Skiles led a study published in Environmental Research Letters that highlights a particularly critical variable needed for snow forecasting models to adapt to a fast changing world: dust.
Cycles of Dusty Snow
Dust, being darker than the underlying snow, absorbs more energy from the Sun and speeds up snowmelt. Fast-melting snow is a problem because mountain snowpack shelters soil from the heat of the Sun, Skiles explained. When snow melts quickly, soil loses that protective blanket and dries out earlier in the season.
Researchers tracked the phenomenon in Utah in 2021 and 2022, when water levels in the Great Salt Lake hit a record low, driven by increased consumption and prolonged drought. Dust from the exposed lake bed blew onto snow in the adjacent Wasatch Mountains.
Great Salt Lake dust accelerated Wasatch snowmelt by 17 days during the 2022 snowmelt season, according to data Skiles and her colleagues published in June 2023.
“The landscape is drier, so any additional moisture that comes after is basically soaked up by the landscape instead of making its way back down to the Great Salt
Lake,” Skiles explained.
The phenomenon is a feedback loop: With less water flowing into the lake, the dry lake bed expands, and more dust is blown onto the Wasatch snowpack. The cycle then repeats.
These findings back up numerous studies conducted from 2010 to 2018 in the Colorado Rockies, Skiles said. In the San Juan Mountains, dusty gusts from the Colorado Plateau accelerated snowmelt by 3–5 weeks and were correlated with snowmelt forecasting errors.
Timely flood warnings and effective reservoir management require accurate snowmelt predictions. NOAA deploys river forecasting models to predict the snowmelt rate and the amount of water that will replenish rivers annually.
But despite dust’s significant impact on the snowmelt rate, many river forecasting models, including NOAA’s, do not account for it.
At the Colorado Basin River Forecast Center (CBRFC), hydrologists are updating models to change that.
One strategy is to turn up the models’ temperature input because adding slightly more heat simulates the impact of dust, said John Lhotak, a hydrologist at CBRFC who was unaffiliated with the study in Utah. Such calibrations are based on data from historic dust events.
“Things keep shifting,” Lhotak said, explaining the need to update the old model. Climate change has altered patterns in precipitation, temperature, and dust. “So when you calibrate this model, you’re calibrating on a historical record that is changing,” he explained.
Lhotak’s office is also testing a more dynamic physical model that relies on additional inputs, such as solar radiation, wind, humidity, and dust on snow, to better simulate the real-time variability in observed snow conditions.
In the Colorado River Basin, “every drop is precious,” Lhotak said, so getting the forecast right, down to the smallest fraction of a percent, is critical for communities to plan and adapt. “That’s where we’re at right now—that’s why every drop is getting scrutinized.”
This piece was produced with support from the National Association of Science Writers’ David Perlman Virtual Mentoring Program.