An Optimal Estimation Retrieval Algorithm for Microwave Humidity Sounding Channels with Minimal Scan Position Bias

• Published in: Journal of atmospheric and oceanic technology 2019-03, Vol.36 (3), p.409-425
• Main Authors: Schulte, Richard M., Kummerow, Christian D.
• Format: Article
• Language: English
• Subjects: Algorithms; Atmosphere; Atmospheric water; Atmospheric water vapor; Bias; Climate change; Clouds; Errors; Estimates; Global positioning systems; GPS; Ground stations; Humidity; Incidence angle; Instruments; Meteorological satellites; Microwave imagery; Microwave radiometers; Oceans; Precipitable water; Radiometers; Retrieval; Satellite constellations; Satellite navigation systems; Satellite tracking; Satellites; Sea surface; Sea surface temperature; Small satellites; Surface temperature; Temperature; Water; Water vapor; Water vapour
• ISSN: 0739-0572; 1520-0426
• DOI: 10.1175/JTECH-D-18-0133.1

A flexible and physical optimal estimation-based inversion algorithm for retrieving atmospheric water vapor and cloud liquid water path from passive microwave radiometers over the global oceans is presented. The algorithm’s main strength lies in its ability to explicitly account for forward model errors that depend on the Earth incidence angle (EIA) at which a given radiometer measurement is made. Validation of total precipitable water (TPW) retrieved from Microwave Humidity Sounder (MHS) measurements against near-coincident estimates of TPW from SuomiNet GPS ground stations shows that retrieved TPW values agree closely with SuomiNet estimates, and somewhat better than values from the Microwave Integrated Retrieval System that are retrieved from the same MHS instruments. More importantly, it is found that the inclusion of appropriate forward model error assumptions, which are tailored to the EIA and sea surface temperature of the scene being considered, are able to almost entirely eliminate EIA-dependent biases in retrieved TPW. This result holds true across all satellites currently carrying an MHS instrument, despite the fact that only measurements from one satellite are used to estimate forward model errors. The consistency achieved by the retrieval algorithm across all view angles suggests that other inversion algorithms, particularly those for cross-track-scanning radiometers and potential future constellations of small satellites, would benefit from the inclusion of nuanced error assumptions that consider the effect of EIA.