Global monthly CMIP6-downscaled high-resolution (1 km) near-surface air temperature dataset (1950-2100)

Temperature projections from general circulation models (GCMs), serving as an important approach of understanding future global warming, are essential for developing adaptation and mitigation strategies of climate change. However, the coarse spatial resolutions (~1-3°) limit their effectiveness at fine-scale (e.g., intra-urban) research. Here, we produced MoCHAT, a global monthly CMIP6-downscaled high-resolution (1 km) near-surface air temperature dataset. We utilized delta downscaling method to generate MoCHAT based on NEX-GDDP-CMIP6 and WorldClim. MoCHAT encompasses mean, maximum, and minimum air temperature of 16 GCMs. It covers both the historical period (1950-2014) and future scenarios (2015-2100) under three Shared Socioeconomic Pathways (SSPs) scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5). Validation with meteorological station observations and existing high-resolution climatic datasets showed that the mean absolute errors for these variables range from 1.60 to 2.38 K and overall biases below 2.0 K. With sufficiently long span and fine resolution, MoCHAT breaks through data resolution limitations and provides solid support for global fine-scale heat risk research.