See the danger before the spark.

PRECEDE scores wildfire danger across 4,809 monitored points in 11 western states, refreshed around the clock over a rolling ~7-day horizon, so the people who protect communities can act before a fire starts, not after.

The problem

Satellites tell you a fire has already started. They cannot tell you that this afternoon, in a particular stretch of country, conditions have lined up so that whatever ignites will be hard to stop. That gap, the hours before a spark, is where lives and homes are saved or lost. PRECEDE was built to close it.

What PRECEDE does

PRECEDE watches the fire-prone parts of 11 western states from 4,809 monitored points and scores wildfire danger continuously, well below the level of a whole county. Each point gets a single risk number, refreshed around the clock and looking ahead over a rolling ~7-day horizon, that reflects whether conditions have entered the zone where a fire would likely turn destructive. Every point also carries building-exposure context from a 26.7-million-building census. When a point crosses that line, it surfaces as an alert someone can act on. The output is simple on purpose: a score, a place, a trend, and a clear threshold for when to pay attention.

Source: Microsoft US Building Footprints

How well it works

In a California validation of 87 destructive-fire events, PRECEDE flagged 73.6% (64/87) versus 50.6% (44/87) for a standard fire-weather-index threshold at the same 6.27% false-alarm rate. PRECEDE scores point-precise weather at each location while the index uses 0.25-degree reanalysis; the asymmetry is a property of the comparison. This is a California result: one operating point, pre-committed before it was scored, on a fixed 87-fire gold set. Outside California the score is monitoring coverage, not a restatement of this result. And it stays quiet on the overwhelming majority of ordinary days, because a warning that fires on every warm afternoon trains people to ignore it.

• The score is an index, not a calibrated probability.
• PRECEDE's approach has been evaluated most rigorously in California, where the historical fire record is richest. Any accuracy figures we publish are specific to that California evaluation and do not transfer to other states.
• The underlying fuel-type term uses LANDFIRE LF2022 labels that post-date part of the fire panel. A checked pre-era relabeling changes no catch and no miss at this operating point.

Full list of caveats: precedefire.com/validation

Knowing when it is safe, too (in development)

A prescribed-burn timing feature — reading the same conditions to flag when it may be safer to put fire on the ground on purpose — is in development. It is not part of the validated core, it is not available today, and we make no capability or performance claim for it yet. Knowing when not to burn is as valuable as knowing when you can.

How you can trust the numbers

We hold ourselves to a written, pre-committed standard. Our 2026 season-end validation was preregistered and frozen on 2026-08-27, with its pass/fail rules fixed before any 2026 outcome was known — so the test cannot be changed after the fact to fit the result. And every published number was independently recomputed from the raw data by an adversarial internal audit, completed 2026-08-28 — zero numbers broke. Full technical documentation is available on request, without us revealing how the model works internally.