulus of elasticity of the specimens were determined according to astm c469. tests on several specimens at elevated temperatures and cooling conditions were done to find out the elastic modulus of self compacted concrete and compared it to the values obtained for the specimens at room temperature to observe the change in elastic modulus at higher temperatures. the change in elasticity with temperature obtained from the investigation is reported in figure 10. the elastic modulus of self compacted concrete decreased at raised temperatures for both cooling conditions. compared to sudden water cooling after heating, air cooling shows better results. the modulus of elasticity dropped from 23310 mpa to 7055 mpa with a maximum reduction of almost 70 % (figure 11) of reference stiffness of scc (30⁰c) when heated to 800⁰c and apply sudden cooling for 1 hour. about 30-40% drop in elasticity was noticed for the elevated temperature of within 300⁰c, and after this temperature, a gradual decrease in elasticity was observed. the residual capacity of the stiffness for self-compacting concrete was dropped to 50% of its original capacity at the temperature ranges from 450-600⁰c (figure 12). figure 10: elastic modulus for different elevated temperatures and cooling conditions figure 11: reduction in elastic modulus for different elevated temperatures and cooling conditions 114 snahashish paul et al. effect of elevated temperature on residual …… figure 12: comparison of residual elastic modulus for different elevated temperatures and cooling conditions 4. conclusions self-compacting concrete is a special type of concrete that can solve the complexities arisen where external compaction can not be applied during the placement of concrete because of its good properties of self compactness. evaluating the residual ability of the self-compacting concrete on its mechanical properties (compressive strength and tensile strength) at various exposures of high temperatures, the following conclusions have been found from experimental evidence.  the compressive strength of scc increased slightly at 150ºc for both cooling conditions and started to decrease continuously on further heating from 300ºc.  the residual strength was dropped significantly at 300ºc by 20.26% for air cooling condition and gradually decreased above 300ºc.  the tensile strength of scc decreased continuously by incremental heating and diminution of 20% residual strength was observed at 450ºc.  the modulus of elasticity for scc made with locally available materials dropped to 50% of its original capacity when heated to a range of 450-600ºc.  cooling at room temperature for 24 hours in the air shows better results than sudden cooling in water for 1 hour. references anand, n., and arulraj g. p., 2014. effect of grade of concrete on the performance of self-compacting concrete beams subjected to elevated temperatures, fire technology, 50(5), 1269–1284. anand, n., arulraj g., and aravindhan c., 2014. stress-strain behaviour of normal compacting and self compacting concrete under elevated temperatures, journal of structural fire engineering, 5(1), 63– 76. annerel, e., taerwe l., and vandevelde p., 2007. assessment of temperature increase and residual strength of scc after fire exposure, 5th international rilem symposium on self-compacting concrete, 715. badogiannis, e. g., sfikas i. p., voukia d. v., trezos k. g., and tsivilis s. g., 2015. durability of metakaolin self-compacting concrete, construction and building materials, 82, 133–141. bakhtiyari, s., allahverdi a., rais-ghasemi m., zarrabi b. a., and parhizkar t., 2011. self-compacting concrete containing different powders at elevated temperatures – mechanical properties and changes in the phase composition of the paste, thermochimica acta, 514(1–2), 74–81. bartos, p. j. m., and grauers m., 1999. self-compacting concrete, concrete, 33(4), 9–13. bingöl, a. f., and gül, r., 2009. effect of elevated temperatures and cooling regimes on normal strength concrete, fire and materials, 33(2), 79–88. botte, w., and caspeele r., 2017. post-cooling properties of concrete exposed to fire, fire safety journal, 92, 142–150. calado, c., camões a., monteiro e., helene p., and barkokébas b., 2015. durability indicators comparison for scc and cc in tropical coastal environments, materials, 8(4), 1459–1481. castillo, c., 1990. effect of transient high temperature on high-strength concrete, aci materials journal, 47-53. chang, y. f., chen y. h., sheu m. s., and yao g. c., 2006. residual stress–strain relationship for concrete after exposure to high temperatures, cement and concrete research, 36(10), 1999–2005. demirel, b., and keleştemur o., 2010. effect of elevated temperature on the mechanical properties of concrete produced with finely ground pumice and silica fume, fire safety journal, 45(6–8), 385–391. journal of engineering science 11(2), 2020, 107-115 115 ding, y., you z., and jalali s, 2011. the composite effect of steel fibres and stirrups on the shear behaviour of beams using self-consolidating concrete, engineering structures, 33(1), 107–117. ding, y., you z., and jalali s., 2010. hybrid fiber influence on strength and toughness of rc beams, composite structures, 92(9), 2083–2089. fares, h., noumowe a., and remond s., 2009. self-consolidating concrete subjected to high temperature, cement and concrete research, 39(12), 1230–1238. fares, h., remond s., noumowe a., and cousture a., 2010. high temperature behaviour of self-consolidating concrete, cement and concrete research, 40(3), 488–496. heikal, m., 2000. effect of temperature on the physico-mechanical and mineralogical properties of homra pozzolanic cement pastes, cement and concrete research, 30(11), 1835–1839. heikal, m., zohdy k. m., and abdelkreem m., 2013. mechanical, microstructure and rheological characteristics of high performance self-compacting cement pastes and concrete containing ground clay bricks, construction and building materials, 38, 101–109. kanadasan, j., fauzi a., razak h., selliah p., subramaniam v., and yusoff s., 2015. feasibility studies of palm oil mill waste aggregates for the construction industry, materials, 8(9), 6508–6530. lenka, s., and panda k. c., 2017. effect of metakaolin on the properties of conventional and self compacting concrete, advances in concrete construction, 5(1), 31–48. mathew, g., and paul m. m., 2012. mix design methodology for laterized self compacting concrete and its behaviour at elevated temperature, construction and building materials, 36, 104–109. noumowé, a., carré h., daoud a., and toutanji h., 2006. high-strength self-compacting concrete exposed to fire test, journal of materials in civil engineering, 18(6), 754–758. okamura, h., and ouchi m., 2003. self-compacting concrete, journal of advanced concrete technology, 1(1), 5–15. pathak, n., and siddique r., 2012. properties of self-compacting-concrete containing fly ash subjected to elevated temperatures, construction and building materials, 30, 274–280. persson, b., 2004. fire resistance of self-compacting concrete, scc, materials and structures, 37(9), 575–584. phan, l. t., and carino n. j., 1998. review of mechanical properties of hsc at elevated temperature, journal of materials in civil engineering, 10(1), 58–65. rama seshu, d., and pratusha a., 2013. study on compressive strength behaviour of normal concrete and selfcompacting concrete subjected to elevated temperatures, magazine of concrete research, 65(7), 415– 421. rukavina, m. j., bjegovic d., and gabrijel i., 2015. mechanical properties of self-compacting concrete with different mineral aditives after high temperature exposure, journal of structural fire engineering, 6(3), 177–184. salhi, m., ghrici m., li a., and bilir t., 2017. effect of curing treatments on the material properties of hardened self-compacting concrete, advances in concrete construction, 5(4), 359–375. seleem, h. e. h., rashad a. m., and elsokary t., 2011. effect of elevated temperature on physico-mechanical properties of blended cement concrete, construction and building materials, 25(2), 1009–1017. tao, j., yuan y., and taerwe l., 2010. compressive strength of self-compacting concrete during hightemperature exposure, journal of materials in civil engineering, 22(10), 1005–1011. tufail, m., shahzada k., gencturk b., and wei j., 2017. effect of elevated temperature on mechanical properties of limestone, quartzite and granite concrete, international journal of concrete structures and materials, 11(1), 17–28. vasusmitha, r., and rao d. p. s., 2012. effect of elevated temperature on mechanical properties of high strength self compacting concrete, international journal of engineering research & technology, 1(8). © 2020 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 02_jes_233_07-04-2020 journal of engineering science 11(1), 2020, 09-17 effect of cement on strength properties of recycled construction and demolition materials a. s. m. riyad* and md. zakir hasan asha department of civil engineering, khulna university of engineering & technology, khulna, bangladesh received: 07 september 2019 accepted: 07 april 2020 abstract evaluation of innovative ways is necessary to use a huge amount of recycled construction and demolition materials generated in bangladesh and require satisfaction considering environmental, economic and engineering perspectives. in that sense, this study evaluates the outcomes when stabilized with locally available ordinary portland cement. furthermore, this study evaluates the research laboratory characterization of physical, geotechnical, and strength properties of unbound construction and demolition materials, for example, reclaimed asphalt from the pavement (rap), recycled crushed concrete (rca) and crushed brick (cb). the results advocated that cement can be a feasible alternative for the stabilization of unbound construction and demolition materials. based on the test outcomes, rap, cb, and rca were found to require 2% of cement to meet the requirements for stabilization with a curing duration of 7 days. keywords: bangladesh; ordinary portland cement; recycled materials; strength. 1. introduction stabilization is referred to a number of differentprocesses which has been adopted to improve the pertinent properties of the unstabilized materials to permit their use in engineering purposes (ali, 1992). better gradation of the particles, decreased plasticity index or swelling potential and improved stability and strength are the most common changes obtained by stabilization (ali et al., 1992; jongpradist et al., 2010; kilic et al., 2015; mohammadinia et al., 2014). the generated solid waste in the universe expected to be about 3.4 billion tonnes in the year 2050 and it is an emerging threat that the generated waste will be reached in the range of more than three times by 2050 at lower-income countries, as general pupils of that regions are not well-concerned about the severity and potentiality of the wastes (kaza et al., 2018). due to tremendous development activity in the world, construction and demolition (c&d) waste materials gained considerable attention among all generated waste in the globe. a large proportion of c&d waste materials have been generated as unwanted materials incidentally or directly from the construction and demolition of new, renovated as well as old structures. the proportion of c&d waste materials is critical in most of the nations that challenged the performance of the development trade as well as its sustainable aims (kulatunga et al., 2005). recently, c&d materials are widely used in several civil engineering operations, for example, backfilling, filter media, embankments of roads, footpath applications, retaining walls, the support structure of pipelines, abutment of the bridge and roads (rahman et al., 2014; 2016). typically, these c&d waste materials include asphalt concrete, metal, wood, asphalt shingles, cardboard, soil, plastic, portland cement concrete, and drywall (ganiron jr, 2015). according to arulrajah et al. (2013), reclaimed asphalt from the pavement (rap), recycled crushed concrete (rca) and crushed brick (cb) are common litters generated among c&d litters in the sphere. the c&d waste materials are about in the range of 25% and sometimes over half of the total municipal solid litters generated in the globe (yeheyis et al., 2013). developed countries produced about 35% of construction wastes, whereas, developing countries like bangladesh, produced about 50% construction wastes of total municipal solid waste generated (bansal and singh, 2014; najafpoor et al., 2014). a study stated that about 2.24 million tons of c&d waste materials such as concrete, plastic, tiles, bricks, steel, aluminum, glass, and timber produced every year in bangladesh. environmental protection agency projected that around 230-530 million tons of c&d waste materials are generated nationwide in the usa in a single year (epa, 2017). the generated wastes mostly dumped into landfills, which create an enormous burden on landfill loading. evidence presents that c&d waste materials contribute approximately 13-60% of all deposited solid waste in landfills in the world (luangcharoenrat et al., 2019). for example, in the uk it is approximately 59% (sharman, 2018), in europe approximately 36% (eurostat, 2019), and 85-90% in bangladesh (islam, 2016; abedin and jahiruddin, 2015), of the total, generated c&d waste.for this, the management of c&d waste materials has gained considerable attention in both economically developed and developing countries. in developing countries, like bangladesh, a significant amount of improvement in waste management is required, including c&d waste management. although these countries are recently experiencing significant development in multiple sectors during the last decades, they are suffering from mismanagement systems of wastes installed in their urban environments. nowadays, the use of the volume of materials increasing for the * corresponding author:riyadtowhid@ce.kuet.ac.bd https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 10 a. s. m. riyad and md. zakir hasan asha effect of cement on strength properties of …… construction of infrastructures of a nation with the pace of the increase of population and industrialization, for example, roads, culverts, bridges, buildings, footpaths, public residents, parking lots in both rural and urban areas. the development in the construction sector is emerging in developing countries, like bangladesh. some mega-developments work, such as padma multipurpose bridge, second meghna-gumti-kanchpur bridge is continuing across the country are enormous as well as many are waiting to start in forthcoming days. from all the available technical resources and research associated with construction and demolition litter management situation in bangladesh, authors have acknowledged that the scenario is very frustrating (chowdhury et al., 2016a). lack of public knowledge, lack of law enforcement, lack of public knowledge and the use of ancient technologies are the key causes of increased c&d waste generation in bangladesh compared to many developed and developing nations in the world (chowdhury et al., 2016b). some literature is available regarding municipal solid litters’ generation and its management, but, in the case of c&d litters, enough statistics are not available to forecast the future scenario. in bangladesh, a large volume of c&d litters is generated in the process of “water line improvement work” in the road and demolition of buildings, especially building the infrastructures of a nation, such as roads, culverts, bridges. the replacement of conventional construction materials with highquality alternative recycled materials, for example, c&d materials are actively being performed by researchers internationally considering environmental (noteworthy savings of carbon), and construction cost perspective (disfani et al., 2014; mohammadinia et al., 2016a; 2016b). furthermore, as discussed earlier, most of these litters are generally disposed to landfills. and so, landfill cost is another considerable concern aside from environmental effects and construction costs to 3r’s (reduce, reuse, and recycling) of c&d materials. for that reason, industries adopted low-carbon material replacement technology to lessen the exhaustion of virgin resources as well as reuse the generated debris from c&d operations (du et al., 2013a; tam and tam, 2006). furthermore, improvement of soil properties for the application in roadway embankments and pavements is one of the emerging issues. chemical stabilization is now extensively adopted for this purpose. different types of additives, such as an emulsion of asphalt, cement, kiln dust of cement, lime, and fly ash for the chemical stabilization of recycled materials has been comprehensively investigated by numerous researchers to improve the efficiency of recycled roadway c&d materials (hoyos et al., 2011; puppala et al., 2011; mohammadinia et al., 2014; latifi et al., 2018). roads and highways department (rhd) of bangladesh summarized that the overall maintenance cost needed for rhd paved road is approximately bdt 26 thousand million in five year period (2018-19 to 2022-23) and insufficient budgetary provision makes it impossible to undertake the maintenance program (rhd, 2018). it is a common belief of engineers of bangladesh, the initial construction cost of rigid pavement is high enough without involving any comparative economic analysis (bhuyan, 2009). although the initial construction cost of rigid pavement is very high, the consideration of some operational and functional benefits as well as the perspective of life cycle cost, the investigation of cement for the stabilization of weak pavement constituents is one of the desired alternatives in modern ages for the construction and refurbishment of roads of urban areas of the major municipalities in the globe (du et al., 1999; 2013b; world highways, 2004). for example, the engineering and geotechnical characterization of australian rap, rca and cb stabilized with general-purpose portland cement has been researched by mohammadinia et al. (2014). based on a personal meeting with rhd personnel in july 2018, rhd estimates that approximately bdt 23 thousand crores can be reserved from 25 years periodical maintenance if the country’s road networks built with cement-treated aggregates. it is an estimation only of rhd roads. if the country's entire road network is taking into consideration, it will be multiple of the mentioned figure. besides, the material cost of cement stabilized roads is almost half of the cost of bituminous roads (bhuyan, 2009). as per the authors' knowledge, generally superior class crushed bricks and stone aggregates treated with cement widely used for cement stabilized roadway bases and sub-bases in bangladesh, and so it is necessary to assess the efficiency of cement stabilized c&d materials as a substitute roadway aggregate material. there is no research work done to determine the properties of c&d waste materials generated in bangladesh using cement as a stabilizer, according to the information available to the authors. however, the diversity of sources of c&d materials, as well as the process of production, may create a detrimental effect on the properties of stiffness, strength, and plastic deformation characteristics of recycled substances (kim et al., 2007; kootstra et al., 2010; al-bared et al., 2018; yaowaratet al., 2018). therefore, laboratory investigation of the locally available c&d waste materials stabilized with cement is focused on this research. 2. methodology in this research, c&d resources encompass of reclaimed asphalt pavement (rap), crushed brick (cb), and recycled concrete aggregate (rca), with a nominal size of the particle of 6.3 mm, are culled from the road under construction of the third-largest city of bangladesh (figure 1). rap is the term given to reused pavement materials containing basalt aggregates and aged asphalt binder. such materials are created when pavements with asphalt are removed for renovation, resurfacing, or for access to buried utilities, conducted regularly. such journal of engineering science 11(1), 2020, 09-17 11 products would end up in landfills without reuse of it by a sustainable process (arulrajah et al., 2013). cb is a by-product of the buildings and other infrastructure construction and demolition operations. cb usually consists of 70 percent brick and 30 percent other non-removed materials, for example, rock, concrete, and asphalt (arulrajah et al., 2011; 2013). rca is created by crushing reclaimed concrete from demolished highways, buildings, bridges, and other structures. the difference between natural aggregate physical properties and rca is highly influenced by the cement paste, which surrounds the aggregate in crushed concrete (langer, 2001). based on the area of application, these concrete chunks are broken into aggregates of varying sizes (arulrajah et al., 2013). impurities include organic materials, gypsum, clay materials, dry mortar paste, and other construction materials that give reduced quality material as compared with the natural aggregate (lemanska, 2019; arulrajah et al., 2013). figure 1: c&d resources used in this study a) rap, b) cb and c) rca laboratory testing program on untreated c&d materials are comprised of grain size analysis with sieve and hydrometer, atterberg limit test, determination of particle density (specific gravity), los angles abrasion, aggregate impact value, flakiness index, water absorption, ph value, atterberg limits, organic content, modified compaction, and unconfined compression strength (ucs) test. the relevant testing standards are followed to perform all the tests as presented in table 1. a minimum amount of 5 kg sample of each category of c&d materials has been sieved using the astm standard sieve [astm d 6913 (astm, 2017a)]. hydrometer analysis has been directed to ascertain the distribution of the size of the particle for the particle sizes smaller than 75-micrometer sieve [astm d 422 (astm, 2007)]. hydrometer analysis has not been performed on rap aggregates as it contained about less than 5% fines content, as suggested by mohammadinia et al. (2014). the static compaction method is selected for the preparation of the samples as it prevented the cracking along with the interfaces of the aggregates layer (mundy, 1991). a split compaction mold having a height to diameter ratio (h/d) of 2, with a collar is used for this purpose. specimens were compacted at a persistent pressure in 8 layers (layers of 1 inch) with the omc obtained from the laboratory compaction curve under modified proctor energy to acquire the target density. ucs tests were conducted using this split mold of the unprocessed and stabilized c&d substances following astm d 5102 (astm, 2009). samples are compacted at static loads to ensure the homogeneity of the mixture, to prevent the damage during the removal, and to maintain the parallel end faces. initially, the water is mixed with aggregates and has been left for about 1-2 hours (based on the type of material) so that the aggregates absorb the water at room temperature just before the compaction. the dry aggregate materials are blended with the relevant moisture content earlier than the addition of additives to keep away from the free water absorption in the mixture in the course of the curing duration. the materials are thoroughly mixed for 2 to 3 minutes to obtain a homogeneous mixture. furthermore, the mixing of the additives and the relevant c&d aggregate are performed before compaction to ensure the adequate hydration process with the available free water. also, a comprehensive laboratory research program is carried out to evaluate the geotechnical and engineering characterization of rap, cb, and rca, when stabilized with ordinary portland cement (opc) type i. the specific gravity (gs) of cement is about 3.15 and loss on ignition (loi) is about 3%. the chemical composition of opc is determined by the x-ray fluorescence method. in this study, 2 and 4% opc are selected for the stabilization of the untreated c&d materials. due to economic considerations, the cement dosage was restricted to a maximum of 4% by dry weight in this research.moreover, the moisture content of the sample was ascertained after ucs tests and instantaneously after compaction of the mixture to check the moisture content of the sample. the stabilized materials were cured in a moist chamber, maintaining the room temperature from 300c to 330c and humidity of 97 to 99%. for cement-treated materials, curing periods of 1, 7, and 28 days were adopted to assess the consequence of curing duration on the development of strength. (a) (b ) (c) 12 a. s. m. riyad and md. zakir hasan asha effect of cement on strength properties of …… 3. results and discussion the engineering properties of the untreated c&d substances are highlighted in table 1. the ph values of the untreated samples point out that the c&d aggregates are alkaline by nature.the observed organic content is highest for rap compared to other c&d materials; this may be due to the presence of carbon-rich bitumen in rap. figure 2 illustrates the grain size analysis plots of the unprocessed c&d substances (before compaction). this figure indicates that the gradation of the cb finest followed by rca and rap.rap, cb, and rca are classified as well-graded sand as per the unified soil classification system (uscs). rca exhibits the highest uniformity coefficient (cu), meaning it is the most well-graded of c&d substances, and cb is more uniform than other c&d materials as it has low cu value.besides, atterberg limit analysis is carried out and the obtained test results indicate that the c&d substances are non-plastic by nature as fine contents are relatively low. particle density (specific gravity) and water absorption measurements were conducted on both coarse (4.75 mm sieve retained) and fine (4.75 mm sieve passed) percentages of c&d substances. from table 1, it is observed that for all the materials tested the particle densities of coarse aggregates are significantly higher than those of the fine aggregates. of the three c&d materials tested, the rca showed the maximum particle density for coarse and fine materials. water absorption of coarse aggregates is smaller than that of fine aggregates for all recycled materials because small particles absorb more water than coarse ones with a larger specific surface. rap reported the lowest water absorption values between the three recycled c&d materials. it is observed that the water absorption values of untreated c&d materials vary from 2.3% to 12.6%, although the value does not exceed 3% for natural aggregate (poon and chan, 2006). table 1: engineering characterization of untreated c&d resources engineering properties testing standards rca cb rap water absorption-coarse (%) astm c 127 (astm, 2015a) 5.2 5.6 2.3 water absorption-fine (%) astm c 128 (astm, 2015b) 12.6 11.3 6.1 specific gravity–coarse astm c 127 (astm, 2015a) 2.56 2.48 2.43 specific gravity–fine astm c 128 (astm, 2015b) 2.54 2.46 2.42 fine content (%) astm d 422-63 (astm, 2007) 4.2 2.9 1.3 sand content (%) astm d 422-63 (astm, 2007) 58.6 65.4 60.2 gravel content (%) astm d 422-63 (astm, 2007) 37.1 31.8 38.5 coefficient of uniformity (cu) astm d 422-63 (astm, 2007) 11.0 10.6 6.0 coefficient of curvature (cc) astm d 422-63 (astm, 2007) 1.3 1.2 1.8 uscs astm d 2487 (astm 2017c) sw sw sw atterberg limit astm d 4318 (astm, 2017b) non-plastic flakiness index bs 812-105.1 (british standards institution, 2000) 14 23 11 aggregate impact value (aiv) bs 812-112 (british standards institution, 1990) 21 17 11 los angeles abrasion loss (%) astm c 131 / c 131m – 14 (astm, 2006) 35 40 25 ph astm d 4972 (astm, 2019) 9.9 10.2 7.7 organic content (%) astm d 2974 (astm, 2014) 1.8 1.0 2.9 maximum dry density (mg/m3) astm d 1557 (astm, 2012) 1.92 1.95 2.02 optimum moisture content (%) astm d 1557 (astm, 2012) 11.52 10.50 5.82 ucs (kpa) astm d 5102 (astm, 2009) 170 160 340 note: sw = well graded sand; ucs = unconfined compressive strength; uscs = unified soil classification system the characteristics of an impact-resistant material are known as toughness. the aggregates are exposed to damage due to the passage of traffic on the road resulting in breaking down into smaller pieces. therefore the aggregates should have enough toughness to withstand their impact-related disintegration. the measure of resistance to sudden shock or impact is aggregate impact value (aiv), which may vary from its resistance to compressive load when applied gradually. of the three c&d materials tested, the rca showed maximum aiv compared to other materials. c&d materials are suitable for construction purposes, as all values are less than 30 (bs 812-112-1990). the presence of flaky particles is considered undesirable for base course and construction of bituminous and cement concrete forms, as these create inherent vulnerability with the possibility of breaking down under heavy loads. of the three c&d materials tested, the cb showed maximum flakiness index value compared to other materials. c&d materials are suitable for construction purposes, as all values are less than 30 (rhd, 2011). the aggregate used in the highway pavements surface course is prone to wear due to traffic movement (hatt, 1939). the los angeles abrasion test (la) is a popular test tool used to demonstrate the aggregate toughness and journal of engineering science 11(1), 2020, 09-17 13 abrasion properties. the findings of the la abrasion test show that rap detects the lowest abrasion and is preceded by rca. except for cb, all c&d materials meet the standard stated value range of 30–35 for traditional quarry substances (rhd, 2011). the compressibility properties of the untreated and treated c&d aggregates are evaluated by the compaction test using a modified effort. the recycled c&d materials are treated with cement. the chemical constituents of this cement are presented in table 2. figure 3 illustrates the grain size analysis plots of the treated c&d substances. figure 4 presented the variation of dry density of materials with moisture content. this figure illustrates that the maximum dry density (mdd) and optimum moisture content (omc) for untreated c&d aggregates and aggregates stabilized with 2 and 4% opc. rap exhibits the highest mdd among the three inspected c&d aggregates, followed by cb and rca. furthermore, with the increase in cement content, the mdd values are increased considerably. the mdd value of untreated cb is slightly higher than the 4% cement-treated sample, which may be due to the variations of the application of modified compaction effort. however, rca presented the highest omc among the three investigated c&d aggregates, followed by cb and rap. the fluctuations of the omc values are negligible with the increase of cement content. the otherness in omc can be due to the variations of water absorption values of the aggregates. figure 2: grain size analysis plots of the untreated recycled samples figure 3: grain size analysis plots of the treated samples figure 4: laboratory compaction curves for unbound and cement treated rap, rca, and cb figure 5: development of ucs in c&d aggregates with curing period table 2: chemical constituents in ordinary portland cement type i chemical composition (%) opc type i sio2 21.62 al2o3 5.24 fe2o3 2.60 sio2 + al2o3 + fe2o3 29.46 cao 63.61 mgo 2.30 na2o3 0.15 k2o 1.03 so3 2.78 the omc value of rca is approximately 9% higher than cb, and the omc value of rap is considerably smaller than the other two resources, which is believed to be as a result of the lower water absorption of rap causing for the presence bitumen coating in the rap materials. the ucs tests can be used to evaluate the performance of the c&d aggregates in the roadway. this test is carried out to research the improvement of 1e-3 0.01 0.1 1 10 100 0 10 20 30 40 50 60 70 80 90 100 fin er (% ) grain size (mm) rca cb rap 1e-3 0.01 0.1 1 10 100 0 10 20 30 40 50 60 70 80 90 100 fin er (% ) grain size (mm) rca cb rap rca +2% opc cb +2% opc rap +2% opc rca +4% opc cb +4% opc rap +4% opc rca +8% rha cb +8% rha rap +8% rha rca +16% rha cb +16% rha rap +16% rha 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1.70 1.75 1.80 1.85 1.90 1.95 2.00 2.05 2.10 rca + 0% cement cb + 0% cement rap + 0% cement rca + 2% cement cb + 2% cement rap + 2% cement rca + 4% cement cb + 4% cement rap + 4% cementd ry de ns ity (m g/m 3 ) moisture content (%) rc a + 2% op c rc a + 4% op c cb + 2 % o pc cb + 4 % o pc ra p + 2% op c ra p + 4% op c0 500 1000 1500 2000 2500 3000 uc s ( kp a) control 1 day 7 days 28 days 14 a. s. m. riyad and md. zakir hasan asha effect of cement on strength properties of …… compressive strength resulting from the stabilization with opc. as per the authors’ knowledge, there are no practical considerations on the ucs requirements of roads in bangladesh. according to ufc (2004), the minimum required ucs test value of 7 days curing varies from approximately 1400 kilopascal for cementtreated pavement sub-bases to 5000 kilopascals for bases. the ucs tests on unstabilized control specimens without a cement treatment and testing are conducted immediately after compaction for comparison of the outcomes of the respective c&d material. furthermore, ucs tests are conducted to evaluate the effects of curing duration and cement content on the development of compressive strength with 2 and 4% cement contents (figure 5). the rap shows the highest ucs value among all untreated c&d materials followed by rca and cb. design standards generally based on 7-day test results; however, for a better understanding of the development of ucs of the cement-stabilized c&d materials, 1 and 28 days of curing are also performed. noteworthy increments of ucs values are evident in the cement-stabilized c&d aggregates concerning the control samples after 1-day curing. subsequently, the ucs values are increased considerably with the increase of the curing period to 7 days. furthermore, the ucs values are increased moderately with the increase of the curing period to 28 days. the higher cement content led to higher ucs value as it increased the bond strength between the materials due to the hydration process with the progression of time. it is clear from the outcomes that the progression of the process of hydration dwindles with time, i.e., the strength values are increased at an impetuous rate at the commencement of the curing duration and start to plateau after 28 days. better performance is observed for rap than cb and rca in all cases with the same opc content under the same duration of curing; however, rca presented higher ucs values concerning cb; which indicated that the quality of rap is best among all the tested materials. the rap presented maximum compressive strength among the three tested c&d materials, with approximately 85% of 28 days strength of curing owing to the inclusion of opc, and the remaining 15% attributed to the initial untreated strength of the material. however, the rca and cb presented lower strength than rap with about 93% of 28 days' strength of curing owing to the inclusion of opc, and the remaining 7% attributed to the initial untreated strength of the material. based on this study, the cement treatment can be a viable option for the improvement of strength of each of the c&d aggregates.the bond formation and hydration process are highest in rap as compared with rca and cb. the rap reached almost 16% of 28 days strength of curing after 1-day curing, whereas the cb and rca reached almost 7% of the 28 days cured strength after the same duration. the development of strength of cement-treated c&d samples is significant after the 7-day curing period (90-95% strength gain). 4. conclusions the laboratory characterization of cement-admixed stabilized c&d aggregates is assessed to evaluate the performance of treated materials compared with the untreated c&d aggregates. the strength development on the treated c&d materials is investigated varying curing duration. based on the ucs test results, rap, cb, and rca are found to require 2% cement to meet the requirements specified by the u.s. army corps of engineers for subbase courses. a curing period of 7 days has been qualified for the requirement. the rap presented better strength compared with rca and cb in all cases with the designatedpercentage of cement under the nominated duration of curing. the initial dry density of the untreated rap sample exhibits higher value that makes the initial strength of the rap aggregates high.the research outcomes indicated that cement-stabilized materials are possible alternatives for the stabilization of unbound c&d aggregates. acknowledgments the authors would like to express profound gratitude to the department of civil engineering, khulna university of engineering & technology for the immense support. references abedin, m.a., and jahiruddin m., 2015. waste generation and management in bangladesh: an overview, asian journal of medical and biological research, 1(1), 114–120, https://doi.org/10.3329/ajmbr.v1i1.25507 al-bared, m. a. m., marto a., latifi n., horpibulsuk s., 2018. sustainable improvement of marine clay using recycled blended tiles, geotechnical and geological engineering, 36(5), 3135-3147. https://doi.org/ 10.1007/s10706-018-0525-8 ali, f.h., 1992. stabilization of a residual soil, soils and foundations, 32(4), 178-185. https://doi.org/ 10.3208/ sandf1972.32.4_178 arulrajah, a., piratheepan j., aatheesan t., and bo m.w., 2011. geotechnical properties of recycled crushed brick in pavement applications, journal of materials in civil engineering, 23(10), 1444–1452. https://doi.org/ 10.1061/(asce)mt.1943-5533.0000319 arulrajah, a., piratheepan j., disfani m. m., and bo m. w., 2013. geotechnical and geoenvironmental journal of engineering science 11(1), 2020, 09-17 15 properties of recycled construction and demolition materials in pavement subbase applications, journal of materials in civil engineering, 25(8), 1077-1088. ,https://doi.org/10.1061/(asce)mt.1943-5533. 0000 652 astm, 2006. standard test method for resistance to degradation of small-size coarse aggregate by abrasion and impact in the los angeles machine, astm c 131/c 131m, astm international, west conshohocken, pa. astm, 2007. standard test method for particle-size analysis of soils, astm d 422-63, astm international, west conshohocken, pa. astm, 2009. standard test method for unconfined compressive strength of compacted soil-lime mixtures, astm d 5102, astm international, west conshohocken, pa. astm, 2012. standard test methods for laboratory compaction characteristics of soil using modified effort (56,000 ft-lbf/ft3 (2,700 kn-m/m3)), astm d 1557, astm international, west conshohocken, pa. astm, 2014. standard test methods for moisture, ash, and organic matter of peat and other organic soils, astm d 2974, astm international, west conshohocken, pa. astm, 2015a. standard test method for relative density (specific gravity) and absorption of coarse aggregate. astm c 127, astm international, west conshohocken, pa. astm, 2015b.standard test method for relative density (specific gravity) and absorption of fine aggregate. astm c 128, astm international, west conshohocken, pa. astm, 2017a. standard test methods for particle-size distribution (gradation) of soils using sieve analysis. astm d 6913 / d 6913m, astm international, west conshohocken, pa. astm, 2017b. standard test methods for liquid limit, plastic limit, and plasticity index of soils. astm d 4318, astm international, west conshohocken, pa. astm, 2017c. standard practice for classification of soils for engineering purposes (unified soil classification system). astm d 2487, astm international, west conshohocken, pa. astm, 2019. standard test methods for ph of soils. astm d 4972, astm international, west conshohocken, pa. british standard 2000. method for determination of particle shape-flakiness index, british standard 812105.1, british standard institution, london, uk. british standard 1990. methods for determination of aggregate impact value (aiv), british standard 812112, british standard institution, london, uk. bansal, s., and singh s.k., 2014. a sustainable approach towards the construction and demolition waste, international journal of innovative research in science, engineering and technology, 3(2), 12621269. bhuyan, m.a., 2009. evaluation of flexible and rigid pavements construction in bangladesh,m.sc. thesis, buet, dhaka, bangladesh. chowdhury, f.h., raihan m.t., islam g.m.s., and ramiz f. 2016a. construction waste management practice: bangladesh perception. proceedings of 3rd international conference on advances in civil engineering, 21-23 december, chittagong, bangladesh. chowdhury, m.a.i., upadhyay a., briggs a., and belal m.m., 2016b. an empirical analysis of green supply chain management practices in bangladesh construction industry, european operation management association (euroma) conference 2016, 17–22 june 2016, trondheim, norway. disfani, m.m., arulrajah a., haghighi h., mohammadinia a., and horpibulsuk s., 2014. flexural beam fatigue strength evaluation of crushed brick as a supplementary material in cement stabilized recycled concrete aggregates, construction and building materials, 68, 667-676. https://doi.org/10.1016/j.conbuildmat. 2014.07.007 du, y., li s., and hayashi s., 1999. swelling–shrinkage properties and soil improvement of compacted expansive soil, ning-liang highway, china. engineering geology, 53(3-4), 351-358. https://doi.org/ 10.1016/s0013-7952(98)00086-6 du, y.j., jiang n.j., liu s.y., jin f., singh d.n., and puppala a.j., 2013a. engineering properties and microstructural characteristics of cement-stabilized zinc-contaminated kaolin, canadian geotechnical journal, 51(3), 289-302, https://doi.org/10.1139/cgj-2013-0177 du, y.j., wei m.l., jin f., and liu z.b., 2013b. stress–strain relation and strength characteristics of cement treated zinc-contaminated clay, engineering geology, 167, 20-26. https://doi.org/10.1016/j.enggeo.2013.10.005 eurostat 2019. waste statistics – statistics explained, available online: https://ec.europa.eu/15haracte/statisticsexplained/index.php?title=waste_statistics (accessed on 27 october 2019). ganiron jr, t.u., 2015. recycling concrete debris from construction and demolition waste, international journal of advanced science and technology, 77, 7-24. http://dx.doi.org/10.14257/ijast.2015.77.02 hatt, w.k., 1939. the cooperative research projectpurdue university and indiana highway commission progress report, highway research board proceedings, 18(1), 255-263. 16 a. s. m. riyad and md. zakir hasan asha effect of cement on strength properties of …… hoyos, l. r., puppala a. j., and ordonez c. a., 2011. characterization of cement-fiber-treated reclaimed asphalt pavement aggregates: preliminary investigation, journal of materials in civil engineering, 23(7), 977-989. https://doi.org/10.1061/(asce)mt.1943-5533.0000267 islam, f.a.s., 2016. solid waste management system in dhaka city of bangladesh, journal of modern science and technology, 4 (1), 192–209. jongpradist, p., jumlongrach n., youwai s., and chucheepsakul s., 2010. influence of fly ash on unconfined compressive strength of cement-admixed clay at high water content, journal of materials in civil engineering, 22(1), 49-58. https://doi.org/10.1061/(asce)0899-1561 (2010)22:1 (49) kaza, s., yao l., bhada-tata p., and van woerden f., 2018. what a waste 2.0: a global snapshot of solid waste management to 2050, world bank publications. kilic, r., kucukali o., and ulamis k., 2015. stabilization of high plasticity clay with lime and gypsum(ankara, turkey), bulletin of engineering geology and the environment, 75, 735-744. https://doi.org/10.1007/ s10064-015-0757-2 kim, w., labuz j.f., and dai s., 2007. resilient modulus of base course containing recycled asphalt pavement, transportation research record, 2005(1), 27-35. https://doi.org/10.3141/2005-04 kootstra, b.r., ebrahimi a., edil t.b., and benson c.h., 2010. plastic deformation of recycled base materials, geo florida, 2682-2691, https://doi.org/10.1061/41095(365)272 kulatunga, u., amaratunga r. d. g., haigh r., rameezdeen r., and rameezdeen d., 2005. sources of construction material wastage in sri lankan sites, proceedings of the 2nd scottish conference for postgraduate researchers of the built and natural environment (probe), glasgow caledonian university, scotland, uk. langer, w.h., 2001. construction materials: crushed stone, sand, and gravel. encyclopedia of materials: science and technology, 1537–1546. latifi, n., vahedifard f., ghazanfari e., and rashid a.s.a., 2018. sustainable usage of calcium carbide residue for stabilization of clays, journal of materials in civil engineering, 30(6), 04018099 (110).https://doi.org/ 10.1061/(asce)mt.1943-5533.0002313 lemanska, j.j., 2019. impurities of recycled concrete aggregate types, origin and influence on the concrete strength parameters, iop conference series: materials science and engineering, 603, 042056 (1-10). luangcharoenrat, c., intrachooto c., peansupap v., and sutthinarakorn w., 2019. factors influencing construction waste generation in building construction: thailand’s perspective. sustainability, 11, 3638, 1-17. mohammadinia, a., arulrajah a., sanjayan j., disfani m.m., bo m.w., and darmawan s., 2014. laboratory evaluation of the use of cement-treated construction and demolition materials in pavement base and subbase applications, journal of materials in civil engineering, 27(6), 04014186 (1-12). https://doi.org/ 10.1061/ (asce)mt.1943-5533.0001148 mohammadinia, a., arulrajah a., sanjayan j., disfani m.m., bo m.w., and darmawan s., 2016a. strength development and microfabric structure of construction and demolition aggregates stabilized with fly ash–based geopolymers, journal of materials in civil engineering, 28(11), 04016141 (1-8), https://doi.org/10.1061/(asce)mt.1943-5533.0001652 mohammadinia, a., arulrajah a., sanjayan j., disfani m.m., win bo m., and darmawan s., 2016b. stabilization of demolition materials for pavement base/subbase applications using fly ash and slag geopolymers, journal of materials in civil engineering, 28(7), 04016033 (1-9).https://doi.org/ 10.1061/(asce)mt.1943-5533. 0001526 mundy, m., 1991. resilient modulus characterization of granular unbound pavement materials, materials, technology research and development program report. najafpoor, a.a., zarei a., jamali-behnam f., vahedian-shahroudi m., and zarei a., 2014. a study identifying causes of construction waste production and applying safety management on construction site, iranian journal of health sciences, 2(3), 49-54. poon, c.s., and chan d., 2006. feasible use of recycled concrete aggregates and crushed clay brick as unbound road subbase, construction and building materials, 20(8), 578–585. puppala, a.j., hoyos l.r., and potturi a.k., 2011. resilient moduli response of moderately cement-treated reclaimed asphalt pavement aggregates. journal of materials in civil engineering, 23(7), 990-998. ,https://doi.org/ 10.1061/(asce)mt.1943-5533.0000268 rahman, m.a., imteaz m., arulrajah a., and disfani m.m., 2014. suitability of recycled construction and demolition aggregates as alternative pipe backfilling materials. journal of cleaner production, 66, 7584. ,https://doi.org/ 10.1016/j.jclepro.2013.11.005 rahman, m.a., imteaz m.a., and arulrajah a., 2016. suitability of reclaimed asphalt pavement and recycled crushed brick as filter media in bioretention applications, international journal of environment and sustainable development, 15(1), 32-48. https://doi.org/10.1504/ijesd.2016.073333 roads and highways department (rhd) (2011). standard tender documents-section 7, roads and highways journal of engineering science 11(1), 2020, 09-17 17 department, ministry of communications, government of the people’s republic of bangladesh. roads and highways department (rhd), 2018. maintenance and rehabilitation needs report of 2018 2019 for rhd paved roads. road transport and highways division, ministry of road transport and bridges, government of the people’s republic of bangladesh. sharman, j. 2018. construction waste and sustainability, available online: https://www.thenbs.com/ knowledge/construction-waste-and-materials-efficiency (accessed on 27 october 2019). tam, v.w., tam c.m., 2006. a review on the viable technology for construction waste recycling, resources, conservation and recycling, 47(3), 209-221,https://doi.org/10.1016/j.resconrec.2005.12.002 ufc, 2004. soil stabilization for pavements, u.s. army corps of engineers, unified facilities criteria (ufc 3250-11). united states environmental protection agency (epa) 2017. the state of the practice of construction and demolition material recovery-final report, office of research and development, national risk management research laboratory, land and materials management division. world highways, 2004. road technology, april issue. yaowarat, t., horpibulsuk s., arulrajah a., mirzababaei m., and rashid a.s.a., 2018. compressive and flexural strength of polyvinyl alcohol–modified pavement concrete using recycled concrete aggregates, journal of materials in civil engineering, 30(4), 04018046 (1-8), https://doi.org/ 10.1061/ (asce)mt. 1943-5533.0002233 yeheyis, m., hewage k., alam m.s., eskicioglu c., and sadiq r., 2013. an overview of construction and demolition waste management in canada: a lifecycle analysis approach to sustainability, clean technologies and environmental policy, 15(1), 81-91. https://doi.org/10.1007/s10098-012-0481-6 microsoft word 06_jes_s5217_11_19_2020 journal of engineering science 12(1), 2021, 43-49 doi: https://doi.org/10.3329/jes.v12i1.53100 modelling of cover concrete cracking due to uniform corrosion of reinforcement sheikh shakib* and abu zakir morshed department of civil engineering, khulna university of engineering & technology, khulna – 9203, bangladesh received: 07 may 2019 accepted: 19 november 2020 abstract cracking of cover concrete due to the corrosion of reinforcing steel is one of the main causes of deterioration in reinforced concrete (rc) structures. an outbound stress is developed in concrete surrounding the reinforcing steels due to the expansive corrosion products of reinforcement leading to cracking of the concrete cover. in this paper, the cracking pressure was simulated through a finite element modeling. the effect of geometrical and material parameters, i.e. concrete cover thickness, bar diameter, and concrete tensile strength, on the cracking pressure was also investigated. abaqus 6.14 was used as modeling platform. the cracking pressure was found to dependent on the cover thickness and tensile strength of concrete. a higher pressure was required to initiate crack for a higher cover thicknesses and tensile strength. the cracking pressure was decreased with the increase in bar diameter. finally the crack initiation and propagation has been simulated successfully for different arrangements of reinforcements. keywords: corrosion; cracking pressure; concrete cover; crack propagation; reinforced concrete. 1. introduction concrete, a durable composite material because of having high alkaline pore solution (mundra et al., 2017; bažant, 1979). the high alkalinity attained as a result of cement hydration. in case of reinforced concrete structures, in the presence of solution, a thin layer of fe2o3 forms on the surface of rebars. this layer, also known as layer of passivation, protects the reinforcement from aggressive weather (bažant 1979; glasser and sagoecrentsil, 1989; val et al., 2009). corrosion is one of the most deteriorative mechanism in the saline weather conditions. the chlorides penetrate through the pore spaces of the concrete and accumulates around the reinforcement. at a certain level of accumulation, the layer of passivation destroyed consequences an initiation if corrosion process. this process is an electrochemical process. in which, electrochemical microcells are produced and the rebars continues to corrode. in the reaction process, different types of oxides are formed which were found to be larger volume than the rebar material (pantazopoulou and papoulia, 2001; liu and weyers, 1998). these expansive products thus impose an outward thrust on the surrounding concrete. tensile stresses are developed in the surrounding concrete because of this expansive thrust. concrete is weak in tension which causes the initiation and propagation of cracks (bhargava et al., 2006). the location of crack initiation depends on the cover thickness (shakib and morshed, 2019). in this research, mechanism of cracking was investigated through finite element modelling. corrosion is a complex mechanism especially in concrete environment. last few decades, a lot of works related to corrosion of reinforcement have conducted both experimentally and numerically (bazant, 1979; bhargava et al., 2006; liu and weyers, 1998; pantazopoulou and papoulia, 2001). the researchers tried to figure out the cracking mechanism and cracking pattern due to corrosion of rebars numerically (dagher and kulendran, 1987; du et al., 2006; val et al., 2009). most of them explained the crack initiation mechanism aswhen the circumferential stress, developed by the expansive oxides, exceeds the concrete tensile strength crack initiates and propagate towards the cover surface. but there is an effect of clear cover thickness on the crack initiation mechanism. this research focused on this point. a commercial software, abaqus was employed as modelling platform. the model regarded a uniform corrosion of reinforcement. the model predicted the initiation and propagation phenomena for a single bar as well as couple of bars successfully. 2. methodology 2.1 constitutive modelling of concrete the software, abaqus, is capable of modelling the fracture of concrete material in three different techniques; smeared crack model (scm), concrete damaged plasticity model (cdp), and brittle cracking model (bcm). of the three approaches concrete damaged plasticity model was selected in this study for its robustness in applying to model cracks. to characterize the constitutive behaviour of concrete, four different laws are needed to be defined; compressive and tensile behaviour of concrete, damage function, yield function, and flow rule. *corresponding author: sheikhshakib10@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 44 sheikh shakib and abu zakir morshed modelling of cover concrete cracking due ……... the behaviour of concrete in uniaxial compression and tension considered in the cdp model is shown in figure 1. in case of compression, a linear-elastic response is considered till the initial yield point, σc0. whereas, in tension, a linear stress strain relationship is considered up to the ultimate stress, 𝜎 . in the post-peak region (plastic region), the unloading curve shows a degraded stiffness both in compression and tension. this degradation is due to the damage of material. the model introduces two damage variables, 𝑑 and 𝑑 , to define this degradation. the range of this variables is 0 to 1 depending on the level of damage. zero represents “undamaged” and one represents “fully damaged”. 𝐸 is the initial (undamaged) modulus of elasticity of concrete. 𝜀 and 𝜀 are compressive inelastic and plastic strain, and 𝜀 and 𝜀 are tensile cracking and plastic strain, respectively. the stress-strain behaviour both in compression and tension are considered as the following equations (abaqus user manual) 𝜎 = (1 − 𝑑 )𝐸 (𝜀 − 𝜀 ) (1) 𝜎 = (1 − 𝑑 )𝐸 (𝜀 − 𝜀 ) (2) figure 1: stress-strain behaviour in tension (a) and compression (b). tensile cracking and compressive crushing are the two failure mechanisms considered in the cdp model. in these mechanisms, the failure surfaces are controlled by the tensile and compressive plastic strains (𝜀 and 𝜀 ) as shown in figure 2. these strains are automatically calculated in the software from the cracking strain and inelastic strain (𝜀 and 𝜀 ) following the equations. 𝜀 = 𝜀 − (3) 𝜀 = 𝜀 − (4) figure 2: definition of (a) cracking strain (𝜀 ) and (b) inelastic strain (𝜀 ) the cdp model uses the modified by lee and fenves (1998) yield function of lubliner et al. (1989) to define the failure surface. two model inputs are needed to define the yield function in the model; the ratio of biaxial to uniaxial compressive stress ( default value is 1.16) and the ratio of the tensile and compressive second invariant (𝐾 , default value is 2/3). 𝐾 controls the shape of failure surface in the deviatory plane having a wide range of values, 0-1, as shown in figure 3. a non-associated plastic flow is presumed by cdp model where the druckerprager hyperbolic function is used as flow potential. the function is as follows 𝐺 = (∈ 𝜎 𝑡𝑎𝑛𝜓) + 𝑞 − 𝑝 𝑡𝑎𝑛𝜓 (5) where, ѱ = dilation angle which is measured in the p–q plane at high confining pressure; ϵ = eccentricity. the default value of eccentricity is 0.1. (a) (b) journal of engineering science 12(1), 2021, 43-49 45 figure 3: yield surfaces for different values of 𝐾 in the deviatoric plane 2.2 input parameters of the model the cdp model requires the values of and 𝐾 to define the yield function as well as ѱ and ϵ to define flow rule. this model made use the default values of , 𝐾 , and ϵ as shown in table 1. a dilation angle of 300 is considered in this study. a special parameter, viscosity parameter, needs to be defined in the model which controls the convergence of model. the value of this parameter is selected by trial method. table 1: parameters for cpdm ѱ ϵ 𝜎 𝜎 𝐾 viscosity parameter 300 0.1 1.16 2/3 0.0001 two different parameters, modulus of elasticity (𝐸 ) and poisson’s ratio are to be input to define the elastic property of the material. in this study, 0.2 is chosen as the poisson’s ratio. the 𝐸 was calculated by the equation given by aci 318 as follows 𝐸 = 57000 𝑓 psi (6) to define the stress-strain relationship in compression, the model proposed by popovics, (1973) was used in this study. the relationship is shown in the following equation, = ( ) ( ) ( ) (7) where 𝑓 , 𝜀 are the compressive strength and strain corresponding to the maximum stress, respectively. the ‘n’ is defined by, 𝑛 = 0.4𝑥10 𝑓 (𝑝𝑠𝑖) + 1.0 (8) whereas, the stress-strain behaviour in tension was presumed linear upto ultimate stress afterward determined by using the following equation, 𝜎 = 𝑓 ( )( . ), 𝜀 = (9) 3. results and discussions 3.1 model verification this study aimed to simulate the mechanical effects (generation of pressure, cracking etc.) of corrosion of reinforcements on the surrounding concrete. a uniform was considered to simplify the model. due to this uniform corrosion, the problem was modelled as two-dimensional frame and formulate as plain strain problem. the 3-node linear plane strain triangle element was choosing to represent concrete. an expansive pressure employs and increases gradually as the corrosion progresses (shakib and morshed, 2019). this pressure consequences a tensile stress on the cover concrete. when this pressure exceeds the resisting capacity (tensile strength) of the concrete material cover crack initiates. this pressure required to initiate crack is named as critical pressure. the model was validated by comparing the critical pressure with the experimental ones from williamson and clark (2000). the test specimens used by williamson and clark (2000) were cubical (with a side of 150 mm) in shape with a hole at one corner as shown in figure 4. the diameter of hole is 8mm. the hole represents as the reinforcement. three different cover thicknesses were used in the experiment; 4 mm, 8 mm, and 16 mm respectively. since the expansive corrosion products exerts an outward pressure on the surrounding concrete, a uniform and gradually increasing hydraulic pressure was applied in the experimentation through the 46 sheikh shakib and abu zakir morshed modelling of cover concrete cracking due ……... hole. in this study, to model this corrosion induced pressure an outward deformation was applied into the hole. the pressure required to crack the surrounding concrete was recorded for different cover thicknesses as shown in figure 5. these pressures were compared with the experimental results obtained from williamson and clark (2000) as tabulated in table 2. critical pressures obtained from the model were comparable with the experimental ones (difference below 25%). the model then used to simulate crack initiation and propagation in structures comprised of single and couple of bars. . (a) (b) (c) figure 4: specimens’ dimensions tested by williamson & clark, (2000) (a) 3d specimen (b) 2d formulation (c) meshing table 2: the critical pressure for concrete cover cracking c/d pmax/numerical (mpa) pmax/exp (mpa) δp (%) 0.5 2.70 2.65 1.9 1.0 4.72 4.08 15.7 2.0 6.00 7.71 22.2 figure 5: comparison of results of finite-element analysis with test results from williamson & clark (2000): for d=8 mm 3.2 crack initiation mechanism in order to investigate the initiation and propagation of cracks due to corrosion of reinforcement, specimens with 200 mm x 200 mm with a clear cover of 20 mm, 37.5 mm, 50 mm, and 75 mm respectively was modelled. from the analysis it was found that for a lower clear cover (20 mm and 37.5 mm), a heaving of cover concrete occurred as shown in figure 6. due to this bending effect, a tensile stress developed on the cover surface of the specimens. it consequence an initiation of crack at the cover surface and propagated inward. this phenomenon was observed experimentally by shakib and morshed (2019). whereas, for the clear covers of 50 mm and 75 mm, crack initiated at the interface of reinforcement and concrete and propagated outwards as no heaving encountered. in this case, circumferential stress (tensile in character) developed due to the pressure was exceeded the tensile strength of concrete. thus, crack initiated at the interface and propagates outward as shown in figure 7. 0 2 4 6 8 10 2 2.5 3 3.5 c ri ti ca l p re ss u re , p m ax (m p a) tensile strength of concrete, ft (mpa) c/d=0.5, test c/d=1.0, test c/d=2.0, test c/d=0.5, fea c/d=1.0, fea journal of engineering science 12(1), 2021, 43-49 47 cover = 20 mm cover = 37.5 mm figure 6: initiation and propagation of crack for cover 20 mm and 37.5 mm cover = 50 mm cover = 75 mm figure 7: initiation and propagation of crack for cover 50 mm and 75 mm figure 8: effect of cover thickness on the critical pressure and comparison with the results of williamson & clark, (2000) 3.3 relation of concrete cracking pressure with concrete cover the variation of the critical pressures depending on the cover thicknesses are provided in figure 8. it can be seen from the figure that, with the increase in cover thicknesses, the critical pressure also increased. the model predicts the pressure required to cover cracking similar to that of test for cover thickness 4 mm. on the other hand, for cover thickness 8mm, the model overestimates for higher tensile strengths (𝑓 = 2.6, 3.0, 3.2 mpa) but underestimate for tensile strength 2.2 mpa. for cover thickness 16mm model underestimate the pressures irrespective of the tensile strengths. 3.4 relation of concrete cracking pressure with bar diameter five finite element models are developed with varying bar (considered in model as a hole) diameters (d), from 10 mm to 25 mm, where all other geometrical and material properties are kept constant (c = 37.5 mm, 𝑓 = 3.4 mpa). the pressure required for cracking of the concrete cover for each model is shown in figure 9. as seen in the figure, the expansive pressure decreases as the bar diameter increases. by increasing of the bar diameter, the 0 2 4 6 8 0 5 10 15 20 c ri ti ca l p re ss u re ( m p a) concrete cover (mm) fea-ft=2.2 mpa fea-ft=2.6 mpa fea-ft=3.2 mpa test-ft=2.2mpa test-ft=2.6mpa test-ft=3.2mpa 48 sheikh shakib and abu zakir morshed modelling of cover concrete cracking due ……... lateral surface of the hole increases which results in higher outward force and consequently lower required pressure for the cracking. figure 9: effect of bar diameter on the required pressure for cracking figure 10: effect of concrete modulus of elasticity on the required pressure 3.5 relation of concrete cracking pressure with modulus of elasticity the variation of concrete cracking pressure with respect to modulus of elasticity of concrete has shown in figure 10. williamson and clark, 2000 didn’t report the effect of modulus of elasticity of concrete on corrosion pressure. as shown in figure the higher the value of 𝐸 the pressure required to cover crack. figure 11: (a) initiation and (b) propagation of cracks for 2-12 mm bar figure 12: (a) initiation and (b) propagation of cracks for 4-12 mm bar figure 13: (a) initiation and (b) propagation of cracks for 8-12 mm bar 3.6 patterns of cracks in beam the model was then used to anticipate the corrosion induced cracking of concrete beams (having cross section of 300 mm × 500 mm) comprised of a different number of bars (ø-12). a clear cover of 37.5 mm was maintained for all the specimens. three different configurations were investigated; 2 and 4 nos. of bars in single layer, and 8 nos. of bar in double layer. the patterns of cracks are shown in figures 11-13. for 2-nos of bars, the crack pattern was similar to that of single bar; crack initiated at the cover surface. but, for 4 and 8-nos of bars, crack occurred 0 5 10 15 20 25 10 12 16 20 25 c ri ti ca l p re ss u re ( m p a) bar diameter (mm) 0 2 4 6 8 18000 20000 22000 24000 26000 28000 c ri ti ca l p re ss u re ( m p a) ec (mpa) c/d=0.5 c/d=1.0 journal of engineering science 12(1), 2021, 43-49 49 in between the reinforcements first and then propagated to the cover surfaces. pressure generally release through shortest distance from the point of generation. the distance between the reinforcements was shorter than the cover thickness. in addition to the distance measurement, the pressure was generated from both direction. this may be the reason behind these crack patterns for 4 and 8 nos. of bars. 4. conclusions in this paper, corrosion induced expansive pressure was modelled through a finite element software abaqus. rebar corrosion was simulated as simplified 2d formulation. the model successfully depicted the crack initiation mechanisms depending on the cover thicknesses. for relatively thinner concrete covers of 20mm and 37.5 mm, initiation of cracks was found from the concrete surface and propagated inward. on the other hand, for thicker concrete covers of 50mm and 75mm, cracks were found to initiate at the interface of steel reinforcement and concrete. the model was also used to predict the cracking patterns of beams comprised of three different combinations of bars. for 2-nos of bars, crack initiated at the cover surface. but, for 4 and 8-nos of bars, crack occurred in between the reinforcements first and then propagated to the cover surfaces. it was found that critical pressure increased with increase of clear cover, whereas, decreased with the increase in bar diameter. references bazant, z. p., 1979. physical model for steel corrosion in concrete sea structures – theory, journal of the structural division-asce, 105(6), 1137–1153. mundra, s., criado m., bernal s. a., and provis j. l., 2017. chloride-induced corrosion of steel rebars in simulated pore solutions of alkali-activated concretes, cement and concrete research, 100, 385–397. https://doi.org/10.1016/j.cemconres.2017.08.006. glasser, f. p., and sagoe-crentsil k. k., 1989. steel in concrete: part ii electron microscopy analysis, magazine of concrete research, 41, (149), 213–220. https://doi.org/10.1680/macr.1989.41.149.213. val, d. v., chernin l., and stewart m. g., 2009. experimental and numerical investigation of corrosion-induced cover cracking in reinforced concrete structures, journal of structural engineering, 135(4), 376–385. https://doi.org/10.1061/(asce)0733-9445(2009)135:4(376). pantazopoulou, s. j., and papoulia k. d., 2001. modeling cover-cracking due to reinforcement corrosion in rc structures, journal of engineering mechanics, 127(4), 342–351. liu, y., and weyers r. e., 1998. modeling the time-to-corrosion cracking in chloride contaminated reinforced concrete structures, aci materials journal, 95(6), 675–680. bhargava, k., ghosh a. k., mori y., and ramanujam s., 2006. model for cover cracking due to rebar corrosion in rc structures, engineering structures, 28(8), 1093–1109. shakib, s., and morshed a. z., 2019. experimental investigation on crack initiation and propagation due to corrosion of reinforcement, advances in civil engineering materials, 8(1), 688–698. https://doi.org/ 10.1520/acem20190161 dagher, h. j., and kulendran s., 1987. finite element modeling of corrosion damage in concrete, aci structural journal, 89(6). du, y. g., chan a. h. c., and clark l. a., 2006. finite element analysis of the effects of radial expansion of corroded reinforcement, computers & structures, 84(13–14), 917–929. popovics, s., 1973. a numerical approach to the complete stress-strain curve of concrete, cement and concrete research, 3(5), 583–599. lee, j. and fenves g. l., 1998. plastic-damage model for cyclic loading of concrete structures, journal of engineering mechanics, 124(8), 892–900. lubliner, j., oliver j., oller s., and oñate e., 1989. a plastic-damage model for concrete, international journal of solids and structures, 25(3), 299–326. © 2021 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 3_jes_349 journal of engineering science 13(1), 2022, 21 29 doi: https://doi.org/10.3329/jes.v13i1.60559 *corresponding author: hasanabir009@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal measurement of efficient travel time of a highway corridor through moving observer method: a case study on savarmanikganj highway in bangladesh md. abir hasan*1, ashraf uddin fahim2 1department of civil engineering, faculty of engineering and technology, pabna university of science and technology, pabna-6600, bangladesh. 2department of urban and regional planning, faculty of engineering and technology, pabna university of science and technology, pabna-6600, bangladesh. received: 22 december 2021 accepted: 24 april 2022 abstract travel time and its variability is one of the biggest controlling factors for trip planning, mode selection and forecasting trip duration. travel time can be affected by various issues. savar-aricha highway is the gateway between south and south-western part of the country. large variability on this corridor can cause great economic losses. to understand this routes travel time, running velocity and level of service moving observer method was conducted. the survey also includes travel time, stopped time and journey time for two types of public transport present on the corridor: ticket-based bus system and without ticket local bus system. the study found that, ticket based bus system has lesser travel time and stopped time resulting is less travel time in comparison with local bus system. speed analysis for both the route also supports this. in consideration of free flow speed, the level of service of this corridor was found a. keywords: highway corridor, level of service, moving observer method, public transport, travel time. 1. introduction travel time, or the amount of time it takes to travel between two sites of interest, is a crucial metric in the transportation planning (arahan, 1986). transportation engineers and planners, business people, commuters, media representatives, administrators, and customers all understand and convey travel time, which is a fundamental concept (box & oppenlander, 1976). since the late 1920s, engineers and planners have employed travel time and delay studies to assess transportation systems and plan improvements. surveys of speed, journey time, and delay are essential in traffic control and roadway upgrades (edie, 1974). roadside friction, vehicle interaction, and the effect of traffic signals all influence traffic flow on arterial roadways (hall, 1994). all of these factors have an impact on the capacity and level of service provided by arterial roadways. capacity of the arterial roads is generally considered by the capacity of signalized intersection along those arterial roads (lum et al. 1998). the amount of access points alongside arterial highways, land use type, pedestrian activity, parking space, and road shape all contribute to roadside friction (mortimer, 1957). the average through-vehicle travel speed is used to determine the effectiveness of arterial roadways (quiroga & bullock, 1998). travel time data can be used to measure a route's efficiency in terms of carrying traffic volume in comparison to other routes, as well as to offer information for capacity analysis of roadway segments (robertson et al. 1994). intersection delay only measures the delay at a single intersection on a given approach, whereas travel time reflects the delay on a larger scale (salter, 1989). by dividing the travelled distance by the journey time, travel speed may be calculated. for corridors, the travel speed can be used to determine the degree of service. estimation of arterial roads level of service depends on free flow speed and average travel speed as shown in table 1 (taylor & abdel-rahim, 1998). 22 m. a. hasan & a. u. fahim measurement of efficient travel time ……… table 1: arterial road los by class. highway corridor 1 2 3 4 range of free flow speed 90-70 km/h 70-55 km/h 55-50 km/h 55-40 km/h range of free flow speed 80 km/h 65 km/h 55 km/h 45 km/h average travel speed los 1 2 3 4 a >72 >59 >50 >41 b >56-72 >46-59 >39-50 >32-41 c >40-56 >33-46 >28-39 >23-32 d >32-40 >26-33 >22-28 >18-23 e >26-32 >21-26 >17-22 >14-18 f ≤ 26 ≤21 ≤17 ≤14 the speed flow relationship is another method to estimate the travel time. the travel speed can be estimated based on traffic flow data. therefore, it is important to estimate a typical speed flow correlation. in order to improve a country's socioeconomic situation and urban areas, transportation and mobility are critical issues to address. a well-designed and constructed transportation infrastructure not only assists people in getting around, but it also has an impact on a city's rate of growth and level of economic activity (fahim, et al. 2022). the countrys overall growth lagely depend on efficient connectivity of different partd with the capital. savarmanikganj route is very important for the connectivity of the southern part with the capital. travel time and its variability of this corridor is of great interest for various road users. the choice of traffic mode has a close connection with travel time, comfort and safety. both ticked based bus and local bus are the main mode of public transport of this route. but the travel time of the bus systems have not been evaluated so far. moreover, traffic flow and speed in this corridor needs to be determined to better understand it’s level of service. the goal of this study is to determine traffic flow and speed on this two-lane interstate highway utilizing the moving observer method. the study then, compares the results from the ticket bus system with the local bus system for determining efficient travel time on this corridor. this study will help policy makers, transport planners and engineers to understand the existing level of service of the corridor. also, the knowledge of journey time and stopped time will help policy makers and planners to optimize bottlenecks and reduce effective travel time. moreover, estimated efficient travel time will help deciding modal choice and trip planning for the road users. 2. methodology 2.1 study route savar to manikganj highway segment is on national highway 5 (n5). this is a 47 km long highway touching savar union, jahangirnagar university, savar cantonment, national martyrs’ monument, dhamrai and manikganj. this segment acts as a link between dhaka and paturia which connects southern and western part of the country with dhaka. the major traffic contributors of this segment are public bus and large to medium truck. the route is crucial for both passenger movement and goods transportation. travel time on this segment is of great interest for transport planners, policymakers and engineers. to conduct a speed, journey time and delay survey the selection of route is an important phenomenon. because it is needed to select a route where various traffic modes play on the road. in this respect, “savar to manikganj” is a suitable route for the study. figure 1 shows the gis map of the selected route. journal of engineering science 13(1), 2022, 21-29 23 figure 1: selected route “savar to manikganj. 2.2 moving observer method data on trip time, speed, average flow rate, and traffic density is collected using the moving observer approach. the original "moving observer approach" was presented by arahan (1986) as a means of measuring the average flow and travel time of vehicles traveling in one direction on a highway segment. the method was developed using data collected by a moving observer in a test vehicle that was immersed in traffic (taylor & abdel-rahim, 1998). the observer drove a test car in the direction of the flow, counting how many vehicles were overtaken and how many were passed (figure 2). figure 2: moving observer method. the number of opposing cars facing in the opposite direction (direction of interest) is counted during the trip by traveling against the flow in the opposite direction (wardrop & charlesworth, 1954). the traveling time of the 24 m. a. hasan & a. u. fahim measurement of efficient travel time ……… test car is recorded in both directions. in addition, the road segment's length is known. the speed-flow relationship for the road segment in the direction of interest is then calculated using these values. the number of runs is selected as per the manual of transportation engineering studies to maintain the accuracy (wright, 1973). the observer in the test vehicle records information about the number of vehicles passing the test vehicle in front of it, the number of automobiles that passed the test, the number of cars met when traveling in the opposite direction, the observer's segment length, trip time in both directions and with and against the flow of traffic for each run. arhan (1986) stated that the advantages of using the moving observer method are the observer can collect data on flow and speed at the same time (this is valuable when seeking the relationship between these two variables), the observer can record trip time, as well as the flow rate and average speed of cars, along the length of the road segment. moreover, when compared to other approaches, the moving observer method requires less manpower and hours to achieve a high degree of accuracy, making it less expensive and vehicles can be grouped, and flow rates for each group can be estimated, if necessary, the observer can record additional information such as the locations and causes of delays. however, box and oppenlander (1976) highlighted a number of problems with the moving observer method, including the following: to achieve a specified level of accuracy, the observer requires a number of tests runs when traffic flows are low (200-300 vehicles per hour for two lanes, one direction), which may be impossible. the traffic volume that moves into the test road section changes according to the total number of intersections (box & oppenlander, 1976). the accuracy of measuring speed and flow is extremely sensitive to changes in the traffic flow along the road segment at various times of the day. 2.3 data collection after selecting the study route, a conceptualization was developed through several literature reviews. the moving observer method is a method for measuring journey time, flow rate, space mean speed, and delay on a roadway section that incorporates the use of a probing vehicle within a traffic stream. from saturday, january 24, to thursday, january 29, 2019, data was collected during off-peak daytime hours between 11:00 a.m. and 1:00 p.m. during the study, a segment of 40.6 km in length was used for data collection. the researcher performed six (6) test runs on each directional segment comprising total 12 test runs in both directions. the observer in the survey vehicle travelled at the average travel speed inside the traffic stream of the chosen stretch under review, according to the manual of transportation engineering studies, which defines the procedure. however, the observer cannot travel for several miles along the portion without slowing down. thus, the number of vehicles passed was subtracted from the number of vehicles overtaken by the observer to solve this problem. during the travel the observer also collected travel time data, stopped time data, journey time data for ticked based bus and local bus system. 2.4 formulation a several case can occur in conducting moving observer survey, a number of vehicle can overtake the observer. secondly, a stationary traffic stream of density 𝑘 can be passed by observer with velocity 𝑣 or observer can move within the stream. this three cases can be expresses generally by the equation 1. 𝑚 = 𝑚 − 𝑚 = 𝑞𝑡 − 𝑘𝑣 𝑡 (1) where, 𝑚 = number of vehicles that will overtake observer 𝑚 = number of vehicles that will be overtaken by observer 𝑞 = traffic flow 𝑡 = time period 𝑘 = flow density 𝑣 = observer speed this equation is the basic equation of moving observer method, which relates 𝑞, 𝑘 to the counts 𝑚, 𝑡 and 𝑣 that can be obtained from the test. for generating two equations, the test vehicle is run twice once with the traffic journal of engineering science 13(1), 2022, 21-29 25 stream and another one against traffic stream, to solve for two unknowns q and k. the above general equation can be written as 𝑚 = 𝑞𝑡 − 𝑘𝑣 𝑡 (2) 𝑚 = 𝑞𝑡 − 𝑘𝑣 𝑡 (3) where, a, w denotes against and with traffic flow. from the equations 2 and 3 the flow of the traffic stream can be calculated using the following formula 𝑞 = (4) where, 𝑞 = flow in the route 𝑚 = average number of vehicles overtaking the test bus minus the number overtaken by the test bus 𝑚 = average opposing traffic count of vehicles met when the test bus was travelling 𝑡 = average journey time when the test bus travelled in the with traffic route 𝑡 = average journey time when the test bus travelled in the against traffic route mean journey time and mean travel speed can also be calculated from the following equations: 𝑡 = 𝑡 − (5) 𝑣 = (6) where, 𝑡 = mean journey time 𝑣 = mean travel speed 𝑙 = length of the roadway segment under observation 3. results and discussion 3.1 travel time, stopped time and journey time the detailed travel time and delay study on the route savar to manikganj shows that, local bus that serves without ticket takes more time to travel compared with ticket-based bus system (figure 3-4). moreover, the study shows stopped time for local bus is significantly greater regarding ticket-based bus system. as a result, local bus takes more time to traverse the savar to manikganj route. study on reverse direction, manikganj to savar route also shows local bus systems takes more travel time and stopped time resulting in more journey time. figure 3: travel time, stopped time and journey time for savar to manikganj. figure 4: travel time, stopped time and journey time for manikganj to savar. 3 9 .0 7 1 2 .3 1 5 1 .3 8 4 5 .1 2 1 8 .2 6 3 .3 2 4 2 .0 9 1 5 .2 6 5 7 .5 5 0 10 20 30 40 50 60 70 travel time (minutes) stopped time (minutes) journey time (minutes) t im e (m in u te s) ticket system bus local bus average 4 4 .0 3 8 .0 7 5 2 .1 5 2 .2 5 1 6 .1 5 6 8 .4 4 8 .1 4 1 2 .1 1 6 0 .2 5 0 10 20 30 40 50 60 70 80 travel time (minutes) stopped time (minutes) journey time (minutes) t im e (m in ut es ) ticket system bus local bus average 26 m. a. hasan & a. u. fahim measurement of efficient travel time ……… the average travel time for savar to manikganj route is 42.09 minutes while average stopped time 15.26 minute resulting average journy time 57.55 minutes. for the opposite direction, manikganj to savar average travel time is 48.14 minutes and stopped time 12.11 minutes so the average journey time is 60.25 minutes. figure 3-4 also shows that for the both ticket and local bus travel time is larger on manikganj to savar route. this is for the presence of side friction and poor traffic management that reduces mean running speed greatly. though the stopped time is less in this direction, total journey time ncreases slightly. from the detailed journey frame study for savar to manikganj route shows that, both in permitted stoppage and non-permitted stoppage exists. the no permitted stoppage contributes significant delay that is almost 20% of total delay. thus, lack of traffic management and delay in non-permitted stoppage increases total delay and journey time as a result. 3.2 traffic flow figure 5 shows that, the flow of the route savar to manikganj was found 616.8 pcus/hour from analysing the survey data using equation 4. similarly, for the opposite direction, manikganj to savar, the flow was found 445.8 pcus/hour. comparison of flow among two direction shows that manikganj to savar route yields less capacity. the potential reasons for this are, flow obstructions, bottlenecks etc. the larger journey time and travel time of the manikganj to savar route supports this finding. figure 5: traffic flow for both direction of savar to manikganj highway. 3.3 journey time, journey speed, running time and running speed the mean journey time is the mean time required to traverse a route. figure 6 shows that, mean journey time of the savar to manikganj route was found 57.40 min. which slightly differs with average journey time. the mean journey speed for this 47 km route was calculated 49.19 kph using mean journey time (figure 7). running time in the time while a vehicle is actually in motion during the journey. mean running time is the difference between mean journey time and stopped delay. for the savar to manikganj route mean running time was found 42.14 min. from that mean running speed was calculated 66.92 kph (figure 6-7). mean running time is the actual mean speed of the vehicle during traversing the route. this exhibits the highway category of the route is class 2 (hcm 2000). as the free flow speed is 66.92, which is greater than 59 k.p.h., the los of the route is a. that means free flow condition exists on the savar to manikganj route. figure 6 also shows that, mean journey time of the manikganj to savar route was found 59.45 min. which slightly differs with average journey time. the mean journey speed for this 47 km route was calculated 47.43 kph using mean journey time (figure 7). for the manikganj to savar route mean running time was found 47.34 min. from that mean running speed was calculated 59.57 kph (figure 6-7). this also exibits the highway category of the route is class 2 (hcm 2000). as the free flow speed is 59.57, which is slightly greater than 59 k.p.h., the los of the route is marginally a and a slight reduction of mean running speed can drag down the los to b. from the comparison of the two direction it was found that, mean journey time is larger for manikganj to savar route resulting a lower mean journey speed. but the stopped delay in the savar to manikganj route is greater, 616.8 445.8 0 100 200 300 400 500 600 700 savar to manikganj manikganj to savar t ra ff ic f lo w ( p c u /h o u r) journal of engineering science 13(1), 2022, 21-29 27 resulting lower mean running time and higher running speed respectively. according to the mean running speed both the corridor is class 2 and the los is a for both ways. figure 6: mean journey time and running time for both direction of savar to manikganj highway. figure 7: mean journey speed and running speed for both direction of savar to manikganj highway. 3.4 overtaking tendency the overtaking tendency of vehicle often creates road accident, traffic congestion etc. to find out the overtaking tendency the number of vehicles those are overtaken by the bus are counted on manual method. figure 8 shows that, at the savar to manikganj route, ticket-based bus system overtaken more vehicles than local buses, as it has greater running speed. figure 9 also shows that, at the manikganj to savar route, local bus system overtaken more vehicles than local buses. the comparison of the two direction also shown that the number of overtaking vehicles are greater on manikganj to savar diretion. thus the accident potential of the local bus on manikganj to savar direction is greated as they try to rush faster to overtake though several bottlenecks restrain the flow of this direction. figure 8: overtaking tendency on savar to manikganj highway. figure 9: overtaking tendency on manikganj to savar highway. 57.4 42.14 59.45 47.34 0 10 20 30 40 50 60 70 mean journey time (minute) mean running time (minute) t im e (m in ut es ) savar to manikganj manikganj to savar 49.19 66.92 47.43 59.57 0 10 20 30 40 50 60 70 80 mean journey speed (kph) mean running speed (kph) s p ee d ( kp h ) savar to manikganj manikganj to savar 13 10 23 11 9 20 0 5 10 15 20 25 ticket system bus local bus total n u m b er o f v eh ic le s overtaking vehicles overtaken vehicles 12 14 26 11 9 20 0 5 10 15 20 25 30 ticket system bus local bus total n u m b er o f v eh ic le s overtaking vehicles overtaken vehicles 28 m. a. hasan & a. u. fahim measurement of efficient travel time ……… 3.5 composition of vehicle from the number of the vehicles met with in the opposing direction provide a useful data about the traffic weight on the road. it represents the amount of traffic flow on the road on the survey period. figure 10 shows that, for savar to manikganj route the opposing vehicle weight mainly consists of large to medium vehicles, which is nearly 96%. almost 26% contribution comes from passenger car and micro. the next higher contributor is truck or pick-up van which is 29% of the opposing vehicle. the public transport, thus the bus occupies 41% vehicle weightage in the route. the leftover 4% contribution comes from cng, taxicab or motorcycle. the bus and truck consist about 70% of the opposing vehicle. figure 11 also shows that, for manikganj to savar route, the opposing vehicle weight also consists of large to medium vehicles, which is also 93%. almost 34% contribution comes from passenger car and micro. the next higher contributor is truck or pick-up van which is 32% of the opposing vehicle. the public transport, thus the bus occupies 27% vehicle weightage in the route. the leftover 7% contribution comes from cng, taxicab or motorcycle. the bus and truck consist about 59% of the opposing vehicle. from the figures, we can see that around 94.5% vehicles are bus, car and truck. figure 10: vehicle composition on savar to manikganj highway. figure 11: vehicle composition on manikganj to savar highway. 4. conclusions this study has employed moving observer method and conducted travel time, speed study and delay study on saval to manikganj route for ticket-based bus system and local bus system. the results show that, ticked based bus system has less stopped time thus greater running speed and less travel time. on the other hand, local bus system costs more stopped time hence increasing travel time. level of service of this class 2 highway corridor was determined according to running speed and found that free flow occurs during commuting. as a result, overtaking tendency was found greater for ticket-based bus systems compared to local buses ignoring the fitness criteria of the two systems. the study also shows that the main contributors of vehicle weight are large to medium vehicles on the both way of this route. acknowledgement the authors would like to express thanks to the department of civil engineering, pabna university of science and technology for technical support for conducting this research. the authors would also like to express gratitude to anonymous reviewers for their continuous assistance. no funding was provided for conducting this research. references arahan t. (1986) a guide on geometric design of roads, kuala lumpur: jabatan kerja raya cawangan jalan. 8/86. 26% 29% 41% 4% car/micro truck/pickup van bus cng/taxi cab/bike 34% 32% 27% 7% car/micro truck/pickup van bus cng/taxi cab/bike journal of engineering science 13(1), 2022, 21-29 29 box p.c. & oppenlander, j.c. (1976) manual of traffic engineering studies, 4th ed. washington: institute of transportation engineers. edie, l.c. (1974) flow theories, traffic science. new york: john wiley & sons, inc. fahim, a. u., rahman, m. m., abir, f. a., & bhuiyan, m. a. f. (2022) an investigation of users’ perception on non-motorized transport services in a municipality area: a cross-sectional study on pabna municipality. case studies on transport policy, 10(1), 657-663. hall, f.l. (1994) traffic stream characteristics. traffic flow theory and characteristics. washington: transportation research board. lum k.m., fan h. s.l., lam s.h. & olszewski p. (1998) speed-flow modelling of arterial roads in singapore. journal of transportation. engineering 124(3): 213-222. mortimer. (1957) moving vehicle method of estimating traffic volumes and speeds. highway research board bulletin 156. quiroga, c.a. & bullock, d. (1998) travel time studies with global positioning and geographic information systems: an integrated methodology. transportation research c 6: 101-127. robertson, h.d., hummer, j.e., & nelson, d.c. (1994) manual of transportation engineering studies. new jersey: prentice hall. salter, r.j. (1989) traffic engineering worked examples 2nd ed. london: macmillan education ltd. taylor w.c. & abdel-rahim a.s. (1998) final report on analysis of corridor delay under scats control (orchard lake road corridor), michigan state university. wardrop, j.g. & charlesworth, g. (1954) a method of estimating speed and flow of traffic from a moving vehicle. proceedings of the institution of civil engineers, part ii, 3: 158-171. wright, c. (1973) a theoretical analysis of the moving observer method. transportation research 7, trb, nrc: washington, dc: 293–311. © 2022 the jes. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 05_jes_274_11-05-2020 journal of engineering science 11(1), 2020, 43-60 study on sensitivity of microphysics for the simulation of rainfall for the month of may 2015 over bangladesh using high resolution wrf-arw model md. salman khan and m. m. alam department of physics, khulna university of engineering & technology, khulna 9203, bangladesh received: 15 march 2020 accepted: 11 may 2020 abstract in this research the advanced research wrf (arw) model v3.8.1 has been used to simulate the rainfall of may 2015 all over bangladesh. the model was configured in nested domain with 18 and 6 km horizontal grid spacing with 100 × 96 and 103 × 127 grids in the east-west and north-south directions, respectively with 30 vertical levels. the lin et al., wsm6, thomson, morrison double-moment (m-2mom), stony brook university (sbu), and wdm6 microphysics schemes coupling with kain-fritsch (kf) cumulus parameterization (cp) scheme have been used to simulate the monthly total rainfall, heavy rainfall, monthly rainy days and heavy rainy days for the month of may 2015 at all meteorological stations of bangladesh. the simulated results are compared with the observed results of 33 meteorological stations of bangladesh meteorological department (bmd) and precipitation estimation from remotely sensed information using artificial neural networks (persiann) output. relative standard deviation of all observed, persiann and model simulated parameters have been analyzed and compared. the maximum monthly observed rain in may 2015 at sylhet was 752 mm but wsm6, m-2mom and wdm6 schemes have simulated 831, 788 and 742 mm for day 1 prediction; wsm6, wdm6 and sbu-lin schemes have simulated 757, 916 and 981 mm for day 2 prediction and wsm6 and wdm6 schemes have simulated 741 and 925 mm for day 3 prediction, respectively and all other mps have simulated much higher rainfall at domain (d1). the wdm6, m-2mom and lin et al. schemes have simulated 744, 807 and 923 mm for day 1 prediction, wsm6 and wdm6 schemes have simulated 714 and 877 mm for day 2 predictions and wsm6, sbu-lin and lin et al. schemes have simulated 802 and 913 and 998 mm, respectively for day 3 predictions at domain (d2). the relative standard deviation (rsd) has minimum at d1 and d2 for wdm6 scheme for day 1 prediction and wsm6 scheme for day 2 and day 3 predictions for the monthly total rainfall and heavy rainfall of may 2015. the results suggest that as the forecast time increased the amount of total rain and also heavy rain is increased. as a result rsd is also increased for all mps. wdm6 scheme gives the better performance of rainfall and rainy days all over the country. keywords: pre-monsoon; persiann; relative standard deviation; microphysics; cumulus parameterization. 1. introduction the mean temperature of bangladesh during the summer months varies between 23-30oc. april and may are the hottest months (khatun et al., 2016). the maximum temperature of 36-40oc was attained in the southwestern (sw) and northwestern (nw) districts. when the maximum temperature goes above 36°c, heat wave situation occurs over bangladesh. due to heavy rainfall associated with severe thunderstorm in the northeastern (ne) part of bangladesh and adjoining ne states of india flash flood occurs in the ne part of bangladesh. only 19 % of the total annual rainfall occurs in this season. matsumoto (1997) studied the onset of rainy season by the index of precipitation over the asian summer monsoon region, and indicated that the assam region of india is earliest region where the onset of rainy season starts. ahasan et al. (2013) conducted research on simulation of high impact rainfall events over southeastern hilly region of bangladesh using mm5 model. the model suggests that the highly localized high impact rainfall was the result of an interaction of the mesoscale severe convective processes with the large scale active monsoon system. shahid (2010) studied rainfall variability and the trends of wet and dry periods in bangladesh. the result shows a significant increase in the average annual and pre-monsoon rainfall of bangladesh. alam (2013) studied the impact of cloud microphysics and cumulus parameterization schemes for the prediction of heavy rainfall event during 15-16 october 2007 over bangladesh. the study showed that the microphysical schemes have a major impact on time and location of rainfall intensity. all mps coupled with bmj schemes fails to simulate heavy rainfall in sitakundu-sandwip-chittagong region but they simulate heavy rainfall in the northnortheastern parts of the country, which does not match with the observation during 15–16 october 2007. alam (2014) have studied the impact of cloud microphysics and cumulus parameterization on simulation of heavy rainfall event during 7–9 october 2007 over bangladesh using 9 and 3 km nested domain. to examine the * corresponding author: malam@phy.kuet.ac.bd https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 44 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. sensitivity of the simulations of six different microphysical schemes and kain–fritsch (kf) and betts–miller– janjic (bmj) schemes were considered. cumulus parameterization (cp) schemes must estimate the rate of subgrid-scale convective precipitation, release of latent heat, and the distribution of heat, moisture, and momentum in the vertical due to convection (kain and fritsch, 1993). cumulus convection modifies the large-scale temperature and moisture fields through detrainment and cumulus-induced subsidence in the environment. the detrainment causes large-scale cooling and moistening, and the cumulus-induced subsidence causes large-scale warming and drying (arakawa and schubert, 1974). precipitation is recognized as one of the most difficult parameters to forecast in numerical weather prediction despite the fact that the accuracy of numerical models has increased during the past several decades (wang and seaman, 1997). prior studies have shown that a model’s microphysical parameterization scheme can strongly influence the magnitude of predicted precipitation (otkin et al., 2006). litta et al., (2012) illustrates that the microphysics scheme can significantly impact the accuracy of quantitative precipitation forecasts during the pre-monsoon season. fihir (2018) has simulated the heavy rainfall events in the southeastern regions of bangladesh during may 2013 using four different mps i.e., lin et al., wsm6, thompson and wdm6 and four different cps i.e., kf, tiedtke (td), zhang-mcfarlane (zm) and multi-scale kf (mskf). the research also suggests that wsm6 and wdm6 schemes coupling with zm and mskf schemes give the better performance on the basis of threat score, equivalent threat score and bias score during may 2013. haney et al. (2018) has been used to simulate the heavy rainfall event in the southeastern regions of bangladesh during 23-26 june 2015 using 12 different mps in wrf-arw model. the microphysics schemes are kessler, lin et al., wsm3, ferrier, wsm6, thomson graupel, mydm, m-2mom, cam v5.12-moment 5 class, sbu, wdm6 and nssl2. country averaged rainfall suggests that among 12 mps, the schemes thompson, sbu, wsm6 and wdm6 have simulated almost similar amount of rainfall as observed during the study period during 23-26 june 2015. the model has also simulated similar amounts of rainfall in the heavy rainfall(hr) area of 5 s i.e., in the se region by sbu, wdm6, lin, th and wsm6 schemes during 23-26 june 2015. saifullah et al. (2018) studied on the simulation of heavy rainfall event (hre) over the south and southeastern part of bangladesh due to monsoon depression using wrf model. their results suggested that, from the enormous area of the bay of bengal a large amount of moisture carried by a strong southwesterly flow towards the southeastern part of bangladesh and the adjoining area and so the hre over these regions might be characterized by the positive vorticity and strong vertical wind shear between 850 to 200 hpa. sumon and alam (2019) studied the impact on environmental moisture during the intensification and movement of tropical cyclone hudhud in the bay of bengal using wrf-arw model. they studied wsm6-class graupel, thomson graupel, wdm6-class graupel and nssl mom-1 microphysics schemes with four different initial conditions. their results suggests that the average track error was minimum for wdm6 schemes with the initial conditions at 7-9 october, which gives the better performance of wdm6 among all used mps. in the present study, wrf-arwv3.8.1 model has been used to simulate the total rainfall, heavy rainfall (hr), total rainy days and hr days for the month of may 2015 all over bangladesh. six mps schemes i.e. lin et al., wsm6, thomson, morrison double-moment (m-2mom), sbu and wdm6-class and kf scheme have been considered to study the monthly rainfall and rainy days of may 2015 and tried to identify the performances of different mp schemes. the primary objectives of this study are to examine which microphysics schemes are suitable for the prediction of monthly rain in the pre-monsoon season. the simulated results have been compared with the observed rainfall of bangladesh meteorological department (bmd) and persiann rainfall. 2. methodology 2.1 model domain and configuration the weather research and forecast (wrf-arw version 3.8.1) model has been used to simulate the premonsoon rainfall all over bangladesh. the model has different microphysics options but in this research 6 microphysics schemes are utilized for the simulation of daily rainfall for the month of may 2015. the model has used initial and lateral boundary conditions (lbcs) from ncep-fnl analysis at six hourly intervals. the model has been configured in double domain, 18 km and 6 km horizontal grid spacing with 103×127 and 100×96 grids in the east-west and north-south directions and 30 vertical levels. the double domain is used to identify, which domain produces more accurate monthly rainfall. time step of integration is set to 30 and 90 seconds for maintaining computational stability as the model uses third-order runge-kutta time integration scheme. the detail of the model and domain configuration is given in table 1. the model domain is given in figure 1 and different physics options are given in table 1: journal of engineering science 11(1), 2020, 43-60 45 table 1: wrf model and domain configurations dynamics non-hydrostatic number of domain 2 horizontal grid distance 6 km and 18 km integration time step 30 s and 90 s number of grid points x-direction 96 and 103 points, y-direction 100 and 127 points initial conditions three-dimensional real-data (fnl: 1° × 1°) microphysics (1) lin et al., (2) wsm6-class graupel, (3) thomson graupel, (4) morrison double-moment, (5) sbu and (6) wdm6-class radiation scheme dudhia (1989) for short wave radiation/ rrtm long wave mlawer et al. (1997) surface layer monin-obukhov similarity theory scheme (hong and pan, 1996) land surface parameterization 5 layer thermal diffusion scheme (ek et al., 2003) cumulus parameterization schemes kain-fritsch (kf) scheme, (kain and fritsch, 1990, 1993; kain, 2004) pbl parameterization yonsei university scheme (ysu) (hong et al., 2006) lin et al. scheme: lin et al. scheme has ice, snow and graupel processes, suitable for real-data high-resolution simulations. in this scheme six classes of hydrometeors are included: water vapor, cloud water, rain water, cloud ice, snow, and graupel. all parameterization production terms are based on lin et al. (1983). the scheme is taken from purdue cloud model and the details can be found in chen and sun (2002) 2-d microphysics scheme. this is one of the first schemes to parameterize snow, graupel, and mixed-phase processes and it includes ice sedimentation and time-split fall terms. it has been extensively used in research studies and in mesoscale nwp model. wrf single-moment 6-class (wsm6) microphysics scheme: this scheme predicts the mixing ratios for water vapor, cloud water, cloud ice, snow, rain, and graupel in six different arrays. the characteristics of the cold rain process in the wsm6 scheme follow the revised ice microphysics process (hong et al., 2004), whereas the warm rain processes are primarily based on the works of lin et al. (1983) and the auto conversion process from tropoli and cotton (1980). a new method for representing mixed-phase particle fall speeds for the snow and graupel processes suitable for high-resolution simulations by assigning a single fall speed to both sedimentation and accumulation processes is introduced. this method uses a large eddy simulation (les)-based approach (khairoutdinov and kogan, 2000) to determine the auto conversion rates and allow for a more sophisticated coupling between cloud field and number concentrations of warm species. double-moment prediction for the warm species in wsm6 scheme will allow more flexibility of the size distribution enabling the mean diameter to evolve in contrast to the one-moment scheme. thompson scheme: thompson scheme is a bulk microphysical parameterization scheme developed to use with wrf or other mesoscale models. the snow size distribution depends on both ice water content and temperature and is represented as a sum of exponential and gamma distribution functions. furthermore, snow assumes a nonspherical shape with a bulk density that varies inversely with diameter. a new scheme with ice, snow and graupel processes suitable for high-resolution simulations. morrison double-moment (m-2mom) scheme: the morrison scheme predicts the mass concentration of cloud water, cloud ice, rain, snow and graupel. the physics are based on the full double-moment version described in morrison et al. (2009). the auto conversion process from cloud ice to snow is triggered when cloud ice mass exceeds a specified threshold related to the maximum permitted size of cloud ice crystals, but it does not require a corresponding minimum size for the snow category. truncation at larger particle sizes would lead to a greater reduction as additional counts of small particles are eliminated. morrison scheme prediction of provided a good fit to aircraft estimates with a mean profile and range that is within the bulk of aircraft estimates below 4 km and also represented the general decrease from cloud top to cloud base associated with continued aggregation. stony brook university (sbu) scheme: stony brook university scheme is a 5-class scheme with riming intensity predicted to account for the mixed-phase processes. in this scheme the ice microphysics is presented, which considers both temperature and riming effects on ice properties. the five prognostic mixing ratios are water vapor, cloud ice, precipitating ice (pi), cloud liquid water, and rain. dry snow, rimed snow, and graupel are included in the pi category through the introduction of a varying riming intensity parameter. the new scheme allows for physically based representation of the ice particles with temperature and riming intensity– dependent properties, such as the mass, cross-sectional area, and fall velocity relationships. riming intensity is diagnosed from liquid water content (lwc), pi mass, and temperature. 46 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. wrf double-moment 6-class (wdm6) scheme: the wdm6 implements a double-moment bulk micro physical parameterization of clouds and precipitation and is applicable in mesoscale and general circulation models. the wdm6 scheme enables the investigation of the aerosol effects on cloud properties and precipitation processes with the prognostic variables of cloud condensation nuclei (ccn), cloud water and rain number concentrations. wdm6 extends the wrf single-moment 6-class microphysics scheme (wsm6) by incorporating the number concentrations for cloud and rainwater along with a prognostic variable of ccn number concentration. moreover, it predicts the mixing ratios of six water species similar to wsm6. prognostic water substance variables include water vapor, clouds, rain, ice, snow, and graupel for both the wdm6 and wsm6 schemes. additionally, the prognostic number concentrations of cloud and rain waters, together with the ccn, are considered in the wdm6 scheme. the number concentrations of ice species such as graupel, snow, and ice are diagnosed following the ice-phase microphysics of hong et al. (2004). 2.2 data final reanalysis (fnl) data (1o x1o) from national centre for environment prediction (ncep) is used as initial and lateral boundary conditions (lbcs) which is updated at six hourly interval i.e. the model will be initialized with 0000, 0600, 1200 and 1800 utc initial field of corresponding date. the ncep fnl data will be interpolated to the model horizontal and vertical grids. bangladesh meteorological department (bmd) observed rainfall and precipitation estimation from remotely sensed information using artificial neural networks (persiann) data will be used for verification. in the present study, the weather research and forecast (wrf-arw version 3.8.1) model has been used to simulate the rainfall of may 2015 over bangladesh. in this research, six different mp schemes and kain-fritsch cumulus parameterization scheme have been used to simulate the monthly rainfall of may 2015. 3-hourly rain gauge data of 33 meteorological stations have been collected from bangladesh meteorological department (bmd) all over bangladesh. in probability theory and statistics, the coefficient of variation (cv), also known as relative standard deviation (rsd) is a figure 1: wrf model domain for the prediction rainfall in bangladesh standardized measure of dispersion of a probability distribution or frequency distribution. relative standard deviation (rsd) = (standard deviation (sd) / mean) ×100= × rsd gives the variability of a parameter. if the value is less it is less variable and vice-versa. 3. results and discussion 3.1 observed rainfall and persiann satellite precipitation for the month of may 2015 special distribution of observed total rainfall and heavy rainfall of may 2015 all over bangladesh is presented in figures 2(a-b). the maximum rain is observed at sylhet (752 mm) and minimum rain at satkhira (16 mm). it is also seen from the spatial distribution pattern that the rainfall increased continuously from southwestern (sw) to northeastern (ne) region of bangladesh. the minimum rainfall is also observed at barishal, bhola, khepupara and patuakhali and is 72, 71, 39 and 38 mm respectively. the second maximum of rainfall is seen at dinajpur region (375 mm). from the spatial distribution pattern it is observed that 100 to 200 mm rainfall is seen in the central to western region of the country and dhaka gets 185 mm of rainfall during this month. from the distribution pattern, the maximum amount of hr is observed at sylhet (379 mm) and minimum in the southern region of bangladesh. the amount of monthly hr of 231 mm is also found at srimangal region. persiann total rainfall and heavy rainfall distribution of may 2015 all over bangladesh is presented in figure 2(c-d). the total rain (figure 2c) is observed maximum at sylhet (556 mm) and minimum at khepupara (48 mm). the persiann rainfall increased continuously from sw to ne and northwestern (nw) regions of bangladesh. the second and third maxima of rainfall are seen at srimangal (446 mm) and rangpur (376 mm) region of bangladesh. the distribution pattern is almost similar to that of bmd observed rainfall. from the distribution pattern, the highest hr (figure 2d) was found at srimangal and its amount is 213 mm. the persiann hr observed at dhaka was 56 mm at the same time the bmd observed rain was 185 mm, which is journal of engineering science 11(1), 2020, 43-60 47 much higher than that of persiann. the persiann distribution of hr indicated that maximum region of the country has no hr. the persiann hr is much lower than that of bmd observed hr. figure 2: (a-b) bmd observed and (c-d) persiann total rainfall and heavy rainfall respectively, (e-f) bmd observed and (g-h) persiann total rainy days and heavy rainy days respectively all over bangladesh for the month of may 2015 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (h) (g) (a) (b) (c) (d) (e) (f) 48 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. 3.2 bmd and persiann observed rainy days and heavy rainy days for the month of may 2015 special distribution of observed and persiann total rainy days and heavy rainy days for the month of may 2015 all over bangladesh is presented in figures 2(e-h). the maximum and minimum rainy days (figure 2e) are 27 and 5 days, respectively observed at sylhet and khepupara. it is seen from the spatial distribution pattern that the rainy days increased continuously from sw to n-ne region of bangladesh due to orography of ne hilly region. the significant number of rainy days is also observed at rangpur (20 days), dhaka (19 days) and mymensingh (19 days) region. the maximum and minimum numbers of persiann total rainy days of 28 and 10 (figure 2f) are observed at sylhet and khepupara, respectively. it is seen from the spatial distribution pattern that the rainy days increased continuously from sw to n-ne region of bangladesh. the maximum rainy days are also observed at bhola, rajshahi, rangamati, satkhira and sitakunda and are 19 days and madaripur and m.court are 18 days and ishwardi, kutubdia and patuakhali regions are 17 days and chittagong and jashore are 16 days. the maximum number of bmd observed hr days (figure 2g) is found at sylhet and no hr days are found at central to south, se, sw and western region of the country. it is also seen from the spatial distribution pattern that the hr days increased continuously from sw to ne region of bangladesh. the second peak of hr days are also observed at dinajpur. the maximum persiann distribution of hr days (figure 2h) is observed at srimangal (4 days) and no hr days are found at maximum region of the country. in the nw and se regions of bangladesh i.e., rangpur and rangamati observed only 2 hr days. 3.3 model simulated monthly total rainfall at d1 for day 1 prediction the spatial distributions of simulated monthly 24 hourly rainfalls in domain 1 (d1) of may 2015 for different mp schemes in combination with kf cp scheme with the initial conditions of 0000 utc of everyday for the month of may are presented in figures 3(a-f). lin et al. scheme (figure 3a) has simulated maximum rainfall at sylhet (861 mm) and minimum rainfall at bhola (45 mm). the distribution pattern shows that the rainfall increased continuously from sw to north-ne region of bangladesh. the second maximum simulated rainfall is at dinajpur region and the third maximum rainfall is at madaripur region. the minimum rainfall is simulated in the southeastern regions of bangladesh. the amounts of minimum rainfall are 113, 115, 133, 134 and 135 mm at rangamati, feni, sitakunda, teknaf and cox-bazar, respectively. the simulated rainfall at dhaka is 393 mm. the model simulated rainfall is matched with the observed rainfall in the northeastern region. wsm6 scheme (figure 3b) has simulated maximum rainfall of 831 mm at sylhet and minimum rainfall of 44 mm at bhola. the second maximum rainfall is seen at srimangal region of bangladesh. it is also seen from the spatial distribution pattern that the rainfall increased continuously from sw to ne region of bangladesh. the simulated rainfall at tangail is 430 mm. the minimum rainfall is seen at teknaf, chuadanga, rangamati, feni, cox-bazar and kutubdia and the amounts are 108, 112, 112, 125, 128 and 131 mm, respectively. the distribution pattern of simulated rainfall in the se and ne region is similar to that of observed rain but the wsm6 scheme has simulated much higher rainfall in the central to southern and western region of the country. thompson scheme (figure 3c) has simulated maximum rainfall of 939 mm at sylhet and minimum rainfall of 50 mm at bhola. from the spatial distribution pattern of rainfall it is found that the rainfall increased continuously from sw to ne region of bangladesh. in the central region of bangladesh i.e., dhaka-tangail regions the simulated rainfall is greater than 300 mm but the observed rain was less than 200 mm. the distribution pattern of simulated rainfall in the se region is similar to that of observed rain but all other region thompson scheme has simulated much higher rainfall. m-2mom scheme (figure 3d) has simulated maximum rainfall of 788 mm at sylhet and minimum rainfall of 33 mm at bhola. it is seen from the spatial distribution pattern that the rainfall increased continuously from south sw to ne region of bangladesh. the rainfall simulated at dhaka is 409 mm. the minimum rainfall also seen at chittagong and cox-bazar is 91 and 93 mm, respectively. the distribution pattern is similar but the scheme has simulated much higher rainfall in the central region. the distribution pattern of simulated rainfall in the se and ne region is similar to that of observed rain but m-2mom scheme has simulated much higher rainfall in the central to southern, west and northwestern (nw) region of the country. sbu-ylin scheme (figure 3e) has simulated maximum rainfall of 941 mm at sylhet and minimum rainfall of 40 mm at bhola. the spatial distribution pattern shows that the rainfall increased continuously from sw to ne and nw regions of bangladesh. the simulated rainfall at dhaka is 325 mm. the minimum rainfall is seen at chittagong, coxbazar and chandpur and are 86, 92 and 98 mm, respectively. the distribution pattern of simulated rainfall in the se region is similar to that of observed rain but all other region sbu-ylin scheme has simulated much higher rainfall. wdm6 scheme (figure 3f) has simulated maximum rainfall of 742 mm at sylhet and minimum rainfall of 46 mm at bhola. the spatial distribution pattern shows that the rainfall increased continuously from sw to ne region of bangladesh. the significant amounts of rainfall equal to 450 and 329 mm have also been simulated in the central and nw regions i.e., tangail and bogura, respectively. the minimum rainfall is also journal of engineering science 11(1), 2020, 43-60 49 seen in the south se regions of the country. the simulated rainfall is found to match with observed rainfall in the ne region i.e., highest rain regions. the errors of monthly total all bmd stations rainfall have simulated 41, 40, 31, 35, 22 and 40% for day 1 prediction; 47, 31, 34, 49, 27 and 38% for day 2 prediction and 42, 37, 29, 50, 28 and 45% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d1. figure 3: model simulated spatial distribution of monthly total rainfall of may 2015 for day 1 prediction in d1 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 3.4 model simulated monthly total rainfall at d2 for day 1 prediction the simulated monthly rainfall distributions of domain 2 (d2) for different mp schemes with the everyday initial conditions for the month of may 2015 are presented in figures 4(a-f). from the spatial distribution it is observed that all mp schemes have simulated maximum rainfall in the ne region and minimum rainfall in the south sw regions. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the highest rainfall at sylhet and the amounts are 923, 999, 1000, 807, 950 and 744 mm and bhola has the lowest rainfall of 33, 25, 47, 30, 22 and 58 mm, respectively and observed rainfall at sylhet is 752 mm and at bhola it is 71 mm. the lin et al. scheme has simulated significant amount of rainfall at dhaka and bogura having 440 and 315 mm (figure 4a), respectively. the minimum rainfall is also seen at teknaf and cox-bazar where the amounts are 107 and 106 mm, respectively. lin et al. scheme has simulated 16 to 18% higher rainfall all over the country in d2 for day 1 prediction. the wsm6 scheme has simulated significant amount of rainfall at dhaka and bogura and the amounts are 423 and 391 mm (figure 4b), respectively. the minimum rainfall of 74 mm is also seen at teknaf. wsm6 scheme has simulated 9 to 24 % higher rainfall all over the country in d2 for day 1 prediction. the thompson scheme has simulated significant amount of rainfall at dhaka and dinajpur where 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 50 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. the amounts are 412 and 310 mm (figure 4c), respectively. the minimum rainfall is also seen at teknaf and cox-bazar regions are 104 and 95 mm, respectively. thompson scheme has simulated 24 to 31% higher rainfall all over the country than that observed in d2 for day 1 prediction. the m-2mom scheme has simulated significant amount of rainfall of 446 and 317 mm at dhaka and dinajpur (figure 4d), respectively. the minimum rainfall of 109 and 82 mm is also seen in teknaf and cox-bazar, respectively. the sbu-ylin scheme has simulated significant amount of rainfall at dhaka and dinajpur where the amounts are 390 and 267 mm (figure 4e), respectively. the minimum rainfall of 108, 86 and 73 mm is also seen at teknaf, cox-bazar and chandpur, respectively. the sbu-ylin scheme has simulated 6 to 20 % higher rainfall all over the country than that observed rainfall in d2 for day 1 prediction. the wdm6 scheme has simulated significant amount of rainfall of 415 and 375 mm (figure 4f) at tangail and bogura, respectively. the minimum rainfall of 128, 116 and 90 mm is also seen at chandpur, mongla and teknaf regions, respectively. the rainfall simulated by wdm6 scheme is almost matched with the observed rainfall except in the central region where the scheme has simulated higher rainfall. the errors of monthly total all bmd stations rainfall have simulated 49, 46, 37, 38, 22 and 41% for day 1 prediction; 49, 42, 26, 48, 28 and 33% for day 2 prediction and 58, 41, 27, 61, 31 and 44% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d2. figure 4: model simulated spatial distribution of monthly total rainfall of may 2015 for day 1 prediction in d2 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 journal of engineering science 11(1), 2020, 43-60 51 3.5 model simulated monthly heavy rainfall at d1 for day 1 prediction the monthly distributions of model simulated heavy rainfall (hr) for day 1 in d1 for different mp schemes with the initial conditions of 0000 utc of everyday for the month of may 2015 are presented in figures 5(a-f). all studied mp schemes have simulated maximum hr in the ne region and no hr in the central to s-se, sw and nw regions of bangladesh. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the highest rainfall at sylhet and the amounts are 654, 499, 619, 523, 781 and 445 mm, respectively and observed rainfall at sylhet is 379 mm. wsm6, m-2mom and wdm6 schemes have also simulated significant amount of hr at bogura having100, 92 and 104 mm, respectively and the observed rain is 73 mm. the distribution pattern of hr for lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes are similar to that of observed hr all over the country except in the central and nw region. lin et al. (figure 5a), thompson (figure 5c) and sbu-ylin (figure 5e) schemes have not simulated hr in the nw region and wsm6 (figure 5b), m-2mom (figure 5d), and wdm6 (figure 5f) schemes have simulated higher hr in bogura but lower in dinajpur. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have also simulated similar rainfall at srimangal and the amounts are 204, 185, 383, 249, 270 and 135 mm, respectively and that of observed rainfall is 231 mm. the errors of monthly all bmd stations heavy rainfall have simulated 35, 39, 37, 34, 34 and 46% for day 1 prediction; 03, 14, 5, 15, 21 and 27% for day 2 prediction and 3, 31, 12, 13, 13 and 22% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d1. figure 5: model simulated spatial distribution of monthly hr of may 2015 for day 1 prediction in d1 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 52 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. 3.6 model simulated monthly total heavy rainfall at d2 for day 1 prediction the monthly distributions of model simulated heavy rainfall (hr) for day 1 in d1 for different mp schemes in combination with kf scheme with the initial conditions of 0000 utc of everyday for the month of may 2015 are presented in figures 6(a-f). all studied mp schemes have simulated maximum hr in the ne region and no hr in the s-se, sw and western regions of bangladesh. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the highest rainfall at sylhet and the amounts are 687, 681, 780, 503, 808 and 386 mm, respectively and the observed rainfall at sylhet is 379 mm. lin et al., wsm6, m-2mom, sbu-lin and wdm6 schemes have also been simulated significant amount of hr at bogura and it is 69, 138, 190, 80 and 123 mm, respectively and the observed rain is 73 mm. the distribution pattern of hr for lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes are similar to that of observed hr all over the country but all the studied schemes have simulated higher rain from central region to north-ne region. lin et al. (figure 6a), wsm6 (figure 6b), thompson (figure 6c), m-2mom (figure 6d), sbu-ylin (figure 6e) and wdm6 (figure 6f) schemes have also simulated similar rainfall at srimangal and the amounts are 251, 236, 362, 270, 256 and 214 mm, respectively and that of observed rainfall is 231 mm. all schemes have simulated lower rainfall at dinajpur region. lin et al., wsm6, thompson and sbu-ylin schemes have simulated higher rainfall and m-2mom and wdm6 schemes simulated lower rainfall at sylhet in d2 than that of d1. the errors of monthly all bmd stations heavy rainfall have simulated 7, 4, 15, 9, 25 and 35% for day 1 prediction; 8, 13, 5, 35, 14 and 26% for day 2 prediction and 44, 6, 0, 58, 8 and 11% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d2. figure 6: model simulated spatial distribution of monthly hr of may 2015 for day 1 prediction in d2 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 journal of engineering science 11(1), 2020, 43-60 53 3.7 simulated rainy days at d1 for day 1 prediction the monthly distributions of model simulated total rainy days for day 1 prediction in d1 for different mp schemes in combination with kf scheme with the initial conditions of 0000 utc of everyday for the month of may 2015 are presented in figures 7(a-f). all studied mp schemes have simulated highest number of rainy days at barishal-madaripur region and next higher number of rainy days have simulated at sylhet region. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the number of rainy days at sylhet having 26, 25 23, 27, 27 and 25 days, respectively and observed is 27 days and at barishal it is 28, 29, 28, 29, 28 and 29 days, respectively and observed is 12 days. the number of rainy days at sylhet is almost matched with the observed rainy days but in the south-se, sw and western region all the mp schemes have simulated much higher rainy days. lin et al. (figure 7a), wsm6 (figure 7b), m-2mom (figure 7d) and wdm6 (figure 7f) schemes have simulated minimum number of rainy days at bhola are 14, 12, 12 and 13, respectively and thompson (figure 7c) and sbu-ylin (figure 7e) schemes at jashore it is 13 days. it is also seen that lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated number of rainy days at sylhet having 28, 28, 29, 28, 28 and 28, respectively for day 2 (figure not shown) prediction and 28, 26, 28, 29, 28 and 27, respectively for day 3 (figure not shown) prediction. the distribution of simulated number of rainy days for all studied mp schemes in day 2 and day 3 (figure not shown) predictions are almost similar but slightly higher than that of day 1. the errors of monthly all bmd stations rainy days have simulated 90, 96, 83, 87, 83 and 92% for day 1 prediction; 94, 87, 92, 82, 84 and 91% for day 2 prediction and 83, 83, 74, 76, 73 and 84% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d1. figure 7: model simulated spatial distribution of monthly total rainy days of may 2015 for day 1 prediction in d1 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 54 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. 3.8 simulated rainy days at d2 for day 1 prediction the monthly distributions of model simulated total rainy days for day 1 prediction in d2 for different mp schemes in combination with kf scheme with the initial conditions of 0000 utc of everyday for the month of may 2015 are presented in figures 8(a-f). lin et al. (figure 8a), wsm6 (figure 8b), thompson (figure 8c), m-2mom (figure 8d), sbu-ylin (figure 8e) and wdm6 (figure 8f) schemes have simulated maximum rainy days in the central to ne regions and minimum in the south-sw regions of bangladesh. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the monthly highest number of rainy days at dhaka, madaripur, cumilla, sylhet, sylhet and sandwip having 27, 26, 25, 26, 26 and 26, respectively. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the number of rainy days at sylhet having 26, 25, 23, 26, 26 and 24, respectively and observed is 27 and at barishal it is 24, 23, 21, 21, 19 and 25, respectively and observed is 12 days. the number of rainy days simulated by different mp schemes at sylhet is almost matched with the observed rainy days but the schemes have simulated much higher rainy days in the south-se, sw and western region. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated minimum number of rainy days at bhola having 12, 10, 13, 10, 9 and 13, respectively for day 1 prediction and observed is 9 days. it is also seen that lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the number of rainy days at bhola having 9, 11, 14, 7, 9 and 12, respectively for day 2 (figure not shown) prediction and 11, 8, 10, 7, 8 and 8, respectively for day 3 (figure not shown) prediction. the distribution of simulated number of rainy days for all studied mp schemes in day 2 and day 3 predictions are almost similar but slightly lower than that of day 1.the errors of monthly all bmd stations rainy days have simulated 73, 73, 67, 69, 60 and 74% for day 1 prediction; 74, 74, 69, 60, 60 and 69% for day 2 prediction and 71, 69, 55, 60, 53 and 68% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d2. figure 8: model simulated spatial distribution of monthly total rainy days of may 2015 for day 1 prediction in d2 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 journal of engineering science 11(1), 2020, 43-60 55 3.9 simulated heavy rainy days at d1 for day 1 prediction the monthly distributions of model simulated hr days for day 1 prediction in d1 for different mp schemes in combination with kf scheme with the initial conditions of 0000 utc of everyday for the month of may 2015 are presented in figures 9(a-f). lin et al. (figure 9a), wsm6 (figure 9b), thompson (figure 9c), m-2mom (figure 9d), sbu-ylin (figure 9e) and wdm6 (figure 9f) schemes have simulated maximum hr days in the central to ne regions and almost zero all other regions of bangladesh. observed hr days is found to increase from central to ne and nw regions but almost all studied mp schemes have simulated hr days from central to ne regions of bangladesh. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the monthly highest number of hr days are 9, 6, 8, 8, 11 and 6, respectively at sylhet and next higher hr days 3, 3, 5, 4, 4 and 2 respectively at srimangal. the observed hr days are 5 and 4 at sylhet and srimangal respectively. wsm6 and wdm6 schemes have simulated almost similar hr days at sylhet but it was less in srimangal. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated number of hr days 10, 7, 11, 13, 9 and 7, respectively for day 2 (figure not shown) prediction and 8, 5, 8, 10, 9 and 7 days, respectively for day 3 (figure not shown) prediction at sylhet. the distribution of simulated number of rainy days for all studied mps in day 2 is higher and day 3 predictions are almost similar to that of day 1. the errors of monthly all bmd stations heavy rainy days have simulated 36, 45, 45, 33, 36 and 52% for day 1 prediction; 3, 9, 15, 3, 24 and 33% for day 2 prediction and 3, 27, 24, 9, 18 and 27% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d1. figure 9: model simulated spatial distribution of monthly hr days of may 2015 for day 1 prediction in d1 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 56 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. 3.10 simulated heavy rainy days at d2 for day 1 prediction the monthly distributions of model simulated hr days for day 1 prediction in d2 for different mp schemes in combination with kf scheme with the initial conditions of 0000 utc of everyday for the month of may 2015 are presented in figures 10(a-f). lin et al. (figure 10a), wsm6 (figure 10b), thompson (figure 10c), m-2mom (figure 10d), sbu-ylin (figure 10e) and wdm6 (figure 10f) schemes have simulated maximum hr days in the ne region and increased from central to north-ne and nw regions and almost zero in the southse and western regions of bangladesh. observed hr days is found to increase from central to north-ne and nw regions and almost similar pattern is simulated for all studied mp schemes. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated the monthly highest number of hr days having 9, 6, 9, 8, 12 and 4, respectively at sylhet and next higher hr days at srimangal having 4, 4, 5, 4, 5 and 3 respectively. the observed hr days are 5 and 4 at sylhet and srimangal respectively. lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated number of hr days 9, 5, 10, 11, 9 and 5, for day 2 (figure not shown) and 7, 6, 7, 12, 9 and 6, respectively for day 3 (figure not shown) prediction at sylhet. wsm6 and wdm6 schemes have simulated almost similar hr days at sylhet and srimangal and all other mps have simulated higher number of hr days for day 1, day 2 and day 3 predictions. similar hr days are also simulated at d2 and less at d1to that of observed. the errors of monthly all bmd stations heavy rainy days have simulated 12, 15, 24, 3, 21 and 45% for day 1 prediction; 6, 21, 3, 21, 9 and 36% for day 2 prediction and 24, 6, 15, 45, 3 and 3% for day 3 prediction by lin et al., wsm6, thompson, m-2mom, sbu and wdm6 schemes, respectively at d2. figure 10: model simulated spatial distribution of monthly hr days of may 2015 for day 1 prediction in d2 using a) lin, b) wsm6, c) thompson, d) m-2mom, e) sbu-ylin and f) wdm6 schemes coupling with kf scheme all over bangladesh. 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip 88 e 89 e 90 e 91 e 92 e 21 n 22 n 23 n 24 n 25 n 26 n barisalbhola bogra chandpur chuadanga comilla dhaka dinajpur faridpur feni hatiya ishurdi jessore madaripur m.court mymensingh rajshahi satkhira sitakunda srimangal sylhet tangail teknaf rangpur khulna mongla khepupara patuakhali chittagong cox's bazar kutubdia rangamatisandwip (a) lin et al. (b) wsm6 (c) thompson (d) m-2mom (e)sbu-ylin (f)wdm6 journal of engineering science 11(1), 2020, 43-60 57 3.11 relative standard deviation of rainfall and rainy days for the month of may 2015 the coefficient of variation (cv), also known as relative standard deviation (rsd), is a standardized measure of dispersion of a probability distribution or frequency distribution. rsd gives the variability of a parameter. if the value is less it is less variable and vice-versa. the rsd for monthly total rainfall and heavy rainfall of may 2015 is presented in figure 11(a-d). the highest and lowest values of rsd for day 1 prediction are found 78 & 59% for in d1 and 79 & 57% in d2 for thompson & wdm6 schemes respectively and for day 2 prediction 94 & 71% for thompson & wsm6 in d1 and 98 & 65% for sbu & wsm6 in d2 and for day 3 prediction 106 & 60% in d1 and 101 & 65% in d2 for thompson & wsm6 schemes, respectively. figure 11: rsd for all bangladesh meteorological stations monthly (a-b) total rainfall, (c-d) heavy rainfall, (e-f) total rainy days and (g-h) heavy rainy days in d1 and d2 respectively for different microphysics schemes of may 2015 (b) (a) (c) (f) (g) (d) (e) (h) 58 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. the rsd for day 1, day 2 and day 3 prediction in d1 & d2 is almost similar to that of observed rainfall. the observed sd of all bmd stations heavy rainfall is 84%. it is found that the rainfall is less variable in day 1 and it increases with the increase in forecasting period (day 2 and day 3) at domain 1 (figure 11a) and domain 2 (figure 11b). the highest and lowest values of rsd of hr is found for day 1 prediction 342 & 249 mm for sbu & m-2mom in d1 and 310 & 192 % for sbu& m-2mom in d2 and for day 2 prediction 288 & 162% for thompson & wsm6 in d1 and 231 & 167 mm for sbu & wsm6 in d2 and for day 3 prediction 277 & 166% for thompson & wsm6 in d1 and 234 & 149% for thompson & lin et al. schemes in d2. the variability of heavy rainfall for day 1 prediction is less variable for m-2mom and wdm6 schemes in d1 & also in d2, for day 2 prediction wsm6 in d1 & d2 and for day 3 prediction wsm6 and lin et al. schemes in d1 & d2 respectively. the observed rsd of all bmd stations heavy rainfall is 131%. it can be seen that in case of heavy rainfall, rsd decreases with the increase of forecasting period at domain 1 (figure 11c) and domain 2 (figure 11d). the rsd of total rainy days and hr days for the month of may 2015 is presented in table 3 and figure 11(eh). the rsd of total rainy days is found for day 1, day 2 and day 3 predictions are 13, 17 and 20% for wsm6 scheme, 14, 18 and 19% for lin et al. scheme, 15, 17 and 22% for thompson scheme, 14, 21 and 24% for m2mom scheme, 15, 20 and 21% for sbu scheme and 14, 17 and 17% for wdm6 scheme at d1. on the basis of the findings it is found that the wdm6 scheme is less variable than the other mp schemes in d1. the rsd of total rainy days is found for day 1, day 2 and day 3 predictions are 16, 20 and 20% for wsm6 scheme, 16, 26and 21% for lin et al. scheme, 15, 24 and 28% for thompson scheme, 16, 27 and 30% for m-2mom scheme, 20, 27 and 24% for sbu scheme and 13, 21 and 21% for wdm6 scheme at d2. on the basis of the findings it is found that the wdm6 scheme is less variable than the other mp schemes in d2 also. the observed rsd is found 39% for total rainy days all over bangladesh. it is found that the total rainy days is less variable in day 1 and it increases with the increase in forecasting period (day 2 and day 3) at domain 1 (figure 11e) and domain 2 (figure 11f). the rsd of heavy rainy days is found for day 1, day 2 and day 3 predictions are 229, 164 and 151% for wsm6 scheme, 269, 210 and 208% for lin et al. scheme, 242, 239 and 187% for thompson scheme, 242, 239 and 187% for m-2mom scheme, 316, 236 and 219% for sbu scheme and 262, 212 and 187% for wdm6 scheme at d1. on the basis of the findings it is found that the wsm6 scheme is less variable than the other mp schemes in d1. the rsd of heavy rainy days is found for day 1, day 2 and day 3 predictions are 164, 162 and 161% for wsm6 scheme, 181, 186 and 135% for lin et al. scheme, 248, 189 and 212% for thompson scheme, 187, 191 and 175% for m-2mom scheme, 284, 208 and 208% for sbu scheme and 191, 156 and 137% for wdm6 scheme at d2. on the basis of the findings it is found that the wdm6 scheme is less variable than the other mp schemes in d2 also. the observed rsd is found 123% for heavy rainy days all over bangladesh. the rsd decreases with the increase of forecasting period at domain 1 (figure 11g) and domain 2 (figure 11h). 4. conclusions on the basis of our findings the following the conclusions have been drawn:  the maximum monthly observed rain of may 2015 at sylhet is 752 mm but wsm6, m-2mom and wdm6 schemes have simulated rainfall of 831, 788 and 742 mm for day 1 prediction wsm6, wdm6 and sbu-lin schemes have simulated rainfall of 757, 916 and 981 mm for day 2 prediction whereas wsm6 and wdm6 schemes have simulated 741 and 925 mm for day 3 prediction, respectively and all other mps have simulated much higher rainfall at d1.  the wdm6, m-2mom and lin et al. schemes have simulated rainfall of 744, 807 and 923 mm for day 1 prediction whereas wsm6 and wdm6 schemes have simulated 714 and 877 mm for day 2 predictions and wsm6, sbu-lin and lin et al. schemes have simulated 802 and 913 and 998 mm, respectively for day 3 prediction at d2.  the maximum monthly observed hr of may 2015 at sylhet is 379 mm but wdm6, wsm6 and m-2mom have simulated 445, 499 and 523 mm for day 1 prediction, wsm6 and wdm6 schemes have simulated 398 & 558 mm and 364 & 619 mm for day 2 and day 3 predictions, respectively at d1.  the wdm6 and m-2mom schemes have simulated 383 and 503 mm hr at sylhet for day 1 prediction, wsm6 and wdm6 schemes have simulated 334 and 476 mm for day 2 predictions and wsm6 and sbulin schemes have simulated 507 and 614 mm for day 3 predictions at d2.  the simulated number of total rainy days at sylhet and bhola for all mps is almost matched with the observed total rainy days. all mps have simulated much higher total rainy days for day 1, day 2 and day 3 predictions with little exceptions all over the country for the month of may 2015. the number of observed rainy days in the south-southeastern regions is very few but different mps have simulated much higher rainy days in those regions. the distribution pattern of hr days for different mps is similar in the central to ne, s-se and sw regions but in the nw regions the number of hr days is insignificant. journal of engineering science 11(1), 2020, 43-60 59  lin et al., wsm6, thompson, m-2mom, sbu-ylin and wdm6 schemes have simulated much higher rain at domain 1 (308, 292, 257, 271, 264 and 310%) and at domain 2 (298, 273, 258, 259, 214, and 283%) at dhaka, faridpur, barishal, patuakhali, khepupara, madaripur, satkhira and tangail and all other stations have simulated 25, 24, 13, 21, 7 and 26% higher at domain 1 and 34, 31, 20, 25, 8 and 31% higher rainfall at domain 2 than that of observed rain at day 1.  the rsd is minimum at d1 and d2 for wdm6 scheme for day 1 prediction and wsm6 scheme for day 2 and day 3 predictions for the monthly total rainfall and total heavy rainfall of may 2015.  the rsd is minimum at d2 for wdm6 scheme for day 1, day 2 and day 3 predictions for the monthly total rainy days and total hr days of may 2015. on the basis of above discussion wdm6 scheme gives the better performance of rainfall and rainy days all over the country. references ahasan, m. n., chowdhury m. a. m., and quadir d. a., 2013. simulation of high impact rainfall events over southeastern hilly region of bangladesh using mm5 model, international journal of atmospheric sciences, id 657108, 13. alam, m. m., 2014. impact of cloud microphysics and cumulus parameterization on simulation of heavy rainfall event during 7–9 october 2007 over bangladesh, j. earth syst. sci., 123(2), 259–279 alam, m. m., 2013. impact of cloud microphysics and cumulus parameterization on simulation of heavy rainfall event during 15-16 october 2007 over bangladesh, journal of engineering science, 4(1), 95-109. arakawa, a., and schubert w. h., 1974. interaction of a cumulus cloud ensemble with the large-scale environment, part i: j. atmos. sci., 31, 674–701. chen, j. y., and sun y., 2002. hydrolysis of lignocellulosic materials for ethanol production, a review, bio resource technology, 83(1), 1-11. dudhia, j., 1989. numerical study of convection observed during the winter monsoon experiment using mesoscale two–dimensional models, j. atmos. sci., 46, 3077–3107. ek, m., mitchell k. e., lin y., rogers e., grunmann p., koren v., gayno g., and tarpley j. d., 2003. implementation of noah land-surface model advances in the ncep operational mesoscale eta model. j. geophys. res., 108, 8851, doi: 10.1029/2002jd003296 fihir, m. a, 2018. study on convective and non-convective rain of different heavy rainfall events in the premonsoon season using wrf-arw model, m.sc. thesis, dept. of physics, khulna university of engineering & technology. haney, s. m., alam m. m., and akhter md. a. e., 2018. sensitivity of microphysics for the simulation of heavy rainfall during 23-26 june 2015 over bangladesh using high resolution wrf-arw model, journal of engineering science, 9(2), 69-84. hong, s. y., noh y., and dudhia j., 2006. a new vertical diffusion package with an explicit treatment of entrainment processes. mon. wea. rev., 134, 2318–2341. hong, s. y., dudhia j., and chen s. h., 2004. a revised approach to ice microphysical processes for the bulk parameterization of clouds and precipitation, mon. wea. rev., 132, 103-120. hong, s.-y., and pan h.-l., 1996. nonlocal boundary layer vertical diffusion in a medium-range forecast model, mon. wea. rev., 124, 2322–2339 kain j. s., 2004: the kain-fritsch convective parameterization: an update, j. appl. meteor., 43, 170-181. kain, j. s., and fritsch j. m., 1993. convective parameterization for mesoscale models: the kain-fritsch scheme, the representation of cumulus convection in numerical models, meteo. monogr, no. 46 amer. meteor. soc., 165 – 170. kain j. s., and fritsch j. m., 1990. a one–dimensional entraining/detraining plume model and its application in convective parameterization. j. atmos. sci., 47, 2684–2702. khairoutdinov, m., and kogan y., 2000. a new cloud physics parameterization in a large-eddy simulation model of marine stratocumulus, monthly wea. rev., 128(1), 229-243. khatun, m.a., rashid m.b., and hygen h.o., 2016. climate of bangladesh, met report no. 08/2016. lin, y.l., farley r. d., and orville h. d., 1983. bulk parameterization of the snow field in a cloud model, j. climate appl. meteor., 22, 1065-1092. litta, a. j., mohanty u. c., and idicula s. m., 2012. the diagnosis of severe thunderstorms with high-resolution wrf model, j. earth syst. sci., 121(2), 297–316. matsumoto, j., 1997. seasonal transition of summer rainy season over indochina and adjacent monsoon region, adv. atmos. sci., 14, 231—245. mlawer, e. j., taubman s. j., brown p. d., lacono m. j., and clough s. a., 1997. radiative transfer for inhomogeneous atmosphere: rrtm, a validated correlated–k model for the longwave, j. geophys. res., 102(d14), 16663–16682. 60 md. salman khan and m. m. alam study on sensitivity of microphysics for the ….. morrison, h., thompson g., and tatarskii v., 2009. impact of cloud microphysics on the development of trailing stratiform precipitation in a simulated squall line: comparison of one-and two-moment schemes, monthly wea. rev., 137(3), 991-1007. otkin, j.a., huang, h.l., and seifert a., 2006. a comparison of microphysical schemes in the wrf model during a severe weather event, preprints, 7th annual wrf user’s workshop, boulder, co., usa. saifullah, mallik m. a. k., alam md. s., and syed i. m., 2018.simulation of heavy rainfall event over the south and southeastern part of bangladesh due to monsoon depression using wrf model, journal of engineering science, 9(2), 23-32. shahid, s., 2010. rainfall variability and the trends of wet and dry periods in bangladesh, international journal of climatology, 30, 2299–2313. sumon k. d., and alam m. m., 2019.study the impact on environmental moisture during the intensification and movement of tropical cyclone hudhud in the bay of bengal using wrf-arw model, journal of engineering science, 10(1), 31-43. tripoli, g. j., and cotton w. r., 1980. a numerical investigation of several factors contributing to the observed variable intensity of deep convection over south florida, journal of applied meteorology 19(9), 1037-1063. wang, w., and seaman n. l., 1997. a comparison study of convective parameterization schemes in a mesoscale model, mon. wea. rev., 125, 252–278. microsoft word 01_jes_e4167_10_20_2020 journal of engineering science 12(1), 2021, 1-8 doi: https://doi.org/10.3329/jes.v12i1.53095 meteorological influences on urban air quality parameters in dhaka city nafisa islam*1, md. golam saroar2 and tanvir ahmed1 1 department of civil engineering, bangladesh university of engineering and technology, dhaka, bangladesh 2 clean air and sustainable environment project, department of environment, dhaka, bangladesh received: 07 may 2019 accepted: 10 november 2020 abstract this study aims at investigating the effect of meteorological parameters on seasonal variation of particulate matter(pm) (both pm2.5 and pm10) using a 4-year (2013-2016) monitoring data of air quality parameters from case project implemented by the department of environment (doe). using monthly data of the continuous air monitoring station(cams) of darus-salam, dhaka, cross correlation analysisis performed between pm and meteorological parameters where inverse relationships of pm with temperature, rainfall and relative humidity are found. increased biomass burning during low temperature period, washout effect of rainfall, wet deposition mechanism of higher humidity may be held responsible for these negative correlations. significant seasonal variation is observed from daily data analysis of darus salam station and it is found that winter pm concentrations are 4.5-5.5 times higher than monsoon pm concentrations. seasonal cross-correlation between pm10 and pm2.5 shows lower correlation during winter (december-february) and monsoon (june-september) seasons. two possible effects can attribute to this seasonal difference: i) presence of biomass burning during winter which increases pm2.5 and ii) presence of rainfall during monsoon which decreases pm10.pm2.5/pm10 ratios for different months indicate the contrasting influences of different mechanisms on different sized pm particles. pm2.5/pm10 ratio is found to be higher during december-february and lower during marchseptember with a rise in august, which indicates the effect of 3 mechanisms: i) dilution effect of wind speed on pm2.5 during december-february, ii) re-suspension effect of wind speed on pm10 during march-september and iii) more pronounced scavenging effect of rainfall on pm10 during august. the study indicates the need for properly accounting the influence of meteorology for better understanding of pm variation in urban areas in bangladesh. keywords: cross correlation; meteorology; pm; pm2.5/pm10 ratio; seasonal variation. 1. introduction particulate matter (pm) is defined as a complex mixture of different sizes of airborne particle having different chemical compositions. they are mainly classified in two categories, the finer particles ranging from 0.005 µm to 2.5 µm which is called pm2.5 and the coarser particles with aerodynamic diameter ≤ 10 µm which is called pm10. like other countries, particulate matter (pm) in ambient air has become one of the major concerns in bangladesh. according to the global air report 2017, dhaka city has become 2nd most air polluted city (hei, 2017). pm concentration in the air has been found to have significant correlation with diseases such as chronic respiratory illness, cardiovascular morbidity etc. (dockery et al., 1993). to fully understand the process responsible for this distribution of particulate matters, analysis of the meteorological condition and detailed study on their influence on pm concentration are required. different studies have shown that particulate matters are highly dependent on specific meteorological parameters (dayan and levy, 2005). it has been reported that wind speed, precipitation, relative humidity, temperature, time of day, atmospheric stability etc. are the major factors to drive the pm10 concentration in germany (gietl and klemm, 2009). several studies have been performed to evaluate the extent of urban pollution in the major cities of bangladesh (begum et al., 2013). although the relationship between meteorology and pm has been investigated, very little information is available on the dependence of urban aerosol on atmospheric parameter in the major cities of bangladesh. in this study, attempt is taken to determine the inherent relation between pm and meteorological parameters in dhaka city using common statistical techniques. the aim is to obtain a deeper understanding of the process involved in the variation of pm concentration over time. 2. methods and data archiving 2.1 data collection under the clean air and sustainable environment (case) project, the department of environment (doe) monitors real-time pm10 (24hr), pm2.5 (24hr) as well as ambient temperature (1hr), rainfall (1hr), relative jes an international journal *corresponding author: nafisa.nikita@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) 2 nafisa islam et al. meteorological influences on urban air quality ……….. humidity (1hr), through 11 continuous air monitoring stations (cams) throughout bangladesh. air quality and meteorological data of cams-3 (darus salam, dhaka) and cams-8 (red crescent campus, sylhet) for the year of 2013-2016 are collected. however, the data of cams-3 (darus salam, dhaka) is used for the analysis. 2.2 approach for analysis single linear regression model is used to quantify the correlation between pm2.5 and pm10 with meteorological parameters. the regression equation is in the form of y = β 0 + β 1 x+ ε (1) here, y is the concentration of pm2.5 or pm10, x is the meteorological parameters (temperature, rainfall, relative humidity and solar radiation), β is the regression coefficients and ε is the error term, where ε= (yi-ŷ), yi =observed y values, ȳ = mean value of series y and ŷ= y values given by the equation. the coefficient of determination r2 measures how related the pm concentration is with response to these meteorological parameters. r2 = st-sr sr (2) here, st is total sum of squares, where st =total sum of squares = yi − ȳ n i=1 2 (3) and, sr is error sum of squares, where sr = error sum of squares = yi − ŷ ² n i=1 (4) r2 continues to increase with increasing terms to the model, which can deter the goodness of fit. hence, adjusted r2 is introduced, which modifies the r2 based on the added terms to the model: adjusted 𝑟 = 1 − (1 − 𝑟 ) ∗ (5) here, m is the number of elements in a series and p is the number of independent variables. besides, cross correlation analysis is performed between pm and meteorological parameters. since the pm variation and meteorological parameters both vary with time, time dynamic analysis of cross correlation would be the best way to represent the actual relationship between pm and weather parameters through a lead-lag relationship. basic cross correlation formula used in the analysis is written below: for 𝑘 ≥ 0, 𝐶 = ∑ [{𝑥(𝑡) − �̅�} ∗ {𝑦(𝑡 + 𝑘) − 𝑦}] (6) for 𝑘 ≤ 0, 𝐶 = ∑ [{𝑦(𝑡) − 𝑦} ∗ {𝑥(𝑡 − 𝑘) − �̅�}] (7) here, x(t) is the concentration of pm at time t, y(t+k) is the respective meteorological parameter at time (t+k), k is the lag between two-time series x and y, t is the total number of elements in series x and y. in order to standardize the correlation values, the cross-correlation coefficient is calculated which is given by: rxy(k) = cxy(k) sx*sy (8) here, sx = cxx(0) (9) and, sy = cyy (0) (10) 3. results and discussions 3.1 average pm concentration and meteorological conditions in bangladesh in bangladesh, the year can be divided into four different seasons: winter (december-february), pre-monsoon (march-may), monsoon (june-september), post-monsoon (october-november) (begum et al., 2014). the climate of bangladesh experiences prominent variation in weather during different seasons. it endures cold and dry air in winter as well as hot and humid air during the other three seasons. however, high temperature and high humidity is observed for most of the year. precipitation shows marked distinction between seasons, maximum rainfall occurs in the monsoon and a minimum in winter. during winter, dry soil condition, scanty rainfall and low relative humidity prevails. during pre-monsoon, rainfall becomes moderately strong and relative humidity increases. during monsoon, moist air condition and high relative humidity prevails. besides, the amount of rainfall also remains at its highest during this season. in the post monsoon, the amount of precipitation starts to decrease and so as the relative humidity. journal of engineering science 12(1), 2021, 1-8 3 figure 1: daily average pm concentration with corresponding bangladesh national ambient air quality standard (bnaaqs) for (a) pm2.5 and (b) pm10 figure 1(a) and (b) show the daily 24hr average concentration for pm2.5 and pm10, respectively, spanning for the year of 2013-2016. the average pm2.5 and pm10 concentrations during the study period are found to be 91.03µg/m3 and 161.69 µg/m3, respectively. to understand the pm variation throughout the entire year, seasonal and annual mean are calculated for the entire study period and the results are shown in table 1. winter pm concentration is found to be considerably higher than the bangladesh national ambient air quality pm2.5 standard of 65 µg/m3 (daily 24hr average) and pm10 standard of 150 µg/m3 (daily 24hr average). for the year of 2013-2016, respectively 172, 194, 174 and 161 daily pm concentrations, corresponding 48, 53, 48 and 44% of the sampling days exceeded the bnaaqs pm2.5 limit value. similarly, for this four-year period, respectively 42, 44, 40 and 43% of the sampling days exceeded the bnaaqs pm10 limit value. from the above statistics, it is evident that pm2.5 concentration is more prone to exceed the limit value compared to pm10. performing the analysis on seasonal basis, the exceedance is found to be highest for winter season (99.45% for pm2.5 and 95.85% for pm10) and lowest for monsoon season (4.5% for pm2.5 and 1.23% for pm10), while exceedance during other seasons are moderate. significant monthly variation has been obtained for both pm fractions. the winter to monsoon ratio of pm2.5 and pm10 concentration during 2013-2016 were 6.09, 5.56, 5.04 and 6.2 as well as 4.22, 4.54, 4.11 and 4.95, respectively (table 1). comparing with other studies, our observations of the difference between pm concentration of winter and monsoon season have been found very high. for example, the winter to monsoon pm ratio has been found to be 2.9 for pm10 and 2.2 for pm2.5 in india (kulshrestha et al., 2009) whereas winter to summer ratio of 2.14 forpm10 has been found in egypt (elminir, 2005). this may be because, during winter, higher atmospheric stability as well as dry weather condition favors suspension of particulate matter in the air. along with it, brick kilns in bangladesh remain operational during this season. aerosol concentration in monsoon was minimum due to scavenging effect of precipitation and the higher winter to monsoon ratio for pm10 indicates that this scavenging effect is more pronounced on pm10 compared to pm2.5. 3.2 seasonal pm concentration prevalence figure 2 shows frequency distribution of pm2.5/pm10 ratio for all seasons which is divided into nine categories starting from <0.2 to <1. here, during high pm prevailing season i.e. in winter, pm2.5/pm10 ratio curve shows a symmetric pattern with its peak at 0.6-0.7 (above 40% of the cases). it indicates that, this ratio fits normal distribution during winter and the contribution of pm2.5 remains higher. similarly, during low pm prevailing season i.e. in monsoon, symmetric pattern is also observed with peak at 0.4-0.5 (above 35% of the cases), which indicates that this distribution too follows normal distribution. however, slightly right skewed distribution is observed for pre-monsoon and post-monsoon season with peak at 0.4-0.5, which indicates that during these seasons, contribution of pm2.5 concentration starts to decrease after winter. during pre-monsoon, monsoon and post-monsoon season, approximately 40, 35 and 45%, respectively are observed to be in the range of 0.4-0.5 4 nafisa islam et al. meteorological influences on urban air quality ……….. whereas it is only 5% during winter. during winter highest ratio is in the segment between 0.6-0.7, however fewer value 8, 12 and 16% are found for pre-monsoon, monsoon and post-monsoon, respectively. high pm2.5/pm10 ratio during winter indicates that significant portion of pollution particles fall under the size distribution of pm2.5. table 1: average pm concentrations, their seasonal ratio and annual exceedance from bangladesh national ambient air quality standard (bnaaqs) year pm2.5 average concentration pm10 average concentration pm2.5 win/ mon. ratio pm10 win/ mon. ratio annual pm2.5 exceedance from bnaaqs (%) annual pm10 exceedance from bnaaqs (%) 2013 winter 187.41 ± 57.37 277.24 ± 84.43 6.09 4.22 47.12 41.87 pre-monsoon 78.96 ± 52.17 155.22 ± 96.12 monsoon 30.77 ± 14.68 65.68 ± 28.22 post-monsoon 78.69 ± 40.88 136.94 ± 80.05 annual 89.61 ± 73.62 152.31 ± 108.65 2014 winter 168.98 ± 47.93 260.55 ± 68.95 5.56 4.54 53.15 43.53 pre-monsoon 89.28 ± 52.55 175.58 ± 89.83 monsoon 30.42 ± 11.55 57.34 ± 20.35 post-monsoon 117.76 ± 69.83 181.72 ± 117.00 annual 95.11 ± 70.50 159.63 ± 108.53 2015 winter 173.48 ± 46.14 258.45 ± 76.65 5.04 4.11 47.67 40.5 pre-monsoon 72.89 ± 43.28 139.16 ± 81.85 monsoon 34.39 ± 16.11 62.92 ± 30.21 post-monsoon 93.32 ± 46.09 169.04 ± 85.18 annual 88.32 ± 65.34 148.13 ± 100.82 2016 winter 188.28 ± 34.84 301.96 ± 76.72 6.20 4.95 44.11 43.01 pre-monsoon 69.70 ± 43.59 151.52 ± 77.20 monsoon 30.38 ± 14.91 61.06 ± 28.68 post-monsoon 79.14 ± 48.10 148.32 ± 78.37 annual 87.88 ± 70.43 158.55 ± 112.23 figure 2: frequency distribution of pm2.5/pm10 ratio figure 3: annual variation of pm2.5/pm10 for the years of 2013, 2014, 2015 and 2016 3.3 meteorological parameters influencing pm levels as the dispersion condition of atmosphere is primarily responsible for the accumulation of pm particle in air, the primary focus is on the role of temperature, relative humidity and precipitation for the variation of pm levels. the results of the correlation analysis between meteorological parameters and different sized pm journal of engineering science 12(1), 2021, 1-8 5 particles are shown in table 2. a significance level of 5% (p=0.05) has been chosen to be the threshold for determining the significance of correlation analysis. table 2 shows that, dominant meteorological parameters those drive pm around dhaka city are temperature (adj. r2= 0.681 for pm2.5; adj. r2=0.566 for pm10) and relative humidity (adj. r2= 0.244 for pm2.5; adj. r2= 0.413 for pm10). solar radiation (adj. r2= 0.232 for pm2.5; adj. r2=0.131 for pm10) and rainfall (adj. r2= 0.141 for pm2.5; adj. r2=0.157 for pm10) are observed to exert weak influence over pm variation. temperature is found to have strong negative relation with particulate matters (rx-corr= -0.825 at lag 0 for pm2.5; rx-corr= -0.752 at lag 0for pm10). this represents the fact that, with increase of temperature, pm decreases and vice versa. when temperature becomes lower i.e. during winter season, formation of stagnant air condition occurs and simultaneous biomass burning activities increases in the kilns, which gives rise to the particulate matter concentration. the correlation coefficient between air pollutants and relative humidity is found to be significant (rx-corr= -0.494 at lag 0 for pm2.5; rx-corr= -0.643 at lag 0 for pm10). high humidity indicates higher precipitation events with in cloud scavenging, reduction in the formation of oc (organic carbon) and ec (elementary carbon), higher moisture absorption and subsequent settling down of particles, all of which eventually result in low concentration of particulate matters. relative humidity has been found to have slightly higher correlation with coarser particle. this is because the wash out effect of humidity and precipitation is more profound for the case of coarser particles. negative correlation is observed between solar radiation and pm (rx-corr= -0.482 at lag 0 for pm2.5; rx-corr= -0.362 at lag 0 for pm10). this relation might indicate the phenomena that, increase of incident solar radiation leads to surface warming which cause rise of boundary layer height (blh). when blh increases, pm gets more space for dispersion. higher dispersion results in decrease of pm concentration in the ambient air. table 2: cross correlation and single linear regression analysis results between pm concentrations and meteorological variables predict-ion variable (cams 3) using monthly data (n=48) cross correlation single linear regression (slr) r(lag) r² adjusted r² error sum of square pm2.5 temperature -0.825(0)** 0.681** 0.674** 57763 rainfall -0.376(0)* 0.141 0.123 155600 humidity -0.494(0)* 0.244* 0.228* 136990 solar radiation -0.482(0)* 0.232 0.215 139210 pm10 temperature -0.752(0)** 0.566** 0.556** 178320 rainfall -0.396(0)* 0.157 0.138 346130 humidity -0.643(0)** 0.413* 0.401* 240850 solar radiation -0.362(0)* 0.131 0.112 356600 statistical significance indicators are as follows: **, p < 0.001; *, 0.01>p > 0.001, otherwise 0.05>p>0.01 table 3: annual cross correlation coefficients between pm and other meteorological parameters (rainfall & temperature) at cams-3 and cams-8 meteorological parameters cams no. pm2.5 pm10 n a correlation coefficients n a correlation coefficients temperature cams-3(dhaka) 48 -0.825** 48 -0.752** cams-8 (sylhet) 47 -0.855** 48 -0.862** rainfall cams-3 (dhaka) 48 -0.340* 48 -0.396* cams-8 (sylhet) 47 -0.731** 47 -0.732** a cross correlation coefficients at zero lag. statistical significance indicators are as follows: **, p < 0.001; *, 0.01>p > 0.001, otherwise 0.05>p>0.01 correlation analysis between pm concentration and rainfall is also carried out where weak negative correlation is found between them (rx-corr= -0.376 at lag 0 for pm2.5; rx-corr= -0.396 at lag 0for pm10). however, a strong negative correlation has been reported between average monthly pm and rainfall in sylhet (cams-8) (table 3). (r= -0.731 at lag 0 for pm2.5; r= -0.732 at lag 0 for pm10) (table 3). in dhaka, amount of rainfall is low compared to sylhet where consistent rainfall is a distinct feature (in northeastern part of bangladesh). in 6 nafisa islam et al. meteorological influences on urban air quality ……….. bangladesh, rainfall is caused by the influence of the southwest monsoon (hossain et al., 2013). total precipitation in dhaka in 2016 is 462.02 mm with highest rainfall recorded in july (157.74 mm) and total precipitation in sylhet in 2016 is 1606.8 mm with the highest rainfall recorded in the month of april (567.4 mm). therefore, lower correlation in dhaka may occur due to the long interval of consistent rainfall occurring in this area. this leads to the conclusion that, rainfall amount and duration both contributes combinedly in pm fluctuation. like rh, higher correlation is observed between pm10 and rainfall which indicates that scavenging effect is more effective on pm10 than on pm2.5. 3.4 global comparison of correlation coefficients a comparative analysis has been conducted between this study and other literature values, based on the calculated correlation coefficients for pm2.5 and pm10 with meteorological parameters and is presented in table 4. from table 4, it is evident that, our results show similarity with the analyses conducted in india, turkey and egypt, whereas contradictory relation has been observed for studies conducted in greece, germany and usa. negative correlation for temperature with pm has been observed for bangladesh, india, turkey and egypt whereas, positive correlation has been obtained for greece, spain, germany and usa. this variation mainly occurs due to the difference in weather condition and chemical composition of particulate matters all over the world. weather patterns are similar for bangladesh, india and turkey, since all of them fall under the subtropical region. biomass burning activities during winter season contribute to the higher concentration of pm. during low temperature period, particulate matter concentration becomes high and thus inverse relationship is formulated between pm and temperature. however, considering usa, greece and germany, high temperature is favorable for atmospheric chemical reaction. hence, secondary particle formation is favored by temperature increase which produces positive correlation between pm and temperature. table 4: pearson’s correlation coefficients between pm10 and meteorological parameters in different regions reference country tempera -ture relative humidity wind speed precipitation pm2.5 (galindo et al., 2011) spain(yearly) -0.016 0.048 -0.496 - (tai et al., 2010) usa(yearly) 0.4-0.7 (-0.1)-(-0.15) (south) (-0.05)-1.0 - 0.05-0.14 (north) (bhaskar, and mehta, 2010) india(yearly) -0.64 --0.53 0.01(northern) 0.74(southern) (akyuz, and cabuk, 2009) turkey(winter) -0.324 -0.108 -0.350 - (chaloulakou et al., 2003) greece(winter) 0.46 --0.54 - -bangladesh -0.681 -0.244 --0.141 pm10 (elminir, 2005) egypt(yearly) -0.48 0.252 -- (galindo et al., 2011) spain(yearly) 0.601 0.189 -0.334 - (hien et al., 2002) germany(yearly) 0.17 -0.15 -0.49 -0.38 (akyuz, and cabuk, 2009) turkey (winter) -0.155 -0.237 -0.409 - (bhaskar, and mehta, 2010) india(yearly) -0.34 -0.44 -0.17 -0.53 (chaloulakou et al., 2003) greece(yearly) 0.39 --0.43 - -bangladesh -0.566 -0.413 --0.157 considering relative humidity (rh), high humidity condition leads to higher moisture absorption and subsequent settling down of particles in the subtropical region. therefore, negative correlation occurs for rh with pm in bangladesh, india and turkey, whereas around the european countries i.e. in usa, greece and germany, ultrafine particulate formation is found to be positively affected in the presence of high humidity, thus resulting in positive correlations between particulate matter and relative humidity. 3.5 relationship between different pm size fractions the relationship between fine particles pm2.5 and coarse particle pm10 is studied using cross correlation coefficients. the data are divided into four seasons for this analysis. the results are presented in the table 5. winter and monsoon correlation coefficients have been found to be lower than the other seasons. these journal of engineering science 12(1), 2021, 1-8 7 differences between the coefficients are due to meteorological conditions those drive the pm concentration to change. it is linked to the seasonal changes of the weather conditions. during winter, there are major natural sources of pm, such as biomass burning activities, fossil fuel burning form vehicles and burning of agricultural soil and clays in the brick kilns which operate mainly during winter. these activities produce significant number of fine particles i.e. pm2.5. therefore, the contribution of fine particles becomes much higher in winter. besides, during monsoon season, significant reduction in pm10 occurs by the wet deposition mechanism of continuous precipitation. thus, increase in pm2.5 in winter and decrease in pm10 in monsoon leads to the reduction of correlation coefficients between pm2.5 and pm10 during these seasons. during other seasons, the contribution of abovementioned sources, which mainly enriches pm2.5, becomes low. as a result, the obtained coefficients are higher for other seasons compared to winter. annual pm2.5/pm10 variability is examined from 2013 to 2016 and it is found that the ratio remains low during the period of march to september i.e. pre-monsoon-monsoon season (figure 3). during this period, wind speed remains at its maximum which induces the re-suspension effect of pm10. therefore, the concentration of pm10 becomes higher in the atmosphere compared to pm2.5. due to these reasons, the pm2.5/pm10 ratio remains low during march-september. besides, a distinct rise of pm2.5/pm10 ratio can be observed in august. during august, rainfall remains at its maximum, which induces scavenging mechanism that works more effectively on pm10. as a result, average concentration of pm10 drops more compared to pm2.5, resulting in a distinct rise in pm2.5/pm10 ratio. therefore, this figure is indicative of 3 mechanisms: i) dilution-effect of wind speed (ws) on pm2.5 during december-february, ii) re-suspension effect of ws on pm10 during march-september, and iii) more pronounced scavenging-effect of rainfall on pm10 during august. table 5: cross correlation analysis results between pm2.5 and pm10 concentrations pm10 winter premonsoon monsoon postmonsoon all data pm2.5 winter 0.792(0)** pre-monsoon 0.922(0)** monsoon 0.812(0)** post-monsoon 0.954(0)** all data 0.944(0)** statistical significance indicators are as follows: **, p < 0.001; *, 0.01>p > 0.001, otherwise 0.05> p>0.01 4. conclusion the influence of meteorological parameters on seasonal variation of particulate matter (pm) is examined using a 4-year (2013-2016) monitoring data of air quality parameters. using monthly-data of the continuous air monitoring station (cams) of darus salam, dhaka, cross-correlation and pearson’s correlation analysis are performed between pm and meteorological parameters. significant seasonal variation is observed and it has been found that winter pm concentrations are 4.5-5.5 times higher than monsoon pm concentrations. major meteorological parameters that control pm in the air of dhaka city are – temperature and relative humidity. inverse relationships of pm with temperature, rainfall and relative humidity have been found. increased biomass burning during low temperature period, washout effect of rainfall, dry deposition effect of higher humidity may be held responsible for these negative correlations. besides, comparison of correlations between rainfall and pm for cams-3 (dhaka) and cams-8 (sylhet) indicates that, rainfall duration along with rainfall amount play major role to dominate pm. dry deposition and scavenging effect are more effective to drive coarser particles (pm10) compared to finer particles (pm2.5). acknowledgements the authors would like to thank the department of environment (doe) for providing the background air quality and meteorological data. references akyuz, m., and cabuk h., 2009. meteorological variations of pm2.5/pm10 concentrations and particleassociated polycyclic aromatic hydrocarbons in the atmospheric environment of zonguldak,turkey, j. hazard. mater., 170, 13-21. begum, b. a., hopke p. k., and markwitz a., 2013. air pollution by fine particulate matter in bangladesh, atmos. pollut. res., 4(1), 75-86. begum, b. a., saroar g., nasiruddin m., randal s., sivertsen b., and hopke p., 2014. particulate matter and black carbon monitoring at urban environment in bangladesh, nucl. sci. appl., 23(1 & 2). 8 nafisa islam et al. meteorological influences on urban air quality ……….. bhaskar, b.v., and mehta v. m., 2010. atmospheric particulate pollutants and their relationship with meteorology in ahmedabad, aerosol air qual. res., 10, 301-315. chaloulakou, a., kassomenos p., spyrellis n., demokritou p., and koutrakis p., 2003. measurements of pm10 and pm2.5 particle concentrations in athens, greece, atmos. environ., 37(5), 649-660. dayan, u., and levy i., 2005. the influence of meteorological conditions and atmospheric circulation types on pm10 and visibility in tel aviv, j. appl. meteorol. 44(5), 606-619. dockery, d., schwartz j., and spengler j., 1993. air pollution and daily mortality: associations with particulates and acid aerosols, environ. res., 59, 362-373. elminir, h. k., 2005. dependence of urban air pollutants on meteorology, sci. total environ. 350(1), 225-237. galindo, n., varea m., gil-moltó j., yubero e., and nicolás j., 2011. the influence of meteorology on particulate matter concentrations at an urban mediterranean location, water air soil poll. 215(1), 365372. gietl, j. k., and klemm o., 2009. analysis of traffic and meteorology on airborne particulate matter in munster, northwest germany, j. air waste manage. 59(7), 809-818. hei, 2017. state of global air 2019: special report, health effect institute, institute of health metrics and evaluation (ihme), boston, ma. hien, p. d., bac v. t., tham h. c., nhan d. d., and vinh l. d., 2002. influence of meteorological conditions on pm2.5 and pm2.5−10 concentrations during the monsoon season in hanoi, vietnam, atmos. environ. 36(21), 3473-3484. hossain, m. a., hossain m. s., bhuyan m. a., karmakar r., and hossain m. m., 2013. annual flood report 2013. flood forecasting & warning centre, processing & flood forecasting circle, bangladesh water development board, ministry of water resources, dhaka. kulshrestha, a., satsangi g., masih j., and taneja a., 2009. metal concentration of pm2.5 and pm10 particles and seasonal variations in urban and rural environment of agra, india, sci. total environ., 407(24), 6196-6204. tai, a., mickley l., and jacob d., 2010. correlations between fine particulate matter (pm2.5) and meteorological variables in the united states: implications for the sensitivity of pm2.5 to climate change, atmos. environ., 44(32), 3976-3984. weather underground, https://www.wunderground.com. last accessesd on 17 dec. 2017. © 2021 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 06_jes_262_11-05-2020 journal of engineering science 11(1), 2020, 61-65 the effect of zno nano particle coating and finishing process on the antibacterial property of cotton fabrics farhana momotaz*, ayesha siddika, md. tashrif shaihan and md. anisul islam department of textile engineering, khulna university of engineering & technology, khulna-9203, bangladesh received: 06 february 2020 accepted: 13 may 2020 abstract in this study, the zno nanoparticles solution was directly applied on to the 100% cotton knit & woven fabric to impart antibacterial property using both spin coater & pad-dry-cure method. then, the disc diffusion method was used to assess the antibacterial activity of the finished fabrics. the topographical analysis of untreated, treated, and washed fabrics of different structures (knit and woven) were studied and compared. the results showed that the finished fabric demonstrated significant antibacterial activity against s. aureus and e. coli. as per result, it was also found that double jersey (rib) fabric showed highest amount of bacterial protection & pad-dry-cure method showed better result against bacterial attack than spin coating method. moreover, gram negative bacteria (e. coli) showed more strength against antibacterial treated unwashed and washed fabrics than gram positive bacteria (s. aureus). keywords: antimicrobial property; nano-particle; pad-dry-cure method; spin coater 1. introduction the nano-world is the intermediary between the atom and the solid, from the large molecule or the small solid object to the strong relationship between surface and volume (nouailhat, 2008). the term nano originated from the greek nanos which suggests ‘dwarf’. it is one billionth of a meter (gazit and mitraki, 2007). nanotechnology is the science that deals with matter at the scale of 1 billionth of a meter. it is also the study of manipulating matter at the molecular and atomic scale. nanoparticle, the most fundamental component in the fabrication of a nanostructure is far smaller than the world of everyday objects, and is described by newton’s laws of motion. however, it is bigger than an atom or a simple molecule that are governed by quantum mechanics (horikoshi and serpone, 2013). when conventional materials formed from nanoparticles, many of its properties changed. nanoparticles are reactive and effective than other molecules because of having a greater surface area per weight than larger particles. nowadays, a very promising scientific research is nanoparticle research because of the wide range of potential and promising applications especially in electronic fields, optical, and biomedical. this new concept changes the use of such material in microscale size to be used in absolutely new and advanced applications by using a nanoscale size of the same material (hashim, 2012). the structure of the nanomaterials can be classified by their dimensions. the zero-dimensional nanostructures are called nanoparticles (alagarasi, 2011). recently, one-dimensional nanostructures with different morphologies (such as nanowires, nano-rods (nrs), and nanotubes) have become the focus of intensive research, because of their unique properties with potential applications. among them, zinc oxide (zno) nanomaterials has been found to be highly attractive, because of the remarkable potential for applications in many different areas such as solar cells, sensors, piezoelectric devices, photodiode devices, sun screens, anti-reflection coatings, and photo catalysis (hatamie et al., 2015). zinc oxide nanoparticles (zno nps), as one of the most important metal oxide nanoparticles, are popularly employed in various fields due to their peculiar physical and chemical properties (smijs and pavel, 2011) (ruszkiewicz et al., 2017). zinc oxide nanoparticles (zno nps) are used in an increasing number of industrial products such as rubber, paint, coating, and cosmetics (jinhuan et al., 2018). in addition, zno nps have superior visible light resistance, antibacterial deodorant, antibacterial, antimicrobial, and excellent uv-blocking properties (hatamie et al., 2015). due to their potential for reducing the transmission of infection in medical and health care environments, antimicrobial textiles have attracted a great interest in recent years. in addition, in these fabrics, the formation of unpleasant odors is reduced. therefore, it is possible to have a pleasant and fresh smell from those textiles still after use. therefore, antimicrobial products have long been used in health care environments for hospital gowns, patient clothes, curtain, bed cover, infection control, wound healing, hygiene, etc. the effectiveness of antimicrobial action on textiles depends on the antimicrobial, the concentration, and the application method of the antimicrobial on the textile (islam and butola, 2019). microbes are very small lives on the earth which can’t be seen by the naked eyes. they may be known as microorganisms like bacteria, fungi, algae, and viruses, etc. these bacteria are sub classified into different groups * corresponding author: farhanatex@yahoo.com; https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 62 farhana momotaz et al. the effect of zno nano particle coating and finishing process….. namely gram positive (staphylococcus aureus), gram negative (escherichia coli), spore bearing or non-spore bearing types. apart from that some are pathogenic and cause cross infection bacteria whereas some microbes like fungi, molds or mildews are shown slower growth rate due to its complex nature (textile news, 2019). various studies have already been done on zinc oxide nanoparticles as antibacterial substance by many researchers. the ability of the antibacterial agent to inhibit bacterial growth was first tested using a disc diffusion method (singh et al., 2012). zno nanoparticles treated fabric shows higher antibacterial activity when compared with zno bulk treated fabrics whereas the untreated fabrics showed no antibacterial activity (rajendran et al., 2010). it has also found that the zno nanoparticles constitute an effective antimicrobial agent against pathogenic microorganisms (gunalan et al., 2013). a zinc oxide nanoparticle is used as an antibacterial substance against e. coli and s. aureus and exhibits the highest toxicity against microorganisms. it has also been demonstrated from sem and tem images that zinc oxide nanoparticles first damage the bacterial cell wall, then penetrate, and finally accumulate in the cell membrane (khwaja et al., 2018). in this study, two different methods have been used to apply the zno nanoparticles named spin coating and paddry-method. spin coating is known as a procedure used to deposit uniform skinny films to flat substrates. in this coating system, very few amount of coating material is applied on the center of the substrate, which is either spinning at low speed or not spinning at all. the substrate is then rotated at high speed in order to spread the coating material by centrifugal force. a machine used for spin coating is called a spin coater, or simply spinner. pad-dry-cure or exhaust-dry-cure is a finishing process applied to textiles to impart different finish treatments, such as waterproofing, softening, antibacterial or anti-odor finishes. a water-based solution bath contains finishing chemicals through which the textile materials pass. the textile is then dried and cured using heat and pressure. this is known as the parent process for various finishing treatments. then antibacterial analysis of the treated and untreated samples was carried out by disc diffusion method (hossain et al., 2014). to determine the effective ness of the finishing process and coating, the diameter of the zone of inhibition produced by the coated samples has been measured and compared with each other. 2. methodology 2.1 raw materials three types of fabric i.e. single jersey (100% cotton, gsm 180), double jersey (100% cotton, 1x1 rib), woven fabric (100% cotton, 1/1 plain weave) have been used for this research. zno nanoparticles (in powder form), acrylic binder, sodium lauryl sulphate, and soap have been used for the treatment. 2.2 preparation of fabric at first, the grey fabrics were taken. then scouring and bleaching were done in all fabrics with required recipe. this rfd fabric of woven, single jersey & double jersey were ready for following process. 2.3 preparation of nanoparticles solution 2% zno solution & 1% acrylic binder solution were taken. both mixed together with the help of stirrer. for fine and uniform mixing, the solution was stirred in ultra-sonicator at 30 ˚c temperature for 20 minutes. 2.4 application of nanoparticles solution in spin coating method, according to fabric weight, the amount of mixed zno & binder solution was applied on circular shaped fabric through one by one drop by using a syringe. the fabric was then rotated at 2000 rpm in order to spread the fluid by centrifugal force. later, the fabric sample was cured at 140˚c for 3 minutes. in pad-dry-cure method, the mixed solution prepared by the ultra-sonicator was taken in a beaker. then fabric samples were soaked in zno & acrylic binder mixed solution for 5 minutes. later, the treated fabric samples were padded by using padder machine. finally, samples were cured at 140˚c for 3 min. 2.5 washing of samples both spin and pad-dry-cured fabrics were washed once with sodium lauryl sulphate 2 times to eliminate the unfix particles from the treated fabric samples. iso c01 method was used for washing the samples. some spin and paddry-cured samples were washed 3 times. each time washing was done at 40˚c for 30 minutes in washing machine by 0.5% soap according to iso c01 test method. later the fabrics were air dried. journal of engineering science 11(1), 2020, 61-65 63 2.6 antibacterial test 36 samples using 12 media plates were tested in disc diffusion method to find antibacterial activity. the sequence of working procedure is given below:  bacteria were sub-cultured from stock by keeping bacteria in incubator for overnight.  nutrient broth and agar media were prepared according to the requirements.  nutrient broth media were prepared at 2 conical flasks (20 ml/conical flask) & 300 ml agar media were prepared in a separate conical flask.  the nutrient broth containing conical flasks were incubated for 24 hours.  for serial dilution purpose 8 dewater test tubes (9 ml/test tube) were taken. all test tubes were closed by foil paper or cotton plug.  2 bacteria’s colony was added to the nutrient broth containing 2 conical flasks media and incubation is done in it for 24 hours.  after that, bacteria containing nutrient broth conical flasks were serial diluted up to 104 times.  on 12 circular plates, firstly agar solution was added and when they turn in semi solid form, 1 ml of bacteria solution was added on each plate respectively & spread the solution with the help of spreader.  in this way, three types (single jersey, double jersey/rib, woven) of fabrics were used. untreated, treated, and 3 times washed fabrics of each type and each method were tested by observing and conditioning in incubator overnight at 37 ˚c temperature.  finally, the result was observed by measuring the area around the fabric which is getting resistance from bacterial attack. figure 1: sample fabrics after incubation overnight at 37˚c temperature. 3. results and disscussion table 1 represents the results of three types of fabric (woven, single jersey, rib/double jersey) treated with zno np by pad-dry-cure method and then applied by s. aureus and e. coli bacteria. it showed the antibacterial activity of untreated samples, and zno nanoparticle coated samples on e. coli and s. aureus bacteria respectively. here, after overnight incubation the resisted area of eighteen samples is found where twelve samples: four woven, four single jersey and four double jersey were treated with zno nanoparticles in pad-dry-method whereas two woven, two single jersey and two double jersey were untreated samples. among these twelve treated samples, six samples were given three times wash. the area of the sample fabric was 352 sq. mm for each. antibacterial activity results showed that the zno nanoparticle has exhibited strong antibacterial activity against both gram-positive (s. aureus) and gram-negative (e. coli) bacteria. however, antibacterial activity of zno particles were greater on grampositive than gram-negative bacteria. it can be assumed that the outer thick peptidoglycan layer and other surface components of gram-positive bacteria may promote zno attachment onto the cell wall whereas the components of gram-negative bacteria may repeal this attachment (getie et al., 2017). in addition, among three samples (woven, single jersey and double jersey/rib) it is found that the treated double jersey has greater antibacterial activity on gram-positive (s. aureus) bacteria. generally, the gram per square meter (g.s.m) of double jersey fabric is higher than the single jersey fabric. it is assumed that, double jersey fabric absorbs high amount of zno np from the solution than single jersey fabric which subsequently shows higher antibacterial ability. after overnight incubation there was a clear visible zone in which the bacteria could not survive and that zone was created around the fabric having an area of 720 sq. mm whereas the untreated double jersey showed zero resisted area. so, it can be said that the treated fabric without wash showed the best anti-biotic properties. and there was a significant reduction in anti-microbial property after three time washing and the non-treated fabric exhibits no resistance at all. 64 farhana momotaz et al. the effect of zno nano particle coating and finishing process….. the comparison of the three types of fabric treated with pad-dry-cure method after applying s. aureus and e. coli bacteria are presented in figure 2. from the chart it is clearly understood that the double jersey/rib fabric showed the maximum anti-bacterial property. and the anti-bacterial treatment resists the s. aureus bacteria better than that of e. coli bacteria. but overall it is being seen that the knitted fabric is slightly better than woven fabric in terms of antibacterial property. it is assumed that this happens due to the looping structure of knitted fabric. table 1: evaluation of anti-bacterial activity in pad-dry-cure method sl no. method of coating fabric type bacteria area of fabric (mm2) resisted area of net resisted area (mm2) untreated sample (mm2) treated sample (mm2) treated sample after 3 times wash (mm2) 1 pad-dry-cure woven s. aureus 352 0 475 408 123 2 woven e. coli 352 0 408 368 56 3 single jersey s. aureus 352 0 567 520 215 4 single jersey e. coli 352 0 475 414 123 5 rib s. aureus 352 0 720 616 368 6 rib e. coli 352 0 450 408 98 figure 2: antibacterial resistance of different fabrics using pad-dry-cure method. figure 3: antibacterial resistance of different fabrics using spin coater likewise table 1, the results of three types of fabric treated with spin coating method after applying s. aureus and e. coli bacteria are presented in the table 2. the area of the sample fabric was 352 sq. mm for each. in the result it is seen that the treated fabric without wash showed the best anti-biotic properties. and there was a significant reduction in anti-microbial property after three time washing and the non-treated fabric exhibits no resistance at all. table 2: evaluation of anti-bacterial activity in spin coating method sl no. method of coating fabric type bacteria area of fabric (mm2) resisted area of net resisted area (mm2) untreated sample (mm2) treated sample (mm2) treated sample after 3 times wash (mm2) 1 spin woven s. aureus 352 0 567 475 215 2 woven e. coli 352 0 414 374 62 3 single jersey s. aureus 352 0 520 432 168 4 single jersey e. coli 352 0 432 414 80 5 rib s. aureus 352 0 621 500 269 6 rib e. coli 352 0 432 391 80 the comparison of the three types of fabric treated with spin coating method after applying s. aureus and e. coli bacteria are presented in the figure 3. from the chart, it is clearly understood that the double jersey/rib fabric showed the maximum anti-bacterial property providing a resisted area of 621 sq. mm. and the anti-bacterial treatment resists the s. aureus bacteria better than that of e. coli bacteria. again, it is being seen that the knitted fabric is slightly better than woven fabric in terms of antibacterial property. however, rib/double jersey fabric treated in pad-dry-method gave 720 sq. mm area which is much higher than spin coating method. in pad-dry-cure method, the sample is immersed into solution which may provide more journal of engineering science 11(1), 2020, 61-65 65 fixations of chemicals into the samples than spin coating method. hence, pad-dry-cure method may have better performance than spin coating. 4. conclusions from this study, it concludes that double jersey fabric showed highest protection against bacteria than single jersey and woven fabric. moreover, pad-dry-cure method gave good antibacterial protection than spin finish. there is more scope to test others structures of fabric like silk, polyester etc. for finding out the best antibacterial coated cloth. although, spin coating takes less time and saves chemical solution uptake than pad-dry-cure. but setting up a spin coater for antibacterial finish is a costlier matter than pad-dry-cure method. moreover, antibacterial activity of zno particles were greater on gram-positive than gram-negative bacteria. therefore, antibacterial treatment resists the s. aureus bacteria better than that of e. coli bacteria. the antimicrobial properties of these fabrics will allow additional protection to them from bio deterioration during storage, transportation and consumption of goods and provide new opportunities to use in medical application. references alagarasi, a., 2011. introduction to nanomaterials in nanomaterials, edition 1, volume 7. gazit, e., and mitraki a., 2007. plenty of room for biology at the bottom: an introduction to bio nanotechnology, london: imperial college press, 105. getie, s., belay a., reddy a.r.c., and belay, z., 2017. synthesis and characterizations of zinc oxide nanoparticles for antibacterial applications, journal of nanomedicine & nanotechnology, 8(5), 1-8. gunalan, s., sivaraj r., and rajendran v., 2013. green synthesized zno nanoparticles against bacterial and fungal pathogens, progress in natural science, materials international, 22(6), 693-700. hashim, a. a., 2012. smart nanoparticles technology, croatia: in tech janeza trdine, 9. hatamie, a., khan a., golabi m., turner a.p.f., beni v., mak w.c., sadollahkhani a., alnoor h., zargar b., azam s., nur o., and willander m., 2015. zinc oxide nanostructure-modified textile and its application to biosensing, photocatalysis, and as antibacterial material, langmuir, 31(39), 10913– 10921. hossain, m. b., salam m. a., yousuf m. a., and rashed a., 2014. characterization and antibacterial studies of the mixed ligands complexes of pd(ii) ion with phthalic acid and heterocylic amines, journal of engineering science, 5(1), 47-53. islam, s., and butola b.s., 2019. the impact and prospects of green chemistry for textile technology, isbn: 978-0-08-102491-1, uk: woodhead publishing, 281. jinhuan, j., jiang p., and jiye c., 2018. the advancing of zinc oxide nanoparticles for biomedical applications, 3, 1-18. khwaja, s.s., azizur r., tajuddin and azamal h., 2018. zinc oxide nanoparticles and their activity against microbes, nanoscale research letters, 13(1), 1-13. nouailhat, a., 2008. an introduction to nanoscience and nanotechnology, united states: john wiley & sons, pp 4. rajendran, r., balakumar c., ahammed, h.a.m., jayakumar, s., vaideki, k. and rajesh, e.m., 2010. use of zinc oxide nano particles for production of antimicrobial textiles, international journal of engineering, science and technology, 2(1), 202-208. ruszkiewicz, j., pinkas a., ferrer b., peres t., tsatsakis a., and aschner m., 2017. neurotoxic effect of active ingredients in sunscreen products, a contemporary review, toxicology reports, 4, 245–259. horikoshi, s., and serpone n., 2013. microwaves in nanoparticle synthesis: fundamentals and applications, 1st edition, usa, wiley & sons, 1-24. singh, g., joyce e.m., beddow j., and mason t.j., 2012. evaluation of antibacterial activity of zno nano particles coated sonochemically onto textile fabrics, the journal of microbiology, biotechnology and food sciences, 2(1), 106-120. smijs, t.g., and pavel s., 2011. titanium dioxide and zinc oxide nanoparticles in sunscreens: focus on their safety and effectiveness, nanotechnology, science and applications, 4, 95–112. textile news, apparel news, rmg news, fashion trends, 2017. antimicrobial finishes for textile materials, available at: https://www.textiletoday.com.bd/antimicrobial-finishes-textile-materials/ (accessed 5 august 2019). microsoft word 08_jes_270_15-05-2020 journal of engineering science 11(1), 2020, 83-91 forecasting the average temperature rise in bangladesh: a time series analysis sneha paul1* and shuvendu roy2 1dept. of urban and regional planning, khulna university of engineering & technology, khulna-9203 2dept. of computer science and engineering, khulna university of engineering & technology, khulna-9203 received: 03 april 2020 accepted: 15 may 2020 abstract global warming has caused a significant increment in surface temperature around the world, including bangladesh. in this study, the temperature data of bangladesh over the past 100 years has been analyzed to see the temperature increment pattern. it has been seen that the average temperature has risen by 10c over the last century. using daily average temperature data of bangladesh, machine learning-based time series forecasting model has been developed to predict the future temperature of bangladesh. the model can predict the minimum, maximum, and average temperatures of any year in the future. this has been treated as a regression problem and linear, polynomial, and support vector regression have been proposed to build the prediction model. the proposed model has a mean square error of 0.00470c which is a good margin for such a model. using the model, the average temperature of bangladesh is predicted over the next hundred years. keywords: global warming; climate change; average temperature rise; prediction; bangladesh. 1. introduction global warming indicates the rise of the mean temperature of the earth's surface. although global warming and climate change are often used as a synonym, global warming is the effect and climate change is the reason (venkataramanan & smitha, 2011). over the last 100 years, the average air temperature near the earth’s surface has risen by a little less than 1 degree celsius (orr, 2007). the rise of the surface temperature of earth plays a vital role in global warming. other factors affecting global warming are the increase of co2 in the atmosphere, burning of fossil fuel, deforestation, extraction of greenhouse gases from the air conditioner, refrigerators, etc. these pollutants absorb and trap the sunlight and solar radiation heating the earth's surface, as such the average temperature of earth increases. climatic phenomena have enhanced due to global warming. sea is getting warmer along with sea-level rise (lindzen, 2008). oceans have warmed and a strong decline has occurred in arctic sea ice. both natural causes and man-made activities are responsible for global warming. among the various natural causes, most influential are volcanic eruptions and internal fluctuations in the climate system. but human activities such as: cutting off trees, burning fossil fuels, using motorized vehicles, refrigerators, air conditioners ejecting co2 and greenhouse gasses, etc. are major responsible for global warming. as human activities are major suspected responsible for climate change and global warming, intergovernmental panel on climate change (ipcc) was established by the world meteorological organization and the united nations environment programme to investigate the factors working behind climate change and global warming (unfccc, 2005). for producing the assessment of ipcc, hundreds of scientists around the world are working to develop advanced mathematical modeling that will be able to predict future changes, monitor historical and present variation in climate (houghton, 2004). as more research is doing on climate change, the occurrence of global warming and its impacts are becoming more and more absolute and clear to the scientists. according to the fourth assessment report of the ipcc, since the mid-20th century the rise in greenhouse gas concentrations is liable to increase in global average temperature. according to the observation of the past 50 years, the external forces (forces that are outside human influence) are also responsible for the widespread warming of the atmosphere and ocean, together with ice mass loss (alley, 2007). global warming affects various aspects of different regions in the world with different geographical contexts. according to tol (2006), countries with higher latitudes such as canada, russia, and scandinavia, climate change results in the rise of temperature by 2 or 3°c which may lead to net advantages, for example, increased agricultural outturn, decreased winter fatality, decreased heating necessity, and a probable uplift to tourism (zholudeva, 2019). along with the positive factors, negative factors also grasp these regions. for example, rapid growth of rates of warming; damage of infrastructure, human health, local establishments, and biodiversity (shahzad, 2015). indirect effects of global warming are seen such as an increase in mortality rate due to hot weather among the old and poor urban dwellers (unfccc, 2005). for example, in shanghai, china will face a rise in the diurnal average mortality (35-63 extra additional deaths per day) due to high average temperature and humidity, while french has faced 15,000 additional deaths due to heat wave in summer 2003 (tol, 2006). bangladesh is situated geographically in the tropical region (islam, 2004). due to its geographical location, * corresponding author: snehapaul@urp.kuet.ac.bd https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 84 sneha paul and shuvendu roy forecasting the average temperature rise in bangladesh….. natural disaster and climate change are general phenomena (daily star, 2019). the geographical features are responsible for making this country vulnerable to natural disasters. the challenges faced by this country due to climate change are elevated due to the enormous level of poverty (icddr b, 2019). according to the research report of action aid, bangladesh ranked fifth most vulnerable country to climate change and hunger. in the world risk index 2011, bangladesh ranked sixth among the most vulnerable countries to natural disasters, while second among the asian countries (daily star, 2019). bangladesh is a low energy-consuming country. it contributes only 0.3% of the amount of emission which contributes to global warming. but the effects of global warming make bangladesh one of the worst victims (elahi, 2015). the average temperature of 30 years in bangladesh is 27.5°c and the average summer temperature in monsoon is relatively higher than the summer (mahmood, 2012). researchers have done various studies and developed mathematical models to predict average surface temperature which portrays global warming and climate change. two observational analysis combining land surface temperature analysis and sea surface temperature analysis is done and a canadian global climate model (cancm4) established. this model incorporates observed anthropogenic concentrations of greenhouse gases, emissions of aerosol precursors, and naturally occurring forcing to the year 2005 (fyfe, 2011). to predict twentieth-century global average temperature changes, analysis is made on three different climate models with very diverse sensitivities and aerosol forcing (al, 2006). a real-time multi-model is also used to see the decadal change in climate (smith, 2012). according to the instrumental record of global surface temperature which began in 1850, eleven years among the last twelve years (1995–2006) has been ranked as the hottest years (alley et al., 2007). further evidence has also been seen that global warming is deteriorating weather and the environment in several ways causing various problems: between the years 1993 and 2003, sea level has been elevated due to melting polar ice caps at a rate of 3.1 mm every year (pender, 2008). furthermore, droughts have been increased in some areas, and rain has become heavier in others due to the change in patterns of rainfall. according to the fourth assessment report of ipcc, it has been expected that global warming will be increased at a rate of 0.2˚c per decade from the next twenty years. it has been predicted that the rise of average global temperature will be between a 1.8˚c and 4˚c by the year 2100 (lindzen, 2008). although the magnitude of temperature rise solely depends on the actions that are taken presently to stop greenhouse gas pollution. however, if the present trends of pollution are not being changed, a 1 3°c rise of temperature is being expected in average global temperature during the next fifty years (tol, 2006). in this study, the past 100 years' temperature rise data of bangladesh has been analyzed to find out the temperature rise trend for this region. three models: linear regression model, polynomial regression model, support vector model (svr) are developed to predict the average temperature rise in bangladesh for the next 100 years. 2. methodology this section describes the methodologies used to build the prediction model to forecast the temperature of bangladesh in the future. the problem is treated as a regression task and experimented with different algorithms to build the best model for this application. following algorithm is used in the experiments: (a) linear regression (b) polynomial regression (c) support vector regression (svr) 2.1 linear regression the linear regression model is implemented as a baseline for predicting the temperature in the future. in this study, the input variable is the year in which a single number and the output variable is the temperature, which is a real number. representing the input variable year as x and the predicted variable temperature as y, the linear regression is expressed as: = + here, and are the coefficients of the linear regression model. the model learns by minimizing the least square errors. the model is never perfect and there is a difference between the predicted value and the actual value. the error is represented as: = − ŷ least square error is squared of the error. the squared error is considered instead of the raw error because the prediction from the model can be above or below the true value, which will cause a positive and negative error which might cancel out when a train in batch. also, the square will penalize the large difference more than that of small differences. so, the model will try to make small mistakes rather than a huge difference between the actual value and the prediction. journal of engineering science 11(1), 2020, 83-91 85 figure 1 explains the linear fit of a dataset. the red dots are the data points and the blue line is the linear fit line to model the dataset. the coefficients of the linear regression model can be calculated with the following equations: = ∑ ( − ̅)( − )∑ ( − ̅ ) = − ̅ experiments are done with different types of predictions with all the models of consideration. we have modeled the mean, minimum, and maximum temperature of every year. as the temperature of consecutive years does not differ too much, we have also experimented with a 10-year average temperature model. figure 1: linear regression fit to a dataset figure 2: comparison of linear regression and polynomial regressions. figure 3: support vector machine figure 4: support vector regression model 2.2 polynomial regression polynomial regression comes into play when the data points are not correlated, which means there does not exist a linear relationship between the independent variable and the predicted variable. polynomial regression (ruppert, 1997) appeared in many other climate researches in the past. polynomial regression considers the higher orders in the data and tries to fit the line in multidimensional space. the formula for polynomial regression of order n is represented as: = + + + + ⋯ + the polynomial regression provides the best approximate fit to the data. it fits a wide range of curvature in multidimensional space. figure 3 illustrates a degree-2 polynomial fit to a data compared with the linear regression fit. figure 2 illustrates that there is no linear line that could fit the data properly. the polynomial regression fits a curved line to the data that most correctly fit the data. the temperature data is not linear either, so the polynomial regression with a higher degree has experimented in this study. 2.3 support vector regression (svr) support vector regression (svr) (drucker, 1997) is built on the concept of the support vector machine. the concept is extended from the support vector classifier, which is a popular classification problem. the concept of a support vector machine is to find the best line that separates the different features in maximum distance. for linearly separable problem can be easily modeled with a linear model. figure 3 and 4 illustrates the concept of the support vector machine. in figure 3, a separator hyperplane constructed with the support vectors shown with the red and green circle. these are the vector closest to the other side and the separator hyperplane is contracted to separate these points. the green line illustrates a linear fit whereas the red line shows a polynomial fit to the data 86 sneha paul and shuvendu roy forecasting the average temperature rise in bangladesh….. the svr shown in figure 4 uses the same concept, which is used by the classification model, with only some minor changes. the regression problem has infinite possible output as the output is a real number. so, the model does not try to output the actual real number. rather a predefined error is allowed, and the model tries to predict values that fall inside the predefined error range. the objective is to minimize the following: 1 2 || || + ( + ) the prediction of the model is represented as: = ( − ∗). < , > + this model can capture the linear patterns of the data. for modeling the nonlinear feature in the data, which is the case with our historical temperature data, svr applies a kernel function to convert the non-linear data into a linear representation in high dimensional space. this is commonly known as kernel trick. some of the most common kernels are linear, polynomial, radial basis function (rbf), and sigmoid. with the introduction of the kernel, the output formula will change a bit to reflect the kernel applied to the input first. = ( − ∗ ). < ( ), ( ) > + here, is the chosen kernel. the formula is represented in a simplified form as: = ( − ∗). ( , ) + the kernel trick is illustrated in figure 5: figure 5: kernel trick experiments are done with different kernels to model the temperature data. finally, the performance can be measured as mean absolute error (mae) or mean squared error (mse). in mse the higher errors are penalized more than small errors. for this reason, mse is considered as a standard measure to report the performance of this model. the mse is formulated as: = 1 ( − ́ ) 2.4 data description in this study, the berkeley earth surface temperature dataset has been used (rohde, 2013). the dataset is constructed from popular land and ocean temperature datasets called noaa's mlost, nasa's gistemp, and uk's hadcrut by lawrence berkeley national laboratory. the dataset contains 1.6 billion temperature reports from 16 pre-existing archives. it contains data from 1750 in average format and 1850 with minimum and maximum temperature. in our model, data from 1900 have been used as there are no missing data points since then. in the dataset, temperature data is considered for only one point (23.31n, 90.00e) in bangladesh. this is a limitation of the study. with more points, the model will be able to predict temperature with more precision. journal of engineering science 11(1), 2020, 83-91 87 3. results and discussion the results and findings of prediction model and the prediction of the temperature of bangladesh in the next 100 years are analyzed in different sections. 3.1 incremental trend in mean temperature the table 1 summarizes the average temperature of the last 120 years separated into a 10-year block. each 10year block is the average of every year's temperature in that range. averaging over 10 years normalize the oscillation in the temperature and shows a clear increase in the temperature of bangladesh. as it can be seen from the temperature data, on average there is around 1-degree temperature rise in the last century. so, the rise in this century is inevitable. there is also an upward trend in the minimum and the maximum temperature. although it is not clear from the table, later it trains the models and fit it to the data to see the upward slope in the fitted curve to the data. several models have been built to predict the temperature of bangladesh in the next 100 years. table 1: average temperature of bangladesh in the past year range mean temperature (0c) minimum temperature (0c) maximum temperature (0c) 1900-1910 24.53 16.78 28.57 1910-1920 24.72 17.04 29.24 1920-1930 24.86 16.62 29.23 1930-1940 24.97 16.46 29.27 1940-1950 25.01 17.49 29.58 1950-1960 25.26 17.27 29.45 1960-1970 25.14 17.16 29.64 1970-1980 25.16 17.06 29.59 1980-1990 25.29 17.39 29.79 1990-2000 25.26 17.13 29.37 2000-2010 25.53 17.23 29.71 2010-present 25.61 17.28 29.87 figure 6 represents temperature data of the year 1900 to present has been plotted to find out the real-time scenario of increment in temperature of bangladesh. figure 6(a) portrays the model on raw data. figure 6(b) shows the model on the mean temperature of each year from which a clear upward trend of temperature rise can be seen. figure 6(c) and 6(d) illustrate the model on the minimum temperature of each year and model on the maximum temperature of each year, respectively. both graphs portray a zigzag trend of temperature, so these figures are not conclusive. figure 6(e) shows every 10 year’s mean which also illustrates an upward trend in temperature rise. these graphs are used to illustrate the fit of every model (linear regression model, polynomial regression model, and svr) to the data to find out the best fit of the models. 3.2 performance of the model the different models are used for prediction of the future temperature of bangladesh. the linear model is the simplest model among them. figure 7 shows the model fit to data. as we can see from the figure, it is hard to predict the raw behavior of the temperature with a linear regression model as there is too much fluctuation in the daily temperature. however, the linear regression model can capture the upwards trends of the mean, minimum, and maximum temperature although there is an error in the prediction. the model best fits the 10-year average data as there is less fluctuation in the data. it also done fairly well in mean temperature data but cannot capture the dumpy nature of minimum and maximum temperature. figure 8 shows that the polynomial regression does a much better fit to the temperature data compared to the linear regression model. this model can capture the complex pattern in the data and fit a curved line to make a close estimation of the data. but the temperature data is still not modeled well with the polynomial regression of degree 3 because the data has a higher dimensional feature than the polynomial regression can fit. here, the model on the minimum and maximum temperature is better than the linear regression model. it is doing slightly better with the mean and 10-year average data too. illustrated in figure 9, the svm fits the complex pattern of the data far better than the linear and polynomial regression modes. especially the model with minimum and maximum temperature data makes a good fit to the data. 88 sneha paul and shuvendu roy forecasting the average temperature rise in bangladesh….. figure 6: temperature data distribution plot of the last century of bangladesh figure 7: linear regression model fits on the temperature data (b) mean temperature (c) minimum temperature (a) raw data (d) maximum temperature (e) 10-year average temperature (e) model fit on 10-year average temperature (d) model fit on maximum temperature (c) model fit on minimum temperature (b) model fit on mean temperature (a) model fit on raw data journal of engineering science 11(1), 2020, 83-91 89 figure 8: polynomial regression model of degree 3 fits on the temperature data figure 9: svm model with rbf kernel fits on the temperature data (b) model fit on mean temperature (c) model fit on minimum temperature (a) model fit on raw data (d) model fit on maximum temperature (e) model fit on 10-year average temperature (e) model fit on 10-year average temperature (d) model fit on maximum temperature (c) model fit on minimum temperature (b) model fit on mean temperature (a) model fit on raw data 90 sneha paul and shuvendu roy forecasting the average temperature rise in bangladesh….. although the complex pattern has the chance of being over fitted by the model and might show unnecessary deviation in future prediction. 3.3 performance comparison of implemented prediction models the performance of the model depends on the error it makes compared to the original output. lesser the error of the model better is that model. the model reports not only the raw data model but also modeled the mean, minimum, and maximum temperature of each year. we have also calculated a separate model for predicting the mean temperature of each decade. the table 2 summarizes the performance of the model on each data type by each model reported as mse. table 2: performance comparison of implemented prediction models in mse model raw data (0c) average of the year (0c) minimum of the year (0c) maximum of the year (0c) 10-year average (0c) linear regression 14.15 0.0566 0.0566 0.3879 0.1663 polynomial regression (degree=3) 0.0054 0.0047 0.0174 0.0043 0.0054 polynomial regression (degree=5) 0.0055 0.0048 0.0175 0.0044 0.0055 svr (kernel = rbf) 3.4509 0.0465 0.2603 0.1202 0.0161 svr (kernel = polynomial) 1.2473 0.4396 2.1434 1.0272 0.5565 as it can be seen from the model that the polynomial regression has better performance in terms of mse compared with other models. the significance of each of the numbers is that the average error of a predicted temperature at a given time from its actual temperature is the root of the reported number in this table. for example, the mse of the average temperature model with the polynomial regression of degree 5 is 0.0048. the root is 0.069. that means if the predicted temperature is 25.000c. the actual temperature lies between 24.93 to 25.0690c, which is a minor difference and can be considered as the near-perfect prediction. as table 2 shows, for predicting all the variations of raw, mean, minimum, maximum, and 10-year average temperature polynomial regression with degree 3 has the lowest prediction error. svr with polynomial kernel has a high mse error in almost all prediction except for the linear regression model has the highest error in raw data prediction. so, the polynomial model with degree 3 can be considered as the best model for predicting the future temperature of bangladesh. from now onward, all the performances are reported with this model. 3.4 temperature prediction from the model as it has been found that polynomial regression model has better performance among the three models, the temperature of bangladesh for the next 100 years has been predicted using this model. table 3 gives the predicted average temperature of bangladesh. table 3: predicted average temperature of bangladesh in the next 100 years year average temperature (0c) year average temperature (0c) 2025 25.62 2075 26.01 2030 25.66 2080 26.05 2035 25.7 2085 26.09 2040 25.74 2090 26.13 2045 25.78 2095 26.17 2050 25.82 2100 26.2 2055 25.85 2105 26.24 2060 25.89 2110 26.28 2065 25.93 2115 26.32 2070 25.97 2120 26.36 as we can see from the table 2 there is a predicted increment in temperature from 25.62 to 26.360c, which is close to 10c increment in one century. apart from the success of the model in predicting the average temperature of each year, there are some limitations in this research. the seasonal temperature variation is not analyzed in our word and month-wise temperature prediction is not possible with our model. 4. conclusions in this study, the past 100 years temperature data of bangladesh has been analyzed and it has been seen that in the past 100 years temperature of bangladesh has risen by 10c on average. linear regression model, polynomial journal of engineering science 11(1), 2020, 83-91 91 regression model and svr are developed to predict the average temperature of each year in bangladesh for the next 100 years. the performances of those models are compared, and it is found that the polynomial regression with degree 3 is the best model for this application. an increase of close to another 10c can be seen in the average temperature of bangladesh from the prediction of the model. reference al, s.e.t., 2006. observational constraints on past attributable warming and predictions of future, 19(13), 3055–3069, https://doi.org/10.1175/jcli3802.1 alley, r.b., hewitson b., hoskins b.j., joos f., jouzel j., kattsov v., lohmann u., manning m., matsuno t., molina m., nicholls n., berntsen t., overpeck j., qin d., raga g., ramaswamy v., ren j., rusticucci m., solomon s., somerville r., stocker t.f., stott p.a., bindoff n.l., stouffer r.j., whetton p., wood r.a., wratt d., chen z., chidthaisong a., friedlingstein p., gregory j.m., hegerl g.c., and heimann m., 2007. summary for policymakers, in s solomon and d qin and m manning and m marquis and kb averyt and m tignor and hl miller and z chen (eds.), climate change 2007: the physical science basis. contribution of working group 1 to the fourth assessment report of the intergovernmental panel on climate change (pp. 1-18). cambridge, uk and ny, usa: cambridge university press. daily star, 2019. human ecology, changing climate and hydrology, retreived from daily star: http://www.thedailystar.net/newdesign/newsdetails.php?nid=209731. daily star, 2019. vulnerable women, children need special attention, retreived from daily star:http://www.the dailystar.net/newdesign/news-details.php?nid=208648. drucker, h., burges c.j., kaufman l., smola a.j., and vapnik v., 1997. support vector regression machines, in advances in neural information processing systems, 28(7), 155-161. elahi, f., and khan n.i., 2015. a study on the effects of global warming in bangladesh, international journal of environmental monitoring and analysis, 3(3), 118-122. fyfe, j.c., merryfield w.j., kharin v., boer g.j., and lee w.s., 2011. skillful predictions of decadal trends in global mean surface temperature, geophys, res. lett, 38, p.l22801. houghton, j., 2004. global warming: the complete briefing (third edition), cambridge university press, 86(30), 282-283. icddr, b., 2019. our strategy, retrieved from icddr, b: https://www.icddrb.org/about-us/strategy islam, s.s., 2004. status of forest genetic resources conservation and management in bangladesh. in forest genetic resources conservation and management: proceedings of the asia pacific forest genetic resources programme (apforgen) inception workshop, kepong, kuala lumpur, malaysia, 15-18 july 2003 (p. 27), biodiversity international. lindzen, r. s., 2008. global warming: what is it all about?, retrieved from epa:https://www.epa.gov/ environmental-economics/global-warming-what-it-all-about mahmood, s.a.i., 2012. impact of climate change in bangladesh: the role of public administration and government’s integrity, journal of ecology and the natural environment, 4(8), 223-240, https://doi.org/ 10.5897/jene11.088 orr, g., 2007. global warming [1], structural engineer, 85(17), 34. pender, j.s., 2008. what is climate change? and how it will affect bangladesh, briefing paper (final draft). dhaka, bangladesh: church of bangladesh social development programme. rohde, r., muller r., jacobsen r., perlmutter s., rosenfeld a., wurtele j., curry j., wickhams c., and mosher s., 2013. berkeley earth temperature averaging process, geoinfor, geostat.: an overview $1: 2$. geoinformatics geostatistics an overview, 1(2), 20-100. ruppert, d., 1997. local polynomial regression and its applications in environmental statistics, statistics for the environment, 3, 155-173. shahzad, u., 2015. global warming: causes, effects and solutions, durreesamin journal, 1(4), 1–7. smith, d. m., scaife a. a., boer g. j., caian m., doblas-reyes f. j., guemas v., hawkins e., hazeleger w., hermanson l., ho c. k., ishii m., kharin v., kimoto m., kirtman b., lean j., matei d., merryfield w. j., müller w. a., pohlmann h., ... wyser k., 2013. real-time multi-model decadal climate predictions, climate dynamics, 41(11-12), 2875-2888, https://doi.org/10.1007/s00382-012-1600-0 tol, r.s., 2006. the stern review of the economics of climate change: a comment, energy & environment, 17(6), 977-981. unfccc, 2005. depledge, j., and lamb r., eds., caring for climate: a guide to the climate change convention and the kyoto protocol (pp. 1-33). bonn, germany: climate change secretariat (unfccc). venkataramanan, m., 2011. causes and effects of global warming, indian journal of science and technology, 4(3), 226-229. zholudeva, v. v., 2019. statistical assessment of the impact of climate change on social and demographic processes (on the example of the yaroslavl region), statistics and economics, 16(6), 57–66. 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 02_jes_287_10_21_2020 journal of engineering science 11(2), 2020, 19-25 doi: https://doi.org/10.3329/jes.v11i2.50894 temporal trend analysis of historical climatic data at northeastern hilly region of bangladesh using mann-kendall test k. k. mondal1*, md. a. e. akhter1 and m. a. k. mallik2 1department of physics, khulna university of engineering & technology, khulna – 9203, bangladesh 2bangladesh meteorological department, agargaon, sher-e-bangla nagar, dhaka – 1207, bangladesh received: 12 june 2020 accepted: 21 october 2020 abstract an attempt has been implemented to find out the temporal trend of climatic data of average temperature and total rainfall for the study period 1980-2016 at north-eastern hilly region in bangladesh. the non-parametric mann-kendall test is used to analyze the trend of climatic data. the objective of the study is to investigate the trend variation in the north-eastern hilly region. results show that in monsoon season, both sylhet and srimangal meteorological stations experience a positive tendency with a rate of 0.037 and 0.0170c/year, respectively which are statistically significant at 99.9% level of significance. monthly significant positive changes are found in all months except november, december and january for sylhet while srimangal indicates significant positive changes except july, september, october and november. the total rainfall at both the stations reveals decreasing trend during maximum seasons and months but the trend is not significant. keyword: hilly; mann-kendall; trend; temporal; temperature; rainfall. 1. introduction climate change issue is apprehensive throughout the world. a scientific observational evident that the global mean temperature has increasing trend at about 0.3 to 0.60c over the last 100 years (wmo 1991). bangladesh is the worst sufferer among the world for its bad impacts. temperature and rainfall are the most important climatic parameters in atmospheric science. temperature is increasing in an alarming rate and rainfall is fluctuating.the climate of bangladesh comprises four seasons: pre-monsoon, monsoon, post-monsoon and winter. most of the rainfall are occurs during pre-monsoon and monsoon seasons in bangladesh according to bangladesh meteorological department (bmd). ahmed (2012) has shown that the variability of rainfall amount was higher in the pre-monsoon season at north-eastern part of bangladesh around sylhet. research was implemented on trend of climatic parameter leading to change the atmospheric conditions by almazroui et al. (2014). karmakar (1997) studied climatic change and its impacts on natural disasters and south-west monsoon in bangladesh and the bay of bengal. they found that especially after 1961-1970 bangladesh have experienced increasing tendency of decadal mean annual temperature. mondal et al. (2018) found in a study that dhaka division (dhaka, faridpur and mymensingh stations) experience a positive trend of average temperature which is high populated-industrial area and situated at the center of the country compare to khulna division. alam et al. (2010) investigated temporal variation of rainfall over south-western part of bangladesh and noticed positive tendency of seasonal rainfall. a study was implemented by khatun et al. (2016) on climate change in bangladesh which reported that southeastern part of bangladesh experiences an extreme rainfall event while negative deviations are found at dhaka, chattogram and sylhet regions during monsoon season. rainfall anomaly of 1984-2016 showed an increasing trend over the south-west bangladesh which is an indicator of changing spatial distribution of rainfall over the country (gupta et al., 2018). reza et al. (2018) showed that excess value of rainfall observed in the fourth decade at north-eastern part of bangladesh during the study period 1948-2015. mannan et al. (2015) indicated that total numbers of rainy days in monsoon season at srimangal & sylhet are 83.8 and 99.7 during 1980-2014 and their stds are 8.5 and 5.7. and the trends of rainfall deviations are 0.05% and -0.04% respectively. the primary objectives of this study are to find out temporal trend of climatic data, such as average temperature ( ) and total rainfall for the study period 1980-2016 at north-eastern hilly region in bangladesh using mannkendall test. 2. data used and methodology 2.1 data used the temperature and rainfall data of northeastern part (figure 1) in bangladesh for the period 1980-2016 are used in this study. these data are collected from bmd which comply the world meteorological organization (wmo) terms and conditions in maintaining the data collection. monthly data are obtained from the daily data. * corresponding author: karnosudra@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 20 karno kumar mondal et al. temporal trend analysis of historical climatic … monthly data of average temperature are evaluated from monthly maximum and minimum temperatures. again, from the monthly data, seasonal mean values are computed for the four seasons such as pre-monsoon (marchmay), monsoon (june-september), post-monsoon (october-november) and winter (december-february). the missing data are filled up by the time mean values of the existing years. the trend of the dataset is analyzed using mann-kendall test. figure 1: selected area of bangladesh used under this study (mannan et al., 2015) 2.2 methodology mann-kendall test is used to find out the temporal variation of climatic data. the following subsections are written for mann-kendall analysis and sen’s slope estimator. 2.2.1 mann-kendall analysis the mann-kendall test was commonly applied where the data do not identical to a normal distribution. this test evaluates a nonparametric form of monotonic trend regression analysis of y values tend to increase or decrease over time. wmo has been extensively recommended this test for public use to evaluate trends. the nonparametric mann-kendall (1945) test is commonly used for hydrologic data analysis; can be used to detect that are monotonic but not necessarily linear. the null hypothesis in the mann-kendall is that the data are independent and randomly ordered. the mann-kendall test does not require the assumption of normality, and not only indicates the direction but not the magnitude of significant trends. the mann-kendall test measured values, ( j− k), where j>k, and test statistics s is computed exerting the formula; s =∑ ∑ (1) where j and k are the annual values in years j and k, j>k, respectively, and = 1 > 0 0 = 0 1 0 (2) the test statistics τ can be computed as; = / (3) it is necessary to compute the probability associated with s and the sample size n, to statistically quantify the significance of the trend. the calculating formula of variance s is denominate as; ( ) = [( −1) (2 +5)∑ t t 1 2t 5 (4) where q is defined as the number of tied groups and p is the number of data in the th group. the values of s and var(s) are accustomed to calculate the test statistics z which is following as; z= > 0 0 = 0 0 (5) journal of engineering science 11(2), 2020, 19-25 21 z score follows a normal distribution. at a choice of α=0.05 (95% confidence interval) and two sided alternative, the critical values of 0.025 are equal to -1.96 to 1.96. the trend is said to be decreasing if z is negative and the absolute value is greater than the level of significance, while it is increasing if z is positive and greater than the level of significance. if n≤10 the normal approximation test is used and a statistically significance trend is computed exerting the z score. mann-kendall & sen’s slope estimator tested the z score significance level at α: 0.001, 0.01, 0.05 and 0.1. 2.2.2 sen’s slope estimator the sen’s slope estimator is a method of nonparametric that was employed to exhibit the linear patterns in this study. this method is most effective than regression equation. if a linear aptitude present in a time series, then the true slope can be calculated by exerting a simple nonparametric procedure. in the case of linear model ƒ(t) can be denominate as ƒ (t) = +b (6) where q is representing slope, b is a constant value and t is time. to derive an estimate of the slope q, the slopes of all data pairs are calculated exert the equation; i= , , = 1,2,3, … … . . , > (7) if there are n values in the time series there will be as many as n = n(n-1)/2 slope estimates i. to estimates of b in equation the n values of differences i– i are calculated. an estimate of b isobtained from the given mean values. data were processed using an excel macro names makesens created by salmi et al. (2002). 3 results and discussion temporal trend variation of climatic data, average temperature (tavg) and total rainfall for the study period 19802016 at northeastern hilly region of bangladesh are discussed in the following subsection. figure 2: seasonal change of average temperature at sylhet during the period 1980-2016 3.1 temporal trend of average temperature at sylhet using mann kendall test seasonal and monthly trends considering tavg is evaluated at sylhet for the study period 1980-2016 and shown in figure 2. calculated values of sen’s slope and z with its significance level using mann-kendall & sen’s slope estimator is tabulated in table 1. it indicates that except winter season the 22 karno kumar mondal et al. temporal trend analysis of historical climatic … other three seasons are experiencing significant positive tendency. sen’s slope values of pre-monsoon and monsoon seasons are 0.036 and 0.037 0c/year, respectively. according to z values these trends are statistically significant at level of 99.9% interval. subsequently the post-monsoon season observed a positive trend with a rate of 0.023 0c/year that is statistically significant at 90% confidence level. 3.2 temporal trend of average temperature at srimangal station using mann kendall test seasonal and monthly trends considering of tavg are evaluated at srimangal for the study period 1980-2016 and the results are shown in figure 3. calculated values of sen’s slope and z with its significance level using mann-kendall & sen’s slope estimator is tabulated in table 1. it indicates that according to z value trend at the winter season for both stations is statistically insignificance but that at the monsoon and post-monsoon seasons for both stations is statistically significant whereas the trend at pre-monsoon season is significant only at sylhet station. monsoon and post-monsoon seasons show a positive tendency with a value of sen’s slope 0.017 and 0.023 0c/year respectively which is statistically significant at level of 99.9% confidence interval. whereas the pre-monsoon and winter seasons imply that the trends are insignificant. figure 3: seasonal change of average temperature at srimangal during the period 1980-2016 monthly trend analysis reveals that tavg rising for all of the months except april. but the values dominate significantly only in july and september to november. the sen’s slope estimates for july, september, october and november are 0.024, 0.027, 0.024 and 0.025 0c/year respectively. all of these values indicate the increasing rate of temperature. again the z values are extending the critical values for all these months but it varies for different level of significance. 99.9% significant are found for the month of july and september. and the trend of temperature in november is significant at level of 90% confidence interval whereas the trend in october is 95% confidence interval according to z value. the sen’s slope estimates for january to june, august and december are 0.008, 0.020, 0.012, -0.003, 0.015, 0.004, 0.008 and 0.010 0c/year respectively. here january, february, march, may, june, august and december are express positive trend whereas the april is negative trend. according to z statistics all of these trends are insignificant. 3.3 temporal trend of total rainfall at sylhet using mann kendall test seasonal and monthly trends considering total rainfall is evaluated at sylhet of bangladesh for the study period 1980-2016 and the results are shown in figure 4. calculated values of sen’s slope and z with its significance level using mann-kendall & sen’s slope estimator is tabulated in table 2. it indicates that all seasons experience negative tendency where all seasons except winter, the trends are journal of engineering science 11(2), 2020, 19-25 23 statistically insignificant. the winter season observed trend to decrease with a value -0.871 mm/year that is significant at 90% confidence level according to z value. table1: monthly and seasonal sen’s slope estimates for tavg with z test value at sylhet and srimangal. season/month sen’s slope q z value significance level sylhet srimangal sylhet srimangal sylhet srimangal premonsoon 0.036 0.008 3.55 0.53 *** monsoon 0.037 0.017 5.18 3.77 *** *** post-monsoon 0.023 0.023 1.84 2.79 + ** winter 0.010 0.004 0.78 0.28 january -0.003 0.008 -0.23 0.33 february 0.028 0.020 1.66 1.46 + march 0.031 0.012 1.65 0.48 + april 0.030 -0.003 2.49 -0.20 * may 0.037 0.015 2.54 1.21 * june 0.018 0.004 1.95 0.55 + july 0.049 0.024 4.93 3.37 *** *** august 0.025 0.008 3.44 1.41 *** september 0.043 0.027 3.32 3.99 *** *** october 0.035 0.024 3.51 2.54 *** * november 0.004 0.025 0.30 1.94 + december 0.018 0.010 0.98 0.50 note: *** significant at the 99.9% confidence level; ** significant at the 99% confidence level; * significant at the 95% confidence level and + significant at the 90% confidence level. figure 4: seasonal change of total rainfall at sylhet during the period 1980-2016 monthly trend analysis reveals that total rainfall has the trend of rising for of the months of january, june, august, november and december. other months are shows negative tendency of rainfall. but the trend in rainfall in march, july and december has only significant values. the sen’s slope estimates for july and december are -8.707 and 0.001 mm/year, respectively. and the trend of december is significant at level of 90% confidence interval whereas july is 95% confidence interval. in sylhet station the rainfall shows negative trends 24 karno kumar mondal et al. temporal trend analysis of historical climatic … for february, march, april, may, july, september and october while positive trends are shows for january, june, august, november and december those are statistically insignificant. 3.4 temporal trend of total rainfall at srimangal using mann kendall test seasonal and monthly trends considering total rainfall is evaluated at srimangal of bangladesh for the study period 1980-2016 and shown in figure 5. calculated values of sen’s slope and z with its significance level using mann-kendall & sen’s slope estimator is tabulated in table 2. it indicates that the rainfall trend in pre-monsoon and monsoon seasons experiences positive tendency while the trends in postmonsoon and winter seasons are negative and are statistically insignificant. monthly trend analysis reveals that total rainfall has rising trends for of the months of january, may-august and december. figure 5: seasonal change of total rainfall at srimangal during the period1980-2016 table 2: monthly and seasonal sen’s slope estimates for total rainfall with z test value at sylhet and srimangal season/month sen’s slope q z value significance level sylhet srimangal sylhet srimangal sylhet srimangal pre-monsoon -3.893 2.038 -0.77 0.52 monsoon -11.026 0.662 -1.12 0.17 post-monsoon -1.211 -0.143 -0.47 -0.10 winter -0.871 -0.479 -1.91 -0.80 + january 0.000 0.000 -0.67 -0.10 february -0.267 -0.085 -1.00 -0.51 march -1.971 -0.485 -2.00 -0.67 * april -2.506 -0.020 -0.59 -0.03 may -0.154 4.994 -0.07 1.54 june 2.056 1.351 0.46 0.72 july -8.707 1.279 -2.21 0.78 * august 3.923 1.000 1.10 0.37 september -5.750 -2.458 -1.45 -1.96 * october -1.793 -0.389 -0.78 -0.18 november 0.000 -0.023 -0.34 -0.65 december 0.001 0.000 -1.88 -1.20 + journal of engineering science 11(2), 2020, 19-25 25 the trends in other months have negative tendency. but only the trend in september has significant value. the sen’s slope estimates for september is -2.458 mm that significant at level of 95% confidence interval according to z value. in srimongal the rainfall experiences negative trends for february, march, april, september, october and november while positive trends are shown in january, may, june, july, august and december those are statistically insignificant. 3.5 temporal variations of & total rainfall at sylhet and srimangal for average temperature 99.9% significant value complies in pre-monsoon at sylhet, whereas it complies in post-monsoon season at srimongal. monthly significant positive changes of average temperature are found in srimangal except july, september, october and november while sylhet except november, december and january. all seasons in sylhet experiences negative trends whereas only post-monsoon and winter seasons experiences negative trends in srimongal. the both stations are perceived positive trends in january, june, august and december and negative trends in february, march, april and september. the trends of average temperature & rainfall in sylhet are mostly significant than that of srimangal. 4 conclusions the following conclusions can be made from studies:  in monsoon season the temperature in both sylhet and srimangal experiences a positive tendency with a rate 0.037 and 0.017 0c/year, respectively which are significant at 99.9% level of significance.  monthly significant positive changes of average temperature are found in all months except november, december and january for sylhet while srimangal indicates significant positive changes except july, september, october and november.  monthly negative changes of average temperature are found in january and august in sylhet and srimangal respectively which are not statistically significant.  the total rainfall in most seasons and months at both stations reveals decreasing tendency dramatically.  there is a clear indication that the trends of sylhet for average temperature & rainfall are mostly significant than that of srimangal station. more investigations are needed to completely identify if there is any climatic change occurred. references alam, m. m., hossain m. a., and ara m., 2010. long term variations of climate over different stations of khulna and barisal divisions of bangladesh, journal of engineering science, 1(1), 1-7. ahmed, s., 2012. seasonal variation of temperature and rainfall characteristics in the northeastern part of bangladesh around sylhet, j. agro for. environ, 6 (2), 81-88. almazroui m., islam m. n., dambul r., and jones p. d., 2014. trends of temperature extremes in saudi arabia. international journal of climatology, 34: 808-826. gupta, a., and moniruzzaman m., 2018. spatio-temporal analysis of changing rainfall pattern of bangladesh using geo-spatial tools, conference on weather forecasting and advances in physics (cwfap), 11-12 may 2018. khatun, a. m., rashid m. b., and hygen h. o., 2016. met report: climate of bangladesh, no.08/2016. mondal, k. k., akhter a. e., and islam m. n., 2018. trends of temperature at highand low-densely populated divisions in bangladesh, the international journal of earth & environmental sciences, 3(1), 1-8 mannan, m. a., chowdhury m. a. m., karmakar s., ahmed s., and mason s. j., 2015. rainfall the prediction over northeastern part of bangladesh during monsoon season, dew-drop 01, 14-25. mann, h. b., 1945. non-parametric test against trend, econometrica, 13, 245–59. karmakar, s., and nessa j., 1997. climate change and its impacts on natural disasters and sw-monsoon in bangladesh and the bay of bengal, journal of bangladesh academy of sciences, 21(2), 127-136 salmi, t., määttä a., anttila p., ruoho-airola t., and amnell t., 2002. detecting trends of annual values of atmospheric pollutants by the mann-kendall test and sen's slope estimatesthe excel template application makesens, publications on air quality no. 31, finnish meteorological institute, helsinki, 35 pp. reza, m. k. h., alam m. m., and rahman m. m., 2018. spatio-temporal variation of pre-monsoon rainfall and rainy days over bangladesh, journal of engineering science, 09(1), 21-27. wmo, 1991. proceeding of the second world climate conference, pp. 24 © 2020 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attribution-noncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. microsoft word 1_jes_315 journal of engineering science 12(2), 2021, 01-10 doi: https://doi.org/10.3329/jes.v12i2.54626 *corresponding author: mmorshed@urp.kuet.ac.bd https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal measuring social license to operate the infrastructure project in bangladesh md manjur morshed*, zaki arafin and salima khan nafi department of urban and regional planning, khulna university of engineering & technology, bangladesh received: 11 march 2021 accepted: 10 may 2021 abstract social license to operate is a measure to engage with and gain acceptance from the stakeholders for large infrastructure projects. the object of this study is to measure social license to operate an infrastructure project in khulna city, bangladesh. predicated on an established model, 16 statements were selected as the basis for measuring the social license to operate. a questionnaire survey was conducted among 44 local stakeholders. using varimax rotation, five different components were identified: socio-political, economic, procedural fairness, interactional trust, and institutional trust. the model was statistically tested and found to be a medium fit explaining the results. the findings are that the project has gained socio-political, economic, procedural fairness, and interactional trust of the stakeholders, yet lacks institutional trust factor to achieve a social license to operate. the paper recommends the inclusion of the social license to operate concept in infrastructure planning and implementation phases in bangladesh. keywords: factor loading; legitimacy; principal component analysis; project; stakeholder. 1. introduction the concept of social license to operate (slo) is traditionally focused on local people directly or indirectly affected by any project. as a response to social risk, the term slo emerged in the mid-1990s from the mining companies (boutilier and thomson, 2011). from that time, the term had been used by a wide range of actors in the resource sector including many development sectors like mining, civil society and ngos, research institutions, governments and consultants (bhpb, 2011; kurlander, 2001; slack, 2009; csiro, 2013; mcnab et al., 2013; australian government, 2006). therefore, slo is used to describe the acceptance or approval of an ongoing project by local community members and other stakeholders (moffat and zhang, 2017). infrastructure is considered is the key to bangladesh's rapid economic development. several mega-projects e.g., padma bridge, nuclear power plant, cng terminal, etc. are providing a boost for the country's drive for rapid economic development. for accelerating growth and empowering citizens in its national development strategies, such as the five year plan (fyp), bangladesh has identified access to infrastructure services as a major component of development. according to the 7th fyp, the country needs more than five percent of gdp as an additional investment on major infrastructure development projects annually. to achieve the sdg targets (infrastructure-related), additional cost has been estimated to be 5.67% up to the fiscal year 2030 (ged, 2017). the southwestern part of bangladesh, namely the khulna division, is expected to get the impetus for economic activity with the development of mongla port, padma bridge connecting dhaka, railway line connecting dhakakhulna-kolkata (india), and road connections to the asian highway. as part of these mega-infrastructure projects, connecting khulna railway station with the asian highway has been a key infrastructure project. to improve the network infrastructure, the khulna railway station had been remodeled and khulna-mongla railway project was passed by ecnec on 21 december 2010. as part of the overall infrastructure, road from the railway station connecting the asian highway (covering gollamari and zero-point intersection) needs to be improved and a road expansion program has been undertaken by local government and engineering departments (lged). for the expansion, land acquisition as well as eviction from the public land is mandatory. infrastructure development project like this where peoples’ land will be acquired, and they are subject to eviction, need to measure the slo for better acceptance of the development project by the local community. the objective of this research is to measure slo for the road expansion project in khulna city in order to measure to see how the project is been received by the local stakeholders. this paper explores the application of 2 md manjur morshed et al. measuring social license to operate ……… thomson and boutilier model, which is one of the most noted methods of measuring slo, in bangladesh contexts. measuring slo is a challenging task in a developing country context like bangladesh as concepts such as consensus building, public participation, and resettlement and rehabilitation are rarely considered in the project conception and implementation phases. one of the major limitations of this study is that it assesses slo of an ongoing infrastructure project, and thus does not consider phases of the project, i.e. conception or implementation. secondly, a small number of directly project-affected stakeholders are selected for the questionnaire survey. against this background, the second section of the paper explains different aspects of slo. the third section is materials and methods. the fourth section is findings and followed by a discussion in the fifth section. the concluding remarks are in the sixth section. 2. the concept of slo the slo concept is a measure to justify large infrastructure investment and activities, and modeling and measuring slo is a pre-condition for any successful project initiation. over the last decade, there has been increasing recognition of the slo to ensure societal support for extractive operations, i.e., mining, which predicated on avoiding protests by the local community and ensuring long-term project sustainability (hanna, 2016; mitchel, 2020). the slo guidelines and criteria noted by different literature include legitimacy, credibility, and trust by the stakeholders (boutilier, 2014; jijelava and vanclay, 2018; vanclay, 2017). on the contrary, different literature also noted the ambiguity of the slo concept (see for example, owen and kemp, 2013; harvey and bice, 2014; de jong and humphreys, 2016). meesters et al. (2020) summarized the key problems of the slo concept and its applicability. firstly, overwhelming literature limit the stakeholders’ engagement to locals who live nearby and to vocal and organized groups while ignoring the non-residents. secondly, stakeholders’ engagement is objective as to run the operation smoothly which is not necessarily inclusive, but can be limited to powerful interest groups only. thirdly, the slo concept is more concerned with the local socio-economic and environmental impacts while ignoring the global consequences. this, however, opens up a new concern with the capacity of the local stakeholders in the hegemonic environment. however, prioritizing local stakeholders in slo is convenient as opposed to considering non-locals and global implications. the pioneering work by thomson and boutilier (2011) defined slo as the community’s or stakeholders’ perceptions regarding the acceptability of a government’s local operations i.e. development projects like road expansion. they identified four levels of the slo: i) withheld/withdrawn; ii) acceptance; iii) approval; and iv) psychological identification. the hypothesis is that that slo granted by stakeholders to a government/company is inversely correlated to the level of socio-political risks and vis-à-vis. the lowest level of slo signifies that the project is being withheld or withdrawn, meaning the project is in danger of accessing public support, access to resources and labor, and in many cases, there is a problem of mass-protest. therefore, the lowest level of slo indicates an extremely high socio-political risk. on the contrary, the highest level of slo is the acceptance of the project. on figure 1, the acceptance of the project covers the highest area which indicates that it is the common level of the social license granted. if trust is established over time, the slo could rise to the highest level – psychological identification, where the level of socio-political risks is very low. figure1: levels of slo (adapted from thomson and boutilier, 2011) generally, slo is measure qualitatively with a pool of statements devised to measure the four-level concept of slo (thomson and boutilier, 2011). the whole set of analyses is validated against the verbal opinion journal of engineering science 12(2), 2021, 01-10 3 (agree/disagree ratings at a 5-point scale) from stakeholders on the project. by applying a varimax factor rotation analysis, four factors (economic legitimacy, socio-political legitimacy, interactional trust , and institutionalized trust) are measured, and then an arrowhead model is proposed. in the arrowhead model, the four factors had been measured independently by the same statements. the average of the ratings on the statements symbolizes the idea that the overall level of slo is a continuum. a high score on the perception of institutionalized trust will always have a high score on other factors. stakeholders with low scores on the perception of socio-political legitimacy and interactional trust will never have high scores on institutionalized trust. likewise, stakeholders with low scores on perceptions of economic legitimacy will never have high scores in any of the other factors according to the arrowhead model. 3. methodology 3.1 study area the study area is khulna city, the third-largest city in bangladesh. for measuring slo, selection of study area was a crucial task as the project needed to have many interactions with the local people and the stakeholders. the road expansion projection from the powerhouse intersection to gollamari, which is directly linked to the khulna-mongla highway, is considered for this study. recently improved khulna railway station is located just beside the powerhouse. in this situation, the railway station will face a huge traffic load from mongla for being the second largest seaport in the country. the project road is sher-e-bangla road-1. a portion of the road expansion, from powerhouse to gollamari is the area of this research. the khulna roads and highways (rhd) is responsible for the construction work, which plans to expand the two-lane roads into four lanes. the construction has started on 17 october, 2020 after a brief halt due to technical difficulties and currently ongoing. the road expansion project is part of the 24 mega-projects that are to shape the future of khulna city. the cost is around 250 crore. the ongoing project is causing a severe problem for traffic congestions, as well as demolition of illegal structures. the khulna city corporation (kcc) is overseeing the clearance activities as well as removing and relocating existing urban infrastructure from the project area (dhaka tribune, 2020). the project area is shown in figure 2. figure 2: study area map (powerhouse to gollamari bridge) 4 md manjur morshed et al. measuring social license to operate ……… 3.2 stakeholders and sample selection stakeholders represent the whole community who are affected directly or indirectly by the project activity. in a broader sense, stakeholders are those who could be affected by the actions of a company or who could have an effect on the company of the development project. thus, there are many stakeholder networks outside a geographic community, and for measuring the slo, they also need to be included. the selected infrastructure development project of this research has many stakeholders: local community, drivers, landowners, daylaborers, traffic police, and so on. the industry association, local shops, residents, motor association, drivers (heavy, medium, light), authority, local institute, community organizations of the area are also directly or indirectly affected due to this construction project, who are the major stakeholders for this study. for this study, a total of 44 local people who lived or worked in the study area within the proximity of the construction site were selected randomly for primary data collection and questionnaire survey. the stakeholders are comprised of: 9 landowners, 5 residential renters, 7 small shop owners, 6 battery-bike drivers and 6 commuters, 3 rickshaw pullers, 4 truck drivers, 2 police officers and 2 illegal occupiers (food sellers). these stakeholders are directly and indirectly affected due to the road construction project. however, understandably, there are many more project stakeholders which are not considered due to lack of data collection. because of the time limitation of the research (funding period between july 2018 to june 2020), as well as a mere lack of cooperation from the stakeholders, a statistically representative sampling was not possible. however, the limited sample size does not put a constraint to testing the applicability of the model for the case study. 3.3 statements selection after selecting the stakeholders, from a pool of various statements related to the stakeholders, 16 were finally selected. these statements had followed thomson and boutilier’s statements and were selected through psychometric item selection process after discussion with the key stakeholders. the statements are as follows: 1. our organization and the infrastructure project have the same vision for the future of this area. 2. the construction project does what it says will do in its relation with the locality. 3. this project is not harmful to our local tradition/culture. 4. this project keeps harmony with the day-to-day life of the local community. 5. this project shares decision-making with the community. 6. the infrastructure project takes account of our interests. 7. there was consultation with the locality before initiating the project. 8. this project will create problems related to the environment and others. 9. in the long run, the project contributes to the well-being of the whole region. 10. the total cost of transportation will be decreased. 11. the project authority has taken mitigating measures to negatively affected people. 12. the project will decrease traffic congestion as well as accidents. 13. land price and house rent will increase because of the road project. 14. the budget and the construction progress of the project seem fair to me. 15. the infrastructure project treats everyone fairly. 16. i think the project has slo in our locality. 3.4 data analysis the collected data was inputted into spss 21 and amos 21 software to run the path model and factor analysis. all the statements were asked to the stakeholders as they were requested to respond on a scale of 5 points (1= strongly disagree, 5= strongly agree). the principal component analysis was performed with varimax rotation to determine the factor affecting slo. a varimax matrix is created to show the correlation among the statements where high scores indicate a highly positive review and vice versa. the rotated component matrix shows the pearson correlation between the items, the statements and the components. these correlations are known as factor loadings. ideally, this technique is used to measure each input variable precisely as one factor. but there are also four cross-loadings here. these cross-loadings are then redistributed over the factors. among many rotation techniques, varimax rotation, short for variable maximization is been used here. it tries to redistribute the factor loadings in such a way that each variable measures precisely one factor. amos software is used to test the path diagram of the slo of the area. the path model provides the result of excellent fitness of the statements with significant and non-significant value. journal of engineering science 12(2), 2021, 01-10 5 4. results and discussions 4.1 principal component analysis with the 15 input variables, that are our statements, principal component analysis initially extracts 16 statements (or “components”). in every factor analysis, there is the same number of factors as there are variables. each factor captures a certain amount of the overall variance in the observed variables, and the factors are always listed in the order of how much variation they explain. the eigenvalue is a measure of how much of the variance of the observed variables a factor explains. only the components with high eigenvalues (>1) are likely to represent real underlying factors. a varimax matrix with all the selected statements which tend to produce factor loading is used to simplify the expression of a particular factor in terms of a few major items. that means each factor has a small number of large loadings. the output of the principal component factor analysis with varimax rotation is shown in table 1. from table 1, it is seen that the 15 statements can be categorized, and are named accordingly, into five components that are affecting the slo of the road expansion project. the factors are: i) socio-political legitimacy, ii) procedural fairness, iii) economic legitimacy, iv) interactional trust; and v) institutionalized trust. these factors are being measured with the statements which had been scaled down and surveyed on the stakeholders. these are described in the path diagram in figure 2 table 1: rotated component matrix. component sociopolitical legitimacy procedural fairness economic legitimacy interaction al trust institution al trust this project keeps harmony with the day to day life of the local community 0.735 compatible with the locality 0.700 the project has same vision for the future 0.621 -0.431 this project accounts for the interest of the community 0.749 the project activities are done in consultation with community 0.677 the project is fair in budgeting and the construction process 0.551 the project will decrease traffic congestion as well as accidents 0.535 0.530 the project shares decision-making with the community 0.445 land price and house rent will increase in the locality 0.793 transportation cost will be decreased 0.651 mitigating measures for the negatively affected people -0.465 -0.621 will not create environmental and other problems 0.712 0.413 the project is not harmful for the local tradition/culture 0.638 the project treats everyone fairly -0.546 the project will bring wellbeing to the whole region 0.862 extraction method: principal component analysis; rotation: varimax with kaiser normalization. 6 md manjur morshed et al. measuring social license to operate ……… figure 2: path diagram 4.2 path analysis to investigate the independent relationships more systematically between the variables, the path analysis is performed. using spss amos 21, the path model was run and analyzed. for having a small sample size, item scores for each variable had been averaged and used as an observed variable in the model. the path diagram presents the model with all the correlations among the statements and factors with errors as well as a relation among the factors. from the path diagram, it is seen that all the five factors are representing their relations with the acceptance of the road expansion project and with the selected statements also. the model data had been standardized to relate the factors to the acceptance of the project. among all the five factors, the only journal of engineering science 12(2), 2021, 01-10 7 institutional trust had shown negative loading (-0.25), which means the stakeholders of the locality are declining the project for not fulfilling their institutional trust. this factor is highly dependent on the regional wellbeing statement. among all the factors, the socio-political legitimacy has shown the highest relationship with acceptance (0.81). the factor itself depends on three statements and all of them have shown positive loading. the factor is related to the other factors also. even though the relation of socio-political legitimacy with institutional trust is negative, the other factors have positive relation over this factor. this represents that, the expansion project is gaining slo over socio-political legitimacy. the second most related factor to the acceptance is economic legitimacy with 0.65 value. though two of the statements have shown negative loading with the factor, the relation with the acceptance is positive. this represents that the stakeholders are accepting the expansion project in their locality. economic legitimacy has shown positive relation with socio-political legitimacy and interactional trust. for not having proper institutional trust and procedural fairness, stakeholders are giving low acceptance over the economic legitimacy. when it comes to procedural fairness and interactional trust factor, both of them have shown positive relations with the acceptance of the project (0.10 and 0.05). procedural fairness has shown positive loading with all the 5 statements under it (gained from factor analysis). on the other hand, interactional trust has shown relation with two statements negatively and one positively. though the relation is low, the stakeholders are positive to slo over these two factors. 4.3 model fitness the goodness of fit of the path model had been assessed using chi-square test, the comparative fit index (cfi), normed fit index (nfi) and root mean error of approximation (rmsea) (hu and bentler, 1999; kenny and mccoach, 2003). chi-square statistics had been used for non-significant values. however, chi-square statistics remains significant and thus absolute fit index and incremental fit index need to be checked. spss amos 21 calculate goodness of fit and badness of fit automatically. on the other hand, for incremental fit index, rmr, gfi, cfi, agfi, nfl test have been checked for badness of fit. for absolute fit index, rmsea (root mean square error of approximation) is been checked in this study. 4.4 chi-square statistics the saturated model shows the best result and the independence model represent the worst result of a model (table 2). npar is the number of parameters in the model. in the saturated model, there are 136 parameters. for the tested (default) model there are 45 parameters. for the independence model, there are 16 parameters where all of the paths have been deleted. cmin is the actual magnitude and df is the degree of freedom. the chi-square value is called cmin. if the chi-square value is significant, the model is regarded as valid. if cmin/df is less than 3.00 then the indicators of model fitness will be good. the value of cmin/df of this default model is 1.192, which is less than 3.00 and it represents a good fitness of the model. table 2: npar, cmin, df & cmin/df of the model model npar cmin df p cmin/df default model 45 108.499 91 0.102 1.192 saturated model 136 0.000 0 independence model 16 247.132 120 0.000 2.059 4.5 rmr, gfi, cfi and nfi rmr indicates the badness of fit and gfi shows the goodness fit of a model. the root mean square residual (rmr) is an index of the amount by which the estimated (by tested model) variance and covariance differ from the observed variance and covariance. for the saturated model, it will be a perfect 0 and 0.05 will be tolerable but not more than 0.1. the standardized rmr should be equal to or less than 0.10 for “a good fitting” model. in this tested model, the value of rmr is 0.10 which showed a moderate fitness of the model. gfi, the goodness of fit index, is the proportion of the variance in the sample variance, and the covariance matrix is accounted for by the model. gfi varies from 0 to 1 but theoretically can yield meaningless negative values. gfi deals with errors in reproducing the variance-covariance matrix. gfi often runs high compared to fit model, using 0.95 as the cutoff, more than 0.95 gives the best result of a model. gfi should be equal to or 8 md manjur morshed et al. measuring social license to operate ……… greater than 0.90 to accept the model. but value between 0.80 and 0.90 of gfi value is also granted and accepted as moderate fit. agfi (adjusted gfi) is an alternate gfi index in which the value of the index is adjusted for the number of parameters in the model. the value of agfi is always less than gfi and the agfi of the default model is 0.70. pgfi (p is for parsimony), the index is adjusted to reward simple models and penalized models in which few paths have been deleted (table 3). table 3: rmr, gfi, agfi, pgfi, nfi and cfi of the model model rmr gfi agfi pgfi nfi delta1 cfi default model 0.107 0.800 0.701 0.535 0.561 0.862 saturated model 0.000 1.000 1.000 1.000 independence model 0.185 0.559 0.500 0.493 0.000 0.000 this goodness of fit indices compares the tested model to the independent model rather than to the saturated model. the normed fit index (nfi) is simply the difference between the two models’ chi-squares divided by the chi-square for the independence model. for our data, that is 0.56. value of 0.9 or higher indicates a good fit. but the value of nfi of this model does not indicate a good fit. the comparative fit index (cfi) uses a similar approach (with a noncentral chi-square) and is said to be a good index. it ranges from 0 to 1, like the nfi, and 0.95 (or 0.9 or higher) indicates a good fit. the cfi value of the default model is 0.86 which seems be a moderate fitness of the model (table 3). 4.6 rmsea the root mean square error of approximation (rmsea) estimates lack of fitness compared to the saturated model. rmsea of 0.05 or less indicates a good fit and 0.08 or less adequate fit. lo 90 and hi 90 are the lower and upper ends of a 90% confidence interval on this estimate. pclose is the p-value testing the null that rmsea is no greater than 0.05. rmsea value of this tested model is 0.06, meaning the model is tolerable and indicating a moderately good fit (table 4). table 4: root mean square error of approximation (rmsea) of the model model rmsea lo 90 hi 90 pclose default model 0.067 0.000 0.111 0.290 independence model 0.157 0.129 0.185 0.000 5. discussion among the five factors, socio-political legitimacy is comprised of three statements defining the slo (table 1). all the statements have positive loadings. however, the project is able to be in harmony with the day-to-day life of the community has the highest factor loading (0.735). this is followed by project compatibility with the locality (0.7) and by the similarity of vision for the future between the stakeholders (0.621). however, when comes to the project’s economic legitimacy, the similarity of vision has negative loading (-0.431), meaning that people are not convinced that the project will benefit economic benefit for all stakeholders. musiyarira et al. (2020) highlight the key challenges of securing slo in developing country context, where companies engaged in project activities grossly ignore slo and secure projects through lobbying resulting in massive corruption. the above findings that the project stakeholders aspire for such development project, they are not convinced of personal economic benefit from the road expansion project. secondly, the procedural fairness has a total of 5 statements. accounting for the interest of the community (0.749), having proper consultation while implementing the project activities (0.677), fairness in the budget and construction process (0.551), decreasing traffic congestion (0.535), and sharing decision making with the community (0.445) – all these have shown positive loadings. these indicate that the project is getting slo over procedural fairness. the road expansion project will decrease traffic congestion, as well as accidents, has shown a positive loading in the interactional trust factor also (0.530), which has almost the same loading as the procedural fairness. this indicates that people are taking decreasing traffic congestion as positivity of the road expansion project. however, road expansion projects can have broader consequences such as, increasing traffic volume as it connects the asian highway and khulna railway station, and loss of local jobs. therefore, broader journal of engineering science 12(2), 2021, 01-10 9 consequences of infrastructure projects are often out-of-sight which the local stakeholders fail to appreciate (see also, brueckner and eabrasu, 2018). thirdly, three statements underline the economic legitimacy factor and two of those have shown positives whereas one negative loading. land price and house rent will gain a significant boost (0.793) and transportation cost will be decreased (0.651) show that people are accepting the road expansion project in their locality. however, mitigating measures for the negatively affected people have negative loading, an indication that the project has less socio-political and economic legitimacy. from the in-depth survey, it was found that very few measures are taken to remedy, such as resettlement and rehabilitation, due to eviction caused by the road expansion project. however, as a whole, the project has slo in terms of economic legitimacy. fourthly, three statements have shown loadings under the interactional trust. among those three, the project will not create environmental and other problems (0.712) and not harmful to the local tradition/culture (0.638) have shown positive relation with the interactional trust factor. however, fair treatment by the road expansion in terms of eviction and resettlement received negative factor loading. such negative loading is an indication that project stakeholders are restrained to approve the project due to the biases in treatment by the implementing authority. finally, the institutional trust, that is, the project will bring wellbeing to the whole region (0.862). the project stakeholders are accepting the expansion project over this statement and relying on the institutionalized trust factor. another factor that has shown positive loading is that the project will not create environmental and other problems (0.413), thus reflecting positive slo of the road expansion project. however, such positivity can also be related to a lack of awareness among the stakeholders as the project is expected to generate larger traffic volume (see also, brueckner and eabrasu, 2018). 6. conclusions in this research we have used the thomson boutilier model to identify major factors affecting the slo of a road expansion project in khulna city, bangladesh. the study suggests that socio-political legitimacy, procedural fairness, economic legitimacy, interactional trust and institutional trust determine the slo of the project. the varimax rotated component matrix has shown the dependency of the statements on various factors. secondly, the path diagram presented all the co-relations with the factors and the acceptance of the project by the stakeholders. socio-political legitimacy has shown the highest determinant for slo of the project. the project has also achieved economic legitimacy, procedural fairness and interactional trust of the stakeholders. however, the project has failed to achieve institutional trust, which is also negatively co-related to economic legitimacy. in conclusion, the project needs to gain institutional trust for higher slo. slo is not a commonly used tool for infrastructure projects in bangladesh. this paper has presented a model for understanding slo with limited data. a comprehensive data collection can significantly improve the model, thus can be used as a tool for large infrastructure projects in bangladesh. additionally, including the slo concept in the project planning and implementation phase can significantly improve the sustainability of infrastructure projects in bangladesh. acknowledgement this research project was funded by the ugc-kuet research grant. the first author was the principal investigator. the second author worked as a research assistant during the project period. the third author collected data as well as did her bachelor dissertation on this very topic, thus making this publication a part of her dissertation. we thank the stakeholders who participated in the questionnaire survey. we acknowledge the two anonymous reviewers and editor of journal of engineering science for their support in the publication process. references australian government, 2006. community engagement and development: leading practice sustainable development program for the mining industry. australian government, canberra, australia. bhpb, 2011. our future: sustainability report 2011. bhpb, melbourne, australia. 10 md manjur morshed et al. measuring social license to operate ……… boutilier, r. g., 2014. frequently asked questions about the social licence to operate. imp ass proj appr, 32, pp. 263-272. boutilier, r. g., thomson, i., 2011. modelling and measuring the slo. in invited paper presented at: the social license to operate seminar. brisbane: centre for social responsibility in mining. university of queensland, 2011. martin brueckner, m., eabrasu, m., 2018. pinning down the social license to operate (slo): the problem of normative complexity. res pol, https://doi.org/10.1016/j.resourpol.2018.07.004 csiro, 2013. social license to operate: minerals down under. csiro, brisbane, australia. de jong, w., humphreys, d. 2016. a failed social licence to operate for the neoliberal modernization of amazonian resource use: the underlying causes of the bagua tragedy of peru. fores, 89, pp. 552-564. http://dx.doi.org/10.1093/forestry/cpw033. dhaka tribune, 2020. road expansion project resumes, to alleviate commuter suffering. https://www.dhakatribune.com/bangladesh/nation/2020/12/19/road-expansion-project-resumes-toalleviate-commuter-suffering (accessed: 22 may 2021) harvey, b., bice, s. 2014. social impact assessment, social development programmes and social licence to operate: tensions and contradictions in intent and practice in the extractive sector. imp ass proj appr, 32, pp. 327335. general economics division (ged), 2017. fy2016-fy2020: accelerating growth, empowering citizens (final draft). planning commission government of the people’s republic of bangladesh. jijelava, d., vanclay, f. 2018. how a large project was halted by the lack of a social licence to operate: testing the applicability of the thomson and boutilier model. env imp ass rev, 73, pp. 31-40. hanna, p., vanclay, f., langdon, e.j., arts, j., 2016. conceptualizing social protest and the significance of protest action to large projects. extr ind soc, 3, pp. 217-239. hu, l. t., bentler, p. m, 1999. cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. structural equation modeling: a multidisciplinary journal, 6(1), pp. 155. kurlander, l.t., 2001. newmont mining: the social license to operate. proceedings of the global executive forum, university of colorado denver. [online] available at: www. ucdenver.edu/academics/internationalprograms/ciber/globalforumreports/ documents/newmont_mining_social_license.pdf (accessed 20 may, 2021) mcnab, k., onate, b., brereton, d., horberry, t., lynas, d., franks, d.m., 2013. exploring the social dimensions of autonomous and remote operation mining: applying social license in design. csiro, brisbane, australia. meesters, m., wostyn, p., van leeuwen, j., behagel, j. h., turnhout, e., 2020. the social licence to operate and the legitimacy of resource extraction. curr opin in env sust, 45, pp. 7–11. mitchell, p. top 10 business risks and opportunities – 2020. ernst & young global mining & metals leader. [online]. available at: https://www.ey.com/en_gl/mining-metals/ 10-business-risks-facing-mining-andmetals 2019 (accessed: 20 june 2020). moffat, k., zhang, a. 2014. the paths to social license to operate: an integrative model explaining community acceptance of mining. res pol, 39, pp. 61-70. musiyarira, h. k., shava, p., dzinomwa, g. the extractive industries and society. https://doi.org/10.1016/j.exis.2020.05.020 owen, j., kemp, d. 2013. social licence and mining: a critical perspective. res pol, 38(1), pp. 29-35. http://dx.doi.org/ 10.1016/j.resourpol.2012.06.016. slack, k., 2009. mining conflicts in peru: condition critical. oxfam america, boston. thomson, i., boutilier, r. g., 2011. social license to operate. sme min engg handbook, 1, pp. 1779-1796. vanclay, f. 2017. principles to gain a social licence to operate for green initiatives and biodiversity projects. curr opin environ sust, 29, pp. 48-56. © 2021 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 5_jes-env 4602-new journal of engineering science 12(3), 2021, 45-55 doi: https://doi.org/10.3329/jes.v12i3.57478 *corresponding author:adibankon50@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) water quality index (wqi) of shitalakshya river near haripur power station, narayanganj, bangladesh rumman mowla chowdhury1, adib ashhab ankon*2, md kamruzzaman bhuiyan3 1 assistant professor, university of asia pacific, dhaka, bangladesh, e-mail: rumman@uap-bd.edu 2 teaching assistant, university of asia pacific, dhaka, bangladesh, e-mail: adibankon50@gmail.com 3 research assistant, university of asia pacific, dhaka,bangladesh, e-mail: bhuiyan8769@gmail.com received: 06 may 2021 accepted: 01 november 2021 abstract the present investigation is aimed at understanding the water quality parameters and the findings of a water quality index (wqi) to assess the characteristics of the shitalakshya river near haripur power station, narayanganj for five different years (2013-2018) considering monsoon, pre-monsoon, post-monsoon seasonal variations. in this study, three different methods were used to evaluate the wqi named as; weighted arithmetic index method, canadian council of ministers of the environment (ccme) wqi method and national sanitation foundation (nsf) method. essential parameters i.e. dissolved oxygen, ph, chloride, turbidity, color, biochemical oxygen demand, total dissolved solids, silica, iron, electrical conductivity, phosphate were considered for calculating the wqi. according to weighted arithmetic index method, the wqi value varied from 80 to 286 for the last five years. from the national sanitation foundation method, the wqi value was found within 36 to 56 for the study duration. the wqi value was varied from 3 to 16 according to the canadian council of ministers of the environment water quality index method. based on wqi values, the shitalakhya river water was being classified as poor water for the above-mentioned different years. among the different parameters, mostly turbidity, electrical conductivity, tss, iron were the parameters that caused the situation worst. keywords:water quality index, dissolved oxygen, total dissolved solids, biochemical oxygen demand. 1. introduction in recent decades,the rapid growth of population and the faster pace of modernization have resulted in a massive increase in demand for freshwater (weeks, 2020). because of the easiest accessibility, the majority of industries are emerging alongside the river. furthermore, the cost of transporting raw resources is substantially lower when done by water. however, because the majority of companies do not have effluent treatment facilities (etp), their effluents are finally dumped straight into rivers without being treated. as a result, resource deterioration and pollution have progressed and continued, particularly in the area of surface water. polluted water is a major route for disease transmission. every year, over 1.8 million people, largely children, die in underdeveloped nations as a result of water-related illnesses (world health organization, 2004). for healthy living, it is critical to have access to safe drinking and usable water (pal et al., 2018). a sufficient quantity of clean and pure drinking water is a basic requirement for all humans. moreover, riverine water quality is a major concern for all users since it affects human activities as well as plant and animal life (goyette et. al., 2016). a wide range of indices has been established to summarize water quality data in a way that is both easy to explain and understand. the water quality index (wqi), which assesses water quality using computed values, is one of the most effective ways to convey water quality. water quality is determined by its physical, chemical, and biological characteristics (chowdhury and hossain, 2012). the wqi is a measure that indicates the combined impact of various water quality parameters (sahu and sikdar 2008). the wqi is a simple and effective approach for determining the appropriateness of water. it's also a great way to get information out to concerned individuals and politicians about water performance and satisfaction. as a response, it becomes an important parameter for water quality evaluation and management (both surface and groundwater). wqi indicates the combined effect of several water quality parameters and is computed from the viewpoint of (both surface and groundwater) acceptability for human consumption. wqi was first established by horton in 1965 and then modified by brown in 1970. the objective of this paper is to determine the wqi of water of the shitalakshya river near haripur power station, narayanganj for five different years (2013-2018). figure 1 shows the research region with the power plant on a topographical map. shitalakhya is the river that is regarded as one of the feeders of the brahmaputra, jes an international journal 46 r. m. chowdhury et al. water quality index (wqi) of shitalakshya……… and the river's initial flow was southwest. after then, it heads east to narayanganj and then south to dhaleswarinear kalagachhiya. near narayanganj, the river stretches for over 110 kilometers (68 miles) and has a width of 300 meters.the brahmaputra river was separated by the shitalakhya river, which subsequently flowed into dhaleshwari. many industries and companies have sprung up along the shitalakhya river's bank as a result of its strategic position. however, these industries do not even follow or implement wastewater and hazardous water treatment methods. as a result of the inefficient discharge method, a significant volume of hazardous and effluent has been mixed up in the shitalakhya river. besides, domestic and urban runoff sludge from the narayanganj urban regions is untreated and dumped into this river. hence, heavy metals, as well as numerous harmful compounds, are carried out by industrial wastes and effluents, and pollution is increasing at an increasing pace day by day (warpo, 2000). figure 1: location of main discharges on sitalakhya river, narayanganj. moreover, this place has been chosen by the government for establishing a power plant for electricity generation. as a result, a 412 mw combined cycle power plant was built here in 2014, and it is still operational today. power plants need a huge amount of water for cooling purposes and therefore power plants are built beside the river. but due to the pollution over the river, power plants required extensive levels of treatment before using river water. subsequently, this increases the cost of treatment units so the rise in the overall cost of power production affects the economy. moreover, local inhabitants of this area are dependent on the water of the shitalakshya river for various purposes which made the analysis inevitable. the single value of the water quality index value will be useful for understanding the actual situation. moreover, the trend in seasonal variation might help different stakeholders for stepping towards necessary actions. foremost, the result might be beneficial for them to decide for treating the worst parameters especially for the time of shortage in the municipal supply water. several wqi’s have been created and utilized correctly by government organizations and researchers throughout the years. the index of river water quality, the overall index of pollution, the chemical water quality index, the iowa water quality index, the universal water quality index-uwqi, the canadian council of ministers of environment water quality index-ccmewqi, and the universal water quality index-uwqi, nsfwqi is for national sanitation foundation water quality index, owqi is for oregon water quality index, and wawqim is for weighted arithmetic water quality index method. the ccmewqi, nsfwqi, owqi, and wawqim are the most regularly utilized of these (oni and fasakin,2016). the wqi was evaluated in this study using three distinct methods: the weighted arithmetic index method, the canadian council of ministers of the environment water quality index method, and the national sanitation foundation method. the index results were computed using water temperature, ph, dissolved oxygen (do), total dissolved solids (tds), total suspended solids (tss), electrical conductivity (ec), hardness (ca and mg), chloride, turbidity, alkalinity, iron, and color. journal of engineering science 12(3), 2021, 45-55 47 2. materials & methodology 2.1 sample collection throughout the year, samples are taken from the power plant's outflow and examined the parameters (ph and temperature) instantly at the sampling location using standard equipment.other major water quality parameters (do, color, conductivity, tds, tss, hardness, cl-, turbidity, alkalinity, fe) were measured in the power plant's laboratory. water samples were taken once a week from haripur (figure 1) in bangladesh's narayanganj district. except for 2017, the analysis spanned five years, from january 2013 to december 2018. the water samples were collected from the shitalakshya river's surface water. 2.2 methodological approach figure 2: methodological approach of the study. methodological flow chart of the study is shown in figure 1. 2.3 weighted arithmetic water quality index method the wqi, which is generated using the weighted arithmetic index algorithm (wawqim), is often utilized by academics in poor nations when data-gathering infrastructure is lacking and valid rating curves are few using the most routinely measured water quality variables, the weighted arithmetic water quality index approach categorized the water quality according to the degree of purity. various scientists have utilized the methodology extensively (balan, shivakumar& kumar, 2012) and the following equation was used to calculate the wqi (brown, mcclelland, deininger& o’connor, 1972): 𝑊𝑄𝐼 = ∑ 𝑄 𝑊 ∑ 𝑊 … … … … … … … … … … … … … . (1) this equation is used to compute the quality rating scale (qi) for each parameter: 𝑸𝒊 = 𝟏𝟎𝟎[(𝑽𝒊 − 𝑽𝟎𝑺𝒊 − 𝑽𝟎)] … … … … … … … . (𝟐) where, vi is the ith parameter's estimated concentration in the examined water. in pure water, vo is the optimal value for this measure. except for ph = 7.0 and do = 14.6 mg/l, vo = 0. this parameter's suggested standard value is si. 48 r. m. chowdhury et al. water quality index (wqi) of shitalakshya……… the following formula is used to compute the unit weight (wi) for each water quality measure: 𝑾𝒊 = 𝑲 𝑺𝒊 … … … … … … . . (𝟑) where, k = the proportionality constant can alternatively be determined using the equation below: 𝐾 = 1 ∑ … … … … … … . (4) table 1 shows the water quality rating according to this wqi. table 1: water quality rating as per weight arithmetic water quality index method. wqi value rating of water quality 0-25 excellent water quality 26-50 good water quality 51-75 poor water quality 76-100 very poor water quality above 100 unsuitable for drinking waterfor supply afterconventional treatment 2.4 national sanitation foundation water quality index (nsf wqi) a common water quality index technique was created by picking criteria with considerable care, constructing a standard scale, and allocating weights. the national sanitation foundation (nsf) backed the effort, which was dubbed nsfwqi to compute the wqi of numerous dangerously contaminated water bodies. ph, temperature, dissolved oxygen, turbidity, fecal coliform, biochemical oxygen demand, total phosphates, nitrates, and total solids are among the nine water quality characteristics used in the proposed method for comparing the water quality of multiple water sources (brown, mcclelland, deininger&tozer, 1970). the data on water quality is captured and transferred to a weighting curve chart, from which a numerical value of qi is calculated. nsf wqi has the following mathematical expression: 𝑊𝑄𝐼 = 𝑄 𝑊 … … … … … … (5) where qi is the ith water quality parameter's sub-index. wi denotes the weight assigned to each water quality metric. the number of water quality parameters is denoted by the letter n. the ratings of water quality for this nsfwqi method have been defined using the following table 2: table 2: water quality rating as per national sanitation foundation water quality index method. national sanitation foundation method (nsf wqi) wqi value rating of water quality 91-100 excellent water quality 71-90 good water quality 51-70 medium water quality 26-50 bad water quality 0-25 very bad water quality 2.5 canadian council of ministers of the environment water quality index (ccme wqi) the ccme wqi is a standardized technique developed by canadian jurisdictions to communicate water quality information to both managers and the general public. moreover, wqi was created by a committee under the canadian council of ministers of the environment (ccme), and it may be used by many water authorities in other nations with minor modifications (boyacioglu, 2010). this method was designed to assess surface water for aquatic life protection in compliance with particular requirements. the parameters associated with various measurements may differ from one station to the next, and the sampling technique calls for at least four journal of engineering science 12(3), 2021, 45-55 49 parameters to be sampled four times (khan, tobin, paterson, khan, & warren, 2005) the following relationship can be used to calculate index scores in the ccme wqi method: 𝑊𝑄𝐼 = 100 − 𝐹 + 𝑥𝐹 + 𝐹 1.732 … … … … … . . (6) where scope (f1) is the number of variables whose objectives have not been reached. f1 = [number of unsuccessful variables divided by the total number of variables] frequency (f2) is the number of times the objectives are not reached out of a total of 100. f2 = [number of failed tests divided by the total number of tests] amplitude (f3) = the percentage of time that the objectives are not reached. a) 𝐸𝑥𝑐𝑢𝑟𝑠𝑖𝑜𝑛𝑖 = [𝐹𝑎𝑖𝑙𝑒𝑑𝑡𝑒𝑠𝑡𝑣𝑎𝑙𝑢𝑒𝑖 /𝑂𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒𝑗] − 1. 𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑧𝑒𝑑𝑠𝑢𝑚𝑜𝑓𝑒𝑥𝑐𝑢𝑟𝑠𝑖𝑜𝑛𝑠, 𝑛𝑠𝑒 = ∑ 𝑒𝑥𝑐𝑢𝑟𝑠𝑖𝑜𝑛 𝑁𝑜. 𝑜𝑓𝑇𝑒𝑠𝑡𝑠 b) 𝐹 = [nse/0.01nse + 0.01] five categories have been suggested to categorize the water qualities which are summarized in table 3. table 3: water quality rating as per canadian council of ministers of the environment water quality canadian council of ministers of the environment water quality index (ccme wqi) wqi value rating of water quality 95-100 excellent water quality 80-94 good water quality 60-79 fairwater quality 45-59 marginal water quality 0-44 poor water quality 3. results & discussion 3.1 assessment of quality using different wqi methods the water quality index of the shitalakshya river was estimated using three distinct methodologies in this study, including the weighted arithmetic index method, the canadian council of ministers of the environment water quality index method, and the national sanitation foundation method. table 4 shows the maximum, minimum, mean, variance, and standard deviation. the standard deviation and variance show that there is a lot of variation in the values of several metrics that affect the river's water quality. table 5 illustrates the correlation matrix of water quality parameters. table 4: maximum, minimum and average values of different water quality parameters. parameter unit ecr-97 (standards) maximum value minimum value mean value variance standard deviation ph 6.5 – 8.5 7.87 7.13 7.47 0.14 0.37 do mg/l 6 5.87 0.23 2.88 0.48 0.69 color pt-co 15 52.50 2.73 20.27 4.72 2.17 conductivity µ/c 550 1151.75 130.50 519.89 680.29 26.08 tds mg/l 1000 637.25 64.40 277.45 1350.76 36.75 tss mg/l 10 136.25 20.20 63.69 30.55 5.53 hardness mg/l 200 – 500 196.50 38.50 107.21 27.32 5.23 50 r. m. chowdhury et al. water quality index (wqi) of shitalakshya……… parameter unit ecr-97 (standards) maximum value minimum value mean value variance standard deviation clmg/l 150-600 147.50 9.20 50.72 141.87 11.91 turbidity ftu 10 146.50 17.60 50.57 37.03 6.08 alkalinity mg/l 128 450.00 35.20 170.08 606.98 24.64 fe mg/l 0.3 – 1.0 10.25 0.02 0.52 0.12 0.34 temperature ℃ 20 – 30 32.08 21.20 28.20 1.26 1.12 bod mg/l 0.2 6.5 2.8 4.9 0.18 0.42 the environment conservation rules, 1997 (ecr-97) establish numerous conditions for acceptable water. ecr-97's standard value is shown in table-4. the tested water quality parameters were then compared to the standard values to determine the existence of being within the limit. table 4 shows that the ecr-97 standards for ph and temperature were satisfied. the average ph is 7.47, with the ecr-97 norm ranging from 6.5 to 8.5. although the maximum temperature was 32.08°c and the usual temperature is 20-30°c, the average water temperature was measured to be 28.20°c. the ecr-97 water hardness limits range from 200 to 500 mg/l, and this study found a mean value of 107.21 mg/l, which is near the limit. color, tss, and bod levels were substantially high in comparison, whereas do rates were significantly lower than the normal threshold. in addition, the variation and standard deviation (sd) of numerous water quality indices are shown in table 4. simply explained, variance is a measure of the dispersion of numbers concerning their average value. the standard deviation, on the other hand, is the degree of variance or dispersion of a collection of numbers. variance and standard deviation are used to determine whether or not a parameter remains consistent throughout the year. the ph, do, temperature, and bod variance readings were found to be near zero in this investigation. this indicates that these characteristics are rather consistent throughout the year. however, conductivity, tds, and alkalinity all show extraordinarily high variation values, indicating that these parameters fluctuate over time or are impacted by other factors. table 5: correlation matrix of water quality parameters of shitalakshya river. p h d o c o lo r c o n d u ct iv it y (c ) t d s t s s h ar d n es s (h ) c l t u rb id it y ( t u ) a lk al in it y (a lk ) ir o n ( f e) t em p er at u re (t ) b o d ph 1 do 0.84 1 color 0.13 -0.03 1 c 0.32 -0.05 0.43 1 tds 0.26 0.15 0.12 0.84 1 tss 0.11 0.41 0.36 -0.60 -0.58 1 h 0.35 0.17 0.92 0.60 0.38 0.25 1 cl-0.24 -0.25 0.01 0.65 0.84 -0.67 0.09 1 tu -0.05 0.12 0.76 0.04 0.05 0.65 0.76 -0.06 1 alk -0.38 -0.73 0.06 0.65 0.48 -0.82 0.10 0.61 -0.23 1 fe -0.08 -0.13 0.95 0.39 0.18 0.34 0.83 0.20 0.80 0.09 1 t 0.47 0.47 -0.19 0.09 0.27 0.03 0.17 -0.19 0.11 -0.10 -0.34 1 bod -0.60 -0.37 -0.78 -0.48 -0.24 -0.32 -0.92 0.20 -0.62 0.06 -0.59 -0.40 1 journal of engineering science 12(3), 2021, 45-55 51 3.1.1 wqi by weighted arithmetic method the water quality of the river might also fluctuate due to seasonal variations. the number of water quality factors changes as the weather changes. variations in sunlight, dissolved oxygen, water temperature, and other factors may induce variations in water quality over sessions. to determine the wqi by weighted arithmetic method (wam), the sub-water quality index for various parameters was estimated. the bar chart (figure 3) compares the seasonal water quality index values for different years. it was found most of the water quality parameters exceed the permissible limit throughout the year. the worst scenario was visible in the post-monsoon season for most of the year. however, water quality parameters were slightly better in the monsoon period which eventually made the index value barely within the limit to be considered as good water quality. according to the rating of arithmetic index value (table-1), only the monsoon season of 2018 showed good water quality. figure 3: seasonal variation in the wqi determined by weight arithmetic method. the bar chart contrasts the seasonal water quality index method for several years as determined by the weighted arithmetic method. throughout the year, most of the water quality metrics were determined to be over permitted levels.according to the arithmetic method, the value of wqi must be as low as feasible to be considered drinking water quality; nevertheless, as the value of wqi rises, the quality will deteriorate from poor to poorer. the method states that when the value of wqi exceeds 100, the water quality will be unsuitable for drinking water supply after conventional treatment, and the figure demonstrates that the value of wqi surpasses 100 in most of the seasons of most of the years. the average value is likewise more than 80.whereas the worstcase scenario was observable for the majority of the year in the post-monsoon season.however, water quality metrics were marginally improved during the monsoon period (2018), resulting in an index value that was just under the limit to be called satisfactory water quality. 3.1.2 wqi by nsf figure 4 shows the seasonal variation in water quality for the different years by the nsf method. according to the nsf method the water quality is degrading along with time. among the 5 different years, the scenario was awful for the year 2018 almost throughout the year. figure 4: seasonal variation in the wqi determined by national sanitation foundation water quality index (nsf) method. 52 r. m. chowdhury et al. water quality index (wqi) of shitalakshya……… as per the nsf method, the value of wqi must be as high as possible to be called drinking water quality; however, as the value of wqi lowers, the quality deteriorates from bad to worse. according to the method, when the wqi number falls below 25, the water quality is very poor. the nsf standard, which defines water quality as 'bad water quality,' was used in the study to determine the value of wqi between 30-50. 3.1.3 wqi by ccme following table 6 shows the calculation of factors for the ccme method for the year 2018. it was found mostly there were three to seven parameters among the twelve parameters which were failed to be within the permissible limit. for the particular year, april was the month which falls within pre-monsoon experienced the poorest quality. according to the following method, f2 expresses the fact that mostly the parameters were far away from the standard values in post-monsoon. table 6: results of wqi founded by ccem method. month failed item total item f1 no. of failed test total test f2 total excursion nse f3 ccme wqi jan 5 13 38.46 23 55 41.8 87.57 1.59 100.01 12.60 feb 5 13 38.46 20 44 45.5 102.13 2.32 100.01 11.56 mar 5 13 38.46 20 44 45.5 159.91 3.63 100.01 11.56 apr 7 13 53.85 21 44 47.7 135.10 3.07 100.01 6.38 may 3 13 23.08 12 44 27.3 21.23 0.48 100.01 19.30 jun 3 13 23.08 9 33 27.3 14.58 0.44 100.01 19.30 jul 6 13 46.15 19 47 40.4 42.95 0.91 100.01 10.85 aug 4 13 30.77 16 44 36.4 54.45 1.24 100.01 15.83 sep 5 13 38.46 17 47 36.2 59.14 1.26 100.01 14.07 oct 5 13 38.46 18 47 38.3 27.67 0.59 100.01 13.54 nov 6 13 46.15 20 45 44.4 61.81 1.37 100.01 9.75 dec 6 13 46.15 22 46 47.8 85.11 1.85 100.01 8.76 in this table, it has been noticed that in the monsoon period number of failures is respectively low, on the other hand in pre-monsoon and post-monsoon the number of failures is high. figure 5: seasonal variation in the wqi determined by the canadian council of ministers of the environment water quality index (ccme) method. figure 5 compares the result for different seasons using the ccme method. for the year 2014, the pre-monsoon season was awful according to table 3.according to the ccme method, the value of wqi must be as high as achievable to be referred to as drinking water quality; unfortunately, as the value of wqi declines, the quality deteriorates from bad to worse. when the wqi value goes below 44, the water quality is considered very poor, journal of engineering science 12(3), 2021, 45-55 53 according to the technique. in the study, the ccme technique, which defines water quality as "bad water quality," was utilized to calculate the value of wqi less than 16. 3.2 comparison of wqi between different methods table-7 compares the seasonal water quality index method for different years for different methods. there was almost no variation among the three different methods for assessing water quality index values. the water quality of the specific river water was found to be unsatisfactory for household, drinking, and aquatic species in practically every season and every technique. however, rendering the method wai, the quality showed good whereas for the same season other two methods showed the opposite result. table 7: wqi value for the period of 2013-2018 according to different methods considering the corresponding rating. w q i m et h o d s ea so n 2013 2014 2015 2016 2018 wqi valu e wqi rating wqi valu e wqi rating wqi valu e wqi rating wqi value wqi rating wqi valu e wqi rating w a i p re -m o n so o n 80 very poor water quality 112 unsuitabl e for drinking water for supply after conventio nal treatment 133 unsuitab le for drinking water for supply after conventi onal treatment 92 very poor water quality 151 unsuitable for drinking water for supply after convention al treatment m o n so o n 77 very poor water quality 77 very poor water quality 56 poor water quality 66 poor water quality 87 good water quality p o st m o n so o n 104 unsuitab le for drinking water for supply after conventi onal treatment 286 unsuitabl e for drinking water for supply after conventio nal treatment 104 unsuitab le for drinking water for supply after conventi onal treatment 84 very poor water quality 108 poor water quality c c m e p re -m o n so on 6 poor water quality 3 poor water quality 9 poor water quality 10 poor water quality 13 poor water quality m o n so o n 16 poor water quality 14 poor water quality 15 poor water quality 14 poor water quality 13 poor water quality p o st m o n so o n 10 poor water quality 5 poor water quality 9 poor water quality 10 poor water quality 10 poor water quality n s f p re -m o n so o n 36 bad water quality 35 bad water quality 39 bad water quality 40 bad water quality 38 bad water quality 54 r. m. chowdhury et al. water quality index (wqi) of shitalakshya……… w q i m et h o d s ea so n 2013 2014 2015 2016 2018 wqi valu e wqi rating wqi valu e wqi rating wqi valu e wqi rating wqi value wqi rating wqi valu e wqi rating m o n so o n 50 bad water quality 52 medium water quality 55 medium water quality 56 mediu m water quality 46 bad water quality p o st m o n so o n 40 bad water quality 41 bad water quality 42 bad water quality 39 bad water quality 37 bad water quality 4. conclusions in this study, the samples were collected from the shitalakshya river once a week at the outlet point of the haripur power plant throughout the year. assessment of shitalkhya river water quality was done for the past five years (january 2013 to december 2018). the comparison was shown considering different seasons; premonsoon, monsoon, post monsoon. water quality parameters such as ph, do, bod, ec, color, turbidity, hardness and some minerals were examined for the evaluation. the main purpose of the research work was to assess the water quality utilizing different water quality index methods; three widely used methods; (wai method, nsf method, ccme method) were used to calculate the wqi. this type of surface water rating might aid individuals in having a clear grasp of the water quality state to make better decisions about its future use. besides, wqi combines the combined effects of many water quality metrics and disseminates water quality data to the general public and legislative decision-makers. after assessing the results, the study reveals that the quality of the shitalakshya river possesses poor water quality. the results were similar for the three different methods which proved the validity of the result. moreover, the water quality status was almost similar throughout the year regardless of seasonal variation. among the different parameters, mostly turbidity, electrical conductivity, tss, iron were the parameters that caused the situation worst. this will eventually affect the aquatic ecosystems, recreational and industrial use. consequently, fish culture has been defused due to this condition. furthermore, the cost of treatment of water to be used in industries is dramatically increasing. indirectly, therefore, the worst quality of surface water helps to increase the cost of production and to affect the economy of the country. acknowledgements foremost, we would like to express our sincere gratitude towards habibur rahman, chemist, new haripur,412 mw ccpp, egcb ltd for the continuous support for this research work. his guidance helped us all the time for all sorts of analysis. this is to declare that a part of this paper has been presented in iccesd 2020 conference. references balan, i. n., shivakumar, m., & kumar, p. d. (2012).an assessment of groundwater quality using water quality index in chennai, tamil nadu, india. chronicles of young scientists, 3(2). boyacioglu, h. (2010). utilization of the water quality index method as a classification tool. environmental monitoring and assessment, 167(1), 115-124. brown, r. m., mcclelland, n. i., deininger, r. a., &tozer, r. g. (1970). a water quality index-do we dare. water and sewage works, 117(10). brown, r. m., mcclelland, n. i., deininger, r. a., & o’connor, m. f. (1972).a water quality index—crashing the psychological barrier.in indicators of environmental quality (pp. 173-182).springer, boston, ma. chowdhury, r. m., muntasir, s. y., &hossain, m. m. (2012).water quality index of water bodies along faridpur-barisal road in bangladesh. glob eng tech rev, 2(3), 1-8. goyette, j. o., bennett, e. m., howarth, r. w., &maranger, r. (2016). changes in anthropogenic nitrogen and phosphorus inputs to the st. lawrence sub‐basin over 110 years and impacts on riverine export. global biogeochemical cycles, 30(7), 1000-1014. journal of engineering science 12(3), 2021, 45-55 55 khan, a. a., tobin, a., paterson, r., khan, h., & warren, r. (2005).application of ccme procedures for deriving site-specific water quality guidelines for the ccme water quality index. water quality research journal, 40(4), 448-456. oni, o., &fasakin, o. (2016).the use of water quality index method to determine the potability of surface water and groundwater in the vicinity of a municipal solid waste dumpsite in nigeria. american journal of engineering research (ajer), 5(10), 96-101. pal, m., ayele, y., hadush, m., panigrahi, s., &jadhav, v. j. (2018). public health hazards due to unsafe drinking water. air water borne dis, 7(1000138), 2. sahu, p., &sikdar, p. k. (2008).hydrochemical framework of the aquifer in and around east kolkata wetlands, west bengal, india. environmental geology, 55(4), 823-835. weeks, j. r. (2020). population: an introduction to concepts and issues. cengage learning. world health organization,&światowaorganizacjazdrowia. (2004). world report on knowledge for better health: strengthening health systems. world health organization. © 2021 the jes.journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. microsoft word 09_jes_246_17-05-2020 journal of engineering science 11(1), 2020, 93-99 characterization of impressed current technique to model corrosion of reinforcement in concrete abu zakir morshed, sheikh shakib* and tanzim jahin department of civil engineering, khulna university of engineering & technology, khulna 9203 received: 15 september 2019 accepted: 17 may 2020 abstract corrosion of reinforcement is an important durability concern for the structures exposed to coastal regions. since corrosion of reinforcement involves long periods of time, impressed current technique is usually used to accelerate the corrosion of reinforcement in laboratories. characterization of impressed current technique was the main focus of this research,which involved determination of optimum chloride content and minimum immersion time of specimens for which the application of faraday’s law could be efficient. to obtain optimum chloride content, the electrolytes in the corrosion cell were prepared similar to that of concrete pore solutions. concrete prisms of 200 mm by 200 mm by 300 mm were used to determine the minimum immersion time for saturation. it was found that the optimum chloride content was 35 gm/l and the minimum immersion time for saturation was 140 hours. accounting the results, a modified expression based on faraday’s law was proposed to calculate weight loss due to corrosion. keywords: corrosion; impressed current technique; minimum immersion time; optimum chloride content. 1. introduction the focus of this research was to study the impressed current technique to model the accelerated corrosion test in laboratory. though the impressed current technique has been extensively used by several researchers (allan and cherry, 1992; andrade et al., 1993; andrade et al., 1990; austin et al., 2004; caré et al., 2008; maaddawy and soudki, 2003; he et al., 2016; kassir and ghosn, 2002; nossoni and harichandran, 2012; tran et al., 2011; val et al., 2009; shakib and morshed, 2016, 2018), literatures describing its appropriateness are very scarce. in particular, suitability of faraday’s law for computing the loss of reinforcement due to corrosion in different test conditions, e.g. immersing water with variable chloride concentrations, immersion time before starting the test to saturate the concrete pores, current density applied during test, techniques applied to ensure sufficient water and oxygen in concrete and etc., were not investigated in detail in the past. in this research, two test conditions were investigated a) different chloride concentrations in synthesized solution similar to that of concrete pore solution, b) time to saturate the concrete. generally, a dc power supply with constant voltage is used to accelerate the corrosion process. the corrosion current amount flowing through the circuit is very much dependent on the chloride content in the electrolytes. it increases with the chloride content till the optimum is attained. so that the time required to obtain chloride saturation in the concrete is important to attain a steady current flow. moreover, in chloride-induced corrosion, prediction of mass loss is also dependent on the chloride content present in the electrolytes (nossoni and harichandran, 2012; tran et al., 2011; shakib and morshed, 2016, 2018). however, so far, there are no specific guidelines regarding the amount of chloride content and immersion time before starting the test to attain saturated chloride content (in terms of time to attain a steady current flow) for reinforced concrete specimen in any codes of standard.thus, application of faraday’s law in different testing environments may end up with different test results. focusing on this issue, a favorable testing environment for applying the faraday’s law was investigated in this research. testing environment was evaluated in terms of efficiency of currenton the basis of gravimetric loss method. 1.1 electrochemical concepts and principle of faraday’s law reinforcing steel remains in stable passive state at very high ph as shown in figure 1. since the ph of concrete pore solutions is very high (austin et al., 2004; nossoni and harichandran, 2012; shakib and morshed, 2020, shakib and morshed, 2018), concretes give a good protection against corrosion (bazant, 1979; broomfield, 2006; maaddawy and soudki, 2003). on the other hand, a passive layer is formed due to hydration of cement on the steel surface, which also protects it from corrosion. but if any corrosive agent, like chloride ions, incorporate into the concrete, passive layer is destroyed and corrosion progresses (bazant, 1979). the passive zone of the steel becomes a pitting zone as shown in figure 2 and the zone expands to high ph with the increase in chloride concentrations. to model this chloride induced corrosion in the laboratory, impressed current technique is used, where the steel corrodes electrochemically. in accelerated corrosion test, a constant dc power supply is used to apply current through the steel rebar as shown in figure 3. in this methodology, presence of chloride ions, water and oxygen * corresponding author:sheikhshakib10@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 94 abu zakir morshed et al. characterization of impressed current technique to model….. are obligatory. considering this obligation, the concrete samples need to be immersed into the water containing chlorides. faraday’s law is employed to evaluate the overall weight loss of the anodic bar which is directly related to current and time. figure 1: pourbiax diagram of steel (nossoni and harichandran, 2012) figure 2: pourbiax diagram of steel in the presence of chloride ions (nossoni and harichandran, 2012) figure 3: schematic diagram of accelerated corrosion test setup the estimated corrosion-induced loss of steel using faraday’s law ிܯ ൌ ெூ௧௭ி ൌ ெ௜೎೚ೝೝ௧஺ ௭ி 1 where, mf = the weight loss of steel bars (gm), m = atomic weight of metal (for fe, m = 56 gm/mole), i = current (amperes), ݅௖௢௥௥ ൌ ூ஺ = current density, a = corroding surface area of the rebar, t = time (seconds), z = ionic charge (for fe, z = 2), f = faraday’s constant = 96,500 coulombs/mole. the ratio of the estimated loss of steel to the actual loss found by gravimetric loss method is defined as current efficiency, ɳ ൌ ெಲெಷ 2 where, ma = actual weight loss (gm). 2. experimental process 2.1 materials and specimens corrosion of reinforcement progresses through consecutive oxidation reduction reaction. the anodic and cathodic reactions are difficult to examine when the electrodes are embedded into the concrete. so, the experimentation was first performed in synthesized solutions. bertolini et al. (2013) studied on the chemical composition and properties of concrete pore solutions. from which three different chemical compositions were considered in this study for simulatinga concrete environment. pallets of naoh, ca(oh)2, and koh were used to prepare the synthesized solutions as shown in table 1. the ph of prepared solutions was 13.7, 13.9, and 13.5 journal of engineering science 11(1), 2020, 93-99 95 respectively. the high ph solutions were similar to that of concrete pore solution (bertolini et al., 2013; nossoni and harichandran, 2012). two 10mm-ø deformed mild steel bars were used as both anode and cathode for the test in synthesized solutions. the length of each bar was 76 mm. the bars were cleaned by steel wire brush. the effect of chloride ions was investigated through this experimentation. the optimum chloride concentration obtained from this investigation was used to simulate the corrosion in concrete prisms. table 1: solution properties (bertolini et al., 2013) [oh-] (mmol/l) [na+] (mmol/l) [k+] (mmol/l) [ca2+] (mmol/l) calculated ph solution-1 470 130 380 1.0 13.7 solution-2 834 271 629 1.0 13.9 solution-3 288 85 228 0 13.5 the concrete used in the prisms was made with ordinary portland cement with w/c of 0.45. the mix was designed according to the aci 211.1 for a c30 grade concrete considering the requirements for corrosion protection of reinforcements in concrete exposed to corrosive environment. 19 mm downgrade stone chips was used as coarse aggregate and river sand with fineness modulus of 2.8 was used as fine aggregate in concrete. the mix proportion is shown in table 2.two different thicknesses of clear cover were used in this study19 mm and 40 mm. a 12 mm-ø plain mild steel bar was embedded in the concrete prisms and a 4 mm-ø mild steel wire used to act as an anode and cathode, respectively. a detail of specimens is shown in figure 4. figure 4: details of prismatic specimens a dc (converted from ac current) power supply was used to supply current through the circuit. the input ac voltage was 200-220v 50/60hz switchable. output dc voltage was variable (3-30 v). a resistor of 4.7ω was used in the circuit and a multimeter was used to measure the voltage drop across the resistor, from which the current flowing through the circuit was calculated by ohm's law (i=v/r), where i is the current (a), v is the voltage drop (v) across a fixed resistor, r (ω). the schematic diagram of test setup is shown in figure 5. table 2: mix proportions for the concrete materials (kg/m3) w/c ratio water cement fine aggregate coarse aggregate fresh density 190 422 677 1056 2345 0.45 2.2 test in synthesized solution in an impressed current technique, there are two electrodesan anode and a cathode. the anode is connected to the positive pole of the power supply and the cathode is connected to the negative pole. the electrodes are placed in an electrolyte through which the ions transfer from one electrode to another. for the tests performed in synthesized solutions, three different solutions were prepared and used, as shown in table 1. the amount of chloride was varied from 0 gm/l to 58 gm/l. the anode bar was cleaned by steel wire brush before and after the test. a new solution and anode bar were used for each trial. a constant 12v current was applied across the circuit. the test setup is shown in figure 5(a). the distance between the anode and cathode was kept constant (20 mm) throughout the test. each test was continued for 18 hours. the actual weight loss was measured from the difference between the weights recorded before and after the test. the actual weight loss was then compared with the theoretical weight loss following faraday’s law and the current efficiency was calculated as well. (a) cross-section (b) longitudinal section 96 abu zakir morshed et al. characterization of impressed current technique to model….. figure 5: schematic diagram of test setup (a) synthesized solution and (b) concrete prisms 2.3 test in concrete prisms a 200 mm by 200 mm by 300 mm concrete prism was used to perform the accelerated corrosion test in concrete environment. the corrosion setup is shown in figure 5(b). in impressed current technique, an artificial supply of chloride ions is needed to increase the conductivity of concrete. on the other hand, a certain amount of chlorides (optimum chloride content) are required at the rebar surfacefor efficient use of the supplied current.the optimum chloride content adjacent to the steel concrete interface can be obtained by saturating the concrete with chlorides.for this reason,the specimens areneeded to immerse in chloride solutionsfor a certain period of time before starting the tests. there are no certain guidelines regarding this. in this context, the optimum chloride content obtained from the tests in synthesized solutions was used to immersethe specimens. the specimens were immersed in salt solutions with optimum chloride content for different period of time (0 day, 1 day and 7 day) before conducting the test to determine the optimum time for which a steady current flow was attained, consequent to a maximum current efficiency. a constant voltage of 12v was applied across the circuit. the actual weight loss and theoretical weight loss was measured as stated before and the current efficiency was also calculated. 3. results and discussion 3.1 tests in solutions the current efficiency for different chloride concentrations is plotted in figure 6 for solution-1. the graph shows that the efficiency of current is highly dependent on the amount of chloride present in solution. the oxidation of anode was almost zero in chloride free solution, but it gradually increased with the increase of chloride ion concentration. from figure 6it is found that the current efficiency reached to 100% at a high chloride concentration of 30 gm/l. the chloride content required to reach 100% current efficiency was dependent on the ph of the solution and current density (nossoni and harichandran, 2012). this phenomenon can also be described by the pourbiax diagram for steel. according to pourbiax diagram, when no chloride ispresent there aretwo possible active zones for corrosion: one for ph lower than 9 and another for ph higher than 14 (figure 1). negligible corrosion of rebars occurred in the solution (ph = 13.7) without having chlorides (figure 6). but due to addition of chloride content, a pitting zone was found and it became more active with increased chloride content (figure 2). at a certain level of chloride (30 gm/l), total supplied current was fully efficacious in oxidizing the rebar. before which, the current was partially used in oxidation process and rest to break the water molecules emmitting different types of gases. they were not chemically identified in this study but the probable reaction might be as follows (nossoni and harichandran, 2012), anode: 2oh= 1/2o2 + 2h2o + 2e 3 cathode: 2h2o + 2e= h2 + 2oh when voltage-controlled power supply is used (as was the case in present study), the amount of current (i) increases with reactivity of the anode, which increases with the increase in chloride content. at a certain level of chloride content (optimum chloride content) a steady current flow was obtained. however, below that level, as the current flow icreased, the demand of chlorides also increased. that, in turn, increased the possibility of occurance of reactions other than the oxidation of steel. this might be the possible reason for the differences between the results of the present study and that of the previous works (figure 6). (a) (b) journal of engineering science 11(1), 2020, 93-99 97 figure 6: variation of current efficiency with respect to different chloride concentration in “solution 1” figure 7: current efficiency for different solutions in chloride free solution all the current was used to break the water molecule and produced oxygen gas at anode and hydrogen gas at cathode. in presence of chlorides lower than a certain amount, partial current was used to oxidize the steel and rest of it was utilized for breaking water molecule. so, for a fixed ph there was a certain chloride content for which a full oxidation of anode occurred. for a solution having ph 13.7, it was found that a 30 gm/l of nacl was needed for 100% efficiency of current. since a 100% current efficiency was found for a nacl concentration of greater than 30 gm/l, tests on the other solutions were conducted only for two different chloride concentration; 30 gm/l and 35 gm/l. the results are shown in figure 7. it is seen that in solution-1 and solution-3, over 100% current efficiency was attained for 35 gm/l of nacl solution irrespective of current density. whereas for solution-2 it was about 95% at 35 gm/l of nacl. the reason for the current efficiency below 100% for solution-2 might be the difference in solution properties (a relatively higher ph of 13.9). the optimum chloride content might also depend on the current density applied through the circuit and ph of the solution (nossoni and harichandran, 2012). it was reported that a higher chloride concentration was required to efficacious corrosion at higher ph. it was found that a nacl of 16.5 to 30 gm/l was required for 100% efficiency depending on the current density and ph of solutions (nossoni and harichandran, 2012). for a higher current density, a higher optimum chloride content was also reported. to investigate the effect of current density, two types of current density were used in this study: one was low (<50 a/sqm) and the other was high (>1000 a/sqm). it is seen from figure 7 that for a higher chloride content (35gm/l), over 100% efficiency was found in both cases; however, for chloride content of 30 gm/l, a lower current efficiency (97%) was found in case of higher current density. so it can be concluded that, 35 gm/l might be the optimum chloride concentration irrespective of current density. 3.2 tests in prisms the high alkaline environment, attained due to hydration of cement in concrete, forms a passive film on the surface of the embedded steel which normally prevents the steel from corrosion. however, under chloride penetration, the passive film is disrupted or destroyed, and the steel is exposed to the harmful attackers and corroded spontaneously. from the investigation in synthesized solutions similar to the concrete pore solutions, it was found that a 35 gm/l of nacl concentration caused a 100% current efficiency. this finding was applied to prepare the electrolytes for concrete specimens [figure 5(b)]. after 28 days water curing, at an average compressive strength of 30.2 mpa, the prism specimens were immersed in the electrolyte solution, so prepared, for three different periods of time0 day, 1 day and 7 day in a steady state condition before starting the test. the results of the tests in concrete prism are presented in figure 8. availability of water and chloride adjacent to the reinforcement is dependent on the concrete property (neville, 2011). so a certain time is needed for the specimens to be immersed in chloride solutions for sufficient availability of water and chloride. it is seen from figure 8 that the current supplied to the anode gradually increased up to a certain period of time and then became stable. the observed time period was 108 to 136 hr for cover = 19 mm [figure 8(a)] and 110 to 140 hr for cover = 40 mm [figure 8(b)]. that was due to the gradual increase in chlorides inside the concrete. the rate of increase of chloride inside the concrete depends mostly on the density and porosity of concrete. since a single type (grade) of concrete was used throughout the experiments, the effect of density and porosity was assumed to be similar for all cases and hence their effect was neglected in this study. after attaining the chloride saturation, the current maintained a steady flow. 98 abu zakir morshed et al. characterization of impressed current technique to model….. figure 8: effect of immersion time before starting tests for (a) 19 mm and (b) 40 mm clear cover the effect of immersion period on the current efficiency is summarized in figure 9. the efficiency of current was 72% when no immersion before starting test. it increased to 74% and 87% when the time of immersion was 1 days and 7 days, respectively for cover thickness is equal to 19 mm. in case of cover of 40 mm, the efficiencies were 69, 72 and 85% respectively.it was seen that the current efficiency was very much dependent on the chloride content adjacent to the reinforcent and the efficiency increased with the increase in chloride content (figure 6). the unsteady flow of current meant that chloride content present adjacent to the reinforement was lower than the optimum, which in turn affected the current efficiency for any period of immersion time. on the other hand, the dependency of clear cover was found to be insignificant as shown in figure 8 and figure 9. this might be due to the fact that the penetration of chloride solution was allowed not only through the clear cover but through the two sides and bottom of the prisms as well in the test setup [figure 5(b)]. in case of penetration through a 44 mm thick cover, a lower immersion time (one hour) was reported by shao (2016). the lower (average 86%) value of the current efficiency was attained might be because of the hidrance of chloride ions from reaching to the rebar surface by the layer of accumulated corrsion products around the rebar. whereas, in a test condition without chlorides inside the concrete, the efficiency wasreported as 34-45% (nossoni and harichandran, 2012; shao, 2016). so, the faraday’s law needs to be calibrated by multiplying with the current efficiency factor (ɳ) with at least 140 hr immersion of specimens in 35 gm/l of nacl solution as follows ிܯ ൌ ɳ ெூ௧௭ி 4 where ɳ = an average of 86% with at least 140 hrs immersion in 35 mg/l salt solution. this proposed value can be reexamined for other test conditions. figure 9: current efficiency in concrete prisms for different immersion time 4. conclusions in impressed current technique, amount of chloride ions is an important factor affecting the corrosion of steel. in solutions similar to that of concrete pore solution, an experimental investigation was carried out to find the optimum chloride content. it was found that at a high chloride concentration of 35 gm/l of mixing water, the actual weight loss met the theoretical one. in concrete, the chloride ions are to be transported through the pore space need to be saturated with the chlorides adjacent to the steel-concrete interface for optimum current efficiency. an investigation was carried out regarding this point. the prismatic specimens were immersed in 35 gm/l chloride solution for different period of time and it was found that an immersion period of at least 140 0 40 80 120 160 200 0 100 200 300 400 500 ele ctr icit y p ass ed (m a) time (hr) 0 day immersion 1 day immersion 7 day immersion 0 40 80 120 160 200 0 100 200 300 400 500 ele ctr icit y p ass ed (m a) time (hr) 0 day immersion 1 day immersion 7 day immersion 0 20 40 60 80 100 0 1 7 cu rre nt eff icie ncy (% ) immersion time before starting test (days) c = 19 mm c = 40 mm (a) (b) journal of engineering science 11(1), 2020, 93-99 99 hours before starting the tests was needed to saturate the specimen. a calibration factor (ɳ) equals to 0.86 was introduced to calibrate the equation of faraday’s law to calculate the actual weight loss of steel bar, embedded in concrete, due to corrosion. references allan, m. l., and cherry b. w., 1992. factors controlling the amount of corrosion for cracking in reinforced concrete, concrete, (may), 426–430. andrade, c., alonso c., and molina f. j., 1993. cover cracking as a function of bar corrosion: part iexperimental test, materials and structures, 26(8), 453–464. https://doi.org/10.1007/bf02472805 andrade, c., alonso m., and gonzalez j., 1990. an initial effort to use the corrosion rate measurements for estimating rebar durability. in corrosion rates of steel in concrete (pp. 29-29–9), astm international. https://doi.org/10.1520/stp25013s austin, s. a., lyons r., and ing m. j., 2004. electrochemical behavior of steel-reinforced concrete during accelerated corrosion testing, corrosion, 60(2), 203–212. bazant, z. p., 1979. physical model for steel corrosion in concrete sea structures – theory, journal of the structural division-asce, 105(6), 1137–1153. bertolini, l., elsener b., pedeferri p., redaelli e., and polder r. b., 2013. corrosion of steel in concrete, weinheim, germany: wiley-vch verlag gmbh and co. kgaa. https://doi.org/10.1002/9783527651696 broomfield, j. p., 2006. corrosion of steel in concrete: understanding, investigation and repair,second edition. https://doi.org/10.4324/9780203414606 caré, s., nguyen q. t., l’hostis v., and berthaud y., 2008. mechanical properties of the rust layer induced by impressed current method in reinforced mortar, cement and concrete research, 38(8–9), 1079–1091. https://doi.org/10.1016/j.cemconres.2008.03.016 el maaddawy, t. a., and soudki k. a., 2003. effectiveness of impressed current technique to simulate corrosion of steel reinforcement in concrete, journal of materials in civil engineering, 15(1), 41–47. https://doi.org/10.1061/(asce)0899-1561(2003)15:1(41) he, j., zhou y., guan x., zhang w., wang y., and zhang w., 2016. an integrated health monitoring method for structural fatigue life evaluation using limited sensor data, materials (basel, switzerland), 9(11). https://doi.org/10.3390/ma9110894 kassir, m. k., and ghosn m., 2002. chloride-induced corrosion of reinforced concrete bridge decks, cement and concrete research, 32(1), 139–143. https://doi.org/10.1016/s0008-8846(01)00644-5 neville, a. m., 2011. properties of concrete 5th edition, pearson education limited. nossoni, g., and harichandran r., 2012. current efficiency in accelerated corrosion testing of concrete, corrosion, 68(9), 801–809. shakib, s., and morshed a. z., 2018. modeling of cover concrete cracking due to uniform corrsion, proceedings of the 4th international conference on civil engineering for sustainable development, kuet, (february), 1–12. shakib, s., and morshed a. z., 2020. experimental and numerical simulation of corrosion induced expansive pressure on concrete cover, engineering solid mechanics, 8, https://doi.org/10.5267/j.esm.2019.9.001 shakib, s. and morshed a. z., 2016. study on prevention of rebar corrosion through cathodic protection by using sacrificial anode, proceedings of the 3rd international conference on civil engineering for sustainable development, kuet, (february), 978–984. shakib, s., and morshed a. z., 2018. initiation and propagation of crack due to corrosion of reinforcement an experimental investigation, iosr-jmce ,15(4), 12–16, https://doi.org/10.9790/1684-1504041216 shao, k., 2016. experimental investigation of non-uniform corrosion of the steel bars in concrete, m. phil thesis, hong kong university of science and technology tran, k. k., nakamura h., kawamura k., and kunieda m., 2011. analysis of crack propagation due to rebar corrosion using rbsm, cement and concrete composites, 33(9), 906–917. https://doi.org/10.1016/ j.cemconcomp.2011.06.001 val, d. v., chernin l., and stewart m. g., 2009. experimental and numerical investigation of corrosioninduced cover cracking in reinforced concrete structures, journal of structural engineering, 135(4), 376– 385. https://doi.org/10.1061/(asce)0733-9445(2009)135:4(376) 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 403 forbidden forbidden you don't have permission to access this resource. apache/2.4.54 (ubuntu) server at www.banglajol.info port 443 microsoft word 03_jes_282_10_28_2020 journal of engineering science 11(2), 2020, 27-35 doi: https://doi.org/10.3329/jes.v11i2.50895 future projection of mean temperature of post-monsoon season over bangladesh using statistical downscaling of global climate models md. bazlur rashid1* and syed shahadat hossain2 1bangladesh meteorological department, agargaon, dhaka-1207, bangladesh 2institute of statistical research and training (isrt), university of dhaka, dhaka-1000, bangladesh received: 2020 accepted: 28 october 2020 abstract in statistical downscaling technique, regional or local information are derived by determining a statistical model which relates large-scale climate variables or predictors generated by global climate models (gcms) to regional and local variables or predictands. in this paper, the results of gcms were statistically downscaled to produce future climate projections of mean temperature in the post-monsoon season (october and november), for the time periods 2021-2050 and 2071-2100 for bangladesh. the future climate projections are based on the three emission scenarios rcp2.6, rcp4.5 and rcp8.5 provided by the fifth coupled model intercomparison project (cmip5). this paper established a method to analyze gcms for use in statistical downscaling and utilized fifteen gcms. the gcms were assessed based upon their performance in simulated past climate in bangladesh and adjoining areas. downscaling was undertaken by linking large scale climate variables, taken from the era-interim (resolution 79 km) reanalysis temperature, a gridded data set incorporating observations and climate models, to local scale observations. overall, all fifteen gcms, via statistical downscaling, show that mean temperature of the post-monsoon season in bangladesh will increase under future climate scenarios. comparing the ensemble of future projections with the reference period (19812010), the mean post-monsoon temperature in bangladesh is projected for rcp8.5 showing warming by 0.31c in near future and 1.79c in far future. on the other hand, estimated warming is 0.39c in near future and 1.14c is far future for rcp4.5. low emission scenarios rcp2.6, near future temperature is nearly same the far future temperature. key words: climate change; post-monsoon season; empirical statistical downscaling; gcm. 1. introduction statistical downscaling is a technique that makes use of dependencies between large scale climate parameters and local surface variables. it involves calibrating a statistical model on historical observations which can then be used to generate future climate data by using in gcms output for future climate scenarios, and is often carried out in order to conduct a climate change impact study on a regional scale (wilby et al., 2004). the first statistical downscaling method was familiarized by klein (1948), and a short description of the early history of statistical predictions was found in klein and bloom (1987). that time, downscaling was usually applied to numerical weather forecasting, and was then referred to as ‘specification’. in the initial 1980s (kim et al., 1984), statistical downscaling was mentioned to as ’statistical problem of climate inversion’. however, a same technique, called ’model output statistics’ (mos) has been used in numerical weather forecasting since the early 1970s (baker, 1982). one reason why downscaling is a comparatively young science is that it depends on the presence of global climate models, which themselves denote recent advances in the climate science community. the literature on statistical downscaling has concentrated on different countries, although there are some publications with a north american focus (lapp et al., 2002; benestad et al., 2015; 2016), as well as for australia/new zealand (kidson and thompson, 1998), africa (reason et al., 2006; penlap et al., 2004), asia (oshima et al., 2002; das and lohar, 2005) and bangladesh (alamgir et al., 2019; hasan et al., 2017; nury et al., 2014; rahaman, et al. 2015; shourav et al., 2016). rahaman et al. (2015) found that statistical downscaling model (sdsm) showed good agreement between observed and simulated maximum and minimum temperature so that it is able to predict future scenarios over bangladesh with confidence. for precipitation, on the other hand, their results were not in such a good agreement. statistical downscaling of modeled temperature and rainfall by gcm can easily be understood and adopted in order to minimize climate changes and its relevant impacts in bangladesh (nury et al., 2014). hasan et al., (2017) found that the severity of summer-day temperatures would be alarming, especially over hilly regions, where winters would be relatively warm. shourav et al. (2016) used sdsm to downscale future climate projections over the city of dhaka, bangladesh. their study indicated that climate change would cause continuous increases of rainfall, temperature and weather-related extreme events. karmakar (2019) found patterns of climate change and its impacts in * corresponding author: bazlur.rashid76@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal 28 md. bazlur rashid and syed shahadat hossain future projection of mean temperature ….. northwestern bangladesh and in the very recent, paul et al. (2020) studied the past 100 years temperature data of bangladesh and it found that in this period temperature of bangladesh was risen by 10c on average and an increment of close to another 10c would be in the average temperature of bangladesh from the prediction of the model in next century. alamgir et al. (2019) analyzed that the greatest increase in maximum and minimum temperature was found in the more northern regions and the lowest increase was found in the southeast coastal regions. in the seasonal pattern, post-monsoon and winter would likely be warmer in the future than other seasons (khan et al., 2020) and at same times during this post-monsoon, many devastative cyclones occurred in coastal regions of bangladesh (khatun et al., 2016). these studies, however, do not address the question of the changes in specific warming levels in post-monsoon season in bangladesh. the main objectives of the present study are to investigate the future projection of mean temperature of postmonsoon season in whole country as well as micro level in bangladesh using 15 gcms model and also find out temperature change based on three scenarios. 2. data and method 2.1 data description the analysis in this paper is based on mean temperature data from 15 gcm simulations from the fifth phase of the coupled model inter-comparison project (cmip5) ensemble (table 1). the gcm data were downloaded from the knmi climate explorer (https://climexp.knmi.nl/start.cgi) which offers data re-gridded to a 2.5degree resolution grid for the time period of 1900-2100. three future scenarios (representative concentration pathways, rcps) were considered: the high emission scenario rcp8.5 (riahi et al., 2007), the medium emission scenario rcp4.5 in which the radiative forcing stabilizes shortly after 2100 (clarke et al., 2007), and the more optimistic peak-and-decline scenario rcp2.6 (van vuuren et al., 2007). local observations of the mean temperature from stations in bangladesh were observed local variables data and are collected from the bangladesh meteorological department (bmd). the era‐interim (resolution 79 km) reanalysis data set is a global atmospheric reanalysis produced by the european centre for medium‐range weather forecasts (ecmwf). the era‐interim project was produced in part to prepare for a new atmospheric reanalysis to replace era‐40. table 1: gcm simulations from the cmip5 ensemble used in this paper. the rip index refers to the realization, initialization method and physics version gcm rip gcm rip mpi-esm-lr r1i1p1 giss-e2-h r5i1p mpi-esm-lr r2i1p1 giss-e2-h-cc r1i1p1 mpi-esm-lr r3i1p1 hadgem2-es r1i1p1 mpi-esm-mr r1i1p1 hadgem2-es r2i1p1 mri-cgcm3 r1i1p1 hadgem2-es r3i1p1 giss-e2-h r1i1p1 hadgem2-es r4i1p1 giss-e2-h r1i1p2 inmcm4.0 r1i1p1 2.2 data analysis statistical downscaling first generates statistical relationship between larger gcm scale variables and observed small-scale (station level) variables. different approaches can be used such as analogue methods (rotation typing), regression analysis, or neural network techniques (wilby et al., 2002). future values of the large-scale variables found from gcm projections of future climate are then used to drive the statistical relationships in order to estimate the smaller-scale particulars of future climate. statistical models usually consist of equations as shown below. − = ℎ + − + = , + where is the predictand (the small-scale climate), is the geography, is the predictor (a quantification of the large-scale patterns of the climate), and n is the noise term. the empirical-statistical downscaling approach used in this study incorporated a form for quality control and bias adjustment through the practice of common empirical orthogonal functions (eofs) in the representation of the large-scale predictors (benestad et al., 2015; 2016). after bias correction, a stepwise multiple regression was used to estimate model parameters, hence downscaling large scale climate variables to local scale. such a journal of engineering science 11(2), 2020, 27-35 29 statistical downscaling approach requires a smaller amount computational effort than dynamic downscaling and can be applied to many scenarios and longtime intervals, rather than the short-term slices of the dynamical downscaling method. in this study principal component analysis (pca) was applied to the observational data before downscaling. the pca decomposed the data into a set of spatial patterns, corresponding time series that describe the temporal variability associated with each pattern, and eigenvalues that represent the relative strength of each pattern. rather than downscaling the observational stations individually, the time series associated with the first spatial patterns (hereby referred to as first principal components) were downscaled. the projected temperature could then be reconstructed from the downscaled principle components combined with the corresponding spatial patterns and eigenvalues. 2.3 analysis software the study was carried out within the r-environment and used empirical statistical downscaling (esd) package (benestad et al., 2015) to analyze the data for attaining the objectives. the development of the esd software fits in with the trend of the r-language increasing role in the climate change debate and as an open science platform. additionally, both r and the esd r-package are appreciated tools for linking high education and research. the wide range of functionalities of the esd tool, including methods for reading and manipulating data, generating various info-graphics, and performing statistical analysis (e.g., calculating eof), pca, canonical correlation analysis (cca) and empirical-statistical downscaling creates it appropriate for processing results from gcms. 3. results and discussions an analysis of the common eofs of the gcm simulations and era‐interim reanalysis temperature was performed to assess the goodness of fit of the gcm output with respect to observation-based data. the residuals from the downscaled information were scrutinized against acceptability. screening of predictors and predictor domains was conducted using gcm simulations, era-interim reanalysis data, observed station data and the best correlated predictors and predictand were selected. results of the statistical downscaling are shown in figures 1, 2 and 3 for rcp2.6, rcp4.5 and rcp8.5 correspondingly. figure 1: downscaled mean temperature of the post monsoon season based on the era‐interim reanalysis and using pca for downscaling a group of stations simultaneously for rcp2.6. figure (a) shows the spatial pattern associated with the leading principle component (pc1) of the predictand, (b) captures the leading spatial pattern of the predictor, (c) shows a cross-validation comparing the original pc1 of the predictand and the corresponding estimated values obtained by empirical-statistical downscaling and figure (d) indicates time series of the estimated and original pc1 of the predictand. the cross validated correlation between the pc1 of the mean temperature and the corresponding pc1 estimated based on empirical-statistical downscaling are 0.82, 0.78 and 0.8 for rcp2.6, rcp4.5 and rcp8.5 respectively. 30 md. bazlur rashid and syed shahadat hossain future projection of mean temperature ….. figure 2: downscaled mean temperature of the post monsoon season based on the era‐interim reanalysis and using pca for downscaling a group of stations simultaneously for rcp4.5. figure (a) shows the spatial pattern associated with the pc1 of the predictand, (b) captures the leading spatial pattern of the predictor, (c) shows a cross-validation comparing the original pc1 of the predictand and the corresponding estimated values obtained by empirical-statistical downscaling and figure (d) indicates time series of the estimated and original pc1 of the predictand. figure 3: downscaled mean temperature of the post monsoon season based on the era‐interim reanalysis and using pca for downscaling a group of stations simultaneously for rcp8.5. figure (a) shows the spatial pattern associated with the pc1 of the predictand, (b) captures the leading spatial pattern of the predictor, (c) shows a cross-validation comparing the original pc1 of the predictand and the corresponding estimated values obtained by empirical-statistical downscaling and figure (d) indicates time series of the estimated and original pc1 of the predictand. journal of engineering science 11(2), 2020, 27-35 31 rcp2.6 rcp4.5 rcp8.5 32 md. bazlur rashid and syed shahadat hossain future projection of mean temperature ….. figure 4: mean temperature of post monsoon season over regional points change for different rcp scenarios run by cmip5 experiments, respectively, relative to the period 1981-2010. the shaded area shows one standard deviation from the mean based on all scenarios for each experiment figures 4 showed downscaled climate projections for different rcps (rcp2.6, rcp4.5 and rcp8.5). the results are summarized in table 2, which includes the average change in the mean temperature of the postmonsoon season relative to 1981-2010 for two periods: the near future (2021-2050) and the far future (2071journal of engineering science 11(2), 2020, 27-35 33 2100). the downscaled projections indicated an increase in the post-monsoon temperature at dhaka by 0.49c for near future and 0.48c for far future with rcp2.6, but for rcp4.5 the near future warming was estimated to be 0.35c and far future was same 0.99c. the scenario, rcp8.5, suggested increases of post-monsoon temperature by 0.28c and 1.85c for the near and far future, respectively. the mean projected change in post-monsoon temperature for bangladesh assuming rcp4.5 was 0.39c for the near future and 1.14c for the far future. for the high emission scenario rcp8.5, the near future estimated warming was near future 0.31c but the far future warming was considerably higher 1.79c. the highest projected warming for rcp2.6 was 0.94c at chandpur for the near future and 0.91c at same place for the fur future and for rcp4.5 captured 0.85c at rajshahi for the near future and 2.05c at chandpur for the far future. for the most severe emission scenario, rcp8.5, the projected warming was 0.53c at chandpur for the near future and 2.82c at same place for the far future. the emission scenario rcp2.6 assumes very low future emissions. in a broad sense, co2 emissions will remain constant until early this century and will become negative by the end of the century. this scenario assumes a sharp decline in the use of cropland for biofuel production, and reduction of methane emissions by 40%. based on these conditions, models nicely captured the mean temperature of the post-monsoon season and indicate lower temperature for the far future over bangladesh with an exception only two places such as tangail and mymensingh regions where it indicates higher temperature increment during far future period. table 2: mean anomaly mean temperature in post-monsoon season based on 1981-2010 division station location emission scenario rcp2.6 rcp4.5 rcp8.5 near future far future near future far future near future far future dhaka dhaka 0.49 0.48 0.35 0.99 0.28 1.85 tangail 0.48 0.49 0.20 0.83 0.30 1.51 faridpur 0.60 0.59 0.39 1.13 0.33 2.24 mymensingh mymensingh 0.15 0.24 -0.73 -0.57 0.35 0.10 chattogram chattogram 0.59 0.56 0.46 1.33 0.35 1.86 cox bazar 0.67 0.64 0.55 1.53 0.37 2.11 chandpur 0.94 0.91 0.69 2.05 0.53 2.82 cumilla 0.58 0.57 0.40 1.17 0.32 2.03 feni 0.58 0.57 0.41 1.22 0.34 1.94 kutubdia 0.66 0.62 0.62 1.57 0.33 2.11 m.court 0.32 0.32 0.14 0.60 0.21 0.86 rangamati 0.41 0.40 0.22 0.83 0.27 1.20 sandwip 0.15 0.24 -0.73 -0.57 0.35 0.10 sitakunda 0.60 0.58 0.44 1.28 0.35 2.02 teknaf 0.56 0.50 0.62 1.74 0.39 1.25 khulna khulna 0.57 0.55 0.43 1.21 0.30 1.89 jashore 0.54 0.53 0.39 1.09 0.29 1.88 satkhira 0.40 0.39 0.33 0.85 0.21 1.51 barishal barishal 0.55 0.53 0.42 1.17 0.30 1.87 patuakhali 0.45 0.44 0.31 0.94 0.25 1.43 bhola 0.46 0.45 0.33 0.97 0.27 1.58 khepupara 0.46 0.45 0.36 1.01 0.25 1.50 rajshahi rajshahi 0.70 0.65 0.85 1.67 0.17 2.52 bogura 0.64 0.63 0.47 1.30 0.33 2.20 rangpur rangpur 0.64 0.62 0.50 1.30 0.30 2.20 dinajpur 0.65 0.64 0.44 1.26 0.34 2.24 sylhet srimangal 0.65 0.61 0.65 1.55 0.29 2.18 country 0.55 0.53 0.39 1.14 0.31 1.79 rcp4.5 describes a scenario of low to moderate future emissions. as per this scenario, co2 emissions will be increased slightly until the mid-century and declined afterwards. the use of energy will decline sharply and there will be large-scale reforestation. in this situation new climate policies will be introduced and methane emissions will be stabilized. table 2 indicates that the post-monsoon mean temperature values in the near future for rcp4.5 will be lowered than that for rcp2.6 and higher than that for rcp8.5. on the other hand, warming condition far future will be higher for rcp4.5 than that for rcp2.6 but lower than that for rcp8.5. 34 md. bazlur rashid and syed shahadat hossain future projection of mean temperature ….. rcp8.5 explains a scenario of very high future emissions by the end of this century and co2 emissions will be expected three times higher than the present status. there will also be a large increase in methane emission. energy use will be further increased and that will be covered mostly by using fossil fuels. as a result, temperatures will be increased considerably in the far future. from table 2 it is seen that the magnitudes of the projected temperature for the far future for rcp8.5 will always be higher than the corresponding values based on emission scenarios for rcp2.6 and rcp4.5. 4. conclusions from this study the following results can be drawn: i. this statistical downscaling is a suitable technique for downscaling the mean temperature to local scale of post-monsoon season in bangladesh. though gcm data are provided with a coarse resolution, but esd technique can capture the local response to large scale climate change variability; ii. the cross validated correlation between the pc1 of mean temperature of bangladesh and the corresponding pc1 estimated based on era‐interim data are caught very high; iii. compared to the reference period of 1981-2010, the projected mean temperature in post-monsoon at country level for rcp2.6, the estimated mean temperatures are nearly same at near and far future overall bangladesh; iv. based on assumptions rcp2.6, models captured very well that the mean temperature of post-monsoon season was lower for the far future than for the near future over different locations in bangladesh, with the exception of tangail & mymensingh; v. results from this study indicate that the mean temperature of post-monsoon season differences of the near future and fur future for rcp4.5 are slight rise at fur future. so, if we would follow strict new climate polices among mid this century, fur future post-monsoon temperature of bangladesh will be increased almost 0.390c in the near future and 1.140c in the far future; vi. as a result of conditions rcp8.5, we shall not follow any climate related polices, mean temperature of post-monsoon season will increase considerably in the far future. from this study, it is seen that projected values of temperature for rcp8.5 will increase by 0.310c in the near future and 1.790c in the far future overall bangladesh. acknowledgements authors are grateful to mr. rasmus e. benestand, abdul kader mezghani & kajsa m. parding who are developer of esd packages and for their great supporting to run it. special thanks mr. hans olav hygen, norwegian meteorological institute due to teach me ‘r’ language which has enriched in this paper. authors are extremely thankful to bmd for providing necessary climate data. references alamgir, m., ahmed k., homsi, r., dewan a., wang j. j. and shahid s., 2019. downscaling and projection of spatiotemporal changes in temperature of bangladesh, earth syst. environ., 3, 381-398. doi: https://doi.org/10.1007/s41748-019-00121-0 baker, d. g., 1982. synoptic-scale and mesoscale contributions to objective operational maximum-minimum temperature forecast errors, monthly weather review, 110, 163–169. benestad, r., and mezghani a., 2015. on downscaling probabilities for heavy 24-hr precipitation events at seasonal-to-decadal scales, tellus a., 67, 25954. doi: http://dx.doi.org/10.3402/tellusa.v67. 25954. benestad, r. e., parding k.m., isaksen k., and mezghani a., 2016. climate change and projections for the barents region: what is expected to change and what will stay the same? environ. res. lett., 11 054017. clarke, l.e., edmonds j.a., jacoby h.d., pitcher h., reilly j. m., and richels r., 2007. scenarios of greenhouse gas emissions and atmospheric concentrations. sub-report 2.1a of synthesis and assessment product 2.1. climate change science program and the subcommittee on global change research, washington dc. das, l., and lohar d., 2005. construction of climate change scenarios for a tropical monsoon region, climate research, 30, 39–52. hasan, m.a., islam a.s., and akanda a.s., 2017. climate projections and extremes in dynamically dowscaled cmip5 model outputs over the bengal delta: a quartile based boas-correction approach with new girdded data, climate dynamics, dio: https://doi.org/10.1007/s00382-017-4006-1 karmakar, s., 2019: patterns of climate change and its impacts in northwestern bangladesh, journal of engineering science, 10(2), 33-48. journal of engineering science 11(2), 2020, 27-35 35 khan, m.j.u., islam a.k.m.s., bala s.k. et al, 2020. changes in climate extremes over bangladesh at 1.50c, 20c and 40c of global warming with high-resolution regional climate modeling, theor apl climatol, 140, 1451-1466, dio: https://doi.org/10.1007/s00704-020-03164-w. khatun, m.a., rashid m.b., and hygen h.o., 2016. climate of bangladesh, met report, no 08, issn23874201, climate. kim, j.-w., chang j.-t., baker n.l., wilks d.s., and gates w.l., 1984. the statistical problem of climate inversion: determination of the relationship between local and large-scale climate, monthly weather review, 112, 2069–2077. kidson, j.w., and thompson c.s., 1998. a comparison of statistical and model-based downscaling techniques for estimating local climate variations. journal of climate, 11, 735–753. klein, w.h., 1948.winter precipitation as related to the 700-millibar circulation, bull. amer. meteor. soc., 9, 439 -4.53. klein, w.h., and bloom h.j., 1987. specification of monthly precipitation over the united states from the surrounding 700 mb height field, mon.wea. rev., 115, 2118–2132. lapp, s., byrne j., kienzle s., and townshend i., 2002. linking global circulation model synoptic and precipitation for western north america, international journal of climatology, 22, 1807–1817. nury a. h., and alam m.j.b., 2014. performance study of global circulation model hadcm3 using sdsm for temperature and rainfall in north-eastern bangladesh, sci. res. 6 (1), 87-96. oshima, n., kato h., and kadokura s., 2002. an application of statistical downscaling to estimate surface air temperature in japan, journal of geophysical research, 107(d10), acl–14. paul, s., and roy s., 2020. forecasting the average temperature rise in bangladesh: a time series analysis, journal of engineering science, 11(1), 83-91 penlap, e.k., matulla m., von storch h., and kamga f.m., 2004. down-scaling of gcm scenarios to assess precipitation changes in the little rainy season (march-june) in cameroon, climate research, 2.6, 8.5– 96. rahaman, a. z., khan f.a., aktar n., and noor f., 2015. climate change scenarios of bangladesh using statistical downscaling model, 5th international conference on water & flood management (icwfm -2015). reason, c. j. c., landman w., and tennantc w., 2006. seasonal to decadal prediction of southern african climate and its links with variability of the atlantic ocean. bull. amer. meteor. soc., 87, 182—190. riahi, k., grübler a., and nakicenovic n., 2007. scenarios of long-term socio-economic and environmental development under climate stabilization, technol forecast soc. chang., 74, 887–935. shourav, m.s.a., mohsenipour m., alamgir m., hadipour s., and ismail t., 2016. historical trends and future projection of climate at dhaka city of bangladesh, journal teknologi (sciences & engineering) 78, 6– 12, 69–75. van vuuren, d.p., den elzen m.g.j., lucas p.l., eickhout b., strengers b.j., van ruijven b., wonink s., and van houdt r., 2007. stabilizing greenhouse gas concentrations at low levels: an assessment of reduction strategies and costs, clim change, 81, 119–159. wilby, r.l., dawson c.w., barrow e.m., 2002. sdsm a decision support tool for the assessment of regional climate change impacts, environmental modeling software, 17, 147–159. wilby, r. l., charles s. p., zorita e., timbal b., whetton p., and mearns l. o., 2004. guidelines for use of climate scenarios developed from statistical downscaling methods, ipcc task group on data and scenario support for impacts and climate analysis (tgica), 27pp. © 2020 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attribution-noncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. microsoft word 8_jes_326-f journal of engineering science 12(2), 2021, 79-92 doi: https://doi.org/10.3329/jes.v12i2.54633 *corresponding author: ashikdas624@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal analysis of magnetohydrodynamic jeffery-hamel flow in a convergent-divergent channel using cu-water nanofluid ashik chandra das*, qazi novera tansue nasa and md. sarwar alam department of mathematics, jagannath university, dhaka, bangladesh received: 28 april 2021 accepted: 16 june 2021 abstract the present study focused on the entropy generation as well as the heat transfer rate and velocity profiles of a nanofluid of the jeffery-hamel flow, especially for convergent-divergent channels. first, the governing dimensional partial differential equations have been transformed into a system of non-dimensional ordinary differential equations. the power series method has been used to solve these non-dimensional governing equations and the hermite-pade approximation has been applied for analyzing them. the effect of various physical parameters such as channel angle, reynolds number, hartmann number, nanoparticle solid volume frictions and eckert number have been investigated for the velocity profiles, heat transfer and entropy generation. here, cu has been used as the solid nanoparticle and water has been used as the base fluid. it is interesting to remark that the entropy generation of the whole system increased at the two walls and a significant effect could be noticed on the heat transfer rate and velocity profiles. keywords: convergent-divergent channel, cu-nanoparticles, hermite-pade approximation, jeffery-hamel flow, mhd, power series method. nomenclature: f fluid  channel angle s solid particle  density nf nanofluid  dimensionless angle ha hartmann number  dynamic viscosity re reynolds number  kinematic viscosity ec eckert number  solid volume friction pr prandtl number  dimensionless temperature ns entropy generation  any angle 1. introduction the dilute suspension of two particles is known as nanofluid. one of them is a solid nanoparticle and the other is a base fluid. different type of solid nanoparticles may be used like cu, al2o3, tio2, zno and so on. these particles are mainly nanosized (less than 100nm). besides these solid nanoparticles behave as good conductors of heat. they also enable the base fluid to enhance its thermal properties. in this case, water, ethylene glycol, oil etc. may be used as the base fluid which is available in (das et al., 2007). but the word nanofluid was first introduced by (choi, 1995). after that, a good number of researchers have been successfully drawn their attention to nanofluid. this is the reason that the applications of nanofluid have become an interesting issue in recent times. nanofluid is used in many industrial and engineering processes like environmental, mechanical, biomechanical, aerospace, chemical and civil engineering. the flux of the rivers and canals is also a splendid application of nanofluid. another real-life example of nanofluid is the blood flow in the human body. the classical jeffery-hamel flow is considered as the flow of a two-dimensional viscous incompressible fluid in a channel. this channel may be parallel or convergent-divergent. in the convergent-divergent channel, the two non-parallel walls are driven by a source or sink and are separated by a fixed angle. it was briefly described in (hamadiche et al., 1994). the concept of the classical jeffery-hamel flow had come to the attention of the researchers since the first published study of (jeffery, 1915) and (hamel, 1916). since then numerous studies have been done on the classical jeffery-hamel flow using nanofluid. later, it can be found that there exists a similarity solution between the solutions of this flow and the navier-stokes equations for two dimensionless parameters. these two dimensionless parameters are mainly the reynolds number and the channel angular width. (fraenkel, 1962) then studied the laminar flow in symmetrical channels with slightly curved walls using 80 a.c. das et al. analysis of magnetohydrodynamic……… the jeffery-hamel solutions for flow between plane walls. it was found in this study that for viscous source or sink flow between straight walls of any plane, the solutions of the jeffery-hamel flow were mostly the same. so, in his analysis, he regarded the jeffery-hamel solutions as the leading term of a series solution for the velocity field of the flow. recently, (jeong et al., 2008) has investigated that in fundamental concepts of spreading and adhesion on a solid surface, nanofluid behaves electrically like classical liquid. the compressible navier-stokes equations are derived for the slip boundary conditions which is available in (aoki, 2017). he has studied it from the boltzmann equation based on the chapman-enskog solution. (sobey and drazin, 1986) have also used analytical, experimental and numerical methods to investigate the instabilities and bifurcations of twodimensional channel flows. they have found a rich and unexpected structure for reynolds numbers of less than a few hundred to the solutions of the navier stoke equation. moreover, (moradi et al., 2015) have studied the non-linear jeffery-hamel flow in a fluid for the effect of heat transfer. three different types of nanoparticles are used in their study. they have observed that the effect of solid nano-particle volume friction on the fluid flow parameters and the heat transfer are more stable. turbulent friction and heat transfer behaviours of dispreads fluids have been investigated by (pak and cho, 1998). a circular pipe has been used in their experiment with alumina (al2o3) and titanium dioxide (tio2) have been used as the solid particles having diameters of 13nm and 27nm, respectively. the results of their study demonstrate that the convective heat transfer coefficient of the dispersed fluid is smaller than the pure water under the condition of constant velocity. presently, many researchers have given their attention to develop the analysis of nanofluid. different types of study have been being continued. in recent times, the combined free and forced convection mhd flow in a rotating channel with a perfectly conducting wall have been studied by (seth and singh, 2008). they have observed that if the fluid is electrically conducting then the current induces joule heating. it is also found that this induced joule heating depends on the strength of the magnetic field. the entropy generation minimization for forced-free convective heat transfer due to viscosity effects and temperature gradient in a fluid was studied by (bejan, 1996). recently, (alam et al., 2017) have studied the entropy generation of steady mhd incompressible flow with viscous dissipation and joule heating using cu-nanoparticles through the convergentdivergent channel. in their study, the remarkable result was that the entropy generation of the system increased at the two walls. they also found that along the centerline of the channel the heat transfer irreversibility and the fluid friction irreversibility were dominant there. the present study is focused on the effect of different thermophysical properties on entropy generation as well as on the velocity and temperature profiles considering viscous dissipation and joule heating in the convergentdivergent channel. the velocity profiles, temperature distributions and entropy generation, as well as the rate of heat transfer, have been discussed graphically for the variation of channel angle  , reynolds number re , hartman number ha , eckert number ec and nanoparticle volume friction  . the hermite-pade approximation (hpa) has been applied to solve the power series. the bifurcation diagrams of the skin friction against divergent channel angle and reynolds number have also been presented here. 2. derivation of mathematical equation in this investigation, a steady two-dimensional incompressible laminar flow of cu-water nanofluid has been considered which places a source or sink in two non-parallel walls. the angle between the two walls is 2 which is shown in figure 1. in this case, a cylindrical coordinate system (r, , z) has been used. here it is assumed that the velocity is purely radial. this radial velocity depends on r and  . this shows that the flow parameter receives no change in the z direction. markedly, it is a presumption that an external magnetic field has been acted vertically downward to the top wall. let the domain of the experimented flow be      . so the semi angle of the channel will be  . now, considering joule heating and viscous dissipation the continuity equation, navier-stokes equation and the energy equation are,   0nf ru r r     (1) 22 2 0 2 2 2 2 2 1 1 1 nf nf nf nf bu p u u u u u u r r r rr r r r                         (2) 2 21 0 nf nf p u r r           (3) journal of engineering science 12(2), 2021, 79-92 81       22 22 2 0 2 2 2 2 2 2 1 1 1 4 nf nf nf p p pnf nf nf bt t t t u u u u r r r rr r rc c c r                                        (4) here it is assumed that the flow is symmetrically radial. so the velocity field has been turned into the form [ , 0, 0]v u where u is the function of both r and  . then the volumetric flow rate through the channel is q ur d       (5) boundary conditions for the problem are as follows, 0,u t t  at    (6) in the above equations, the velocity components and temperature of the base fluid are u and t respectively. also 0 , p, , , ,nf nf nf nfb     represent the electromagnetic induction, the fluid pressure, the effective density, the effective dynamic viscosity, the electrical conductivity and the kinematic viscosity of the nanofluid respectively and the variations of these physical parameters are given as (aminossadati and ghasemi, 2009),  1nf f s       ,  2.51 f nf      , nf nf nf     (7) 1 3 1 / 2 1 nf s s s f f f f                                              now the corresponding heat capacity and the effective thermal conductivity of the nanofluids are,     2 2 2 s f f s nf f s f f s                    ,       1p p pnf f sc c c       (8) here  is the nanoparticle volume friction of the nanoparticles. s and f represent the thermal conductivities of solid particle and the base fluid respectively. also  p nfc is the heat capacity of the nanofluid. the thermophysical properties of the nanoparticle and the base fluid are given in table 1. figure 1: diagram of the flow problem. table 1: thermophysical properties of the cu-nanoparticle and the water (das et al., 2005). physical properties cu water  3/kg m 8933 997.1  /pc j kgk 385 4179  /w mk 401 0.613  m  59.6 x 106 0.05 if it requires 0q  , then 0  , the flow is diverging from a source at 0r  . let the stream function be  ,r   , then it becomes source or sink 0b 0u  at   0u  at    r 82 a.c. das et al. analysis of magnetohydrodynamic……… 0, ur r          now the new non-dimensional variables have been established,      , , tf q t           (9) the reduced ordinary differential equation of the governing equations (2) – (4) can be written in the following form,     2.5 2.5(iv) 2 22 re 1 ' '' 4 1 '' 0,f a f f d ha f         (10)      2 2.52 2 2 2 22.5 pr '' 4 ' '' 1 ' 0 1 b ec f f d ha f c             (11) the corresponding boundary conditions have been reduced to the form,      1 1, 1 1, ' 1 0f f f      ,  1 1   (12) re is the reynolds number and it is defined as re f q   . furthermore,   pr p f f c   is the prandtl number,   2 max p u ec c t  is the eckert number and 2 0f f f b ha     is the hartman number. and some more assumed constants are             2 2 1 , 1 , 2 p s f f ss s f p s f f sf c a b c c                               1 3 1 / 2 1s s s f f f d                                            3. series analysis in this investigation, the equations have been solved using the series solution method. these series have been analysed by considering different non-linear terms and then the equations are also expressed through a parameter. for solving these, the power series expansion in terms of the parameter  have been considered first. for initially performing this problem, equation (10) and equation (11) can be written in the non-linear form for stream function and temperature profile. these non-linear terms are as follows,         0 0 ,i ii i i i f f               as 1  (13) now by substituting equation (13) into equation (10) and equation (11) and using the corresponding boundary conditions given in equation (12), the dimensionless governing equations are then solved into a series solution. first of all, the coefficients of powers of  have been equated in both sides of the equation. we have then computed it for the series of the stream function  f  and temperature    . for solving this more accurately, the programming software maple has been used. by using this software, the first 12 coefficients for these series in terms of , , re, , , pr, , , ,ha ec a b c d  have been computed. the different non-dimensional parameters are also expressed here. the expressions of shear stress and skin friction are as follows, 1 w nf u r          and  2max w f f c u    (14) journal of engineering science 12(2), 2021, 79-92 83 using the non-dimensional transformation equation (9) in the above equation (14), the reduced form is obtained from equation (14).    2.5 1 '' 1 re 1 fc f    (15) 4. hermite pade approximant hermite-pade approximant is a method of solution of a non-linear system of equation. this method is first introduced by (hermite, 1893) and (padé, 1892). in this study, conditions of criticality along with earlier mentioned irreversibility in the system by the hermite-pade method have been computed accurately. this approximation method is as follows, for 1  , s be the partial sum of the series. by applying rapid polynomial approximation, desired sum can be obtained.   1 1 0 n n n n n s a       (16) let d be any positive integer and (d 1) tuple of polynomials then hermite-padé form is      0 1deg deg ... deg , d n n np p p d n     if (17)      [i] 0 d n n i i p s o     as 1  (18) where      0 1, ,..., ds s s   different forms of a single series or independent series. after solving equaton (17) and (18) got the polynomials [i]np which are determined by their coefficient. the unknowns in the equation (18) is [i] 0 deg 1 1 d n i p d n      (19) here it is confirmed by the equation that the coefficient matrix related to this system is square. now the left side of the equation (18) has been expanded in power of  and then equated n equations of the system equal to zero. in this case, a system of linear homogeneous equations can be found. to perform the next step of the solution, some kind of normalization is needed to calculate the coefficients of the hermite-padé polynomials such as,  [ ] 0 1inp  for some integer 0 i d  (20)      0 1, ,..., ds s s   are the first n coefficients of the series, which are required for the calculation of the hermite-padé polynomials. for that, the linear equations have been solved by gaussian elimination or gaussjordan elimination. in this case, if the singularity is of an algebraic type, then the exponent  may be approximated by     [ 1] , [ ] , 2 d n c n n d n c n p d dp        here two unique moderately approximants hoda (high-order differential approximant) and hpda (highorder partial differential approximants) have been carefully introduced by (khan, 2002) and (rahman, 2004) respectively. with the above-mentioned methods, a particular algebraic approximant have been used additionally which was introduced by (drazin and tourigney, 1996). 5. analysis of irreversibility in this century, the world is going to face a large problem as people are using non-renewable energy without any limit. this limitless use of energy causes much wastage too. the result goes this renewable energy is being lost day by day from the world. if it is continuing, it will be impossible to live in the world. so, it is one of the major concern of the engineers and the researchers is to find the methods that can control the wastage of useful energy. 84 a.c. das et al. analysis of magnetohydrodynamic……… it is interesting that there is a significant influence of mhd on this system. a great example of this system is that in a convergent-divergent channel with the isothermal walls, the properties of flow in presence of a magnetic field and viscous dissipation are irreversible. at the solid boundaries, the temperature in equilibrium conditions arises because of the exchange of energy and momentum within the fluid. this energy produces continuous entropy generation. for the fully developed flow, the volumetric entropy generation rate in cylindrical coordinates is given as, 2 2 2 0 2 2 nf nf nf g bdt du e u d t d tt                      (21) here, entropy generation is determined by the summation of three terms. the first term of the right side of equation (21) is the irreversibility due to heat transfer. the second and third terms of the right side of the above equation (21) are the irreversibility due to viscous dissipation and the local entropy generation due to the effect of the magnetic field, respectively. the non-dimensional form of the entropy generation number can be reduced as,   222 2 2 2 1 2 32.5 2 pr pr 1 nfg s f f e b ecd d f df n bd ec ha n n n d dd                             (22) where,   22 2 2 2 1 2 32.5 2 pr , , pr 1 nf f b ecd d f df n n n bd ec ha d dd                        in this paper, the focus of the study is observing the entropy generation for the optimum values of the parameters and discussing the obtained outcomes. 6. results and discussion in the present study, the influence of different parameters of interest in the convergent-divergent channel have been observed for the problem. entropy generation, temperature distribution and velocity profiles are performed here for the parameters. the parameters which are investigated here are the nanoparticle volume friction  , magnetic hartman number ha , reynolds number re, channel angle  and eckert number ec . also, the solid volume friction for cu-nanoparticles is studied in the range of 0 0.2  and the prandtl number pr for the base fluid is kept at 6.2 here. the graphical representations of the outcomes are presented and discussed for this study. 6.1 stability analysis it is mentioned earlier that in this study the centerline axial velocity and the radial velocity will be investigated as two series in powers of , , re, ,ha a  and d . here, table 2 exhibits the convergence of c for 3d  using 12n  up to 12 decimal places. this study has been performed for a different amount of the presence of nanoparticle in the nanofluid. the variation in the critical values c with critical exponent  for four different values of solid volume friction 0.00, 0.07, 0.14, 0.20  . if the present outcome is observed, there is an indication that there is a great influence of the presence of nanofluids in the instability of the flow process. it is seen in the flow that these nanofluids result from early development in the system. table 2: numerical values of critical angles c and their exponent  at 1ha  and re 20 .  c  0.00 0.271769826156 0.450319299852 0.07 0.210881499718 0.507663726998 0.14 0.187752358428 0.428501987679 0.20 0.183265993141 0.421205605495 here, the algebraic approximation method has been applied to our series. by applying this method, the bifurcation graphs of skin friction coefficient against  are obtained. figure 2 denotes the bifurcation graph of skin friction against  for the cu-nanoparticles. it is interestingly noticed from figure 2 that there is a turning point or a saddle-node bifurcation at c  . it is also noticeable from the bifurcation diagram that there is a growing effect for the increasing amount of cu-nanoparticles in the fluid. journal of engineering science 12(2), 2021, 79-92 85 here two solutions branches for the skin friction for different values of nanoparticle volume friction have been observed. one is for c  and the other marginal solution is for c  . but there is no solution for c  . it is clear from the above figure that the bifurcation points have become change for various values of  . the bifurcation points are obtained at re = 20, ha = 1. here, it is also noted that the skin friction coefficient is increased by the solid volume friction. figure 2: bifurcation diagram of skin friction coefficient versus divergent channel angle  at re = 20, ha = 1 for cu-nanofluid. (a) (b) (c) figure 3: (a) velocity profile (b) temperature distribution (c) entropy generation for different values of  at 0.07, re 20, 10, pr 6.2, 0.1ha ec      86 a.c. das et al. analysis of magnetohydrodynamic……… 6.2 diverging channel a non parallel channel having a source by which the fluid enters into the channel is known as a diverging channel. the effect of different parameters for the entropy generation as well as velocity profiles and temperature distribution have been discussed for the divergent channel here. below the key figures, the unique results have come out from the study of the diverging channel. figure 3(a) represents the effect of channel angle on velocity profiles for cu-nanofluid in the divergent channel. here it can be noticed that the presence of cu-nanoparticles  0.07  expedites the velocity at the centerline of the channel. it is also observed that a major backflow occurs for the large channel angel in the two walls of the channel. the effect of increasing channel angle is also noticed on temperature distribution for the divergent channel. figure 3(b) represents the temperature distribution for the increasing channel angle. it is significantly noticed that for the large angle / 18 , the nanofluid generates the highest temperature values and / 180 exhibits the lowest heat transfer rate. there is also seen that at the centerline of the channel, the heat transfer rate reach the highest. it is mentioned in many affiliated textbooks and research papers that the heat transfer property in a flow region is enhanced by the characteristics of nanofluids. the influence of nanofluid in entropy generation is also noticed here. the impact of channel angle on the entropy generation is shown in figure 3(c). it is interesting to percept here that the entropy generation has a great influence on the channel opening. the increasing channel angle provides a more entropy rate for the cu-nanofluid. here it is seen that in the channel angle / 18 , the entropy gives the highest rate followed by a small channel angle. but the entropy is high at the two walls more than the centerline of the channel. it is a splendid effect that the lowest channel opening has not only the lowest velocity profiles but also the lowest channel angle exhibits the lowest temperature and entropy generation. the lowest channel angle is kept / 180 in this study. it is interesting to notice that the outcomes of the increasing channel angle has also an increasing effect on this study. (a) (b) (c) figure 4: (a) velocity profile (b) temperature distribution (c) entropy generation for different values of  at / 36, re 20, 10, pr 6.2, 0.1ha ec      journal of engineering science 12(2), 2021, 79-92 87 figure 4(a) represents the velocity profiles with the variations of nanoparticle volume friction  . it is seen from the figure that the presence of nanoparticles has a great influence on the velocity profiles. without any nanoparticles which means only base fluid water has the lowest velocity rate. perhaps the increasing amount of nanoparticles in the nanofluid produces high velocity in the channel. the effect of nanoparticles in the temperature distribution is presented in figure 4(b). a noticeable effect for nanoparticles volume friction is found in this temperature distribution. the temperature rises high at the centerline of the channel. it is also significant in figure 4(b) that at the two walls of the channel, the temperature remains almost the same. the entropy generation for the variations of nanoparticles volume is plotted in figure 4(c). it is interesting to notice here that the velocity profiles and temperature distribution 0.20  gives the highest entropy. the effects of reynolds number on velocity profile, temperature distribution and entropy generation are shown in figure 5(a)-5(c) respectively. with the fixed nanoparticle volume friction and channel angle, the velocity profile rises high with the increasing reynolds number. the same changes can be noticed in figure 5(b). the temperature distribution for different re shows different outcomes here. the highest re results at the highest temperature for the divergent channel. but the temperature of the fluid near the non-parallel walls persists identical. there is also an opposite influence of re on the entropy generation for the cu-based nanofluid. from figure 5(c) it is clear that re=10 gives the highest entropy for the nanofluids. a significant change can be noticed that at the centerline, the entropy rates for re=30 is the fastest followed by the others. (a) (b) (c) figure 5: (a) velocity profile (b) temperature distribution (c) entropy generation for different values of re at / 36, 0.07, 10, pr 6.2, 0.1ha ec       88 a.c. das et al. analysis of magnetohydrodynamic……… (a) (b) figure 6: (a) velocity profile (b) temperature distribution for different values of ha at / 36, 0.07, re 20, pr 6.2, 0.1ec       figure 6(a) and 6(b) represents the velocity profiles and temperature distribution respectively. in the effect of hartman number, ha has a conversed effect on the velocity profiles. for the divergent channel with a fixed channel angle and flow rate, the velocity profiles have reduced for the increasing hartman number. it is seen in the above figure that for ha = 0, the velocity reaches the highest level at the centerline and then for different values of ha greater than 0 show their result. the temperature distribution for the increasing ha is also plotted in figure 6(b). it is interestingly noticed that with the presence of an external magnetic field, the rate of heat transfer increases rapidly due to a faster rate of ha. (a) (b) figure 7: (a) temperature distribution (b) entropy generation for different values of ec at / 36, 0.07, re 20, pr 6.2, 10ha       the effect of the eckert number on temperature distribution and the entropy generation for the divergent channel is demonstrated in figure 7(a) and 7(b) respectively. the heat transfer rate of the nanofluid gains the maximum level at the centerline of the channel which is plotted in figure 7(a). it is clear that there also exists an increasing trend of increasing eckert number. the increasing value of ec results in the fastest rate of temperature distribution. entropy generation for various eckert number is also presented in figure 7(b). here, it can be noticed that at the centerline of the channel the entropy decreases significantly with the increasing value of the eckert number. highest ec generates the highest entropy with irreversibility for the nanofluid. journal of engineering science 12(2), 2021, 79-92 89 6.3 converging channel the effect of various physical and thermal parameter for the converging channel on velocity profiles and heat transfer rate are represented graphically and discussed here. (a) (b) figure 8: (a) velocity profile (b) temperature distribution for different values of  at 10, 0.07, re 20, pr 6.2, 0.1ha ec     the possession of channel angle for the converging channel on velocity profiles and temperature distribution are represented graphically in figure 8(a) and 8(b), respectively. from figure 8(a), it can be noticed that in the convergent channel the velocity of the fluid descends for the accelerating values of the channel closing angle. the heat transfer rate in this channel is also presented in figure 8(b). here, it is immersing to notice that an antipodean effect created for the temperature at the centerline of the channel. convergent channel angle / 180 exhibits the highest temperature for the nanofluid. on the contrary, the smallest heat transfer rate can be exposed to / 18 in the channel. (a) (b) figure 9: (a) velocity profile (b) temperature distribution for different values of  at / 36, 10, re 20, pr 6.2, 0.1ha ec       the effect of nanoparticle volume friction for the convergent channel is plotted in the above figures. figure 9(a) denotes the velocity profiles and figure 9(b) denotes the temperature distribution for the convergent channel. it can be noted here that for the expanding amount of volume friction the velocity of the fluid shows the reverse conclusion. only the base fluid outcomes the highest velocity and base fluid with cu-nanoparticles show the less. nanofluid of 20% nanoparticles terminates the lowest velocity for the convergent channel. the rate of heat transfer with the same thermo physical properties are presented in figure 9(b). here the base fluid also generates 90 a.c. das et al. analysis of magnetohydrodynamic……… the highest temperature in the centerline of the channel and then the nanofluid with cu-nanoparticles of different volume friction. (a) (b) figure 10: (a) velocity profile (b) temperature distribution for different values of re at / 36, 10, 0.07, pr 6.2, 0.1ha ec        figure 10(a)-10(b) represent the influence of reynolds number in the convergent channel on velocity profiles and temperature distribution. in figure 10(a) it is seen that the rising fluid flow results in less velocity. nanoparticles volume friction is kept fixed at 0.07 in this case. in this investigation reynolds number 30 causes the lowest velocity at the centerline. the heat transfer rate behaves the same characteristic for the same thermo physical properties. also, the highest fluid flow generates the lowest temperature in the convergent channel which can be cleared from figure 10(b). here the centerline temperature also rises high compared to the temperature of the channel walls. also, the same growing effect of increasing reynolds number can be explored for the heat transfer rate. (a) (b) figure 11: (a) velocity profile (b) temperature distribution for different values of ha at / 36, re 20, 0.07, pr 6.2, 0.1ec        figure 11(a)-(b) stands for the effect of the magnetic hartmann number on the velocity profiles and the temperature profiles for the convergent channel. from the above figure 11(a), it can be observed that the hartmann number has a very minimal effect on the velocity profiles for the convergent channel. in this case, the flow rate of the nanofluid reduces because of the rising magnetic field in the channel. the velocity reaches the highest level at the centerline of the channel. but the velocity has the lowest performance at the two walls of the channel for the convergent channel. the temperature distribution is seen in figure 11(b). for the heat transfer rate, the magnetic hartmann number has an opposite characteristic compared to the velocity profiles. the journal of engineering science 12(2), 2021, 79-92 91 highest temperature is recorded for the highest ha among all ha which has been used in this study. this shows that ha = 20 exhibits the highest temperature and ha = 0 exhibits the lowest. here, it is also significant that the temperature rises in the centerline of the channel. 7. conclusions in the current study, we have investigated the domination of cu-nanoparticles on the entropy generation in the magnetohydrodynamic flow of viscous incompressible fluid. here the water has been used as the base fluid. the effects of various physical parameters on the entropy generation with velocity field and temperature distribution are discussed in detail with the graphical representation. based on the obtained tabular and graphical results the main inferences of this study are as follows:  at the two walls, in the solution of the shear stress, two bifurcations at the critical angle are found namely the upper branch and lower branch. in this investigation, it has been initiated that with the increases in nanoparticles volume friction, the value of skin friction coefficient also amplifies.  enlargement in channel opening causes increasing in entropy generation, velocity and temperature distribution for both convergent and divergent channel but the fluid flow has an opposite behavior for the convergent channel.  significantly, there is a conflicting effect on the increasing nanoparticles volume friction between the convergent and divergent channel.  the reynolds number has an increasing influence with the increment of re for the divergent channel and an opposite effect is observed for the convergent channel.  the velocity profiles at the centerline against hartmann number has a rising effect but the temperature distribution has the reversed characteristics in both channels.  the behavior of the entropy generation for the eckert number has an increasing effect at the centerline of the channel and the heat transfer rate has the same for the divergent channel. references das s.k., choi s.u.s., yu w. and pradeep p.t., 2007. nanofluids: science and technology, wiley. choi s., 1995. enhancing thermal conductivity of fluids with nanoparticles, in proceedings of the 1995 asme international mechanical engineering cong and exposition, san francisco, usa, 66, 99-105. hamadiche m., scott j., and jeandel d., 1994. temporal stability of jeffery-hamel flow, journal of fluid mechanics, 268, 71-88. jeffery g., 1915. the two -dimensional steady motion of a viscous fluids, phil. mag, 6, 455-465. hamel g., 1916. spiralformige bewgungen zaher flussigkeiten, jahresbericht der deutschen math, 25, 34-60. fraenkel l.e., 1962. laminar flow in symmetrical channels with slightly curved walls. i: on the jeffery-hamel solutions for flow between plane walls, proceedings of the royal society, 267, 119-138. jeong y.h., chang w.j. and chang s.h., 2008. wettability of heated surfaces under pool boiling using surfactant solutions and nano-fluids, international journal of heat and mass, 51(11-12), 3025-3031. aoki k., baranger c., hattori m. and kosuge s., 2017. slip boundary conditions for the compressible navier stokes equations, journal of statistical physics, springer verlag, 4, 744-781. sobey i.j. and drazin p.g., 1986. bifurcation of two-dimensional channel flows, journal of fluid mechanics, 171, 263-287. moradi a., alsaedi a. and hayat t., 2015. investigation of heat transfer and viscous dissipation effects on the jeffery-hamel flow of nanofluids, thermal science, 19(2), 563-578. pak b.c. and cho y.i., 1998. hydrodynamic and heat transfer study of dispersed fluids with submicron metallic oxide particles, experimental heat transfer, 11, 151-170. seth g.s. and singh m.k., 2008. combined free and convection mhd flow in a rotating channel with perfectly conducting wall, indian journal of theoretical physics, 56, 203-222. bejan a., 1996. entropy generation minimization, crc press, new york. alam md.s., hakim m.a.h. and makinde o.d., 2017. magneto-nanofluid dynamics in convergent-divergent channel and its inherent irreversibility, defect and diffusion forum, 377, 95-110. aminossadati s.m. and ghasemi b., 2007. natural convection cooling of a localized heat source at the bottom of a nanofluid filled enclosure, european journal of mechanics-b/fluids, 9, 630-640. das s., chakraborty s., jana r.n. and makinde, o.d., 2005. entropy analysis of unsteady magneto-nanofluid flow past accelerating stretching sheet with convective boundary condition, applied mathematics and mechanics, 36(12), 1593-1610. 92 a.c. das et al. analysis of magnetohydrodynamic……… padé h., 1892. sur la représentation approchée d'une fonction pourdes fractions rationnelles, annales scientifiques. i’école normale supérieure, 9(3), 1-93. hermite c., 1893. sur la généralisation des fractions continues algébriques, annali di mathematica pura e applicata, 21(2), 289-308. khan m.h.k., 2002. high-order differential approximants, journal of computational and applied mathematics, 149, 457-468. rahman m.m., 2004. a new approach to partial differential approximants [m. phil thesis], bangladesh university of engineering & technology, dhaka. drazin p.g. and tourigny y., 1996. numerically study of bifurcation by analytic continuation of a function defined by a power series, siam journal of applied mathematics, 56, 1-18. © 2021 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. microsoft word 3_jes_316-f journal of engineering science 12(2), 2021, 23-28 doi: https://doi.org/10.3329/jes.v12i2.54628 *corresponding author: ch031dulal@gmail.com https://www2.kuet.ac.bd/jes/ issn 2075-4914 (print); issn 2706-6835 (online) jes an international journal health risk assessment of black carbon emission from fossil fuel md. dulal hossain khan1,2*, mahima sultana sarkar1, syeda sadika haque2 and md. amjad hossain3 1department of chemistry, comilla university, cumilla-3506, bangladesh 2department of chemistry, university of dhaka, faculty of science, dhaka 1000, bangladesh 3institute of leather engineering and technology, university of dhaka received: 15 march 2021 accepted: 04 may 2021 abstract fossil fuel combustion is one of the major sources of carbonaceous emission throughout the world. in this study, two light absorbing carbonaceous aerosol namely black carbon (bc) and brown carbon (brc) from fossil fuel combustion under controlled laboratory condition was reported. four different fossil fuels; octane, petrol, diesel and kerosene was taken as samples (four different fossil fuels; octane, petrol, diesel, and kerosene samples were collected from filling station of nilkhet, dhaka city. two wavelengths aethalometer (ot21) had been taken for systematic analysis of black carbon and brown carbon. bc and brc particulates were determined in terms of density, concentration, emission and emission factor. the concentrations of black carbon in mgm-3 for respective fuel samples were kerosene (3.83), diesel (4.59), petrol (7.94), octane (13.18) while concentrations of brown carbon were kerosene (7.77), diesel (7.98), petrol (13.61), octane (20.46). brc concentrations were found to be higher than those of bc for all the fuel samples. average concentrations of black carbon and brown carbon were 7.38 mgm-3 and 11.46 mgm-3 respectively. thereafter, health risk assessment for chronic exposure to black carbon was done (estimated/ evaluated/ calculated) according to the u.s. epa human health risk assessment protocol. experimental results were correlated with the data given by the exposure factors handbook of epa for assessing carcinogenic and non-carcinogenic risk associated with bc. total carcinogenic risk (cr) was found to be 3.27 for adults and 1.34 for children. while total noncarcinogenic risk i.e hazard quotient (hq) for adults and children were 243.32 and 594.32 respectively. both cr and hq values crossed the safe limit given by the us epa protocol indicating high probability of the occurrence of adverse health effects. keywords: black carbon, brown carbon, fine particulates, exposure, aethalometer, health risk. 1. introduction black carbon is a distinct type of carbonaceous material that is formed primarily in flames during combustion of carbon-based fuels. it strongly absorbs visible light with a mass absorption cross section of at least 5 m2g-1 at a wavelength of 550 nm (olson et al. 2015). bc particles fall under inhalable fine particulates and thus can be deeply inhaled and deposited in the lungs or other airways. causing many serious respiratory problems such as oxidative stress damage, respiratory irritation symptom (olson et al. 2015 and salam et al. 2013). strong solar radiation region such as tropical area are particularly especially at risk from black carbon emission (salam et al. 2013). primary source of bc is the incomplete combustion of biomass and fossil fuel in the absence of oxygen. black carbon stays in the atmosphere for just days to weeks (usually 7 to 10 days), but it can do a lot of lasting damage. the contribution to warming by one gram of bc is 100 to 2,000 times more than one gram of co2 on a 100-year time scale (vanloon et al. 2011). in the year 2011 scientists from nasas goddard institute for space studies found that as much as a quarter of arctic warming is caused by bc (bond et al. 2004). brown carbon on the other hand is a fraction of organic carbon (oc) that can share primary sources with bc but also can originate from soil humic matter or biogenic sources (e.g. plant debris and fungi). like bc, the primary sources of brown carbon are biomass burning, fossil fuel combustion, forest fire, soil eruption etc. although less, they also contribute to light absorption in atmospheric aerosols. this particulate matter appears light brown to yellowish (bond et al. 2001 and patterson et al. 1984). particles from smoldering combustion or from residential coal combustion (bond et al. 2001) can contain substantial amounts of brc. most aerosols in the smoke of combustion are an internal mixture of black and brown carbon. brc is present independently it has nearly 15% potential to warm the atmosphere by absorbing light. however, health risk 24 md dulal hossain khan et al. health risk assessment of black carbon……… associated with brc has not yet been found (adler et al. 2011). a number of motor vehicles are restlessly running in dhaka city. most of which are run through incomplete combustion of fossil fuels releasing dense black smoke into the atmosphere (brta, 2019). two major carbonaceous aerosols namely black carbon (bc) and brown carbon (brc) are notably present in this smoke which are receiving utmost concern due to disastrous environment and health issues recent years. therefore the aim of this study is to systematic determination of bc and brc in combustion smoke as well as studying their health risk associated with inhalation of bc in terms of carcinogenic risk (cr) and hazard quotient (hq). 2. methodology 2.1 sample collection and study site in this study, four different fossil fuels; octane, petrol, diesel and kerosene were collected from pother bondhu filling station, nilkhet, dhaka (figure 1). sampling was done in the inorganic and analytical chemistry laboratory, department of chemistry, university of dhaka and combustion of fossils were done under controlled laboratory conditions. one stage open face 9633, “nilu” filter holder was used for sampling of different fossil fuels. figure 1: images of sampling location messers pather bandhu filling station, nilkhet, dhaka (source: google map) 2.2 experimental procedure particulate matter was collected in quartz filters. (german, membrane filters, tissue quartz 2500 qat-up, 47 mm diameter). every filter paper was heated at 800 °c for 4 hours before sampling to eliminate all organic impurities. about 50.0 ml of each fossil fuel was taken for combustion. the measurement of deposited pm weight from difference between loaded and unloaded filters was carried out. from the difference between initial and final gas meter reading the amount of air is measured. the pm loaded filters which were collected for 5 seconds were used for characterization of bc and brc. 2.3 characterization technique the soot scan™ model ot21 transmissometer bench top analyser was used for measuring black carbon and brown carbon particulate matter (pm) from a variety of sample filters. aethalometer contains a 2 wavelength light source; 880 nm providing the quantitative measurement of black carbon and 370 nm for quantitative measurement of brown carbon (adler et al. 2011 and cheng et al. 2015). the density, concentration, emission and emission factor of black and brown carbon were determined by using aethalometer reading at 880 and 370 nm, respectively. all equations related to the determination of bc and brc parameters are suggested from arcadis, usa (lin et al. 2019 and usepa, 2019). 2.4 health risk evaluation of bc health risk evaluation was done on the basis of the u.s. epa human health risk assessment model. theoretical carcinogenic and non-carcinogenic risk was calculated using the, cdi (chronic daily intake in mg/kg/day) (us epa 2009). cdi = c × i r × ef × ed / (bw × at); c is the concentration of bc in mgm-3 and other parameters with their reference values are shown in table 1. journal of engineering science 12(2), 2021, 23-28 25 table 1: required parameters for health risk assessment study (deng et al. 2016 and epa, 2000 ) parameters recommended value children adult ir (inhalation rate) 7.6 m3day-1 20 m3day-1 ef (exposure frequency) 346 days year-1 ed (exposure duration) 6 years 26 years bw (body weight) 15 kg 70 kg at (averaging time) 365 days year-1× 70 years 2.5 characterization of carcinogenic risk to determine the carcinogenic risk, the lifetime carcinogenic risk (cr) was measured which is defined as the possibility of identifying cancer over a lifetime exposure (navid et al. 2019). cr = cdi × csf where, csf is cancer slope factor = 1.1 (mgkg-1day-1)-1 (epa 2000). 2.6 characterization of non-carcinogenic risk non-cancer risk was assessed by evaluating the hazard quotient (hq). where the inhalation toxicity reference doses for bc in our study is 5×10-3 mg m-3. hq = cdi / rfc (feng et al. 2019) hq value larger than 1 signifies that the exposed population is anticipated to have adverse non-cancer effects (epa, 2009). 3. results and discussions 3.1 density and concentration factors table 2 summarizes the experimental results of bc and brc parameters obtained from aethalometer reading. for all fossil fuels concentration of all brc was much greater than those of bc. both bc and brc concentrations levels were highest in octane and lowest in kerosene. the average concentration of bc was 7.38 mgm-3, while that of brc was 11.46 mgm-3. average emission of bc was 0.10 mgj-1 and that of brc was 0.16 mgj-1. the brc emission was 1.57 times higher for diesel, 1.17 times higher for octane, 1.58 times higher for petrol and 2.8 times higher for kerosene. table 2: density (in μgcm-3, concentration (in mg/m3), emission (in mgj-1) and emission factor (in mg.g-1) of bc and brc for all fossil fuels used in this study. 3.2 health risk evaluation table 3 summarizes the cancer risk (cr) as well as non-cancer risk (hq) caused by bc for children and adults. as recommended by us epa, the acceptable risk levels for carcinogens should be larger than 10-6 (epa, 2000). carcinogenic risks (cr) of adults and children due to bc exposure was higher than the permissible limit indicating that black carbon exposure may bring about considerable carcinogenic health hazard in this region. adults (3.27) were found to have approximately 2.4 times higher cr level as compared to children (1.34). the total non-carcinogenic risk in terms of hazard quotient (hq) caused to different populations was adult (243.32) and children (594.32) indicating that exposure to bc is strongly responsible for producing adverse non-cancer fossil fuel density concentration emission emission factor bc brc bc brc bc brc bc brc diesel 4.09 7.10 4.59 7.98 0.07 0.11 0.80 1.27 octane 9.39 14.58 13.18 20.46 0.17 0.20 1.93 3.01 petrol 5.66 9.69 7.94 13.61 0.12 0.19 1.28 2.18 kerosene 2.05 4.45 3.83 7.77 0.05 0.14 0. 62 1.25 26 md dulal hossain khan et al. health risk assessment of black carbon……… effects. children have 2.4 times higher hq values as compared to adults. which predicts that children are more prone to suffer non-cancer effects. table 3: carcinogenic risk and non-carcinogenic risk assessment for black carbon figure 2: comparison of chronic daily intake (cdi) between children and adults from figure 2, it is evident that adults were found to have higher cdi level as compared to children. it predicts that adults will have comparatively higher cancer risk (cr) values than children. total cdi for children was 1.22 mgkg-1day-1 and that for adults it was 3.08 mgkg-1day-1. adults were found to have 2.52 times higher chronic daily intake than children (table 3). figure 3: comparison of cancer risk (cr) between children and adults as recommended by u.s.epa protocol, acceptable risk levels for carcinogens should be larger than 10-6 (epa 2000). carcinogenic risk for adults and children due to bc exposure have been found to be much higher than the permissible limit 10-6 as shown in table 3. this indicates that black carbon exposure may bring about considerable carcinogenic health hazard. the order of cr caused by these fuels was cr(octane) > cr(petrol) > fossil fuel cdi (mg/kg/day) cr hq adults children adults children adults children diesel 4.62×10-1 1.89×10-1 5.08×10-1 2.08×10-1 37.79 92.35 octane 13.3×10-1 5.43×10-1 14.6×10-1 5.97×10-1 108.52 265.17 petrol 7.99×10-1 3.27×10-1 8.79×10-1 3.60×10-1 65.37 159.75 kerosene 4.85×10-1 1.58×10-1 4.24×10-1 1.73×10-1 31.53 77.05 total 3.08 1.22 3.27 1.34 243.32 594.32 journal of engineering science 12(2), 2021, 23-28 27 cr(diesel) > cr(kerosene). the total carcinogenic risk caused to exposed population was children(1.34) and adult (3.27). adults were found to have approximately 2.4 times higher cancer risk than children (figure 3). figure 4: comparison of non-cancer risk (hq) between children and adults the hq value greater than unity suggests that the exposed pollution likely to have harmful non-cancer effect, according to usepa protocol (epa, 2000). in our study, hq values obtained from experimental data were much larger than 1 (unity). total non-carcinogenic risk in terms of hazard quotient (hq) caused to exposed population was adult (243.32) and children (594.32) (table 2). these results indicate that exposure to bc is quite responsible for producing adverse non-cancer effects. moreover, (figure 2) displays that, children have 2.4 times higher hq values as compared to adults; which suggests that children are more prone to suffer non-cancer diseases than adults. 4. conclusions our findings indicate that the black carbon (bc) emission from various vehicles running on dhaka city street create environmental and health effect. the average concentrations of black carbon (bc) and brown carbon (brc) in the combustion smoke of fuel samples were found to be 7.38 and 11.46 mgm-3, respectively. brown carbon concentrations were higher than black carbon concentrations for all the four fuel samples. our results show that, carcinogenic risks of adults and children due to bc exposure are higher than the acceptable risk level recommended by us epa protocol. total cancer risk was found to be 3.27 for adults and 1.34 for children. also, hazard quotient (hq) which defines the non-carcinogenic risk was much larger than ‘one’ indicating that the exposed population is likely to have adverse non-cancer effects. total non-cancer risk was found to be 243.32 and 594.32 for adults and children, respectively which implies that children are more prone to suffer from non-cancer diseases than adults which cross it permissible level 1 (unity) set by us epa in 2009. in view of the magnitude of our reported effects, reduction of bc emission could lead to substantial health benefits. references adler, g., flores, j.m., riziq, a.a., borrmann, s, rudich, y., 2011. chemical, physical and optical evolution of biomass burning aerosols: a case study. atmos. chem. phys. 1491-1503. bond,t.c., 2001. spectral dependence of visible light absorption by carbonaceous particles emitted from coal combustion, geophys. res. lett. 28, 4075–4078. bond,t.c., streets, d.g., yarber,k.f., nelson, s.m., woo,j.h., 2004. a technology based global inventory of black and organic carbon emissions from combustion. j. geophys.res.atmos. 1-43. brta(bangladesh road transport authority), 2020. air pollution in bangladesh, brta associated with greater dhaka metropolitan area integrated transport study, working paper no. 23. 28 md dulal hossain khan et al. health risk assessment of black carbon……… cheng, y.h., yang, l.s., 2015. correcting aethalometer black carbon data for measurement artifacts by using inter-comparison methodology used on two different light attenuation increasing rates. atmos.meas.tech. 2851-2879. coz, e., guldris, j.p., calvo, a.l., alves, c., tarelho, l.a.c., ramos, g., 2014. a study on the structural properties of aerosols from biomass combustion for domestic heating. italian association of chemical engineering; 811-816. deng, w.j., zheng, h.l., tsui, a.k.y., 2016. measurement and health risk assessment of pm2.5, flame retardants, carbonyls and black carbon in indoor and outdoor air in kindergartens in hong kong. environ.int. 96, 65-74. feng, z.b.; cao, s.j.(2019). a newly developed electrostatic enhanced pleated air filters towards the improvement of energy and filtration efficiency.sustainable cities and society, 546-569. lin,w., dai,j., liu,r., zhai,y., yue, d., hu,q., 2019. integrated assessment of health risk and climate effects of black carbon in the pearl river delta region, china. environ.res.2019, doi: https://doi.org/10.1016/j.envres.2019.06.003. navid ghanavati, ahad nazarpour, michael j. watts. (2019). status, source, ecological and health risk assessment of toxic metals and polycyclic aromatic hydrocarbons (pahs) in street dust of abadan, iran. catena, 177(7): 246-259. olson, m.r., garcia, m.v., robinson, m. a., van rooy, p., dietenberger,m. a., bergin, m., schauer,j.j., 2015. investigation of black and brown carbon multiple-wavelength-dependent light absorption from biomass and fossil fuel combustion source emissions.geophys. res.atmos. 120, 6682– 6697,doi:10.1002/2014jd022970. patterson, e. m. and mcmahon, c. k., 1984. absorption characteristics of forest fire particulate matter. atmos. environ. 18, 2541–2551. salam, a., ullah, m.b., islam, m.s., salam, m.a., ullah, s.m., 2013. carbonaceous species in total suspended particulate matters at different urban and suburban locations in the greater dhaka region, bangladesh. air qual. atmos.health, 239-245. tracy,t.l., kirchstetter,t.w., malejan,c.j., ward,c.e., 2014. infiltration of black carbon particles from residential wood smoke into nearby homes. open j. air pollut. 3,111-120. u.s. environmental protection agency (epa): risk characterization handbook. usepa office of research and development, washington, dc, epa-100-b-00-002, 2000. u.s. environmental protection agency (epa): risk assessment guidance for superfund volume i: human health evaluation manual (part f, supplemental guidance for inhalation risk assessment). usepa office of research and development, washington, dc, epa-540-r-070-002 0swer 9285 7-82, 2009. u.s. epa. human health risk assessment. https://www.epa.gov/risk/human-health-risk-assessment. (accessed nov 15,2019). vanloon, g.w. and duffy, s.j., 2011. environmental chemistrya global perspective, 3rd ed.; oxford university press: new york. pp 131-133. © 2021 the authors. journal of engineering science published by faculty of civil engineering, khulna university of engineering & technology. this is an open access article under the terms of the creative commons attributionnoncommercial-noderivatives license, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.