The use of remote sensing for reliable estimation of net radiation and its components: a case study for contrasting land covers in an agricultural hotspot of the Brazilian semiarid region

摘要

This study aims to ascertain the uncertainties related to the spatiotemporal estimation of net radiation, and its components, using remote sensing data. Geographical focus is an irrigated agricultural hotspot of the Brazilian semiarid region, for which we also investigate the impact that contrasting land-cover types have on the upwelling radiation balance components, and hence on net radiation. Instantaneous (R-n) and daily (R-n,R-24) values of net radiation were estimated based on OLI/TIRS-Landsat-8 images and key weather variables. In addition, we evaluated two models for downwelling shortwave (R-sw), ten models for downwelling longwave radiation (R-1w), and two models for derivation of R-n,R-24. The accuracy of each model was evaluated with radiation measurements obtained from research quality sensors installed in micrometeorological towers. The best performances were found for the Allen model, Duarte model, and De Bruin model for R-sw, R-1w,R- and R-n,R-24, respectively. The contrasting land-use types exhibited substantial differences in the biophysical variables and radiative properties that affect R-n. The albedo for the irrigated crops has average absolute values that are 0.01-0.03 greater than those found for the pristine caatinga, whereas the land surface temperature, LST, is 3-5 degrees smaller. However, R-n for these two distinctly different surface types was similar, as a result of a considerably lower surface emissivity in the caatinga. For rangeland, the albedo, LST, and hence the upwelling radiation had greater values than those found for the caatinga, which caused reduced values of R-n. The urban areas exhibited the lowest values of R-n, mainly as a consequence of their high albedo values. We show that when in-situ net radiation data are not available, remote sensing data combined with more readily available in-situ weather data can be used to derive spatiotemporal estimates of R-n. This facilitates the identification of anthropogenic impacts on the radiation at the land-surface and ultimately the energy balance, including the short-term seasonal and long-term effects.