Evaluation of Downward Surface Longwave Flux Estimates Using Meteosat Cloud Observations

Abstract

The downward surface longwave flux (DSLF) plays a relevant role in the Earth’s surface radiative budget, which is crucial to monitor, understand and model the impact of changes at local and global scales on surface temperature and surface conditions. This study focuses on the evaluation and intercomparison of four DSLF products: (a) a recently developed all‐weather DSLF product based on the multivariate adaptive regression splines (MARS) algorithm driven by satellite cloud information from the Meteosat Second Generation (MSG) and ERA5 reanalysis screen variables; (b) the Satellite Application Facility on Land Surface Analysis (LSA SAF); (c) CERES Synoptic top‐of‐atmosphere and surface fluxes and clouds (CERES‐SYN1deg) and (d) ERA5 reanalysis. The study covers the period 2005–2021 and the MSG region focusing on monthly means. The evaluation performed against 48 ground stations from the Baseline Surface Radiation Network (BSRN) and FLUXNET2015 networks showed that the MARS product outperforms the remaining products, particularly the LSA SAF, while ERA5 and CERES show similar performance metrics. The fours products are intercompared in terms of their mean spatial variability and temporal mean annual cycles and inter‐annual variability in four selected regions, showing a high level of agreement, particularly between MARS, ERA5 and CERES. Our results highlight the clear added value of MARS with respect to LSA SAF, while providing higher spatial resolution (0.05°), constrained by satellite cloud information, when compared with ERA5 (0.25°) or CERES (1°).