There are many ways EarthCARE can be used to evaluate atmospheric models for both weather and climate prediction, providing insights into how the models can be improved. Direct evaluation of individual weather systems can be carried out both in model space, in which retrieved atmospheric properties are compared to model variables, or in observation space, in which model fields are used to calculate observables that are then compared to the actual observations. EarthCARE observations can also be used to derive relationships between quantities that can be used to improve parameterizations, for example between ice cloud effective radius and temperature. All these approaches will be pursued further in the EarthCARE-ORCESTRA Model Intercomparison Project (ECOMIP).
Here we provide an initial taster of some early comparisons of retrieved cloud, precipitation and aerosol properties with values from the ECMWF model.
The figure below compares total cloud water content and precipitation rate between EarthCARE's ACM-CAP synergy retrieval and the 9-km resolution operational ECMWF forecast, for the same 18 September 2024 case that was considered here, which spans from Sweden on the left to the Gulf of Guinea on the right. We see that the model predicts the rain cloud over Italy at 45°N well, but in parts of the scene it appears to predict clouds too frequently, and the large thunderstorm at around 11°N has been misplaced. We also see a tendency for the stratocumulus cloud on the right of the figure to be raining too frequently, a problem that has been identified previously in several global models including ECMWF.
The high-spectral resolution capability of ATLID enables cloud and aerosol extinction coefficient to be retrieved unambiguously, without the need to make assumptions on the scattering properties of the particles. Extinction coefficient is the most important radiative property of a cloud or aerosol field, and can be thought of as the range-resolved optical depth. The top panel of the figure below shows aerosol extinction coefficient at the wavelength of the lidar (355-nm) as retrieved by EarthCARE's A-EBD product, for the same case as considered above. The target classification information has been used to screen out only the aerosol signal. Three very different aerosol types are present in this scene: the left third of the image consists of largely of European continental aerosol, the central third is Saharan dust and the right third is smoke originating from biomass burning in central Africa.
The bottom panel in the figure above shows the corresponding 355-nm aerosol extinction coefficient from the ECMWF air quality forecasts produced as part of the Copernicus Atmosphere Monitoring Service (CAMS). This version of the model is run with 40-km resolution and has around 15 prognostic aerosol species, in addition to around 100 gases including those that are precursors to aerosol formation.
A quantitative comparison is shown below: profiles of extinction coefficient at three latitudes where different aerosol species dominate. The left panel shows that in this case the extinction of European aerosol is underestimated. By contrast, the middle panel shows that both the shape and the magnitude of the profile through desert dust is very well captured. The right panel shows that the biomass burning plume advected from central Africal is also rather well captured. In this case ATLID cannot measure aerosol extinction inside the stratocumulus cloud below around 1.8 km, but in fact this case illustrates the power of high spectral resolution lidar: we can accurately retrieve the optical depth of aerosol above cloud, something that is impossible with MODIS or AERONET.
The EarthCARE-ORCESTRA Model Intercomparison Project (ECOMIP) is coordinating the evaluation of many models using EarthCARE
A preliminary radiative closure assessment has been performed for this case