Projecting climate change in the United States: A technical document supporting the Forest Service RPA 2010 Assessment

Abstract

A set of climate change projections for the United States was developed for use in the 2010 USDA Forest Service RPA Assessment. These climate projections, along with projections for population dynamics, economic growth, and land use change in the United States, comprise the RPA scenarios and are used in the RPA Assessment to project future renewable resource conditions 50 years into the future. This report describes the development of the historical and projected climate data set. The climate variables are monthly total precipitation in millimeters (mm), monthly mean daily maximum air temperature in degrees Celsius (°C), and monthly mean daily minimum air temperature in degrees Celsius (°C). Downscaled climate data were developed for the period 2001-2100 at the 5-arcminute grid scale (approximately 9.3 km by 7.1 km grid size at 40 degree N) for the conterminous United States. These data were also summarized at the U.S. county level. Computed monthly mean daily potential evapotranspiration (mm) and mean grid cell elevation in meters (m) are also included in the data set. The scenarios used here from the IPCC Special Report on Emissions Scenarios are A1B, A2, and B2. The A1B and A2 scenarios were used to drive three climate models: the Third Generation Coupled Global Climate Model, version 3.1, medium resolution; the Climate System Model, Mark 3.5 (T63); and the Model for Interdisciplinary Research on Climate, version 3.2, (T42), all used in the Fourth IPCC Assessment. The B2 scenario was used to drive three earlier generation climate models: the Second Generation Coupled Global Climate Model, version 2, medium resolution; the Climate System Model, Mark 2; and the UKMO Hadley Centre Coupled Model, version 3, all used in the IPCC Third Assessment. Monthly change factors were developed from global climate model output using the delta method. The coarse-resolution change factors were downscaled to a 5-arcminute resolution grid using ANUSPLIN. The 30-year mean historical climatology (1961-1990) was developed using the Parameter-elevation Regressions on Independent Slopes Model (PRISM) data at 2.5-arcminute resolution and aggregated to the 5-arcminute resolution grid. The downscaled change factors were combined with the PRISM observed climatology to develop nine future climate projections for the conterminous United States. These projection data and the change factor data are available through the U.S. Forest Service data archive website (http://www.fs.usda.gov/rds/archive/).