Air Quality in Europe: Daily, Monthly, and Annual Maps of NO2, O3, PM10, and PM2.5 between 2015 and 2023

Introduction This repository contains spatial datasets describing air quality in Europe for different temporal aggregation levels between 2015 and 2023. Air quality is specifically expressed as atmospheric concentration of nitrogen-dioxide (NO2), ozone (O3), and particulate matter 10 (PM10) and 2.5 (PM2.5). All maps are results from the Open Earth Monitor & Cyberinfrastructure (OEMC) project and its use case on Air Quality Monitoring at Continental Scale. Stakeholder in this use case was the European Environmental Agency's Topic Center on Human Health and the Environment (EEA-HE), who produces annual air quality reports for the EEA member states. The objective of the use case was to implement the stakeholder's mapping method using open source software and enable transparent mapping at different (higher) temporal resolutions. View all air quality maps interactively in the OEMC app Details Regression Interpolation Merging Mapping (RIMM) Map predictions were performed using the RIMM method, which was develeped by EEA-HE. It leverages official EEA measurement station data and covariates related to terrain, weather, or land cover in a universal kriging approach. Map values represent the sum of two components: the prediction from a linear regression model and kriging output fom model residuals. All maps come with an uncertainty estimate defined as the sum of standard errors of the two components. This approach is applied for separate groups of air quality measurement stations based on their location: rural vs. urban, background vs. traffic. Maps from all groups are unified to a final product by applying a weighted average based on spatial weight layers for urban character (population density) and traffic (road network density). For more detail see Horálek et al., 2023. Input Data & Code The EEA measurement station data was preprocessed to meet the needs of downstream modeling tasks and was published seperately on Zenodo. Code to retrieve and preprocesses station data, covariates and to apply the RIMM method can by found in the Github repository. Results The data in this repository is structured by air quality indicator and temporal resolution. Monthly and annual maps use daily aggregates of indicators, which may vary by indicator. For example, while a simple mean is sufficiently meaningful for NO2, it is more appropriate to use the 8-hour running mean for O3. See below for a list of available data products and their temporal extents and aggregation levels. Variable Spatial Resolution Temporal Resolution N Temporal Extent Temporal Aggregation no2_mean 10 km daily 2066 2018-05-01 - 2023-12-31** daily mean o3_max 10 km daily 3287 2015-01-01 - 2023-12-31 daily maximum of 8-hour running mean pm2.5_mean 10 km daily 3287 2015-01-01 - 2023-12-31 daily mean pm10_mean 10 km daily 3287 2015-01-01 - 2023-12-31 daily mean no2_mean 1 km monthly 68 2018-05-01 - 2023-12-31** monthly mean o3_perc 10 km* monthly 108 2015-01-01 - 2023-12-31 monthly 93.4th percentile of daily o3_max pm2.5_mean 1 km monthly 108 2015-01-01 - 2023-12-31 monthly mean pm10_mean 1 km monthly 108 2015-01-01 - 2023-12-31 monthly mean pm10_perc 1 km monthly 108 2015-01-01 - 2023-12-31 monthly 91.2th percentile of daily mean no2_mean 1 km annual 6 2018-05-01 - 2023-12-31** annual mean o3_perc 10 km* annual 9 2015-01-01 - 2023-12-31 annual 93.4th percentile of daily o3_max pm2.5_mean 1 km annual 9 2015-01-01 - 2023-12-31 annual mean pm10_mean 1 km annual 9 2015-01-01 - 2023-12-31 annual mean pm10_perc 1 km annual 9 2015-01-01 - 2023-12-31 annual 91.2th percentile of daily mean * O3 is always mapped at 10 km spatial resolution. ** NO2 mapping relies on TROPOMI/Sentinel-5P data, which was only available from May 2018. Data Format Maps are shipped as Zarr stores, an efficient data format particularily suited for multidimensional data. For better compatibility with Zenodo these archives are zipped. Check the supplementary Python notebook for instructions on how to open Zarr stores.