Representing spatial variability of forest snow: Implementation of a new interception model

Abstract

A new interception model was integrated in a snowmelt model and for the first time the spatial variability of forest snow was effectively represented due to the inclusion of new forest structure metrics. The model was tested at 1273 field points surrounding Davos, Switzerland, that feature an extremely wide range of canopy and forest structure. The behavior of the new model was compared against a widely applied interception model. Due to the inclusion of novel forest structure parameters (mean distance to canopy and total gap area) in the new model, simulated interception mimicked the horizontal layout of canopy structure, while the standard interception model yielded fairly homogeneous interception estimations even under highly heterogeneous canopy conditions. The large variance of estimated interception between points using the new model translated into significant effects on under-canopy snow water equivalent and snow depth. Precipitation conditions were also analyzed, and further differences between the models were related to storm intensity. In climates characterized by large storm events, the new interception model provides significantly higher interception estimations (i.e., lower under-canopy snow) than the standard model, whereas in climates characterized by small storms events, the new model yields lower interception estimations (i.e., higher under-canopy snow) in areas with moderately sized to large canopy gaps.