PDF Version

Evaluation of MOPITT and TROPOMI Carbon Monoxide Retrievals using AirCore in Situ vertical Profiles

S. Martínez-Alonso1, I. Aben2,3, B. Baier4,5, T. Borsdorff2, M. Deeter1, K. McKain4,5, C. Sweeney5 and H. Worden1

1National Center for Atmospheric Research (NCAR), Atmospheric Chemistry Observations and Modeling Laboratory, Boulder, CO 80307; 303-497-1862, E-mail: sma@ucar.edu
2Space Research Organization Netherlands (SRON), Netherlands Institute for Space Research, Utrecht, The Netherlands
3Vrije Universiteit Amsterdam, Department of Earth Sciences, Amsterdam, The Netherlands
4Cooperative Institute for Research in Environmental Sciences (CIRES), University of Colorado, Boulder, CO 80309
5NOAA Global Monitoring Laboratory (GML), Boulder, CO 80305

AirCore in situ vertical profiles sample the atmosphere from near the surface to the lower stratosphere, making them ideal for the validation of satellite tropospheric trace gas data. Here we present intercomparison results of AirCore carbon monoxide (CO) measurements with respect to retrievals from MOPITT (Measurements of Pollution In The Troposphere; version 8) and TROPOMI (TROPOspheric Monitoring Instrument), onboard the NASA Terra and ESA Sentinel 5-Precursor satellites, respectively. Mean MOPITT/AirCore total column bias values and their standard deviation (0.0±0.9, 0.3±1.0, and 0.1±1.0 for MOPITT thermal-infrared, near-infrared, and multispectral retrievals, respectively; all in units of 1017 molec cm-2) are similar to results obtained in MOPITT/NOAA aircraft flask data comparisons from this study and from previous validation efforts. MOPITT CO retrievals are systematically validated using in situ vertical profiles from a variety of aircraft campaigns. Because most aircraft vertical profiles do not sample the troposphere’s entire vertical extent, they must be extended upwards in order to be usable in validation. Here we quantify for the first time the error introduced in MOPITT CO validation by the use of shorter aircraft vertical profiles extended upwards by analyzing validation results from AirCore CO vertical profiles. Our results indicate that the error is small, affects mostly upper tropospheric retrievals (at 300 hPa: ~2.6, 0.8, and 3.2 percent points for MOPITT thermal-infrared, near-infrared, and multispectral, respectively), and may have resulted in the overestimation of MOPITT retrieval biases in that region. TROPOMI can retrieve CO under both clear and cloudy conditions. The latter is achieved by quantifying interfering trace gases and parameters describing the cloud contamination of the measurements together with the CO column; then, the reference CO profiles used in the retrieval are scaled based on estimated above-cloud CO rather than on estimated total CO. We use AirCore measurements to evaluate the error introduced by this approach in cloudy TROPOMI retrievals over land after accounting for TROPOMI’s vertical sensitivity to CO (relative bias and its standard deviation = 2.02 % ± 11.13 %). We also quantify the null-space error, which accounts for differences between the shape of TROPOMI reference profiles and that of AirCore true profiles (for TROPOMI cloudy enull = 0.98 % ± 2.32 %).

Figure 1

Figure 1. CO biases for the 2012-2019 period from the comparison of MOPITT with respect to NOAA aircraft flask data (blue), AirCore profiles (pink), and truncated AirCore profiles extended upwards (purple). (a) For MOPITT TIR. (b) For MOPITT NIR. (c) For MOPITT multispectral. Relative bias in %. Column bias in units of 1017 molec cm-2. The ±10 % CO bias range is equal to the MOPITT target accuracy.