Post‐fire recovery of sagebrush‐steppe communities is better explained by elevation than climate‐derived indicators of resistance and resilience

Abstract

Listen

Translate article

Abstract More landscapes require restoration than can feasibly be treated, and so decision‐support tools to prioritize areas for treatment are needed. Moreover, restoration is complicated by the threat of biological invasion in disturbed areas, and so indicators of ecosystem resistance to invasion and resilience to disturbance (hereafter R&R) are important candidate criteria for prioritizing sites for restoration. We asked how climate‐based R&R indicators that differed in being either categorical or continuous compared in their ability to explain plant‐community recovery after six wildfires that collectively encompassed >750,000 ha and 7803 plot‐year observations in sagebrush steppe of the western USA. Unique associations of species that most frequently co‐occurred were identified using structural topic modelling. Mixed effect random forests were used to identify the relative importance of various R&R indicators in explaining post‐fire plant associations compared with weather, landscape characteristics and treatment history. Simple metrics (elevation, latitude, longitude and year of monitoring) were more informative predictors of post‐fire recovery than climate‐based R&R indicators. However, small differences in the abundances of perennial grass and especially annual grass associations were predicted by the spring modified Thornthwaite Moisture Index (difference between precipitation and potential evapotranspiration). Synthesis and applications: The convenience of categorical resistance and resilience indicators has led to their widespread adoption for large‐scale planning of restoration. Our results reveal that none of the resistance and resilience indicators assessed effectively explained post‐fire restoration better than elevation, although a simple continuous resistance and resilience indicator describing water balance performed better than categorical indicators for explaining small but critical differences in cheatgrass association abundances.