Influences of fire–vegetation feedbacks and post‐fire recovery rates on forest landscape vulnerability to altered fire regimes

Abstract In the context of ongoing climatic warming, forest landscapes face increasing risk of conversion to non‐forest vegetation through alteration of their fire regimes and their post‐fire recovery dynamics. However, this pressure could be amplified or dampened, depending on how fire‐driven changes to vegetation feed back to alter the extent or behaviour of subsequent fires. Here we develop a mathematical model to formalize understanding of how fire–vegetation feedbacks and the time to forest recovery following high‐severity (i.e. stand‐replacing) fire affect the extent and stability of forest cover across landscapes facing altered fire regimes. We evaluate responses to increasing burn rates while varying the direction (negative vs. positive) of fire–vegetation feedbacks under a continuum of values for feedback strength and post‐fire recovery time. In doing so, we determine how interactions among these variables produce thresholds and tipping points in landscape responses to changing fire regimes. Where the early‐seral vegetation was less fire‐prone than older forests, negative feedbacks limited the reductions in forest cover in response to higher fire frequency or slower forest recovery. By contrast, positive feedbacks (more flammable early‐seral vegetation) produced a tipping point beyond which increases in burn rates or a slowing of forest recovery drove extensive forest loss. With negative feedbacks, the rates of forest loss and expansion in response to variation in fire frequency were similar. However, where feedbacks were positive, the conversion from predominantly forested to non‐forested conditions in response to increasing fire frequency was faster than the re‐expansion of forest cover following a return to the initial burn rate. Strengthening the positive feedbacks increased this asymmetry. Synthesis . Our analyses elucidate how fire–vegetation feedbacks and post‐fire recovery rates interact to affect the trajectories and rates of landscape response to altered fire regimes. We illustrate the vulnerability of ecosystems with positive fire–vegetation feedbacks to climate change‐driven increases in fire activity, especially where post‐fire recovery is slow. Although negative feedbacks initially provide resistance to forest loss with increasing burn rates, this resistance is eventually overwhelmed with sufficient increases to burn rates relative to recovery times.