Wildfire smoke in Montana carried more pollutants than models predicted

What the smoke contained

Smoke that had traveled for at least three days from California and the Pacific Northwest increased carbon monoxide, fine particulate matter known as PM2.5, and total measured volatile organic compounds by factors of 2 to 8. Hourly peaks reached 800 parts per billion for carbon monoxide, 120 micrograms per cubic meter for PM2.5, and 85 parts per billion for the measured volatile compounds.

Nitrogen oxides were not higher than the urban background. Ozone responded differently depending on smoke intensity: its daily maximum eight-hour average increased under light smoke but flattened or declined when PM2.5 exceeded about 30–40 micrograms per cubic meter. GEOS-Chem did not reproduce this pattern.

The researchers estimated that, if a 2020-style wildfire season occurred every year, the excess lifetime cancer risk would be 100 in 1 million people—about seven times the non-smoke baseline. About 90% of that estimated cancer risk came from PM2.5. Formaldehyde, benzene, acrolein and acetaldehyde dominated the estimated chronic non-cancer hazard, which reached a hazard index of 3.0.

The model captured the major smoke intrusions but underestimated carbon monoxide, PM2.5 and volatile organic compounds by 30%–90%. The resulting health estimates were also lower: cancer risk by about 40% and the chronic non-cancer hazard index by about 10 times. The researchers attribute the model errors to underpredicted fire emissions and missing chemical reactions involving volatile organic compounds.

Why the estimates matter

The findings indicate that atmospheric models can understate the pollution in aged wildfire smoke and therefore understate health-risk estimates derived from those concentrations. In this study, most of the estimated cancer risk came from PM2.5, while several individual chemicals drove the chronic non-cancer hazard. Better representation of fire emissions and smoke chemistry could make assessments of wildfire pollution more accurate.

Evidence and caveats

The study used hourly measurements of pollutants in Missoula during one strong smoke event in 2020 and compared them with results from the GEOS-Chem atmospheric model. The health findings are estimates based on pollutant concentrations and a scenario in which a 2020-style wildfire season recurs every year, not observations of cancer or other illnesses in people. The evidence comes from one city and one event, so it does not establish that the same concentrations or risks would occur in every wildfire season or location.