Estimation of evapotranspiration through surface energy balance model using satellite data

Abstract

Water is a precious natural resource and that required effective and sustainable management. Agriculture sector is the largest consumer of water in India and has a substantial contribution to the GDP of India. At the same time in agriculture sector, demand of fresh water is increasing and its estimation involves computation of many water balance factors and evapotranspiration is one of them. Evapotranspiration is combined loss of water from soil as well as plants and it is also a crucial component of Hydrologic Cycle. But at the same time ET depends upon several factors and it is very difficult to calculate it precisely. Estimation of Evapotranspiration is important for determining the agro-climatic potential of a particular region, water requirement of field crops, irrigation scheduling and suitability of crops or varieties, which can be grown successfully with the best economic returns. Actual evapotranspiration represents over 60% of precipitation water loses. Direct measurements of ET by Lysimeters are difficult, expensive and at same time they provide ET results for small and homogeneous areas (point data). The present study aims to estimate evapotranspiration spatially by Surface Energy Balance model using satellite data and to generate spatial ET maps after validation. Surface Energy Balance Algorithm for Land (SEBAL) model is one of the best model to estimate spatial ET because it calculates the fluxes independently from land cover and, can handle thermal infrared images at resolutions between a few meters to several kilometres. The empirical relationships of SEBAL model are adjusted to different geographical regions and time of image acquisition. Due to these major bottlenecks of this surface energy balance model, this method was adopted for the spatial Evapotranspiration estimation of the study area (Hisar and Bhiwani districts, Haryana, India). The Estimated ET was also validated by four methods; by comparing (i) actual observed ET with Lysimeter data (ii) actual crop ET estimated by Penman Montieth method (iii) actual wheat crop ET with ground truthing and (iv) crop growth biophysical parameters viz Surface temperature, Albedo and Normalised Difference Vegetation Index. Long term LANDSAT ETM+ satellite data was used for the study starting from 2004 to 2014 which include five wheat crop seasons. In each crop season ET was estimated for 3 cloud free images taken at initial, mid and end growth stages. Weather Data was collected for the same years from Indian Meteorological Department (IMD), Chandigarh. Ground Truthing was carried out by measuring coordinates by Global Positioning System (GPS) at both the districts. Estimated Evapotranspiration through SEBAL model compared with Lysimeter experimental results, model fairly predicted the high degree of correlation with experimental values, the coefficient of determination found around 0.91. SEBAL Model also gave fair prediction when compared to Penman monteith, FAO standard method and the coefficient of determination was found 0.681. Validation of SEBAL ET by ground truthing fields at Bhiwani and Hisar district, the coefficient of determination was found 0.850 and 0.835 respectively. Biophysical Parameters NDVI (Normalized Difference Vegetation Index), Albedo, Surface Temperature estimated through SEBAL were also analysed and found according to growth stages of wheat crop. It is found that SEBAL model has ability to estimate spatial ET for composite terrain and this model can be used in policy making for water management in Haryana State. Based on the outcome data, spatial ET maps were generated for study area (Hisar and Bhiwani districts) which will be beneficial for irrigation scheduling and crop water management. newline