Estimation of all-sky downwelling longwave radiation over the Tibetan Plateau using an improved parameterization scheme

Downwelling longwave radiation (DLR) is crucial for the global energy cycle. The Tibetan Plateau (TP) is a focal point in global energy cycle research owing to its distinct geographical position and remarkably high elevation. At present, the DLR estimates under clear-sky conditions are relatively mature, but only a few studies have specifically estimated the DLR under all-sky conditions. Moreover, some of these methods still need to be improved with regard to spatial resolution and accuracy when applied over the TP. The primary challenge is the uncertainty of cloud radiation effects and the atmospheric conditions beneath the clouds. Current parameterization schemes often rely solely on near-surface meteorological parameters and the cloud fraction, which are insufficient for characterizing the thermal differences between the cloud base and the surface. Additionally, optical sensors are limited by their penetration depth and cannot directly provide information on the cloud base. By combining satellite data, meteorological forcing data, reanalysis temperature profiles, and land surface temperature datasets and simultaneously considering the thermal radiation contributions from both the atmosphere below clouds and the cloud layer itself, the all-sky DLR over the TP was estimated. With the introduction of a low-cloud correction scheme and the incorporation of multiple temperature and humidity input parameters when estimating the radiation contribution from the atmosphere below clouds, these improvements further enhance the accuracy. The precision of this study is comparable to that of CERES-SYN, with RMSEs below 30 W m⁻² at any timescale, and more detailed spatial variations can be presented due to the higher spatial resolution. A comparison with existing DLR estimation schemes shows that this study achieves more accurate results without the need for local calibration with preobtained in situ data. Therefore, this method shows the potential for application across various regions globally to further improve the precision of DLR estimation.