Scientific Investigations Report 2008–5093
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2008–5093
Back to Table of Contents
The Coeur d’Alene River flows in a westerly direction from its headwaters in the mountainous, coniferous forests near the Idaho-Montana border to its outlet in Coeur d’Alene Lake (fig. 1). The North Fork (NF) and South Fork (SF) Coeur d’Alene Rivers converge downstream of the town of Enaville, Idaho. From there, the main stem Coeur d’Alene River flows about 35 mi westward where it empties into Coeur d’Alene Lake. The Coeur d’Alene River terminates at the inlet to the lake; the lake’s outlet is on its north end and forms the beginning of the Spokane River. The Post Falls Dam is about 9 mi downstream of the lake (fig. 1). The dam regulates the Coeur d’Alene Lake water level from mid-June to mid-November. The Spokane River continues flowing westerly through Idaho and part of Washington, passing through several dams before joining the Columbia River at Franklin D. Roosevelt Lake (fig. 1).
More than 100 years of mining in the upper Coeur d’Alene River basin have resulted in the transport and deposition of large quantities of metal-enriched sediments into the lower Coeur d’Alene River and onto its floodplain. The SF Coeur d’Alene River flows through the Coeur d’Alene Mining District, one of the world’s largest producers of silver and one of the Nation’s largest producers of lead and zinc (Long, 1998). Mining produced more than 130 million tons of lead-zinc-silver sulfide ores from the district since the mid-1880s (Long, 1998). From the 1880s to 1968, mine waste (tailings) was dumped into the river (Ellis, 1940). The river carried the tailings downstream, especially during moderate to high flows, and deposited them in the lower reaches. Previous studies have shown that the highest concentrations of contaminated sediments are in the streambed. The sediments are entrained and redistributed during high-flow events (Bookstrom and others, 1999; Box and others, 2001; Box and others, 2005). Some of the highest concentrations of lead in the streambed were detected in the reach near Dudley, Idaho (Bookstrom and others, 2004). During times of flooding, metal-enriched sediments are deposited and (or) remobilized in the flat lower Coeur d’Alene River valley (Cataldo to Harrison) and its floodplain. The flood history of the lower valley—1933, 1949, 1956, 1964, 1974, 1995, 1966, and 1996 (Woods and Beckwith, 1997; and Box and others, 2005)—includes the most recent large peak flow, which occurred in 1996.
The Coeur d’Alene River flows naturally because no dams or major diversion structures lie upstream of Coeur d’Alene Lake. Eleven peak flows greater than 30,000 ft3/s occurred before 2005, as measured at the U.S. Geological Survey (USGS) Coeur d’Alene River near Cataldo gaging station (12413500). The largest peak flow (79,000 ft3/s) was measured on January 16, 1974. A peak flow of 70,000 ft3/s was measured during the most recent large flood on February 9, 1996 (Brennan and others, 2006; updated from Beckwith and Berenbrock, 1996). Average spring runoff for the period of record is about 15,000 ft3/s, and summer discharges usually are less than 600 ft3/s. The annual mean discharge for the period of record is about 2,500 ft3/s. Without flood control structures in the channel, periodic flooding of the lower valley occurred naturally during high flow events. This led to the construction of levees in flood prone areas to protect resources, the railroad, and nearby communities. These levees, especially in the lower valley, have limited the lateral movement of the river. River discharges must exceed 20,000 ft3/s to overtop levees and overflow riverbanks.
The Post Falls Dam creates a backwater effect in the Coeur d’Alene River that reaches upstream to Mission Flats (fig. 2). This backwater effect, in turn, decreases the river’s energy gradient and flow velocity, which affects erosion and sediment transport processes. Sediments move downstream from upper reaches of the basin to the lower energy gradient reaches of the lower Coeur d’Alene River valley. Decreased velocities in the lower valley limit the river’s ability to meander and change course, and decrease the transport of sediments. For floods at bankfull and less, sediment transport takes place only in the river. Sediments are mobilized and transported downstream in the streambed. When the river banks are overtopped, sediments in the river and the floodplain remobilize and move downstream.
Sediment transport is a physical process that depends on many hydraulic factors in the river and surrounding areas. The Coeur d’Alene River transports sediments and unconsolidated material from the streambed, banks and floodplain. Upstream of river mile (RM) 159.8 near Mission Flats, the streambed is composed mostly of gravels and cobbles. Downstream of RM 159.8, the composition is mostly sands and silts. These sediments vary in their degree of mobility.
The USGS, in cooperation with the Idaho Department of Environmental Quality, Basin Environmental Improvement Commission, and the U.S. Environmental Protection Agency developed flow hydraulic and sediment-transport models of the Coeur d’Alene River between Enaville and Pinehurst, Idaho to Coeur d’Alene Lake near Harrison, Idaho. These models will help improve the understanding of streamflow, hydraulics, bed shear stress, erosion and deposition, and sediment-transport processes. The models also will provide insight into the effects of proposed recovery actions in the river.
One model, a one-dimensional (1D) sediment-transport model, encompasses the reach beginning upstream at the streamflow-gaging stations on the NF Coeur d’Alene River at Enaville (12413000) and the SF Coeur d’Alene River near Pinehurst (12413470) and ending at the gaging station on the Coeur d’Alene River near Harrison (12413860) near the inlet to the lake. This model incorporated 234 cross sections, about 1,000 ft apart, in the reaches downstream of Mission Flats (fig. 2) and at variable intervals (200–2,000 ft) upstream to the Enaville and Pinehurst gaging stations.
The USGS also developed a multi-dimensional hydraulic and bed shear stress model for a smaller reach in the lower valley near Dudley. This model was used to simulate flow hydraulics and sediment mobility potential based on shear stress and bedform geometry in the lower Coeur d’Alene River. The study reach contains highly contaminated streambed sediments; part of the streambed in this reach may be removed by dredging because of the contaminated sediments (Bookstrom and others, 2004).
This report documents the development and calibration of a 1D sediment-transport model and a multi-dimensional hydraulic and bed shear stress model of the lower Coeur d’Alene River. The study reach (figs. 1 and 2) encompasses the lower Coeur d’Alene River from Enaville and Pinehurst to about Coeur d’Alene Lake, with a detailed focus in the Dudley reach near Rose Lake (fig. 2). Understanding the hydraulic and sediment-transport characteristics of the Dudley reach is of primary interest because of proposed dredging to remove contaminated streambed sediments (Bookstrom and others, 2004).
The scope of this report (1) defines flow types, river stages, river discharge, channel geometry, streambed sediments, and total sediment load; (2) documents the development and calibration of models of the study reach; and (3) examines the use of these models to simulate the response of the hydraulic and sediment system to river discharges, lake levels, and sediment input scenarios. The model simulation results can help facilitate the evaluation of the feasibility and potential effects of remedial actions (management alternatives) on the streambed.
The 1D sediment-transport model used in this study simulates a movable streambed, in contrast to many models in which the streambed is fixed and does not change. This 1D sediment-transport model averages parameters across the cross sections and calculates erosion, deposition, and sediment transport. One-dimensional models usually require less setup, run-time, and data than two-dimensional (2D) models. However, 2D models are able to simulate uneven water surface elevations, varying velocities, and flows in more than one direction in a cross section. For example, a 2D model that incorporated sediment transport was used to describe erosion and deposition of the Kootenai River in Idaho and the Platt River in Colorado (McDonald and others, 2006; Nelson and others, 2006).
The Coeur d’Alene River originates in the upper part of the Coeur d’Alene and St. Joe Mountains near the Idaho-Montana border and flows westward into Coeur d’Alene Lake (fig. 1). The river basin drains an area of about 1,475 mi2 and consists of two main branches of the Coeur d’Alene River, the North Fork (NF) and the South Fork (SF). These branches converge at RM 167.7 and then follow a sinuous path to the lake. The river crosses through Shoshone and Kootenai Counties, passing a few old mining and logging towns along the way. The study reach is about 35 mi long—starting upstream of Enaville on the NF and downstream of Pinehurst on the SF to Springston to about 1.5 mi upstream from the inlet to Coeur d’Alene Lake (fig. 2).
The Coeur d’Alene Mountains, at the headwaters of the river, reach elevations of 6,650 ft. The lowest point in the study area (2,128.8 ft at normal summer levels) is at the Coeur d’Alene River inlet to the lake, (NAVD 88 datum; Avista Corporation, 2005, p. B-2). The St. Joe Mountains near Cataldo reach elevations of 6,400 ft. These mountains are part of the Bitterroot Range, and consist of high forested peaks and steep intermontane valleys that cut through thick deposits of volcanic ash and metasedimentary rock. The study area is in the Coeur d’Alene metasedimentary zone (McGrath, 2002). Fractured quartzite and argillaceous rocks of Precambrian age underlie the study area (McGrath, 2002). Alluvium of Holocene and Pleistocene fills the valley and consists of unconsolidated fluvial and lacustrine deposits. The percentage of lacustrine deposits (silts and clays) increases toward the lake.
Vegetation changes with elevation. Only northern climate flora exists in the riparian zones. The valley areas consist of agricultural and open range land, wetlands, and lakes. Tree density in the valley becomes sparse with grasses and scrub vegetation dominating, especially in and around the wetlands and lateral lakes. Tree species include conifers, cottonwoods, willow, aspen, and birch. The lateral lakes feature a large variety of aquatic plants that provide habitat for avian and amphibious species. Around the turn of the 20th century, modification of natural levees formed by the river altered the hydrologic regime, and the reclaimed wetlands allowed for expansion of farming and cattle operations in the valley. At higher elevations, the study area primarily consists of coniferous forest comprised of Douglas fir, white pine, grand fir, western red cedar, and western hemlock. Mountain hemlock, subalpine fir, Engelmann spruce, and white bark pine dominate farther up the mountains.
The river is unrestricted by any dam or flood control structures upstream of Coeur d’Alene Lake and is subject to seasonal and peak flows. The annual historical mean flow for the main stem of the Coeur d’Alene River near Cataldo (RM 162.9) is about 2,500 ft3/s. Precipitation in the basin is highly variable due to orographic effects in the upper reaches—annual averages range between 26 and 40 in. of rain and 50 to 70 in. of snow. High flows in the winter can occur when heavy, warm rains from the Pacific Ocean rapidly melt the snowpack in the basin. The rapid snowmelt and heavy rains abruptly increase runoff from low winter base flows to high flows as during the winter high flows of 1995, 1996, and 1997 (Box and others, 2005). In contrast, gradual melting of snow in the basin causes a slow rising and falling runoff curve (hydrograph) for spring, represented by flows from April through June 1997 (Box and others, 2005). Warm days and spring rainstorms cause small sharp peaks on the hydrograph, represented by flows from April through June 2006.
Interstate Highway 90 lies to the north and to some extent is parallel to the river until it reaches Cataldo where it follows the SF eastward toward Wallace. The Union Pacific Railroad, recently converted to a bicycle trail in the “Rails to Trails” Program, borders the river. Elevation of the bicycle trail usually is above the floodplain and acts like a levee restricting flooding.
The floodplain starts at the confluence of the NF and SF near Enaville and extends to Coeur d’Alene Lake. The floodplain width varies with the local topography and ranges from about 1,000 ft at the confluence of the NF and SF (RM 167.7) to more than 3 mi near Dudley and Rose Lake. From Dudley, the floodplain narrows to about 1 mi near Lane, then widens to about 2 mi near Medimont, and narrows again at State Highway 97 to about 1 mi (fig. 2).
Downstream of Cataldo, wetlands and lateral lakes border the Coeur d’Alene River (fig. 2). Eleven lakes and numerous adjoining wetlands are in the floodplain. These areas provide habitat for various aquatic and terrestrial species including zooplankton, amphibians, fish, muskrats, and beavers, as well as local and migratory fowl. Based on average depth, two of the lateral lakes (Black and Blue) are deeper than 20 ft, and six lakes (Anderson, Thompson, Swan, Killarney, Rose, and Porter) are between 10 and 20 ft deep. Three lakes (Cave, Medicine, and Bull Run) are less than 10 ft deep (U.S. Fish and Wildlife Service, written commun., 2006). Most wetlands have depths of 5 ft or less. Inundation of the floodplain occurs at varying discharges due to lateral lake and wetland expansion as well as overtopping of the levee system. Lateral lakes with connecting inlet channels provide hydraulic connectivity with the river. The wetlands and lateral lakes also act as terminal or semi-terminal sinks for the contaminated sediments.
The levee system extends from the town of Cataldo downstream to the outlet at Coeur d’Alene Lake, with an average height of 7.5 ft. The original levees formed naturally from sediment deposition during floods when the river flowed out of its banks. These natural levees were highest near Cataldo and decreased in height in the downstream direction. The levee system changed as the valley population increased and land use changed. Many levees were artificially raised during the early 1900s to reclaim land for agriculture and cattle grazing. The Union Pacific Railroad Company also enhanced the levees along the southern edge of the river to support a rail line to the mining camps and mills. These levees constrained the river’s natural meandering course and stabilized the once dynamic riparian ecosystem.
Two major geomorphic reaches were identified in the study reach: a braided reach and a meandering reach (Bookstrom and others, 1999; Box and others, 2005). The braided reach includes the SF and NF and the main stem to RM 159.8 near Mission Flats (fig. 2). In the study area, the braided reach is about 11 mi long, and the river usually consists of multiple channels. Composition of the streambed is mostly gravels and cobbles. Gravel bars and terraces also are present along this reach. Considerable quantities of fines (clay, silt, and sand) also may be present in the SF and in the main stem because of poor mine-waste practices in the upper SF basin. Overall, the braided reach is a gravelbed river with median particle sizes (d50) ranging from 54 mm on the NF, 19 mm on the SF, and 23 mm on the main stem (Borden and others, 2004). The surface layer of the streambed between Latour Creek (RM 160.340) and RM 159.8 (end of the braided reach) has a pavement of overlapping cobbles. Upstream of Latour Creek, the armored surface layer has mixed amounts of sands, gravels, cobbles, and boulders. The NF produces natural quantities of sediment from a relatively pristine upper basin. However, the SF produces excessive quantities of metal-contaminated fines—lead, zinc, silver, antimony, arsenic, cadmium, and copper (Woods and Beckwith, 1997; Bookstrom and others, 1999; and Bookstrom and others, 2001; Woods, 2001). Mean annual discharge from the NF (about 1,900 ft3/s) is about 4 times greater than from the SF (about 510 ft3/s) (Brennan and others, 2006). Whereas, total sediment load from the SF is about 20 times greater than from the NF (Clark and Woods, 2001). Average flow depths in the braided reach usually are less than 10 ft, and water-surface slope is about 6.5 × 10-4 ft/ft.
The meander reach of the Coeur d’Alene River is a single channel with gentle bends, although the bends at RM 135.25 and from RM 149 to 152 are sharp. The meander reach covers the main stem from RM 159.8 at Mission Flats to the Coeur d’Alene River inlet (RM 131) to the lake, a distance of about 29 mi. This reach cuts through a nearly flat floodplain, and channel widths vary between 200 and 350 ft. This reach is relatively stable due to the river location, levees, and railroad embankments that limits the lateral migration of the river. The streambed consists primarily of silts and sands (Bookstrom and others, 1999), and d50 ranges from 0.25 to 0.10 mm. Over time, these fine, heavy metal-contaminated sediments from mine tailings settled in the meander reach. Contaminated sediments were detected to a depth of about 10 ft below the river bottom and in meander point bars (Box and others, 2001; Bookstrom and others, 2004). Flow velocities are low in this reach because of backwater conditions. Average flow depths in the meander reach usually are 20 ft, and water-surface slope is about 6.5 × 10-5 ft/ft, about 10 times flatter than the braided reach.
Calibration of the 1D model used stage or flow information from five USGS gaging stations in the study reach (fig. 2): Coeur d’Alene River near Harrison (12413860), at Rose Lake (12413810), and at Cataldo (12413500); NF Coeur d’Alene River at Enaville (12413000); and SF Coeur d’Alene River near Pinehurst (12413470). The Cataldo gaging station has been in continuous operation since 1920; Enaville since 1939, Pinehurst since 1987, and Harrison since 1991. The Rose Lake gaging station was in operation from 1994 to 2000. Only stage data are available at the Harrison and Rose Lake gaging stations because of backwater conditions at the sites. Discharge is not a function of stage alone in backwater conditions. A 1D hydraulic-flow model (Woods and Beckwith, 1997) based on water-surface elevation data for the Harrison gaging station and discharge data for the Cataldo gaging station was used to estimate discharge at the Harrison gaging station. An acoustic Doppler velocity meter (ADVM) was installed at the Harrison gaging station on February 25, 2004. A velocity-stage-discharge relation has been used since March 2004 to determine discharge. Morlock and others (2002) provide a more complete discussion on computing discharge at sites using ADVMs. Discharges for the Cataldo, Enaville, and Pinehurst gaging stations are calculated by a stage-discharge relation.
Back to Table of Contents