Daily Evapotranspiration over Lebanon by Fusing VIIRS and PROBA-V Satellite Imagery

Abstract

During the last few decades, water demand in the whole world has been increasing due to the increase in population, climate change, global warming, and restriction of natural resources. With this increase, monitoring water resources has become more significant and concerns of agriculture have been increasing daily since it’s the largest consumer of water. Evapotranspiration is the main element in the water budget and energy balance. Therefore, accurate measurements of ET are essential for evaluating water resources and quantifying the irrigation water efficiency at global and local scales. In this study, daily ET maps for Lebanon were generated by applying the pySEBAL model using VIIRS thermal imagery, PROBA-V surface reflectance imagery, three hourly GLDAS NOAH v2.1 meteorological data, and FAO soil grids. Daily analysis of three potato fields in the Bekaa Valley of Lebanon was conducted for the year 2016, and ET of different summer crops was compared to literature values for the year of 2016. In addition to that, monthly and annual ET maps were produced for the period 2015-2017 in ArcMap. Monthly analysis throughout the study period was done by comparing the minimum and maximum monthly ET, and annual analysis by comparing annual mean ET and annual precipitation were conducted. Peak mean monthly ET in 2015, 2016, and 2017 was recorded to be respectively 150, 152, and 153 mm/ month in May, and annual ET mean in 2015, 2016, and 2017 was respectively 1070, 1124, and 1112 mm/year. VIIRS-PROBA-V monthly and annual ET values were overestimated by the pySEBAL model at the coast. Moreover, pySEBAL based VIIRS-PROBA-V and Landsat ET were compared over different land use land covers by analyzing evaporation, transpiration, and evapotranspiration, seasonal analysis of ET, and water consumption calculation of each cover in the year 2016. Results show that VIIRS-PROBA-V daily ET helps monitor ET at higher temporal scales. Annual ET exceeded annual precipitation indicating that the model overestimates ET for some land-use types, particularly evaporation from bared soils. Also, the comparison between Landsat and VIIRS-PROBA-V showed that while their ET is highly correlated, Landsat tends to overestimate monthly and annual ET due to its smaller overpass frequency. The study shows the need for a higher-spatiotemporal frequency of thermal satellites for accurate monitoring of water use and water stress in the MENA region.