What You Don’t Know About Snow: The USDA’s SNOTEL Network is Playing a Critical Role in Protecting Water Resources in the Western United States

Bookmark and Share

Written by Keith Bellingham
Stevens Water Monitoring Systems, Inc.

There are many complex political, social, environmental and scientific challenges surrounding water resources in the western United States. Water is scarce and does not occur when and where people need it the most. According to the US Geological Survey (USGS), about 80 percent of fresh water used in the western United States is appropriated for irrigation. Over the past 150 years, society has diverted streams, built wells and created reservoirs to better distribute water to where it is needed the most.

Over the past 30 years, the impact to the environment of water withdrawals from natural sources has become more evident with increasing demand. Some of the environmental challenges associated with water withdrawals from streams include maintaining enough flow to support the habitats, increase temperature, and nutrient loading.

For more than 100 years, knowing and predicting stream flow has been very important for water managers, not only to provide enough water for human consumption, but to protect the aquatic habitats in our natural water ways as well. Companies such as Stevens Water have been providing instrumentation that is critical to the prediction of water resource availability in the western United States and throughout the world.

History of SNOTEL

Much of the water in the western United States comes from the winter snowpack in the mountainous regions. The snowpack in the mountains of western US can range from nothing or very little up to 30 or 40 feet of snow in the high Cascades.

In 1906, a Hydrologist at the University of Nevada, Dr. James Church, began to document the relationship between winter snowpack in the mountains and stream flow throughout the year for certain watersheds. Dr. Church enhanced existing Russian technology for measuring snow water equivalent (SWE). Shortly after Dr. Church developed these snow measurement techniques, the US Department of Agriculture began to construct “Snow Courses” in the mountainous areas of the west so that hydrologists could make stream flow predictions from snow data.

Dr James Church
Dr. James Church in 1906. Picture compliments of the USDA National Resources Conservation Service

These snow courses were areas free of trees where the snow survey staff could take manual measurements of the snowpack. About that same time, the USGS began installing stream gauging stations so that stream data could be compared to the snow data. In 1911, these USGS Stream gauging stations began using mechanical chart recorders, an innovative new technology for automatically measuring water level developed by J. C. Stevens, one of the founders of Leupold & Stevens, which later became Stevens Water Monitoring Systems.

Stevens Type F Chart Recorder from the 1960s  Current Stevens Type F Chart Recorder
Left: a Stevens Type F chart recorder from the 1960s. Right: the current production model of the Stevens Type F chart recorder.

Starting in the 1980s, the USDA’s snow courses became more sophisticated, adding an array of weather sensors, data loggers and telemetry systems. These snow course telemetry sites were named SNOTEL. Today, the US Department of Agriculture, National Resources Conservation Service, manages and operates over 600 (and growing) SNOTEL stations. The hourly data is now displayed on the internet for every station. The data from SNOTEL is of high quality, and SNOTEL is known the world over for having the one of best quality control protocols of any environmental network.

What is a Snow Course and what is the significance of Snow Water Equivalent?

For the traditional stream flow prediction models, the parameter Snow Water Equivalent (SWE) was needed. SWE is the amount of water contained within a core of snowpack. This is a manual measurement where a technician would push a preweighed cylindrical tube into the snow. The tube is then weighed to get the weight of the snow. From this weight, they are able to determine the amount of water in the snow. The density of snow can change with temperature and precipitation throughout the year. The same depth of snowpack can yield different water amounts depending on the density. With the SWE measurement, apples to apples comparisons can be made with snow data across regions and time.

SNOTEL Components

While this manual SWE measurement method is still used on most snow courses several times a year, SNOTEL has automated ways for collecting information about snowpack. Each SNOTEL site is equipped with a radar sensor that can provide snow depth, a precipitation gauge that that measures the total amount of precipitation (both solid and liquid) using a pressure transducer inside of a collector, and a snow pillow. A snow pillow is a big bladder filled with a non-toxic liquid antifreeze solution. As the snowpack builds on a snow pillow throughout the winter, the antifreeze is displaced up a stand pipe. From the pressure of antifreeze in the stand pipe, a SWE is calculated. SNOTEL sites also collect air temperature, wind speed and direction, relative humidity, barometric pressure, and solar radiation data.

Typical configuration of SNOTEL site
Typical configuration of a SNOTEL Site. Picture complements of the USDA Nation Resources Conservation Service.

How Are Stream Flow Forecasts Calculated?

The stream flow predictions from the USDA’s SNOTEL data are derived from the statistical relationship between the SWE on April 1st and the stream flow throughout the summer. The snow courses are standardized plots of ground where a transect of SWE measurements can be taken. These SWE measurements need to be collected consistently year after year so that the hydrologic trends can be statistically quantified.

Based on many years of historical (antecedent) data between the snow courses, SNOTEL and the USGS stream flow data, a mathematical algorithm can be generated from a matrix method to correlate the data so that a stream flow prediction can be generated. The comparison between the stream flow prediction and the actual flow is called “skill”. The closer the skill is to 1.0, the closer the prediction was to the actual stream flow. The skill is the same as a Pearson correlation coefficient or r squared value used in statistics. Many stream flow forecasts provided by SNOTEL have a skill of 0.9 or greater.

USDA’s SNOTEL Network across the western United States.

How Does Stream Forecasting Work?

The stream flow forecasts or Surface Water Supply Index (SWSI) is calculated for specific points along streams and is SNOTEL’s main publically available product. Stream flow forecasts are available for over 750 locations in the western US. The SWSI is given as a distribution of probabilities that accommodates a wide number of applications depending on the user’s interest.

USDA stream flow forecast example

The above table is an example of a typical stream flow forecast from April to July in units of 1000 acre feet of water. The 50% exceedance forecast provides the percent of average. In this example, the out flow of Emma Lake is 65% of average for April to July of this particular year. An irrigation district that withdraws water directly downstream of Emma Lake outlet would plan for the 70% exceedance. There is a 70% chance that there will be at least 345,000 acre feet of water flowing out of the lake from April through July and a 30% chance that there would be less than 345,000 acre feet of water. Because in this example, the 50% exceedance forecast only shows 65% of average, the irrigation district would use caution and use the 90% exceedance forecast to be on the safe side. There will be a 10% chance that there will only be 285,000 acre feet of water flowing out of the lake from April to July.

A hydrologist at the highway department planning for flood conditions would use this forecast differently. There is only a 10% chance that there will be more than 485,000 acre feet draining out of the lake for this time period. Like the irrigation district, the highway department can look at the USGS stream gauging station to get actual daily rates. In this example, the forecast is 65% of average, so the highway department should not be that concerned about flooding, while the irrigation district needs to use more caution with water allocations.

Stream Flow and Soil Moisture

Traditionally, stream flow forecast models primarily use the antecedent SWE values in their calculations and ignore soil moisture. Soil represents a huge water reservoir and can introduce errors into the forecasts. Excluding runoff, water from snow melt and precipitation will first percolate through the soil before entering the water table or saturated ground water zone. Once the water is in the saturated zone, it will travel down gradient and eventually discharge into a stream or lake. The vadose zone is the soil above the water table and represents a hydrological regime that can hold large amounts of water.

Once water enters the vadose zone, it will only move in one dimension, up and down. In unsaturated soils, water will migrate upward because of evaporation and the uptake of water by plants and trees. This upward movement of water is called evapotranspiration. Evapotranspiration is the primary mechanism responsible for removing water from soil. During the winter months with a snowpack, evapotranspiration is almost zero, and the soil moisture values stay relatively constant.

Groundwater flow diagram
Groundwater flow diagram, taken from the USGS Report 00-4008.

The downward movement of water in soil obeys an entirely different set of rules. Water will suspend itself and adhere to soil particles. This attraction between water and soil particles is called capillary force. As the soil moisture increases, gravity will pull the water downward. The point at which the gravitational influence exceeds the capillary influence is called field capacity. Above field capacity, water will be conducted downward through the soil and will discharge into the water table. If the soil moisture stays below field capacity, water can only travel upward due to evapotranspiration.

Field capacity is an important parameter in determining the fate and direction of water transport in soil. The field capacity of a particular soil is usually a function of the soil texture. In general, the more clay-rich the soil is the more water it will hold on to.

water availability graph for various soil types
A graph illustrating the availability of water at different soil moisture levels across various types of soil.

The soil moisture value before the winter snow arrives will be the same soil moisture value in the spring when the snow begins to melt because evapotranspiration will be negligible in the winter. If there is a dry fall, and if winter arrives quickly, the soil moisture under the snowpack will be low.

When the snow melts in the spring, much of the water will be retained by the soil, and not as much water will reach the streams. There can be a below- average stream flow even if there is an above-average snowpack because the soils dried out the previous fall. Conversely, if there is a rainy autumn, the soil will already be at field capacity when the snow melts in the spring, and all of the water from the snowpack will enter the water table pushing an equal amount of water out into the streams. A wet fall can cause flooding in the spring even if there is a below average snowpack.

Because soil moisture was recognized as a major factor of the hydrology of a watershed starting in the late 1990s, STOTEL began installing high quality soil sensors at SNOTEL sites. The soil sensor selected to go into SNOTEL sites was the Stevens Hydra Probe. Now there are almost 300 sites nationwide that are equipped with Stevens Hydra Probes.

Even though many SNOTEL sites are equipped with Hydra Probe Soil Sensors, the official stream flow forecasts do not yet include soil moisture data in their models. Recently, the SNOTEL office in Utah began evaluating soil moisture under snowpack. They suggest a correction to the forecasts based on a parameter called soil moisture deficit. The soil moisture deficit is the difference between the current soil moisture and the soil’s field capacity and represents the amount of water that can enter into the soil before migrating downward to the water table. The soil moisture deficit can then be used to adjust the chance of exceedance forecasts. While this technique shows promise in improving forecasts by incorporating soil moisture data, it is not yet used system wide. More new SNOTEL sites equipped with Hydra Probes are scheduled to be installed.

Installation of a Stevens Hydra Probe into native soil at a SNOTEL site
Installation of a Stevens Hydra Probe into native soil at a SNOTEL site.

SNOTEL, Climate Change and Other Environmental Issues

The Clean Water Act’s “303d stream listing” is a list of streams that do not meet the water quality standards required by law. Streams on the 303d list are said to be “impaired”. While streams in the US no longer catch on fire from excessive pollution, they can still become impaired from storm water runoff and from agricultural/irrigation land uses.

As water is withdrawn from streams for irrigation purposes, the temperature may increase, as well as the salinity and nutrient level. This loading of temperature, nutrients and salinity causes algal blooms and threatens the aquatic ecosystems. The demand for water from urban areas and for irrigated crops is increasing every year. The steam forecasts provided by the NRCS offer part of a potential solution for water conservation in western US.

Climate change is also a concern for many water managers and stake holders. Climate change can of course affect the quantity of the snowpack, but, it also impacts the timing of the melt. Timing of the availability of the water dictates the management and the operation of reservoirs, which in turn affects power generation.

The SNOTEL Network also provides high quality data for scientists modeling climate change. SNOTEL offers uniform, hourly data in real time for hundreds of watersheds in mountainous regions, and it is all available to the public for free.

The quality control policies for SNOTEL are mimicked by other network and mesonets in the US and in Europe such as the NRCS’s Soil Climate Analysis Network (SCAN), NOAA’s Climate Reference Network and the Kentucky Mesonet. Some of the policies include strategic placement of sensors to capture the particular characteristic of the site. For example, air temperature sensors and precipitation gauges are placed high enough off the ground so that they do not become buried in snow. The precipitation gauges use elaborate wind shields that help diminish the effects of wind on precipitation measurements. Each station is also ground truthed several times a year by highly trained staff that can collect measurements as well as perform maintenance on the stations, ensuring accurate readings year-round.

The data is also evaluated with special modeling software called Parameter-Elevation Regression on Independent Slopes Model (PRISM) developed at Oregon State University. PRISM statistically examines the data from SNOTEL sites and can calculate the variability of the data. PRISM can make regional prediction of the parameters. For example, if an air temperature gauge is suspected of malfunctioning, PRISM software can take temperature data from other nearby SNOTEL sites and other sources and calculate what the expected temperature would be at the suspect site. If necessary, the data archives can be edited to provide the highest quality data possible.

Water Rights in the Western United States

Water rights in the western United States are exceedingly complex and vary from region to region. In general, water rights are based on Prior Appropriation which means whoever got there first has the senior water rights.

An example of irrigation district river diversion in Oregon
An example of irrigation district river diversion in Oregon.

The availability of water determined where people settled in the west in the 1800s. Reservations were created for the Indian tribes which often time included large amounts of land high in the mountains that hold a snowpack much of the year. In 1908 the US Supreme Court ruled that the Indian tribes own the water rights on their reservations. Because snow occurs up gradient from the irrigated lands, as the snow melts, it travels though several if not many jurisdictions before returning to the natural water body.

Often times, the first water jurisdiction is snowpack on Indian Reservations, however many tribes have chosen not to exercise their water rights. Structures that divert state or tribe owned natural waters are built and paid for by the US Bureau of Reclamation.

Measurement of water flow in an irrigation ditch, using the Stevens AxSys Flowmeter
Measurement of water flow in an irrigation ditch, using the Stevens AxSys Flowmeter.

The US Bureau of Reclamation will then sell the water to irrigation districts, municipalities, and other water users that hold or service holders of water rights. Under the US Department of the Interior, the US Bureau of Reclamation is the nation’s largest wholesale supplier of water and the nation's second largest producer of hydroelectric power. Its facilities also provide substantial flood control, recreation, and fish and wildlife benefits.

When water passes though jurisdictions, the stakeholders and the holders of the water rights don’t always come to full agreement on the pricing, usages, Total Maximum Daily Loads, and ownership of the water. These disagreements can erupt into a “water war” where SNOTEL and USGS data are often times used in litigation. In 1969, the Nation Environmental Policy Act (NEPA) was established to help settle disputes while using sound environmental assessments of the particular issues.

The Snow Science Community

The Western Snow Conference started in the early 20th century as the American Geophysical Union’s committee on snow hydrology, and Dr. Church was the chair. In the 1930’s, a special conference was held highlighting the advances in stream forecasting from snow surveys. Since the 1930’s, the Western Snow Conference was held every year.

Western Snow Conference visit to the Columbia Ice Sheet in 2009
Western Snow Conference visit to the Columbia Ice Sheet in 2009.

To this day, the yearly Western Snow Conference continues to promote snow science and stream forecast modeling. At the yearly conference there are technical exhibits, a forum of talks and a peer reviewed journal. The organization is divided into four sub-organizations based on region. Some of the topics at this year’s Western Snow Conference included effects of dust on the melt rate of snow, water management, the challenges of hydrological modeling, the bark beetle and other invasive species, and prediction of landslide hazards based on SNOTEL data.

For almost 100 years, snow surveys in the western US have helped resolve some fundamental issues regarding the availability of water. With increasing water demand, climate change, and changes in land use, we will need to rely more and more on monitoring and modeling technology to help protect valuable water resources. SNOTEL, in cooperation with the USGS, Bureau of Reclamation, NOAA and other agencies, is providing products such as real time data and forecasts free of charge to help water resource managers meet the new challenges facing water in the west.