Observing, modelling, and validating snow redistribution by wind in a Wyoming upper treeline landscape

Abstract

Redistribution of snow by wind is a defining feature of the physical environment in high-elevation upper treeline and alpine landscapes where most precipitation is snow and wind speeds are frequently high. In these landscapes, snow, wind, topography, and vegetation interact to produce a pattern of snow deposition where wind-driven snow is eroded from exposed areas and deposited on the lee sides of hills, trees, rocks, and other obstructions. During winters on Libby Flats, a broad, high-elevation ridge located in southeastern Wyoming, USA, snow is heterogeneously distributed with snow depths ranging from 0.1 to 7 m over a lateral span of a few meters. Snow depth variations are principally controlled by the aerodynamic influence of trees on wind, with the deepest drifts forming on the lee sides of krummholz patches and ribbon forests. While precipitation varied from year to year, spatial patterns of snow scouring, drifting, and snowmelt were nearly identical. This characteristic and chronic pattern alters ecosystem structure (e.g., plant species distributions and soil characteristics) and function (e.g., decomposition, primary production, nutrient cycling, and water balance). To better comprehend and predict how ecosystem structure and function change with respect to snow distribution in treeline ecotones, an understanding of the snow transport process, ablation, and subsequent water distribution is required. Since snow accumulation and melt are difficult to observe spatially, we adapted and validated snow transport and land surface models to simulate the heterogeneous snow distributions produced during three consecutive water years (1998–2000). Comparisons of field measurements with model predictions indicate that the linked models generally mimicked the temporal and spatial characteristics of snow accumulation, redistribution, and ablation on Libby Flats. However, considerable modelled and observed discrepancies were associated with shallow or deep snow depths. Model agreement with observations could be improved with higher temporal resolution meteorological data, better resolved terrain and vegetation data, and model improvements.