While most ecological studies focus on the impact of climate change on some populations or species, broader, ecosystem-based information is needed to better predict the future state of the Arctic. We therefore conducted a large-scale study aimed at characterizing the vulnerability of tundra ecosystems to climate change through modelling. As part of this study, we obtained variables describing the potential climate of the province of Québec north of the 50th parallel during the 21st century at a fine resolution (10km x 10km grid cells). Raw climate data (monthly minimum and maximum temperatures and total precipitation from January 2011 to December 2100) come from 10 climate simulations obtained from the CORDEX program (Giorgi et al. 2009; data available at http://cordex.org/data-access) and Ouranos (Separovic et al. 2013, Martynov et al. 2013; data available at https://www.ouranos.ca/en/program/climate-simulation-and-analysis). These simulations come from regional climate models driven by global climate model under one RCP (RCP4.5 or RCP8.5). From these raw layers (for one simulation) we derived and interpolated (Natural neighbor interpolation, see the Supplementary documentation section for further details) 27 annual climate variables, including mean annual temperature, total annual precipitation, temperature/precipitation of the coldest/warmest/driest/wettest month, growing/freezing degree-days, growing/freezing dates, and water balance (see exhaustive list in the Supplementary documentation section). Finally we averaged annual values to build climate normals for three future periods (2011-2040, 2041-2070 and 2071-2100). The climate layers are available at a 10km x 10km spatial resolution in NetCDF format (see the Supplementary documentation section to use NetCDF files with R). A total of 30 NetCDF, each containing 27 variables, is available (one file per simulation and per future period).
