Estimations of net radiation flux distributions from automatic weather stations and satellite data in the Upper Blue Nile (UBN) basin, Ethiopia

Net radiation is the main energy balance component of land surfaces. It is an important factor in the studies of land-atmosphere processes, water resources management and so on. This is particularly true in the UBN basin where significant parts of the basin are dry and evapotranspiration (ET) is a major path of water loss. In this paper, we have estimated instantaneous net radiation distributions in the basin from MODIS Terra satellite and Automatic Weather Station (AWS) data. As downward shortwave radiation and air temperature usually vary spatially due to topographic effects, which are common features of our study area, we had applied residual kriging spatial interpolation approaches in the conversion processes of point weather data to surface data. Validation attempts of the simulated net radiation outputs with an independent field measurement at Choke flux tower site, which is in the central part of the basin, has shown that our method were able to reproduce downward shortwave, upward shortwave., and net radiation flux with a statistical metrics of Mean bias (MB) and Root Mean Square (RMSE) lesser than other studies done in similar physiographic regions in several parts of the world. It looked that the use of AWS data and residual kriging spatial interpolation technique made our results robust and even comparable to works done using finer spatial resolution satellite data than MODIS. The estimated net shortwave, net longwave and overall net radiations were in close agreement with ground truth measurements with MB of -14.84, 5.7 & 20.53 Wm^-2 and RMSE 83.43, 32.54 & 78.07 Wm^-2 respectively. The method has potential applications in research works like energy balance, ET estimation, and weather predictions in regions with similar physiographic features as that of the Nile basin.