Question and Answer - Setting Instream Flows in Dungeness River Basin

This page is being maintained as an archive of past information and will not be updated.

In an effort to provide more detail about the use of scientific studies related to instream flow, Ecology presents the following questions and answers written by Brad Caldwell. (Brad no longer works for Ecology)

Q:  Are the instream flows proposed for the Dungeness River supported by scientific, peer-reviewed instream flow studies?

A:  Yes. There have been several scientific studies over the last 20 years to answer questions on how salmon and steelhead (salmonid) habitat is related to flow in the lower 10 miles of the Dungeness River.  In the following narrative I will list and describe 4 of the many reports that have been useful for determining the instream flows proposed for the Dungeness River.

In 1987 the Dungeness River Management Team (DRMT) was formed to bring all water users to the table to discuss how to tackle the problem of the degraded Dungeness River spring Chinook and pink salmon runs.  Water withdrawals were identified as the primary reason preventing upstream migration of adult pink and spring Chinook salmon along with reduced spawning and rearing habitat for salmonids.

In 1988 the U.S. Fish and Wildlife Service was requested to do an instream flow study to determine the flow required in the lower Dungeness River to restore the degraded pink and spring Chinook salmon runs.  The study was done under guidance from an interagency technical team. This Dungeness Instream Flow Group helped design, implement, and peer review the study using the Instream Flow Incremental Methodology (IFIM).  This group consisted of Hal Beecher PH.D. and Tim Rymer with Wa. Dept. of Wildlife, Brian Winter Ph.D. with National Marine Fisheries Service, Brad Caldwell M.S. with Wa. Dept. of Ecology, Joe Hiss and Phil Wampler with U.S. Fish and Wildlife Service, Randy Johnson with Wa. Dept. of Fisheries, and Mike Reed and Ann Seiter and Brad Sele with Jamestown S’Klallam Tribe.

This interagency technical work group had an unusually large amount of expertise with the PHABSIM/IFIM hydraulic model. Dr. Hal Beecher and Dr. Brian Winter and Brad Caldwell and Phil Wampler all had several years of experience using this particular computer model on rivers in Washington State to determine instream flows needed by salmonids.

Field data was gathered in 1988 and 1989 by measuring the depths and velocities along 8 transects at a representative site at river mile (RM) 4.2 and another 8 transects at RM 2.3.  These sites represent the braided reach from RM 3.3 to 6.4 and the diked single channel from RM 1.8 to 2.5.  The representative sites were selected by walking the lower 6 miles of the river and then rafting from RM 6 to 4 to quantify the amount of each habitat type (pool, riffle, etc) to determine transect weighting in the hydraulic model.  This study went beyond most IFIM studies by gathering site specific fish preference data.  Dr. Beecher and others snorkeled the river to collect observations on the depths and velocities used by juvenile steelhead, coho, Chinook, and dolly varden (bull trout).  Dr. Beecher reviewed the creation of the fish preference data and final curves used in the hydraulic model. Calibration of the 10 hydraulic models for the upper site and 4 hydraulic models for the lower site was reviewed by Dr. Brian Winter and Brad Caldwell.  This final report is available here:

The model is complex and requires extensive training to understand the numbers input into the computer model and the how the models calculates its regressions.  For instance, the raw data input of depths, velocities, substrates, cover, water surface elevations, streamflows, roughness coefficients, etc. is listed on pages 79 to 130 with around 50 lines per page with around 10 numbers per line.  So there are around 25,000 numbers that are input into the model so it can accurately predict the depths and velocities over useable substrate over a wide range of streamflows.  However, those resulting graphs of how much fish habitat one gains or loses for each increment of streamflow have to be interpreted by knowledgeable fish biologist to determine what streamflow is best for fish.  There are many species (steelhead, Chinook, coho, pink, chum, dolly varden) and multiple lifestages (spawning, rearing).  Each species and lifestage has a different streamflow where habitat peaks.  One has to prioritize month by month and by species and by lifestage and by PHABSIM/IFIM site.  Clearly, there is no one streamflow that is optimum for all the fish species and lifestages that exist simultaneously in the stream.

The process of determining how to use the fish habitat relationships in the IFIM report is documented in the following report:

The Dungeness River Instream Flow Group consisted of Hal Beecher PH.D. and Tim Rymer with Wa. Dept. of Wildlife, Brian Winter Ph.D. with National Marine Fisheries Service, Brad Caldwell M.S. with Wa. Dept. of Ecology, Joe Hiss and Phil Wampler with U.S. Fish and Wildlife Service, Randy Johnson with Wa. Dept. of Fisheries, and Mike Reed and Ann Seiter and Brad Sele with Jamestown S’Klallam Tribe.

These technical experts reviewed the IFIM report over 1992 and 1993.  They incorporated their knowledge of other streamflow requirements for fish survival (migration, incubation, past hydrology, gravel deposition and scouring, etc.) and with the fish habitat information from the IFIM report arrived at the following recommended instream flows at a location just downstream of the irrigation diversions: November through March = 575 cfs, April through July = 475 cfs, and August through October 180 cfs.

The 14 hydraulic models at the 2 sites allowed comparisons of the how the fish habitat changed in the Dungeness River with and without the 4 side channels at the upper site (the lower site did not have any side channels).  The fish habitat results for the upper site at RM 4.2 compliments the results at RM 2.3.  The lower site is diked so there is only one channel whereas the upper site is not diked so it has 4 side channels and a main channel.  These 2 independent hydraulic models predict that the peak habitat flow for Chinook spawning is 200 cfs at the lower site and 220 cfs at the upper site.  One would expect to get similar results for the two mainstem channels if the side channels are not included.  This is an indication that both models were done well since they were done independent of each other.

However, there were still questions about how well the proposed instream flows would provide fish habitat in the all the side channels of the Dungeness River.  To answer that question in 2003 the Bureau of Reclamation did a very extensive peer-reviewed study of all the side channels.  Their report is available here:

This study provided detailed information on the flow interaction between the main and side channels and their impact on fish habitat.  The 10 large side channels downstream of the fish hatchery RM 10.6 to Woodcock Bridge at RM 3.3 were analyzed.  They measured when the channels became connected and disconnected from the main channel and how the fish habitat changed in the side channels as the mainstem flow changed.  The 1991 PHABSIM/IFIM study only measured side channels at river mile 4.2 and these side channels totaled about 600 linear feet whereas the 2003 Bureau of Reclamation study covered 10.6 miles of river with 24,080 linear feet of side channels.

Further analysis on the flow interaction between the main and side channels and quantification of the amount of fish habitat in the side channels was done in 2007.  This was a study on the potential effects of surface water withdrawal on the freshwater habitats in the Dungeness River. T he following link contains the report Task 4 Dungeness River Aquifer Recharge Habitat Technical Memorandum by Ron Campbell (R2 Consultants, Inc.) done May 31, 2007. 

After review by the agencies’ technical experts the graphs showed the proposed instream flows of 180 to 575 cfs would protect salmonid habitat in the side channels throughout the river.

Q:  Does the PHABSIM/IFIM hydraulic model have a 100 foot error in elevation between the side channels and the mainstem channel at the upper site?

A:  No. The elevations of the water surface at each transect are surveyed relative to a benchmark.  This benchmark is often the top of a nail in a tree alongside the channel.  The benchmark is surveyed and then the water surface is surveyed so one can determine how much the water surface has dropped from the last time you measured the streamflow and surveyed.  The calculations are based on how much the water surface has dropped relative to the height of the top of the nail.  The water surface elevations is subtracted from the arbitrary benchmark number to get the new water surface elevation.  In this case a benchmark of 100 was used for the side channel and 200 on the main channel.  There is no error with the surveyed elevations because the elevations are relative to the benchmark.  They are not the true elevation relative to mean sea level.

Q:  Is the PHABSIM/IFIM model flawed because I can’t tell exactly when the side channel at RM 4.2 connects and disconnects from the main channel?

A: No, the model is not flawed. The PHABSIM/IFIM model’s purpose was to calculate the amount of fish habitat at RM 4.2 at different flows. It was not intended to be a definitive study of when side channels connect and disconnect in the Dungeness River.  However, that definitive side channel study was done in 2003 by the Bureau of Reclamation in 2003, and reanalyzed with the PHABSIM/IFIM study in 2007 by R2 Consultants.  These studies are available at the following links:

Q:  Is there an inconsistency between the PHABSIM/IFIM report and the Bureau of Reclamation Side Channel Study?  The IFIM report shows no side channel fish habitat until flows reach 240 cfs whereas the Side Channel Study shows lots of side channel habitat at 110 to 480 cfs.

A:  There is no inconsistency between the two reports. The links to the reports are listed above. As noted above, the 1991 PHABSIM/IFIM study only measured side channels at river mile 4.2 and these side channels totaled about 600 linear feet whereas the 2003 Bureau of Reclamation study covered 10.6 miles of river with 24,080 linear feet of side channels.  See Table 3 from the Bureau of Reclamation study:

Table 3 from the Bureau of Reclamation study

One would expect the 2003 Bureau of Reclamation (BOR) study of 24,080 feet of side channels covering 10.6 miles of river to give a wider range of when side channels are connected to the mainstem than a 1991 study with 600 feet of side channel at one spot in the river.  The 2003 BOR Side Channel Study found (to pick one species for example) bull trout rearing habitat starts at 160 cfs at RM 4.7, at 110 cfs at RM 5.4, at 115 cfs at RM 6.9, at 120 cfs at RM 10.2 west side, and at 10 cfs at RM 10.2 east side.  However, when one analyzes the BOR study with PHABSIM/IFIM study you will find that they support each other very well in determining that the instream flows proposed of 180 to 575 cfs are correct.