Abstract

Regional groundwater-flow simulations for a 30,000-square-mile area of the High Plains aquifer, referred to collectively as the Elkhorn-Loup Model, were developed to predict the effects of groundwater irrigation on stream base flow in the Elkhorn and Loup River Basins, Nebraska. Simulations described the stream-aquifer system from predevelopment through 2005 [including predevelopment (pre-1895), early development (1895–1940), and historical development (1940 through 2005) conditions] and future hypothetical development conditions (2006 through 2033 or 2055). Predicted changes to stream base flow that resulted from simulated changes to groundwater irrigation will aid development of long-term strategies for management of hydrologically connected water supplies.

The predevelopment through 2005 simulation was calibrated using an automated parameter-estimation method to optimize the fit to pre-1940 groundwater levels and base flows, 1945 through 2005 decadal groundwater-level changes, and 1940 through 2005 base flows. The calibration results of the pre-1940 period indicated that 81 percent of the simulated groundwater levels were within 30 feet of the measured water levels. The results did not indicate large areas of simulated groundwater levels that were biased too high or too low, indicating that the simulation generally captures the regional trends. Calibration results using 1945 through 2005 decadal groundwater-level changes indicated that a majority of the simulated groundwater-level changes were within 5 feet of the changes calculated from measured groundwater levels. Simulated groundwater-level rises generally were smaller than measured rises near surface-water irrigation districts. Simulated groundwater-level declines were larger than measured declines in several parts of the study area having large amounts of irrigated crops. Base-flow trends and volumes generally were reproduced by the simulation at most sites. Exceptions include downward trends of simulated base flow from the 1970s to the end of the calibration period for the Elkhorn River at Norfolk, Beaver Creek at Genoa, and Cedar River near Fullerton.

Effects of groundwater irrigation on stream base flow were predicted using several methods: (1) simulated base-flow depletion was mapped to represent the percentage of water pumped from a hypothetical well during 2006 through 2055 that corresponds to base-flow depletions at the end of that 50-year period; (2) the groundwater-flow simulation predicted changes in stream base flow that result from modifying the number of irrigated acres in a 25-year period (2009 through 2033); and (3) a simulation-optimization model determined the minimum reduction of groundwater pumpage that would be necessary in the Elkhorn River Basin in a 25-year period (2009 through 2033) to comply with various hypothetical base-flow requirements for the Elkhorn River. The results are not intended to determine specific management plans that must be adopted, but rather to improve the understanding of how base flow is affected by irrigation.

A 50-year simulation (2006–55) indicated that depletions of less than 10 percent of pumpage mainly occur in areas that are about 10 miles or farther from the Elkhorn and Loup Rivers and their tributaries.

The calibrated simulation was used to predict the 25-year effect on base flow of a 10 percent decrease in irrigated acres and the effect of increasing acres at the presently (2010) allowed rate. Hypothesized changes to irrigated acres were applied only to areas where mapped base-flow depletions were at least 10 percent of pumpage. The effect of changes in irrigated acres includes the combined effects of changes to pumpage and additional recharge from irrigated acres. When irrigated acres were decreased by 10 percent within selected areas of four Natural Resources Districts (a total reduction of about 120,000 acres and a 5 percent reduction in irrigation pumpage), simulated base flow was predicted to increase by as much as 13.0 cubic feet per second in the Loup River Basin and by as much as 23.8 cubic feet per second in the Elkhorn River Basin. The largest increases to base flow were simulated at downstream locations. When irrigated land was increased by about 25,000 acres, predicted base flow decreased by a maximum of 2.9 cubic feet per second in the Loup River Basin and by as much as 6.9 cubic feet per second in the Elkhorn River Basin. Changes to base flow were related to the proximity of the hypothetical newly irrigated acres to a stream.

When a simulation-optimization model was formulated to maximize pumpage, while maintaining base flow in two key reaches of the Elkhorn River at the estimated rate needed to support the average historical frequency (about 70 percent) at which streamflow met or exceeded an in-stream flow criterion for a downstream reach in the Platte River, maximum overall pumpage was optimized when simulated pumpage within the Lower Elkhorn Natural Resources District was reduced by approximately 40 percent and pumpage within the Upper Elkhorn Natural Resources District was reduced by about 8 percent. When the simulation-optimization model was formulated to maximize pumpage, while maintaining base flow in the two key reaches of the Elkhorn River at a more conservative estimated rate (at 65 percent of 2005 simulated base flow) needed to support the in-stream flow criterion, maximum overall pumpage was optimized when simulated pumpage within the Lower Elkhorn Natural Resources District was reduced by about 49 percent and pumpage within the Upper Elkhorn Natural Resources District was reduced by about 17 percent. Neither of these two base-flow target rates are regulatory requirements, but the results provide benchmarks for the Natural Resources Districts for determination of appropriate management goals.

Accuracy of the simulations is affected by input data limitations, system simplifications, assumptions, and resources available at the time of the simulation construction and calibration. Most of the important limitations relate either to data used as simulation inputs or data used to estimate simulation inputs. Development of the regional simulations focused on generalized hydrogeologic characteristics within the study area, and did not attempt to describe variations important to local-scale conditions. For example, a single unconfined layer was used to simulate the aquifer. Therefore, these simulations are most appropriate for analyzing groundwater-management scenarios over large areas and long time periods, and are not reliable for analysis of small areas or short time periods.

First posted September 30, 2010