Data sources and linkage limits¶
Choose a source that matches the chemical, route, place and period in the research question. The links below are entry points, not a claim of complete national or historical coverage. Record the data product, release, years, geographic unit, units of measurement and processing method used in each analysis.
Air and industrial sources¶
| Source | What it provides | Use and limitation |
|---|---|---|
| EPA AirData / AQS | Historical outdoor monitor measurements and downloadable summaries | Useful for monitoring-based analyses and model evaluation. Coverage varies by pollutant, place and date; a monitor value is not a personal dose |
| EPA AirToxScreen | Screening estimates for outdoor hazardous air pollutants | Use pollutant concentration fields for relevant research questions. Cancer-risk estimates are not ALS-risk estimates. Geographic resolution and methods differ across releases |
| EPA Toxics Release Inventory | Reported chemical releases from covered facilities | Retain chemical, year and release medium. Released mass does not establish human exposure; reporting does not cover every source |
| EPA Superfund site search | Site locations and links to site records | Review contaminants, affected media, dates and remediation. The mapped point may not represent a plume or the full affected area |
EPA describes AirToxScreen as a screening resource with data and modeling limitations. Do not interpolate its releases into an apparent continuous exposure history without evaluating changes in models, inventories and geography. For particulate pollution, consider a validated spatiotemporal model with documented geographic and temporal coverage.
Agriculture¶
| Source | What it provides | Use and limitation |
|---|---|---|
| USGS Pesticide National Synthesis Project | Estimated annual agricultural use by chemical, with county data and maps | Appropriate primarily for broad patterns. Read the methods and limitations before individual-level linkage. Within-county allocation is modeled rather than observed application |
| California Pesticide Information Portal | Pesticide use records searchable by time, chemical, location and treated site or crop | Useful where record detail supports the question. Geographic detail varies, and California reporting cannot be assumed for other states |
For USGS data, preserve the distinction between EPest-low and EPest-high and examine discontinuities in source coverage, including the exclusion of seed-treatment estimates beginning in 2015. The two estimates are not statistical confidence bounds. Retain application units rather than converting use to a biological dose without a validated model.
Water and blooms¶
| Source | What it provides | Use and limitation |
|---|---|---|
| EPA CyAN | Satellite-derived cyanobacterial indicators for resolvable water bodies | Use the selected product's actual dates, spatial resolution, quality flags and valid observations. Coverage depends on water-body size and sensor |
| EPA HAB monitoring and trends | An overview of satellite bloom resources | Satellite estimates describe biomass-related conditions and cannot detect toxin concentrations |
| EPA HAB methods | Methods for sampling and measuring specified cyanotoxins | Choose the assay for the toxin and sample medium. A microcystin assay does not measure BMAA |
| EPA service-line inventory | Public-water-system counts by pipe-material category | System counts do not identify a participant's pipe or water concentration. Seek local service-area and address-level records where authorized |
For bloom histories, document the sensor and gaps between products. Water-body proximity, recreational contact, source-water contamination and finished drinking-water concentrations are distinct variables. Public-supply and private-well histories require different linkage information.
Minimum record for a derived estimate¶
| Field | Example |
|---|---|
| Participant and location interval | Study ID, address-period ID, start date, end date |
| Variable and assessment type | PM2.5 concentration; modeled outdoor estimate |
| Source and coverage | Product name, release, covered years, source-file identifier |
| Spatial linkage | Grid-cell assignment, coordinate system, geocode quality |
| Temporal linkage | Prespecified average window, available days, missing days |
| Result and unit | Value in µg/m³, with uncertainty where available |
| Quality and provenance | Missingness reason, processing version, linkage date |
These fields are a proposed documentation minimum. Store exact location identifiers within the approved secure system. Use the standards guidance to implement a shared representation and the privacy guidance before releasing maps.