Search

Mean tropospheric hydroxyl radical (OH) abundance is often used as a measure of the oxidation capacity (or “self-cleansing”) of the atmosphere. The primary mechanism by which atmospheric pollutant gases are removed from the atmosphere is initiated by the reaction with OH. As a result, large interannual or decadal variations in OH concentrations, as suggested in recent reports, are of great concern. In addition, an important method for discerning tropospheric OH burdens and variability, the analysis of methyl chloroform (MCF) observations, will soon become less useful as the concentration of this industrial gas approaches zero. With these concerns in mind, a workshop focusing on global OH trends and variability was convened in Boulder, Colorado, on 28–30 November 2005. Although the concept of tropospheric mean OH does not do justice to regional OH differences, and ignores the less significant contributions by other oxidants, global OH changes integrate the response to large-scale atmospheric chemistry forcings. The latter include the tendencies of atmospheric water vapour, solar radiation and notably the human-induced emissions of NOx, CO, CH4 and other hydrocarbons. Analogously, the global mean surface temperature change is an integral climate response to natural and anthropogenic forcings by greenhouse gases, aerosols, etc. Furthermore, the variability of mean OH is an indicator of the sensitivity of atmospheric chemistry to global air pollution and natural events (e.g., large volcano eruptions, El Niño). A large response to a small forcing is typical for a system that is not well buffered and vice versa. The analysis of relatively long-lived gases for which emission magnitudes are well characterized can provide insight into the interannual and decadal variability of tropospheric OH. Analysis of MCF measurements, a tracer with a lifetime of about 5 yr owing to its removal by OH, points to a substantial OH growth in the 1980s, a decline in the 1990s and a recovery after 1998, indicating decadal OH changes of 10–15% (R.G. Prinn et al., 2005, Geophysical Research Letters 32, L07809, doi:10.1029/2004GL022228). MCF analysis furthermore suggests a large interannual OH variability of 8.5±1.0%; however, this may reflect uncertainties in the MCF emission inventory (P. Bousquet et al., 2005, Atmospheric Chemistry and Physics 5, 2635–2656). Using radiocarbon 14CO as a diagnostic for OH gives additional evidence of 10% variability in OH over timescales of less than a year, although the 14CO measurements are only representative of the extratropical southern hemisphere (M.R. Manning et al., 2005, Nature 436, 1001–1004). Even though chemistry–transport models fail to reproduce the large OH variability, many studies point to relatively large OH changes after the 1991 Mt. Pinatubo eruption and during the 1997/8 El Niño event. The likelihood of large OH variability is challenged by CH4 mass balance calculations based on the NOAA network measurements. Emissions of CH4 (E) can be derived from the measured global burden [CH4] and rate of CH4 increase, and an estimate of the CH4 lifetime: E=d[CH4]/dt+[CH4]/τ, where d[CH4]/dt is the observed rate of increase and τ the CH4 lifetime. Since τ is not constant in the real atmosphere, fixing it in the equation means that E includes variability of the sink, i.e. changes in OH. Calculation of E based on yields a mean source of 556±10 Tg CH4 yr−1 for 1984 to 2004, with a trend of 0.1±0.4 Tg yr−1. Maximum deviations from E are 18.4 Tg in 1991 and 27.0 Tg in 1998 (see Fig. 1). Assuming that emissions are uncorrelated with OH, these anomalies provide upper limits of the interannual variability of OH, e.g., 3–5% in 1991 and 1998. In other years the OH variability is typically less than 2%, in agreement with chemistry–transport models. Yet, there is little doubt that global OH decreased immediately after the Mt. Pinatubo eruption, as this is evident in both the CH4 and 14CO measurements. The large anomaly in CH4 in 1998 had contributions by increased emissions from wetlands and biomass burning (S. Morimoto et al., 2006, Geophysical Research Letters 33, L01807, doi:10.1029/2005GL024648), and decreased OH resulting from the Indonesian biomass burning emissions (T.M. Butler et al., 2005, Journal of Geophysical Research 110, D21310, doi:10.1029/2005JD006071).