Older Featured Articles July 2008 June 2008 May 2008

August 2008 Featured Articles

A Review of the Biology and Management of Umpqua River Wild River Steelhead
E. Eric Knudson, Ph.D. - July 30, 2008

 Although there is no specific documentation, anecdotal information indicates that Umpqua  steelhead populations could be smaller than they were in the past.

 There is no apparent frame of reference for estimating habitat-based carrying capacity for steelhead in the Umpqua basin to understand whether the physical habitat is being fully utilized.

 The ability of spawner-recruit models to detect the true carrying capacity of Umpqua steelhead habitat is questionable and therefore may lead to an underestimate of the population potential.

 The effects of salmon-delivered, marine-derived nutrients on Umpqua steelhead are unknown, but declines in nutrients could have contributed to the perceived reduced steelhead populations.

 Relatively greater harvest rates in years of smaller run sizes may potentially reduce the spawning stock to below habitat capacity.

 There could be important discrepancies in some of the methodologies applied to stock assessment, which should be further investigated, such as:

o Trend-related discrepancies in the dam counts because of changes in counting methods over time.

o Potential biases in the population estimates, including the influence of changing harvest regulations.

o Potentially inappropriate extension of the 2003 telemetry and markrecapture results to earlier years.

o Possible violation of the Peterson mark-recapture assumptions.

o Difficulties with using spawner redd surveys for abundance estimates.

o There are a number of apparent weaknesses in the methods used to estimate Umpqua steelhead exploitation rates.

o Details of the population viability analysis

o Inherent inaccuracies of creel surveys

 Umpqua steelhead management would benefit from additional studies on population differentiation and life-history and population diversity among the major population segments, particularly to determine whether some segments should be managed separately from others.

 It is essential that Umpqua habitat management be focused on maintaining any possible sanctuaries, protecting existing good habitat form deterioration, and restoring habitat where it has been degraded, especially with regard to temperature problems in many stream reaches.

 The currently used warning criteria for Umpqua steelhead management action are a cause for concern. The first warning sign is if the 6-year average in the Winchester Dam count falls to an average of 1,667 fish. This is only about 22% of the long-term average count. Management action to maintain the population should begin well before the declines ever became this severe. The second criteria is a continued downward trend in counts at Winchester Dam when other stocks in the winter steelhead ESU continue to remain stable. Management action to preclude Umpqua steelhead declines should not be dependent on the performance of neighboring stocks.

 If harvest is to be implemented, a management scheme that includes greater harvest in years of greater abundance, and less harvest in poor years, would be preferable to a constant percent harvest rate.

 Hatchery operations may be influencing wild steelhead when a) the actual proportion of hatchery-origin fish incorporated back into the hatchery broodstock may be too high in some years, b) the incidence of interbreeding of hatchery origin fish with natural-origin fish in the wild may be reducing the fitness of the natural population, c) straying of hatchery-bred fish into the wild population may be exceeding the target 10% acceptable rate, especially in the area around Rock Creek and the adjacent North Umpqua River, and d) some hatchery steelhead and their offspring may be contributing to wild winter steelhead population declines through competition for spawning and rearing habitats.

 Effective management could benefit from greater monitoring at all life stages. In particular, smolt trapping on the mainstem of the North Fork, South Fork, and Smith River could greatly assist in understanding the relative freshwater survival and condition of each wild and hatchery population.

 Umpqua steelhead could benefit from advanced, life-history and habitat-based population modeling.

 There are a number of cases where institutional policy is inadequate to support effective management of Umpqua steelhead.

 A conservation plan, as described under the NFCP, is clearly needed for Umpqua steelhead management.

Comments on the Eric Knudsen review by Bill Bakke
- August 05, 2008

The section of the report in the Executive Summary on major findings is sound and presents a course of action with ODFW.  I hope the Steamboaters will discuss this document with ODFW and it may be necessary to arrange a discussion between Knudsen and ODFW staff.  Certainly, this report provides the basis for developing a conservation plan under the Native Fish Conservation Plan and the Steamboaters should make that request of the agency.

Comments:

Page 3:  I agree that harvest rate management is a poor policy as indicated by Knudsen, however, I do question the assertion that harvest should automatically increase when abundance increases.  Observations confirm that fish agencies increase harvest when run strength increases after a period of low returns.  The fish need an increase in adult spawners for all the genetic, ecological, and nutrient enrichment factors important to their breeding and rearing success that Knudsen identifies.  For example, Columbia River sockeye are primarily a wild run and one population is listed as an endangered species.  The river has been closed to directed harvest for some years, however, when the sockeye return suddenly increased this year, the agencies moved in with a directed commercial fishery.  This harvest may not provide an increase in spawner abundance, even though established escapement goals may be achieved.

This kind of harvest management is call abundance based harvest.  When a run gets to a certain size it is harvested at a certain rate.  The Willamette spring chinook are listed under the Endangered Species Act and their status is described as “very high risk of extinction.”  This fishery is managed on abundance with a harvest rate rather than a spawner abundance goal by population and watershed.   A harvest rate sets a particular rate regardless of the run size.  This policy is based on hatchery production rather than on wild spawner abundance goals.  In Alaska harvest is based on escapement goals for wild populations and they have more fish to work with than we do in Oregon and Washington. We made a decision to replace wild production with hatchery fish and we have designed our harvest management program to fit the hatchery model. 

In 2008 the Willamette spring chinook run flopped and the ODFW became concerned about falling short on the hatchery egg supply, so they closed the fishery.  They were not willing to risk a hatchery failure by sticking with harvest rate management.  However, the Molalla River was left open for harvest of hatchery spring chinook which are outplanted into that river. It was the only river in the entire Willamette Basin open for a chinook kill fishery. Since there is no hatchery at risk in the Molalla, the ODFW allowed a fishery even though it would impact wild spring chinook.  The Molalla River wild spring chinook run status is very high risk of extinction, yet the state allowed a barbed hook and bait fishery for hatchery fish that would increase the incidental mortality on wild chinook.  This is a good example illustrating the ODFW abundance based harvest management program tied to hatchery production and insensitive to the needs of wild salmon. 

On the Umpqua, the state will manage for hatchery production and not wild fish.  Their management policy is to maximize a kill fishery.  When I was provided a tour of the Umpqua a few years ago by ODFW, the hatchery program was a problem. A clear conservation management program for wild winter and summer steelhead should be the priority, but as Knudsen points out there are problems that need correction.

Page 5:  Knudsen points to work by Reisenbichler, saying that electrophoretic analysis of steelhead on the Oregon coast suggested that steelhead should be managed in each river as separate populations.  In large systems such as the Umpqua, there may be more than one wild winter steelhead population.  Steelhead and spring chinook tend to form distinct breeding populations and in large rivers like the Umpqua, they can form several distinct populations.  The goal would be to determine the stock structure of Umpqua wild winter steelhead using DNA analysis, and establish a management program that maintains the integrity of that stock structure. Lacking that data, the precautionary approach would be to manage wild steelhead as distinct breeding units in the Umpqua basin.

Page 8:  The life history table presents a comparison between wild and hatchery steelhead.  The hatchery fish have diverged phenotypically from the wild in at least 8 of the 9 characteristics identified.  Changes in phenotypic characteristics influence survival and in a few generations can become fixed as genetic traits.  The hatchery program is producing fish that are distinct from the wild form and when they interbreed with the wild fish, they will degrade the reproductive success of the wild steelhead.  In a recent public comment, Mark Chilcote, ODFW, said the reproductive success of the wild run declines in direct proportion to the number of hatchery fish spawning naturally.  So if there are ten percent hatchery strays spawning naturally with wild fish, the wild run’s reproductive success is reduced by ten percent.  If wild fish are taken for an egg supply as is common in native broodstock hatcheries, the hatchery program is responsible for decreasing the production of wild fish in order to produce hatchery fish.  For more information on the problems associated with native broodstock hatcheries see my paper on the topic on the NFS web page.

Page 9:  Hatchery fish did not survive well until the 1960s with improvements in feed and pathology.  Prior to that time the runs were primarily composed of wild origin fish.  The count of steelhead over Winchester Dam from 1946 through 1962 would primarily be wild steelhead.  The paper I have does not provide the number of steelhead counted each year over the dam during this period.  However, that will give you an estimate of run size for part of the basin to compare with today’s run size.  Bill McMillan has been doing run reconstruction for some rivers in Puget Sound.  By examining commercial catch records in the 1800s, he was able to estimate that the wild steelhead run in the Stillaguamish River range between 70,000 and 90,000.  Today the wild run is about 500 fish and the management goal is just 2,000.  The state did not look at the historical record in developing its management goals for the river, so committed one of the problems we are faced with today, that is, a travelling base line.  Each generation of fish managers assumes the run size they have is what was always present.  The Umpqua is a larger more diverse watershed than the Stillaguamish River so it is reasonable to expect the historical wild steelhead run size to be much larger.  If the Stillaguamish is functioning now with a wild run size of 0.05% then the Umpqua wild steelhead run size is probably comparable.  This means that the wild steelhead run size was historically greater than today by 95%.  Knudsen did not look at the historical commercial catch, so he was unable to make an estimate of wild steelhead abundance prior to modern time. Also, given the information in the paper, it is not clear to me when ODFW began to make a distinction between wild and hatchery steelhead passing Winchester Dam.  It would have happened when all the hatchery fish were externally marked with a clipped adipose fin. 

Page 12:  Knudsen states that “…there are a number of stream reaches that are considered to be wild steelhead sanctuaries.” However, those stream reaches are not listed in the report.  It is important to know their name and location.  It would be a good idea to expand the wild steelhead management waters in the basin (using weirs) to prevent hatchery fish from spawning naturally with wild steelhead. 

Page 12:  Knudsen says that the ODFW habitat survey has not been used to “develop capacity estimates.”  He is correct in his assertion that ODFW needs to develop habitat capacity estimates for all life history stages of steelhead in the basin and establish spawner abundance goals based on those estimates.  Lacking that, wild steelhead are not being managed properly for reproductive success and the fishery will suffer.  The hatchery program is the basis of ODFW management not wild steelhead and this has to be changed if the Umpqua watershed is to remain productive to its full potential.

Page 13:  Kundsen’s discussion of marine derived nutrients from salmon in the basin is very important.  Spring chinook and coho are the two species that primarily benefit steelhead because they penetrate the watershed more than other species, going into upper reaches and small tributaries.  While some ODFW biologists have embraced the importance of nutrient enrichment from salmon carcasses, it has been largely ignored by the leadership as an important ecological value.  The literature indicates carcass density of 1-3 carcasses per meter squared is important nutrient level.  For coho, an estimated 200 spawners per mile are needed to provide adequate carcass enrichment.  In Washington State, Bilby says that coho carcass levels are less than one per square meter.  There has been an interest in placing hatchery carcasses into streams or using processed salmon pellets for the same purpose.  However, the distribution of those carcasses and pellets is a problem because they are placed in certain locations rather than along the entire stream length.  The best way to get adequate nutrient enrichment is to have adequate spawner abundance for these fish will distribute the nutrients throughout the system.  I spoke with one biologist in Washington State who has been testing the idea of managing harvest to achieve nutrient enrichment, but was told by the harvest managers that it was too much of a constraint on harvest to achieve the level of nutrient enrichment needed.   This indicates that harvest management is a key institutional limitation on fully seeding our streams with carcasses for stream enrichment.  Also, because coho are in low abundance and now listed as a threatened species, it is unlikely that coho will provide the nutrient enrichment from carcasses that are needed. 

In speaking with Mark Chilcote he alerted me to the issue of escapement and nutrient seeding.  His premise was that at a certain spawner abundance level, using the Ricker curve, that the productivity of a run declines.  This is believed to be due to density dependent mortality.  He postulates, however, that we should test that assumption by having more spawners over several years to see if the productivity of the stream increases as nutrient levels increase over time.  The stream could become more productive with more spawners and the threshold for density dependence would increase above what it was prior to the increase.

Page 14:  Knudsen says that “reliable smolt estimates could eventually be used to differentiate freshwater from marine survival limitations.”  The ODFW should be tracking the productivity of wild steelhead in the freshwater environment by counting smolt yield throughout the basin and using adult traps and dam counts to estimate marine survival.  Only in this way can the agency effectively measure the productivity of the freshwater and salt water environments.  Without this it is impossible to evaluate habitat improvements and to effectively manage for adequate adult spawners.  Right now we get what we get and that is a rather passive and stupid approach to management.  

Page 14:  Knudsen states, “Relatively little is known about the genetic and biological diversity among and within the various Umpqua steelhead populations.”  It is not an uncommon situation, but it is deplorable for it means the conservation management program is ineffective and incompetent.  It will only mean loss of important attributes and diminished abundance and fisheries.  In forestry, they know enough through their genetic inventory that a tree is not a tree.  Trees are selected for site, elevation, aspect and microclimates.  They have learned that planting a tree in the forest is not enough for the yield in fiber (harvest) is low and may be non-existent due to stunted growth or mortality.  Using trees from a watershed is the preferred approach to reforestation.  In fish management we have not come that far, thinking for the last 150 years that we can replace wild diversity with domesticated hatchery products.  While fish management is based in an agricultural model, it is not a very sophisticated one and is far behind the thinking in agriculture today with its emphasis on genetics and locally adapted species.   However, there is a strong corporate movement to replace locally adapted races of plants with generic “green revolution” stock and it operates like the modern salmon hatchery program.  It is expensive, and less productive.  Maybe the Steamboaters should become the equivalent of seed savers in agriculture and become advocates of locally adapted wild steelhead populations and the habitat diversity that supports them. 

Page 15:  Knudsen mentions repeat spawners and there abundance.  While Canada pays attention to this life history attribute of wild steelhead, it is seldom an item on the agenda among USA biologists.  Only recently, has there been any investigation of steelhead kelts and repeat spawners in the Columbia.  This life history trait is important for it is risk adverse and increases egg deposition.  Repeat spawners are survivors and effective spawners so they should be an important component of steelhead management.  A decline in repeat spawners is recognized as an important issue on wild steelhead streams in Canada, but in the US they are largely unnoticed. 

Page 16:  Knudsen asks a basic question of management: “Does management maximize production.”  This is a key question to always ask land, water, and fish management agencies, for they are responsible under state and federal law to prevent the depletion of indigenous (native) species. I remember a biologist said one time that a hatchery can increase production but it cannot increase productivity.  The research into hatchery and wild fish performance, reproductive success and survival has only confirmed that statement.  ODFW is more concerned with providing production for harvest than they are in maintaining the basis for productivity of our watersheds and wild stocks.  The record is very revealing and should be a major issue in salmonid management.  Wild production is from 3-5% of historic abundance and hatcheries have not made up for that loss.  In fact, we now know that they interfere with it.  The pattern has been set for extinction of salmonids on the West Coast, but managers have not done what is needed to reverse that pattern and trend.  I did an evaluation of coho salmon plans by Oregon and found that the biologists were deeply concerned about the decline of wild spawners, but their adopted plans over 82 years have been ineffective in reversing that decline.  The dodge is to blame the ocean or habitat, but fish management is also responsible.  Until the state accepts that premise, there will be no decision to have management maximize productivity of salmonids. 

Page 17:  Knudsen says that the “common goal of harvest management is to allow a reasonable amount of harvest, while optimizing total production, and without negatively affecting future production.”  Unfortunately, this common goal is not followed.  Harvest management is not based on delivering a spawner abundance goal for each fished population, and overharvest has become a major player in the demise of wild salmonids.  I like his description of the common goal, but it is impossible to find it operating in fisheries today.  In fact, the long standing policy of harvest management in Oregon and other states is to purposely overharvest salmonids, for having fewer spawners means that there will be less competition for food and space for juveniles and more juveniles will survive  to smolts and produce adults.  This premise ignores the ecological value of salmonids and the genetic realities of managing a wild population in a fluctuating environment.  

Page 18:  The discussion of the formula for a productive NF Umpqua steelhead population assumes that the formula is able to deliver abundance, diversity, and distribution needed to maintain a productive population over the long term in a changing environment.  There are examples in the literature that show an intrinsic production level of 25 rather than 1.25.  The minimalist formula for a managed population is at the extreme edge and just over replacement.  By managing on that premise, the ODFW can justify its fishery program at the expense of a healthy wild population that is resilient and even large.  I have argued hopelessly with ODFW scientists that the agency should stop managing against extinction and begin managing for abundance and diversity.  The intrinsic productivity of 1.25 is managing just above extinction and does nothing to achieve the potential of the population and the ecological benefits and services it can provide.

Page 20.  Kundsen’s critique using studies by Araki, Chilcote, and Kostow are right on target.  The harvest augmentation hatchery program is a threat to wild steelhead in both the South and North forks of the river.  ODFW has adopted the premise that because it is an integrated hatchery program it is clean and safe.  Nothing could be further from reality given what their research shows.  It is interesting that a stray factor of 10% is the same as the limit adopted as policy under the Native Fish Conservation Policy.  It is likely that the stray rate is both higher and lower than 10% if there was an attempt to evaluate the strays by stream in the basin, especially in Rock Creek and the North Umpqua.  The NMFS recommend a stray rate of no more than 5% for fish that are non-native to the watershed they stray into.  Since steelhead are locally adapted and develop distinct subpopulations any strays from a hatchery population are a problem for the reproductive success of the wild run. Even though they are of native stock, they have diverged from the native fish due to their hatchery selection and are actually a foreign and distinct stock.  Given the fact that hatchery fish become domesticated and are changed so that they are a distinct hatchery population in the first generation, a problem that only increases with more generations under artificial production, the stray rate of 10% is of concern.  It is a concern because these strays reduce the reproductive success in proportion of their abundance on the spawning grounds.  Also, it must be kept in mind that the stray rate is not calculated for each distinct wild breeding population, but by the number of strays in a watershed.  So the overall stray rate for a watershed may be 10%, but certain streams may have a stray rate in excess of 10%, but that is okay because it falls within the rule.  This means that some wild populations have to deal with the degrading effects of strays at a much higher rate than 10% but because the estimated stray rate for the whole watershed is consistent with the rule, ODFW concludes there is no problem.  That is politics not science running the management program.

Page 21-22:  Knudsen’s discussion of spawner-recruit models is very good and important for developing a management plan for the Umpqua.  ODFW is fond of using the Ricker curve and Maximum Sustained Yield as the basis of their fish management program, but for all the reasons laid out by Knudsen, this approach is not a sound one for conservation.  The purpose of this approach is to get more fish in the boat rather than on the spawning grounds.  It has served the agency well, as long as one does not happen to notice the number of ESA-listed populations and declining runs.

Page 23-24:  Knudsen points out that ODFW has adopted a statewide steelhead plan (1986), but has done little to follow it or implement it in the last 22 years.  When Lichatowich asked ODFW biologists about their implementation of the Oregon Wild Fish Policy, he found that it was unevenly applied around the state.  We now have the Native Fish Conservation Policy that will replace the Wild Fish Policy once a conservation plan is developed for a watershed.  As far as I can tell, the Steelhead Plan (1986), the OWFP (1978), and the NFCP(2003) are not being applied to management.  The Steamboaters should insist that these be applied to the Umpqua and that the state develop a conservation plan for steelhead consistent with the NFCP and the Steelhead Plan. 

Knudsen’s evaluation is very helpful in identifying problems that should be fixed to improve conservation management of Umpqua steelhead.  However, his evaluation is often quite technical, so it would be necessary to have him deliver his evaluation to ODFW with the goal of making improvements in management that the Steamboaters could follow up on to make sure they are actually applied.