You are here

Fire refugia and forest resilience

Date: September 02, 2020

The occurrence and pattern of small forest remnants after high severity fire can determine the rate and extent of forest recovery.


A gully with mostly burned trees with a few patches of live, unburned forest.
Fire refugia are more likely to occur and persist in gullies and adjacent to rocky fuel breaks that limit fire spread and intensity. Photo by Will Downing.
In forested landscapes across the western United States, the combined effects of fire suppression, logging, and grazing have altered forest structure and composition. These effects and recent climate-induced changes in fire regimes are resulting in uncharacteristically severe fire effects, dramatically affecting forested landscapes. In some cases, large fires have led to large-scale conversion of vegetation type (e.g., forest to non-forest), with potentially dire implications for existing biota and fundamental ecosystem processes [see previous Spotlight “Living on the edge: Trailing edge forests are at risk”].

Research suggests that fire refugia—places within high-severity burns that are unburned or that burn less severely—may help buffer forests from change and serve a key role in forest resilience. Not only do they provide a place for individual plants and animals to survive, but they harbor legacies (e.g., canopy shade and seed sources) that allow repopulation of surrounding areas and ecosystem recovery. Managers need to know how to identify fire refugia, understand their ecological functions, and make the best use of this knowledge to promote ecosystems that are resilient to fire.

Some fire refugia endure on the landscape for long time periods, remaining unburned even through multiple successive fires. These persistent fire refugia occur in places that are protected from fire such as topographically defined microclimates or in locations that are surrounded by nonflammable landscape features (e.g., lakes or exposed rock). Persistent refugia may be particularly important in retaining old-growth dependent communities, as well as rare species associated with microclimates not found in the surrounding burn mosaic. Other, more transient fire refugia are formed by chance coincidence with stochastic factors, such as variation in fuel continuity or unpredictable processes such as wind shifts during burning. Although less enduring, they nevertheless serve important ecosystem functions.

We developed methods to detect and map fire refugia and to predict their occurrence from topographic and weather information. To learn where fire refugia are most likely to occur, we used satellite-derived imagery for a set of large burns to develop predictive models based on a suite of topographic variables. We also developed methods using high-resolution (1-meter) aerial imagery so we could map smaller refugia.

A single standing tree surrounded by mostly fallen, dead trees and grass, and a few seedlings.
Even small patches of fire refugia that contain seed sources can be important for tree regeneration. Photo by Ryan Walker.
To investigate the role of fire refugia in promoting tree regeneration, we conducted field studies of thirteen large burns in Oregon, Idaho, Colorado, Arizona, and New Mexico. We collected data on tree species composition and tree seedling abundances within fire refugia and in areas that burned at high severity. We asked how patterns of seed sources within fire refugia influence post-fire tree regeneration and developed a spatially explicit simulation model to explore how refugia patch size affects the rate and extent of forest recovery.

We expected distance to nearest surviving seed source would be important for post-fire conifer regeneration, but we also developed a continuous metric to describe the density of refugia within a 9-ha neighborhood. This density metric allowed us to consider the additive effect of isolated or scattered seed sources on regeneration.

Key Findings

  • The occurrence of fire refugia is influenced by topography and the weather at the time of burning. Although specific topographic controls on fire refugia differ among environments, analyses suggest that fire refugia are more commonly formed in valley bottoms and in areas with higher soil moisture (Krawchuk et al. 2016). Fire refugia are least likely to form under extreme fire weather conditions as extreme dry and windy conditions can override the influence of topography on fire behavior (Krawchuk et al. 2016 and Haire et al. 2017).
  • High resolution maps of post-fire landscapes allow small patches of refugia to be detected. We found that within areas of high severity, the median patch size of fire refugia is 0.01 ha. These small patches of refugia may be undetected with widely used coarse-scale (30-m resolution) satellite imagery, leading to underestimation of persistent forest and capacity for regeneration (Walker et al. 2019).
  • Fire refugia were found to promote tree recovery on burned landscapes. Field studies showed that regeneration by ponderosa pine and other mixed-conifer tree species depends not only on proximity to seed source, but also on the abundance (density) of fire refugia (Coop et al. 2019 and Downing et al. 2019).
  • Simulations based on the field results demonstrate that small, isolated patches of surviving trees are disproportionately important for tree regeneration (Coop et al. 2019).

Additional Publications

Walker, RB; Coop, JD; Downing, WM; Krawchuk, MA; Malone, SL; Meigs, GW. 2019. How much forest persists through fire? High-resolution mapping of tree cover to characterize the abundance and spatial pattern of fire refugia across mosaics of burn severity. Forests 2019, 10, 782.

Featured Publications

Downing, William M. ; Krawchuk, Meg A. ; Meigs, Garrett W. ; Haire, Sandra L. ; Coop, Jonathan D. ; Walker, Ryan B. ; Whitman, Ellen ; Chong, Geneva ; Miller, Carol L. , 2019
Haire, Sandra L. ; Coop, Jonathan D. ; Miller, Carol L. , 2017
Krawchuk, Meg A. ; Haire, Sandra L. ; Coop, Jonathan ; Parisien, Marc-Andre ; Whitman, Ellen ; Chong, Geneva ; Miller, Carol L. , 2016


Principal Investigators: 
Principal Investigators - External: 
Meg A. Krawchuk - Oregon State University
Jonathan D. Coop - Western Colorado University
Sandra L. Haire - Haire Laboratory for Landscape Ecology