Abstract
Although the Coordinated Regional Climate Downscaling Experiments (CORDEX) consistently provide dynamically downscaled high-resolution climate data, their biases are nonetheless corrected against subjectively selected observational datasets and varying bias-correction methods, making impact assessments inconsistent. Impact assessment communities, such as the Intersectoral Impact Model Intercomparison Project (ISIMIP), prefer consistent climate simulations that are corrected for biases using the same method against a common set of observational datasets. We present the quasi-global datasets of GloBCORD-HD, prepared at a daily temporal and 0.5° spatial resolution for the historical period (1950-2019) and the future period (2020-2099) for three Representative Concentration Pathways (RCPs) 2.6, 4.5, and 8.5. Three datasets were developed by bias-correcting CORDEX experiments, which dynamically downscaled the Earth System Models of MPI-M-MPI-ESM-LR, ICHEC-EC-EARTH, and NOAA-GFDL-GFDL-ESM2M over the maximum number of regional domains, against the single observational dataset GSWP3-W5E5 v2.0, using the same ISIMIP3BASD v2.5 bias correction algorithm. Comprehensive validation demonstrates that GloBCORD-HD reduces biases, improves regional extremes representation, enhances climate signals consistency, with seamless boundary transitions, and enables robust global impact assessments.