Estimation of Regional-Scale Near Real Time Reference Evapotranspiration Using Remote Sensing and Weather Data to Improve Agriculture Advisory

Abstract

Abstract Accurate and timely information of evapotranspiration (ET0) is essential for multiple agricultural applications, including irrigation scheduling, studying crop-specific water loss at different growth stages, predicting crop yields, and forecasting drought conditions. The aim of this study is to examine the spatiotemporal patterns of ET0 and facilitate the monitoring of crop water demand, optimizing irrigation water usage, and enhancing agricultural advisory services. This paper estimates regional-level daily ET0 gridded data with a spatial resolution of 12.5 km by integrating observed weather data, IMD GFS-T1534 reanalysis data, and INSAT-3D satellite-based insolation product using the standard FAO56 Penman-Monteith empirical equation. The estimated monthly mean of ET0 across India ranged from 10 to 400 mm. ET0 exhibited an increasing trend from January to May, reaching its maximum values in May. In June, ET0 significantly decreased as the monsoon arrived, coinciding with the movement of rainfall patterns. The month of December exhibited the lowest ET0 values. The estimated daily gridded ET0 was compared with station-based ET0, resulting in daily correlation coefficient R2 and daily maximum absolute percentage errors ranging from 0.34 to 0.90 and 10% to 27% respectively. However, these errors decreased to a large extent when considering multiday accumulated values. A comparison was conducted between the GLDAS model ET0 and the station-estimated values, revealing an overestimation of ET0 by the GLDAS model. Additionally, significant variations were observed among the meteorological subdivisions. This highlights the necessity for proper calibration of the GLDAS model ET0 or its effective agricultural application.