Regional Downscaling for Air Quality Assessment

Assessing future changes in air quality using downscaled climate scenarios is a relatively new application of the dynamical downscaling technique. This article compares and evaluates two downscaled simulations for the United States made using the fifth-generation Pennsylvania State University–NCAR Mesoscale Model with the goal of understanding how errors in the downscaled climate simulations may introduce uncertainty in air quality assessment. The two downscaled simulations were driven by boundary conditions from the NCEP–NCAR global reanalysis and a global climate simulation generated by the Goddard Institute for Space Studies global circulation model, respectively. Comparisons of the model runs are made against the boundary layer and circulation characteristics of the North American Regional Reanalysis, and also against observed precipitation. The relative dependence of different simulated quantities on regional forcing, model parameterizations, and large-scale circulation provides a framework to understand similarities and differences between model simulations. Results show significant improvements in the downscaled diurnal wind patterns, in response to the complex orography, that are important for air quality assessment. Evaluation of downscaled boundary layer depth and winds, precipitation, and large-scale circulation shows larger biases related to model physics and biases in the GCM large-scale conditions. Based on the comparisons, recommendations are made to improve the utility of downscaled scenarios for air quality assessment.