Effect of Changing Climate on Water Requirement of Chickpea in North Interior Karnataka: Cropwat Model based Assessment

Prediction of the crop water requirement is of vital importance in water resource management. Crop water requirements are normally expressed by the rate of evapotranspiration (ET) in mm day^-1. The level of ET has been shown to be related to the evaporative demand of the air. The evaporative demand can be expressed as the reference evapotranspiration (ET_o) which when calculated predicts the effect of the climate on the level of the crop evapotranspiration. The effect of the weather variable on the water demand of the crop can be analysed by the CROPWAT model which was developed by the FAO-Land and Water Development Division (Allen et al., 1998).  Agrometeorological variables are one of the key inputs required for the operation of the crop simulation models. These include the maximum and minimum temperature, air humidity, wind speed, sunshine hours, solar radiation and total rainfall. Out of these variables temperature and rainfall have direct and maximum impact on crop production.
       
North Interior Karnataka (NIK) also locally known as ‘Uttara Karnataka’ is a geographical region consisting of mostly semi-arid plateau from 300 to 730 metres (980 to 2,400 ft) elevation that constitutes the northern part of the South Indian state- Karnataka. It constitutes 12 districts namely Bagalakote, Ballari, Belagavi, Bidar, Dharwad, Gadag, Haveri, Kalaburgi, Koppal, Raichur, Vijayapura and Yadagiri. This region is largely covered with rich black cotton and red sandy loamy soils, gently sloping lands and plains, summits of plateau and table lands. NIK is one of the drier regions of India receiving on an average just 731 mm rainfall per annum (Anonymous, 2016). Chickpea is one of the major pulse crop grown under rainfed conditions in NIK region. Some of the farmers go for one or two irrigation at flowering and pod filling stages if necessary.
       
The seasonal analysis was done for all the 12 districts of NIK (Fig 1) for chickpea using district level historical weather data of past 30 years (1991-2020) as well as projected weather data of 30 years (2021-2050) to bring about crop evapotranspiration (ET_c), effective rainfall (ER) and irrigation requirements (IR) across different dates of sowing (DOS) i.e., four dates of sowing starting from 01st October to 15th November at quarterly interval on black clay soil. As above mentioned parameters varied distinctly under decadal scenario for each district and the best date of sowing as an adaptation strategy was identified for chickpea under past and projected weather conditions. These results are represented for each district separately.
 

The immediate past weather data (rainfall, minimum and maximum temperature) for 12 district of NIK (Fig 1) was collected from NASA POWER web portal (https://power.larc.nasa.gov) (Sparks, A. H, 2018) for the period of 30 years from 1991 to 2020 and the projected data for the period of coming 30 years from 2021-2050 was collected from Copernicus Climate Change Service (IPSL-CM5A model) (https://climate.copernicus.eu). According to RCP 6.0 scenario, there would be an increase of 97.4 mm rainfall and 0.1  C temperature while number of rainy days decrease by 12 during the chickpea cropping period (Oct-Feb) under the projected climates (2021-2050) compared to past climates (1991-2020) (Table 1).
 

       
The field experiment was conducted at University of agricultural sciences, Dharwad in rabi seasons of 2019-20 and 2020-21. The phenological data for initial, mid and late growth stages of chickpea variety BGD-103 collected from the field experiment were used in the model. The salient details of chickpea crop required for the study i.e., crop coefficients (Kc), phenological days, critical depletion fraction (p), yield response factor (Ky) were also taken from the available 18 published data of FAO (Allen et al., 1998). The soil data on total available soil moisture content (SMC), initial soil moisture depletion, maximum rooting depth and maximum rain infiltration rate for black clay and red sandy loam soils for all the 12 districts of NIK were collected from the world bank sponsored Sujala Project at UAS, Dharwad (Table 2). The CROPWAT 8.0 model suited for windows was used for the simulation of crop and irrigation water requirements based on soil, climate and crop data for the study. It is a computer program developed by land and development division of FAO. The spatial interpretation of the parameters for all the 12 districts of NIK was done using ArcGIS software.
 

Correlation analysis
 
Higher positive correlation at 1% level of significance between ET_o and ET_c was observed for all the 12 districts of NIK. Similarly higher positive correlation at 1% level of significance was observed between ET_c-ER and IR for all the districts of NIK. The relation between ER and IR was non-significant for Ballari district while it was negatively correlated at 5% level of significance for Bidar (0.9), Kalaburagi (0.89) and Raichur (0.878) districts. The remaining districts have shown negative correlation between ER and IR at 1% level of significance Table 3.
 

 
Reference evapotranspiration (ET_o) in the past and projected climates
 
The highest ET_o was recorded for Vijayapur district (1806.8 mm year^-1) and the lowest was for Ballari (1598.7 mm year^-1) under the past climate (1991-2020). Under projected climate (2021-2050) the highest ET_o was recorded for Kalaburagi district (1648.6 mm year^-1) and the lowest was for Koppal district (1526.9 mm year^-1). The ET_o under the projected climates was simulated to decrease by 50-200 mm year^-1 across districts except for Ballari which showed decrease of only 4.9 mm year^-1. The highest reduction in ET_o under projected climate compared to past climate was observed in Koppal district (210 mm year^-1) (Table 4).
 

       
The ET_o for the past climate (1991-2020) showed the highest value (6.4 mm day^-1) in the month of April and the lowest value (3.8 mm day^-1) in the month of November for NIK.  The highest (6.1 mm day^-1) ET_o for projected climate (2021-2050) was for May and the lowest (3 mm day^-1) was for December. The average ET_o under past climate for NIK was 1715 mm year^-1 and under projected climate it was 1587.8 mm year^-1. The ET_o simulated decreased by 127.2 mm year^-1 under the projected climate compared to the past (Table 4).
       
The higher ET_o during March to May can be explained by the rising temperature in that particular period. Thus, the air temperature has a direct effect on ET_o. Relative humidity is a function of air temperature. Higher the temperature more is the amount of water vapour that can be held by the atmosphere. The extent of evaporation and transpiration depend on the amount of moisture present in the atmosphere. Ali et al., (2009) concluded that the ET_o estimates are most sensitive to maximum temperature and least sensitive to minimum temperature. The order of sensitivity noticed was; maximum temperature > relative humidity > sunshine duration > wind speed > minimum temperature. The decline in ET_o under projected climate for all districts except Ballari was mainly due to increased amount of rainfall and decreased sunshine hours in the projected climates.
       
There was no significant change in the ET_o during the projected climates showing decreased ET_o in all the districts of NIK compared to past (Table 5) for Chickpea cropping period (October- February). Vijayapur district (1 mm day^-1) has shown highest decrease in ET­_o, while lowest was for Ballari (0.6 mm day^-1). This was because of their respective highest and lowest ET_o respectively among the 12 districts of NIK in past climates.
 

 
Crop evapotranspiration (ET_c) in the past and projected climates
 
The model simulated the highest average ET_c (303.9 mm) in Vijayapur district for the past climate (1991-2020) across four DOS followed by Bagalakote (301.2 mm) (Table 6). This was because of the highest average temperature recorded and average ET_o simulated in the cropping season. The lowest average ET_c of 271.1 mm was simulated for Ballari district followed by Haveri (282.1 mm). This can be explained by the lowest average ET_o during the cropping period in Ballari across districts. The average ET_c simulated for chickpea by Desta et al., (2015) in Ethiopia for the period 1973-2007 was 366.6 mm. In the projected climates (2021-2050) the highest ET_c was simulated for Bidar district (230.5 mm) followed by Kalaburagi (230.2 mm) and the lowest was for Belagavi district (221.1 mm) followed by Koppal (221.7 mm). The highest average temperature and average ET_o (13.3 mm day^-1) during cropping period (Oct-Feb) in Bidar district and the lowest average temperature and average ET_o (12.8 mm day^-1) during cropping period in Belagavi district among the 12 districts in the projected climates of NIK are the influential parameters. All the districts showed decreased ET_c for the projected climates compared to the past due to decreased ET_o­ and increased rainfall during the cropping period (Table 6 and Fig 2). The highest decrease in ET_o in the projected climates was simulated for Vijayapur district (76.6 mm day^-1) as it showed the highest decrease in ET_o among 12 districts of NIK whereas the lowest was for Ballari (46.4 mm day^-1). Gilanipour and Gholizadeh (2016) estimated that in the predicted climatic period of 2016-2045, the rice water requirement and irrigation water requirement decreased by more than 9.9%. Further, the rise in rainfall during rice growth period may be the main reason for the decline in crop water requirement, while the significant decrease in irrigation water requirement can be attributed to combined action of rising precipitation and a slight increase in temperature.
 

 

 
Simulated outputs across dates of sowing suggest that ET_c increased with delay in sowing in all the 12 districts of NIK under past climates (Table 7 and Fig 3). This is due to increase in average ET_o and decreased rainfall with delay in sowing i.e., October sowing received more rainfall than November sowing. In case of projected climates there was no significant change in ET_c with delay in sowing (Table 8 and Fig 3). However, the highest ET_c was for Nov-15 sowing because of minimum rainfall during second fortnight of November.
 

 

 

 
 
Effective rainfall (ER)
 
The highest ER simulated in the cropping season was for Haveri district (72.6 mm) during the cropping period followed by Ballari (67.1 mm) and lowest was for Vijayapur district (47.4 mm) followed by Bidar (49 mm) in the past climate (Table 6 and Fig 4). Haveri district recorded rainfall (RF) of 175.8 mm during cropping period (Oct-Feb) with 24 rainy days (RD) in past climates, the highest among all the 12 districts in the past climates while, for Vijayapur district it was 112.1 mm of rainfall during cropping period with 16 rainy days, the lowest rainfall and rainy days among the 12 districts of NIK in the past climates.
       

 
For the projected climates all the districts showed increased ER compared to past with the highest ER for Belagavi and Dharwad districts (116.3 mm), whereas the lowest was for Kalaburgi district (75.2 mm) (Table 6 and Fig 4). Belagavi and Dharwad districts recorded rainfall of 281.1 mm during cropping period (Oct-Feb) with 39 RD, the highest among all the 12 districts in the projected climates. Bidar district recorded 159.7 mm rainfall during cropping period with 20 RD (Table 1), the lowest RF and RD among the 12 districts of NIK. The highest increase in ER in the projected climates compared to past was for Belagavi district (61 mm) and lowest was for Kalaburgi (24.3 mm). The increase in ER is proportional to their respective increase in rainfall in projected climates. Kyu and An (2019) from Vietnam found that the ER for the summer-autumn rice crop significantly increased by 6.2, 16.9 and 15.4 per cent, respectively in 2020, 2055 and 2090 (RCP 4.5 scenario) compared to baseline period (2002-2017).
       
The CROPWAT simulated decline in ER with delay in sowing in both past and projected climates and in all the districts of NIK (Table 7 and 8) (Fig 2). This can be explained by higher North-East monsoon rainfall in the early months (October) and rainfall dissipation of rainfall towards December month.
 
Irrigation requirement (IR)
 
Among all the 12 districts of NIK, Vijayapur district (263 mm) simulated the highest IR for chickpea followed by Bagalakote (255.6 mm) in the past climates because of higher ET_c and lowest ER during the cropping period. The Lowest IR was simulated for Ballari district (216.8 mm) followed by Haveri (228.9 mm) (Table 6 and Fig 5). This can be explained by higher ER and lower ET_c during the cropping period. In case of projected climates all the districts showed to require lower IR compared to past climates. Bidar district (168.7 mm) followed by Kalaburagi (167.6 mm) showed to require the highest IR for chickpea and the lowest for Belagavi district (141.5 mm) followed by Dharwad (142.5 mm). Kalaburagi and Bidar districts recorded the highest ET_c, lowest ER and increased average temperatures in the projected climates. The highest decline in IR for projected climates compared to the past was simulated for Belagavi district (108.1 mm) and the lowest was for Ballari (68.1 mm). The decrease in IR was simulated directly proportional respective decrease in ET_c and inversely to increased ER.
 

       
The IR increased with delay in sowing in both past and projected climates in all the districts of NIK because of decreased ER and increased ETc during the cropping period (Table 7 and 8) (Fig 2). Early sowing (October) receives more rainfall due to North-East monsoon onset which dissipates towards December. Desta et al., (2015) observed that IR increased in the range of 134-372 mm with delay in sowing (01-July to 30-Aug, quarterly interval) for the period 1973-2007 in DebreZeit, Ethiopia.

The study for Northern Interior Karnataka revealed that during chickpea cropping period (October- February) increased rainfall under projected climate (2021-2050) compared to past climate decreased crop evapotranspiration and irrigation requirement. Sowing chickpea early i.e., on 01^st October under projected climate simulated the lowest water requirement and irrigation requirement for all the 12 districts of NIK. In this context, however, further research needs to be taken up on adaptability of pulses to the climate variability expected under the future climates for their, performance, sustenance and improved productivity. 

None.