Large Overestimation of Projected Western U.S. Wildfire Burned Forest Area With Warming

Abstract Wildfires are projected to increase with warming in the western United States. Since vapor pressure deficit (VPD) is highly correlated with wildfire burned area historically, many studies have argued that large projected increases in VPD with warming imply large increases in burned area. Here, we argue that those projections are overestimated by as much as an order‐of‐magnitude. First, we show that both soil moisture and VPD are well correlated with historical burned forest area. Second, we demonstrate that projected changes in VPD with warming are much larger than those in soil moisture, leading to wildly divergent projections of burned forest area: with 3 K (4 K) of warming relative to pre‐industrial, the VPD‐based projection is about 16 (66) times the historical burned area, whereas the soil moisture‐based projection is only 2 (3) times the historical burned area. A similar divergence arises in more complex models that include VPD as only one of many explanatory variables. Third, we argue that the VPD‐based projections are incorrect. VPD is used as a measure of atmospheric evaporative demand, but recent advances have demonstrated that VPD and related quantities are actually poor measures of atmospheric evaporative demand and overstate projected drying with warming. We conclude that the rate at which wildfire burned forest area will increase with warming has been greatly overestimated by some studies.