PM2.5 Plume Forecast

How It Works

This forecast uses a reduced-form Gaussian puff dispersion model to simulate how PM2.5 pollution spreads through the atmosphere. Individual “puffs” of pollution are released from known sources and tracked as they move with forecast wind fields, dispersing over time based on atmospheric stability conditions.

Data Sources

Emissions are drawn from the EPA National Emissions Inventory (point sources and county-level nonpoint sources) and active wildfire detections from NIFC. Meteorological data comes from NOAA's HRRR forecast model. Results are aggregated into H3 hexagonal grid cells at resolution 4 for visualization.

Known Limitations

• •Does not include all emission sources — mobile sources (vehicles), residential wood burning, and some smaller industrial sources are not yet modeled
• •Secondary PM2.5 formation (atmospheric chemistry converting gases to particles) uses simplified parameterizations rather than full chemical transport
• •Wildfire emissions are estimated from satellite detections and may not capture all active fires or accurately reflect fire intensity
• •Concentrations should be treated as relative indicators, not precise measurements — absolute values may differ from monitored readings

What's Next

We're actively working on improving source coverage, adding data assimilation from ground monitors, and refining the dispersion physics. This beta release is an early look at what we're building. Feedback is welcome.