Periodic Reporting for period 1 - FORECAST-AIR (Open-Access Forecasting System of the Health Effects of Air Pollution)
Reporting period: 2024-04-01 to 2025-09-30
Air pollution represents a major global environmental health problem. Currently, 99% of the world population lives in locations where international air quality guidelines are not met. Overall, ambient air pollution causes 4.2 million premature deaths every year worldwide, half a million of which in Europe alone.
Despite this, existing air quality early warning systems are largely based on pollutant concentration thresholds and atmospheric forecasts. These systems do not reflect the actual health impacts of pollution nor account for differences in population vulnerability, limiting their effectiveness for public health protection. Alerts are often triggered by severe episodes only, overlooking the broader and more frequent effects of moderate pollution events.
The ERC-funded project FORECAST-AIR seeks to overcome these limitations. It integrates high-resolution air quality forecasting, environmental epidemiology, and socioeconomic vulnerability data to create a next-generation health early warning system. The project has developed epidemiological models linking pollution to health outcomes across different causes of disease and population groups. These models are applied to air quality forecasts to predict health impacts, not just pollutant concentrations.
Despite this, existing air quality early warning systems are largely based on pollutant concentration thresholds and atmospheric forecasts. These systems do not reflect the actual health impacts of pollution nor account for differences in population vulnerability, limiting their effectiveness for public health protection. Alerts are often triggered by severe episodes only, overlooking the broader and more frequent effects of moderate pollution events.
The ERC-funded project FORECAST-AIR seeks to overcome these limitations. It integrates high-resolution air quality forecasting, environmental epidemiology, and socioeconomic vulnerability data to create a next-generation health early warning system. The project has developed epidemiological models linking pollution to health outcomes across different causes of disease and population groups. These models are applied to air quality forecasts to predict health impacts, not just pollutant concentrations.
FORECAST-AIR aimed to innovate health early warning systems by integrating air quality forecasts with health risk modelling, addressing a key gap in current systems that rely solely on pollutant thresholds and ignore population vulnerability. The project advanced technically and scientifically in the following areas:
Database generation and formatting: FORECAST-AIR exploited the EARLY-ADAPT database -a harmonised, continental-scale dataset combining cause-specific mortality data with environmental and socioeconomic variables. This database enabled fine-scale, high-resolution analyses across Europe and laid the foundation for pan-European epidemiological modelling.
Epidemiological modelling: We developed and validated models that quantified the short-term effects of air pollutants (PM2.₅, NO2, O₃, etc.) on mortality, stratified by demographic and socioeconomic groups and causes of death.
Health-based forecasting: We transformed air quality forecasts into health impact estimates using our epidemiological models. This represents a shift from conventional physical forecast systems to health-driven early warnings that better reflect population-level impacts.
Predictability and system design: We assessed the forecast window for both pollutant concentrations and predicted health outcomes, identifying lead times sufficient for decision-making. This work supports trust and usability by public health authorities and enables evidence-based alert thresholds.
Online platform: We expanded the early warning system platform that was originally created in a previous ERC PoC project (HHS-EWS) to assess the health impacts of environmental temperatures (Forecaster.Health https://forecaster.health/(opens in new window)). Within FORECAST-AIR, we added and integrated the counterpart air pollution component, making Forecaster.Health a more complete, multi-exposure tool that now also captures the health effects of air quality. The health predictions that appear in the early warning system platform are automatically updated and uploaded every day by a series of protocols and scripts that we wrote, and that are synchronised in real time with the time of release of the daily air quality forecasts in the CAMS data store.
Scientific outputs and impact: FORECAST-AIR led to publications in Nature Medicine, Nature Communications, Nature Climate Change, The Lancet Planetary Health, and Science Advances.
FORECAST-AIR provides a robust, health-centred methodology for early warnings of air pollution impacts, supporting more targeted and equitable public health interventions. By combining environmental epidemiology with real-time forecasting and vulnerability mapping, the project offers a breakthrough framework in climate services for health, aligning with EU goals on air quality, health equity, and early action.
Database generation and formatting: FORECAST-AIR exploited the EARLY-ADAPT database -a harmonised, continental-scale dataset combining cause-specific mortality data with environmental and socioeconomic variables. This database enabled fine-scale, high-resolution analyses across Europe and laid the foundation for pan-European epidemiological modelling.
Epidemiological modelling: We developed and validated models that quantified the short-term effects of air pollutants (PM2.₅, NO2, O₃, etc.) on mortality, stratified by demographic and socioeconomic groups and causes of death.
Health-based forecasting: We transformed air quality forecasts into health impact estimates using our epidemiological models. This represents a shift from conventional physical forecast systems to health-driven early warnings that better reflect population-level impacts.
Predictability and system design: We assessed the forecast window for both pollutant concentrations and predicted health outcomes, identifying lead times sufficient for decision-making. This work supports trust and usability by public health authorities and enables evidence-based alert thresholds.
Online platform: We expanded the early warning system platform that was originally created in a previous ERC PoC project (HHS-EWS) to assess the health impacts of environmental temperatures (Forecaster.Health https://forecaster.health/(opens in new window)). Within FORECAST-AIR, we added and integrated the counterpart air pollution component, making Forecaster.Health a more complete, multi-exposure tool that now also captures the health effects of air quality. The health predictions that appear in the early warning system platform are automatically updated and uploaded every day by a series of protocols and scripts that we wrote, and that are synchronised in real time with the time of release of the daily air quality forecasts in the CAMS data store.
Scientific outputs and impact: FORECAST-AIR led to publications in Nature Medicine, Nature Communications, Nature Climate Change, The Lancet Planetary Health, and Science Advances.
FORECAST-AIR provides a robust, health-centred methodology for early warnings of air pollution impacts, supporting more targeted and equitable public health interventions. By combining environmental epidemiology with real-time forecasting and vulnerability mapping, the project offers a breakthrough framework in climate services for health, aligning with EU goals on air quality, health equity, and early action.
FORECAST-AIR has resulted in a breakthrough innovation in the area of early warning systems due to its multiplicative effect. The driving force of state-of-the-art early warning systems is air quality forecasting, from which operational threshold-based early warnings are now issued. FORECAST-AIR changes this paradigm by moving the centre of gravity of the system from meteorology and atmospheric chemistry to environmental epidemiology. In this innovative, inter-disciplinary concept, physical variables from air quality forecasts are not at the core of the early warnings. Instead, we applied epidemiological models to health data disaggregated by causes of disease and sociodemographic groups to calculate group-specific health indicators, which will be used to activate public health and emergency plans. In other words, FORECAST-AIR uses epidemiological models to transform the physical information of air quality forecasting schemes into a range of early warnings representing the real health risks of, for example, the elderly, children, women, low-income people or people with allergies, asthma or cardiorespiratory diseases. FORECAST-AIR thus represents a new generation of early warning systems better fitted for the purpose of explicitly accounting for the real risks and impacts of air pollution on the vulnerable population, including the inequalities and gaps in vulnerability. In turn, these more precise early warnings can better inform potential end-users, such as public health agencies and governments, to activate fit-for-purpose emergency plans and real-world measures directly targeting vulnerable groups to protect their health, and at the same time, minimise the impact on the general population. In summary, FORECAST-AIR delivers a transformative, health-centred solution to address the societal burden of air pollution. With appropriate support, it has the potential to become a cornerstone of future climate-health services and policy-making.