Author(s):
Jeffrey S. Deems - Colorado State University
Thomas H. Painter - National Snow and Ice Data Center
Christopher C. Landry - Center for Snow and Avalanche Studies
Abstract:
Snowpack properties such as depth, water equivalent, stability, or vertical structure, when observed on the hillslope scale (1 – 100 m), often exhibit dramatic spatial variability. This variability is compounded in the temporal domain. Snow system processes operate over multiple spatial and temporal scales, with nonlinearities, feedbacks, and threshold behaviors adding to the complexity. For example, snow depth depends on the sequence of precipitation and wind redistribution at the point of interest, and is often only weakly correlated to larger-scale patterns due to local effects of vegetation, terrain, or wind turbulence. The importance of processes at a particular location are contingent upon the local process history, and differing scales of observation are likely to exhibit different levels of contingency. This hysteresis portends an irreducible level of uncertainty, and has important ramifications for snowmelt and avalanche forecasting, extrapolation of point data, and process modeling. While the complexities of the snow system are well recognized, the concept of spatial and historical contingency must be addressed explicitly, and may guide development of methods for dealing with spatial and temporal variability in snowpack properties. This study explores an approach combining probabilistic prediction with synoptic typology to quantify variation in snowpack properties due to wind redistribution. This holistic mode of inquiry avoids reductionist, single-scale simplifications and addresses complexities in snow system processes within the context of spatial and historical contingency.
