Validation of MOPITT CO Mixing Ratio and Column Retrievals

In order to make a quantitative validation of the retrieved CO mixing ratios, the averaging kernels and a priori data must be used to transform the high resolution in situ validation data.  This is discussed further on the Data page .

Comparisons have been made to aircraft profiles measured regularly at 5 sites (in Hawaii, Alaska, Massachusetts, Rarotonga and Colorado) by Paul Novelli, NOAA/CMDL. Aircraft profiles were also made available from a number of campaigns, including MOVE, SAFARI-2000, TRACE-P, TOPSE, BIBLE-C, and COBRA. The following figures summarize the agreement between MOPITT CO and the aircraft measurements for the first year of data (Phase 1: Mar 2000-May 2001).

Correlations between MOPITT CO profile and column retrievals and the in situ measurements (after taking into account the averaging kernels)

The bias (MOPITT - insitu) in % for the retrieved profiles at each site.

The bias (MOPITT - insitu) in % for the column amounts at each site.

All of the biases for each retrieval level and the column have been averaged to give an overall bias with uncertainty for each L2 product, and are summarized in the following table.

Retrieval Product	Absolute Bias	Bias (%)
Surface	9.2 +/- 21.6 ppbv	11.1 +/- 28.7
850 mbar	9.7 +/- 18.3 ppbv	11.2 +/- 19.6
700 mbar	7.9 +/- 14.2 ppbv	9.0 +/- 14.3
500 mbar	4.4 +/- 12.3 ppbv	5.4 +/- 11.9
350 mbar	2.6 +/- 14.3 ppbv	3.8 +/- 13.6
250 mbar	0.5 +/- 13.3 ppbv	1.9 +/- 13.7
150 mbar	-0.5 +/- 10.9 ppbv	0.5+/- 14.7
Column	0.95 +/- 1.99 molec/cm2	6.2 +/- 11.4

Presentations and Publications

Emmons, L., Validation of MOPITT Retrievals of Carbon Monoxide, IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Toronto, June 24-28, 2002. Poster, Extended abstract

Emmons, L., The Distribution of Tropospheric Carbon Monoxide Observed by MOPITT, IGAC Conference, Sep 18-25, 2002. (pdf)