An AI System to Predict Street-Level Air Quality: Introduction to the QUALARIA Project

Here we present the joint German-Brazilian project QUALARIA (Artificial Intelligence based system for sub-urban scale air quality prediction/Sistema baseado em Inteligência Artificial  para previsão de qualidade do ar em escala sub-urbana). 

Through joint effort between research and business partners in Brazil and Germany, QUALARIA proposes to develop an operational artificial intelligence-based system for monitoring, simulating and predicting air quality in urban environments with unprecedented spatial resolution availability (https://meteoia.com/qualaria/). New downscaling approaches based on artificial intelligence have recently shown promising performance to simulate sub-grid atmospheric processes. This approach is designed to monitor and predict atmospheric pollutant concentrations and air quality indexes with high spatial resolution, through the development of the QUALARIA system. Advanced global and regional chemical-meteorological models, such as reanalysis data from ERA5 and EAC4, and WRF-Chem simulations are applied to derive the climatological state of air composition, specifically the average levels of air pollutant based on existing emission inventories. Measurements of PM10, PM2.5 NO2, and O3 concentrations from Air Quality Automatic Stations of the Environmental Company of São Paulo State (CETESB) are used to train the downscaling AI algorithm to capture the sub-grid spatial variations of the pollutant concentrations. Low-cost sensors are deployed to increase and complement the spatial coverage of the CETESB network. Artificial intelligence will transform air quality maps at a horizontal resolution of 10 km to street-level maps with an increased resolution of 100 m. From its simulated and predicted downscaled pollutant concentration fields, QUALARIA will provide its users with relevant air quality indicators, informing about the impacts of air pollution in human health and activities via an online dashboard. To achieve the optimal dashboard design, public and private sector stakeholders are being engaged and consulted for the co-development of the dashboard design and features.