Projections of future global climate have been developed by numerous climate modeling groups around the world; however, this data is often at spatial scales much larger than the spatial scale of resource management. This study develops a set of change factors that can be used with a user-selected historical climate data set to create climate change projections at the spatial scale of approximately 9.25 kilometer grid. Climate projection output was obtained from four well-established general circulation models (GCM) forced by each of three greenhouse gas (GHG) emissions scenarios namely A2, A1B, and B1, from the Special Report on Emissions Scenarios, and used in the Intergovernmental Panel on Climate Change Fourth Assessment Report. Monthly data for the period 1961-2100 were downloaded mainly from Third Coupled Model Intercomparison Project (CMIP3) through the Program for Climate Model Diagnosis and Intercomparison (PCMDI) web portal. Climate variables included monthly mean daily maximum and minimum temperatures, precipitation, solar radiation, wind speed, and vapor pressure (used to calculate specific humidity). All variables are expressed as changes relative to the simulated monthly means for 1961-1990, which corrected for GCM bias in reproducing past climate and allowed future projected trends to be compared directly. Each month value at each GCM grid node was normalized either by subtracting (for temperature variables) or dividing by (for other climate variables) the mean of that month's value for the 30-year baseline period 1961-1990. The normalized data (or "deltas") we then formatted for input to ANUSPLIN thin-plate software. The downscaling procedure used ANUSPLIN software package to fit a two-dimensional spline function to each month's change data for each of the six normalized climate variables at a spatial resolution of 5 arcminutes (0.0833 degrees) longitude and latitude.
This research was a collaborative effort between scientists at the Canadian Forest Service and scientists at the USDA Forest Service. Data for the United States and Canada were extracted. Data for Alaska are contained in this dataset.