Long-term evaluation of surface air pollution in CAMSRA and MERRA-2 global reanalyses over Europe

<p>In the last two decades reanalyses have become a powerful tool for modern geosciences as they combine both model- and observation-based (mostly from remote-sensing sources) information to provide physically-consistent data of land, ocean and atmospheric fields with continuous spatial and temporal coverage. In the frame of the ERC project EARLY-ADAPT (https://early-adapt.eu/), a pioneer health dataset is currently being collected over Europe to investigate the time-varying health effects of climate and air pollution, which will shed light into the early adaptation response to climate change in the area of human health. This impact will be quantified by fitting epidemiological models on historical local health, climate and air pollution data, which thus requires a long-term air quality database at daily-scale over all of Europe. In this context, atmospheric composition reanalyses provide highly valuable information, but remain subject to biases and errors both in terms of spatiotemporal variability and long-term trends. It is therefore key to determine whether these reanalyses correctly capture the values, trends and cyclic processes of the different aforementioned fields.</p><p>Our work aims to evaluate how the Copernicus Atmosphere Monitoring Service global reanalysis (CAMSRA), developed by the ECMWF, and the Modern-Era Retrospective Analysis for Research and Applications v2 (MERRA-2), produced by NASA, perform against independent ground-level in-situ observations over Europe for a period of 18 years (2003 &#8211; 2020). We analyse these reanalyses products considering the most harmful pollutants for human health, namely O₃, NO₂, SO₂, CO, PM_2.5 and PM₁₀. A careful quality-assurance filtering of the surface observations is performed using GHOST, which stands for Globally Harmonised Observational Surface Treatment, a BSC in-house project dedicated to the harmonisation of global air pollution surface observations and its metadata, with the purpose of facilitating a greater quality of observational/model comparison in the atmospheric chemistry community. This study considers a domain extending from 25&#176;W to 45&#176;E in longitude, and from 27&#176;N to 72&#176;N in latitude, thus covering all continental Europe as well as the Canary Islands, Iceland, Western/European Russia, North Africa and the westernmost regions of the Middle East and the Caucasus.</p><p>CAMSRA and MERRA-2 reproduce the observational values, trends and seasonal cycles with a varying degree of accuracy, depending on the pollutant considered, though significant and persistent biases are found in almost all cases. As the computed statistics present strong spatiotemporal dependencies, given the long-term scope of the evaluation, a regional and country-level analysis has been performed in order to provide a more exhaustive and complete evaluation. An intercomparison between CAMSRA and MERRA-2 has also been conducted for the pollutants available in both reanalyses. The obtained results highlight the necessity of applying bias correction schemes when working with air pollution reanalysis data, and open the door for improved continental-wide, regional-scale, environmental epidemiological analyses of the health impacts of air pollutants.</p>