Harmony-TIR: Production and Assessment of Cloud-top Winds and Heights from GEO-GEO and SLSTR Stereo Retrieval Using the MISR Approach
This study introduces a framework of retrieving cloud-top winds and heights from stereo EO measurements. This framework uses the M2 (Multipoint Matcher) method, which was developed for the NASA MISR instrument onboard the EOS-Terra platform, to derive pixel-level disparities from stereo measurements. The disparities can then be converted to cloud-top heights using the observational geometries, and can also be converted to horizontal wind fields using the timestamps from consecutive stereo measurements. This framework also employs the approach described in, which uses the decreasing rate of local brightness temperature as a proxy, to derive the vertical wind velocities after co-registration of the cloud-field in a moving frame-of-reference.
This framework is first applied to retrieve cloud-top winds and heights using inputs from two different large-eddy simulation (LES) simulations, which are described in a companion paper (Muller et al., this workshop). This retrieval framework was then applied to retrieve cloud-top winds and heights from real satellite observations: 1) GOES-17 and GOES-16 stereo measurements, and 2) Sea and land Surface Temperature Radiometer (SLSTR) stereo measurements. The GOES-GOES stereo was first applied to Hurricane Iselle over the Gulf of Mexico on 26-Aug-20 @15:10 UTZ. The experimental results demonstrate that cloud-top heights are retrieved from band-13 of GOES-GOES stereo at 2-km resolution, and cloud-top winds can be retrieved every 2 minutes from both band-2 at 500m and band -13 at 2km resolution. The GOES-GOES stereo was then applied to very low Marine Atmospheric Boundary Layer (MABL) type clouds over the Peruvian/Humboldt cold current, to retrieve the cloud-top heights and cloud-top winds within a ten-minute interval. The SLSTR instruments are part of the payloads of Sentinel-3a launched in February 2016 and Sentinel-3b in April 2018 respectively. In this study, this retrieval framework is used to derive low-level cloud-top heights from nadir and oblique aft view of SLSTR 1km Band S8 (10.85µm) data over 1) the entire coastline of the Peruvian coast, and 2) the entire length of the Californian coast. The implications of these results for the candidate ESA Harmony-TIR mission will be discussed as well as for future production of SMVs from GEO-GEO intersections.