Skip to main content
Dryad

Data from: New methods provide a 300–year perspective on modern area burned in two wilderness areas of the southwest United States

Farris, Calvin1; Margolis, Ellis2; Allen, Craig D.3

Published Jun 02, 2025 on Dryad. https://doi.org/10.5061/dryad.k6djh9wh2

Data files

Jun 02, 2025 version files 3 MB

Click names to download individual files

Abstract

Climate change and increased fuels from fire exclusion drive recent increases in area burned and fire severity in dry conifer forests of the western United States. Recent increases in area burned are occurring against the backdrop of a large fire deficit caused by over a century of fire exclusion. A key land management question is whether fire regimes can be restored – i.e., can the fire deficit be paid off? Accurate estimates of historical area burned prior to fire exclusion (circa 1900) are difficult to derive and have rarely been calibrated or validated against modern fires, leaving their accuracy uncertain. In this study, we applied multiple time-series methods to tree-ring fire-scar data for the first time, including a multi-model ensemble, time-varying predictor subset calibration, and multiple measures of model uncertainty and validation. We focused on two southwestern U.S. wilderness areas – Saguaro National Park (SAGU) and the Gila Wilderness (GILA) – that have abundant and well-documented modern fires due to decades of active fire restoration, allowing us to test whether the area burned has been restored to historical levels.

Our ensemble of ten fire-scar models accurately estimated area burned by modern fires with no consistent biases. While individual models did not differ significantly from mapped fires, the ensemble solution reduced overall bias and uncertainty. Each member model has distinct strengths that make them suitable for specific applications (e.g., the synchrony model is easily applied, and Thiessen polygons provide spatially explicit estimates). The accurate reconstruction of the modern area burned from relatively sparse fire scar data at GILA suggests that dense grids are not necessary for accurate reconstructions. Our findings reveal that despite dropping to near zero in the early 20th century, the area burned in recent decades is within historical levels at GILA and trends toward historical levels at SAGU. These results highlight the magnitude of the post-1900 fire deficit and demonstrate that fire management can restore the historically prevalent, ecologically important process of widespread, frequent, low- to moderate-severity fire in dry conifer forests.