Incorporating uncertainty into air quality modeling and planning - A case study for Georgia
Antara Digar1, Jim Boylan2, Maudood Khan2, Wei Zhou1, Dennis Cox3, and Daniel Cohan1 1 Department of Civil and Environmental Engineering, Rice University 2Georgia Environmental Protection Division (GA EPD) 3Department of Statistics, Rice University

The sensitivity of atmospheric pollutants to emissions changes is crucial to informing sound choices for air pollutant control strategies. However, there is no direct way to gauge the accuracy of photochemical sensitivities. We will present a novel method to gauge how uncertainty in photochemical model input parameters (reaction rates, emission inventories, deposition velocities, etc.) leads to uncertainty in the sensitivities of ozone and fine particulate matter to precursor emissions. The method uses High-order Decoupled Method (HDDM) and other sensitivity analysis in the Community Multi-scale Air Quality (CMAQ) model to develop a surrogate model characterizing the responsiveness of sensitivities to parameter uncertainties. Monte Carlo analysis can then readily be applied to the surrogate model, sampling from the probability distributions of the input parameters, to develop probability distribution functions for the output sensitivities. We will illustrate the methodology in detail and furnish some preliminary results from applying this method to a case study of ozone and PM attainment planning in Georgia. In that study, the uncertainties in photochemical sensitivities will be linked with uncertainties in control measure cost-effectiveness and health effects to explore the range of uncertainties that influence attainment strategy development.