A framework for estimating cloudy sky surface downward longwave radiation from the derived active and passive cloud property parameters

Abstract

The cloud-base temperature (CBT) is one of the parameters that dominates the cloudy sky surface downward longwave radiation (SDLR). However, CBT is rarely available at regional and global scales, and its application in estimating cloud sky SDLR is limited. In this study, a framework to globally estimate cloud sky SDLR during both daytime and nighttime is proposed. This framework is composed of three parts. First, a global cloudy property database was constructed by combing the extracted cloud vertical structure (CVS) parameters from the active CloudSat data and cloud properties from passive MODIS data. Second, the empirical methods for estimating cloud thickness (CT) under ISCCP cloud classification system and MODIS cloud classification system were developed. Additionally, the coefficients of CERES CT estimate models were refitted using the constructed cloud property database. With the estimated CT and reanalysis data, calculating the CBT is straightforward. The accuracy of the estimated CT for ISCCP cloud type is compared with the existing studies that were conducted at local scales. Our CT estimate accuracy is comparable to that of the existing studies. According to the validation results at ARM NSA and SGP stations, the CT estimated by the developed CT model for MODIS cloud type is better than that estimated by the original CERES CT model. Finally, the cloudy sky SDLR values were derived by feeding the estimated CBT and other parameters to the single-layer cloud model (SLCM). When validated by the ground measured SDLR collected from the SURFRAD network, the bias and RMSE are 5.42 W∙m−2 and 30.3 W∙m−2, respectively. This accuracy is comparable to the evaluation results of the mainstream SDLR products (Gui et al. 2010), the new evaluation results of SLCMs (Yu et al. 2018), and the accuracy of a new cloudy sky SDLR estimate method (Wang et al. 2018). All the derived CBTs improve the SDLR estimate accuracy more than the SLCM that directly uses cloud-top temperature (CTT). We will collect more ground measurements and continue to validate the developed framework in the future.