Monte Carlo uncertainty analysis of a regional‐scale transport chemistry model constrained by measurements from the Atmospheric Pollution Over the Paris Area (ESQUIF) campaign

Abstract

A Bayesian Monte Carlo uncertainty analysis constrained by measurements is applied to emission scenario calculations with the chemistry transport model CHIMERE within the Ile‐de‐France region. The overall uncertainty with respect to the following model input parameters is evaluated: anthropogenic and biogenic emissions, meteorological parameters such as wind speed and mixing layer height, actinic fluxes, quantum yields, and chemical rate coefficients. Airborne and ground‐based ozone and nitrogen species, volatile organic compounds (VOCs), and wind measurements from the Atmospheric Pollution Over the Paris Area (ESQUIF) campaign are used to construct an agreement function that assigns a larger weight to Monte Carlo simulations closer to observations. The observational constraint reduces the uncertainty in the simulations of daily surface ozone maxima (O3max) and the differential sensitivity of ozone formation (DSO) to NOx and VOC emissions reductions for 3 polluted days in the Ile‐de‐France region by a factor of between 1.5 and 3. Constrained uncertainties in O3max (expressed as relative differences between the 50th and the 10th or 90th percentiles) range from 15 to 30%, both for a baseline and for a 50% reduced emission scenario. Uncertainty in the DSO averaged over the plume ranges from 4 to 10 ppb. Including the observational constraint in the Monte Carlo analysis shifts the DSO in different directions for different days. Sensitivity tests with different input parameter distributions and agreement functions indicate the robustness of the above results.