Two reduced form air quality modeling techniques for rapidly calculating pollutant mitigation potential across many sources, locations and precursor emission types

Abstract

Due to the computational cost of running regional-scale numerical air quality models, reduced form models (RFM) have been proposed as computationally efficient simulation tools for characterizing the pollutant response to many different types of emission reductions. The U.S. Environmental Protection Agency has developed two types of reduced form models based upon simulations of the Community Multiscale Air Quality (CMAQ) modeling system. One is based on statistical response surface modeling (RSM) techniques using a multidimensional kriging approach to approximate the nonlinear chemical and physical processes. The second approach is based on using sensitivity coefficients estimated with the Decoupled Direct Method in 3 dimensions (CMAQ-DDM-3D) in a Taylor series approximation for the nonlinear response of the pollutant concentrations to changes in emissions from specific sectors and locations. Both types of reduced form models are used to estimate the changes in O3 and PM2.5 across space associated with emission reductions of NOx and SO2 from power plants and other sectors in the eastern United States. This study provides a direct comparison of the RSM- and DDM-3D-based tools in terms of: computational cost, model performance against brute force runs, and model response to changes in emission inputs. For O3, the DDM-3D RFM had slightly better performance on average for low to moderate emission cuts compared to the kriging-based RSM, but over-predicted O3 disbenefits from cuts to mobile source NOx in very urban areas. The RSM approach required more up-front computational cost and produced some spurious O3 increases in response to reductions in power plant emissions. However the RSM provided more accurate predictions for PM2.5 and for predictions of very large emission cuts (e.g. −60 to −90%). This comparison indicates that there are some important differences in the output of the two approaches that should be taken under consideration when interpreting results for a given application.