Unveiling the landscape predictors of resilient vegetation in coastal wetlands to inform conservation in the face of climate extremes

Unveiling spatial variation in vegetation resilience to climate extremes can inform effective conservation planning under climate change. Although many conservation efforts are implemented on landscape scales, they often remain blind to landscape variation in vegetation resilience. We explored the distribution of drought-resilient vegetation (i.e., vegetation that could withstand and quickly recover from drought) and its predictors across a heterogeneous coastal landscape under long-term wetland conversion, through a series of high-resolution satellite image interpretations, spatial analyses, and nonlinear modelling. We found that vegetation varied greatly in drought resilience across the coastal wetland landscape and that drought-resilient vegetation could be predicted with distances to coastline and tidal channel. Specifically, drought-resilient vegetation exhibited a nearly bimodal distribution and had a seaward optimum at ~2 km from coastline (corresponding to an inundation frequency of ~30%), a pattern particularly pronounced in areas further away from tidal channels. Furthermore, we found that areas with drought-resilient vegetation were more likely to be eliminated by wetland conversion. Even in protected areas where wetland conversion was slowed, drought-resilient vegetation was increasingly lost to wetland conversion at its landward optimum in combination with rapid plant invasions at its seaward optimum. Our study highlights that the distribution of drought-resilient vegetation can be predicted using landscape features but without incorporating this predictive understanding, conservation efforts may risk failing in the face of climate extremes.