Groundwater development, mainly for large-scale irrigation, increased substantially in the Harney Basin since 2010. At the same time, groundwater-levels declined by more than 100 feet in some areas in the basin, and some shallow wells have gone dry. Using computer model simulations, scientists show that groundwater use has exceeded recharge since the 1980’s, causing groundwater storage to decrease by 840,000 acre-ft, that’s more than three times the volume of Oregon’s Paulina Lake. 

A new report detailing the design and application of the Harney Basin groundwater-flow model was released today by the U.S. Geological Survey and Oregon Water Resources Department. 

The Harney Basin Groundwater Model (HBGM) incorporates nearly a century of data and research in the basin and is an important tool for understanding the groundwater-flow system. Managers and stakeholders can use the HBGM to help navigate future groundwater resource needs and challenges. 

The model simulates movement of water within the groundwater system from recharge to discharge areas. Simulations of groundwater flow and discharge characterize both natural processes (streams, springs, evapotranspiration) and human use (wells). 

What can the model do? 

The HBGM can be used to assess groundwater availability based on potential future scenarios that describe different groundwater-management strategies, and changes in natural recharge patterns such as impacts due to climate change. The HBGM can also be used to understand the historic changes in the groundwater-flow system as precipitation has fluctuated and pumpage has increased over the past century. 

Model scenario results: 

The groundwater model was used to forecast how groundwater levels, discharge, and storage might change depending on groundwater usage in two hypothetical future scenarios: 

Scenario 1 assumes irrigation pumpage remains the same as in 2018 until 2100. By the end of this 82-year simulation, the areas experiencing groundwater decline have worsened and expanded due to ongoing pumpage, leading to further depletion of groundwater storage. For example, the model indicates that groundwater levels decline over 210 feet in the Weaver Spring area, while groundwater levels in the northern lowland area decline over 170 feet by 2100. Although the rate of decline slows in some areas, most still show decreasing groundwater levels at the end of the simulation. 

Scenario 2 assumes no pumpage after 2018 and average recharge rates from 1982 to 2016 to evaluate how quickly the groundwater system in the lowlands would return to levels within 5 feet of those observed in 1990, before substantial pumpage occurred. The model indicates that groundwater levels take more than 60 years to recover to within 5 feet of 1990 levels in certain areas such as the upper Silver Creek floodplain and the Weaver Spring area. Groundwater levels in most of Crane and  the northeast part of the lowlands recover in 20 to 30 years while parts of Crane are simulated to take over 40 years to recover. Groundwater levels in much of the Virginia Valley area recover within 20 years, and along the Silvies River and Donner und Blitzen River floodplains, they are simulated to return to within 5 feet of 1990 levels in less than 10 years. 

While these scenarios do not account for impacts due to climate change, such as decreased precipitation and warmer temperatures, the HBGM is capable of doing so. 

Model design and limitations:  

The HBGM is a three-dimensional numerical model covering the entire 5,240-square-mile Harney Basin and surrounding areas, utilizing data from 1930–2018. The groundwater model divides the area into a grid of square cells. Each cell represents a small part of the landscape, such as terrain, rock units, streams, springs and more. Values of aquifer properties, such as groundwater flow rates through a given material, are assigned to each cell to describe the groundwater framework. The model covers an area of about 11,269 square miles and has 10 vertical layers of varying thickness based on the geologic features and topography. 

Numerical models are approximations of complex natural systems, and their limitations must be kept in mind. Models inherently contain errors and uncertainties. The model divides the area into cells, averaging conditions within each cell over time, which can overlook changes at smaller scales. Additionally, monthly time intervals may not capture shorter-term hydrological changes accurately.  

The model provides insights into regional groundwater trends, but it may not precisely predict individual well behavior. All wells that lie in an individual cell are lumped together so results are generalized. Uncertainties exist due to various assumptions and data limitations, but ongoing data collection can help reduce these uncertainties over time. 

Only trained scientists will be able to run the model and the process requires hours to simulate results and then additional time to extract and interpret results. 

While the HBGM isn't perfect for predicting precise groundwater levels, it offers insights into the basin's hydrogeologic system and can help compare different water-management strategies. It's currently the best tool available for understanding the basin's groundwater dynamics, though refinements and updates are possible as more data becomes available and modeling techniques improve.  

For more information about the HBGM please refer to the report, published on March 22, 2024, Groundwater Model of the Harney Basin, Southeastern Oregon.