USGS - science for a changing world

Scientific Investigations Report 2008–5093

Scientific Investigations Report 2008–5093

Back to Table of Contents


The Coeur d’Alene River originates in the Coeur d’Alene and St. Joe Mountains near the Idaho-Montana border and flows westward into Coeur d’Alene Lake. At the headwater of the river, elevations in the Coeur d’Alene Mountains reach 6,650 ft. The basin drains an area of about 1,475 mi2 and consists of two branches, the North Fork (NF) and the South Fork (SF). These rivers reach a confluence downstream of Enaville, and then the main stem follows a sinuous path to the lake. The study reach is about 35 mi long, starting near the towns of Enaville and Pinehurst and continuing downstream to Springston near the river inlet to the lake. More than 100 years of mining in the upper Coeur d’Alene River Basin have resulted in large quantities of metal-contaminated sediments transported to and deposited in the lower Coeur d’Alene River and its floodplain.

In the study reach, the Coeur d’Alene River is characterized by a braided reach and a meander reach. The braided reach extends upstream of river mile (RM) 159.8 near Mission Flats, and the meander reach extends downstream from RM 159.8 to the river inlet. Many exposed islands and (or) bars are in the braided reach and divide the river into multiple channels especially at low flows; the streambed primarily is composed of gravels and cobbles. The meander reach is a single channel with gentle bends except at RM 135.25 and RM 149. The streambed in the meander reach primarily is composed of sand and silt and is contaminated with heavy metals transported from mine tailings in the SF basin. Previous studies have shown the highest concentration of contaminated sediments is in the streambed in the lower Coeur d’Alene River. Some of the highest concentrations of lead in the streambed are in the Dudley reach.

A computer sediment-transport model, HEC-6, was used to simulate water-surface and streambed elevations, erosion and deposition of the streambed, and sediment transport. This steady-state, 1D model simulates the resultant erosion and deposition averaged across the streambed at any cross section in the river reach. The model was calibrated to water-surface elevations and discharges at four historical stage-discharge conditions. Model calibration was considered acceptable when the difference between measured and simulated water-surface elevations was ±0.15 ft. Actual differences were ±0.11 ft, except for one value. Also, the model was used to simulate river discharges and stages for calendar year 1999 as a check on the reasonableness of sediment transport. Measured and simulated sand transport were compared at the Harrison and Rose Lake gaging stations, and the comparisons generally agreed with one another.

The calibrated model was used to evaluate the feasibility and potential effects of remedial actions (management alternatives) on the streambed. Four management alternatives were simulated to understand effects from dredging the streambed and reducing sediment discharge input. Management alternatives 1 and 3 used river discharge and stage data from calendar year 2000, and alternative 2 and 4 used data from calendar year 1997. Calendar year 2000 was simulated because flows stayed within the confines of the channel, and only one large peak flow (representing about a 4-year flood) occurred during that year. Calendar year 1997 was simulated because flows stayed within the confines of the channel, and several large peak flows occurred. Also, twice as many consecutive days that exceeded bankfull flow occurred during calendar year 1997 than during 2000. Before the start of simulation in these alternatives, seven cross sections in the Dudley reach were deepened 20 ft to simulate dredging. About 296,000 yd3 of streambed sediments were removed by dredging.

Results from simulation 1 showed sediments being deposited in the dredged reach. About 6,530 yd3 of sediments were deposited during 2000, about 2 percent of the total volume that was dredged. Thus, about 45 years of stage-discharge conditions similar to those that occurred in calendar year 2000 would be needed to fill up the dredged reach. For simulation 2 (1997), about 12,300 yd3 of sediment was deposited in the dredged reach, about twice as much as in 2000. About 24 years of stage-discharge conditions similar to those that occurred in calendar year 1997 would be needed to fill up the dredge reach.

For alternative 3, model conditions in simulation 1 (2000) were used except that the incoming total sediment discharge from the SF was decreased by one-half. Results from this simulation showed neither differences in deposition nor in the elevation of the streambed from alternative 1 in the dredged and Dudley reaches because of the reduction of incoming total sediment discharge. The distance between the SF and Dudley reach is long enough for the sediment supply, transport capacity, and channel geometry to be balanced with river discharge before reaching the Dudley reach. Alternative 4 used the same model conditions as those used in alternative 2 (1997), except that the incoming total sediment discharge from the SF was decreased by one-half. Results from this simulation also showed no differences in deposition or streambed elevation between alternatives 2 and 4 in the dredged and Dudley reaches. It may take many years or even decades for the river to reach equilibrium conditions after incoming total sediment discharge is decreased. Effects from extreme events such as the 1996 flood (about a 100-year flood) on the channel and floodplain are unknown.

The FASTMECH computer model was used to increase understanding of two-dimensional flow hydraulics, in particular velocities and shear stresses, as they vary across the channel and in river bends. The model covers a 5.3-mile reach near Dudley, between river miles 153.9 and 159.2. The U.S. Geological Survey Multi-Dimensional Surface Water Modeling System (MD_SWMS) was used to build the computational grid for FASTMECH. The MD_SWMS also was used to edit model input, run the model, and view model simulation results. The computational grid was 27,730-ft long with 3,371 nodes in the streamwise direction and 1,148.3-ft wide with 141 nodes in the cross-stream direction, forming an approximately 2.5×2.5 m grid. The computational grid included 475,311 nodes.

Calibration of the FASTMECH model included five historical discharge conditions ranging from 10,500 to 28,900 ft3/s. Only drag coefficients were adjusted during calibration. Water-surface elevations at the downstream boundary were based on results from the HEC-6 model in fixed-bed mode. Results from FASTMECH were compared to results from HEC-6, used instead of measured data. Differences in water-surface elevation between FASTMECH and HEC-6 were less than 0.15 ft, except for one value at 0.18 ft. The mean absolute difference error and maximum difference in water-surface elevation also were less than ±0.15 ft except for one value.

Results from the calibration simulations of FASTMECH also included simulated depth and velocity. The model showed that flow depths increased as river discharges increased, except for the fourth simulation where high lake elevations cause water-surface elevations in the river to be high due to backwater conditions. The water-surface elevation at the downstream boundary in simulation 4 was about 2 ft higher than in simulation 5. However, the river discharge for simulation 5 was about 6,000 ft3/s greater than for simulation 4. The reach from RM 157.0 to RM 157.7 was the shallowest in all simulations. In the dredged reach, average simulated depth ranged from 21 to 31 ft. Simulated velocities also increased as river discharges increased. Average simulated velocities along the thalweg ranged from about 3 to 5.3 ft/s, and maximum simulated velocities ranged from 3.9 to about 7 ft/s. In the dredged reach, average simulated velocities along the thalweg ranged from 3.5 to 6 ft/s.

The FASTMECH model also showed several areas where reverse flows (back-eddies) occurred. These back-eddies usually occurred near river bends. A large back-eddy near the left bank near RM 158.3 was about 120 ft in width and about 500 ft in length and encompassed about one-third of the river width. Other back-eddies occurred throughout the modeled reach but were much smaller. No back-eddies occurred in the dredged reach.

FASTMECH also simulated bed shear stress. Simulations showed that as river discharge increased, bed shear stress increased. Average stress along the thalweg increased more than 200 percent from simulation 1 (river discharge of 10,500 ft3/s) to simulation 5 (river discharge of 28,900 ft3/s). Bed shear stress was used to assess sediment mobility potential. The potential of sediment mobility occurs when the bed shear stress exceeds the critical shear stress of the particle. This only determines whether a given particle is potentially mobile. Simulated sediment mobility indicated the transport of very coarse sand to fine gravel in these simulations. In simulation 5, simulated sediment mobility potential indicated the transport of very fine gravel to fine gravel. Also these simulations showed areas of greater sediment mobility potential near the thalweg where bed shear stresses were greater. Areas of lower sediment mobility potential were near the banks and back-eddies where bed shear stress was lower.

The models developed and calibrated for this study closely duplicate measured water-surface elevations and sediment discharges. Because the FASTMECH model was calibrated only to water-surface elevations from the HEC-6 model, a better understanding of flow and hydraulics in this reach would improve the model if water-surface elevations were measured at different flows. Greater understanding and improvement to the model would also occur if cross-stream and longitudinal velocities were measured throughout the study reach at different flows. The FASTMECH model is limited to river discharges contained within the river banks. Flood flows that overtop the banks and spread onto the floodplain would have to be addressed by a multi-dimensional floodplain model with sediment transport capability.

Back to Table of Contents

AccessibilityFOIAPrivacyPolicies and Notices

Take Pride in America logoU.S. Department of the Interior | U.S. Geological Survey
Page Contact Information: Publications Team
Page Last Modified: Thursday, 10-Jan-2013 18:49:17 EST