Model Analysis of Fast Photochemistry Over the Arctic Using In-Situ ARCTAS and TOPSE observations

Aircraft observations of constituents and meteorological quantities observed during the two seasonal Arctic phases of ARCTAS (Arctic Research of the Composition of the Troposphere from Aircraft and Satellites) and during the 2000 TOPSE (Tropospheric Ozone Production About the Spring Equinox) are analyzed using an observationally-constrained steady state box model. An examination of the springtime Arctic portion of the 2000 TOPSE program shows a highly similar meteorological background and chemical composition relative to ARCTAS-A, with the exception of peroxides. Concentrations of H2O2 observed during ARCTAS-A were 2-3 times larger than those during TOPSE. The cause of this discrepancy is unresolved, and it will be shown to have important implications for conclusions related to the Arctic HOx budget. Measurements of HOx from the Penn State ATHOS instrument are available during ARCTAS and are compared to box model predictions. Model predictions show striking inconsistencies during both phases of ARCTAS between observed concentrations of HO2 and of HOx precursors, primarily H2O2 and CH2O. Using observations of precursors in the box model results in predictions of HO2 that are up to nearly a factor of 2 larger than observed. An estimated temperature-dependent terminal loss rate of HO2 to aerosol [Mao et al., 2010] was shown to be insufficient to reconcile model predictions and observations of HO2. When the terminal losses from GEOS-Chem are directly inserted into the fully constrained boxmodel, predictions of upper tropospheric HO2 decrease by no more than 15-25%. Steady state predictions of upper tropospheric CH2O are lower than observations by factors of 2-4 during both phases of ARCTAS. Likewise, steady state predictions of H2O2 are lower than observations by factors of 2-3, and are similar to concentrations measured during TOPSE. Global models suggest that there is an important transport component to the Arctic H2O2 budget not captured by steady state models. An examination of back-trajectories and observations from ARCTAS for in-situ evidence of transport on Arctic peroxide concentrations does not find a widespread persistent signal from transport, although short-lived transient features are evident. Additional possible explanations for the inconsistencies between HO2 and precursor observations are explored. A comparison of calculated and observed OH during the spring (ARCTAS-A) phase shows that median observed-to-calculated ratios are near one, but have large scatter. 40% of OH measurements below 2 km were at the limit of detection (LOD) during the spring, and analysis indicates that the scatter of raw observations at these very low concentrations is larger than the ambient variability of OH, limiting the practicality of further analysis such as finding observational evidence of the impact of halogens on the HO2/OH ratio. Alternately, during ARCTAS-B, model predictions of OH were persistently lower than observations.