Water resources Research Papers - Academia.edu (original) (raw)

Distribution of snow water equivalence (SWE) was measured in the Emerald Lake watershed located in Sequoia National Park, California, by taking hundreds of depth measurements and density profiles at six locations during the 1986, 1987 and 1988 water years. A stratified sampling scheme was evaluated by identifying and mapping zones of similar snow properties on the basis of topographic parameters that account fur variations in both accumulation and ablation. Elevation, slope, and radiation values calculated from a digital elevation model were used to determine the zones. Of the variables studied, net radiation was of primary importance. Field measurements of SWE were combined with the physical attributes of the watershed and clustered to identify similar classes of SWE. The entire basin was then partitioned into zones for each survey date. Statistical analysis showed that partitioning the watershed on the basis of topographic and radiation variables does produce superior results over a simple random sample. I N'I'R() I) UCTION Recent pressure on hydrologic resources caused by population influx and resource development increases the need for accurate measurement ol'snow water equivalence (SWE) in alpine regions. In Calil•)rnia, for example, agricultural and metropolitan areas depend on water obtained from the Sierra Nevada to supply a large portion of their water needs. Most of the runoff !Yore alpine environs is melt from the seasonal snowpack. To understand the timing and volume of runoff, a good understanding of the spatial variation of snowpack properties is needed. With the use of both established and recently developed techniques, SWE measurements at a given location are not difficult to obtain [U.S. Army ('orps o./' Engineers, 1956; Dunne and Leopold, 19781. Several accurate methods for measuring density exist, ranging from those involving excavation and sampling pits [Perla and Martinelli, 1978] to the isotope profiling gauge [Kattelmann et al., 1983]. Depth measurement requires only a robust probe and some experience in use. The persistent question is, How do we accurately interpolate between measurements at points to estimate the total volume of water stored in the snowpack over an entire drainage basin? Snowpack properties may vary greatly over small distances. Numerous studies have been conducted in prairies or regions of mild relief [Steppuhn and Dyck, 1974; Adams, 1976; Granberg, I979], and snowpack variation in these places is better understood than spatial and temporal variations of snow cover in alpine regions. The factors contributing to variation in SWE (slope, aspect, elevation, vegetation type, surface roughness, energy exchange) are exaggerated in alpine areas, resulting in a heterogeneous snowpack that changes markedly in space and time. We need sampling methods that can capture snowpack variability and characterize it over an area, that have reasonable time and manpower re, quirements, and yet can accurately assess the snowpack. An approach that requires many samples throughout a basin is seldom practical, given logistical constraints of safety and time. In this study we Copyright 1991 by the American Geophysical Union. Paper number 91WR00506. 0043-1397/91/91 WR-00506505.00 attempted to accurately determine the distribution of SWE over a small alpine basin by identifying and mapping zones of similar snow properties on the basis of topographic and radiation parameters that account for variations in both accumulation and ablation. Parameters used were elevation, slope, and daily integrated net solar radiation calculated for clear atmospheric conditions and an assumed surface albedo. We tested different classification combinations of net radiation, slope, and elevation and varied the number of classes between eight and twelve. The objective was to see if an improved sampling scheme could be developed by stratifying the sample on the basis of physical parameters that control the accumulation and distribution of snow in an alpine watershed. Snow depth and density measurements were obtained in eleven intensive snow surveys over three melt seasons, providing a large sample of spatial point measurements for model development and testing. FACTORS AFFECTING SNOW DISTRIBUTION Investigations on snow accumulation and distribution in the last two decades have focused on elevation, vegetation, and topography [Meiman, 1968]. Although much of the work has been done in regions of low elevation and minimum relief, many of the results apply to alpine areas. Properties of the snowpack (e.g., depth, density, temperature, chemistry) vary in space and time. Snow depth and density are controlled by both accumulation and ablation. On a large scale these processes are controlled by meteorological patterns and major terrain features and on a small scale by redistribution, new snow properties, and micrometeorology. Accumulation consists of two processes: snowfall itself and redistribution of the original snowfall by wind transport or by sloughing and avalanching. Ablation occurs by melting, sublimation, and deflation. Accumulation Snowfall. Precipitation, including snowfall, is a stochastic process, and its variability must be considered on a wide range of scales. Regional climate and latitude affect snowfall, but neither of these vary significantly within most alpine basins. Elevation is considered the single most important factor in snow cover distribution by most of the recent 1541 1542 ELDER ET AL.: SNOW DISTRII[IIJTION IN AI.PINt; WA'II:,RSH}•I)S studies, but orographic effects depend more on slope and wind speed than on elevation [Gray, 1979]. Wind. Much of the spatial heterogeneity of SWE in alpine regions is the result of redistribution by wind. Even if snowfall were uniform over an area, the final deposition pattern would be irregular, because snow is typically moved by wind and redeposited during the storm. Snow is analogous to other sediments and tends to accumulate in areas where flow decelerates or diverges, and it tends to erode in areas of accelerated or convergent flow. Where terrain irregularities and wind patterns are consistent in time, drifts and scoured areas tend to repeat in form and location, year after year. Drifts may shift between storms as the storm track changes, but over a season, consistent patterns still often emerge. Transport of snow by wind is reviewed in detail by Schmidt [1982]. Avalanches and sloughs. Considerable volumes of snow may be moved by avalanches in a watershed, and snow tends to repeatedly slough i¾om slopes that are sufficiently steep. Avalanching does not change the total mass of snow in a drainage basin but does change the distribution. Correct estimates of the volume in avalanche deposits are hydrologically important because they may contain large amounts of water. Zalikhanov [1975] found that 30-64% of the alpine snow cover in the Caucasus may be transported to valley bottoms by avalanches. Ablation A common method to evaluate ablation and snowmelt is through evaluation of the surface energy exchange. Snowpack ablation is controlled by energy exchanges at the air/snow and snow/ground interfaces. Energy inputs may come from solar and emitted atmospheric radiation, sensible heat exchange, latent heat exchange, heat flux from the underlying substrate, and advective heat transfer. Of the available energy sources, it is well documented that solar and longwave radiation usually dominate [Zuzet and Cox, 1975]. Radiation affects net accumulation through ablation at the surface. If the melt only percolates into the snowpack and refreezes, then depth and density have changed but SWE has not. Once meltwater reaches an ice lens or the ground, however, it may move laterally, and the SWE at that point will change. Radiation thus influences the spatial element of accumulation because it may effectively move SWE from discrete parts of the basin where the energy balance is positive or remove SWE when runoff leaves the basin. ..