摘要：A state-of-the-art inverse model, CarbonTracker Data Assimilation Shell (CTDAS), was used to optimize estimates of methane (CH4) surface fluxes using atmospheric observations of CH4 as a constraint. The model consists of the latest version of the TM5 atmospheric chemistry-transport model and an ensemble Kalman filter based data assimilation system. The model was constrained by atmospheric methane surface concentrations, obtained from the World Data Centre for Greenhouse Gases (WDCGG). Prior methane emissions were specified for five sources: biosphere, anthropogenic, fire, termites and ocean, of which biosphere and anthropogenic emissions were optimized. Atmospheric CH4 mole fractions for 2007 from northern Finland calculated from prior and optimized emissions were compared with observations. It was found that the root mean squared errors of the posterior estimates were more than halved. Furthermore, inclusion of NOAA observations of CH4 from weekly discrete air samples collected at Pallas improved agreement between posterior CH4 mole fraction estimates and continuous observations, and resulted in reducing optimized biosphere emissions and their uncertainties in northern Finland.
