Wind-transport impacts on snow accumulation and water available for runoff at the watershed scale

Abstract:

Patterns of snow accumulation drive local water availability from snowmelt; these patterns are dominated by wind redistribution at the hillslope scale and orographic precipitation at the basin scale. Realistic representation of wind transport is challenging and requires high resolution (10m) to be able to represent small features in topography that will affect wind deposition. Operational models (e.g., the National Water Model and SNODAS) are commonly implemented at coarse resolution (1000 m) and do not represent wind redistribution.

Our goal is to evaluate the importance of wind transport to snow distributions at the scales of small headwater basins (<10 km²) to larger mountain watersheds (>100 km²). To obtain realistic snow depth distributions, it is critical to have spatially distributed snow data (e.g., from airborne lidar). The distributed snow data enable selection of optimal parameter values. A key consideration is the physical meaning of parameters and how well these selected parameters transfer to different basin that lack spatial snow data.

Here we test how different terrain parameters influence wind-driven snow patterns in Snowmodel, using the high resolution (3m) NASA Airborne Snow Observatory data sets for parameter selection and validation. We first develop model parameters in a small headwater catchment (Senator Beck Basin), and then will test how well they transfer to other headwater catchments (Bradley Creek) and to larger watershed scales (East River). Initial results show that the most sensitive parameters were the curvature and slope weights in wind acceleration over the topography and the curvature length scale. We selected the best values for these parameters based on two evaluation criteria: a point to point match using the root mean squared error (RMS) and the spatial pattern representatively using the Spatial efficiency metric (SPAEF).

To estimate the impact of model resolution on wind transport, we compare the fluxes summed over the watersheds using 10m and 100m simulations. We found different impacts of snow depth and SWE spatial distributions on the timing and magnitude of ablation fluxes and hydrological cycle. These results highlight the importance of scale and wind process representation, which can support water resources management.