Global sensitivity analysis of ozone, HO2, and OH during ARCTAS campaign

Abstract:

Modeling the chemical state of the atmosphere is a complicated endeavor due to the complex, non-linear interactions between meteorology, emissions, and kinetics that govern trace gas concentrations. Given the rapid environmental changes taking place, the Arctic is one area of particular interest with regards to climate and atmospheric composition. To observe these changes to the Arctic atmosphere, NASA funded the Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS) campaign (2008). As part of the mission, measurements of oxidative factors (hydroxyl (OH) and hydroperoxyl (HO₂) abundances) were taken using the Airborne Tropospheric Hydrogen Oxides Sensor (ATHOS) aboard the NASA DC-8. Using GEOS-Chem, a popular global chemical transport model, we perform a global sensitivity analysis for the period of the ARCTAS campaign, allowing for non-linear interactions between input factors to be accounted and quantified in the analysis. Sensitivities are determined for around 50 model input factors and for combinations of pairs of input factors using the Random Sampling – High Dimensional Model Representation (RS-HDMR) method. We calculate the uncertainty in these oxidative factors, and in ozone, ozone production rate, and hydroxyl production rate and find the sensitivity of these oxidative factors and the differences between the measured and modeled oxidative factors to model inputs in meteorology, emissions, and chemistry. This presentation will include a solid estimate of GEOS-Chem model uncertainty for the period of the ARCTAS campaign, the emissions, meteorology, or chemistry to which oxidative properties are most sensitive for these periods, and the factors to which the differences between the modeled and measured oxidative factors are most sensitive.