ETC/ATNI Report 17/2019: Potential use of CAMS modelling results in air quality mapping under ETC/ATNI

The report examines potential use of modelling outputs from the Copernicus Atmospheric Monitoring Service (CAMS) in the air quality annual spatial interpolation mapping. Specifically, CAMS Ensemble Interim Reanalysis and CAMS Ensemble Forecast modelling results have been examined, based on 2017 data.

15 Feb 2021

Prepared by:

Jan Horálek (CHMI), Paul Hamer (NILU), Markéta Schreiberová (CHMI), Augustin Colette (INERIS), Philipp Schneider (NILU), Laure Malherbe (INERIS)

Air quality European-wide annual maps based on the Regression – Interpolation – Merging Mapping (RIMM) data fusion methodology have been regularly produced, using the Air Quality e-Reporting validated (E1a) monitoring data, the EMEP modelling data and other supplementary data. In this report, we examine the use of the preliminary (E2a) monitoring data as provided up-to-date (UTD) by many European countries and as also stored in the Air Quality e-Reporting database, together with the EMEP or the Copernicus Atmospheric Monitoring Service (CAMS) modelling data in two variants (i.e. CAMS Ensemble Interim Reanalysis and CAMS Ensemble Forecast) for potential preparing of preliminary spatial maps. With respect to the availability, the CAMS Ensemble Forecast is the most useful in the potential interim mapping. Such preliminary maps could be constructed approximately one year earlier than the validated maps. Even though we have demonstrated the feasibility, the mapping performance presented in the report is influenced by the lack of the E2a data in some areas.

Next to the evaluation of potential interim maps, regular RIMM maps based on the validated E1a measurement data using three different chemical transport model outputs have been compared, i.e. using the CAMS Ensemble Forecast, the CAMS Ensemble Interim Reanalysis and the EMEP model outputs. Based on the evaluation of the results presented, it is not possible to conclude that any of the three model datasets gives definitively better results compared to the others. The results do not provide strong reasons for a potential change of the model used in the regular mapping.

In addition, the RIMM mapping results have been compared with the CAMS Ensemble Forecast and the CAMS Ensemble Interim Reanalysis outputs. The comparison shows that the data fusion RIMM method gives better results, both in the rural and urban background areas, presumably because of the higher spatial resolution, introduction of additional ancillary data in the data fusion and not fully reduced bias in some data assimilation methods used in CAMS.