2018 ESA Annual Meeting (August 5 -- 10)

OOS 35-9 - Topographic fire refugia in late successional forests of the US Pacific Northwest

Thursday, August 9, 2018: 4:20 PM
343, New Orleans Ernest N. Morial Convention Center
Garrett W. Meigs and Meg A. Krawchuk, Forest Ecosystems and Society, Oregon State University, Corvallis, OR
Background/Question/Methods

Fire refugia are an important element of the burn mosaic in late-successional and old-growth forests of western North America. The spatial and temporal predictability of fire refugia is an important characteristic of the fire ecology of these forest ecosystems, and from a land management perspective provides information relevant to conservation of species that rely on late successional forest structure. Previous work shows that fire refugia – here defined as forests that remain unburned or burn less severely than their surroundings – are linked to specific topoedaphic and weather conditions. Here, we investigate how predictability of fire refugia varies with topographic complexity and fire weather for a population of large wildfires that burned between 2004 and 2015 in late successional conifer forests of Washington, Oregon, and northern California. We classified fire mosaics as refugia and non-refugia by relating a satellite-based RdNBR burn severity metric to field-measured indices of tree mortality from fire. We used boosted regression tree models to assess the relationship between fire refugia and a suite of metrics representing terrain, fuel and non-fuel, fire weather conditions, and antecedent climate. We used model selection tools to identify a parsimonious set of explanatory variables underlying the probability of topographically mediated fire refugia.

Results/Conclusions

Our analyses demonstrate the probability of fire refugia persistence through contemporary fires in late successional forests varies with topography and burning conditions in forests of the Pacific Northwest. The results highlight the distinction between fire refugia that are topographically mediated and predictable, from those that are stochastically arranged. We observed the highest predictability under moderate terrain and fire weather conditions. Multiple fine- and intermediate-scale topographic metrics explain the variability of refugia occurrence, and their relative importance depends on burning conditions. Our findings aim to inform ongoing conservation planning initiatives and bioregional assessment for the Pacific Northwest by identifying locations more likely to resist stand replacing fire under a range of weather and climate conditions. We propose these areas more resistant to burning confer resilience to the ecosystem, particularly in late-successional old growth forest environments critical to the survival of threatened species.