Palmer-type soil modelling for evapotranspiration in different climatic regions of Kenya 365Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.DOI: 10.15201/hungeobull.71.4.4 Hungarian Geographical Bulletin 71 2022 (4) 365–382. Introduction Kenya’s mainstay of the economy is pre- dominantly rainfed agriculture. Droughts of various severities, frequencies, timings, dura- tion, intensity, and spatial extent vary from one climatic region to another and threaten food security in the country (Huho, J.M. and Mugalavai, M.E. 2010; Bowell, A. et al. 2021; Kipkemboi, K.B. et al. 2021). Therefore, a bet- ter understanding of hydrological processes and the distribution of water balance com- 1 Department of Geophysics and Space Science, Institute of Geography and Earth Sciences, ELTE Eötvös Loránd University, H-1117 Budapest, Pázmány Péter sétány 1/C, Hungary. Corresponding authors e-mail: musyimipeter@student.elte.hu 2 Department of Humanities and Languages, Karatina University, P.O BOX 1957-10101, Karatina, Kenya. 3 Department of Meteorology, Institute of Geography and Earth Sciences, ELTE Eötvös Loránd University, H-1117 Budapest, Pázmány Péter sétány 1/A, Hungary. Palmer-type soil modelling for evapotranspiration in different climatic regions of Kenya Peter K. MUSYIMI1,2, Balázs SZÉKELY1, Arun GANDHI3 and Tamás WEIDINGER3 Abstract Reference evapotranspiration (ET0) and real evapotranspiration (ET) are vital components in hydrological processes and climate-related studies. Understanding their variability in estimation is equally crucial for micro- meteorology and agricultural planning processes. The primary goal of this study was to analyze and compare estimates of (ET0) and (ET) from two different climatic regions of Kenya using long-term quality controlled synoptic station datasets from 2000 to 2009 with 3-hour time resolution. One weather station (Voi, 63793) was sought from lowlands with an elevation of 579 m and characterized by tropical savannah climate while the other (Kitale, 63661) was sought from Kenya highlands with humid conditions and elevation of 1850 m above sea level. Reference evapotranspiration was calculated based on the FAO 56 standard methodology of a daily basis. One dimension Palmer-type soil model was used for estimating of real evapotranspiration using the wilting point, field capacity, and soil saturation point for each station at 1 m deep soil layer. The ratio of real and reference evapotranspiration dependent on the soil moisture stress linearly. Calculations of estimated evapotranspiration were made on daily and monthly basis. Applications of the site-specific crop coefficients (KC) were also used. The result indicated that the differences among daily and monthly scale calculations of evapotranspiration (ET) were small without and with an application of crop coefficients (ETKc). This was due to high temperatures, global radiation, and also high soil moisture stress due to inadequate precipitation experienced in the tropics where Kenya lies. Results from Voi showed that mean monthly ET0 ranged from 148.3±11.6 mm in November to 175.3±10.8 mm in March while ET was from 8.0±4.5 mm in September to 105.8±50.3 mm in January. From Kitale, ET0 ranged from 121.5±8.5 mm/month in June to 157.1±8.5 mm/month in March while ET ranged from 41.7±32.6 mm/month in March to 126.6±12.2 mm/month in September. This was due to variability in temperature and precipitation between the two climatic regions. The study concludes that ET0 and calculated evapotrans- piration variability among the years on a monthly scale is slightly higher in arid and semi-arid climate regions than in humid regions. The study is important in strategizing viable means to enhance optimal crop water use and reduce ET losses estimates for optimal agricultural yields and production maximization in Kenya. Keywords: crop coefficient, climatic regions, Kenya, reference evapotranspiration, real evapotranspiration, soil model Received April 2022, accepted November 2022. Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.366 ponents is important (Okello, C. et al. 2020; Ferina, J. et al. 2021) to cushion inhabitants against extreme meteorological events. The components which include reference evapo- transpiration (ET0) and real evapotranspira- tion without and with an application of the crop coefficient (ET and ETKc), precipitation (P), soil moisture content (θ), soil recharge, soil surface runoff (R), and soil moisture loss coupled with other soil parameters should be given an in-depth insight to aid in operation- alizing water management decisions. Hydrological processes are crucial in plant developmental stages in times of water ex- cess and/or stress. This is because crops have different rates of transpiration at dif- ferent stages compounded by other factors such as environment and management prac- tices (Zotarelli, L. et al. 2010; Ngetich, K.F. et al. 2012; Djaman, K. et al. 2017; Macharia, J.M. et al. 2021). For instance, during early crop developmental stages that is stages be- tween vegetative emergence (VE) and veg- etative tasseling (VT) (Ransom, J. et al. 2014) evaporation becomes the major process in water loss. For a fully grown crop, at repro- ductive stages (silking to physiological ma- turity) transpiration plays a major role and water stress causes more harm at the initial seedling stage and continued damage as crops near tasseling (Allen, R.G. et at. 1998). Soil water deficiency caused by unpredict- able precipitation is an impediment to high yields in agriculturally potential areas. This prompts timely planting to ensure optimum utilization of available soil water during the rainy season (Ferina, J. et al. 2021). Since ET0 and evapotranspiration are key determi- nants, long term modeling studies in Kenya and Africa are vital because of variations in water demand and soil characterization (Omondi, J.O. et al. 2017). A wide range of scientific methods have been used to estimate ET0 (Penman, H.L. 1948) from different cli- matic components. This study used the FAO 56 standard methodology to estimate ET0 on daily basis and one dimension Palmer-type soil model (Palmer, W.C. 1965; Ferina, J. et al. 2021) was used to estimate real evapo- transpiration. The main aim of the study was to model ET0, ET and ETKc in different climat- ic regions of Kenya. This was geared towards comparing changes in their estimates since they are influenced by climatic parameters and soil parameters which differ seasonally and from one climate region to another. Recent studies (Hao, X. et al. 2018; McColl, K.A 2020) identified incorrectness of vital lim- iting cases and surface energy imbalances. This is due to the heterogeneity of regional characteristics in the Penman-Monteith evapotranspiration method. They provid- ed a counter equation to correct the errors. McColl, K.A. (2020) suggests that it is more accurate in real-world conditions and it is not bound to additional assumptions, empiri- cism, or computational cost. This implies the complexity of the estimation of evapotranspi- ration since it relies on the heterogeneity na- ture of land surface features. As a key compo- nent of the hydrological cycle and its critical role in various sectors such as water resource management and agriculture (McColl, K.A. and Rigden, A.J. 2020), its study in various climate regions in Kenya which vary spa- tially resource-wise, is also very important in the current regime of climate change and variability. However, in our study, we relied on the standard and traditional Penman- Monteith method, because of its accuracy and ease of application to compute potential evapotranspiration. The goal of this study is to evaluate es- timates of reference (ET0), and real evapo- transpiration from two climatic regions of Kenya for proper planning and management of water resources using the traditional meth- odology (Zotarelli, L. et al. 2010; Ferina, J. et al. 2021) for present and future agricultural processes across water and agricultural sec- tors in daily and monthly time scale. Geography and climate of Kenya Kenya is geographically located at a longitude 34° E – 42° E, and latitude 5° S – 5° N. It has rich, diverse, and complex geomorphologi- 367Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. cal features which are key modifiers of the climate system. The highest point is Mount Kenya (5,199 m) above sea level, while ranges, arid and semi-arid plains, and plateaus dom- inate the majority of the land. To the south, it is the Indian Ocean that regulates coastal cli- mate (Ayugi, B. et al. 2020). In the western part of the country lies a complex rift valley lakes system. The climate varies from the modified tropical climate of the Kenya highlands to the desert climate of Central Northern Kenya (Obiero, J. and Onyando, J. 2013). Study area and data sources Different climatic regions of selected counties The study was carried out in the different cli- matic regions of Kenya and from two counties (Figure 1, Table 1). One, Trans-Nzoia County, is mountainous and climatically characterized by humid conditions, and the other, Taita-Tav- eta County, is lowland comprising of Taita, Mwambirwa and Sagalla hills with an altitude of 2,208 m a.s.l., and characterized by arid and semi-arid to tropical savannah climate. Trans- Nzoia County is humid, highland equatorial, mild, and generally warm and temperate. The Köppen-Geiger climate classification is Cfb (Peel, M.C. et al. 2007; Beck, H. et al. 2018). The annual average temperature is approximately 16 °C around Mount Elgon, and 28 °C in the lower areas. The diversity of agroecological factors coupled with agro-climatic zones has influenced spatial variation in the rainfed ag- riculturally productive region (Mbaisi, C.N. et al. 2016). Annual rainfall amount ranges between 1,267 mm to 1,808 mm while its ele- vation is between 1,800–2,000 m a.s.l. (Nyberg, J.M. et al. 2020). Taita-Taveta is 89 percent arid and semi- arid. It is characterized by a tropical savannah climate (Aw). Mean monthly temperature is approximately 23 °C while the maximum and minimum are approximately 18 °C and 25 °C (Ogallo, L.A. et al. 2019). Its climate is influ- enced by south-easterly winds. On average, the county highlands receive 265 mm of pre- cipitation, while the lowlands receive 157 mm during long rains between March, April, and May (MAM) while during short rains between October, November, and December (OND), rainfall amounts range from 341 mm in low- lands to 1,200 mm in highlands. Annual av- erage precipitation amounts to 650 mm. The county is divided into three major topograph- ical zones namely upper zone, comprising of Taita, Mwambirwa, and Sagalla hills region Fig. 1. Sketch map of Africa (a); Weather stations and their elevations (b); The two Kenyan counties under study: Trans-Nzoia, and Taita-Taveta in a sketch map of Kenya (c) Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.368 with altitudes ranging between 304 and 2,208 m a.s.l., the lower zone consists of plains and the zone of national parks and mining areas (Government of Kenya, 2013; Mwakesi, I. et al. 2020). It is dominated agriculturally by maize, beans, and peas. Maize crop is the staple food from both counties as well as the whole of Kenya. Dataset and quality control Data with 3-hour time resolution was down- loaded from Voi and Kitale synoptic weath- er stations and arranged into datasets from 2000–2009 (Meteomanz.com). Methodology of linear interpolation was used to check if the missing measurement periods were smaller than 12 hours. Mean daily course of the meteorological elements combined with the measured variables before and after the data gap was used for longer missing pe- riods. If the lack of data was between half a day and 5 days, then the missing period was replaced with the average daily course from the data of the days (1 or 2 depending on the length of the data gap) before and af- ter the missing period. If the data gap was even longer then we replaced the averages of 9 years for the given measuring period (8 measuring period each day). The meas- ured and gap-filled time periods have been aligned during the initial and final 12 hours of the data-deficient period (5–5 data points) with a linear or exponential approximation. Errors in the SYNOP messages (for in- stance bad digits) were also filtered in the temperature, relative humidity, pressure, wind speed, direction, and time series from the Meteomanz database based on a Visual Basic macro. 2.6 and 4.5 percent accounted for the missing data from Kitale and Voi SYNOP station daily data. The quality-assured data- base was arranged in Excel tables. After the data set was cleaned, step by step analysis of ET0 (FAO 56 methodology, Zotarelli, L. et al. 2010; Lakatos, M. et al. 2020), calcu- lated evapotranspiration using soil param- eters ET and extended with maize coefficient, ETKc was undertaken using own Visual Basic Macro programmes developed in MS Excel. Methodology Due to complexity of the climate param- eters required, FAO 56 standard methodol- ogy (Equation 1) of the daily base was used to estimate reference evapotranspiration, ET0 (Allen, R.G. et al. 1998). There are also many methods for the estimation of poten- tial evapotranspiration (Epot), for instance, temperature as well as both temperature and terrestrial radiation-based methods (McMahon, T.A. et al. 2013; Lang, D. et al. 2017; Musyimi, P.K. et al. 2021). The defini- tion of potential evapotranspiration is that from a surface of unlimited water but in this definition of potential evapotranspiration, the evapotranspiration rate does not relate to a specific crop while for the definition of reference evapotranspiration is that from a well-watered grass surface (Irmak, S. and Haman, D.Z. 2003). Penman-Monteith reference evapotranspi- ration method is accepted as accurate, adopt- ed, and recommended worldwide as a stan- dardized method for ET0 estimation across Table 1. Synoptic stations of counties under study, their geographical locations and duration of data set County Weather station WMO-ID Latitude Longitude Altitude, m Duration of data set Highland Trans-Nzoia Kitale 63661 0.9733°N 34.9588°E 1,850 2000–2009 Lowland Taita-Taveta Voi 63793 –3.3981°S 38.5581°E 579 2000–2009 369Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. various climatic characterization for instance in Muranga County of Kenya (Shilenje, Z.W. et al. 2015). One dimension Palmer-type soil model (Equations 2–7) was used for estimat- ing evapotranspiration using site-specific soil parameters (Table 2) which included wilting point, field capacity, and soil saturation point for each station at 1 m deep soil layer (Ács, F. and Breuer, H. 2006; Ács, F. et al. 2007; Dy, C.Y. and Fung, J.C.-H. 2016; Ferina, J. et al. 2021). This is obviously a rough approach since this study did not take into account the areal variability of the depth of the root zone. Therefore, the most commonly used 1 m depth value was applied. The Palmer-type evapotranspiration model has also been applied in Kenya in previous studies to compute climate-based indices and evaluate meteorological and agricul- tural droughts (Marshall, M.T. et al. 2012). The soil types with the parameters were obtained from a soil map of Kenya with a 5 km space resolution which was taken us- ing the Weather Research and Forecasting (WRF) model due to the scarcity of soil data parameters (Dy, C.Y. and Fung, J.C.-H. 2016) (see Table 2). The integration of the two models in the methodology was ap- plied because the use of evapotranspiration- driven models in agricultural studies is still in its initial stages due to data scarcity in sub-Saharan Africa (Marshall, M.T. et al. 2012). Ratio of real and reference evapo- transpiration dependent on the region- specific soil moisture stress. Application of the site-specific crop coefficients (KC) were also applied in the model (Equation 3). This study used the maize coefficient as specified by FAO (Allen, R.G. et al. 1998; Tyagi, N.K. et al. 2003), because it is the staple food of Kenya and widely grown in the counties un- der study (Luciani, R. et al. 2019). The param- eters used in Equation (1) are for daily time step by step in which analysis of each station climate data was done using Visual Basic Macro programmes developed in MS Excel. Considering the ith day of the year. The ref- erence evapotranspiration for the ith day of the year is [mm day–1]: Meteorological variables for the given day of the year (without the notation i) are: Rn = net radiation at the crop surface [MJ m–2 day–1], G = soil heat flux density [MJ m–2 day–1], T = mean daily air temperature at 2 m height [°C], u2 = wind speed at 2 m height [m s-1], calculated from the reference wind measure- ment in 10 m height (Zotarelli, L. et al. 2010; Lakatos, M. et al. 2020), es = saturation water vapour pressure [kPa], ea = actual water vapour pressure [kPa], es – ea = saturation water vapour pressure deficit [kPa], Δ = slope of the water vapour pressure curve [kPa °C–1], γ = psychro- metric constant [kPa °C–1]. The daily evapotranspiration without (LE) and with the crop coefficients (LEKc) calculat- ed by the 1D Palmer model for the ith day of the year is determined by the soil type (see Table 2), the plant constant (Kci) and the param- eterization of βi–1 function respectively. The latter is considered a simple linear function of the available soil moisture (θi–1 – WLT) (see Equation 4). There are other approaches as an exponential form of parametrization of function (Mintz, Y. and Walker, G.K. 1993). During the test calculations, no significant differences were observed among the different methodologies, so we kept the linear approximation. The initial soil moisture [in mm] in the upper 1 m deep soil layer is the previous daily (i – 1) value, θi–1 where ETi = real evapotranspiration, βi–1 = soil moisture availability parameter in a one-meter- deep layer of soil, WLT = wilting point, FC = field capacity, and Kci = specific crop coefficient for a given day. Parameterization knowing the amount of daily precipitation (Pi), of Runoff (Ri) provide the base on the daily water balance equation. (Units are mm in our cases.) (1) (2) (3) (4) Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.370 A simple 1D bucket model was applied. If, at the end of the day, the estimated soil moisture value exceeds the saturation point (θi > FC), then the remainder is considered as a runoff (Ri). The model does not take into account the terrain conditions nor the depth of groundwater and the surface water move- ment: The value of soil moisture, θi and corrected evaporation ETi (practically near zero) can also be easily calculated at the end of the ith day, when the soil moisture is near wilting point, as we know that: WLT ≤ θi–1 ≤ SAT: Although the Palmer-type soil model is globally and regionally used for its suitabil- ity in the analysis of hydrological processes, it has a limitation in that it does not consider the application and use of other soil proper- ties such as textural variation among soils, physical and chemical composition of vari- ous soils which vary from one region to the other (Ferina, J. et al. 2021). This study also did not put into consideration such inputs due to a range of issues such as scarcity, un- certainty, and unavailability of the properties of soil data from Kenya. However, the model is useful as it forms a basis for future studies of other soil properties as well as provides room for its improvement. A normality and hypothesis test The methods to analyze the normality of the time series used in the study are Kolmog- orov-Smirnov (K-S) test, and Shapiro-Wilk test (Ghasemi, A. and Zahediasl, S. 2012). The importance of the normality test was to help decide the statistical significance test for mean and standard deviation from the two counties. This study relied on the Shapiro- Wilk test as it is recommended for small samples. The distribution of 10-year rainfall data portrayed a normal distribution while monthly precipitation between the two coun- ties showed variation in normality among the months regardless of the season. A sim- ple F-test was used to compare the standard deviation of annual precipitation, ET, ET0, ETKc, from the two counties while the T-test (was used when the distribution between the months was normal) and Mann-Whitney U- test (was used when the distribution between months was skewed) were used to compare the monthly mean of precipitation. This was due to its applicability in determining the stability of time series data, its ease, and simplicity of use (Lim, G.-K. et al. 2020). The following four hypotheses were tested based on the normality of the data: – H0: There is no statistical significant dif- ference of annual precipitation, ET, ET0, ETKc, between the two climatic regions of the two counties. – H1: There is statistical significant difference of annual precipitation, ET, ET0, ETKc, between the two climatic regions of the two counties. – H0: There is no statistically significant dif- ference between the two climatic regions of Table 2. Soil characteristics* of synoptic stations in regions under investigation Stations Soil type for WRF model Saturation point (SAT), % v/v Field capacity (FC), % v/v Wilting point (WLT), % v/v Lowland Voi Sandy clay loam (7) 40.4 31.5 6.9 Highland Kitale Loam (6) 43.9 32.9 6.6 *According to Dy, C.Y. and Fung, J.C.-H. 2016. (5) (6) (7) 371Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. the two counties based on mean monthly precipitation data. – H1: There is statistically significant differ- ence between the two climatic regions of the two counties based on mean monthly precipitation data. Results and discussion This section describes the pattern of tempera- ture and precipitation and compares spatial variation of the mean monthly ET0 and esti- mates ET without crop coefficient (Kci ≡ 1) and evapotranspiration with crop (maize) coeffi- cient, ETKc from arid and semi-arid Taita-Taveta and humid Trans-Nzoia counties of Kenya. It also examines decadal and annual means (x) and standard deviations (σX) of precipitation (P), reference and estimated real evapotranspi- ration (ET0, ET, ETKc), ET/ET0 and ETKc/ET0 ra- tios. The importance of analyzing ratios of the regions under study was to determine evapora- tive stress indices which are also synonymous with drought index, a reflection of temperature properties on the surface. Evaporative stress indices have been previously used to examine droughts of various durations more so short- term, crop growth and irrigation demands as well as water stress (Yao, A.Y.M. 1974; Choi, M. et al. 2013; Anderson, M.C. et al. 2016; Liu, Y. et al. 2019). It was also important to compare the evaporative index using ET with and with- out the application of the maize coefficient. Temperature and rainfall pattern of the counties under study Results indicated that the mean monthly temperature ranged from 22.7 to 28.4 °C in Taita-Taveta County while in Trans- Nzoia County the range was between 17.8 to 21.9 °C. (There are tropical regions.) Mean annual precipitation in Taita-Taveta was 574.2±205.8 mm while absolute minimum and maximum were 212.3 mm in the year 2003, and 801.4 mm in the year 2004 respec- tively. The rainfall amount of 212.3 mm was too low compared to the mean of 574.2 mm, indicating drought. Similar droughts were experienced across Kenya in the 2000s. Arid and semi-arid regions which cover 80 percent of land mass were highly affected (Nyaoro, D. et al. 2016; Venton, C.C. 2018). In Trans-Nzoia County, the mean annual pre- cipitation was 1,200±174 mm while maximum and minimum absolute values were 1,014.2 mm in the year 2000 and 1,460.3 mm in the year 2001 respectively. There was a noticeable variation in precipitation in the two regions under study (Figure 2). This conforms to the results of Huho, Fig. 2. Monthly precipitation variation (2000–2009) in Voi synoptic station, Taita-Taveta County (a), and in Kitale synoptic station, Trans-Nzoia County (b) Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.372 Ta bl e 3 . D ec ad al m ea n m on th ly E T 0 i n Lo w la nd , V oi st at io n, T ai ta -T av et a Co un ty (A w ) Ye ar /M on th 1 2 3 4 5 6 7 8 9 10 11 12 M ea n (x ) SD ev . ( σ X ) 20 00 16 2. 9 17 6. 1 16 6. 4 15 8. 2 14 7. 3 13 5. 7 14 3. 6 15 5. 9 15 7. 8 17 8. 1 14 4. 2 15 0. 5 15 6. 4 13 .0 20 01 14 6. 3 15 8. 6 18 0. 2 15 4. 2 16 2. 8 14 3. 8 14 9. 0 17 0. 4 17 8. 1 19 1. 0 15 3. 1 13 9. 6 16 0. 6 16 .2 20 02 14 4. 4 15 9. 8 16 5. 3 15 7. 3 16 9. 9 17 7. 7 17 7. 3 15 7. 7 15 6. 4 16 4. 8 15 7. 7 15 1. 0 16 1. 6 10 .0 20 03 17 0. 1 16 4. 6 19 8. 3 17 7. 4 16 4. 7 16 7. 3 18 2. 2 17 5. 3 17 5. 3 19 1. 2 17 0. 4 18 2. 4 17 6. 6 10 .5 20 04 15 2. 7 14 9. 6 18 3. 0 16 0. 6 18 8. 2 16 0. 2 17 7. 1 17 2. 9 17 9. 2 16 7. 6 15 1. 6 14 7. 4 16 5. 8 14 .1 20 05 16 1. 2 16 9. 3 17 6. 4 16 5. 0 18 2. 3 17 4. 8 16 5. 5 16 4. 4 16 9. 4 18 1. 5 14 8. 5 17 2. 9 16 9. 3 9. 4 20 06 18 2. 3 17 3. 4 16 9. 4 15 3. 6 16 5. 8 16 6. 8 16 7. 6 17 4. 9 16 2. 2 15 6. 4 13 3. 1 13 7. 3 16 1. 9 14 .1 20 07 14 4. 5 15 5. 9 17 0. 9 17 3. 8 15 7. 3 16 7. 0 16 7. 3 15 7. 2 16 6. 7 17 4. 4 15 1. 1 15 0. 1 16 1. 4 10 .1 20 08 15 0. 8 15 6. 0 16 2. 1 14 7. 8 15 4. 4 15 0. 3 15 2. 3 15 2. 8 16 2. 4 18 3. 4 13 0. 3 14 0. 9 15 3. 6 12 .8 20 09 16 4. 4 14 9. 1 18 1. 3 16 8. 8 17 0. 7 15 4. 8 16 0. 1 15 2. 2 17 2. 0 15 4. 4 14 3. 5 14 0. 2 15 9. 3 12 .5 M ea n (x ) 15 7. 9 16 1. 2 17 5. 3 16 1. 7 16 6. 3 15 9. 9 16 4. 2 16 3. 4 16 7. 9 17 4. 3 14 8. 3 15 1. 2 – SD ev . ( σ X ) 12 .5 9. 4 10 .8 9. 4 12 .3 13 .6 12 .9 9. 3 8. 2 13 .2 11 .6 15 .0 Ta bl e 4 . D ec ad al m ea n m on th ly E T 0 i n H ig hl an d, K ita le st at io n, T ra ns -N zo ia C ou nt y (C fa ) Ye ar /M on th 1 2 3 4 5 6 7 8 9 10 11 12 M ea n (x ) SD ev . ( σ X ) 20 00 15 8. 0 16 2. 7 16 6. 0 12 8. 4 12 4. 7 11 9. 7 11 1. 0 12 2. 7 13 8. 8 12 7. 0 11 7. 4 13 6. 5 13 4. 4 18 .5 20 01 12 8. 4 14 7. 7 13 6. 8 12 3. 3 12 3. 3 11 2. 0 11 3. 9 12 3. 2 13 1. 9 12 5. 0 11 0. 4 13 9. 3 12 6. 3 11 .3 20 02 13 4. 3 15 1. 4 14 0. 8 12 9. 6 12 8. 0 12 3. 5 13 3. 7 13 4. 6 14 4. 2 13 4. 6 12 8. 1 12 9. 8 13 4. 4 7. 8 20 03 15 0. 7 15 5. 4 16 4. 4 13 1. 1 12 8. 5 11 9. 5 12 5. 2 12 4. 9 14 3. 0 14 2. 2 12 9. 1 15 0. 1 13 8. 7 14 .3 20 04 15 1. 1 15 2. 6 16 5. 8 12 5. 9 13 9. 4 12 6. 8 13 3. 5 13 4. 8 13 4. 8 14 6. 3 12 7. 3 14 3. 4 14 0. 1 12 .2 20 05 15 7. 3 16 3. 2 15 9. 9 14 6. 3 12 0. 4 12 4. 3 12 8. 7 13 6. 7 13 2. 8 14 1. 4 14 3. 9 16 6. 3 14 3. 4 15 .6 20 06 16 4. 4 15 6. 9 15 1. 9 13 9. 8 13 9. 3 12 9. 6 13 0. 8 13 2. 4 14 4. 8 15 0. 3 11 8. 2 12 3. 4 14 0. 2 14 .0 20 07 15 0. 3 13 5. 1 16 2. 7 14 1. 8 13 5. 4 10 4. 2 11 9. 1 12 8. 9 12 9. 4 14 3. 5 13 7. 0 15 1. 3 13 6. 5 15 .4 20 08 15 5. 8 15 5. 1 14 9. 9 14 1. 8 12 9. 2 12 1. 4 12 2. 6 12 3. 7 13 6. 9 13 2. 5 13 9. 7 15 3. 9 13 8. 5 12 .9 20 09 15 2. 5 16 5. 0 17 3. 2 12 3. 4 12 5. 5 13 3. 6 13 3. 0 14 1. 3 14 4. 9 14 0. 6 14 6. 1 13 4. 5 14 2. 8 14 .3 M ea n (x ) 15 0. 2 15 4. 5 15 7. 1 13 3. 1 12 9. 4 12 1. 5 12 5. 1 13 0. 3 13 8. 1 13 8. 4 12 9. 7 14 3. 8 – SD ev . ( σ X ) 11 .0 8. 8 11 .8 8. 5 6. 6 8. 5 8. 3 6. 6 5. 9 8. 3 12 .0 12 .8 373Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. J.M. (2017) who mentioned that the amount of rainfall received in a given region differs from year-to-year. For instance, in Machakos County, the coefficient of variation of 42 and 41 percent for MAM and OND rainfall sea- sons respectively for KARI Katumani station and 39 and 54 percent for Mutisya Mango Farm station respectively were experienced (Huho, J.M. 2017). Similarly, Ghaedi, S. (2021) noted that patterns of precipitation variability were evident across Iran from one year to another. Consequently, higher variability dominates the arid and semi-arid climatic regions across the world, which are characterized by low, unpredictable, and er- ratic rainfall amounts. Spatial variation of mean monthly ET0 in lowland, Taita-Taveta County (Aw) Reference evapotranspiration (ET0) was estimated using FAO 56 standard method- ology (Equation 1). Results indicated that decadal mean monthly reference evapotran- spiration varied from one year to the other and from one month to the other. This vari- ation was dependent on the seasons of the year since a greater percentage of Kenya exhibits two major rainy seasons, the MAM long rain season and OND, short rain season, both related to the influence of the ITCZ, but differing in the amount of precipitation re- ceived and its interannual and inter-seasonal variability (Camberlin, P. and Wairoto, J.G. 1997). For instance, Taita-Taveta experiences two rainy seasons, MAM and OND. The pre- cipitation climatology of countries near the Equator where Kenya lies is heterogeneous due to influences of topography, lakes, and seasonal dynamics of tropical winds (Ni- cholson, S.E. 2017). The highest decadal mean monthly val- ue of the reference evapotranspiration was 175.3 ±10.8 mm in March, while the low- est decadal mean monthly value was 148.3±11.6 mm in November for the 10 years of analysis (Table 3) during the two rainy seasons. The highest annual mean value was 176.6±10.5 mm/year in 2003, while the lowest was 153.6±12.8 mm/year in 2008 (see Table 3). During dry seasons, which have its peak from July to September and December to February experience no precipitation or very little amounts, there were relatively small differences of ET0 among months and high values of ET0 as shown in Table 3. This implies that ET0 estimates depend mostly on high temperatures and solar insolation but in cases where ET0 is larger than precipitation more so in the dry months irrigation is an op- tion to substitute the insufficient amount of precipitation and evaporative requirements by crops (Sadick, A. et al. 2015). Spatial variation of mean monthly ET0 in highland Trans-Nzoia County (Cfa) Results from humid Trans-Nzoia County in- dicated that the decadal mean monthly ET0 varied from one year to the other and from one month to the other but the estimates were lower than in Taita-Taveta County. This was because of lower temperatures experienced in Trans-Nzoia than in Taita-Taveta County. The highest mean value was 157.1±11.8 mm/month in March while the lowest mean monthly val- ue was 121.5±8.5 mm/month in June (Table 4). It was also evident that there were moderate differences among the months. The mean differences among the months were also moderate with the largest mean difference of 11.1 mm/month between November and December (see Table 4). Similar variations were observed by Djaman, K. et al. (2018) who stated that there were temporal and spatial variations in the month- ly average ET0 from January to December across Madagascar with January average ET0 (less than 5 mm/day), the highest more so in regions characterized by hot and dry climates while the lowest ET0, ranging from 3.27 to 3.70 mm/day, evident in the central- eastern humid region of Madagascar. Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.374 the lowlands during long rains (MAM); while during short rains (OND) the range is from 341 mm in lowlands to 1,200 mm in highlands (Mwakesi, I. et al. 2020). Textural characteristics vary greatly from one soil type to the other as stated by Mugo, J.W. et al. (2016) in a study in Kitui County of Kenya. During dry spells ET mostly depends on both soil texture and amount of precipi- tation which vary across different climatic regions (Ács, F. et al. 2007). For instance, in Taita-Taveta County, the type of soil was sandy clay loam while in Kitale it was loam (see Table 2). This implies that the amount of precipitation received in the lowlands is not sufficient enough to meet the requirements of ET0 and ET unlike in highlands or mountain- ous regions where rainfall is reliable. There were some few cases where small differences or equal estimates of ET0 and ET were equal, for example in January 2001, 2003, and 2007. In this month temperatures as well as pre- cipitation were low. These results concur with Ferina, J. et al. (2021) who stated that differences in ET0 and ET are small when precipitation is adequate and equally large when precipitation decreases. The variabili- ty in ET0 and ET is dependent on soil mois- ture, recharging of lost soil moisture through precipitation, nature, and type of land cover among other heterogeneous land character- Fig. 3. Mean monthly variability of ET0 and ET in Voi synoptic station, Taita-Taveta County (a), and in Kitale synoptic station, Trans-Nzoia County (b) Mean monthly reference and real evapotranspiration estimation using soil parameters from the two counties Results of the estimates indicated variation in ET0 and ET from two counties. In Taita-Taveta County the differences in mean monthly ET esti- mates were small and almost followed the same trend varying from month to month and year to year (Figure 3, a). Mean monthly ET estimates ranged from 8.0±4.5 mm/month in September to 105.8±50.3 mm/month in January. This was due to varying precipitation amounts and si- multaneously soil moisture contents. In Trans- Nzoia County, the differences in estimates were slightly high and indicated a noticeable difference (see Figure 2, b). Mean monthly ET (without application of plant constant, Kc) es- timates ranged from 41.7±32.6 mm/month in March to 126.6±12.2 mm/month in September. Mean monthly ET dependent on the season of the year and varied from one climatic region to the other but greater variation was experienced in arid and semi-arid climates. However, in long and short rainy seasons real evapotranspiration is nearly independent of soil textural charac- teristics because of adequate precipitation. For instance, in Trans-Nzoia County, precipitation ranges from 1,267 mm to 1,808 mm (Nyberg, J.M. et al. 2020) while Taita-Taveta County re- ceives 265 mm in the highlands and 157 mm in 375Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. istics. However, its estimation is of greater importance to water and agricultural sectors as well as ecosystem stability and well-being (McColl, K.A. and Rigden, A.J. 2020). In Taita-Taveta County, Voi station (Figu- re 3, a), in many instances, ET was too low compared to ET estimates of Trans-Nzoia County (Figure 3, b). For instance, ET ranged from 8.0±4.5 mm/month in September to 105.8±50.3 mm/ month in January. This was be- cause of high temperatures and unreliable pre- cipitation amounts. This increased soil mois- ture content stress brings a deficit due to aridity conditions of the region and this impacts agri- culture and brings potentially adverse effects to the yields hence food insecurity. Contrary to this, in Kitale station (see Figure 2, b), Trans- Nzoia County, ET was relatively higher and ranged from 41.7±32.6 mm/month in March to 126.6±12.2 mm/month in September and varying annually which also influence maize yield. According to Marshall, M.T. et al. (2012), variability in ET and maize yield correlations and incompatibility with the awaited growing season are highest in Western Kenya where Trans-Nzoia County lies geographically. There were some few cases where small differences or equal estimates of ET0 and ET were equal, for example in January 2007 because the region receives an adequate amount of precipitation, and the temperatures are usually low. These results are in tandem with Ferina, J. et al. (2021) who stated that different climatic regions vary significantly in terms of agricultural essentials, more importantly, soil moisture content, pre- cipitation amount, temperature, and sufficient water. Therefore, the estimation of ET0 and ET is of fundamental importance in the agricultur- al potential region as well as the other sectors considering their variability. Daily reference evapotranspiration, real evapotrans- piration without and with maize coefficient (ETKc) Daily results in Taita-Taveta County in- dicated that, daily averages, ET0, ET and evapotranspiration with maize coefficient, ETKc was 5.3±0.9 mm/day, 1.6±1.2 mm/day and 1.6±1.2 mm/day respectively. There was practically no significant difference between evapotranspiration without and with maize coefficient as their estimates were almost the same but ET0 was higher in Taita-Taveta County (Figure 4, a) than in Trans-Nzoia (Figure 4, b) because of the high temperatures experienced in lowland Taita-Taveta County. The daily maximum value of ET0 was 8.5 mm/day and was ob- served on 13 January 2006, while the daily minimum absolute estimate was 1.7 mm/day and recorded on 31 May 2003 in Taita-Taveta County. This was because January is among the hottest months in the dry season of Janu- ary and February, while May is a month of the long rainy season, hence the variations among values in Taita-Taveta County. Similarly, in Trans-Nzoia County, the daily maximum absolute estimate of ET was 6.2 mm/day on 29 May 2001, whilst the ab- solute minimum value was 0.0 mm/day on three days: 19, 20 and 21 October 2003. On the other hand, daily evapotranspiration with modified maize coefficients, ETKc Kcini (initial period, Kcmid mid-season, the crop growth development period and Kcend, late season period) of 0.3, 0.75, 1.2, and 0.4 and 0.3, 0.8, 1.2, and 0.6 (Allen, R.G. et al. 1998) in Taita- Taveta and Trans-Nzoia County respectively. For the Kcend, the modification was done in this study as the average between crop de- velopment period and Kcend that was 0.8 and 0.9 in Taita-Taveta and Trans-Nzoia County respectively (Figure 5) were used to compute the estimations. As stated by Guerra, E. et al. (2011), crop coefficients are fundamentally vital for estimating the evapotranspiration of crops. They computed approximately simi- lar maize coefficients in Kenya of Kcini of 0.5, Kcmid of 1.0 and Kcend of 0.8. These values were computed for three crop stages, initial, mid- season, and end-season without considera- tion of the development stage. The reason for using the maize crop coeffi- cient is because of the importance of soil mois- ture content to crop growth and development and its deficiency can highly impact the yield of maize crop hence food insecurity in Kenya. Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.376 The government’s food security depends on the availability of enough quantity of maize to meet food demands. Further, it is the most important and staple food for over 90 percent of population, grown on 1.6 million hectares of land and 80 percent of its farming is prac- ticed by small-scale farmers (Wambugu, P.W. et al. 2012). As stated by Shanahan, J.F. and Nielsen, D.C. (1987), despite water availabil- ity’s importance in every stage of crop devel- opment, from germination to harvest, many crops are most sensitive to moisture deficits during the reproductive stages. This occurs mostly during the duration of tasseling, silk- ing, and pollination since water stress during the reproductive stages revitalizes deeper root growth (Mayaki, W.C. et al. 1976). Similarly, soil moisture distress can high- ly influence real evapotranspiration since if there is a moisture deficiency, evapotran- spiration requirements are not satisfied. For instance, Stegman, E.C. (1982) noted that a 1 percent decrease in seasonal real evapotranspiration led to an average loss of 1.5 percent in maize yield, whereas water stress more so during the reproductive stages more so the blister stage (10–14 days after silking) led to a 2.6 percent decline in maize crop yield. This can as well proportionate- ly explain the reduction in maize yield in Taita-Taveta County which experiences two growing seasons (MAM and OND). Daily maximum estimate of ETKc was 5.7 mm/day on 10 April 2009 while the daily absolute minimum estimate was 0.1 mm/day from 29 August to 28 of September 2003, 2005 and 2006 respectively. This was because April is a month of the long rainy season of Kenya while July, August, and September (JAS) are dry seasons with August and September re- cording high temperatures in the arid and semi-arid climatic region of Kenya. Daily estimates of Trans-Nzoia County showed that the mean daily average of, ET0 and ET was 4.5±0.9 mm/day, 3.1±1.1 mm/day and 3.2±1.2 mm/day respectively. These averages Fig. 4. Daily variability of ET0, ET and ETKc in Voi synoptic station, Taita-Taveta County (a), and in Kitale syn- optic station, Trans-Nzoia County (b) Fig. 5. Maize coefficients in Voi (two rainy seasons), and Kitale (single rainy season) annually 377Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. were moderately low as compared to Taita Taveta County (see Figure 3, b) because these regions received a high amount of precipitation and low temperatures hence meeting the re- quirements of ET. The daily absolute maximum ET0 for the 10 years (2000–2009) was 8.5 mm/ day on 21 February 2009 whilst the minimum absolute value was 1.6 mm/day on 4 July 2007. Similarly, evapotranspiration without and with maize coefficient maximum absolute daily es- timates was 7.0 mm/day and 6.6 mm/day on 1 November 2008 and 27 July 2007 respec- tively. These results are in tandem with those reported by Hobeichi, S. et al. (2018) who indi- cated that evapotranspiration had higher val- ues in the Sahel from September to November. The analysis also showed the daily abso- lute minimum values of ET without, and with Kc were 0.51 mm/day and 0.36 mm/day on the same day (27 March 2008) respectively. The daily, monthly, and annual variations of ET0, ET and ETKc was due to varying daily, monthly and annual precipitation amounts (see Figure 2) and deficiency of precipitation mean soil moisture content and evapotranspi- ration deficiency hence agricultural drought. This concurs with Marshall, M.T. et al. (2012) who indicated that deficits in estimated real evapotranspiration are a direct measure of crop stress and can be integrated into agricul- tural drought monitoring systems. Annual comparison of trends of P, ET, ET0, ETKc from the two counties Analysis also shows an annual variation of ET0 in the two climate regions, with a range of 1,837.6 mm/year to 2,119 mm/year in Taita-Taveta and 1,515.1 mm/year to 1,721.1 mm/year in Trans-Nzoia, and a dec- adal average of 1,950.7 mm/year in Taita- Taveta County and 1,650.3 mm/year in Trans-Nzoia County. These results are in agreement with Djaman, K. et al. (2018) who observed that across Madagascar, annual ET0 varied from 1,081 mm/year to 2,239 mm/year and averaged 1,620 mm/year. The highest value range of the long-term average annual ET0 was between 1,891 and 2,111 mm/year on the west- ern coast and northwestern coast (Djaman, K. et al. 2018). Further, from Table 5, it was evident that higher annual precipitation (P) amounts led to higher ET and lower annual precipitation amounts led to lower. For instance, in 2003, in Taita-Taveta County, the annual precipitation amount was 212.3 mm and the same year had the lowest ET estimate of 447.4 mm/year. This was due to the 2003 drought which was expe- rienced across the whole lower eastern. This result conforms with Yang, Z. et al. (2016) who mentioned that the decrease in ET is attributed to a decrease in precipitation amounts, and re- gions with less annual precipitation depict less. Analysis based on standard deviation (σ) from the annual mean precipitation, ET, ET0, ETKc was carried out for the whole decade (10 years) for the two counties. A simple F-test (Lim, G. et al. 2020) was used to determine whether the standard deviation between the two counties was statistically different. At a significance level (α) of 0.05, results showed that p values of precipitation, ET, ET0, ETKc from the two counties were greater than 0.05. This means we do not reject the null hypoth- esis and there is no statistically significant difference between the standard deviation of precipitation, of the two counties. Contrary to the expected results, this outcome implies a similarity of tropical annual precipitation cycles between the two counties and from the two climatic regions (Ilyés, C. et al. 2021). A further test of significance by T-test and Mann-Whitney U-test showed that there is no statistically significant difference between the monthly mean precipitation of January, February, March, and November. There exist a statistically significant difference between the monthly mean precipitation of April, May, June, July, August, September, October, and December. These differences and similarities may be attributed to climate variability and change which causes shifts in air and ocean currents circulation linked to ITCZ hence the anomalies (Ayugi, B. et al. 2016; Obwocha, E.B. et al. 2022) hence change in monthly weath- er pattern since precipitation distribution is mostly irregular Kenya in time and space. The Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.378 similarity of the monthly mean precipitation was regardless of the season since it did not follow the seasonal pattern of Kenya. Conclusions In this study, we estimated long-term (2000– 2009) reference and real evapotranspiration using a one-dimensional Palmer-type soil model for two synoptic stations found in two different climatic regions of Kenya. It was established that: 1. Annual ET0 estimates were high in Taita- Taveta County with values ranging from 1,838 mm/year to 2,119 mm/year compared to Trans-Nzoia County where its estimates ranged from 1,515 mm/year to 1,721 mm/year. 2. Annual evapotranspiration without (ET) and with maize coefficient (ETKc) var- ied in the two climatic regions under study. The estimates ranged from 447 mm/year to 737 mm/year and 461.6 mm/year to 748.8 mm/year in Taita-Taveta County re- spectively, whilst in Trans-Nzoia the range was from 972 mm/year to 1,453 mm/year and 958 mm/year to 1,560 mm/year respectively. 3. The evaporative stress indices were low in Taita-Taveta County with and without the maize coefficient. It ranged from 0.22 to 0.38 and from 0.21 to 0.38 with and without maize coefficient respectively. This was because of the low amount of annual precipitation and high temperatures. However, in Trans-Nzoia County, evaporative stress indices were high with a range of 0.60 to 0.95 and 0.57 to 0.89 with and without maize coefficient due to varying annual precipitation amounts among the years. 4. The highest evaporative stress index was experienced in Trans-Nzoia County when the annual precipitation was high. For instance, in 2001 and 2007 the annual rain- fall amount was 1,460.3 mm and 1,460 mm respectively, and the evaporative stress indi- ces were 0.84 and 0.81 with and without maize coefficient in 2001 and 0.95 and 0.89 with and without crop coefficient in 2007 respectively. The study concludes that these results are important to different climatic conditions Ta bl e 5 . D ec ad al a nn ua l m ea n an d st an da rd d ev ia tio n of a nn ua l p re ci pi ta tio n (P ), re fer en ce ev ap ot ra ns pi ra tio n (E T 0) an d th e e st im at ed ev ap ot ra ns pi ra tio n (E T) w ith ou t an d w ith p la nt (m ai ze ) c oe ffi ci en t ( ET Kc ) f ro m th e t w o co un tie s C ou nt ie s In di ca to r Ye ar s M ea n (x ) SD ev . (σ X ) 20 00 20 01 20 02 20 03 20 04 20 05 20 06 20 07 20 08 20 09 Ta ita -T av et a (T ro pi ca l s av an na h, ar id a nd s em ia ri d) P* ET 0 ET ET K c ET /E T 0 ET Kc /E T 0 69 6. 4 18 76 .5 59 8. 7 66 9. 9 0. 32 0. 36 70 4. 8 19 27 .1 73 7. 3 74 8. 8 0. 38 0. 38 76 3. 8 19 39 .4 67 5. 9 66 7. 8 0. 35 0. 34 21 2. 3 21 19 .0 44 7. 4 46 7. 2 0. 21 0. 22 80 1. 4 19 85 .5 67 7. 7 66 5. 0 0. 34 0. 33 32 6. 2 20 31 .1 45 0. 8 46 1. 6 0. 22 0. 22 76 5. 5 19 42 .8 50 7. 0 49 4. 2 0. 26 0. 25 46 3. 0 19 36 .2 60 4. 1 59 8. 4 0. 31 0. 31 43 2. 7 18 37 .6 50 5. 4 51 0. 8 0. 28 0. 27 57 6. 0 19 11 .4 46 6. 1 46 9. 9 0. 24 0. 25 57 4. 2 19 50 .7 56 7. 0 57 5. 4 0. 29 0. 29 20 5. 8 79 .4 10 5. 9 10 6. 8 0. 06 0. 06 Tr an s- N zo ia (H um id ) P* ET 0 ET ET K c ET /E T 0 ET Kc /E T 0 10 14 .2 16 12 .8 10 05 .6 10 88 .7 0. 62 0. 67 14 60 .3 15 15 .1 12 29 .4 12 73 .1 0. 81 0. 84 10 68 .2 16 12 .8 10 56 .1 10 77 .6 0. 65 0. 66 13 35 .7 16 64 .1 12 12 .1 13 25 .4 0. 72 0. 80 10 42 .0 16 81 .7 97 2. 0 95 7. 6 0. 57 0. 57 10 25 .5 17 21 .1 11 23 .7 11 35 .0 0. 65 0. 66 12 82 .0 16 81 .7 10 45 .9 10 08 .5 0. 62 0. 60 14 60 .0 16 38 .8 14 52 .7 15 60 .3 0. 89 0. 95 11 69 .9 16 61 .5 11 26 .4 11 45 .8 0. 68 0. 69 11 38 .6 17 13 .6 98 8. 0 99 4. 6 0. 57 0. 58 12 00 .0 16 50 .3 11 21 .2 11 56 .7 0. 68 0. 70 17 4. 0 60 .1 14 6. 5 18 3. 9 0. 11 0. 12 *m m /y ea r 379Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. in Kenya which are endowed with various agricultural necessities in terms of topogra- phy, amount of precipitation received, and different soil characteristics such as moisture, drainage, and depth among others. The vari- ation in estimated real evapotranspiration in various climatic regions should be applied in farmers’ decision-making in their choice of crops to be planted, variety of farming systems, choice of planting seasons, and duration of crops to maturity. Therefore, it is fundamentally important to estimate ET0, ET and ETKc with and without specific crop coefficients considering their daily, monthly and annual variability, more so in agricul- turally potential regions of Kenya. Further, the study is important for the shifting plant- ing seasons of various crops which differ in terms of soil moisture requirements. Acknowledgement: The authors are grateful to the editors, deputy editors, and anonymous reviewers for their useful, constructive comments and suggestions which contributed to the improvement of the quality of this scientific article. The corresponding author is also grateful to the Stipendium Hungaricum Doctoral Research Scholarship, Tempus Public Foundation of the Hungarian Government which financially sup- ports his Ph.D. research. REFERENCES Ács, F. and Breuer, H. 2006. Modelling of soil respira- tion in Hungary. Agrokémia és Talajtan 55. (6): 59–68. Available at https://doi:10.1556/Agrokem.55.2006.1.7 (in Hungarian) Ács, F., Breuer, H. and Szász, G. 2007. Estimation of actual evapotranspiration and soil water content in the growing season. Agrokémia és Talajtan 56. (2): 217–236. Available at https://doi.org/10.1556/ agrokem.56.2007.2.3 Allen, R.G., Pereira, L.S., Raes, D. and Smith, M. 1998. Crop evapotranspiration: FAO Irrigation and Drainage Paper 56. Rome, UN FAO. Available at http://www.fao.org/3/X0490E/X0490E00.htm Anderson, M.C., Zolin, C.A., Sentelhas, P.C., Hain, C.R., Semmens, K., Yilmaz, M.T., Gao, F., Otkin, J.A. and Tetrault, R. 2016. The Evaporative Stress Index as an indicator of agricultural drought in Brazil: An assessment based on crop yield impacts. Remote Sensing of Environment 174. 82–99. Available at https://doi.org/10.1016/j.rse.2015.11.034 Ayugi, B., Wen, Y.W. and Chepkemoi, D. 2016. Analysis of spatial and temporal patterns of rain- fall variations over Kenya. Environmental Earth Sciences 6. (11): 69–83. Available at https://www. researchgate.net/journal/Environmental-Earth- Sciences-1432-0495 Ayugi, B., Tan, G., Niu, R., Dong, Z., Ojara, M., Mumo, L., Babaousmail, H. and Ongoma, V. 2020. Evaluation of meteorological drought and flood scenarios over Kenya, East Africa. Atmosphere 11. (3): 307. Available at https://doi.org/10.3390/ atmos11030307 Beck , H., Zimmermann , N., McVicar, T.R., Vergopolan, N., Berg, A. and Wood, E.F. 2018. Present and future Köppen-Geiger climate clas- sification maps at 1-km resolution. Scientific Data 5. 180214. Available at https://doi.org/10.1038/ sdata.2018.214 Bowell, A., Salakpi, E.E., Guigma, K., Muthoka, J.M., Mwangi, J. and Rowhani, P. 2021. Validating commonly used drought indicators in Kenya. Environmental Research Letters 16. (8): 084066. Available at https://doi.org/10.1088/1748-9326/ac16a2 Camberlin, P. and Wairoto, J.G. 1997. Intra-seasonal wind anomalies related to wet and dry spells dur- ing “long” and “short” rainy seasons in Kenya. Theoretical Applied Climatology 58. 57–69. Available at https://doi.org/10.1007/BF00867432 Choi, M., Jacobs, J.M., Anderson, M.C. and Bosch, D.D. 2013. Evaluation of drought indices via re- motely sensed data with hydrological variables. Journal of Hydrology 476. 265–273. Available at http:// doi.org/10.1016/j.jhydrol.2012.10.042 Djaman, K., Irmak, S. and Futakuchi, K. 2017. Daily reference evapotranspiration estimation under lim- ited data in Eastern Africa. Journal of Irrigation and Drainage Engineering 143. (4): 0001154. Available at https://doi.org/10.1061/ (ASCE) IR.1943-4774.000115 Djaman, K., Ndiaye, P.M., Koudahe, K., Bodian, A., Diop, L., O’Neill, M. and Irmak, S. 2018. Spatial and temporal trend in monthly and annual refer- ence evapotranspiration in Madagascar for the 1980–2010 periods. International Journal of Hydrology 2. (2): 110–120. Available at https://doi.org/10.15406/ ijh.2018.02.00058 Dy, C.Y. and Fung, J.C.-H. 2016. Updated global soil map for the Weather Research and Forecasting model and soil moisture initialization for the Noah land surface model. Journal of Geophysical Research: Atmospheres 121. 8777–8800. Available at https://doi. org/10.1002/2015JD024558 Ferina, J., Vučetić, V., Bašić, T. and Anić, M. 2021. Spatial distribution and long-term changes in water balance components in Croatia. Theoretical Applied Climatology 144. 1311–1333. Available at https://doi. org/10.1007/s00704-021-03593-1 Ghaedi, S. 2021. Anomalies of precipitation and drought in objectively derived climate regions https://akjournals.com/view/journals/0088/55/1/article-p59.xml https://doi.org/10.1556/agrokem.56.2007.2.3 https://doi.org/10.1556/agrokem.56.2007.2.3 http://www.fao.org/3/X0490E/X0490E00.htm https://doi.org/10.1016/j.rse.2015.11.034 https://www.researchgate.net/journal/Environmental-Earth-Sciences-1432-0495 https://www.researchgate.net/journal/Environmental-Earth-Sciences-1432-0495 https://www.researchgate.net/journal/Environmental-Earth-Sciences-1432-0495 https://doi.org/10.3390/atmos11030307 https://doi.org/10.3390/atmos11030307 https://doi.org/10.1038/sdata.2018.214 https://doi.org/10.1038/sdata.2018.214 https://doi.org/10.1088/1748-9326/ac16a2 https://doi.org/10.1007/BF00867432 http://doi.org/10.1016/j.jhydrol.2012.10.042 http://doi.org/10.1016/j.jhydrol.2012.10.042 https://ascelibrary.org/doi/10.1061/%28ASCE%29IR.1943-4774.0001154 https://medcraveonline.com/IJH/spatial-and-temporal-trend-in-monthly-and-annual-reference-evapotranspiration-in-madagascar-for-the-1980-2010.html https://medcraveonline.com/IJH/spatial-and-temporal-trend-in-monthly-and-annual-reference-evapotranspiration-in-madagascar-for-the-1980-2010.html https://doi.org/10.1002/2015JD024558 https://doi.org/10.1002/2015JD024558 https://doi.org/10.1007/s00704-021-03593-1 https://doi.org/10.1007/s00704-021-03593-1 Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.380 of Iran. Hungarian Geographical Bulletin 70. (2): 163–174. Available at https://doi.org/10.15201/ hungeobull.70.2.5 Ghasemi, A. and Zahediasl, S. 2012. Normality tests for statistical analysis: a guide for non-statisticians. International Journal of Endocrinology and Metabolism 10. (2): 486–489. Available at https://doi.org/10.5812/ ijem.3505 Government of Kenya, 2013. County Integrated Development Plan, Taita Taveta County. Government of Taita Taveta. Nairobi, Government Printers. Guerra, E., Snyder, R.L. and Ventura, F. 2011. Crop coefficients to adapt for climate change impact on evapotranspiration. Italian Journal Agrometeorology 14. (2): 15–16. Available at https://www.research- gate.net/publication/256462155 Hao, X., Zhang, S., Li, W., Duan, W., Fang, G., Zhang, Y. and Guo, B. 2018. The uncertainty of Penman-Monteith method and the energy balance closure problem. Journal of Geophysical Research: Atmospheres 123. 7433–7443. Available at https://doi. org/10.1029/2018JD028371 Hobeichi, S., Abramowitz, G., Evans, J. and Ukkola, A. 2018. Derived Optimal Linear Combination Evapotranspiration (DOLCE): A global gridded synthesis ET estimate. Hydrology and Earth System Sciences Discussions 22. 1317–1336. Available at https://doi.org/10.5194/hess-22-1317-2018 Huho, J.M. and Mugalavai, M.E. 2010. The ef- fects of droughts on food security in Kenya. The International Journal of Climate Change: Impacts and Responses 2. (2): 61–72. Available at https:// doi:10.18848/1835-7156/CGP/v02i02/37312 Huho, J.M. 2017. An analysis of rainfall character- istics in Machakos County, Kenya. IOSR Journal of Environmental Science, Toxicology and Food Technology 11. (4): 64–72. Available at www.iosrjour- nals.org. https://doi.org/10.9790/2402-1104026472 Ilyés, C., Wendo, V.A.J.A., Carpio, Y.F. and Szűcs, P. 2021. Differences and similarities between precipi- tation patterns of different climates. Acta Geodaetica et Geophysica 56. 781–800. Available at https://doi. org/10.1007/s40328-021-00360-6 Irmak, S. and Haman, D.Z. 2003. Evapotranspiration: Potential or Reference. Agricultural Engineering Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences. Gainesville, University of Florida, ABE. Kipkemboi, K.B., Kumar, L. and Koech, R. 2021. Climate change and variability in Kenya: a review of impacts on agriculture and food security. Environment, Development and Sustainability 23. 23–43. Available at https://doi.org/10.1007/s10668-020-00589-1 Lakatos, M., Weidinger, T., Hoffmann, L., Bihari, Z. and Horváth, Á. 2020. Computation of daily Penman-Monteith reference evapotranspiration in the Carpathian Region and comparison with Thornthwaite estimates. Advances in Science and Research 16. 251–259. Available at https://doi. org/10.5194/asr-16-251-2020 Lang, D., Zheng, J., Shi, J., Liao, F., Ma, X., Wang, W., Chen, X. and Zhang, M. 2017. A comparative study of potential evapotranspiration estimation by eight methods with FAO Penman-Monteith method in Southwestern China. Water 9. 734. Available at https://doi.org/10.3390/w9100734 Lim, G.-K., Kim, B.-S., Lee, B.-H. and Jeung, S.-J. 2020. Effect of climate change on annual precipitation in Korea using data screening techniques and climate change scenarios. Atmosphere 11. (10): 1027. Available at https://doi.org/10.3390/atmos11101027 Liu, Y., Hao, L., Zhou, D., Pan, C., Liu, P., Xiong, Z. and Sun, G. 2019. Identifying a transition climate zone in an arid river basin using the evaporative stress index. Natural Hazards Earth System Science 19. 2281–2294. Available at http://doi.org/10.5194/ nhess-19-2281-2019 Luciani, R., Laneve, G. and Jahjah, M. 2019. Agricultural monitoring, an automatic procedure for crop mapping and yield estimation: The Great Rift Valley of Kenya case. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 12. (7): 2196–2208. Available at http://doi.org/10.1109/ JSTARS.2019.2921437 Macharia, J.M., Ngetich, F.K. and Shisanya, C.A. 2021. Parameterization, calibration and validation of the DNDC model for carbon dioxide, nitrous oxide and maize crop. Heliyon 7. (5): e06977. Available at https:// doi.org/10.1016/j.heliyon.2021.e06977 Marshall, M.T., Funk, C. and Michaelsen, J. 2012. Agricultural drought monitoring in Kenya using evapotranspiration derived from remote sensing and reanalysis data. USGS Staff - Published Research 978. Available at http://digitalcommons.unl.edu/ usgsstaffpub/978 Mayaki, W.C., Stone, L.R. and Teare, I.D. 1976. Irrigated and non-irrigated soybean, corn, and grain sorghum root systems. Agronomy Journal 68. 532–534. Available at https://doi.org/10.2134/agronj1976.0002 1962006800030028x Mbaisi, C.N., Kipkorir, E.C. and Omondi, P. 2016. The perception of farmers on climate change and variability patterns in the Nzoia River Basin, Kenya. Journal of Natural Sciences Research 6. (20): 89–97. Available at https://core.ac.uk/download/ pdf/234656697.pdf McColl, K.A. 2020. Practical and theoretical benefits of an alternative to the Penman-Monteith evapo- transpiration equation. Water Resources Research 56. e2020WR027106. Available at https://doi. org/10.1029/2020WR027106 McColl, K.A. and Rigden, A.J. 2020. Emergent simplic- ity of continental evapotranspiration. Geophysical Research Letters 47. e2020GL087101. Available at https://doi.org/10.1029/2020GL087101 https://doi.org/10.15201/hungeobull.70.2.5 https://doi.org/10.15201/hungeobull.70.2.5 https://doi.org/10.5812/ijem.3505 https://doi.org/10.5812/ijem.3505 https://www.researchgate.net/publication/256462155 https://www.researchgate.net/publication/256462155 https://doi.org/10.1029/2018JD028371 https://doi.org/10.1029/2018JD028371 https://doi.org/10.5194/hess-22-1317-2018 https://cgscholar.com/bookstore/works/the-effects-of-droughts-on-food-security-in-kenya https://cgscholar.com/bookstore/works/the-effects-of-droughts-on-food-security-in-kenya https://doi.org/10.9790/2402-1104026472 https://doi.org/10.1007/s40328-021-00360-6 https://doi.org/10.1007/s40328-021-00360-6 https://doi.org/10.1007/s10668-020-00589-1 https://doi.org/10.5194/asr-16-251-2020 https://doi.org/10.5194/asr-16-251-2020 https://doi.org/10.3390/w9100734 https://doi.org/10.3390/atmos11101027 http://doi.org/10.5194/nhess-19-2281-2019 http://doi.org/10.5194/nhess-19-2281-2019 https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=4609443 https://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=4609443 https://ieeexplore.ieee.org/document/8745683 https://ieeexplore.ieee.org/document/8745683 https://www.sciencedirect.com/science/article/pii/S240584402101080X https://doi.org/10.1016/j.heliyon.2021.e06977 https://doi.org/10.1016/j.heliyon.2021.e06977 http://digitalcommons.unl.edu/usgsstaffpub/978 http://digitalcommons.unl.edu/usgsstaffpub/978 https://doi.org/10.2134/agronj1976.00021962006800030028x https://doi.org/10.2134/agronj1976.00021962006800030028x https://core.ac.uk/download/pdf/234656697.pdf https://core.ac.uk/download/pdf/234656697.pdf https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020WR027106 https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2020WR027106 https://doi.org/10.1029/2020GL087101 381Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382. McMahon, T.A., Peel, M.C., Lowe, L., Srikanthan, R. and McVicar, T.R. 2013. Estimating actual, po- tential, reference crop and pan evaporation using standard meteorological data: a pragmatic synthesis. Hydrology and Earth System Sciences 17. 1331–1363. Available at https://doi:10.5194/hess-17-1331-2013 Meteomanz.com by NOAA. Available at http://www. meteomanz.com Mintz, Y. and Walker, G.K. 1993. Global fields of soil moisture and land surface evapotran- spiration derived from observed precipitation and surface air temperature. Journal of Applied Meteorology 32. (8): 1305–1334. Available at https:// doi.org/10.1175/1520-0450 (1993)032%3C1305: GFOSMA%3E2.0.CO; 2: Mugo, J.W., Kariuki, P.C. and Musembi, D.K. 2016. Identification of suitable land for green grain production using GIS based analytical hierar- chy process in Kitui County, Kenya. Journal of Remote Sensing & GIS 5. 170. Available at https:// doi:10.4172/2469-4134.1000170 Musyimi, P.K., Székely, B. and Weidinger, T. 2021. Long term (1901–2016) temperature based po- tential evapotranspiration and aridity index analysis for lower eastern region of Kenya. Egyetemi Meteorológiai Füzetek / Meteorological Notes of University 33. (9): 74–83. Available at https://doi. org/10.31852/EMF.33.2020.074.083 Mwakesi, I., Wahome, R. and Ichang’i, D. 2020. Mining impacts on society: A case study of Taita Taveta County, Kenya. Journal of Environmental Protection 11. 986–997. Available at https://doi. org/10.4236/jep.2020.1111062 Ngetich, K.F., Raes, D., Shisanya, C.A., Mugwe, J., Mucheru-Muna, M., Mugendi, D.N. and Diels, J. 2012. Calibration and validation of AquaCrop model for maize in sub-humid and semiarid regions of central highlands of Kenya. In Third RUFORUM Biennial Meeting 24–28 September 2012, Entebbe, Uganda. Kampala, Uganda, RUFORUM, 1528–1548. Nicholson, S.E. 2017. Climate and climatic vari- ability of rainfall over eastern Africa. Review of Geophysics 55. 590–635. Available at https://doi.org/ 10.1002/2016RG000544 Nyaoro, D., Schade, D. and Schmidt, K. 2016. Assessing the Evidence: Migration, Environment and Climate Change in Kenya. Geneva, International Organization for Migration IOM. Available at https://www.preventionweb.net/files/50534_as- sessingtheevidencekenya.pdf Nyberg, J.M., Ambjörnsson, E.L., Wetterlind, J. and Öborn, I. 2020. Smallholders’ awareness of adapta- tion and coping measures to deal with rainfall vari- ability in Western Kenya. Agroecology and Sustainable Food Systems 44. (10): 1280–1308. Available at https:// doi.org/ 10.1080/21683565.2020.1782305 Obiero, J. and Onyando, J. 2013. Climate, develop- ments in earth surface processes. In Kenya: A Natural Outlook Geo-Environmental Resources and Hazards. Developments in Earth Surface Processes 16. Eds.: Paron, P., Olago, O. and Omuto. T., Amsterdam, Elsevier, 39–50. Available at https://doi.org/ 10.1016/ B978-0-444-59559-1.00005-0 Obwocha, E.B., Ramisch, J.J., Duguma, L. and Orero, L. 2022. The relationship between climate change, variability, and food security: Understanding the impacts and building resilient food systems in West Pokot County, Kenya. Sustainability 14. 765. Available at https://doi.org/10.3390/su14020765 Ogallo, L.A., Omay, P., Kabaka, G. and Lutta, I. 2019. Report on Historical Climate Baseline Statistics for Taita Taveta, Kenya. Nairobi, IGAD Climate Prediction and Application Centre. Available at https://doi. org/10.13140/RG.2.2.25814.68165 Okello, C., Greggio, N., Giambastiani, B.M.S., Wambiji, N., Nzeve, J. and Antonellini, M. 2020. Modelling projected changes in soil water budget in coastal Kenya under different long-term climate change scenarios. Water 12. 2455. Available at https://doi. org/10.3390/w12092455 Omondi, J.O., Mungai, N.W., Ouma, J.P. and Baijukya, F.P. 2017. Shoot water content and reference evapo- transpiration for determination of crop evapotrans- piration. African Crop Science Journal 25. (4): 387–403. Available at http://doi.org/10.4314/acsj.v25i4.1 Palmer, W.C. 1965. Meteorological Drought. U.S. Research Paper 45. Indicator codes CLIM 029, LSI 007. Washington, DC, US Weather Bureau. Peel, M.C., Finlayson, B.L. and McMahon, T.A. 2007. Updated world map of the Köppen-Geiger climate classification. Hydrology and Earth System Sciences Discussions EGU 11. 1633–1644. Available at https:// doi.org/10.5194/hess-11-1633-2007 Penman, H.L. 1948. Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society of London, Series A. Mathematical and Physical Sciences 193. (1032): 120–145. Available at https:// doi.org/10.1098/rspa.1948.0037 Ransom, J., Endres, G.J., Berlund, D.R., Endres, G.J. and McWilliams, D.A. 2014. Corn Growth and Management: Quick Guide. Fargo, ND, North Dakota State University Extension Service. Sadick, A., Gaisie, E., Adjei, E.O., Agyeman, K., Nketia, K.A. and Asamoah, E. 2015. Assessment of the relationship between actual evapotranspiration, reference evapotranspiration and precipitation. A case study of Tono irrigation scheme. International Journal of Scientific Research in Science and Technology 1. (5): 307–317. Available at https://www.research- gate.net/publication/317167487 Shanahan, J.F. and Nielsen, D.C. 1987. Influence of growth retardants (anti-gibberellins) on corn veg- etative growth, water use, and grain yield under different levels of water stress. Agronomy Journal 79. 103–109. Available at https://doi.org/10.2134/ AGRONJ1987.00021962007900010022X https://hess.copernicus.org/articles/17/1331/2013/hess-17-1331-2013.html http://www.meteomanz.com http://www.meteomanz.com https://journals.ametsoc.org/view/journals/apme/32/8/1520-0450_1993_032_1305_gfosma_2_0_co_2.xml https://journals.ametsoc.org/view/journals/apme/32/8/1520-0450_1993_032_1305_gfosma_2_0_co_2.xml https://journals.ametsoc.org/view/journals/apme/32/8/1520-0450_1993_032_1305_gfosma_2_0_co_2.xml https://doi:10.4172/2469-4134.1000170 https://doi:10.4172/2469-4134.1000170 https://doi.org/10.31852/EMF.33.2020.074.083 https://doi.org/10.31852/EMF.33.2020.074.083 https://www.scirp.org/journal/paperinformation.aspx?paperid=104345 https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016RG000544 https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2016RG000544 https://www.preventionweb.net/files/50534_assessingtheevidencekenya.pdf https://www.preventionweb.net/files/50534_assessingtheevidencekenya.pdf https://www.tandfonline.com/doi/full/10.1080/21683565.2020.1782305 https://www.tandfonline.com/doi/full/10.1080/21683565.2020.1782305 https://www.sciencedirect.com/science/article/abs/pii/B9780444595591000050?via%3Dihub https://www.sciencedirect.com/science/article/abs/pii/B9780444595591000050?via%3Dihub https://doi.org/10.3390/su14020765 https://doi.org/10.13140/RG.2.2.25814.68165 https://doi.org/10.13140/RG.2.2.25814.68165 https://doi.org/10.3390/w12092455 https://doi.org/10.3390/w12092455 http://doi.org/10.4314/acsj.v25i4.1 https://doi.org/10.5194/hess-11-1633-2007 https://doi.org/10.5194/hess-11-1633-2007 https://doi.org/10.1098/rspa.1948.0037 https://doi.org/10.1098/rspa.1948.0037 https://www.researchgate.net/publication/317167487 https://www.researchgate.net/publication/317167487 https://acsess.onlinelibrary.wiley.com/doi/10.2134/agronj1987.00021962007900010022x https://acsess.onlinelibrary.wiley.com/doi/10.2134/agronj1987.00021962007900010022x Musyimi, P.K. et al. Hungarian Geographical Bulletin 71 (2022) (4) 365–382.382 Shilenje, Z.W., Murage, P. and Ongoma, V. 2015. Estimation of potential evaporation based on Penman equation under varying climate, for Murang’a County, Kenya. Pakistan Journal of Meteorology 12. (23): 33–42. Available at http:// repository.usp.ac.fj/id/eprint/11497 Stegman, E.C. 1982. Corn grain yield as influenced by timing of evapotranspiration deficits. Irrigation Science 3. 75–87. Available at https://doi.org/10.1007/ BF00264851 Tyagi, N.K., Sharma, D.K. and Luthra, S.K. 2003. Determination of evapotranspiration for maize and berseem clover. Irrigation Science 21. (4): 173–181. Available at https://doi.org/10.1007/s00271-002- 0061-3 Venton, C.C. 2018. Economics of Resilience to Drought Kenya Analysis. USAID Center for Resilience. Nairobi, USAID Kenya. Available at https://www. usaid.gov/sites/default/files/documents/1867/ Kenya_Economics_of_Res i l i ence_F ina l_ Jan_4_2018_-_BRANDED.pdf Wambugu, P.W., Mathenge, P.W., Auma, E.O. and van Rheenen, H.A. 2012. Constraints to on-farm maize (Zea mays L.) seed production in Western Kenya: Plant growth and yield. International Scholarly Research Notices Article ID 153412. Available at https://doi.org/10.5402/2012/153412 Yao, A.Y.M. 1974. Agricultural potential estimated from the ratio of actual to potential evapotranspira- tion. Agricultural Meteorology 13. 405–417. Available at https://doi.org/10.1016/0002-1571 (74)90081-8 Yang , Z., Qiang , Z. and Xiaocui , H. 2016. Evapotranspiration trend and its relationship with precipitation over Loess Plateau during the last three decades. Advances in Meteorology 2016 Article ID 6809749, 10. Available at http://doi. org/10.1155/2016/6809749 Zotarelli, L., Dukes, M.D., Romero, C.C., Migliaccio, K.W. and Morgan, K.T. 2010. Step by Step Calculation of the Penman-Monteith Evapotranspiration (FAO-56 method). Institute of Food and Agricultural Sciences. Gainesville, University of Florida, AE 459. Available at https://edis.ifas.ufl.edu/pdf/AE/AE45900.pdf http://repository.usp.ac.fj/id/eprint/11497 http://repository.usp.ac.fj/id/eprint/11497 https://doi.org/10.1007/BF00264851 https://doi.org/10.1007/BF00264851 https://doi.org/10.1007/s00271-002-0061-3 https://doi.org/10.1007/s00271-002-0061-3 https://www.usaid.gov/sites/default/files/documents/1867/Kenya_Economics_of_Resilience_Final_Jan_4_2018_-_BRANDED.pdf https://www.usaid.gov/sites/default/files/documents/1867/Kenya_Economics_of_Resilience_Final_Jan_4_2018_-_BRANDED.pdf https://www.usaid.gov/sites/default/files/documents/1867/Kenya_Economics_of_Resilience_Final_Jan_4_2018_-_BRANDED.pdf https://www.usaid.gov/sites/default/files/documents/1867/Kenya_Economics_of_Resilience_Final_Jan_4_2018_-_BRANDED.pdf https://doi.org/10.5402/2012/153412 https://www.sciencedirect.com/science/article/abs/pii/0002157174900818?via%3Dihub https://www.hindawi.com/journals/amete/2016/6809749/ https://www.hindawi.com/journals/amete/2016/6809749/