Technical Attachment T-01-01

(ARH 02/26/01)

The Natural Resource Conservation Service (NRCS) has established automated snow telemetry (SNOTEL) sites to monitor the seasonal snowpack in the western United States and to obtain an index of snow water equivalence over a region. SNOTEL sites consist of snow pillows, a storage precipitation gauge, and an air temperature sensor. Some SNOTEL sites also have wind, snow depth, and soil temperature sensors. There is an increasing number of automated snow measuring sites in Alaska.

Snow water equivalence (SWE) is simply the amount of water that would be produced if the snow were instantaneously melted at a point. The water equivalence of a snowpack is measured by snow pillows. Snow pillows are synthetic rubber envelopes about six feet in diameter containing an antifreeze solution. Accumulating snow exerts pressure on the pillow and solution. A transducer converts the weight of the snowpack into a measurement of SWE in inches. Hydrologists are concerned with SWE on the ground as this variable is a prime indicator of the volume of water that will reach the rivers and streams during the melt season.

Snow depth sensors are being installed at various locations throughout Alaska. These acoustic sensors measure the vertical distance downward to the snow surface and then calculate the snow depth beneath the sensor. Such snow depth sensors are very accurate during calm conditions. However a few hourly readings may be erroneously high during storms as the falling snowflakes can impede the distance measurement. With no snow cover present and in locations where the depth sensor is not directly above a snow pillow, the depth sensors will measure grass height. Therefore, some summer and autumn readings of snow depth can be in error.

Uses of Snow Data

Hydrologic Forecasting

Since snow depth and SWE exhibit great spatial variation due to topography, radiation, and wind redistribution, point values of snow depth and SWE should be used as indices rather than absolute values. The NRCS uses the measured SWE at SNOTEL sites and other variables such as antecedent streamflow, soil moisture, temperature, and fall precipitation in a regression analysis to predict snowmelt runoff for water supply forecasting and drought monitoring. Additional information on NRCS streamflow and water supply forecasts can be found at http://www.wcc.nrcs.usda.gov/. Specific Alaska snowpack reports can be found at http://ambcs.org/.

The Alaska-Pacific River Forecast Center (APRFC) utilizes data from SNOTEL sites to help estimate basin-wide SWE in our hydrologic models. Good estimates of areal SWE are necessary to accurately forecast streamflow. The APRFC begins to issue streamflow forecasts for selected river basins at the beginning of the melt season. These forecasts rely on current conditions of the snowpack, soil moisture, mean areal precipitation and temperature. During the melt season, hydrologists at the APRFC carefully watch the snowmelt rates (as evidenced by declining values of SWE) at SNOTEL sites to ensure that our hydrologic models are handling snowmelt adequately.

As hydrologists and water supply forecasters are concerned with the seasonal accumulation and ablation of the snowpack, others are more concerned with daily or incremental interpretations of snow data.

Avalanche Forecasting

Precipitation intensity is just one of many meteorological factors that are considered in snowpack stability assessments. Precipitation intensity is a direct measure of the load applied to the snowpack and is important since failure that causes avalanches is only produced when a critical rate of loading is applied. Avalanche forecasters may look at the average precipitation intensity as well as the storm duration and maximum or hourly intensities by reviewing hourly data from SNOTEL sites. In general, a prolonged period of hourly SWE increases of 0.1" or more is a sign of very heavy snowfall and possible increased snowpack instability. Air temperature trends during a storm are also considered in snowpack stability assessments. Warm temperatures at the onset of the storm followed by cooling during the snow event has different avalanche hazard implications than a warming trend during the snow event. The latter temperature trend can lead to an "inverted cake" scenario in the new storm layer. The increasing air temperatures during the snow event could form a slab with higher density snow overlying lower density snow, increasing the potential for avalanches.

Weather Forecasting/Updating

Weather forecasters use automated snow data (SNOTEL and snow depth sensors) to update or verify snowfall forecasts. If a SNOTEL site does not have a snow depth sensor, then one must back-calculate depth by dividing the SWE by the estimated new snowfall density. Snow density is expressed as the percent of density of liquid water, and most new snowfall has densities between 0.04 and 0.10. Generally, the lower density values are found in colder, drier, continental climates and higher densities are prevalent in warmer, wetter, maritime climates. However, average snowfall density is a function of cloud temperature, air temperature and available atmospheric moisture, so regardless of location, new snowfall densities will vary depending on the storm track. Therefore, forecasters would not necessarily apply the commonly used 10:1 ratio of snowfall to water equivalent in their conversions of SWE data to new snowfall. Densities of new snowfall on the order of 0.10 are often found in the coastal mountains of Alaska, but new snowfall densities are often lower (0.06-0.08) elsewhere in Alaska. Looking at air temperature data from SNOTEL sites during a storm can give a rough indication of the new snowfall density. A storm that produces 1" of SWE can yield 8-10" of new snow at lower elevations along the Gulf of Alaska coast or Southeast Panhandle, or upwards of 14-18" of new snow at interior locations.

General Use

Backcountry recreationalists may want to use such automated snow data as a starting point in their own weather, snowpack, and avalanche hazard assessment. For instance, consider two storms at the popular recreational area of Turnagain Pass on the Kenai Peninsula. The Turnagain Pass SNOTEL site at an elevation of 2000' above sea level (pass elevation is 1000'), records 2.0" SWE during a storm where air temperatures hovered in the low 30's (F). This storm likely produced a heavy, wet snow and, depending on the rate of snowfall (precipitation intensity), there may be high instability in the snowpack due to rapid loading. Now consider a different scenario where the SNOTEL site received only 0.4" of SWE but air temperatures during the snowfall event were near 20 (F). While not adding a heavy load to the snowpack, this new snowfall is probably much drier, and the light and powdery snow could provide excellent skiing conditions. Of course, here we are only considering new SWE and air temperatures. Terrain, wind speeds and directions, air temperature trends during the storm, and old snow surface conditions also play major roles in determining the quality and safety of skiing/climbing/snowmachining conditions. Backcountry recreationalists may want to track the snow data through the course of the season to understand how various weather patterns and trends affect the snowpack conditions and verify this by digging snowpits.

As the accumulation season progresses and the snowpack becomes deeper, the density of the snowpack will increase due to compaction and changes in the snow crystal structures. By April, deep mountain snowpacks may have densities between 0.30 and 0.50. Therefore the 10:1 ratio cannot be applied when converting total SWE on the ground to total snow depth or vice versa.

It is important to note that the automated snow data is point data, and snow conditions can have extreme variations over short distances. The snow depth or snowfall at one location can be very different from the depth or snowfall at another location-depending on elevation, aspect, wind redistribution, etc. The automated snow data provides valuable information in remote locations, but values should only be used as an index of conditions in the area.