Evaluation of Landsat and MODIS imagery fusion for high-resolution evapotranspiration mapping over large agricultural area

Abstract

Regarding to severe droughts and increasing shortages in water resources, in particular in the Middle East, evapotranspiration estimation is essential for irrigation management in agricultural lands. Evapotranspiration mapping is currently possible in a suitable spatial resolution from Landsat imagery; however, these data do not allow to obtain an optimal temporal resolution, required in agriculture management. In other hand, Modis imagery provide a dense temporal resolution, with however a deficiency in spatial resolution. In this study, we evaluated the fusion of 16-dayily Landsat and daily Modis imageries in order to obtain a sufficient spatio-temporal resolution for evapotranspiration mapping required in water management over agricultural areas. The image fusion is applied on the six-month Landsat and Modis images over large sugarcane farms in south-western Iran. The Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM) are applied on the images. The evapotranspiration estimations are based on two algorithms: SEBAL and METRIC. The point extracted results are highly consistent with the evapotranspiration product from Modis in terms of temporal resolution. As well as, the Kappa index indicates on an acceptable accuracy in spatial resolution of estimations. Due to the smoothness of the region, no significant differences between SEBAL and METRIC estimations are resulted.