Evaluation of satellite-derived HCHO using statistical methods
1Department of Atmospheric Science, Pusan National University, Korea
2Department of Atmospheric Science, University of Alabama in Huntsville, USA
3Harvard-Smithsonian Center for Astrophysics, Harvard University, USA
4Belgian Institute for Space Aeronomy (BIRA-IASB), Brussels, Belgium
Abstract. Most previous evaluations of satellite performance relied on pair wise comparisons with limited spatial and temporal coverage of surface-based measurements such as soundings and in-situ measurements. Especially, validation of satellite HCHO measurements is very difficult because ground-based HCHO measurements are extremely sparse. We use a new scientific approach, statistical analyses with empirical orthogonal function (EOF) and singular value decomposition (SVD), to provide three-dimensional results of comparisons with a global picture over long measurement periods. The EOF and SVD analyses with GOME, SCIAMACHY and OMI HCHO, and MOPITT CO show dipole distributions oscillating between northern and southern equatorial Africa with an annual cycle. This feature is exactly coincident with the spatial and temporal pattern of biomass burning occurring over tropical Africa. The double-peaked maximum seen in OMI HCHO is only marginally observed in SCIAMACHY and GOME HCHO seasonality over the northern tropical region during northern biomass-burning season. Spatial and temporal difference between two datasets may cause this discrepancy, but the detailed analysis for the cause requires an examination with a chemical model. The statistical analyses of all data indicate that biomass-burning activity over South America is responsible for the HCHO seasonality over that continent. We have not observed any evidence to support the influence of biogenic activity on HCHO over these regions; however, we find robust evidence that biomass burning is the strongest source of HCHO over tropical Africa and South America. We also found that these statistical tools are a very efficient method for evaluating satellite data.