Assessment of spatiotemporal variability and trend analysis of reference crop evapotranspiration for the southern region of Peninsular India

Accurate estimation of reference evapotranspiration (ET0) is an essential requirement for water resource management and scheduling agricultural activities. Several empirical methods have been employed in estimating ET0 across diverse climate regimes over the past decades. In this study, the Python implementation for estimation of daily and monthly ET0 values of representative stations of ten agro-climatic zones of Karnataka from 1979 to 2014 using the standard FAO Penman-Monteith method was carried out. The assessment of temporal and spatial variability of monthly ET0 values across the various agro-climatic zones done by the various statistical measures revealed that the variation in spatial ET0 values was higher than temporal variation, indicating major difference in ET0 values was with respect to the stations rather than years under study. The nonparametric Mann-Kendall test conducted at 1% significance level on the annual ET0 values revealed a statistically significant increasing trend for all the ten stations during the study period. The trend test conducted on the climate variables like mean air temperature, wind speed, relative humidity, and solar radiation signifies their influence on the annual ET0 values. The magnitude changes in the trends detected by the Theil Sen’s slope indicated that increasing values of mean temperature, solar radiation, and decreasing values of relative humidity predominantly contributed to the annual upward trend in ET0 values for the 10 stations. A trivial impact of wind speed on annual ET0 values was observed for the stations. Kalburgi and Udupi stations exhibited a positive ET0 trend with the highest and lowest annual values among ten stations.


Introduction
Reference crop evapotranspiration (ET 0 ) is a significant parameter in the estimation of crop water requirement and irrigation scheduling. The crop water need or crop water requirement (ET crop) is defined as the amount of water consumed to meet the water loss through evapotranspiration. Reference crop evapotranspiration is defined as "the rate of evapotranspiration from a hypothetical reference crop with an assumed crop height of 0.12 m (4.72 in), a fixed surface resistance of 70 sec m −1 (70 sec 3.2ft −1 ) and an albedo of 0.23, closely resembling the evapotranspiration from an extensive surface of green grass of uniform height, actively growing, well-watered and completely shading the ground" (Allen et al. 1998). ET 0 is the rate at which water vapor is evaporated from a hypothetical green grass with an abundant soil moisture supply. The irrigation scheduling based on evapotranspiration is of major concern for sustainable agricultural development across the world. Hence, it is necessary to quantify accurate reference evapotranspiration values by considering different agro-climatic conditions for the best irrigation planning and management practices.
The ET 0 computation requires both meteorological and nonmeteorological data of climate stations. The direct method of quantifying the ET 0 using the lysimetric approach is expensive, time-consuming, and lacks precision. Among the several existing empirical methods, the Food and Agriculture Organization (FAO) adopted the Penman-Monteith (P-M) equation as the standard one, which can work in a wide variety of meteorological conditions across the world.
The FAO P-M method was found to be the most preferred and desirable indirect approach over the other empirical methods for accurate estimation of ET 0 across a wide range of climate conditions (Allen et al. 1998;Heydari et al. 2014;Lopez-Urrea et al. 2006;Mohan 1991;Nandagiri and Kovoor 2006;Pandey et al. 2016;Poddar et al. 2018;Tabari et al. 2013). The FAO P-M method was used in the estimation of ET 0 with more precision and the values were further used to assess both temporal and spatial variabilities (Łabedzki et al. 2014;Yadav et al. 2017). Several researches have been conducted to demonstrate widespread climatic variations in view of trends of annual ET 0 values and the corresponding climatic variables which influence the ET 0 using used Mann-Kendall and Sen's slope methods (Bandyopadhayay et al. 2009;Jhajharia et al. 2012;Liu and Zhang 2013;). An increasing trend in ET 0 values in the study conducted at different regions with time scales (Azizzadeh and Javan 2015;Djaman et al. 2018;Gocic and Trajkovic 2014;Łabedzki et al. 2014;Liu et al. 2020;Ndiaye et al. 2020;. Researchers have also reported a decreasing trend in ET 0 with respect to climate variables across different parts of the world (Gao et al. 2016;Pandey and Pandey 2013;Xie and Zhu 2013;Yadav et al. 2016). Recently, the influence of ET 0 trends on climate change, irrigation design, and ecosystem has gained importance (Dinpashoh and Babamiri 2020;Jerin et al. 2021;Pour et al. 2020;Tomas-Burguera et al. 2021;Wang et al. 2021). Several studies were carried out on the assessment of ET 0 and detection of trends of meteorological variables in different regions, mostly northern parts of India (Bandyopadhayay et al. 2009;Jhajharia et al. 2012;Jhajharia et al. 2014aJhajharia et al. , 2014bMonia and Jhajharia 2016). However, studies on ET 0 trends and their influence on climate change, irrigation design, and crop yield for the southern regions of India are very limited. Hence, the estimation of ET 0 and the assessment of spatial and temporal variables that ultimately influence the irrigational requirements and crop yield for southern states of India is the need of the hour.
Karnataka is the largest state in the southwestern part of India. It comprises various agro-climatic zones with diverse climate conditions, which supports the raising of varieties of crops including cash crops. The primary occupation of rural Karnataka is agriculture and the farmers cultivate a variety of crops. Agriculture is the most essential sector of Karnataka state's economy and is characterized by both irrigated and drought-prone regions. Hence, a large portion of the cultivable land in Karnataka is rain-fed and the rest use groundwater resources with severe agro-climatic constraints. Karnataka is considered as the study area in the present research work, as water management is one of the most important components for irrigation and ultimately yield prediction in the state. Karnataka state includes ten agro-climatic zones comprising of varied agriculture practices that contribute to the economy of the state.
The objective of the present study is to estimate the daily ET 0 values of ten agro-climatic zones of Karnataka State, India for a duration of 35 years (1979 to 2014) using the FAO-56 P-M method. The temporal and spatial variability of the agro-climatic zones of Karnataka on monthly basis is assessed and the annual trends of ET 0 corresponding to different climate variables were examined using Mann-Kendall and Sen's slope estimator methods.

Methodology
The outline of the research work is briefed in Fig. 1. The input variables like temperature, relative humidity, wind speed, sunshine hours, and solar radiation intensity need to be collected. The sequential protocol of ET 0 estimation and assessment of the temporal and spatial variability of ET 0 is shown in Fig. 2. The description of the study area and the climate data collection are presented in the following sections.

Study area
Karnataka exhibits considerable variation in the distribution of temperature and rainfall with varying topographic boundaries. It has been divided into ten agro-climatic zones based on variations in climatic factors. The ten agro-climatic zones are viz., North Eastern Transition Zone, North Eastern Dry Zone, Northern Dry Zone, Central Dry Zone, Eastern Dry Zone, Southern Dry Zone, Southern Transition Zone, Northern Transition Zone, Hilly Zone, and Coastal Zone (Energy and wetlands research @ ces, iisc, Bangalore, India. 2005). The map of Karnataka, India showing its agro-climatic zones, is given in Fig. 3.
Karnataka observes four climatological seasons: winter (January and February), summer (March to May), a monsoon (June to September), and a postmonsoon period (October to December). Karnataka has good variability in its rainfall pattern. The highest rainfall with an average of 3,638.5 mm (143 in) per annum is noticed in the coastal zone of Karnataka. The variability of rainfall is reported to be 800 mm in the western area to 1600 mm in the eastern area of Karnataka. Therefore, it is necessary to estimate ET 0 by considering different agroclimatic conditions for best irrigation planning and management practices.

Climate data collection
The daily climatic data for each station in different agroclimatic zones of Karnataka for a span of 35 years (1979 -2014) were collected from the website Global Weather Data for SWAT (https://globalweather.tamu.edu/). The climate data comprises daily maximum temperature, minimum temperature, mean relative humidity, wind speed, and solar radiation for periods of 35 years. The ten climate stations representing each agro-climatic zones have been considered in the present study. Even though there are many stations located in the agro-climatic zones, stations that have typical  Table 1 presents the salient features of climate stations of 10 agro-climatic zones of Karnataka.
The daily climate data sets which are tabulated in a CSV file comprise maximum and minimum temperature (°C), mean relative humidity (%), wind speed (m/sec), solar radiation (MJ/m 2 /day), and other site data. The quality and integrity of the data set were performed as per the guidelines provided by (Allen 1996). The number of days having complete climatic data for all stations considered in this study, excluding the outliers is found to be 12928. The daily reference evapotranspiration (ET 0 ) values from January to December between 1979 and 2014 were computed for ten climate stations representing various agro-climatic regions of Karnataka using Python implementation. The developed program for ET 0 estimation was validated with respect to the results given in the FAO-56 document.

Calculation of reference Evapotranspiration (ET 0 )
The reference evapotranspiration values are influenced by the meteorological variables viz., maximum temperature, minimum temperature, wind speed, relative humidity, solar radiation, and sunshine hours of the locations taken for study. The standard FAO P-M method has been used to calculate daily ET 0 of different agro-climatic zones of Karnataka. The general form of the P-M equation is given as follows where, ET 0 is evapotranspiration, R n is the net radiation, G is the soil heat flux, (e s -e a ) is the vapor pressure deficit of the air, γ is the psychometric constant, Δ is the slope of the vapor pressure curve, λ is the latent heat of vaporization, T is the air temperature at 2 m height, u 2 is the wind speed at 2 m height, e s is saturation vapor pressure in kPa, e a is actual vapor pressure in kPa, D is slope vapor pressure curve in kPa°C, g is psychometric constant in kPa°C. ET 0 can be estimated through the FAO P-M method which involves certain calculation procedures given in as Appendix-I in the supplementary file.

Time series trend analysis
The ET 0 values of the agro-climatic zones can be analyzed to detect the presence of significant trends and their magnitudes during the study period. The trends of climate variables of the stations which influence the ET 0 values can also be investigated. The presence of Fig. 3. Karnataka Map showing its agro-climatic zones specific types of trends in the meteorological time series data is best detected by the nonparametric tests. Nonparametric tests are simple and robust in design and are easy to understand. They are distribution-free and hence, fewer assumptions are required for the data, and they can cope up with outliers/missing values in the dataset (Crawford et al. 1983). The most widely accepted nonparametric method is the Mann-Kendall test. The Mann-Kendall test (Mann 1945) detects long-term trends in the hydro-meteorological time series data (Azizzadeh and Javan 2015;Shadmani et al. 2012).

Mann-Kendall (MK) test
The Mann-Kendall is a statistical hypothesis test procedure to determine the presence of trends in the given time series data. It does not estimate the magnitude of the trends. The test procedure is given as follows. Let (x 1 , x 2,… x n ) be a sequence of data samples measured over the time period. As per the test, the null hypothesis (H 0 ) is that the data are independent and identically distributed (meaning there is no trend). The alternate hypothesis (H 1 ) is that there is a monotonic trend over the time period. The Mann-Kendall test statistic S is where, x i and x j are the values in the dataset (1<i<j<n); n is the size of the dataset; sgn () is the sign function which can be computed as A positive value of S indicates that there is an increasing trend and a negative value is an indicator of a decreasing trend. S is a normal distribution with mean value of 0 and a variance of where, t denotes the tie, and ∑ t indicates the summation over all ties.
The exact distribution of S for n≤10 was given by (Mann 1945) and (Kendall 1975). So, the significance of the test can be tested by computing the standard statistic test (Z) and is given by Var(S) is the variance of statistics S and Z is the Mann-Kendall test statistics with standard normal distribution with mean 0 and variance 1. This Z statistic can be used when the number of samples n>10. The trend's significance is assessed by comparing the Z value with the standard normal variation at the pre-specified level of statistical significance (Hamed and Rao 1998). In a two-sided trend test, with α representing the significance level, the null hypothesis should not be accepted if |Z| > Z α/2 , suggesting that the trend is significant. A positive Z value at the significance level implies a positive trend, whereas a negative value indicates a negative trend. The significance of a trend can be verified by the p-value (probability value) obtained from the M-K Z value. If the pvalue is less than the predetermined significance level (e.g., α= 5%) or greater than the confidence level (if α= 5%, confidence level = 95%), the null hypothesis of the trend cannot be accepted.

Theil-Sen's estimator
The Mann-Kendall test doesn't estimate the slope of trends. Sen's slope is the most popular nonparametric method to estimate the magnitude of trends in a sample of n pairs of data (Sen 1968). It can handle missing values, insensitive to outliers, and requires no assumptions on the probability distribution of data. The major steps in this method are: (i) Calculate the slopes for pairs of data points, (ii) Determine the median of all slopes from step(i). The slope of n pairs of data points was estimated using the following formula: where, 1<j<i<n and β is the estimate of the trend magnitude. A positive value of β indicates an upward trend, and a negative value of β indicates a downward trend.

Results and discussions
The monthly average reference evapotranspiration, ET 0 of ten stations representing the agro-climatic zones of Karnataka calculated from 1979 to 2014 was observed to have the higher values in the month of March for stations like Bangalore, Chikmagalur, Hassan, and Madikeri. The daily ET 0 values were estimated to be above 6.0 mm/day for all these four stations mentioned above. However, for climate stations like Belgaum, Bijapur, Mysore, and Udupi, the monthly average ET 0 values were found to be higher in April, ranging from 4.06 mm/day to 7.22 mm/day. Whereas, in stations like Bidar and Kalburgi the monthly average ET 0 values were found to be higher in the month of May and the daily ET 0 values were found to be 7.0 mm/day and above. These higher ET 0 values for above the mentioned climatic stations fall during the summer months characterized by higher temperatures, lower relative humidity values, higher wind speeds, and longer sunshine hours. The minimum daily ET 0 values ranged between 1.59 mm/day and 4.0 mm/day between July and August for all ten climatic stations which may be due to change in the season from summer to monsoon characterized by a decrease in temperature, wind speed, and sunshine hours and higher values of relative humidity. Similar studies carried out by (Tellen 2017) for 15 years reported the highest daily ET 0 value was around 3.16 mm/day in the hottest month of February and the lowest ET 0 value of 2.59 mm/day in the coldest month of August in few stations of Yaoundé, Cameroon during the summer which is due to uniform wind speed observed throughout the year. The mean daily ET 0 value estimated using the FAO P-M method for 6 stations of subhumid subtropical agro-climatic locations of the western Himalayas between 2011 and 2016 was observed to be 3.75 mm/day. It was found that the variations in values of solar radiation, temperature, and relative humidity have an implicit effect on ET 0 values of the regions under study (Poddar et al. 2018).
The temporal and statistical variability of ET 0 values across the various agro-climatic zones was assessed by various statistical measures. The mean, minimum, maximum, standard deviation, coefficient of variation, and median values characterize the spatial and temporal distributed ET 0 values.

Temporal variability of ET 0
The monthly average of ET 0 across ten climate stations under the agro-climatic zones of Karnataka characterizes the multiyear mean reference evapotranspiration that ranged between 85.49 mm in August and 189.38 mm in April as shown in Table 2. This implies that the mean daily ET 0 ranges from 2.8 mm/day to 6.31 mm/day. The coefficient of variation (CV) ranges from 9 to 37% for monthly sums in the multiyear period in the case of temporal variability analysis which is higher than that observed in Madhya Pradesh, India which was reported to be between 6 and 20% (Yadav et al. 2017). (Łabędzki et al. 2011) reported the daily mean ET 0 values between 2 and 4 mm/day for 40 climatic stations for 34 years across Poland. The CV falls within the range of 2 to 15%, which is quite narrow compared to the temporal variations of the regions mentioned in the present study. (Garg et al. 2016) detected quite higher mean daily ET 0 values across 22 stations which varied from 6 to 9 mm/day and the CV values between

Spatial variability of ET 0
The monthly average reference evapotranspiration ET 0 , over 35 years characterize the spatial variability of multiyear mean reference evapotranspiration. The lowest monthly sums of ET 0 were found in Chikmagalur station in July (49.19 mm) and highest in Kalburgi in April (225.12 mm). The coefficient of variation (CV) ranges from 18 to 40%, which is higher than that of temporal variability. Hence the spatial variability of ET 0 is higher than the temporal variability indicating there is a higher variation of ET 0 in the space than in the years. This contrasts with the results obtained by (Łabedzki et al. 2014;Yadav et al. 2017) where the difference in ET 0 values is more in the period of study than among the regions. However, the spatial variability was higher than the temporal variability with the CV ranging from 30 to 57% in Himachal Pradesh, India (Garg et al. 2016) where the climatic variations with respect to hot subhumid tropical tracts of the southern parts to that of the cold, dry, and mountains in the northern/eastern parts contribute to the above values. The average monthly reference evapotranspiration in different climate stations is as depicted in Fig. 5.

Trend analysis of reference Evapotranspiration (ET 0 )
The annual P-M ET 0 values for 10 stations of agro-climatic zones of Karnataka between 1979 and 2014 were analyzed using the Mann-Kendall test and Theil Sen's slope estimator methods. Both the tests were implemented using Python version 3.9.1 in the PyCharm editor, which offers sophisticated functionalities for time series analysis. Trends of both ET 0 and climate variables were determined statistically at a 1% significant level. The outcomes of the test include the normalized test statistic Z, Kendall's Tau statistic which measures the significance of the test, and a 2-sided p-value for testing the hypothesis as presented in Table 3. The positive value of Tau indicates an upward trend and a negative value indicates a downward trend. From the table, it can be noted that all the 10 stations showed an increasing trend in annual ET 0 values during the study period.
The results of the Mann-Kendall test and Sen's slope estimator for the annual ET 0 time series and the corresponding meteorological variables for all 10 stations under study are given in Table 3. Time series analysis of the annual ET 0 trends for ten stations during the period 1979-2014 are presented in Figs. 6, 7, 8, 9, 10, 11, 12, 13, 14 and15.   Table 3, it can be seen that a significant positive trend in annual mean ET 0 was observed in Bangalore though there was a considerable decrease in annual ET 0 values in the middle of the entire study period. The increased values of mean temperature, solar radiation, wind speed, and decreased vales of relative humidity would have contributed to the positive annual ET 0 trend. These results are found to be in agreement with the significant positive ET 0 trend observed in Southern Spain between 1960 and 2005 (Espadafor et al. 2011). The average annual ET 0 value ranged between 1271 mm and 1661 mm for Bangalore as shown in Fig. 6.
An overall positive trend in annual mean ET 0 was observed in Belgaum, mainly due to the influence of increasing mean air temperature and solar radiation. However, the lowest value of annual ET 0 of 1291 mm was observed in the year 1990 which is still on the higher side compared to the average lowest value of ET 0 reported at Udupi station in the 35 years of the study period. The annual ET 0 variation for the station is plotted in Fig. 7.
The annual ET 0 values of Bijapur revealed an upward trend which is mainly because of the fall in relative humidity and rise in mean temperature, solar radiation, and wind speed values. The lowest and the highest average annual ET 0 values observed were 1463 mm and 1777, respectively, as depicted in Fig. 8. However, Chikmagalur, located in the central dry zone, also experiences hot and dry summers indicate an upward trend in annual ET 0 values even though with decreasing trend in wind speed values. The annual ET 0 values ranged from 1186 to 1392 mm. Similar results were reported by (Bandyopadhayay et al. 2009) for the arid climatic conditions of the Deccan Plateau in India which exhibited a positive annual ET 0 trend. The increasing trend in annual ET 0 values is presented in Fig. 9 for the station.
From Table 3, a positive trend in ET 0 values of Hassan is observed with the average annual ET 0 values varying between 1149 mm and 1460 mm. Even though the decrease in the wind speed over the 35 years of the study period was observed, the increase in climatic variables like mean temperature and solar radiation, decreased values of relative humidity have predominantly contributed to the increasing trend in ET 0 values. These results are found to agree with ET 0 trends of the yellow river basin in China (Liu 2004), the western coast of Madagascar (Djaman et al. 2018), and the northwest region of Iran . Figure 10 shows the positive trend in annual ET 0 values at Hassan. The Kalburgi station situated in the northeastern arid zone is mostly hot year-round (annual temperature range: 18 o C-40 o C), with a dry climate in summer and winter. The station witnessed the minimum annual mean ET 0 value of 1522 mm and the maximum value of 1841 mm, which is the highest annual mean ET 0 value observed amid all the 10 stations as presented in Fig. 11. A positive trend in average annual ET 0 was observed during the study period majorly due to the simultaneous occurrence of rising mean air temperature, solar radiation, and decrease in relative humidity values.
The hilly station Madikeri with a tropical highland climate observes fairly moderate weather with constant temperature throughout the year. The positive annual trends of ET 0 were contributed by the increasing temperature and solar radiation values and decrease in relative humidity for the study duration. The mean annual ET 0 ranged between 1248 mm and 1473 mm as depicted in Fig. 12.
It can be seen that from Table 3, Bidar station, which is in a semiarid climate zone indicated a significant upward trend in average annual ET 0 values. It is attributed to the increasing trends in mean temperature, solar radiation, relative humidity values, and decreased trends in wind speed during the study period. The annual ET 0 value ranged from 1436 to 1754 mm. The trend on ET 0 for Bidar is presented in Fig. 13. Similar results were reported in the northeastern part of Iran (Azizzadeh and Javan 2015).
Mysore station characterized by the dry climate in southern Karnataka reveals an upward trend in the mean   Figure 14 shows the annual ET 0 time series. The increased values in meteorological variables like mean temperature, wind speed, and solar radiation, decreased values of relative humidity have shown a significant influence on the positive ET 0 trend. These outcomes are found to be in agreement with the annual ET 0 trends in dry regions of Rajasthan, Gujarat, and western Madhya Pradesh of India (Goroshi et al. 2017).
From Table 3, the Udupi station, characterized by the tropical monsoon climate, located in the coastal zone of the state observed an upward trend in average annual ET 0 during the period taken for study. The minimum mean annual ET 0 recorded was 943 mm, which is the lowest of the mean annual ET 0 among the 10 stations. The maximum mean annual ET 0 was 1106 mm. Even though there was a feeble increase in the relative humidity and decrease in wind speed, the increased solar radiation and mean temperature would have majorly contributed toward the increasing trend in the ET 0 values. The annual ET 0 variation for the station is plotted in Fig. 15.
The Sen's slopes of annual ET 0 of the ten stations during the period 1979 -2014 are represented in Fig. 16. The solid line in Sen's slope plot indicates the least-squares fit for the comparison (regression line) and the regular square dots show the various data points.

Trend analysis of climate variables
The major climatic variables affecting the evapotranspiration process are air temperature, wind speed, relative humidity, and solar radiation. Hence, it is necessary to investigate the impact of these climate variables on the annual ET 0 values of the stations under study. The annual trends in climatic parameters for the study period in 10 stations are investigated by the Mann-Kendall test. The corresponding magnitudes of the slopes detected by Sen's estimator are presented in Table 3. The trends on maximum average temperature, minimum average temperature, wind speed, sunshine hours and relative humidity for ten stations are presented in Figs. 17,18,19,20,21,22,23,24,25 and 26.

Trend of mean air temperature
The increase in air temperature increases the water vapor in the atmosphere which in turn results in a higher rate of evapotranspiration. As per the M-K test results, all the 10 stations showed a significant upward trend in mean annual air temperature during the study period. The heat radiated from the earth, exothermic chemical reactions, increased concentrations of greenhouse gasses generated by various human activities, increased urbanization may be the sources for the incessant increase in mean air temperature of the stations under study (Duhan et al. 2013;Soltani and Soltani 2008;. Significantly increasing trends in the mean annual temperatures for Bidar and Hassan were observed and to be 8775°C to 9933°C (Fig. 17). Hassan recorded the lowest and highest mean annual temperatures of 7374°C and 8356°C, respectively, (Fig. 18). The increasing values of mean temperature influence the annual ET 0 values in Bidar and Hassan stations to a larger extent during the study period. Past researchers (Dinpashoh and Babamiri 2020;Wang et al. 2012aWang et al. , 2012b observed an increasing mean annual temperature trend which was the primary cause for the positive trend in ET 0 values at the western half of Iran, the Yellow River Basin and the Urmia Lake basin, respectively.

Trend of relative humidity
The relative humidity is the ratio between the amount of moisture in the air and the highest probable level of moisture present in the air at a specific temperature. As the relative humidity of the air increases the evapotranspiration rate decreases. The trend test on climate variables revealed that annual relative humidity has an equivalent number of increasing and decreasing trends among the 10 stations. The increase/decrease in relative humidity in the meteorological stations is attributed to the temperature drop/rise of a constant amount of water vapor in the atmosphere. A consistent decreasing trend in mean annual relative humidity was observed which caused a substantial upward trend in annual ET 0 values for Kalburgi (Fig. 19) and Mysore stations (Fig. 20) during the time taken for study. Similar results were reported in The Three Gorges Reservoir project of China and the Upper Yangtze River Basin of China during the study period (Ma et al. 2018;Wang et al. 2021). The diminishing values of relative  (Liu et al. 2020;). An increasing trend in annual relative humidity with a positive annual ET 0 trend was detected in Belgaum (Fig. 21) and Udupi (Fig. 22) which contradicts the results obtained at The Senegal River basin (Ndiaye et al. 2020).

Trend of wind speed
As the wind speed increases, the rate of evapotranspiration also increases. The trend test results showed that about 70% of the stations exhibited a declining trend in wind speed on an annual scale. This may be attributed to the influence of urban development which has an impact on vegetation growth, environmental factors (Xu et al. 2006;Zhang et al. 2009), impacts of human-specific land usages (Bandyopadhayay et al. 2009;Lopes et al. 2011;McVicar and Roderick 2010). Significantly decreasing and an increasing trend in annual wind speed with an upward ET 0 trend were observed in Chikmagalur (Fig. 23) and Bijapur (Fig. 24), respectively, for the period 1979-2014. The wind speed does not seem to have a major influence on annual ET 0 values of the meteorological stations in the present study, which contradicts the results reported by (Gao et al. 2015) wherein wind speed was the key factor toward the variations of ET 0 values of 15 climate stations in The West Liao River for the duration of 51 years. (Liu et al. 2017) found out that wind speed did not have a significant effect on the ET 0 values at the stations of the Yellow River catchment which supports the results obtained by the current work. Past research (Zeng et al. 2019) found that wind speed decreased for most of the 79 stations of southwest China on the entire study duration and had a very minor

Trend of solar radiation
Solar radiation is the main energy source for converting liquid water into water vapor, thus increasing the air temperature. The evapotranspiration increases with the increase in solar radiation. The M-K test revealed that all the 10 stations showed an increasing trend in annual solar radiation during the 35 years of the study period and can be considered as the dominating factor in view of the positive annual ET 0 trend. The increased solar radiation at the ten stations may be due to atmospheric variations such as cloud cover, water vapor, and day length at specific latitudes. However, an upward trend of solar radiation does not always imply the decreasing trend of relative humidity, and vice versa, as observed in the present study. A major increasing trend in annual solar radiation was observed in Bangalore (Fig. 25) and Madikeri (Fig. 26) for the study period. The increasing trends of ET 0 were explained by the significantly increasing trends of sunshine hours during April-September at 18 stations in Poland (Łabedzki et al. 2014). The results obtained by (Ndiaye et al. 2020) showed that the increased solar radiation had a positive influence on annual ET 0 values. The sunshine duration (solar radiation) was the key climate variable in governing the trend of annual ET 0 of the 19 stations in the Guizhou Province of southwest China during 1959-2011 (Gao et al. 2016). On the contrary, the mid-Himalayan region located in Sikkim state in India observed a downward trend in sunshine duration, and hence, leading to a falling trend of ET 0 during Spring (March to May) and monsoon (June to November) season for the study period 1985(Yadav et al. 2016. The major contributors for the reduction of net solar radiation and the consequent reduction in ET 0 may be the various factors such as the increased amount of aerosols present in the atmosphere, greenhouse gasses, and cloud coverage.