Atmospheric carbon gases retrieved from SCIAMACHY by WFM-DOAS: improved global CO and CH 4 and initial verification of CO 2 over Park Falls (46° N, 90° W)

The three carbon gases carbon monoxide (CO), carbon dioxide (CO 2 ), and methane (CH 4 ) are important atmospheric constituents affecting air quality and climate. The nadir spectra of reflected and scattered solar radiation in the near-infrared region, as observed by SCIAMACHY/ENVISAT, contain information on the vertical columns of these gases. A modified DOAS algorithm (WFM-DOAS) has been developed to retrieve this information. The main SCIAMACHY/WFM-DOAS data products are CO vertical columns and dry-air column averaged mixing ratios of methane and CO 2 , denoted XCH 4 and XCO 2 , respectively. For CO and methane we present new results obtained with an improved version of WFM-DOAS (v0.5). The SCIAMACHY data products have been compared with global reference data (MOPITT for CO, TM5 model simulations for XCH 4 ). The comparisons indicate that major problems of the previous version of WFM-DOAS (v0.4x) related to the varying ice-layer on the SCIAMACHY channel 8 detector have been solved. On average, the SCIAMACHY CO agrees within 10% (standard deviation 30%) with MOPITT but regionally, especially over northern South America, large differences have been found (up to about 80%). For methane we present global and regional maps which are compared to TM5 model simulations performed using standard emission inventories. Overall, there is good agreement but regionally there are substantial differences, e.g., due to limitations of current methane emission inventories. It still needs to be assessed by how much emission inventories can be improved by using the SCIAMACHY data. Concerning CO 2 we present a comparison of SCIAMACHY XCO 2 (WFM-DOAS v0.4) with TM3 model simulations over Park Falls, Wisconsin, USA. The peak-to-peak XCO 2 variability as measured by SCIAMACHY (seasonal cycle of year 2003?2005 data) is ~13 ppmv, in good agreement with preliminary analysis of ground-based Fourier Transform Spectrometer (FTS) measurements, which is a factor of 2.3 larger than the XCO 2 variability of TM3 model simulation for 2003. Park Falls is one of the few FTS ground stations which measure column averaged CO 2 and detailed comparison with these measurements (after data release) will help identifying the reason for the observed differences between SCIAMACHY and global (atmospheric) carbon models such as TM3 as reported here and in previous studies. For all three carbon gases we present regional results including seasonal variation focusing on China.