Prize Competition Launched To Improve Snowpack Water Forecast Techniques For U.S. West

The Bureau of Reclamation is launching a new prize competition for improved snowpack water forecast techniques throughout the West. Developing better techniques to determine the amount of water stored as snowpack provides water managers more accurate information to make better water management decisions.

This competition is divided into two tracks. In track one, participants develop a model and calibrate it using historical information. The effectiveness and accuracy of the test model will be evaluated during the winter and spring using real-time snowpack measurements. For track two, models in the first track are eligible to submit a report that discusses their solution and approaches to solving the problem in track one.

Reclamation is partnering with Bonneville Power Administration, NASA – Goddard Space Flight Center, U.S. Army Corps of Engineers, USDA – Natural Resources Conservation Service, U.S. Geological Survey, National Center for Atmospheric Research, DrivenData, HeroX, Ensemble and NASA Tournament Lab.

To learn more, visit

Reclamation conducts prize competitions to spur innovation by engaging a non-traditional, problem-solver community. In the past six years, it has awarded more than $4 million in prizes through 29 competitions. Please visit Reclamation’s Water Prize Competition Center to learn more.

Water resource managers use measurements and estimates of the amount of water stored in a snowpack (SWE) for streamflow and water supply forecasts, which then inform a wide range of management decisions, including managing reservoir storage levels, setting seasonal water allocations, and responding to extreme weather events such as floods and droughts.

Streamflow and water supply forecasts rely primarily on measuring SWE on the ground and air, which are limited in areas covered and times measured. High resolution satellite imagery offers promising opportunities to improve snow monitoring—using satellite imagery to estimate SWE remains an active research area. This challenge focuses on using machine learning methods that provide flexible and efficient algorithms for data-driven models and real-time prediction.

This challenge will include two tracks:

Track 1: Prediction Competition where solvers will develop and train machine learning models to estimate the current spatial distribution of SWE over the West. Models will be executed on a weekly basis to generate near real-time estimates of SWE throughout the winter and spring seasons. Models will be evaluated against ground truth data, and prizes will be awarded based on model performance. The total prize purse for the prediction competition is $440,000.

Track 2: Model Report Competition where solvers will submit additional documentation of their models. Documentation must contain additional model analysis and discussion of solution methodology, including detailed discussion of the robustness and interpretability. The total prize purse for the model report competition is $60,000.

Solvers must participate in Track 1 to be eligible for Track 2. To compete in Prediction Competition (Track 1), solvers must register and submit their ideas no later February 15, 2022 at 6:59 PM EST. To compete in the Model Report Competition(Track 2), solvers must submit their model reports no later than March 15, 2022 at 6:59 PM EST.

Visit Snowcast Showdown at

to learn more and ask your competition-specific questions. 

More news from CBB: