Wildfire Source Library

How to evaluate a wildfire evidence source

The same incident can appear in a satellite thermal-detection feed, an agency situation report, a perimeter map and a post-fire severity assessment. Those records answer different questions. An active-fire detection is not a confirmed ignition location; a mapped fire perimeter is not the exact minute-by-minute path of flames; a severity map is not an incident-cause investigation.

A practical source-checking protocol

1. Identify the observation. Is the source reporting thermal anomalies, burned area, ground measurements, an official incident estimate, evacuation instructions or an investigator’s conclusion?
2. Identify the spatial scale. A satellite pixel, an administrative boundary, an approximate fire perimeter and a structure address have fundamentally different location uncertainty.
3. Separate event time from publication time. A bulletin published on Tuesday can describe a perimeter mapped on Monday. Keep both timestamps when available.
4. Preserve the version. Save a source URL, publication date and, where permitted, an archived copy or snapshot reference. Updated records may legitimately supersede earlier measurements.
5. Trace the claim. Connect each numerical or causal statement to its direct source instead of treating a chain of secondary summaries as independent confirmation.
6. Report the limit. Write explicitly what the source cannot establish. Missing data is not evidence that an event did not occur.

Worked example: what a satellite hotspot does—and does not—establish

Suppose a VIIRS-based NASA FIRMS product flags a thermal anomaly near a reported wildfire. It can support a statement that a sensor detected a temperature-related signal in a location and observation window. It does not, on its own, prove the first ignition point, the fire’s cause, exact burned perimeter or the number of affected structures. Cloud cover, revisit timing, sensor resolution and false positives also constrain interpretation.

For a more complete reconstruction, compare that detection with official incident updates, time-stamped perimeter data, applicable weather observations and later investigation findings. A historical burned-area product such as MTBS can add retrospective context, but it answers a different question than a near-real-time hotspot map. Do not treat different products as interchangeable measurements.

Use this library as a provenance register, not a list of endorsements

Each source profile explains a dataset’s purpose and boundaries. Selection for this library does not mean every source supports every claim. For terminology, begin with the Wildfire Glossary. For claim-by-claim reconstruction, visit Wildfire Black Box. Our editorial policy explains how corrections and uncertainty are handled.

Safety boundary: this is a research reference, not a live emergency-information channel. In an active incident, use the responsible fire and emergency authorities for operational decisions.