Abstract:
|
The Alaskan landscape has undergone substantial changes in recent decades, most notably the expansion of shrubs and trees across the Arctic. We developed a dynamic statistical model to quantify the impact of climate change on the structural transformation of ecosystems using remotely sensed imagery. Our model accommodates changes in temperature and precipitation to infer and predict rates of land cover transitions while accounting for spatio-temporal heterogeneity. Transition types are highly correlated at both plot and subplot levels in our study system, therefore we characterized multiscale spatial correlation using Gaussian processes. Imagery pairs were collected at irregular time intervals, therefore we modeled dynamic state probabilities that evolve annually using a hierarchical framework. We developed a PĆ³lya-Gamma representation of our model to improve computation. Our model facilitates inference on the response of ecosystem state probabilities to shifts in climate and can be used to predict future land cover transitions under various climate scenarios.
|