Joint inversion of SO2 and NOx emissions using OMI observations in China

Nitrogen oxides (NO_x) and sulfur dioxide (SO₂) are both precursor gases for PM_2.5. Accurate estimates of their emissions are important for improving air quality in China. To incorporate chemical interactions of SO₂ and NO_x through their reactions with O₃ in top-down emissions estimates, we assimilate NO₂ and SO₂ observations from OMI simultaneously and optimize their emissions at the same time using the GEOS-Chem adjoint model at the nested 0.5° x 0.667° resolution. Optimization of NO_x emissions provides larger corrections for O₃ and OH concentration than that from SO₂ emission optimization, and we therefore expect this joint inversion to be more beneficial in improving SO₂ emissions. We also develop a multi-species mass balance approach and combine it with 4D-Var to estimate long-term SO₂ emissions from 2005 - 2012. In pseudo observation tests with known true emissions (wherein when prior emissions are half the true emissions) the accuracy of posterior NO_x emissions are degraded (e.g., NMSE increases from 1.61 for NO_x single species inversion to 2.54 after incorporating SO₂ observations) due to the larger uncertainties in SO₂ observations than those of NO₂, whereas errors in SO₂ posterior emissions decrease by 52% after incorporating NO₂ observations. Comparisons of surface measurements with GEOS-Chem simulations show slightly reduced error for posterior SO₂ (from -21% in single species inversion to -19% in joint inversion) and NO_x (from -53 % to -53%) in the joint inversion compared to single species inversion. We further estimate long-term SO₂ and NO_x emissions using the joint inversion framework and compare their seasonality and inter-annual variation with emissions constrained by single species observations alone. Impact of the changes in SO₂ and NO_x emissions and meteorology on PM_2.5 concentrations will also be discussed.