A Multi-Component Proxy for OH Variability Measured from Space: Evaluation and Evolution

Abstract:

Observed changes in the decay rate of methyl chloroform (MCF; CH₃CCl₃) have been the traditional top-down constraint for variability in tropospheric abundances of the hydroxyl radical (OH), and thereby the oxidative capacity of the atmosphere. However, as atmospheric MCF concentrations approach detection limits following its ban under the Montréal Protocol, it is necessary to identify new proxies for global OH and its variability. We present here a novel proxy for tropospheric OH using convolved observations of total and tropospheric columns of ozone, water vapor, NO₂, and CO, available from the Aura and Aqua satellites in the NASA Earth Observing System constellation. Derived from photochemical steady-state assumptions, the satellite proxy generates spatiotemporally coherent monthly percent anomalies in column OH for Oct. 2004 through Sept. 2012. We demonstrate that the temporal evolution of the globally integrated signal is statistically consistent with changes inferred by the MCF decay rate during this period. The magnitude of interannual variability in the satellite OH proxy is smaller than that from MCF, more consistent with global chemical transport models (CTMs). We evaluate the proxy by comparing to 9-year hind-cast simulations of the GEOS-Chem global CTM driven by MERRA reanalysis meteorology at 2° x 2.5° horizontal resolution, and find that the satellite proxy correlates with the airmass-weighted tropospheric monthly mean OH anomalies (R = 0.6; n = 96 months). The satellite proxy and model simulations indicate regional-scale coherent OH anomalies for many regions and seasons; where discrepancies occur, we highlight possible causes. We present a source attribution of the spatial and temporal patterns in the satellite OH proxy, informed by the individual components of the satellite proxy, a series of zero-emission sensitivity simulations with GEOS-Chem, and independent proxies for major modes of climate variability.