The Measurements of Pollution in the Troposphere (MOPITT) satellite instrument has measured tropospheric CO continuously since March 2000, providing the longest continuous dataset of CO from space. During its long mission, data processing has been developed and updated to improve the quality of CO retrievals and the sensitivity to lower troposphere. It has a global successful retrieval rate that is about 30%. The spatial seasonal variations demonstrated that while the data coverage in some places reaches 30% in summer, this number drops to less than 10% in winter due to significant cloud cover. Therefore, we investigate the current MOPITT cloud detection algorithm and consider approaches to enhance the data coverage.

The MOPITT CO total column (TC) data were modified by turning off the cloud detection scheme which means all MOPITT data were used to produce retrievals regardless of the cloud detection. Analysis of the CO TC standard (cloud filtered) and non-standard product (non-cloud masked) are conducted for selected days. Results showed some coherent structures that were observed frequently in the non-masked CO that were not present in the standard product and could be CO features. An analysis was conducted using Moderate Resolution Imaging Spectroradiometer (MODIS) cloud height and cloud mask along with MOPITT cloud flag description in order to explore the situation behind those observations. Results illustrated that, a significant number of low cloud cases were rejected in the standard product. Those missing areas match the coherent patterns that were detected in the non-masked CO product. An analysis of IASI CO TC indicated an evidence of the strong CO features that are observed in the non-cloud-masked product. The use of the radiance ratio threshold test as an independent method for MOPITT cloud detection in the future V9 will significantly increase the number low cloud cases over land.