Evidence of systematic errors in SCIAMACHY-observed CO₂ due to aerosols

Abstract

Abstract. SCIAMACHY CO2 measurements show a large variability in total column CO2 over the Sahara desert of up to 10% that is not anticipated from in situ measurements and cannot be explained by results of atmospheric models. Comparisons with colocated aerosol measurements by TOMS and MISR over the Sahara indicate that the seasonal variation of SCIAMACHY-observed CO2 strongly resembles seasonal variations of windblown dust. Correlation coefficients of monthly datasets of colocated MISR aerosol optical depth and SCIAMACHY CO2 vary between 0.6 and 0.8, indicating that about half of the CO2 variance is explained by aerosol optical depth. Radiative transfer model calculations confirm the role of dust and can explain the size of the errors. Sensitivity tests suggest that the remaining variance may largely be explained by variations in the vertical distribution of dust. Further calculations for a few typical aerosol classes and a broad range of atmospheric conditions show that the impact of aerosols on SCIAMACHY retrieved CO2 is by far the largest over the Sahara, but may also reach significant levels elsewhere. Inverse modelling calculations indicate that continental scale source and sink estimation on the basis of SCIAMACHY CO2 data without aerosol correction leads to significant errors. To improve terrestrial CO2 flux estimates by inverse modelling using SCIAMACHY measurements at 1.6μm, aerosol correction will be needed. Methods for correcting aerosol-induced errors exist, but so far mainly on the basis of theoretical considerations. As demonstrated by this study, SCIAMACHY may contribute to a verification of such methods using real data.