The dataset presented here was developed in a collaborative effort between The World Bank, The Nature Conservancy, Climate Central, and Santa Clara University to produce the first standardized set of daily downscaled GCM projections that span the entire globe. This includes all terrestrial daily data archived in CMIP3-the standard for raw GCM data distribution from the IPCC Fourth Assessment Report: nine different GCMs, some with multiple model runs, across three different greenhouse gas emissions scenarios (SRES A2, A1b, B1), totaling 53 future projections, all downscaled to a 0.5 degree resolution (~50 km) for the time periods of 1961-1999, 2046-2045, and 2081-2100.
The downscaling method used was a daily-timescale variant of a method known as Bias Correction/Spatial Downscaling (BCSD) that has been widely applied to produce monthly downscaled quantities based upon monthly GCM results. The monthly version of the method is described by Wood et al. (2002 and 2004).The downscaling and bias-correction was done using historical observed daily gridded observations. The base meteorological data consists of daily time-series for the period of 1950 through 1999 of precipitation, maximum temperature and minimum temperature. Monthly station data from a variety of sources (including the Global Historical Climatology Network (GHCN) version 2 data) were compiled and gridded over all global land areas. The daily variability of precipitation, maximum and minimum temperature was constructed using other global daily datasets, which were scaled to match the monthly values. Click HERE (PDF 2.03MB) for more a more complete information on the methodology used.