A robust method for mapping refugial capacity in montane forests

Abstract

Climate change and its associated disturbances, such as droughts and wildfires, threaten ecosystems and the services they provide. One strategy to mitigate these impacts is identifying refugia, areas predicted to experience less severe disturbances than their surroundings. However, refugia mapping often lacks validation due to limited data and resources, underscoring the importance of approaches that explicitly address the inherent uncertainty in refugia models. We developed a quantitative analysis to identify areas with high refugial capacity, while accounting for uncertainty, using a fuzzy logic model to combine climate forecasts, topographic features, and anthropogenic factors in southern California's montane forests. These unique ecosystems protect neighboring watersheds, sequester carbon, and provide other services to the surrounding communities. We tested the model across three future climate scenarios and six cases with uncertain ecological thresholds to evaluate its robustness. Key areas with high refugial capacity were located near the coast (e.g., Southern Coast Range Mountains) and at higher elevations (e.g., southeastern San Gabriel Mountains). Results of future scenarios and uncertain parameters were good predictors of reference results ( and , respectively), meaning that modeled refugial capacity was stable to uncertain forecast data and parameters. The findings also indicated strong spatial congruence across future scenarios (>60% for pessimistic, >95% for optimistic) and cases with uncertain parameters (>80%). Given the expected increase in drought and wildfire in southern California, our analysis can support a proactive management framework to preserve a healthy forest with valued ecosystem services. The sensitivity analyses presented here can increase the utility of refugia mapping products to guide climate change adaptation strategies despite a lack of empirical data and uncertain parameters and forecasting data.