Estimation of 1km downwelling shortwave radiation over the Tibetan Plateau under all-sky conditions

Abstract

Downwelling shortwave radiation (DSR) is the basic driving force for the energy and water cycles of the Earth's climate system. Called the Third Pole of the Earth, the Tibetan Plateau (TP) absorbs a large amount of shortwave radiation and exerts important impacts on global weather and climate change. However, due to coarse spatial resolution and insufficient consideration of factors influencing radiative transfer processes, DSR parameterization schemes still need to be improved when applied to the TP. Based on satellite datasets and meteorological forcing data, all-sky DSR over the TP at a spatial resolution of 1km was derived using an improved parameterization scheme. The influence of topography and different radiative attenuations were comprehensively taken into account. Specifically, the introduction of cloud multiscattering and topography factors further improves the DSR estimation accuracy. The validation results indicated that the developed parameterization scheme showed reasonable accuracy. By comparing with current, widely used DSR products based on the same in situ observations, the derived DSR performed much better on different spatial and temporal scales. On instantaneous, 10d and monthly timescales, the root-mean-square errors (RMSEs) of the derived DSR are 132.8-158.2, 70.8-76.5 and 61.3-67.5Wm-2, respectively, which are much smaller than those of current DSR products. The derived DSR not only captured the temporal-variation characteristics that are more consistent with the in situ measurements, but also provided reasonable spatial patterns. Meanwhile, the proposed parameterization scheme demonstrated its superiority in characterizing more details and high dynamics of the spatial pattern of DSR due to its terrain correction and high resolution. Moreover, this parameterization scheme does not need any local correction in advance and has the potential to be extended to other regions in the world.