Global downscaled projections are now some of the most widely used climate datasets in the world, however, they are rarely examined for representativeness of local climate or the plausibility of their projected changes. Here we apply steps to improve the utility of two such global datasets (CHELSA and WorldClim2) to provide credible climate scenarios for climate change impact studies in Hawaii. Our approach is based on three steps: 1) Using a standardized baseline period, comparing available global downscaled projections with regional observation-based datasets and regional downscaled datasets (if available); 2) bias correcting projections using observation-based data; and 3) creating ensembles to make use of the differential strengths of global downscaling datasets. This data release includes a baseline (1983-2012) model as well as future climate projections for mid- (2040-2059) and late-century (2060-2079) for three regionally-adapted global datasets (CHELSA, WorldClim2, and a combined ensemble). We considered mean annual temperature (MAT) and mean annual precipitation (MAP) as our primary variables for comparison since they are the most widely used and desired datasets for climate impact studies. These regionally-downscaled future climate projections are available for various individual Global Circulation Models (GCMs) under four representative concentration pathways (RCPs; 2.6, 4.5, 6.0, and 8.5) for each global dataset.