Understanding Snow

With winter snowpack ranking as Colorado’s largest reservoir — and its least understood — scientists and water managers are developing new insights into the factors influencing snow accumulation and runoff. The upshot? Communities get invaluable glimpses into water futures affecting farmers, recreationists, aquatic wildlife, and everyone else.

Every April 1, when most Americans are plotting April Fool’s pranks, water managers across Colorado fixate on the season’s streamflow forecasts. Initial predictions are released as early as January, but April 1 reigns as the most pivotal forecasting date because it’s late enough in the spring to mark the snowpack’s peak but early enough that runoff hasn’t started. It’s also early enough that irrigators and water managers can use these predictions to plan ahead. So for many, April 1 is considered the day when they find out how much water they will likely have that summer and when it will arrive. It’s like learning the balance of your bank account—because for this state’s farmers, cities, river outfitters and others, water is akin to wealth. And most of Colorado’s water starts as snow.

The blanket of frozen water crystals that covers our high country from November through May ranks as the state’s largest water reservoir by volume. And that melting snow has tremendous influence over our summertime streamflow patterns: Snowmelt runoff provides as much as 80% of Colorado’s annual surface water supply. 

So assessing the snowpack and how it will feed Colorado’s streams, rivers and lakes gives people a glimpse into the future. Food producers choose which crops to plant based on expected water deliveries, recreational guides may tweak their scheduled trips and staffing levels to suit projected water levels, and dam operators decide how to budget their releases to fortify reservoirs and regulate downstream flows. 

River outfitters rely on runoff forecasts when scheduling trips and staffing for the season. Getty Images

At the Colorado-Big Thompson Project—which is operated by Northern Water and comprised of 12 reservoirs and 35 miles of tunnels that deliver water to communities and farms along the northern Front Range—managers use snowpack data to predict how much water they will need to tap from West Slope sources, and how much they can provide to their municipal and agricultural shareholders on the plains. Streamflow forecasts guide managers in making allocations, which tell Broomfield, Boulder, Fort Collins and other cities whether they will have excess water available to lease back to farmers. And with ample snowmelt in ditches and canals, farmers might opt to plant thirstier crops that offer a higher profit margin. 

“Snow [data] also tells us how to move water through the system,” explains Northern Water spokesperson Jeff Stahla. Last winter, for example, snow monitoring confirmed that Lake Granby would fill by natural means, so water managers never turned on the pumps that draw water to Lake Granby from Windy Gap Reservoir. “That allowed us to avoid the unnecessary cost of pumping,” says Stahla.

But measuring snowpack and translating that to predicted runoff volume isn’t easy.

The challenge stems, in part, from Colorado’s high elevations and dramatic terrain, which create significant variations in precipitation patterns across relatively short distances. The same storm that deposits a scant inch of snow on my driveway in Steamboat Springs may dump two feet of powder on the summit that rises nearly 4,000 vertical feet above my home. It’s hard to account for all those variations in accumulation.

Yet the current boom in snowpack monitoring and modeling technologies is already providing water managers with a better crystal ball. And with a sharper view into the summer’s water supplies, planners can make every drop count for both residents and the environment.

Gathering snowpack data 

In 1903, Enos Mills, the naturalist and mountain guide who was instrumental in creating Rocky Mountain National Park, became Colorado’s first snow observer, skiing or snowshoeing into the high country and reporting on snow and forest conditions. 

Monitoring grew more formalized in the 1930s, when the Natural Resources Conservation Service (NRCS) began to create “snow course” sites where observers manually collected snow samples and calculated their water content. 

These days, the NRCS continues to employ some of those snow courses but also uses automated Snowpack Telemetry (SNOTEL) stations equipped with “snow pillows” that calculate snow water equivalent (SWE), or the amount of water contained in snowpack. These broad, 10-foot diameter bladders are filled with liquid antifreeze and buried level with the ground, so that the weight of falling snow exerts pressure on the antifreeze, which is then measured and converted into SWE. Typically located in sheltered forest clearings on flat ground between 8,500 and 11,600 feet in elevation, SNOTEL sites look like buried alien landings with variously shaped metal towers reaching out of the snow to measure conditions such as temperature, wind, precipitation and snow depth.

Brian Domonkos with the Natural Resources Conservation Service takes measurements at the Wolf Creek SNOTEL site in southwestern Colorado. Credit: Christi Bode Skeie

Such stations feed a data set that’s prized for its longevity. The record measuring snowpack and its moisture content now extends for nearly 100 years in some locations, providing historical averages that lend valuable context to the current year’s snowpack. Water users and managers rely on station data and the resulting forecasts to anticipate operations for the coming spring and summer.

But the measurements from SNOTEL stations and snow courses, while hugely valuable, only reveal so much about the state’s snow and how it may melt into waterways. They’re like scattered individual pixels in a big picture that forecasters and water managers would prefer to see in its entirety, at as high a resolution as possible. That’s especially true because most of Colorado’s water managers now struggle to consistently deliver desired water quantities to all stakeholders. “When you have less, you’d better know more, because every drop counts,” explains climate researcher and consultant Jeff Lukas.

Not only are water managers operating on ever-smaller margins of error, but they’re also experiencing greater climate variability from year to year. “Even if there’s no shift in the mean, the variance is getting bigger, and the year-to-year swing is getting bigger,” explains Jeff Deems. Deems helped NASA employ LiDAR sensors, or Light Detection and Ranging, for snow depth mapping, before co-founding Airborne Snow Observatories, Inc. (ASO), a relatively new player on the snowpack monitoring, modeling and forecasting scene.

Indeed, snowpack research and data gathering is surging. The NRCS is updating its SNOTEL stations and forecast models. Other entities, such as the U.S. Geological Survey (USGS) and ASO, have launched new monitoring options and related predictions. Advances are being made with research too, as lasers, artificial intelligence, satellite imagery, and cosmic ray neutron sensors all provide novel ways of assessing the volume and release patterns of Colorado’s largest reservoir. “We’re in a really exciting period right now,” says Deems, describing a revolution of snowpack assessment that’s generating new ideas, new sensor technologies, new runoff models, and new ways of collaborating over the emerging paradigms.

Forecasts for the future

Until 2013, two agencies, NRCS and the National Oceanic and Atmospheric Administration, or NOAA, collaborated to produce unified runoff forecasts at locations across the Upper and Lower Colorado River Basins and the eastern half of the Great Basin. NRCS processed the data from its SNOTEL stations into a statistical regression model that predicted the current season’s runoff based on past years’ snowpack amounts and runoff. NOAA fed NRCS data a more complex forecast model that assessed the processes acting upon the snow, such as temperature and snowmelt rate, and offered a prediction for the coming spring. The agencies compared their two forecasts and offered the public a single monthly forecast for each location from January 1 through June 1.

But starting in the 2013 water season, the agencies began publishing their own independent projections, marking the beginning of data multiplicity in streamflow forecasting. “We realized that we were washing out some important information from both our models, so we stopped issuing one single number,” says Paul Miller, hydrologist for NOAA’s Colorado Basin River Forecast Center. “The vast majority of the time, [NRCS and NOAA] issue very similar forecasts, but having more than one source gives users a greater degree of confidence,” Miller explains. Even when those forecasts don’t align, having multiple predictions helps users better prepare for runoff.

Both agencies have continued to refine their forecast models using the latest capability in computation. NRCS hopes to expand the SNOTEL network in Colorado over the next three years with a handful of new locations that impact multiple large watersheds. 

NRCS also continues to update its SNOTEL stations. For the first time, this past winter from 2023 to 2024, three sites featured an expanded suite of sensors that include solar radiation and snowpack temperature. Says NRCS snow survey supervisor Brian Domonkos, “By beefing up these sensors, we’re watching more than just SWE but also impacts to it.” Understanding the factors that cause snow to melt and sublimate, turning directly into a gas, at different rates is now a major focus of snowpack monitoring.

Insights from dust on snow

Senator Beck Basin is a perennially cold, treeless cirque in the headwaters of the San Miguel, Uncompahgre and Animas rivers. This is Jeff Derry’s field office: He frequently drives up Red Mountain Pass and treks into study plots located above 12,000 feet. Here, an array of air, snow and radiation sensors record factors that influence snow’s behavior once it reaches the ground. But as executive director of the Silverton-based Center for Snow and Avalanche Studies, Derry also seeks information that instruments can’t record automatically. 

Researchers slice into snowpack, revealing layers of dust on snow. Courtesy Center for Snow and Avalanche Studies

As snowmelt season approaches, Derry digs pits in the snow to scrutinize layers of dust that have become a common snowpack feature in southwest Colorado. Wind patterns carry this dust out of neighboring Utah and Arizona and deposit it on the sky-kissing summits of the San Juans (and commonly on mountain ranges located across Colorado). When sun hits the dark-colored dust, it accelerates snowmelt—so Derry regularly slices into the layer cake to give water managers a prediction on when they’re likely to see related surges in runoff. Using dust on snow data from satellites as well as Derry’s in-field observations, NOAA’s Colorado Basin River Forecast Center tweaks its streamflow forecast model to account for the dust influence. Reservoir operators use Derry’s information to know when and how fast water will come down the mountain and impact the resources that they manage.

Dust covering snow adds to rapid melt. Courtesy NASA

These snow scientists and water managers are craving insights into storm and moisture trends that don’t always follow historical patterns. Dust, for example, has become more prevalent in Colorado’s snowpack as increasing human disturbances on desert soils loosen particles to prevailing winds. Dry soils in the mountain watersheds themselves are also influencing runoff volume. Extended periods of drought have left soils thirstier than they used to be, so they soak up more snowmelt and return less water to streams.

The monitoring equipment installed at various locations across Senator Beck Basin helps to quantify such unknowns. It’s an extraordinary site which, until the recent uptick in snow study, was one of just two sites in the West — the only site in the Upper Colorado River Basin — that collected the entire energy budget of the snowpack. This means that it measures all the variables, such as solar radiation and wind speed, that alter the snow after it has fallen. “Until very recently, we were the only high-elevation study basin collecting this data in Colorado,” says Derry.

The Center for Snow and Avalanche Studies has been doing this work for almost two decades: In 2004, the center received a grant from the National Science Foundation to purchase instruments and conduct groundbreaking research into the effects of dust on snow that are now well established. But Derry continues to monitor dust deposition and analyze the snow’s reflective power, known as albedo, so he can relay that data to regional forecasters. 

Water managers learn about snowpack data sources by doing fieldwork at the Center for Snow and Avalanche Studies’ “snow school” in Senator Beck Basin. Credit: Christi Bode Skeie

Although Senator Beck Basin is best known for its dust studies, the site also hosts snow researchers that test new instrumentation and monitoring approaches. The Durango-based Mountain Studies Institute, for example, initiated a new “snowtography” study this winter in an effort to understand how forest structures influence snowpack. Located in the trees adjacent to Senator Beck Basin, the instruments measure factors such as soil moisture, wind and ablation (snow loss) to observe how the forest canopy affects snow accumulation in open clearings, semi-forested areas, and dense stands. 

Given the prevalence of beetle kill, wildfire and other changes to Colorado’s forests, Derry says, “We need to account for how that influences snow and soil moisture in the high country.” 

Volunteers with the San Juan Outdoor Club take measurements at a snowtography site near Pagosa Springs. The snowtography equipment uses automated time-lapse photography to observe snow depth. Credit: Christi Bode Skeie

Next-gen monitoring 

Senator Beck Basin is also the location for a new USGS monitoring site—one of 14 new sites across the Upper Colorado River Basin that the USGS installed for full operation this past winter. It’s one focus of the agency’s national effort to establish Next Generation Water Observing Systems (NGWOS) that provide high-res temporal and spatial data for complex water environments. 

“We’re not trying to replace the monitoring that SNOTEL does,” says USGS research hydrologist Graham Sexstone. Instead, he explains, the USGS is striving to expand upon that system with stations situated in places where SNOTEL hasn’t traditionally operated—such as extremely high and unusually low elevations. Instead of positioning them in sheltered locations designed to measure snowfall, USGS monitoring sites occupy zones that are subject to wind redistribution, sun-exposed meltoff, and other physical factors influencing snow and the “snow to flow” equation. Says Sexstone, “We’re trying to represent conditions that we don’t already have a lot of data for.” 

Each site is equipped with new sensors to measure SWE, snow depth, blowing snow, soil moisture, meteorological variables, and more. USGS staff also visit each site periodically with a sled-based ground-penetrating radar unit to measure the snow depth, and to fly a LiDAR-equipped drone to track changes in snow depth. Researchers also physically dig snow pits at the sites to compare those findings against station sensor data. 

By using new sensor technology and monitoring zones that have either been lower or higher than the SNOTEL’s reach, the USGS hopes to give forecasters insight into marginal—but potentially influential—aspects of the snowpack. “When factors influence places where SNOTEL sites can’t be, you miss anomalies,” explains consultant Jeff Lukas. 

Lukas remembers how, in April 2010, when SNOTEL sites in Summit County did not capture the full extent of the disproportionately large alpine snowpack sitting above the sites, the magnitude of peak runoff in early June wasn’t well anticipated. Faced with surprisingly large flows pouring into an already-full Dillon Reservoir, Denver Water sent an unplanned slug of water to flow through the Roberts Tunnel to the South Platte River in order to avoid flooding downstream of Dillon. “Even if they don‘t happen very often, those near-crises are unnerving,” continues Lukas.

The USGS’ new monitoring efforts should provide more data points, and more warning, of such situations. Their data is fed into computational models that water managers use to crunch the numbers and predict when and how much water will enter area streams. Modeling turns discrete data points into real-world scenarios that irrigators, reservoir managers, cities and water utilities, and others can interpret and plan for. But as SNOTEL stations collect increasingly sophisticated data, that information is outpacing some models’ ability to digest and translate the data into forecasts. 

Theoretically, the growth in snow monitoring data will make forecasting better, says Lukas, “if you have more detailed information about the snow and are able to translate that into information about runoff.” Models must be adapted for the new data sets that emerging research provides, and those evolutions can take time. 

Yet even with more information and an expanded suite of sensors in more locations than ever before, point stations still represent just one pixel point in a vast and varying landscape of snow. That’s why water managers are extremely excited about a new option that may prove to be a forecasting game-changer: Called Airborne Snow Observatories, the flyover technology promises to provide Coloradans with their first big-picture, high-resolution view of the snowpack and highly accurate estimates of the water volume contained there.

The era of high-resolution imagery

Colorado holds the distinction of having hosted the first-ever ASO flight, from Grand Junction in 2013, when the technology was under development at NASA’s Jet Propulsion Laboratory. But it was California that first seized upon ASO’s potential to transform snowpack forecasting: That state’s Department of Water Resources has been commissioning flights over almost the entire Sierra Nevada range and Shasta River Basin since 2013. ASO data has helped California agricultural irrigators and aquatic wildlife managers budget their water and brace for anomalies. Now, the technology is here in Colorado.

“It’s the first operational advance in snow water equivalent monitoring since SNOTEL,” says ASO co-founder Jeff Deems. The flights read the terrain using two scanning technologies: A LiDAR laser maps the elevation when it’s dry and again when it’s covered with snow in order to measure snow depth, and an imaging spectrometer photographs the invisible wavelengths of light that indicate varying textures and impurities on the snow surface (and up to a foot below it). “When you know how much snow is there, and what color it is because of dust or other impacts to solar radiation, you can better quantify runoff,” Deems explains. And unlike point stations, ASO provides a spatially continuous picture over all aspects and elevations to provide the high-resolution portrait that water managers have yearned for.

ASO aircrafts map snow depth and albedo with laser scanners and imaging spectrometers. Courtesy of Jeff Deems, Airborne Snow Observatories, Inc.

In California, February 2022 ASO flights over the Feather River Basin documented just half the snow that had been suggested by point stations. By March, the ASO team documented that half of that already meager snowpack had melted off. “We triple-checked the data, but it ended up being correct because 60% of that watershed had burned within five years, and the soot from the standing dead trees had melted out the snow as the sun angles got higher,” Deems explains. That variability wouldn’t have been caught or predicted by existing melt models. So although ASO ended up being the bearer of bad news, “bad news early is better than bad news late,” he concludes. Advance warning gives communities the power to adapt to changing conditions.  

In Colorado, ASO flights have been providing water managers with SWE data since 2019, and surveys have grown in scope each year. For winter 2023-24, ASO has been surveying 6,200 square miles of watershed. The flights encompass the Upper Colorado River above Windy Gap, the Blue River, Clear Creek north through the South Platte and into the main fork of the Poudre River, the Roaring Fork, the Upper Gunnison, the Conejos, the Dolores, and the Yampa and Elk rivers. 

A 3D rendering of ASO snow depth data from the Elk Mountains in the Roaring Fork River Basin, taken on May 28, 2023. Courtesy of Jeff Deems, Airborne Snow Observatories, Inc.

Funding for these efforts comes from individual water providers, such as Northern Water and Denver Water, as well as the Colorado Water Conservation Board (CWCB) and a growing number of local, state and federal partners. These partners have been collaborating and funding the program through what’s known as the Colorado Airborne Snowpack Measurement (CASM) workgroup since 2019. 

The technology is so expensive that individual communities and regions often can’t afford to commission their own flights. However, the benefits are great—and, says Deems, they’re well worth the cost. He points to a 2019 success in California’s Kings River Basin where ASO data saved stakeholders approximately $100 million in avoided water lease costs to meet water delivery obligations. 

Colorado is already seeing its own gains. Last winter’s ASO flights across the Fraser River watershed provided Denver Water with better-than-ever data on how much snow remained on that region’s mountains. Using that knowledge, the utility was able to meet its needs while leaving more water in the Fraser River. The higher resolution snowpack data gave Denver Water the confidence to make localized decisions that benefitted the whole system. 

“Previously we relied on only one SNOTEL site and one streamflow forecast point when determining how to manage that system of 30 diversion points,” explains Denver Water spokesperson Todd Hartman. But ASO data convinced Denver Water that it could safely bypass unusual quantities into an upstream section of the Fraser River. Says Deems, “It provided more surety of operations for the diverter, while also benefiting communities by keeping water in the stream that wouldn’t otherwise be there.”

Yet to be solved is how ASO data will fit into existing forecasting models, if at all. Most current operational models can’t readily handle the spatial data produced by ASO, which also offers its own standalone runoff forecast. Plus, ASO mapping is most helpful when paired with measurements collected in the field.

“ASO leverages the SNOTEL measurements to confirm the measurements they get from the plane,” says Taylor Winchell, climate adaptation specialist for Denver Water. He embraces multiple insights into the snowpack as a net gain rather than an either-or choice. “Working in tandem [with the SNOTEL network] will paint the better picture moving forward.”

In Colorado, ASO and the CASM group got a boost in 2021 thanks to a Water Supply Reserve Fund grant via the CWCB, which facilitated a project to answer participants’ questions and identify a sustainable, long-term funding mechanism for facilitating ASO statewide. Those administrative pathways have not yet been finalized, but Winchell is already excited about what ASO offers. “Colloquially, we refer to the snowpack as our largest reservoir, but historically, we haven’t had an accurate volume estimate of that reservoir,” he explains. Lake Mead, Lake Powell, Dillon Reservoir—all these bodies of water represent known quantities. ASO stands to unlock that number for Colorado’s snow.

Future tools and magic wands

Despite the recent proliferation of data describing Colorado’s snowpack, researchers and water managers still wish for more insight. Meanwhile climate change challenges runoff forecasters with quickly shifting norms. 

“You can’t look at the snowpack’s percentage of normal and say, that’s the runoff we should expect. There’s a correlation, but the percentages don’t quite align,” explains Lukas. Warming trends make that relationship even more approximate because the recent climatic baselines used in forecasts always lag behind the real-time warming. Continues Lukas, “One challenge for forecasting in the future is making sure to incorporate up-to-date temperatures and their effects on snowpack, both with sublimation and runoff.”

NOAA has looked at swapping out historical records (from 1991, as an example) for temperature data that’s more in line with recent observations. But that hasn’t proven to be a magic bullet when accounting for climate change, says Paul Miller of NOAA’s Colorado Basin River Forecast Center. “Interestingly, the impact to our model was not as great as we expected,” he explains. “Temperature affected the runoff timing [making it earlier] but not magnitude, because we weren’t able to adjust evaporation to increase along with temperature,” he says. Making those dynamic temperature adjustments to modeled evaporation requires reconfiguring how evaporation is calculated in the model. With those adjustments now underway, Miller expects to publish findings within the coming year.

In some ways, snowpack represents the low-hanging fruit of water resources: Measuring its water content requires no insights into the future, as weather forecasting does. “It’s just sitting there,” says Deems. “If we know the snowpack really well, we can confine the uncertainties to harder-to-measure things, like future precipitation.” 

A freelance writer living in Steamboat Springs, Kelly Bastone covers water, conservation and the outdoors for publications including Outside, AFAR, 5280, Backpacker, Field & Stream, and others. She is a regular contributor to Headwaters magazine. 

Translate »