Vegetation structure and microclimate influence on fire spread along an ecotonal gradient

Background/Question/Methods

In pyrogenic ecosystems, fire, vegetation, and microclimate interact in a fire-promoting, positive feedback. However within the broader matrix, non-pyrogenic communities may also be found. Fires extinguish somewhere between these adjacent communities, generating a dynamic ecotonal boundary. Some boundaries are sharp (e.g. tropical savanna – forest), whereas others have less abrupt transitions. While drivers of the feedback at sharp boundaries have been suggested, the specific impacts of these drivers on fire spread in broader ecotones are unknown. Our objectives were to determine the influence of vegetation structure and microclimate on ecotonal fire spread. We established 53 transects, along a longleaf pine savanna – wetland ecotone. We recorded height, percent cover, vertical density, and canopy cover for 12 vegetation structure classes within 269 plots. We collected living and dead fuels along a subset of 24 transects. Microclimate was quantified at five representative ecotones from continuously collected data on precipitation, soil moisture, fuel moisture and temperature, air temperature, relative humidity, wind speed and direction, and net radiation. Following prescribed fire, we measured fire spread along each transect. We used regression analyses and model selection with measures of vegetation structure and microclimate as predictors of the probability of fire spread. 

Results/Conclusions

Vegetation structure and microclimate variables significantly predicted the probability of fire spread through the savanna-wetland ecotone. Our best model using vegetation structure predictors included evergreen shrub and fern cover. Our best microclimate model included fuel moisture and soil moisture as predictors. Our best combined (vegetation structure and microclimate) model used all of the predictors from the two separate models. The probability of burning was greatest under low cover of evergreen shrubs and ferns and low soil and fuel moisture levels. The relative fit of each of the three models was assessed with receiver operating characteristic curves. We found that the vegetation structure and the combined model performed well in predicting the probability of ecotonal fire spread. It appears that vegetation structure and microclimate strongly influence fire spread, but should not be considered absolute, especially with threats of environmental change. This feedback is critical for the maintenance of vegetative structural boundaries and distinct ecosystems (e.g. savanna – wetland).