B31A-0528
Modeling Regional Climate Responses to 17th-20th Century Land Use Change over the Southeastern United States
Abstract:
Regional climate change is the combined response to global and regional radiative forcing, circulation and interactions between the atmosphere and Earth surface. A potentially key component of regional climate change derives from natural and human-caused land use and land-use change (LULC) and related atmosphere-surface feedbacks. In the first step of disentangling and quantifying the effect of LULC over the southeastern United States, we conducted a series of simulations with the RegCM4 regional climate model which includes the Biosphere Atmosphere Transfer Scheme (BATS) surface physics package. Land use and vegetation types determine the specified values of a number of biophysical and physical parameters in BATS such as albedo, roughness length, stomatal resistance, and rooting depth. We ran four simulations in which we specified BATS land use types derived from harmonized reconstructions for 1650, 1850, 1920 and Landsat-based observations for 1992 (Steyaert and Knox, 2008). The simulations were run over a model domain comprising a 20 km × 20 km horizontal grid and 23 atmospheric levels for the period 1990-2000. The boundary conditions for the simulations were derived from the NCAR-NCEP Reanalysis and the NOAA Optimum Interpolated Sea Surface Temperature global data sets.Depending on the time periods considered, the simulations indicate regionally coherent patterns of April-September changes in surface radiation, warming and cooling of up to 2 °C or more, substantial changes in soil moisture and atmospheric humidity associated with conversion of native vegetation to agriculture and agriculture back to forest, draining of wetlands and marshes, and urbanization. Extensive draining of wetlands in the lower Mississippi Valley during the 20th century induced a robust feedback with the atmosphere which suppressed convective summertime precipitation. Our joint analysis of LULC classes and the model simulations suggests a need to be able to disaggregate land use to a finer grid and expand BATS surface types to encompass a more diverse range of wetland and soil types.