Developing ensemble mean models of satellite remote sensing, climate reanalysis, and land surface models

This study aims to access the selected satellite remote sensing, climate reanalysis, and land surface models to estimate monthly land surface air temperature (LSAT), solar radiation (SR), and precipitation (P) at the global scale. To this end, we apply six datasets including Modern-Era Retrospective Analysis for Research and Applications-version 2 (MERRA-2), European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis-version 5 (ERA-5), ERA-5-Land version (ERA5-Land), Global Land Data Assimilation System (GLDAS), Famine Early Warning Systems Network (FEWS NET) Land Data Assimilation System (FL and Global Precipitation Climatology Project (GPCP). In terms of SR, we compare the selected products against the National Oceanic and Atmospheric Administration (NOAA)-Cooperative Institute for Research in Environmental Sciences (CIRES)-Department of Energy (DOE) Twentieth Century Reanalysis (20CR) (NOAA-CIRES-DOE 20CR) dataset from 1982 to 2015. For LSAT and P, we consider NOAA Climate Prediction Center (CPC) (NOAA-CPC) as the reference dataset in the periods of 1982–2020 and 1983–2019, respectively, based on available data. ERA5-Land, MERRA-2, and GLDAS show the best results with root mean square difference (RMSD) equal to 19.03 W/m ², 1.93 °C, and 37.61 mm/month for SR, LSAT, and P estimates compared to NOAA datasets. Since there are uncertainties in all of the products, here we introduce new datasets based on merging the best products concerning their accuracy. The evaluation results can be used also as feedback to developers to improve the products and to facilitate the users to understand the status of the products and better use them for practical applications on a global scale.