Open Access

Souleymane SY, Fabio Madonna, Federico Serva, Ismaila Diallo, Benjamin Quesada

• Climate models' capability of reproducing the present climate at both globaland regional scales still needs improvements. The assessment of model perfor-mance critically depends on the data sets used as comparators/references. Rea-nalysis and gridded observational data sets have been frequently used for thispurpose. However, none of these can be considered an accurate reference dataset because of their associated uncertainties and full representativity. Thispaper, for the first time, uses in-situ measurements from National Oceanic andAtmospheric Administration U.S. Climate Reference Network (USCRN) span-ning the period 2006–2020 to assess daily temperature and precipitation from asuite of dynamically downscaled regional climate models (RCMs; driven byERA-Interim) involved in NA-CORDEX. The assessment is also extended tothe most recent and widely used Earth system reanalyses (ERA5, ERA-Interim,MERRA2 and NARR) and a few in situ-based gridded data sets (Daymet,PRISM, Livneh and CPC).Climate models' capability of reproducing the present climate at both globaland regional scales still needs improvements. The assessment of model perfor-mance critically depends on the data sets used as comparators/references. Rea-nalysis and gridded observational data sets have been frequently used for thispurpose. However, none of these can be considered an accurate reference dataset because of their associated uncertainties and full representativity. Thispaper, for the first time, uses in-situ measurements from National Oceanic andAtmospheric Administration U.S. Climate Reference Network (USCRN) span-ning the period 2006–2020 to assess daily temperature and precipitation from asuite of dynamically downscaled regional climate models (RCMs; driven byERA-Interim) involved in NA-CORDEX. The assessment is also extended tothe most recent and widely used Earth system reanalyses (ERA5, ERA-Interim,MERRA2 and NARR) and a few in situ-based gridded data sets (Daymet,PRISM, Livneh and CPC). Results show that biases for the different data setsare seasonally and subregionally dependent. On average, reanalysis and insitu-based gridded data sets are warmer (with biases exceeding 0.3� C) thanUSCRN year-round, while RCMs are colder (warmer) in winter (summer) withbiases ranging from−0.5 (0.9)� C for RCMs at 0.44� to−0.2 (1.4)� C forCRCM5-UQAM-11. In situ-based gridded data sets provide the best perfor-mance in most of the Contiguous United States (CONUS) regions compared toreanalyses and RCMs, but still have biases in regions such as the Westernmountains and the Pacific Northwest. Furthermore, in most US subregions,reanalysis data sets do not outperform reanalysis-driven RCMs. Likewise, forboth reanalysis data sets and RCMs, temperature and precipitation biases varyconsiderably depending on the local orography, with larger temperature biasesfor coarser model resolutions and precipitation biases for reanalysis.…