Estimation of crop water requirements within Wainganga sub-basin for Kharif and Rabi season using spatial analysis

In India, largest user of the water is agriculture sector is the, so it is important to do the proper management of available water. Aim of this study to estimate crop water requirements (CWR) for Kharif and Rabi seasons within Wainganga sub-basin by using remote sensing and GIS technique. For this, reference evapotranspiration (ET0) is estimated by using the Food and Agricultural Organization (FAO) Penman-Monteith method. Weather data from 8 weather stations has been collected to estimate ET0. The crop coefficient (KC) is estimated by using linear relationship with Normalized Difference Vegetation Index (NDVI). The MODIS NDVI dataset is used for calculation of crop coefficient. The effective precipitation (Pe) has been calculated to estimate CWR by using FAO recommended empirical method. The spatial variation maps for ET0, KC, actual evapotranspiration (ETa) and CWR are generated using Inverse Distance Weightage (IDW) interpolation technique in ArcGIS software. The results show that ET0 and ETa are higher in Kharif season than the Rabi season. The KC is also found higher in kharif than in rabi. Since the Pe is negligible in Rabi season, CWR in Rabi season is found to be higher than kharif season. The CWR in Rabi varies between 320 mm to 378 mm and in kharif season it varies between 94 mm to 263 mm. It is concluded that the seasonal estimation of CWR helps in understanding the peak water demand in that season in better way. It is required to provide sufficient irrigation to the crops in Rabi season especially as the agricultural production completely depends on the irrigation facilities in the study area because there is no rainfall in this season.


Introduction
Water is the most important component for the agriculture production. There is increase in demand of freshwater due to rising population, urbanization and industrialization. Due to global warming, temperature increases rapidly and hence, demand of freshwater increases as well. Due to the rise in temperature and variation in climatic factors, crop production and crop water requirements are significantly affected [1]. As the effective precipitation is the key factor for the crop water requirement estimation, it is important to analyse the effects of climate change on precipitation. There is need to get thorough knowledge about the climate change effect on reference evapotranspiration, actual evapotranspiration and effective precipitation to monitor and control the future agricultural policy, research and development [2].
In recent studies, researchers have estimated reference crop evapotranspiration by using many methods such as crop models based on climatic parameters or surface observations, remote sensing  [3] used residual of energy balance approach for estimation of ET a and K C in California with the aim to better manage limited water resources by gaining more accurate knowledge regarding evapotranspiration of crops.
Many researchers give the different models and methods for the crop water requirements. Li et al. 2020 [4] estimated irrigation water requirements based on GIS techniques and Penman-Monteith formula for three main crops (wheat, cotton and corn) during their growing periods in Xinjiang province, China. It was concluded that planting area of crops were generally more sensitive to irrigation water requirements than rising temperature. Wang et al. 2020 [1] used CROPWAT 8.0 to estimate crop water requirements in Heilongjiang Province, China. Mann-Kendall trend test was used to analyse the changing trend of ET 0 , ET a and P e . Usman et al. 2021 [5] used surface energy balance system model to estimate Ea and FAO-56 Penman-Monteith method to estimate ET 0 and comparison between them.
The main objective of the study is to calculate the CWR within Wainganga river sub-basin, Nagpur, Maharashtra. For this weather data from 8 weather stations has been obtained and analysed. The ET 0 was calculated by using FAO Penman-Monteith method and K C was estimated by using MODIS NDVI data. P e was calculated by using FAO based empirical method. IDW method is used for interpolation of weather data in GIS environment. The spatial variation of K C , ET 0 and CWR for kharif and Rabi season was also estimated.
The objectives of the present study are as follow: 2) To calculate and analyse spatial variation of ET 0 , K C and ET a in the study area.
3) To calculate effective precipitation within study area. 4) To analyse the spatial variation of crop water requirement in study area due to climate change for Kharif and Rabi seasons.

Study Area
Wainganga River sub-basin is located in Nagpur district, Maharashtra, India. It lies between 21º10'0"N to 21º40'0"N latitude and 78º30'0" E to 79º20'0" E longitude ( Figure 1). The Wainganga River subbasin has elevation range from 304 m to 467 m above mean sea level. The total area is 1677.2911 Sq. Km. Wainganga River basin within Nagpur District has 40 sub-watersheds [6], out of them 8 subwatersheds are selected for study. The geographical location map of the study area is shown in Figure  1.

Meteorological Data
Meteorological data is obtained from CFSR (Climate Forecast System Reanalysis) which is coupled atmosphere, ocean and land system developed at NOAA-NCEP. The spatial resolution of CFSR data is 0.35 0 . CFSR provides the maximum and minimum temperatures (°C), precipitation (mm), wind velocity (m/s), humidity (%) and solar radiation (MJ/m 2 ) values from 1979 [7]. Meteorological data obtained from CFSR is useful to estimate ET 0 by FAO Penman Monteith method.

MODIS NDVI Data
MODIS NDVI data was downloaded and analysed. There are two platforms from where datasets are collected viz. Aqua and Terra. MOD13Q1 data has spatial resolution of 250 m and temporal resolution of 16 days. The NDVI is a vegetation index, which gives the information about the actual growth condition of the vegetation. NDVI value varies between -1 to +1. During a year, minimum NDVI occurs in May and maximum NDVI occurs in September.

Methodology
The main objective of study is to estimate the CWR for the part of Wainganga sub-basin which is situated within Nagpur for kharif and rabi seasons. For this, precipitation, wind speed, relative humidity, solar radiations and temperature were collected. ET 0 was estimated by using Penman-Monteith method [8,9]. The Penman-Monteith method is widely used all over the world for estimating ET 0 . For the estimation of K C , MODIS NDVI data is used. The ET a was calculated by using K c and ET 0 . The P e was calculated by using empirical formula which is suggested by FAO using total monthly precipitation. Finally, the CWR is estimated by using Et a and P e .
The overall methodology adopted for the calculation of CWR is given in flowchart ( Figure 2) as follows: Overall methodology adopted for calculation of crop water requirements

FAO Penman-Monteith method for Calculation of ET 0
The FAO Penman-Monteith method is now mostly recommended method for estimation of ET 0 which is applicable in almost all regions and climatic conditions. The Penman-Monteith method uses precipitation, daily mean temperature, relative humidity, solar radiations and wind speed. Zotarelli, et.al (2010) [10] provides steps for calculation of ET 0 . The equation for the estimation of ET 0 is given as Eq. 1.

Estimation of crop coefficient (K C ) by using satellite-based vegetation index
The crop coefficient is determined by using NDVI. The different models were developed to estimate crop coefficient directly with the help of NDVI. Kamble et al. (2013) [11] developed simple linear relation model for crop coefficient estimation. The linear relation between K C and NDVI is given as Eq. 2.

Calculation of Crop water Requirements (CWR)
The crop water requirement is the actual water required to that crop during its growth period. The crop water requirement estimation formula [4] is given as follows, CWR=ET a -P e Eq. 3 Where, ET a -Actual evapotranspiration in mm/day; P e -Effective precipitation in mm.
The ET a in mm for 3 months was calculated and also effective precipitation for 3 months for each season was calculated. The obtained CWR values were interpolated by IDW tool in ArcGIS software. The spatial variation maps of CWR for kharif and Rabi season were generated.

Results and Discussion
The estimated results of ET 0 obtained from the FAO Penman-Monteith method are analysed for kharif and rabi seasons for years 2001, 2005, 2010 and 2013. The K C estimated by using MODIS NDVI data, ET a by using ET 0 and CWR using ET a and P e within study area for kharif and rabi seasons are compared and analysed.
Statistical summary of the ET 0 (mm/day), K C , ET a ((mm/day)) and CWR (mm) over study area for kharif and rabi seasons are summarized in Table 1. The statistical values of CWR in Table 1 indicate that the CWR for rabi season is much more as compared to the kharif season. This is because of the effective precipitation is zero in rabi season as there is no precipitation and more effective precipitation during kharif season due to monsoon rainfall. This affects the requirement of the crop water in the different seasons in the study area.

Spatial variation of ET 0 for Kharif and Rabi season
The accurate estimation of ET 0 is important for ET a estimation. These weather data  The rate of ET 0 is slightly lower at western part of the study area. As the available moisture content in atmosphere is less in rabi season than kharif season, ET 0 rate is lower in Rabi season.

Spatial variation of crop coefficient (K C ) for Kharif and Rabi season
The K C depends on the growth stages of the crop, canopy of crop and area under crop. The K C is higher at peak growth stage and then decreases when the crop attends maturity. The MODIS data was processed in ArcGIS and K C was calculated by using Raster Calculator tool in ArcGIS. Spatial variation maps of K C are shown in Figure 4. Results of spatial variation of K C show that the K C value varies from -0.047 to +1.10 in kharif season.
The negative values are indicating the presence of water body. The spatial variation maps of K C indicate that in Rabi season K C value ranges from -0.30 to +0.86. The result shows that the K C is higher in Kharif season than Rabi season. The reason behind this may be the initial growth stage of the crop. The ET a is affected by the K C value. K C is the simply ratio of ET a to ET 0 .

Spatial variation of ET a for Kharif and Rabi season
The ET a is the loss of water due to evaporation and transpiration from actual crop field. Spatial variation maps of ET a are shown in Figure 5. . CWR for given study area was estimated with the help of actual evapotranspiration rate for 3 months in mm and effective precipitation for equivalent 3 months in mm. The spatial variation maps for CWR for kharif and Rabi season are given in Figure 6. The results show that the maximum CWR varies from 320 mm to 378 mm for Rabi season and it varies from 94 mm to 263 mm in Kharif season. The results show that water requirement for crops is maximum in rabi season as compared to Kharif season. The precipitation is the important affecting factor. During the Rabi season the precipitation is very less or negligible; so effective precipitation is found to be zero and hence, CWR is maximum in rabi season than kharif season. Also in Kharif season, the precipitation is quite sufficient to fulfil the water requirements of the crops.The results also show that the CWR in Rabi season decreases gradually from year 2001 to year 2013. This is because of the changing climate and other anthropogenic activities related to the agriculture.

Conclusion
The main objective is to estimation of crop water requirements within Wainganga sub-basin for kharif and rabi season. And the spatial variation maps for ET 0 , ET a and CWR are also prepared and analysed. The results show that ET 0 is lower in Rabi season than because of the variation in soil moisture content in Kharif and Rabi season.
The spatial variation maps of K C show that value of Kc is higher in Rabi season than Kharif season. The K C directly affects the actual evapotranspiration rate. The K C was estimated by using MODIS NDVI data. As NDVI directly related to the crop coefficient, it represents the growth stages of crop and crop type. The ET a map shows that ET a is higher in kharif season than Rabi season as ET 0 is maximum in kharif season because of higher temperature.
The spatial variation maps of CWR shows that the CWR in Rabi season is higher than Kharif season. The maximum CWR found in rabi season which range from 320 mm to 378 mm and in kharif season, CWR is range from 94 mm to 263 mm. The rainfall reduces the amount of irrigation water requirement as water requirements is fulfilled by rainfall mostly. It is concluded that the seasonal estimation of CWR helps in better understanding the peak crop water demand. It is required to select sufficient irrigation discharge to the crop field in rabi season specially as the agricultural production is completely depends on the irrigation water because of no rainfall in this season. Hence, the remote sensing and GIS techniques proved to be most reliable technique as all the data required for CWR estimation is easily available and it can be used to generate spatial variation maps of all the parameters.