Regional clear sky evapotranspiration over agricultural land using remote sensing data from Indian geostationary meteorological satellite

Regular and rapid monitoring of evapotranspiration (ET) at regional scale is required to optimize hydrological resources for irrigation scheduling and water management in agricultural systems. A simplified single-source energy balance scheme was implemented to estimate regional clear sky ET using noon-midnight data acquired from Indian geostationary meteorological satellite (Kalapana-1) sensor (VHRR) (hereafter termed K1VHRR). The core inputs to ET model such as land surface temperature (Ts), surface albedo, insolation and air temperature were retrieved using data of visible (VIS), water vapour (WV) and thermal infrared (IR) bands. These were further used to estimate available energy at surface. Evaporative fraction was estimated from image–albedo two-dimensional (2D) space to convert available energy into latent heat fluxes image. The validation of coarse resolution image flux from K1VHRR was a two-step approach comprising (1) comparison of moderate resolution (0.01°) image flux from MODIS AQUA with in situ measurements over five different agricultural land uses, and (2) upscaling validated moderate resolution image to compare coarser resolution (0.08°) image fluxes. The net error in estimated daily ET from K1VHRR varied between 25% and 32% of measurements. The errors in estimates of eight-day ET sum were found to be little less (26%) as compared to daily estimates. Regional validation of K1VHRR eight-day ET from wall-to-wall comparison with aggregated moderate resolution ET estimates yielded a correlation coefficient (r) of 0.80 from 52,853 paired datasets over Indian agricultural land. In general, the error in ET estimates increased with increasing surface heterogeneity. The uncertainty in ET fluxes from K1VHRR due to inherent retrieval errors in core inputs was also assessed. The overall errors in image and ET estimates from K1VHRR were found to be at par with globally available experimental results using data from sensors on geostationary platform.