Information content of MOPITT CO profile retrievals: Temporal and geographical variability. Issue 24 (17th December 2015)

• Record Type: Journal Article
• Authors: Deeter, M. N.; Edwards, D. P.; Gille, J. C.; Worden, H. M.
• Is Part Of: Journal of geophysical research. Volume 120:Issue 24(2015:Dec.)
• Page Start: 12723
• Page End: 12738
• Publication Date: 2015-12-17
• Subjects: carbon monoxide -- satellite remote sensing -- optimal estimation; Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1002/2015JD024024
• Languages: English
• ISSNs: 2169-897X

Abstract

Abstract: Satellite measurements of tropospheric carbon monoxide (CO) enable a wide array of applications including studies of air quality and pollution transport. The MOPITT (Measurements of Pollution in the Troposphere) instrument on the Earth Observing System Terra platform has been measuring CO concentrations globally since March 2000. As indicated by the Degrees of Freedom for Signal (DFS), the standard metric for trace‐gas retrieval information content, MOPITT retrieval performance varies over a wide range. We show that both instrumental and geophysical effects yield significant geographical and temporal variability in MOPITT DFS values. Instrumental radiance uncertainties, which describe random errors (or "noise") in the calibrated radiances, vary over long time scales (e.g., months to years) and vary between the four detector elements of MOPITT's linear detector array. MOPITT retrieval performance depends on several factors including thermal contrast, fine‐scale variability of surface properties, and CO loading. The relative importance of these various effects is highly variable, as demonstrated by analyses of monthly mean DFS values for the United States and the Amazon Basin. An understanding of the geographical and temporal variability of MOPITT retrieval performance is potentially valuable to data users seeking to limit the influence of the a priori through data filtering. To illustrate, it is demonstrated that calculated regional‐average CO mixing ratios may be improved by excluding observations from a subset of pixels in MOPITT's linear detector array. Key Points: Instrumental and geophysical parameters strongly affect MOPITT retrieval performance DFS variability is analyzed over the CONUS region and the Amazon Basin Maximum DFS values are observed over the Amazon Basin near the end of the burning season