Preliminary Simulation of Chloride Transport in the Equus Beds Aquifer and Simulated Effects of Well Pumping and Artificial Recharge on Groundwater Flow and Chloride Transport near the City of Wichita, Kansas, 1990 through 2008

Abstract

The Equus Beds aquifer in south-central Kansas is a primary water-supply source for the city of Wichita. Water-level declines because of groundwater pumping for municipal and irrigation needs as well as sporadic drought conditions have caused concern about the adequacy of the Equus Beds aquifer as a future water supply for Wichita. In March 2006, the city of Wichita began construction of the Equus Beds Aquifer Storage and Recovery project, a plan to artificially recharge the aquifer with excess water from the Little Arkansas River. Artificial recharge will raise groundwater levels, increase storage volume in the aquifer, and deter or slow down a plume of chloride brine approaching the Wichita well field from the Burrton, Kansas area caused by oil production activities in the 1930s. Another source of high chloride water to the aquifer is the Arkansas River. This study was prepared in cooperation with the city of Wichita as part of the Equus Beds Aquifer Storage and Recovery project.

Chloride transport in the Equus Beds aquifer was simulated between the Arkansas and Little Arkansas Rivers near the Wichita well field. Chloride transport was simulated for the Equus Beds aquifer using SEAWAT, a computer program that combines the groundwater-flow model MODFLOW-2000 and the solute-transport model MT3DMS. The chloride-transport model was used to simulate the period from 1990 through 2008 and the effects of five well pumping scenarios and one artificial recharge scenario. The chloride distribution in the aquifer for the beginning of 1990 was interpolated from groundwater samples from around that time, and the chloride concentrations in rivers for the study period were interpolated from surface water samples.

Five well-pumping scenarios and one artificial-recharge scenario were assessed for their effects on simulated chloride transport and water levels in and around the Wichita well field. The scenarios were: (1) existing 1990 through 2008 pumping conditions, to serve as a baseline scenario for comparison with the hypothetical scenarios; (2) no pumping in the model area, to demonstrate the chloride movement without the influence of well pumping; (3) double municipal pumping from the Wichita well field with existing irrigation pumping; (4) existing municipal pumping with no irrigation pumping in the model area; (5) double municipal pumping in the Wichita well field and no irrigation pumping in the model area; and (6) increasing artificial recharge to the Phase 1 Artificial Storage and Recovery project sites by 2,300 acre-feet per year.

The effects of the hypothetical pumping and artificial recharge scenarios on simulated chloride transport were measured by comparing the rate of movement of the 250-milligrams-per-liter-chloride front for each hypothetical scenario with the baseline scenario at the Arkansas River area near the southern part of the Wichita well field and the Burrton plume area. The scenarios that increased the rate of movement the most compared to the baseline scenario of existing pumping between the Arkansas River and the southern boundary of the well field were those that doubled the city of Wichita’s pumping from the well field (scenarios 3 and 5), increasing the rate of movement by 50 to 150 feet per year, with the highest rate increases in the shallow layer and the lowest rate increases in the deepest layer. The no pumping and no irrigation pumping scenarios (2 and 4) slowed the rate of movement in this area by 150 to 210 feet per year and 40 to 70 feet per year, respectively. In the double Wichita pumping scenario (3), the rate of movement in the shallow layer of the Burrton area decreased by about 50 feet per year. Simulated chloride rate of movement in the deeper layers of the Burrton area was decreased in the no pumping and no irrigation scenarios (2 and 4) by 80 to 120 feet per year and 50 feet per year, respectively, and increased in the scenarios that double Wichita’s pumping (3 and 5) from the well field by zero to 130 feet per year, with the largest increases in the deepest layer. In the increased Phase 1 artificial recharge scenario (6), the rate of chloride movement in the Burrton area increased in the shallow layer by about 30 feet per year, and decreased in the middle and deepest layer by about 10 and 60 feet per year, respectively. Comparisons of the rate of movement of the simulated 250-milligrams-per-liter-chloride front in the hypothetical scenarios to the baseline scenario indicated that, in general, increases to pumping in the well field area increased the rate of simulated chloride movement toward the well field area by as much as 150 feet per year. Reductions in pumping slowed the advance of chloride toward the well field by as much as 210 feet per year, although reductions did not stop the movement of chloride toward the well field, including when pumping rates were eliminated. If pumping is completely discontinued, the rate of chloride movement is about 500 to 600 feet per year in the area between the Arkansas River and the southern part of the Wichita well field, and 70 to 500 feet per year in the area near Burrton with the highest rate of movement in the shallow aquifer layer.

The averages of simulated water-levels in index monitoring wells in the Wichita well field at the end of 2008 were calculated for each scenario. Compared to the baseline scenario, the average simulated water level was 5.05 feet higher for the no pumping scenario, 4.72 feet lower for the double Wichita pumping with existing irrigation scenario, 2.49 feet higher for the no irrigation pumping with existing Wichita pumping scenario, 1.53 feet lower for the double Wichita pumping with no irrigation scenario, and 0.48 feet higher for the increased Phase 1 artificial recharge scenario.

The groundwater flow was simulated with a preexisting groundwater-flow model, which was not altered to calibrate the solute-transport model to observed chloride-concentration data. Therefore, some areas in the model had poor fit between simulated chloride concentrations and observed chloride concentrations, including the area between Arkansas River and the southern part of the Wichita well field, and the Hollow-Nikkel area about 6 miles north of Burrton. Compared to the interpreted location of the 250-milligrams per liter-chloride front based on data collected in 2011, in the Arkansas River area the simulated 250-milligrams per liter-chloride front moved from the river toward the well field about twice the rate of the actual 250-milligrams per liter-chloride front in the shallow layer and about four times the rate of the actual 250-milligrams per liter-chloride front in the deep layer. Future groundwater-flow and chloride-transport modeling efforts may achieve better agreement between observed and simulated chloride concentrations in these areas by taking the chloride-transport model fit into account when adjusting parameters such as hydraulic conductivity, riverbed conductance, and effective porosity during calibration.

Results of the hypothetical scenarios simulated indicate that the Burrton chloride plume will continue moving toward the well field regardless of pumping in the area and that one alternative may be to increase pumping from within the plume area to reverse the groundwater-flow gradients and remove the plume. Additionally, the results of modeling these scenarios indicate that eastward movement of the Burrton plume could be slowed by the additional artificial recharge at the Phase 1 sites and that decreasing pumping along the Arkansas River or increasing water levels could retard the movement of chloride and may prevent further encroachment into the southern part of the well field area.

First posted August 7, 2014