Tropospheric CO vertical profiles deduced from total columns using data assimilation: methodology and validation.

This paper presents a validation of a method to derive the vertical profile of carbon monoxide (CO) from its total column using data assimilation. The main motivation of this study is twofold. First, to deduce both the vertical CO profiles and the assimilated CO fields with good confidence. Second, for chemical species that can be measured only as the total column, this method provides an attractive alternative for estimating their vertical profiles in the troposphere. We choose version 3 (V3) of MOPITT CO total columns to validate the proposed method. MOPITT has the advantage of providing both the vertical profiles and the total columns of CO. Furthermore, this version has been extensively validated by comparison with many independent datasets, and has been used in many scientific studies. The first step of the paper consists in the specification of the observation errors based on the Chi-square (χ²) test. The observations have been binned according to day, night, land and sea (LAND_DAY, LAND_NIGHT and SEA, respectively). The respective optimal observation error values for which the χ² metric is the closest to 1 are: 7%, 8% and 11% for SEA, LAND_DAY and LAND_NIGHT, respectively. In a second step, the CO total column, with its specified errors, is used within the assimilation system to estimate the vertical profiles. These are validated by comparison with vertical profiles of MOPITT V3 retrievals at global and regional scales. Generally, both datasets show similar patterns and good agreement at both global and regional scales. Nevertheless, the total column analyses (TOTCOL_ANALYSES) slightly overestimate CO concentrations compared to MOPITT observations. In a third step, vertical profiles calculated from TOTCOL_ANALYSES have been compared to those calculated from the assimilation of MOPITT V3 vertical profiles (PROFILE_ANALYSES). Both datasets shows very good agreement, but TOTCOL_ANALYSES tend to slightly overestimate CO concentrations. The mean bias between both datasets is 6% and 8% at the pressure levels 700 and 200 hPa, respectively. In terms of zonal means, the CO distribution is similar for both analyses. The mean bias between these datasets is low and doesn't exceed 12 %. These results confirm that both analyses (total column and vertical profiles) are in very good agreement at global and regional scales. [ABSTRACT FROM AUTHOR]