Preliminary Analysis of Wind Speed Simulations over the US Produced by Regional Climate Models

Regional climate models (RCMs) are finding increased use for studying climate variability and change at resolutions higher than are available from routine long-term simulations with global models. Most analyses of RCM results have focused on temperature and precipitation, but little attention has been devoted to simulations of wind speed. Pryor et al. (2008) have shown that studies of long-term trends in surface wind speeds from surface observations are complicated by changes in instrumentation and changes in the near- field environment of anemometers, among other factors. Long-term RCM simulations from the NARCCAP archive offer an alternative way to study diurnal, seasonal, and long-term trends in winds without these complications. We examine both model-reported 10-m winds and also the lowest-model-level winds. The near surface wind speeds are of high importance due to their role in regulating the exchange of heat, momentum, and moisture with the Earth's surface, but they are approximated by use of extrapolations from the model lowest level. The influence of these approximations is minimized for studying trends by focusing on the lowest-model-level winds themselves. Analysis of MM5 model wind results from the NARCCAP archive show generally decreasing winds over the period 1982-2004 across the new wind-farm region of the upper Midwest except for a weak increase across a narrow band from central Minnesota to northern Kansas. Largest contribution to this decrease comes in winter when changes on the order of -0.05 m/s/yr are produced by the model. The winter decrease is partially offset in the annual average by a small increase in fall. Additional analyses of results are in progress.