Structural Similarity Assessment of Precipitation and Temperature in NEX-GDDP-CMIP6 Simulations over the Contiguous United States

In a warming climate, spatiotemporal changes in precipitation and temperature can impact hydrologic processes and thus the quality of freshwater ecosystems. This study evaluates the performance of 35 downscaled and biascorrected Coupled Model Intercomparison Project phase 6 (CMIP6) general circulation models (GCMs) from NASA’s Earth Exchange Global Daily Downscaled Projections (NEX-GDDP, 0.258 resolution) by comparing their historical simulations (1950–2014) of precipitation and near-surface air temperature with those from the fifth generation European Centre for Medium-Range Weather Forecasts (ECMWF) atmospheric reanalysis (ERA5) reanalysis across the contiguous United States. Two complementary metrics are employed: the Wasserstein distance (WD) to assess distributional similarity and the structural similarity index measure (SSIM) to quantify spatial consistency on a monthly scale. The results reveal systematic uncertainties: The NEX-GDDP-CMIP6 simulations struggle to reproduce the warm-season spatial organization and temporal trends of precipitation in regions predominantly influenced by mesoscale convective systems. Additionally, they exhibit the lowest SSIM scores for cold-season temperatures, highlighting challenges in simulating snow–albedo feedback mechanisms, cloud-radiative processes, and boundary layer dynamics. The models are ranked based on the similarity of their historical simulations for both precipitation and temperature. In summary, IPSL-CM6A-LR leads in monthly distributional proximity, while MIROC-ES2L leads in monthly spatial similarity for total precipitation and mean temperature. Regionally, MIROC-ES2L performs best in the Midwest, Northern Great Plains, Northwest, and Southern Plains, whereas NorESM2-LM, FGOALS-g3, and GFDL-CM4-gr2 lead in the Northeast, Southeast, and Southwest, respectively. The spatial similarity analysis of maps of temporal trends reveals that BCC-CSM2-MR is the top model across contiguous United States (CONUS). SIGNIFICANCE STATEMENT: Reliable projections of space–time dynamics of precipitation and temperature are essential for predicting risks associated with floods, droughts, and eutrophication. This study evaluates historical simulations (1950–2014) from 35 climate models by NASA’s Earth Exchange Global Daily Downscaled Projections (NEXGDDP) over the contiguous United States and its subregions. The assessment quantifies the ability of climate models to reproduce the distributions and spatial patterns of monthly precipitation and temperature relative to observationally constrained Earth system model historical simulations. The findings quantify differences in model performance, providing practical guidance for the regional selection of the most suitable models for future predictions of precipitation and temperature changes, as well as their spatiotemporal trends, with respect to their impacts on eutrophication of lakes and river systems.