Nonlinear response of ozone to precursor emission changes in China: a modeling study using response surface methodology

Abstract

Statistical response surface methodology (RSM) is successfully applied for a Community Multi-scale Air Quality model (CMAQ) analysis of ozone sensitivity studies. Prediction performance has been demonstrated through cross validation, out-of-sample validation and isopleth validation. Sample methods and key parameters, including the maximum numbers of variables involved in statistical interpolation and training samples have been tested and selected through computational experiments. Overall impacts from individual source categories which include local/regional NO_x and VOC emission sources and NO_x emissions from power plants for three megacities – Beijing, Shanghai and Guangzhou – were evaluated using an RSM analysis of a July 2005 modeling study. NO_x control appears to be beneficial for ozone reduction in the downwind areas which usually experience high ozone levels, and NO_x control is likely to be more effective than anthropogenic VOC control during periods of heavy photochemical pollution. Regional NO_x source categories are strong contributors to surface ozone mixing ratios in three megacities. Local NO_x emission control without regional involvement may raise the risk of increasing urban ozone levels due to the VOC-limited conditions. However, local NO_x control provides considerable reduction of ozone in upper layers (up to 1 km where the ozone chemistry is NO_x-limited) and helps improve regional air quality in downwind areas. Stricter NO_x emission control has a substantial effect on ozone reduction because of the shift from VOC-limited to NO_x-limited chemistry. Therefore, NO_x emission control should be significantly enhanced to reduce ozone pollution in China.