1. Retrieval of Vertical Columns of Sulfur Dioxide from SCIAMACHY and OMI: Air Mass Factor Algorithm Development, Validation, and Error Analysis Chulkyu Lee 1, Randall V. Martin 1,2, Aaron van Donkelaar 1, Gray O’Byrne 1, Nickolay Krotkov 3, Andreas Richter 4, Greg Huey 5, and John S. Holloway 6 1 Department of Physics and Atmospheric Science, Dalhousie Univ., Canada; 2 Harvard-Smithsonian Center for Astrophysics, USA, 3 Goddard Earth Sciences and Technology Center, Univ. of Maryland, USA, 4 Institute of Environmental Physics and Remote Sensing, Univ. of Bremen, Germany, 5 School of Earth and Atmospheric Sciences, Georgia Institute of Technology, USA, 6 NOAA Earth System Research Laboratory, USA We develop an improved retrieval of sulfur dioxide (SO 2 ) vertical columns from two satellite instruments (SCIAMACHY and OMI) that measure ultraviolet solar backscatter. For each SCIAMACHY and OMI observation, a local air mass factor (AMF) algorithm converts line-of-sight “slant” columns to vertical columns using altitude-dependent scattering weights computed with a radiative transfer model (LIDORT), weighted by relative vertical SO 2 profile (shape factor) determined locally with a global atmospheric chemistry model (GEOS-Chem). The scattering weights account for viewing geometry, surface albedo, cloud scattering, absorption by ozone, and scattering and absorption by aerosols. Mean SO 2 shape factors simulated with GEOS-Chem and used in the AMF calculation are highly consistent with airborne in situ measurements (INTEX-A and INTEX-B); differences would affect the retrieved SO 2 columns by 10%. The retrieved vertical columns are validated with coincident airborne in situ measurements (INTEX-A, INTEX-B, and a campaign over East China). The annual mean AMF errors are estimated to be 35-70% in polluted regions (e.g., East Asia and the eastern US) and less than 10% over clear ocean regions. The overall SO 2 error assessment is 45-80% for yearly averages over polluted regions. Seasonal mean SO 2 columns retrieved from SCIAMACHY and OMI for 2006 are significantly spatially correlated with those from GEOS-Chem, in particular over the United States (r = 0.85 for SCIAMACHY and 0.82 for OMI). A sensitivity study confirms the sensitivity of SCIAMACHY and OMI to anthropogenic SO 2 emissions. ABSTRACT Scattering weight IoIo IBIB Earth Surface Radiative Transfer Model (LIDORT)  (η) is temperature-dependent cross-section Calculate ω(η) as function of: solar and viewing zenith angle (θ s, θ v ) surface reflectivity pressure, aerosol clouds O 3 column: For individual scenes Shape factor SO 2 mixing ratio, C SO2 (η) Atmospheric Chemistry Model (GEOS-Chem) Local AMF Calculation eta ( η ) dt(η) Cloud Radiance Fraction < 0.2; SZA < 70 GC at OMI overpass within 10% of GC at SCIAMACHY overpass GEOS-Chem SO 2 changes by ×2 OMI SO 2 AMF changes < 30%  Local AMFs improve agreement of OMI and SCIAMACHY with in-situ  Validation of SO 2 vertical columns from SCIAMACHY and OMI with airborne in-situ measurements for INTEX-A and B (r = 0.9)  Validation of GEOS-Chem SO 2 shape factors with airborne in-situ measurements for INTEX-A and B (<10% change in AMF)  Large signal from anthropogenic emissions in retrieved SO 2 columns Evaluation of GEOS-Chem Simulation of Shape Factor Slant Columns with Reference Sector Method Seasonal Air Mass Factors for 2006 Seasonal Mean Vertical Columns for 2006  Calculated at 313.2 nm  Cloud Radiance Fraction < 0.2; SZA < 70  SCIAMACHY AMF at 319.7 nm within 25% Local AMF Increases Agreement with INTEX Aircraft Measurements Summary Sensitivity of Retrieved SO 2 to ΔEmissions Differences between modeled and measured shape factors would change AMF by <10% Scatter plots of tropospheric SO2 vertical columns versus those from in-situ measurements during INTEX-A (triangles) and INTEX-B (squares). The constant AMF refers to the OMI operational PBL product. The local AMF is developed here. Sensitivity of SO2 columns to errors in anthropogenic emissions. The top panel is the difference between GEOS-Chem simulations of SO2 columns using meteorology for 2006 with anthropogenic emissions for 2006 minus one with emissions for 1998. The bottom panel is the difference between OMI SO2 columns for 2006 retrieved using SO2 shape factors from the GEOS-Chem simulations with anthropogenic SO2 emissions for either 2006 or 1998. Funding: NSERC and NASA