USGS Home
Contact USGS
Search USGS

U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2007–5044

Back to Table of Contents

Introduction

The Spokane Valley-Rathdrum Prairie (SVRP) aquifer supplies water to more than 500,000 residents in Spokane County, Washington, and Bonner and Kootenai Counties, Idaho (fig. 1). In 1978, the U.S. Environmental Protection Agency designated the aquifer as a “Sole Source Aquifer” in response to concerns about aquifer vulnerability to water-quality degradation (Federal Register, 1978). Recent and projected urban growth in the aquifer area, which includes Spokane, Spokane Valley, and Liberty Lake, Washington, and Post Falls and Coeur d’Alene, Idaho, has raised additional concerns about the effects of increased ground-water withdrawals from the aquifer. To address these concerns, a comprehensive hydrologic study was developed in 2004 by the Idaho Department of Water Resources, the Washington Department of Ecology, and the U.S. Geological Survey (USGS) to improve the understanding of ground-water flow in the aquifer and of the interaction between ground water and surface water. The purpose of the comprehensive study is to provide a scientific foundation for management of the aquifer and to provide tools that are needed to evaluate aquifer management strategies. Development of a ground-water flow model of the aquifer is one component of the comprehensive study.

The ground-water flow model presented in this report was developed by the Modeling Team formed within the comprehensive study. The Modeling Team (authors of this report) consisted of staff and personnel working under contract with the Idaho Department of Water Resources, personnel working under contract with the Washington Department of Ecology, and staff of the USGS. To arrive at a final model that has the endorsement of all team members, decisions on modeling approach, methodology, assumptions, and interpretations were reached by consensus. The Modeling Team operated under the management of the Project Technical Leadership Team and received advice and comments from the Technical Advisory Committee. In addition to undergoing the USGS report review process, the model was reviewed by the study’s Peer Review Team.

Purpose and Scope

This report describes the development of a computer model to simulate ground-water flow in the SVRP aquifer. Steps in model development include: (1) defining the areal and vertical extents of the model, (2) defining boundary conditions, (3) estimating components and rates of inflows to and outflows from the aquifer, and (4) calibrating the model by adjusting model parameters (such as hydraulic conductivity and specific yield) so the differences between simulated and measured quantities (such as water levels and flows) are minimized with respect to an objective function. Model calibration uses a nonlinear least-squares regression method, which enables quantification of parameter uncertainty within the context of the regression problem.

The primary purpose of the model is to serve as a tool for analyzing SVRP aquifer inflows and outflows, simulating the effects of future changes in ground-water withdrawals from the aquifer, and evaluating aquifer management strategies. The scale of the model and the level of detail are intended for analysis of aquifer-wide water-supply issues. The model presented in this report is not intended for application to contaminant-transport issues such as the prediction of contaminant traveltimes or flow paths. A contaminant-transport model would require a substantially greater amount of hydrogeologic detail for the contamination site.

Although the areal extent of the model differs slightly from the areal extent of the SVRP aquifer as defined by Kahle and others (2005), this report does not redefine the aquifer boundary. Because of the paucity of data for the northern extreme of Rathdrum Prairie, the model does not include Spirit and Hoodoo Valleys. Ground-water flow directions in Spirit and Hoodoo Valleys and the degree of hydraulic connections between those valleys and northern Rathdrum Prairie cannot be determined with certainty with the presently available data. Therefore, during model calibration, the uncertainty in ground-water inflow from the valleys is treated by assuming different inflow values along the model boundary.

Description of Study Area

The areal extent of the SVRP aquifer model is shown in figure 1. The model encompasses an area of approximately 326 mi² in eastern Washington and northwestern Idaho. For the most part, the model extent coincides with the 2005 revised extent of the SVRP aquifer as defined by Kahle and others (2005). However, the model excludes Spirit and Hoodoo Valleys and three areas where bedrock is close to land surface and the aquifer sediments likely are unsaturated. Within the model extent, land-surface altitude ranges from about 2,600 ft in northern Rathdrum Prairie to about 1,500 ft at the western limit of the model near Long Lake. The climate varies from subhumid to semiarid and is characterized by warm, dry summers and cool, moist winters. Mean annual (1971–2000) precipitation is 16.7 in. at the Spokane International Airport, Washington; 25.9 in. near Bayview, Idaho; and 28.1 in. at the Coeur d’Alene Airport, Idaho (Kahle and others, 2005, p. 6).

The SVRP aquifer is divided into several subregions by bedrock outcrops and subsurface bedrock ridges. On the east side of the aquifer, a crystalline rock outcrop known as Round Mountain and a less prominent bedrock ridge divide the aquifer into three channels that connect northern and southern Rathdrum Prairie. From west to east, these channels are known as West Channel, Ramsey Channel, and Chilco Channel. On the west side of the aquifer, two subsurface bedrock ridges extend from a basalt highland known as Five Mile Prairie. The first ridge extends to the south and the second extends to the west. The south-extending ridge, along with Five Mile Prairie, divides the aquifer into two arms. The eastern arm is known as Hillyard Trough, and the western arm is known as Western Arm. At the north end of Hillyard Trough, the aquifer continues west in the valley containing the Little Spokane River. This part of the aquifer is referred to as the Little Spokane River Arm. At the north end of Western Arm, the aquifer terminates against the subsurface bedrock ridge that extends west from Five Mile Prairie (see Kahle and others, 2005, p. 18-19). The narrow channel between Western Arm and the rest of the aquifer is known as Trinity Trough.

Nine lakes are located along the perimeter of the model area. Because the water levels of those lakes are higher than the ground-water level in the SVRP aquifer, water seeps from the lakebed and recharges the aquifer. The two largest lakes are Lake Pend Oreille and Coeur d’Alene Lake. Water levels in those lakes are regulated by dams on the respective outlet rivers. The outlet of Lake Pend Oreille is the Pend Oreille River, which is north of the area shown in the location map in figure 1. The outlet of Coeur d’Alene Lake is the Spokane River. The seven smaller lakes are Fernan Lake, Hauser Lake, Hayden Lake, Liberty Lake, Newman Lake, Twin Lakes, and Spirit Lake. These lakes do not have perennial outlet streams. However, during wet seasons, if the lake level rises above the outlet structure, lake water spills over the outlet structure and exits the lake as surface flow. Because of the highly permeable nature of the surficial and aquifer material, the surface flow soaks into the ground within a short distance of the lake.

The Spokane and Little Spokane Rivers are the major surface-water drainages in the model area. The Spokane River originates at Coeur d’Alene Lake. Water discharge from that lake is regulated by Post Falls Dam. Downstream of the dam, the Spokane River is free flowing for about 16 river miles until it reaches Upriver Pool near the Centennial Trail Bridge. Downstream of Upriver Dam, the river is again free flowing for about 4.5 river miles until it enters downtown Spokane. In the next 2.5 river miles, the river flows over basaltic rocks outside the SVRP aquifer and is regulated by dams in the Spokane Falls area. The river re-enters the aquifer at the south end of Western Arm and is again free flowing for about 10 river miles as it meanders northwest to Nine Mile Reservoir. Downstream of Nine Mile Dam, the Spokane River is joined by the Little Spokane River, which enters the model area at the north end of Hillyard Trough and flows toward the west. Downstream of the confluence of the two rivers, the Spokane River becomes a reservoir known as Long Lake, which is regulated by Long Lake Dam.

Previous Investigations

In preparation for developing the ground-water flow model presented in this report, Kahle and others (2005) compiled geologic and hydrologic information available as of June 2005 for the SVRP aquifer. The geologic information includes the pre-Tertiary, Tertiary, and Quaternary geology of the aquifer and results of previous geophysical investigations. The hydrologic information includes the current understanding of the hydrogeologic framework, ground-water movement, water-budget components, and ground-water/surface-water interactions. The compilation also includes descriptions of previous ground-water flow models by Bolke and Vaccaro (1981), CH2M Hill (1998, 2000), Buchanan (2000), and Golder Associates, Inc. (2004). In another report, Kahle and Bartolino (2007) refined the hydrogeologic understanding of the aquifer. Using drillers’ records and results of available geophysical investigations, they developed a contour map showing the altitude of the base of the aquifer, mapped the extent of a clay layer in Hillyard Trough and in the Little Spokane River Arm, and updated the aquifer water budget using recently compiled information.

Because previous ground-water flow models of the SVRP aquifer are described in detail in the report by Kahle and others (2005), the following discussion highlights only the similarities and differences among the models. The models by Bolke and Vaccaro (1981), CH2M Hill (1998, 2000), and Golder Associates, Inc. (2004) encompass the western (mostly Washington) part of the aquifer from Post Falls or the Idaho-Washington State line to the eastern end of Long Lake. The model by Buchanan (2000) encompasses the entire aquifer in both States. All four models simulate the interaction between the aquifer and the Spokane and Little Spokane Rivers. In addition, the model by Golder Associates, Inc. (2004) simulates overland flow, river flow, and subsurface flow in the unsaturated and saturated zones of the aquifer.

Although evidence available since the late 1990s (see, for example, Gruenenfelder, 1997) indicates that an extensive clay layer divides the SVRP aquifer into an upper, unconfined unit and a lower, confined unit in Hillyard Trough and the Little Spokane River Arm, all four previous models treat the aquifer effectively as a single, unconfined, hydrogeologic unit. The clay layer is not mentioned in the reports by Bolke and Vaccaro (1981) and Buchanan (2000). In the reports by CH2M Hill (1998, 2000), the clay layer is discussed but the model excludes the lower, confined unit. In the report by Golder Associates, Inc. (2004), the clay layer is treated as a geologic lens within a model layer rather than as a separate model layer. Therefore, a single model layer represents both the upper and lower units.

The four models are calibrated using different amounts of measured data for different time periods. Bolke and Vaccaro (1981) presented both a time-averaged simulation and a transient simulation for May 1977–April 1978. The model by Bolke and Vaccaro (1981) is calibrated using measured water levels in 73 wells and ground-water discharge from the SVRP aquifer to the Little Spokane River (that is, streamflow gain on the river). The calibration is checked by a water-balance calculation and by comparing simulated streamflow with measured streamflow for three sites on the Spokane River. The model by CH2M Hill (1998, 2000) is a steady-state model and is calibrated using measured water levels in about 110 wells and streamflow gains and losses on the Spokane River during September 1994. The calibration is checked by comparing simulated and measured water levels and streamflow gains and losses for April 1995. The model by Buchanan (2000) also is a steady-state model and is calibrated using measured water levels in 15 wells during unspecified time periods. The calibration data do not include measured streamflow gains or losses. The model by Golder Associates, Inc. (2004) is a transient model and is calibrated using measured water levels in about 20 wells and measured streamflow on the Spokane and Little Spokane Rivers during 1994, 1997, and 1999.

An important assumption in the models by Bolke and Vaccaro (1981), CH2M Hill (1998, 2000), and Golder Associates, Inc. (2004) is the boundary condition at the eastern terminus of the model. This boundary condition controls ground-water flow from Rathdrum Prairie to the model. The model by Bolke and Vaccaro (1981) specifies the hydraulic head along the eastern boundary. The calibrated model simulates a ground-water inflow of about 400 ft³/s across the boundary. By contrast, the model by CH2M Hill (1998, 2000) specifies the ground-water inflow across the eastern boundary and inflow rate is adjusted during calibration. The calibrated model simulates a ground-water inflow of 385 ft³/s. The model by Golder Associates, Inc. (2004) specifies a time-varying hydraulic head along the eastern boundary. The report by Golder Associates, Inc. (2004, fig. 9.8) does not give the simulated flow across the boundary but does indicate that the simulated flow across the Washington-Idaho State line, about 3.5 mi west of the eastern boundary, ranges from about 30 to 850 ft³/s.

Components of aquifer inflows and outflows are treated in somewhat different manners in the four models. In all four models, aquifer inflows include recharge from precipitation (that is, precipitation minus evapotranspiration), leakage from the losing segments of the Spokane River, and varying amounts of inflows from tributary basins, adjacent uplands, or lakes along the aquifer perimeter. The models by Bolke and Vaccaro (1981), CH2M Hill (1998, 2000), and Golder Associates, Inc. (2004) also include return percolation from irrigation, effluent from septic systems, and inflow across the eastern boundary as discussed in the previous paragraph. In all four models, ground water discharges to the Little Spokane River and to gaining segments of the Spokane River. The models by Bolke and Vaccaro (1981), CH2M Hill (1998, 2000), and Golder Associates, Inc. (2004) also include ground-water withdrawals by water purveyors and by industrial, commercial, agricultural, and domestic users. In the model by Bolke and Vaccaro (1981), ground-water outflow also occurs along a specified-hydraulic head boundary near the confluence of the Spokane and Little Spokane Rivers.

Back to Table of Contents