Investigation of the Discrepancies in Cloud Top/Base Height Retrievals Between Geostationary and Polar-Orbiting Satellites

This study investigates the discrepancies in cloud top height (CTH) and cloud base height (CBH) retrievals between five geostationary (GEO) satellites (GOES-16, GOES-18, Himawari-9, Meteosat-9, and Meteosat-10) and the polar-orbiting Suomi NPP (SNPP) satellite.

The CTH and CBH data come from a parallax-corrected database that collocates GEO and SNPP satellite cloud data. The cloud retrievals are produced for the individual sensors using the Clouds for AVHRR Extended (CLAVR-x) processing system, a development tool for many NOAA Enterprise cloud retrieval algorithms, and processed onto a global grid of 7000 x 18000, spanning latitudes of -70° to 70° as part of global 3D cloud data that we are developing. The dataset covers the period from November 27, 2023 to April 28, 2024. The differences in CTH and CBH between the two satellite types were calculated by subtracting the SNPP measurements from the GEO measurements. If data is missing from one or both satellites, then that difference is not considered in the calculation.

The initial analysis averaged these differences over three time periods: the entire dataset timeframe, the December-January-February (DJF) season, and the March-April (partial spring) season. Results indicate that the highest positive differences (>5500 m) in CTH and CBH occur near the equator, particularly in the Indian Ocean and Western Pacific regions, and are most pronounced during the partial spring period. Histograms reveal a slight positive shift in CBH differences compared to CTH differences across all regions and timeframes, especially notable in the Antarctic region, where a ring of relatively high positive CBH differences appears at the data cutoff of -70°.

Next, a case study focused on the Indian Ocean was conducted to further explore the large positive discrepancies in CTH and CBH. The results show that GEO satellites recorded CTH/CBH up to 15,000 meters higher than those from SNPP in this region. Further analysis of the meteorological conditions that may lead to these types of discrepancies is underway.

The impact of scan and time differences between GEO and SNPP will be examined by using a subset of cases featuring closely matched sensor zenith angles, which can be used to isolate systematic differences between retrieval for different sensors. Ongoing research and further case studies will aim to identify the root causes of these CTH/CBH differences between GEO and polar-orbiting satellite cloud measurements and allow us to accommodate the associated uncertainties in blended, multi-sensor products.