Relative Snow Accumulation Patterns and Inconsistencies Across the Southern Rockies, U.S.A.

Abstract:

Understanding spatial patterns and variability in the distribution of snow is critical in determining the timing and magnitude of snowmelt runoff. For snowmelt hydrology modeling, due to cost limitations only several point measurements are taken within a given watershed to represent basin-scale snow processes. To necessitate this, it is often assumed that at coarse resolution there is consistency in the spatial distribution of snow. This study examines this assumption of temporal consistency in spatial snowpack patterns. The focus areas is the Southern Rockies which is a key region for water supply to millions, as it is the headwater of four major western U.S. watersheds (Colorado, Platte, Arkansas, and Rio Grande Rivers). Within this area there are 90 long-term snow telemetry (SNOTEL) stations operated by the Natural Resources Conservation Service (NRCS) that have measured and recorded snow water equivalent (SWE) for at least 26 years. For each year, the daily SWE time series for each SNOTEL station was compared to each other station’s time series for the period when both stations were accumulating snow, yielding an annual relative accumulation rate between each station pair. The inter-annual variability of the accumulation rate among all combinations of SNOTEL pairs was computed and evaluated as a function of distance between SNOTEL station pairs, analogous to a semi-variogram. A general additive model approach was used to explain the variations in relative accumulation rates among the SNOTEL station pairs. Variation in accumulation among station pairs is highly correlated to distance, although there is no clear abrupt change in accumulation rates explained by distance, as is often present as the correlation length in semi-variograms. Elevation and land cover differences further explain observed variations in relative accumulation rates.