Climate change is making it harder to estimate Sierra snowpack. AI may be able to help

By Heather Waldman

Click here for updates on this story

    SIERRA NEVADA, California (KCRA) — Changes in California’s climate patterns are making it more difficult for scientists to accurately estimate the state’s snowpack. This could present a problem for water managers, who rely on accurate numbers to plan for spring flood control and annual water supply allocations.

Historically, a large network of automatic sensors called “snow pillows” has been used to take automated measurements of the weight of the snow at a specific location. Knowing the weight allows scientists to calculate the amount of water in the snow.

Decades of that measured data combined with an understanding of historical snow trends have allowed scientists to make informed estimates about the amount of snow in certain areas. As California’s climate changes, the relationships informing those estimates may become less useful.

“When we use all of this knowledge that we’ve gained from our really long record of the past and try to apply it in a new situation that doesn’t look quite like the past, we don’t trust that application as well,” said Marianne Cowherd.

Cowherd is currently a PhD candidate at the University of California, Berkeley, studying environmental science policy and management. As part of her program, Cowherd is working with UC Berkeley’s Central Sierra Snow Lab to help find ways to factor climate change into snowpack estimates.

“We were really interested in situations where the snow pillow itself would be experiencing a decrease [in snowpack],” Cowherd said.

She added that areas at elevations above those snow pillows may also be experiencing a decrease in snowpack, but in some cases, there could be an undetected increase.

“And if we’re just using that snow pillow record and our expectations based on what’s happened in the past, we might miss something that’s happening in a region that we’re not measuring,” said Cowherd.

Artificial intelligence could be part of the solution. Cowherd and her colleagues have been exploring different AI algorithms that can learn from past snow pillow data and incorporate climate change models to produce a more accurate picture of future snowpack scenarios.

“AI is definitely a potential helper for solving this problem of wanting to have better understanding of where our snow is based on a sparse number of observations,” said Cowherd, adding that AI should be seen as a suite of possible tools. The trick is finding which tool is best for this situation.

That will be part of future research along with getting a better understanding of how different water users rely on snow pillow data.

California currently utilizes several methods to get a picture of the snowpack each winter. They include manual snow surveys which are conducted each month. Images from high-resolution satellites and LIDAR measurements are also being used in addition to the automated snow pillow network.

Please note: This content carries a strict local market embargo. If you share the same market as the contributor of this article, you may not use it on any platform.