Advances in representing interactive methane in ModelE2-YIBs (version 1.1)

Abstract

Abstract. Methane (CH4) is both a greenhouse gas and a precursor of tropospheric ozone, making it an important focus of chemistry–climate interactions. Methane has both anthropogenic and natural emission sources, and reaction with the atmosphere's principal oxidizing agent, the hydroxyl radical (OH), is the dominant tropospheric loss process of methane. The tight coupling between methane and OH abundances drives indirect linkages between methane and other short-lived air pollutants and prompts the use of interactive methane chemistry in global chemistry–climate modeling. In this study, an updated contemporary inventory of natural methane emissions and the soil sink is developed using an optimization procedure that applies published emissions data to the NASA GISS ModelE2-Yale Interactive terrestrial Biosphere (ModelE2-YIBs) global chemistry–climate model. Methane observations from the global surface air-sampling network of the Earth System Research Laboratory (ESRL) of the US National Oceanic and Atmospheric Administration (NOAA) are used to guide refinement of the natural methane inventory. The wetland methane flux is calculated as a best fit; thus, the accuracy of this derived flux assumes accurate simulation of methane chemical loss in the atmosphere and accurate prescription of the other methane fluxes (anthropogenic and natural). The optimization process indicates global annual wetland methane emissions of 140 Tg CH4 yr−1. The updated inventory includes total global annual methane emissions from natural sources of 181 Tg CH4 yr−1 and a global annual methane soil sink of 60 Tg CH4 yr−1. An interactive methane simulation is run using ModelE2-YIBs, applying dynamic methane emissions and the updated natural methane emissions inventory that results from the optimization process. The simulated methane chemical lifetime of 10.4±0.1 years corresponds well to observed lifetimes. The simulated year 2005 global-mean surface methane concentration is 1.1 % higher than the observed value from the NOAA ESRL measurements. Comparison of the simulated atmospheric methane distribution with the NOAA ESRL surface observations at 50 measurement locations finds that the simulated annual methane mixing ratio is within 1 % (i.e., +1 % to −1 %) of the observed value at 76 % of locations. Considering the 50 stations, the mean relative difference between the simulated and observed annual methane mixing ratio is a model overestimate of only 0.5 %. Comparison of simulated annual column-averaged methane concentrations with SCIAMACHY satellite retrievals provides an independent post-optimization evaluation of modeled methane. The comparison finds a slight model underestimate in 95 % of grid cells, suggesting that the applied methane source in the model is slightly underestimated or the model's methane sink strength is slightly too strong outside of the surface layer. Overall, the strong agreement between simulated and observed methane lifetimes and concentrations indicates that the ModelE2-YIBs chemistry–climate model is able to capture the principal processes that control atmospheric methane.