Investigating the performance of AEOLUS L2A products over Europe with EARLINET ground-based lidars

<p>In this study, we present a comparison of the AEOLUS satellite L2A product with the retrievals of the ground-based lidar systems of EARLINET (European Aerosol Research Lidar Network), part the European Research Infrastructure for the observation of Aerosol, Clouds and Trace Gases (ACTRIS). Dedicated ground‐based measurements during AEOLUS overpasses have been performed among the 29 member stations since the beginning of the mission, however, we have included only the stations that have gathered a significant number of collocations in the analysis. The satellite timeseries we deployed covers the period 2019-2020 that correspond to the best available version of the satellite processing algorithms. We harvest the collocations using the following spacio-temporal criteria. Only overpasses that fall within a radius less than 100km around the station are included. Using this criterion, the AEOLUS L2A climatology is generated per station independently of the ground-based measurements. To isolate collocated data we reject all AEOLUS data with a time interval between the overpass and the central time of the ground-based measurement that is greater than 3 hours. The ground based lidar climatology is also computed per station. AEOLUS L2A products include aerosol extinction coefficient profiles and aerosol co-polar backscatter coefficient profiles from circularly polarized light emission. While the extinction profiles are directly comparable with the ground-based lidars, this is not the case for the backscatter profiles since AEOLUS cannot measure the cross polar component of the aerosol backscatter. The co-polar backscatter is close to the total backscatter only in the absence of depolarizing scatterers such as dust, pollen, volcanic ash, and cirrus ice crystals. Ground-based measurements are divided in two categories for the evaluation depending on whether aerosol depolarization measurements have been performed. If the particle linear depolarization ratio (PLDR) is available, it can be applied to convert the lidar total backscatter to an AOLUS-like co-polar backscatter coefficient. This category is applied for the direct evaluation of the satellite product. Cases that lack PLDR information assist to quantify the uncertainties introduced by using the AEOLUS co-polar backscatter as a substitute for the total backscatter. The analysis includes both an indirect climatological comparison and a direct collocation comparison between the ground based and satellite datasets. Via the collocation comparison, random and systematic uncertainties in the satellite product are identified and quantified. A climatological comparison can show the potential of AEOLUS to capture annual cycles despite its intrinsic random errors. In the future, the analysis will be further supported with auxiliary data such as sunphotometer measurements, aerosol classification flags, modeled backward trajectories, and satellite cloud fraction data.</p>