Demonstration optimization analyses of pumping from selected Arapahoe aquifer municipal wells in the west-central Denver Basin, Colorado, 2010–2109

Abstract

Declining water levels caused by withdrawals of water from wells in the west-central part of the Denver Basin bedrock-aquifer system have raised concerns with respect to the ability of the aquifer system to sustain production. The Arapahoe aquifer in particular is heavily used in this area. Two optimization analyses were conducted to demonstrate approaches that could be used to evaluate possible future pumping scenarios intended to prolong the productivity of the aquifer and to delay excessive loss of saturated thickness. These analyses were designed as demonstrations only, and were not intended as a comprehensive optimization study. Optimization analyses were based on a groundwater-flow model of the Denver Basin developed as part of a recently published U.S. Geological Survey groundwater-availability study. For each analysis an optimization problem was set up to maximize total withdrawal rate, subject to withdrawal-rate and hydraulic-head constraints, for 119 selected municipal water-supply wells located in 96 model cells. The optimization analyses were based on 50and 100-year simulations of groundwater withdrawals. The optimized total withdrawal rate for all selected wells for a 50-year simulation time was about 58.8 cubic feet per second. For an analysis in which the simulation time and headconstraint time were extended to 100 years, the optimized total withdrawal rate for all selected wells was about 53.0 cubic feet per second, demonstrating that a reduction in withdrawal rate of about 10 percent may extend the time before the hydraulichead constraints are violated by 50 years, provided that pumping rates are optimally distributed. Analysis of simulation results showed that initially, the pumping produces water primarily by release of water from storage in the Arapahoe aquifer. However, because confining layers between the Denver and Arapahoe aquifers are thin, in less than 5 years, most of the water removed by managedflows pumping likely would be supplied by depleting overlying hydrogeologic units, substantially increasing the rate of decline of hydraulic heads in parts of the overlying Denver aquifer. Introduction The Denver Basin bedrock aquifers are a primary source of water for much of the population of the southern part of the Denver metropolitan area and adjacent rural areas. Because of the slow rate of natural recharge, groundwater in the Denver Basin commonly is regarded for practical purposes as nonrenewable. Municipal-well pumping in the west-central part of the basin has contributed to water-level declines and, in some areas, declining well yields. Concerns related to the rates of water-level decline and the longevity of the usefulness of the aquifers (Topper and Raynolds, 2007) have prompted investigations (Hydrosphere Resource Consultants, Inc., and others, 1999; Black and Veatch and others, 2003) with the goal of prolonging productivity of the aquifers for beneficial use. A recent report on groundwater availability (Paschke, 2011) presents a three-dimensional groundwater flow model of the Denver Basin and describes the decline of water levels in wells in these areas as the rate of withdrawal of water from wells in the area has increased. The U.S. Geological Survey (USGS), in cooperation with the Colorado Water Conservation Board (CWCB), developed a series of groundwater-flow simulations to demonstrate how optimization techniques could be used to design pumping schemes that may prolong the productivity of the Arapahoe aquifer. These analyses were designed as demonstrations only and were not intended as a comprehensive optimization study. The optimization analyses were based on the groundwaterflow model of the Denver Basin bedrock aquifers and overlying alluvial aquifer described by Banta and others (2011). The analyses demonstrate the applicability of optimization technology to the problem of maximizing production of water while maintaining saturated thickness of the Arapahoe aquifer of at least 90 percent of the aquifer thickness for 50 or 100 years. 2 Demonstration Optimization Analyses of Pumping from Selected Arapahoe Aquifer Municipal Wells Purpose and Scope This report describes results of a groundwater-flow modeling investigation designed to demonstrate the application of optimization techniques to understand the effects of various possible future pumping schemes and to delay excessive loss of saturated thickness in the Arapahoe aquifer. The optimization goals were developed in cooperation with CWCB and the Colorado Division of Water Resources (CDWR). The investigation focused on pumping from the Arapahoe aquifer in the metropolitan area south of Denver in the west-central part of the Denver Basin. The other bedrock aquifers of the Denver Basin and the overlying alluvial aquifer were included in the modeling simulations because of the hydrologic connections among the aquifers. These analyses were designed as demonstrations only and were not intended as a comprehensive optimization study. Study Area The project study area is the area underlain by the Denver Basin bedrock aquifer system, which occupies about 6,700 mi2 in eastern Colorado (fig. 1). The metropolitan area south of Denver (hereinafter, the south Denver metropolitan area) is the main area of interest for this study because it contains a large number (119) of municipal wells completed in the upper and lower Arapahoe aquifers (fig. 2). For the purposes of this study, the south Denver metropolitan area is considered to include northern Douglas County and southwestern Arapahoe County. The groundwater model encompasses the entire study area; however, the optimization analyses only involve wells operated by municipal water providers in the south Denver metropolitan area. Two wells included in the optimization analyses are east of Denver and somewhat isolated from the other selected wells, but they belong to a water provider that operates wells within the area of greatest interest. Denver Basin Bedrock Aquifer System As residential development has expanded beyond Denver, particularly to the south of Denver, newly developed areas have become increasingly dependent on water from Denver Basin bedrock aquifers. Estimated production from all wells completed in the Denver Basin bedrock aquifers increased from about 15 ft3/s in 1958 to about 41 ft3/s in 1978 (Robson, 1987) and about 118 ft3/s in 2003 (Paschke and others, 2011a). In 2003, municipal water providers in the study area withdrew an estimated 48 ft3/s (35,000 acre-ft/yr) from wells completed in the upper and lower Arapahoe aquifers, the aquifers of interest in this investigation (Paschke and others, 2011a). A detailed description of the hydrogeologic framework is provided by Paschke and others (2011b) and is summarized in the following paragraph. The bedrock aquifers of the Denver Basin are in wateryielding sandstones of Cretaceous and Tertiary age in a 6,700mi2 area in eastern Colorado located between Greeley on the north, Colorado Springs on the south, Limon on the east, and the Rocky Mountain Front Range on the west (fig. 1). Rules relating to withdrawal of groundwater of the Denver Basin (Colorado Division of Water Resources, 1985) are based on a hydrogeologic framework in which four primary bedrock aquifers (the Dawson, Denver, Arapahoe, and Laramie–Fox Hills aquifers) are defined; the Dawson and Arapahoe aquifers locally are further differentiated into upper and lower aquifers (fig. 2, table 1). The aquifers generally are separated by confining units. The lower Dawson, Denver, and Arapahoe aquifers represent the synorogenic deposition of sediments in the Denver Basin during Laramide uplift of the Rocky Mountain Front Range (Raynolds, 2002). The prominent Wildcat Mountain alluvial fan mapped in the lower Arapahoe aquifer is also present in the Denver and lower Dawson portions of the sequence, and the Denver confining units are thin (mean thicknesses of 50 ft) compared to the thickness of the Denver aquifer (mean thickness of 740 ft). A confining unit separates the upper Arapahoe aquifer from the lower Arapahoe aquifer in approximately the northern one-third of the basin (fig. 3). In the southern two-thirds, the confining unit is absent and, for the purposes of this study, the Arapahoe aquifer is considered undifferentiated. Alluvial sediments of the South Platte River and numerous tributaries of the South Platte and Arkansas Rivers, some of which are ephemeral or intermittent, overlie substantial areas of the bedrock basin and, where saturated, form an unconfined alluvial-aquifer system. This investigation uses the same hydrogeologic framework as the groundwaterflow model described by Paschke and others (2011b). Numerical Model For the optimization analyses, the MODFLOW-2000 (Harbaugh and others, 2000) model of the Denver Basin aquifer system (Banta and others, 2011) was converted to a MODFLOW-2005 (Harbaugh, 2005) format. The finitedifference model grid (fig. 4) encompasses the Denver Basin bedrock-aquifer system and immediately adjacent areas of the alluvial aquifer of the South Platte River (fig. 2). The grid has 84 columns and 124 rows of square cells; each cell represents a 1-mile by 1-mile area. In the vertical dimension, the aquifers and confining units are represented by 12 model layers, as indicated in table 1. Details of model conceptualization and parameterization are fully described in Banta and others (2011). This section only describes aspects of the model that are particularly relevant to the optimization analyses or that differ from the model described in that report. The calibration period for the Denver Basin groundwater model extended from 1880 through 2003 (Banta and others, 2011). In that model, one MODFLOW-2000 (Harbaugh and others, 2000) steady-state stress period was used to simulate