Reclamation launches $500k snowpack estimation competition

The Bureau of Reclamation has used crowdsourcing to launch the "Snowcast Showdown" competition, offering a $500k prize purse for snowpack estimation across the western U.S.

Dec. 7, 2021

3 min read

Today, the Bureau of Reclamation announced a new crowdsourcing competition, the "Snowcast Showdown," which is being implemented by DrivenData, Ensemble, and HeroX.

The competition will look for innovative methods of estimating a large-scale snow water equivalent (SWE), tasking competitors to provide spatially-distributed estimates of SWE over the western U.S. during the 2022 winter season. Participants will compete for a prize purse of $500,000.

Seasonal mountain snowpack is a critical water resource throughout the Western US. During winter, snow accumulates in mountain headwaters. When temperatures rise in the spring and summer, snowpacks melt, serving as a freshwater source for major streams and rivers which feed Reclamation reservoirs. Therefore, knowing how much water is stored in winter snowmelt is essential to forecasting spring and summer water supplies.

The Bureau of Reclamation is the largest wholesaler of water in the western U.S. and the second largest producer of hydropower in the U.S. Reclamation delivers water to more than 31 million people per year and provides irrigation water to one out of five Western farmers.

“Our technology must advance to meet the particular challenges of climate change and extreme weather,” said Greg Lipstein, Principal at DrivenData. “These kinds of Challenges channel the efforts of skilled experts around the world to push forward the tools available to our nation's public agencies and water managers.”

While there are multiple technologies to monitor snowpack, none adequately address all of the needs of Reclamation. Over the past several decades, ground-based instruments including SNOwpack TELemetry(SNOTEL) stations have been used to monitor snowpacks.

Ground measures can provide accurate SWE estimates but they are spatially limited, especially at high elevation, and are challenging to maintain. Airborne methods to estimate SWE are highly accurate, but are expensive and currently have limited spatial and temporal coverage. Satellite remote sensing methods are also emerging, but are not yet operational. Thus, new and improved methods are needed to accurately measure SWE and inform critical water management decisions.

The Prize

Prizes will be split across two challenge tracks. The Prediction Competition Track is a machine learning competition, where participants train models to estimate SWE at 1km resolution across 17 states in the Western US. $440,000 will be awarded in the Prediction Competition Track, with separate prizes for overall performance and performance across three specific geographies.

The second challenge track, the Model Report Competition, is a model analysis competition which will award a $60,000 prize purse to the top three teams.

The competition’s hosts have stated that the the competition is open to anyone aged 18 or older and can participate as either an individual or as a team. Participants may originate from any country, as long as United States federal sanctions do not prohibit participation. Additional eligibility requirements or limits can be found in the challenge rules. More information on the competition can be found at DrivenData's page for the event.