United States of America - (BioClim) Climate Scenarios annual steps 2010 to 2100 - NASA

This dataset comprises two climate scenarios for the contiguous United States at a resolution of ~800m x 800m, with annual time slices from 2010 to 2100. Data include nineteen bioclimatic variables that are commonly used in ecological analyses.

As described in Supplementary Material to the above paper: The procedure for generating an annual time series of climate variables comprised three steps: First, (MAGICC/SCENGEN 5.3), a coupled gas cycle/aerosol/climate model used in the IPCC Fourth Assessment Report1, was used to generate an annual time series of future climate anomalies (2010 – 2100) using an ensemble of five atmosphere-ocean general circulation models (GCMs). Fordham et al.2 have highlighted the advantages of working within the MAGICC/SCENGEN framework, rather than using GCM data from the Coupled Model Intercomparison Project 3 (CMIP3) archive. We used two strongly contrasting greenhouse gas emission scenarios: a Reference scenario that assumes high CO2 concentration (WRE750;3) and a Policy scenario that assumes CO2 stabilization at about 450 ppm (MiniCAM LEV1;4). GCMs were chosen according to their superior skill in reproducing seasonal precipitation and temperature across North America. Model performance was assessed following already published methods5. The five GCMs were: UKMO-HadCM3 (UK); CGCMA.31(T47) (Canada); MRI-CGCM2.3.2 (Japan); ECHAM5/MPI-OM (Germany); IPSL-CM4 (France). Model terminology follows the CMIP3/AR4 multi-model data archive. Four of these models have been shown elsewhere to have good retrospective skill in reproducing recent climates at a global scale, as well as for North America2. GCM skill assessment results can be quite different depending on the variable considered, the region studied, the month or season examined, or the comparison metric used5. However, ensemble forecasts that include five or more GCMs tend to be more robust to GCM choice6. Second, climate anomalies were downscaled to an ecologically relevant spatial resolution (~800m x 800m)7, using the “change factor” method, where the low-resolution climate signal (anomaly) from a GCM is added directly to a high-resolution baseline observed climatology (we used PRISM 1971-2000 normals;8,9. Bi-linear interpolation of the GCM data (2.5 x 2.5º longitude/latitude) to a resolution of 0.5 x 0.5º longitude/latitude was used to reduce discontinuities in the perturbed climate at the GCM grid box boundaries2. One advantage of this method is that, by using only GCM change data, it avoids possible errors due to biases in the GCMs baseline (present-day) climate5. Third, we generated 19 bioclimate variables10 from monthly estimates of minimum temperature, maximum temperature, and mean precipitation generated by the above steps.