Foehn model forecasts air pollution more precisely days ahead
Global probabilistic forecasting of air pollution
Artificial Intelligence
Summary
Air pollution from fine particles harms many people every year, but current forecasts are often too coarse for local details. The authors present Foehn, a model that predicts tiny particle pollution with much finer detail and up to four days ahead. It learns from global data and tracks where the polluted air comes from, improving forecast accuracy and alerting people to pollution spikes missed by other systems. This could help warn communities at high risk from local pollution well before it happens.
What this means in practice
- •For environmental monitoring agencies: Provide timely, localized pollution forecasts to alert high-risk populations not detected by coarse grid systems.
- •For urban planning teams: Incorporate detailed pollution forecasts into planning tools to better assess local air quality impacts of traffic and industry.
Authors
Ammar Kheder, Michael Boy, Zhi-song Liu
Abstract
Fine particulate matter (PM2.5) refers to particulate matter with an aerodynamic diameter of 2.5 micrometre or less. It comes from human activities, such as traffic and industry, and from wildfires, desert dust and gases emitted by vegetation. Short-term exposure to it accounts for about a million deaths a year. Operational forecasts run on grids tens of kilometres wide, which hide the pollution breathed by nearly two hundred million people a day. Machine learning has transformed weather forecasting, but data-driven forecasts of air pollution still blur extremes and lose skill as the air travels. Here we introduce Foehn, a generative model that learns from global kilometre-scale estimates of PM2.5 and computes where the air comes from. Over a year of retrospective forecasts, it outperforms the operational forecast at three stations in four and forecasts at the scale of this hidden exposure. Its ensembles are calibrated, and their mean detects local pollution extremes four days ahead as well as the operational forecast does one day ahead. Such forecasts could warn the millions whom coarse maps miss, once kilometre-scale data are available in real time.