Measuring What Matters: Upper Missouri River Basin Studies Improve Drought and Flood Forecasts with Snow and Soil Moisture Data

In the Upper Missouri River Basin, which includes sections of Montana, Wyoming, Nebraska, and the Dakotas, one of the most important reservoirs is underneath your feet. Snowpack and soil moisture form enormous natural reservoirs whose water storage can be challenging to estimate. The accuracy of these estimates, however, are critical for drought and flood preparedness. 

Led by NOAA’s National Integrated Drought Information System (NIDIS), the Upper Missouri River Basin (UMRB) Data Value Study is a collaboration across the federal government, with Tribal Nations, states, and academic institutions. The UMRB Data Value Study advances understanding of soil moisture and plains snowpack to improve the Nation’s ability to predict and provide warning of flood and drought. 

Collaborating to Improve Snow and Soil Moisture-Based Forecasts 

In May 2011, heavy rains combined with melt-out from a record-high winter snowpack to trigger flooding across the Missouri River Basin, forcing the U.S. Army Corps of Engineers to conduct record-high releases from six mainstem dams. Across the region, floods caused over $2 billion in losses and damages. Six years later, severe drought struck the upper basin, causing an estimated $3.3 billion in economic costs for this agriculturally critical region. In 2019, the weather pendulum swung back to wet, causing widespread flooding again. Extreme weather is beyond human control, but regional assessment teams and post-event reports indicated that additional monitoring of variables associated with drought and flood had the potential to better support early warning and inform decision-making.

In response, Congress provided funding for the UMRB Soil Moisture and Plains Snowpack Project, an initiative to improve comprehensive water data collection. The project was split into three components: a build out of 540 new stations led by the U.S. Army Corps of Engineers in partnership with state mesonets; a data acquisition and use pilot led by the NOAA National Weather Service’s National Mesonet Program; and a Data Value Study led by NIDIS to evaluate and explore applications of the data generated by these new stations. All three components of this project involved collaboration from the National Coordinated Soil Moisture Monitoring Network (NCSMMN), state climate offices, Tribal Nations, academic institutions, and multiple federal agencies, including NASA, U.S. Geological Survey, and the U.S. Department of Agriculture’s Agricultural Research Service. 

NIDIS’s project, the Data Value Study, highlights current and potential benefits of expanded monitoring in the UMRB, including better representation of snowpack and soil moisture variability; capacity to advance satellite soil moisture products and modeled representation of soil moisture; and decision support for Weather Forecast Offices and hazards planning. The Data Value Study is now nearing completion, with findings that highlight the importance of the expanded station data for the UMRB, as well as for the broader national soil moisture community.

Exploring the Value of Expanded Water Monitoring

The Data Value Study was split into ten projects, overseen by NIDIS and carried out by partners from the U.S. Army Corps of Engineers, the Cooperative Institute for Research to Operations in Hydrology (CIROH), NOAA’s National Centers for Environmental Information (NCEI) and National Weather Service Missouri Basin River Forecast Center, NASA, JG Research, RTI International, South Dakota State University, Texas A&M, University of Montana, and the Ohio State University. Each project investigated a different component of four overarching themes:

• The contribution of soil moisture, snowpack, and other new data to hazards and water supply forecast products on local, regional, and national levels
• Enhancements made to models utilized by NIDIS, the National Water Model, the U.S. Drought Monitor, and others
• Utility of data generated by the network to improve remote sensing products, such as soil moisture products from satellites
• Lessons learned from a partnered approach to installation and operation of the network

Turning Data into Decisions

As of January 2026, the Army Corps and state mesonets have installed 366 new stations that are now transmitting data.  The nine completed Data Value Study projects provide a window into how this expansive effort is already providing benefits across sectors. The tenth, ongoing, project is a viability assessment of the larger effort, which will provide lessons learned for other networks interested in undertaking a basin-wide approach to water monitoring.

Key findings from the Data Value Study cover a range of applications. A literature review and survey of data users revealed that multiple sectors are interested in using soil moisture information to inform their work, including: water supply management, land management, municipalities, hazards planning, agricultural production, recreation, and conservation. However, soil moisture and plains snowpack have been under-monitored historically, and as new sources of these data become available, investment in communication and education on how to use them will be needed. These findings provide important guidance to the NCSMMN to support the Network in advancing quality and usability of soil moisture information. 

Even in their current format, data from the stations are already used in multiple ways. Precipitation data are being integrated by the Missouri Basin River Forecast Center (MBRFC) to improve streamflow forecasts. Additionally, new snowpack data provide a more spatially representative picture of snowpack and are already being integrated into the commonly used Snow Data Assimilation System (SNODAS) model. Meanwhile, initial tests of soil moisture data show value as a potential means to improve accuracy in future runoff forecasts both for the National Weather Service and the U.S. Army Corps of Engineers. 

Data from the new stations are also being used to enhance modeled soil moisture products and land surface models. A sensitivity analysis carried out as part of the Data Value Study showed that accuracy of drought representation, as indicated by soil dryness, improved significantly with denser networks as opposed to sparser networks. Similarly, machine learning methods applied to the UMRB stations significantly improved accuracy of modeled soil moisture in the UMRB in the Short-Term Prediction and Transition Center – Land Information System (SPoRT- LIS), a frequently used soil moisture product. Using machine learning to calibrate and validate remote-sensed soil moisture products from NASA’s Soil Moisture Active Passive (SMAP) satellite mission likewise led to improvements in product accuracy. Models that more accurately represent soil moisture variability across landscapes can help researchers, decision-makers, and others to monitor drought more effectively, and better track drought onset and improvement. 

Challenges to application of new UMRB data still remain: For example, many indices of droughts or other events rely on comparison of current conditions to averages across a longer baseline time period, often several decades. As the period of record available for these new stations grows, so too will the value of the data. In the meantime, near-term uses for datasets of this type are being tested. To address this problem, the Data Value Study supported development of new approaches to communicate soil moisture or snowpack information with short periods of record, such as calculating water available to plants by soil type or tracking change over time. These metrics also provide helpful context for interpreting soil moisture data.

Advancing Data for the Nation, Starting in the Upper Missouri

Throughout the Data Value Study, the National Coordinated Soil Moisture Monitoring Network (NCSMMN) provided guidance on project selection and approach to ensure the study's outcomes translate into broader national improvements for soil moisture data quality and usability. All data generated from the expanded UMRB station networks will be incorporated into the NCSMMN, and findings from the Data Value study—particularly the need for better communication and education on data use—offer critical guidance for the NCSMMN as it seeks to improve the usability of soil moisture information. 

“The Data Value Study team was able to develop answers to these questions quickly with the directed resources available to UMRB,” said Michael Cosh, Supervisory Research Hydrologist at the U.S. Department of Agriculture Agricultural Research Service and chair of the NCSMMN Coordinating Team. 

The techniques developed, whether for using machine learning to improve modeled products or using change over time to communicate soil moisture, will help guide soil moisture research and applications nationwide. Although the build out, data acquisition, and research of the Soil Moisture and Plains Snowpack Project focused on the Upper Missouri River Basin, the work can provide methodologies, approaches, and lessons learned that can be applied nationwide to improve our ability to predict and warn for floods and droughts earlier than previously possible, supporting risk-informed decision-making.