Future vulnerability mapping based on response to extreme climate events: Dieback thresholds in an endemic California oak

Aim: This study presents a bioclimate modelling approach, using responses to ex-treme climate events, rather than historical distributional associations, to project fu-ture species vulnerability and refugia. We aim to illustrate the compounding effects of groundwater loss and climate on species vulnerability. Location: California, USA. Methods: As a case study, we used the 2012–2015 California drought and resulting extensive dieback of blue oak (Quercus douglasii). We used aerial dieback surveys, downscaled climate data and subsurface water change data to develop boosted re-gression tree models identifying key thresholds associated with dieback throughout the blue oak distribution. We (1) combined observed dieback–climatic threshold rela-tionships with climate futures to anticipate future areas of vulnerability and (2) used satellite- derived measurements of subsurface water loss in drought/dieback model-ling to capture the mediating effect of groundwater on species response to climatic drought. Results: A model including climate, climate anomalies and subsurface water change explained 46% of the variability in dieback. Precipitation in 2015 and subsurface water change accounted for 62.6% of the modelled probability of dieback. We found an interaction between precipitation and subsurface water in which dieback proba-bility increased with low precipitation and subsurface water loss. The relationship between precipitation and dieback was nonlinear, with 99% of dieback occurring in areas that received <363 mm precipitation. Based on a MIROC_rcp85 future climate scenario, relative to historical conditions, 13% of the blue oak distribution is pre-dicted to experience more frequent years below this precipitation threshold by mid- century and 81% by end of century. Main conclusions: As ongoing climate change and extreme events impact ecological processes, the identification of thresholds associated with observed dieback may be combined with climate futures to help identify vulnerable populations and refugia and prioritize climate change-related conservation efforts.