Colorado Wildfire, Measured: The Season Now Runs Into December

Part two of three. Part one covers the largest fires since 1984; part three covers where FEMA has been called.

The fire perimeter record says how much burned. It is thin on when: discovery dates exist only for fires since 2020. For the season we turn to the National Weather Service's own record, NOAA's Storm Events Database, which logs each wildfire the Weather Service reported, by date, since 1996.

Where the reports fall in the year

Period	Reports	June–August	September–December	October–December
1996–2005	201	170 (85%)	11 (5%)	3 (1%)
2006–2015	131	74 (56%)	15 (11%)	12 (9%)
2016–2025	329	197 (60%)	106 (32%)	58 (18%)

Summer is still the heart of the season. What has changed is the tail. In the decade from 1996 a Colorado wildfire report after Labor Day was rare. In the decade to 2025, nearly a third of them came in September or later, and one in six in October, November or December. The 2006–2015 decade also shows a spring shoulder: 32% of its reports fell between January and May.

Three late fires

East Troublesome, October 2020. Discovered October 14 in Grand County, it grew to 188,915 acres, the second largest in the state's record, destroyed 555 structures and killed two people, according to the NOAA record.

Cameron Peak, October 2020. Discovered in August, it was still making major runs in October, until snow on October 23–25 slowed it. The largest fire in Colorado's record at 208,913 acres.

Marshall, December 30, 2021. Downslope winds gusting near 100 mph drove a grass fire into Superior, Louisville and unincorporated Boulder County. NOAA records $2.0 billion in property damage and two deaths. That single event is more than half of the $3.83 billion of property damage across all 668 Colorado wildfire reports in the database.

Damage does not follow acres

The costliest Colorado wildfire records in NOAA's database are not the largest fires. Marshall burned 6,080 acres. Waldo Canyon (2012) and Black Forest (2013) in El Paso County, each a fraction of Hayman's size, are the next costliest, because they burned where people live. The large backcountry fires of part one dominate acreage; the fires at the edge of towns dominate damage.

What the record cannot tell you

These are reports, not fires. The Storm Events Database logs a wildfire per National Weather Service forecast zone, so one fire spanning three zones is three reports, and reporting practice has changed over thirty years. Read the table for the shape of the season, not as a fire count.

2026 is incomplete and left out of the decade comparison.

Nothing here is attribution. The record shows that late-season fires have become common. It does not, by itself, say why.

Why a longer season matters for an EOC

Staffing plans, red flag briefings and mutual-aid agreements are often written around a summer season. When a quarter or more of the activity falls after September, the October duty officer is working a fire season, not the end of one. Calchis watches red flag warnings, fire weather watches and active fires against your own jurisdiction all year, and its conditions do not switch off on a calendar date.

Sources: NOAA National Centers for Environmental Information, Storm Events Database, event type Wildfire, Colorado, 1996–2025, as loaded into Calchis and queried October 6, 2026. National Interagency Fire Center, WFIGS Interagency Perimeters, for acreage and discovery dates.