Air Quality & Health Challenge

Overview

The AI4EO Air Quality & Health challenge stems from a need expressed by ECMWF and focuses on downscaling Earth Observation (EO) data, which is especially pressing for air quality data and products.

The Copernicus Atmosphere Monitoring Service (CAMS) provides global information about air pollutants at the surface with a spatial resolution of about 40 km and with a resolution of about 10 km over Europe.

With the help of local information on topography, emission sources and their proxies, as well as additional in situ data and meteorology they can be downscaled. The problem lies in the incompleteness of in situ datasets, which are not always available, well connected, or in compatible formats. It is therefore necessary to developing AI-powered solutions to meet these needs.

The challenge is to downscale air quality products to a resolution that can be used on a local level. The focus is on the pollutants PM2.5 and NO2, which are amongst the most important to consider for health effects.

The goal is to reconstruct past time-series data using observations and demonstrate a downscaling method that can potentially be applied anywhere in the world for assessing the health impacts of air pollution.

The participants are requested to produce output datasets on three areas of interest (AoI, 500 km x 500 km each): North Italy, California and South Africa.