This repository contains gridded raster datasets used to quantify climate velocity and its components for annual potential evapotranspiration (PET) across the North American boreal region. The data support analyses presented in the associated pre-print “Large lakes may moderate projected climate-change velocity in boreal North America”. Climate velocity was calculated using the gradient method, defined as the ratio of the temporal gradient (rate of change over time) to the spatial gradient (rate of change across space). The datasets enable comparison of climate velocity patterns across alternative climate data products, downscaling approaches, spatial resolutions, and global climate model (GCM) inputs, with particular emphasis on lake-mediated climate buffering. Included data For each combination of climate product, spatial resolution, and GCM, the repository includes GeoTIFF rasters for: Climate velocity (cv, km yr⁻¹) Spatial gradient of PET (sg, mm km⁻¹) Temporal gradient of PET (tg, mm yr⁻¹) Climate data products include: ClimateNA (station-interpolated baseline with statistical downscaling) ERA5 reanalysis (model-based baseline with statistical downscaling) CRCM5 regional climate model (dynamical downscaling) Spatial resolutions: 1 km 22 km Global climate models: CanESM2 MPI-ESM-LR CNRM-CM5 All rasters are spatially aligned and projected consistently within each resolution. Methods summary PET was calculated from monthly temperature and precipitation using Hogg’s (1997) modified Penman–Monteith formulation. Climate velocity and its components were calculated using the gradient-based method implemented in the VoCC R package, following Loarie et al. (2009) and Burrows et al. (2011). Spatial gradients were derived from baseline climatologies using a moving-window approach, and temporal gradients represent mean projected change between historical (1981–2010) and late-century (2071–2100) periods under the RCP 8.5 scenario. Intended use These data are intended for: Climate-change exposure and vulnerability assessments Evaluation of climate velocity metrics across alternative baseline representations and downscaling approaches Identification of areas with low climate velocity and potential climate-change refugia Comparative or methodological studies of climate velocity in lake-rich or data-sparse regions Velocity estimates are sensitive to baseline climate representation, spatial resolution, and downscaling method, as discussed in the associated manuscript. Citation Stralberg, D., Sang, Z., Hillman, A., Morrison, K., Beckers, J., Parisien, M.-A., Nielsen, S. E., Paquin, D., & Logan, T. (2026). Large lakes may moderate projected climate‐change velocity in boreal North America. Zenodo. https://doi.org/10.5281/zenodo.18066597