Surface downward longwave radiation estimation from new generation geostationary satellite data

Surface downward longwave radiation (SDLR) is a key component of the Earth's energy budget. To use newgeneration geostationary satellite data with high spatial and spectral resolution for SDLR estimation, we developed a clear-sky SDLR algorithm based on multi-channel brightness temperature (BT) and precipitable water vapor (PWV), named as MBTW, for strictly clear conditions, and a clear-sky SDLR algorithm based on atmospheric parameters (named as NewAtmos) for probably clear conditions. Then, we developed a framework to estimate all-sky SDLR; the framework combined the two clear-sky algorithms and an existing cloudy-sky algorithm. Finally, the clear-sky and all-sky SDLRs were calculated using the proposed algorithms and framework for the FY-4A and GOES-16 data. The validation results against field measurements indicated that the MBTW and NewAtmos algorithms performed well under the expected clear-sky conditions. The clear-sky SDLRs combined using MBTW and NewAtmos algorithms had RMSEs of 24.8 W/m2 and 18.3 W/m2 for the FY-4A and GOES-16 data, which is comparable to those derived using a single algorithm based on atmospheric parameters; the former portrayed better consistency with field measurements than the latter. The RMSE of instantaneous all-sky SDLR calculated using the proposed method were 24.6 W/m2 and 20.1 W/m2 for FY-4A and GOES-16 data, and they were 23.7 W/m2 and 18.3 W/m2 for the hourly SDLR. The hourly SDLR from the combined algorithms showed a better accuracy than those of the ERA5 SDLR at the same resolution, and also yielded more details of spatial variation than those from ERA5 and from single atmospheric-parameter-based algorithm, because highresolution thermal infrared (TIR) BT was used in our method. Furthermore, using ERA5 derived PWV, cloud liquid water path and ice water path produced very similar SDLR to those produced using satellite parameters; therefore, for the conditions missing satellite parameters, the ERA5 parameters can be an alternative to achieve continuous SDLR estimation.

Automated summary

This study developed clear-sky and all-sky algorithms for estimating surface downward longwave radiation (SDLR), incorporating high-resolution satellite data. The proposed algorithms showed good performance in estimating SDLR under clear-sky conditions and provided more accurate and detailed results compared to single algorithms and ERA5 data.