Simulation of ground-water flow and solute transport in the Glen Canyon aquifer, East-Central Utah

The extraction of methane from coal beds in the Ferron coal trend in central Utah started in the mid-1980s. Beginning in 1994, water from the extraction process was pressure injected into the Glen Canyon aquifer. The lateral extent of the aquifer that could be affected by injection is about 7,600 square miles. To address regional-scale effects of injection over a decadal time frame, a conceptual model of ground-water movement and transport of dissolved solids was formulated. A numerical model that incorporates aquifer concepts was then constructed and used to simulate injection.

The Glen Canyon aquifer within the study area is conceptualized in two parts—an active area of ground-water flow and solute transport that exists between recharge areas in the San Rafael Swell and Desert, Waterpocket Fold, and Henry Mountains and discharge locations along the Muddy, Dirty Devil, San Rafael, and Green Rivers. An area of little or negligible ground-water flow exists north of Price, Utah, and beneath the Wasatch Plateau. Pressurized injection of coal-bed methane production water occurs in this area where dissolved-solids concentrations can be more than 100,000 milligrams per liter. Injection has the potential to increase hydrologic interaction with the active flow area, where dissolved-solids concentrations are generally less than 3,000 milligrams per liter.

Pressurized injection of coal-bed methane production water in 1994 initiated a net addition of flow and mass of solutes into the Glen Canyon aquifer. To better understand the regional scale hydrologic interaction between the two areas of the Glen Canyon aquifer, pressurized injection was numerically simulated. Data constraints precluded development of a fully calibrated simulation; instead, an uncalibrated model was constructed that is a plausible representation of the conceptual flow and solute-transport processes. The amount of injected water over the 36-year simulation period is about 25,000 acre-feet. As a result, simulated water levels in the injection areas increased by 50 feet and dissolved-solids concentrations increased by 100 milligrams per liter or more. These increases are accrued into aquifer storage and do not extend to the rivers during the 36-year simulation period. The amount of change in simulated discharge and solute load to the rivers is less than the resolution accuracy of the numerical simulation and is interpreted as no significant change over the considered time period.