|
Abstract:
|
Ecohydrological studies often incorporate datasets that are collected at various temporal and spatial scales and then combined to elucidate underlying ecosystem processes. While Bayesian statistics have enabled analyses of disparate datasets through process based models, it is often necessary to compartmentalize these analyses into subsets and then convey the results from one into another. A solution for this is the Bayesian cut function, which allows information to flow in one direction through a hierarchy, but not in the other direction. Here we describe the application of the Bayesian cut function for an ecohydrological study of snow sublimation in the mountains of southeastern Wyoming. The analysis investigates the impact of a spruce beetle outbreak via a mulita-decade sublimation dataset, and uses the Bayesian cut function to incorporate the uncertainty associated with satellite leaf area index measurements and net radiation into a forested canopy. Though the cut function led to more variability in all model process parameters, the results of the Bayesian analysis were robust and indicated that the impact of beetles on sublimation was statistically significant.
|
