A unique approach to testing quality of cloud retrievals

Abstract

The EarthCARE satellite mission’s objective is to retrieve profiles of aerosol and water cloud physical properties from measurements made by its cloud-profiling radar, backscattering lidar, and passive multi-spectral spectral imager (MSI). These retrievals, together with other geophysical properties, are input into broadband (BB) radiative transfer (RT) models that predict radiances, and fluxes, commensurate with measurements made, and inferred from, EarthCARE’s BB radiometer (BBR). The scientific goal is that modelled and “observed” BB fluxes differ, on average, by less than ±10 W m-2. When sound synergistic retrievals from the ACM-CAP process are available, they are acted on by the RT models. When they are not available, the RT models act on “composite” atmospheric profiles of retrievals from individual sensors. “Compositing” is performed in the ACM-COM process as described in this report. The majority of this report describes the RT models, and their products, that make-up Earth-CARE’s ACM-RT process. Shortwave (SW) and longwave (LW) flux and heating rate (HR) profiles are computed by 1D RT models for each ~1 km nadir column of inferred properties. 3D RT models compute radiances for the BBR’s three viewing directions, with the SW model also computing flux and HR profiles; the 3D LW model produces upwelling flux at just one level. All 3D RT products are averages over 5 x 21 km “assessment domains” that are constructed using MSI data. A subset of ACM-RT’s products is passed forward to the “radiative closure assess-ment” process that quantifies, for each assessment domain, the likelihood that EarthCARE’s goal has been achieved. As EarthCARE represents the first mission to make “operational” use of 3D RT models, emphasis in this report is placed on differences between 1D and 3D RT results. For upwelling SW flux at 20 km altitude, 1D and 3D values can be expected to differ by more than EarthCARE’s scientific goal of ±10 W m-2 at least 50 % of the time.