1 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) ISSN (Print) 2313-4410, ISSN (Online) 2313-4402 © Global Society of Scientific Research and Researchers http://asrjetsjournal.org/ Modeling of Dissolved Oxygen (O2) Dynamics on the Left Bank of the Congo River; Port Ex Onatra River City Stéphane Mbuyambaa, Thierry Tangoub, Crispin Mulajic, Marc klesh Maya- Vanguad, Fifi Muderwae, Céline Sikulisimwae* a,c,d,e,fDepartment of Chemistry, University of Kinshasa, B.P. 190, Kinshasa XI, Democratic Republic of Congo bDepartment of the Environment, University of Kinshasa, B.P. 190, Kinshasa XI, Democratic Republic of Congo . Abstract The water quality of the Congo River is affected by industrial discharges. In this study, the evaluation of their impact was carried out using a mathematical approach. Used as water quality management tools, mathematical models are developed through an iterative process. The Streeter-Phelps model was used as a base model, and the methodological approach was that of Eckenfelder. This consisted of a campaign of parameter measurements that influence the dissolved oxygen content, a statistical treatment of data, a determination of the different global kinetic constants (K1 and K2) and the evaluation of the model by the Nash criterion. The results obtained are 2.26 j-1 and 13.26 j-1 respectively for the K2 reaeration and deoxygenation K1 constants; 498 mg / L for the initial oxygen biochemical demand (L0) and 3.82 mg / L for the initial oxygen deficiency (OD); 1.4767 j-1 for the degradation constant of the organic matter kd and 1.000621805 for the ratio α (L0/L(5)); generated a theoretical dissolved oxygen profile different from the experimental profile but which nevertheless allowed to simulate the fluctuations of the dissolved oxygen with the organic load. The statistical processing of the data has increased the reliability of the measurement campaign. The low value found for the Nash criterion (0,1756506) showed that this model should be adjusted to better evaluate the oxygen content. Key words: management ; evaluation ; industrial discharges ; reaeration ; deoxygenation. 1. Introduction According to the reports of the United Nations Environment Program (UNEP), the Democratic Republic of Congo (DRC) has a low water supply [1] while it has the second largest watershed important in the world, the Congo Basin [2]. ------------------------------------------------------------------------ * Corresponding author. http://asrjetsjournal.org/ American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 2 The use of surface water for domestic, agricultural, fish and other uses is very common. Thereby ; the quality of surface waters, especially those of the Congo River, is of crucial importance for the country [3,4]. Unfortunately in some respects, anthropogenic activities alter the quality of the latter restricting its uses ([31,33]). This deterioration is due to an overload of organic matter of agrifood and domestic origin, to toxic substances coming from agricultural and industrial activities [5,11]. The section of the Congo River from the Cité du fleuve to the SCPT port (formerly Onatra) in Kinshasa concentrates a few agro-food companies (breweries, lemonades, flour mills, etc.), emptying septic tanks and a few river ports that pour their water directly into it. used with or without treatment affecting the quality of the water [6, 7]. Doulaye reveals in his work, that nearly 78% of African cities did not have any formal service of evacuation and wastewater treatment [6]. Assessment of stream quality is done through a key parameter, the dissolved oxygen content, whose role for aquatic life and the health of the population is important [8, 12, 28]. This parameter makes it possible to estimate the conditions of pollution, of the degradation of the discharges and the self-purifying capacities of a watercourse [9, 10, 33]. In the Democratic Republic of Congo, this evaluation is very often done using a descriptive approach which consists of analyzing a certain number of physical, chemical and biological parameters which describe the good ecological state and comparing the values obtained with the target values called norms [10, 11]. Streeter and Phelps have developed a mathematical relationship that allows the assessment of surface water quality by the profile of dissolved oxygen; this is a function of the degradation level of the organic matter and the rate of diffusion of oxygen between / at the water-atmosphere interface [12, 13,14, 15, 32]. The Streeter- Phelps model was used as a basic model in the iterative process of developing a water quality model for the left bank of the Congo River (Kinshasa). This simple model is a forerunner of current river quality models [10]. Eckenfelder is one of the researchers who has used the Streeter-Phelps model extensively to determine the oxygen profile of several rivers and whose approach inspired us to evaluate the global impact of industrial discharges on this section [ 3, 5, 12, 13, 16]. The statistical evaluation of the data was made to assess the degree of reliability of the measurement campaign before the determination of the global kinetic constants, followed by the profile determination by the Excel software and finally the evaluation of the model obtained by the criterion of Nash. The results obtained could contribute to the optimal management of the Congo River, which is increasingly subject to increasing anthropic pressure. 2. Material and methods 2.1. Experimental site and sampling The measurement campaigns were carried out during the low-water period (August and September 2015 and 2016) at a frequency of two samples per month. The samples were taken immediately 30 cm from the surface and about 100 meters off the left bank of the American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 3 Congo River (Kinshasa) with coordinates at the starting point latitude 4 ° 19'37.6 '' and longitude 15 ° 20'58.0 '' and at the point of arrival latitude 4 ° 17'49.6''and longitude 15 ° 19'13.3 ''. Two samples were taken, upstream of the rejection point of a brewing enterprise rated EA (EA1 and EA2); and ten others downstream from the same point denoted EV (EVi, i = 1 to 10), in one liter flasks made of new and cleaned plastics (Figure 1). These were stored in the cooler before being sent to the laboratories of the DRC's Regie (Regideso) and the Regional Postgraduate School of Integrated Management and Planning of Forests and Tropical Territories (ERAIFT). The sampling points were chosen according to their accessibility and their position, that is to say they are located downstream of all new discharges and in such a way that the dissolved oxygen content is not zero according to Eckenfelder [16] . The GARMIN Global Positioning System (GPS), GPSMAP 78 series allowed the determination of the position and distance of each sampling point and the mapping of the site. Nylon wire with a weight and a graduated cane was used to measure the average depth of the water column on the section. Figure 1: Experimental site [Cité du fleuve-Port SCPT (ex Onatra) in Kinshasa] Legend: a. Map Democratic Republic of Congo b. Study area Cité du fleuve-Port SCPT (formerly Onatra) in Kinshasa c. Malebo Pool d. Map of the city of Kinshasa. PD Starting point, PR: Discharge point, PS: Ending point, EA: Upstream sample EV Downstream sample 2.2. Measurement campaign The measurement campaign focused on parameters describing the state of oxygenation, organic pollution and nutrient loading [8, 10, 17, 18]. The parameters measured in situ and in the laboratory of Regideso and Eraift, the conditions of conservation of the samples, the methods of analysis [30] are shown in Table 1. American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 4 2.3. Statistical treatment of data The statistical treatment of the results of the measurement campaign consisted of the determination and analysis of the correlation between the key variable of the model (dissolved oxygen) and the other variables measured. The existence of a relationship between dissolved oxygen and other variables, the form (or sense) of it (positive or negative, linear or nonlinear, monotonic or non-monotonic) and its intensity were determined by calculating the covariance ),( yxCov and Pearson coefficient ( )yx,ρ estimated by calculating the correlation coefficient ( )yxr , . A strong link between variables results in a high value of covariance. The Pearson test revealed whether there is a significant linear link [20]. Tableau 1: Parameters analyzed and methods used for modeling Parameters References AFNOR / ISO Conservation Materials and Methods pH NFT 90008 In situ Portable pH Meter Pen Series Pen Meter (OHAUS) Electrical conductivity NFT 90031 In situ Conductivity meter Waterproof HI 98311 Ammonium NF T90-015- 2 4°C Nitroprusside sodium and phenol, NaOH (HACH); Spectrophotometry ; λ= 630 nm Phosphates NF EN ISO 6878 4°C ammonium molybdate, antimony and potassium double tartrate, ascorbic acid, H2SO4; Spectrophotometry; λ = 700-800 nm COD NFT 90 101 4°C/ H2SO4 Acidic medium, K2Cr2O7, AgSO4, HgSO4 BOD5 4°C Respirometry / BODTrak ™ II / HACH T° Sonde Conductivity meter and pH meter O2 dissous In situ Portable Oximeter / HI 9146 / Hanna, Electrochemical Methods Nitrates ISO 7890-3 4°C sodium salicylate, disodium salt of ethylene diamine tetracetic acid, NaOH, H2SO4; sodium azide (Nitraver / HACH) / Spectrophotometry; λ = 415 nm Nitrites NFTEN 26777 4°C Orthophosphoric acid; Diazotization reagent (4- aminobenzenesulfonamide and N- (1-naphthyl) 1,2- diamino ethane), H3PO4 (NitriVer / HACH) / Spectrophotometry; λ = 510 and 543 nm American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 5 The regression made it possible to check whether the variable (dissolved oxygen content), depended on the other variables or better predict the future values of y (dissolved oxygen) knowing the corresponding x (pH, NH4 + content, PO4 3-, BOD5, COD) [19, 22, 26]. The contribution of the different points to the regression was obtained by calculating ( ) 2 , yxr [22]. The covariance, the Pearson coefficient and the correlation coefficient are calculated as follows: ( ) ))(( 1 1, 1 yyxx n yxCov i n i i −− − = ∑ = (1) ( ) ( ) yx yx yxCov σσ ρ . , , = (2) ( ) ( )∑ ∑ −− −− = i ii i ii yx yyxx yyxx r 22 , )( ))(( (3) Graphical analysis was used to visualize the shape of the connection, the proximity between the points, the overall shape of the points, and to detect the points that deviate from the others, to make the atypical observations and to check if there is any no suspicious groupings implying that there is in fact a third variable that influences the positioning of individuals. The line for which is small indicates a strong dispersion of points with respect to the line [21, 22]. The significance test t was performed to check whether the links between the variables are linearly significant or not, and deduce if the amount of data is sufficient [19, 20, 22]. This test is an assessment of the reliability of the results of the measurement campaign. 2.4. Modeling proper The construction of a model does not rest on a formal theoretical basis but rather on an iterative method consisting of carrying out successive tests and comparing field-collected data and models formulated in the symbolic domain [25]. The Streeter-Phelps model used considers that the fluctuations of dissolved oxygen in the aquatic environment are due solely to two phenomena: deoxygenation K1 and re aeration K2 [12, 13, 14, 15]. In modeling, the degradation of organic matter and the re-generation at the air-water interface are simulated by first-order kinetics [30, 32]. The methodological approach used to determine the overall kinetic constants of the section was that of Eckenfelder. The oxygen deoxygenation constant K1 is obtained by measuring the biochemical oxygen demand (BOD, American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 6 represented by L in the rest of this presentation) due solely to the biochemical oxidation of the organic matter) as a function of the distance traveled or the time. This degradation process is described by equation (1) v xK x eLL . .0)( 1 . − = or x v KLL x 1 0)( lnln −= (4) ([13, 16, 24, 25]) L (X) is the organic charge at the distance X; L0 the organic load at the point of departure and v the flow velocity. The graphical method according to Thomas made it possible to determine kd by plotting on a graph (t / y) 1/3 (L = y = biochemical oxygen demand) as a function of time t. The line obtained made it possible to deduce the angular coefficient b and the ordinate at the origin has; and calculate kd as follows: a bkd .6= (5) Experimental measurements make it possible to determine L (5). Knowing kd and L (5), L0 is calculated using the relation: ( ) α= − = − tkdeL L . 5 0 1 1 (6) The consumption of oxygen during the degradation of organic matter creates a deficit (D) which is compensated at all times by atmospheric oxygen. The oxygen deficit is calculated as follows: )()( tSt CCD −= (7) With C (t): concentration of dissolved oxygen at time t and CS the concentration of dissolved oxygen at saturation at temperature T. The reaeration coefficient of the section k2 has been calculated by the relation: tkDD t .lnln 20)( −= (8) While the global constant of reaeration K2 is obtained by making the ratio of the réaération coefficient k2 by the average depth H of the section. H kK 2 2 = (9) The variation of the kinetic constants with the temperature is expressed by the Van't Hoff equation: American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 7 )20_( )20()( . T Ct kk θ°= (10) With θ = 1,047 (Phelps), only for temperatures above 15°C [16,23] The determination of these kinetic constants, by the graphical method, required the elimination of the extreme or aberrant points after calculation of the residuals (residual differences between the measured y and the estimated by the regression) .The latter is calculated in the following way : )( bmxyR iii +−= (11) A good correlation was obtained by eliminating the points (xi, yi) whose residues are greater than the average of squares of the residues [19, 21, 22]. The oxygen profile of the section is obtained by replacing the different kinetic constants in the Streeter-Phelps equation:         +        −      − −=−= −−− v xK v xK v xK SxSx eDee KK LKCDCC . 0 .. 12 01 )()( 221 ..)( (12) [13, 14, 15, 24] 2.5. Assessment of the quality of the model The modeling errors are the differences between the calculated values and the observed values. They were evaluated by determining the precision which translates the dispersion of the modeling errors εi by the Nash criterion [32] as follows: ∑ ∑       −       − −= ∧ ∧ 21 yy yy CN i i (13) With observed (or experimental) values and values calculated by the relation (12). When the adjustment between simulated and observed values is perfect CN = 1 while CN is negative when the model gives poorer results. 3. Results and discussion 3.1. Results of the measurement campaign The measurement campaign gave the results shown in Table 2: American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 8 Tableau 2 : Results of the measurement campaign (august-september 2015 and 2016) Samples Distance (Km) pH T (°C) Conductivity (μs/cm at 20 ° C) dissolved O2 (mg / L) COD (mgO2 / L) BOD5 (mgO2 / L) NH4 + (mg/L) Ptotal (mg/L) EA1 0,0 7,40±0,12 25,9±1,1 38,4±1,5 4,07±0,73 1022±20 498±10 10,24±1,12 9,23±1,11 EA2 1,7 7,30±0,11 26,1±1,0 36,4±2,1 3,89±0,73 780±14 390±15 8,94±1,22 5,78±0,49 Ebrute 3,8 12,55±0,25 28,60±0,61 14850,00±1200,51 0,84±0,12 1393,0±5 697±3 27,18±2,32 14,98±1,32 EV1 3,9 8,14±1,22 26,1±1,2 45,3±6,4 2,12±0,59 978±12 481±8 16,47±1,42 9,97±1,22 EV2 4,1 7,48±0,25 26,1±1,2 57,3±18,3 5,56±0,79 1185±18 653±8 5,64±0,65 6,22±1,11 EV3 4,2 7,22±0,27 26,0±1,1 40,5±13,8 5,29±0,65 11051±34 536±17 5,37±0,67 5,86±0,87 EV4 4,4 7,23±0,20 26,1±1,4 32,8±3,1 5,44±0,78 930±13 490±8 11,76±1,32 7,38±1,14 EV5 4,7 7,16±0,23 26,3±1,2 31,2±1,7 5,62±1,02 854±13 417±14 15,23±1,34 9,04±2,02 EV6 4,9 7,15±0,22 26,0±1,4 31,6±2,2 5,72±1,09 716±16 274±8 5,03±0,45 5,45±1,09 EV7 5,1 7,17±0,15 26,2±1,4 32,9±4,0 5,91±0,90 658±7 349±5 4,32±0,56 4,89±0,89 EV8 5,3 7,17±0,11 26,0±1,5 35,4±5,5 5,72±0,73 601±12 332±13 4,26±0,63 4,6±0,32 EV9 5,4 7,08±0,24 26,3±1,2 33,2±2,8 5,88±0,78 570±43 308±23 3,91±0,51 4,58±0,44 EV10 5,6 7,27±0,20 26,3±1,3 33,6±2,6 6,41±1,02 511±49 276±27 3,74±0,45 4,32±0,25 Norm sewages * 6–9 <35°C < 500 <300 Norms waters of surface * * 6,5-8,5 0–20 3 mg/L 0,5 mg/L 0,02 mg/L Legend :* : ** : American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 9 3.2. Results of the correlation analysis 3.2.1 Analysis of covariance and correlation coefficient Table 3: Statistical Processing of Data The results recorded in this table show that: • The calculated covariance values are all negative, which implies that a negative linear bond, that is to say that the increase of the variable (COD, BOD5, PO4 3-, NH4 + ...) in the abscissa is accompanied by a drop in dissolved O2 content, ordinate. • The values for COD and BOD5 with dissolved oxygen are high. In other words, the increase in COD and BOD5 is accompanied by a significant decrease in dissolved O2 content. The relationship or bond between the dissolved O2 and these two parameters is stronger. This decrease in dissolved oxygen is lower for the other variables. • However, the correlation coefficient gave high values for ammonium and phosphate involving a strong bond between these parameters and dissolved oxygen, and a statistically linear link. Which was not the case with COD and BOD5 • The coefficient of determination calculated for the pair of dissolved pH-O2 variables is higher than the coefficient of determination of the other pairs of variables. The value obtained for the pH shows that 70.46% of the values found make it possible to explain the O2 variation with the pH, conversely. The values of 0.4928 for dissolved PO4 3--O2, dissolved NH4 + -O2 and dissolved DBO5-O2 simply reflect the fact that almost half of the values found, ie only 49.28%, explain the link between these variables and the dissolved oxygen content. • Very low values of 2 ),( yxr (<0.2) for dissolved T-O2 couples, dissolved O2-mineralization, dissolved O2- Conductivity and dissolved COD -O2 imply that few results (less than 20%) can explain the intensity and direction of the connection. 3.2.2. Graphical Analysis N° Variables X-Y ),( yxCov ),' yxr ),( 2 yxr 1 pH- O2 dissous -0,50521515 -0,83943 0,7046 2 Cond- O2 dissous -5,65878788 -0,35541 0,169 3 Min- O2 dissous -7,9936 10-15 -0,38168 0,1457 4 COD - O2 dissous -184,939091 -0,44199 0,0171 5 BOD5-O2 dissous -76,2465152 -0,35789 0,4928 6 NH4 + -O2 dissous -6,39918636 -0,68594 0,4928 7 PO4 3-- O2 dissous -2,52228333 -0,70201 0,4928 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 10 Graphical analysis with Excel software has led to the figures below: Figure 1: pH and dissolved O2 correlation Figure 2 : Mineralization and dissolved O2 correlation Figure 3: Conductivity and dissolved O2 correlation Figure 4: Organic load and dissolved O2 correlation Figure 5: COD and dissolved O2 correlation Figure 6: NH4 + and O2 dissolved correlation Figure 7: Phosphorus and dissolved O2 correlation Figures 3, 4 and 5 show a strong dispersion of the points which affect the regression (with 22 rR = ) compared to the other figures. These results confirm those calculated from Table 2. However, covariance alone better illustrates the relationship between biochemical demand, chemical oxygen demand and dissolved oxygen. 3.2.3. Significance test t The results of the statistical test t show that the pairs of pH-O2, NH4 +-O2 and Ptotal-O2 variables have statistically significant linear links compared to the other pairs. This could be due to the sampling method, American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 11 conservation or method of analysis of these parameters. Special and sustained attention should be given to the measurement campaign. The creation of Congo River quality monitoring stations is a necessity to build a reliable and up-to-date database. Table 4: Results of the significance test t Parameters Correlation coefficient r(X, Y) test t │t│ tseuil Conclusion │t│> tseuil for α = 0,05 (risk 5%), dismissal of the hypothesis H0 càd not pH-O2 -0,839422313 - 4,88424 2122 4,8842 42122 2,22 8 H0 rejection, existence of a linear link statistically significative COD -O2 0,006301546 0,01992 7632 0,0199 27632 2,22 8 No H0 rejection, no statistically significant linear link BOD5-O2 -0,357894035 - 1,21204 3391 1,2120 43391 2,22 8 No H0 rejection, no statistically significant linear link NH4 +-O2 -0,685944604 - 2,98101 5702 2,9810 15702 2,22 8 H0 rejection, existence of a linear link statistically Ptotal-O2 -0,702006464 - 3,11714 5589 3,1171 45589 2,22 8 H0 rejection, existence of a linear link statistically 3.3. Determination of kinetic constants The physicochemical side of the modeling was the determination of the constants K1, K2, Lo and Do of the section [12, 25, 27]. This was done according to the Eckenfelder approach which uses the graphical method [16] 3.3.1. Determination of the kinetic constant of degradation of the organic matter kd and the ratio α The determination of the kinetic constant of degradation of the organic matter kd was made by the graphical method of Thomas. It consisted in measuring the L (BOD) of at least one sample taken at random for 7 days in the bottle. BOD and then graphed (t / y) 1/3 as a function of time (t). Note that y represents L (BOD) and t is time. The results obtained are shown in Table 5 and in Figure 9: Table 5: L (or BOD) as a function of time (day) Temps, t (jour) L (mg O2/L) (t/y)(L=y) (t/y) 1/3 1,0 239 0,00418 0,16110 2,0 222 0,00901 0,20810 3,0 211 0,01422 0,24220 4,0 208 0,01923 0,26790 5,0 205 0,02439 0,29000 6,0 202 0,02970 0,30970 7,0 200 0,03500 0,32710 From the equation of the line y = 0.0268x + 0.151 (of the form y = bx + a), we deduce a and b in order to calculate kd by the relation (5), 1.064900662 j-1 at 20 ° C and 1.47670113 at 26 ° C by the relation (10). Eckenfelder asserts that when kd > 0.15 d-1, the discharges are unpurified, while for kd <0.10 d-1, the effluents American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 12 are purified [16]. In view of this assertion, the waters of the section of the left bank of the Congo River (Kinshasa) are unpurified since kd = 1.064 d-1 > 0.15 d-1 and therefore very rich in organic matter. The value of kd and the experimental time to measure BOD5 (t = 5 days) made it possible to calculate the ratio between the organic load at the start of the degradation process in the BOD (L0) flask and that remaining after 5 days. (L (5)) using the relation (6): α = L0 / L (5) = 1.000621805. The value of α gives the possibility of knowing the quantity of initial organic matter L0. Figure 9: (t / y) 1/3 as a function of time (t) 3.3.2. Determination of the global deoxygenation constant K1 The determination of K1 was made by plotting on graph ln BOD5 (kg / d) as a function of the distance traveled x (Km). Table 6: Conversion of BOD5 (mg / L) to BOD5 (kg / d) and calculation of residues Xi yi (mesuré) Ri Distance traveled (Km) Débit m3/sec BOD5 mg/L BOD5 kg/j ln BOD5 kg/j Résidus Ri 2 0,0 41000,0 498 1764115200 21,2909151 -0,049084901 0,002409328 1,7 41000,0 390 1381787234 21,0466436 -0,115026405 0,013231074 3,9 41000,0 481 1703894400 21,25618229 0,325292292 0,105815075 4,1 41000,0 362 1282939200 20,97241953 0,062509533 0,003907442 4,2 41000,0 232 821836800 20,52705239 -0,370584307 0,137332728 4,4 41000,0 428 1516737600 21,13982755 0,264849249 0,070145125 4,7 41000,0 483 1710979200 21,26033168 0,408116675 0,166559221 4,9 41000,0 305 1081848960 20,80193741 -0,027514286 0,000757036 5,1 41000,0 297 1050675840 20,77269945 -0,034093849 0,001162391 5,3 41000,0 287 1015960320 20,73910013 -0,05017487 0,002517518 5,4 41000,0 254 898352640 20,61607324 -0,153270756 0,023491925 5,6 41000,0 228 808375680 20,51053746 -0,240973541 0,058068247 Average Ri 2 0,048783092 The line obtained, y = 21.282-0.1033x, is identical to equation (1) which describes the degradation of organic matter. American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 13 Figure 10: Determination of J1 by carrying ln BOD5 (kg / d) as a function of distance traveled From this line, the ratio is 0.1033 Km-1 at 20 ° C and 0.1432 Km-1 at 26 ° C. K1 is 12.03 d-1 at 26 ° C for an average flow rate of 3.5 Km / h. 3.3.3. Determination of the overall kinetic constant of K2 reaeration The transfer coefficient k2 was determined by comparing the line, y = 1.0913 - 25.37x, obtained by carrying lnD as a function of time t as illustrated in FIG. 11 to expression (8). K2 is calculated using relation (9). For an average depth of 10 m, the value of K2 is 2.54 d-1. Tableau 7 Dissolved Oxygen vs. Time and Deficit Calculation Table 7 Xi Yi Time in (min) Time in (day) D=Cs-C lnD Ri Ri 2 5 0,003472222 5,22 1,65 -0,16891649 0,02853278 10 0,006944444 7,75 2,05 0,3149801 0,099212461 15 0,010416667 5,58 1,72 0,0754637 0,00569477 20 0,013888889 4,28 1,45 -0,10267033 0,010541196 25 0,017361111 4,10 1,41 -0,05680773 0,003227118 30 0,020833333 4,05 1,40 0,02040672 0,000416434 35 0,024305556 3,91 1,36 0,07395314 0,005469067 40 0,027777778 3,55 1,27 0,06450303 0,004160641 45 0,03125 2,91 1,07 -0,04431512 0,001963829 50 0,034722222 2,15 0,77 -0,25853799 0,066841892 55 0,038194444 2,27 0,82 -0,11750743 0,013807996 60 0,041666667 2,10 0,74 -0,10660563 0,011364759 65 0,045138889 2,24 0,80 0,04466067 0,001994576 70 0,048611111 2,14 0,76 0,09087893 0,008258979 75 0,052083333 2,12 0,75 0,16998631 0,028895344 Ymoy Average Ri 2 1,20 0,01935879 σ 0,42111898 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 14 Figure 11: Determination of K2 by carrying lnD as a function of time (days) The value obtained is in the range of 2.0-3.0 d-1 and corresponds to the reoxygenation coefficients of large rivers or rivers with moderate velocities according to Ugberbor [15]. 3.4. Oxygen profile The expression of the dissolved oxygen on the section, after calculations of the different kinetic constants, is the following one: S v x S v x v x xSx CeCCeeLDCC +−−      −=−= −−− 26,2 0 26,226,13 0)()( )(205,1)( (14) From expression (14), the distance at which the dissolved oxygen concentration is critical (almost zero) Xc and its Dc value were calculated as follows:               − −= 0 0 26,13 )(004,1111704,0ln 004,11 L CCvx S c (15) v x c eLD 26,2 08673,5 − = (16) where cx : distance at which the concentration of dissolved oxygen is critical (almost zero); SC : dissolved oxygen content at saturation; 0C : dissolved oxygen content at the point of departure and )( xC the dissolved oxygen content at the distance x from the starting point; Dc: Oxygen deficit at the critical point cx , x and v are respectively the distance traveled and the speed . Equation (14) and the Excel software gave the following profile: American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 15 Figure 12: Dissolved O2 profile of the section This theoretical profile shows that the oxygen content varies between 4.08 and 5.18 mg / L on this section while the experimental profile reveals a content of 4.08 mg / L dissolved O2 at the start and a maximum value of 6.41 mg / L of dissolved O2 downstream with a minimum of 2.12 mg / L of O2 dissolved at 3.9 Km of the path. Equations (15) and (16) allow us to say that the distance at which we will have a critical oxygen deficiency (= 0.34 mg / L) is 13.64 Km. Equation 10 also made it possible to simulate the variation of dissolved oxygen as a function of the organic load. The results obtained are illustrated in figure. 13: Figure 13: Variation of dissolved O2 as a function of organic load The higher the organic load at the point of departure, the greater the oxygen deficiency. The self-purification factor calculated for the section is 4.74; value close to that of torrents according to Cluis D. [25] but this can be affected by the morphology and the nature of the effluents. 3.5. Evaluation of the quality of the model The Nash criterion calculated by relation (13) gave the value of 0.1756506. This indicates that the adjustment between simulated and observed values is unsatisfactory. The two curves therefore have different and marked paces. These differences, relative to the theoretical profile, reveal the simplicity of the Streeter-Phelps model which does not integrate the physical phenomena (diffusion and dispersion: description of the movements of the masses) [29] and the biomass of the studied section (origin, quantity, type and special distribution of organic matter) [8, 24, 26, 31] but also the amount of input data as shown by the statistical results. American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 16 Cluis, in his work had shown that the assumptions emitted by Streeter-Phelps discard several factors that influence the level of dissolved oxygen in a watercourse. As a result, this model can only be used, at best, for preliminary estimates of the response of a watercourse to organic loads [24]. This corroborates with the results obtained. These deviations are also to be found in the morphology and the composition of the benthic bottom of the section which varies with the heterogeneity of the organic matter present in the medium [18, 24, 29]. The low value of the Nash criterion (0,1756506) confirms the results obtained. Eckenfelder also states that the size and age of the companies on this section determine the flow and characteristics of wastewater; and the quality of the receiving waters [18]. Edeline adds that the complexity of situations (reconciliation of discharge points, tributaries, diversity of port activities, etc.) is difficult to model, and as a result forecasts are rarely very accurate. According to Sophie S. and his colleagues the predictive aspect of modeling, although sought after, is not necessarily the most available because of the amount of input data that it imposes and the impossibility of strict verification [10]. The statistical test has shown that the correlation coefficient ),( yxr Pearson is -0.35789 for the pair of dissolved BOD5-O2 variables, an important parameter of the model. This value actually shows that the content of dissolved O2 varies inversely with the but the intensity of this relationship is low in our case. The reason for these discrepancies is the sampling technique and the amount of data. Continuous sampling is more recommended than instantaneous sampling. The establishment of a network monitoring the quality of rivers along the Congo River, however, will have sufficient data. 4. Conclusion The study of the impacts of discharges on receiving watercourses using the models gives satisfactory results because they take into account both the quality of watercourses and discharges, and particularly the fact that river is actually taken a continuum. The model obtained in this study is adapted to simulate the fluctuations of dissolved oxygen as a function of the organic load. The point located at 3.9 Km reveals a severe and old pollution at this place (2.12 mg / L O2 dissolved compared to the theoretical value found 4.97 mg / L O2 dissolved). The low value of the Nash criterion shows that the Streeter-Phelps model, although satisfying in some respects, shows some weaknesses and needs improvement. This model constitutes for us a (basic) test model from which the physicochemical, ecological, biological and microbiological parameters influencing the oxygen balance in the watercourse will be progressively incorporated. This work is only the first step in the iterative process of constructing a quality model of a section of the Congo River because there is no universal model that can be applied indiscriminately, as Villeneuve would say . References [1] Hassan Partow, Problématique de l’Eau en République Démocratique du Congo, Défis et Opportunités, Rapport Technique, Programme des Nations Unies pour l’environnement(PNUE), Kenya, 2011 [2] Takoy Lomema. ’’Les ressources hydriques de la République Démocratique du Congo et les aspects connexes.’’ Actes du premier Colloque International sur la problématique de l’eau en R.D.Congo de American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 17 Kinshasa, Annales de la Faculté des Sciences, Presses de l’Université de Kinshasa, 2008, pp 51-59 [3] Yoann RAINAT, Implémentation du modèle de qualité des eaux QUAL2K sur le bassin versant de l’Arroux, M1 Sciences de l’Environnement, Université de Bourgogne ; 2015, pp 1-3. [4] Céline Jacmain, Adaptabilité des contrats de Rivières au Bassin versant de la Lukaya en République Démocratique du Congo, TFE MC, Belgique, 2010 pp 2. [5] Nada Nehme. Evaluation de la qualité du bassin inférieur de la rivière Litani, Liban, Approche environnementale (Thèse), Université de Lorraine ,2014 pp 1. [6] Doulaye Kone, Epuration des eaux usées par Lagunage à microphytes et à macrophytes en Afrique de l’Ouest et Centre : Etat des lieux, performances épuratoires et critères de dimensionnement (Thèse), Ecole Polytechnique Fédérale de Lausanne, 2002 pp 2. [7] Sikulisimwa Pole Céline, Diagnostic des biomasses épuratrices des stations d’épuration et des systèmes d’assainissement non collectif (ANC) (Thèse), Université de Liège, 2011, pp 12-14. [8] Laure Pierron. Etat Chimique des rivières, Gazette Rivière n°26 ,2012 pp 1-4. cpie-val-de- gartempe@cpa-lathus.asso.fr [9] Serge Hébert et Stéphane Légaré, Suivi de la qualité de l’eau des rivières et petits cours d’eau, Ministère de l’Environnement Gouvernement du Québec, Octobre 2000, pp 9-22. [10] Sophie S. et al. Catalogue de modèles de qualité des cours d’eau, Revue des principaux outils de modélisation de la qualité des cours d’eau, Centre d’Etudes Techniques Maritimes Et Fluviales (CETMEF) ; France, 2010, pp 21- 43 / 71-72. [11] Micheline Poirier et al. Lignes Directrices pour l’Utilisation des Objectifs environnementaux de Rejet relatifs aux rejets industriels dans le milieu aquatique, Canada, 2008, page 9-10 / 29-31. [12] Dubin Ch. La mise en modèle mathématique de l’autoépuration des rivières, la Houille Blanche /N°8, 1971, pp 705-710. . Article published by SHF and available at http://www.shf-lhb.org or http://dx.doi.org/10.1051/lhb/1971047 [13] Thomas J.Grizzard. Deoxygenation – Reaeration and the Streeter-Phelps Equation, CEE 5134, 25 October 2007, pp 2-10/10-35 [14] Gotovtesev A. V., Modification of the Streeter Phelps System with the Aim to Account for the Feedback between Dissolved Oxygen Concentration and Organique Matter Oxidation Rate, Water resources Vol.37 N°2, pp. 245–251, Russie, 2010. [15] J.N. Ugbebor, J.C. Agunwamba , V.E Amoh , Detrermination of reaeration coefficient K2 for pollued mailto:cpie-val-de-gartempe@cpa-lathus.asso.fr mailto:cpie-val-de-gartempe@cpa-lathus.asso.fr http://dx.doi.org/10.1051/lhb/1971047 American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 18 stream as a function of de depth , hydraulic radius , temperature and velocity ; Nigerian Journal of Technology (NIJOTECH) , Vol 31 ,N° 2 July , 2012 , pp 174-180. [16] Eckenfelder W.W, Gestion des eaux usées urbaines et industrielles ; Caractérisation-Techniques d’épuration-Aspects économiques, Technique et Documentation, Paris, 1982, pp 11-37, 38-70 [17] Stéphanie Laronde et Katell Petit, Bilan national des efforts de surveillance de la qualité des cours d’eau (Rapport final), Office national de l’eau et des milieux aquatiques (Onema) et Office Internationale de l’Eau (OIE), 2010 , pp 881-85. [18] Jean Rodier Jean RODIER, Bernard LEGUBE, Nicole MERLET et coll. L’analyse de l’eau, 9ième édition entièrement mise à jour. Dunod, Paris ,2009 . [19] Panaretos V.M., Davison A.C. et Kuonen. Probabilités et Statistique pour Sciences de l’Environnement ; Ecole Polytechnique Fédérale de Lausanne, 2013 , pp 74-78. [20] Bardel C. La corrélation : Présentation et test du coefficient de corrélation de Pearson .Master 2 Neurosciences, Septembre 2016 pp 1-30 ; Source : http://xkcd.com/552 [21] Jacques Nicolas, Outils et statistiques et mathématiques pour le traitement des données environnementales, Surveillance de l’Environnement, ULG, page 123-159, 2005. [22] Ricco Rakotomalala. Analyse de corrélation, Etude des dépendances-Variables quantitatives, version1.1. Université Lumière Lyon 2, 2017, pp 3-17. http://dis.univ-lyon2.fr/ [23] Edeline F. Pouvoir auto épuratoire des rivières, Une introduction critique. Cebedoc, Paris 2001, pp 180 -200. [24] Ellis, T.G. ; D.S. Barbeau ; B.F. Smets et C.P.J. Grady, Détermination des paramètres cinétiques décrivant la biodégradation par respirométrie, Recherche environnementale en eau 68(5) pp 917- 926.1996. [25] Cluis D. Analyse des réactions en rivière : Modèles mathématiques de qualité, Revue de littérature, Rapport scientifique n°23, Institut National de la Recherche Scientifique (INRS)-Eau, Université du Québec ,1973 pp : 3-4 / 22-40. [26] Daniel Maurer, Modélisation de la dispersion des effluents rejetés au wharf de la Salie, Agence de l’Eau 1997 pp 1-2. [27] Kabele Ngiefu C. et Munginda Yong F. Pollution et degré d’auto-épuration des rivières de Kinshasa Actes du premier colloque international sur la problématique de l’eau en République Démocratique du Congo (RDC), 2008 pp 156-174. http://xkcd.com/552 http://dis.univ-lyon2.fr/ American Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) (2018) Volume 45, No 1, pp 1-19 19 [28] Lefort D., Modèles mathématiques de pollution de rivière, Centre de Recherche de Pont-à-Mousson, la Houille Blanche N°8, 1971, pp 669-704. [29] Pierre-Alain Roche, Jacques Miquel, Eric Gaume ; Hydrologie quantitative, Processus modèles et aide à la décision ; Springer Verlag France, 2012 pp 242-244. [30] Liechti Paul, Méthodes d’analyse et d’appréciation des cours d’eau. Analyses physico-chimiques, nutriments. L’environnement pratique n°1005. Office fédéral de l’environnement, Berne. 44 p. 2010 pp 11-36. ; [31] Jérôme Dispan, Biodégradabilité de la matière organique dans le continuum aquatique, réseau d’assainissement-station d’épuration-milieu naturel récepteur (Thèse), Ecole National des ponts et chaussées, 2003, pp 29-36. [32] Villeneuve J.P., Hubert P., Mailhot A. et Rousseau A.N. La modélisation hydrologique et la gestion de l’eau. Journal des Sciences de l’Eau, Vol 11, 1998, pp19- 39 URL id.erudit.org/iderudit/705327ar Document generated on 10/24/2017 6:34 a.m. [33] Figuet C. et Frangi J.P. Les cours d’eau réceptrice de rejets de stations d’épuration : le cas de la Mauldre (Ile-de-France), un milieu sous très haute pression. Revue des Sciences de l’Eau 13(2) ; 2000 pp 119-138. [34] Tangou Tabou Thierry, Etude et modélisation de la contribution des macrophytes flottants (Lemna minor) dans le fonctionnement des Lagunes naturelles (Thèses), Université de Liège 2014 pp 2 : pp 91.