Description
The CACES data portal provides publicly available, high?resolution estimates of air pollution exposures and related metrics developed by the Center for Air, Climate, and Energy Solutions. It offers gridded and geographically linked datasets (e.g., by census tract, county, or other units) for pollutants like fine particulate matter (PM?.?), ozone, and nitrogen dioxide over multiple years, along with documentation and code. Researchers, agencies, and advocates can download these modeled exposure surfaces to analyze trends, assess health and equity impacts, and link air quality estimates to demographic, health, and policy data.
Questions this resource can help answer
How have PM2.5, ozone, or NO2 exposures changed over time in my city, county, or state?
Which neighborhoods or demographic groups experience the highest long?term air pollution exposures?
How do changes in policy or emissions line up with trends in modeled exposures?
What is the relationship between pollution exposure and health outcomes (e.g., mortality, asthma) across areas?
Are there disparities in exposure by race, income, or other social factors, and how large are those gaps?
How do I use this resource?
To use the CACES data portal, go to the Data page, choose the pollutant (e.g., PM2.5, ozone, NO2), spatial unit (grid, county, tract, etc.), and time period you need, then download the corresponding files and documentation. Open the data in your analysis or GIS software, join it to geographic identifiers (like FIPS codes) or shapefiles, and link it with other datasets (e.g., demographics, health outcomes, policies) to explore exposure levels, trends, and disparities.
Pro tips
Always download and read the documentation/codebook for each dataset so you understand units, years, and model versions.
Be consistent about spatial scale (e.g., always tract-level) when comparing across pollutants or years.
For small areas, consider using multi?year averages to reduce noise when studying chronic exposures.
Compare CACES estimates with local monitoring data (where available) to build intuition about biases and uncertainty.
Keep track of dataset version, date downloaded, and variable names in your project notes for reproducibility.
Resource information
Technical specifications
CACES air pollution exposure surfaces are created using statistical models that fuse multiple data sources—ambient monitor measurements, satellite remote sensing products, chemical transport model outputs, land?use and meteorological variables—into high?resolution predictions. For each pollutant and year, the team builds and validates regression/ensemble models that relate observed monitor concentrations to these predictors, then applies the fitted models across the full spatial domain (e.g., national grid). The resulting gridded estimates are then aggregated or linked to geographic units such as census tracts and counties. Extensive cross?validation against withheld monitoring data is used to assess model performance and guide refinements, and detailed metadata document inputs, modeling choices, and uncertainty.
