Click for next page ( 303


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 302
1 ~ Hatcheries . - Anadromous-salmon hatcheries constitute a primary human intervention in the Pacific Northwest, and in many areas of the Pacific Northwest hatchery fish make up the majority of salmon in rivers and streams (see Box 12-13~. In this century, the intended goal of most hatchery programs has been mitigation. Miti- gation aims to lessen the immediate impact of human actions through definition of a "socially acceptable" altered state (Christie et al. 1987~. Hatcheries were expected to lesser the impact of numerous human actions that have dramatically altered freshwater ecosystems of Pacific salmon. For instance, the impact of construction of mainstem dams on the Columbia River was the decline or loss of many upriver populations. The socially acceptable state chosen was large-scale hatchery production in the lower Columbia, below the dams, to substitute for lost upriver populations. Both degraded environments and overfishing have contrib- uted to the dramatic decline in the numbers of mature salmon that escape capture or death arid reach natural spawning areas. Hatchery systems have been expected to compensate for this decline in escapements. A number of long-existing Pacific salmon art~f~cial-propagation programs have been touted as successes, but these claims have increasingly been called into question (e.g., Riddell 1993a, Gharrett and Smoker 1993, BowIes 1995, AlIendorf end Waples in press). Most artificial-propagation programs have not undertaken long-term evalua- tion and documentation of the extent to which intended goals were reached (e.g., increase the catch for a given population, prevent extinction of populations whose spawning grounds were destroyed by dams) and unintended risks were imposed (e.g., adverse genetic or ecological impacts on naturally reproducing fish). For many artificial-propagation facilities, this lack of long-term monitoring makes it 302

OCR for page 302
HATCHERIES .................................................................. ... ......... ....................... . ~t,.~-,.,, ~: .......................... .................................. ..... . . .. ..... .. .. ... ......... .. ..... ... . ~.~. ~` =~es o~ once o' "~ - se" ~' ..... ...... .. .. .. .. ........ .. ... .......... ... ....... ............ . . . . ........................ . ......................... ............ . ........ ............... . ~ ~. ....................................................................................... . ~. ~. t ~. ~ ., I .............. ................................ ~ . .. . . . ~.~n, ~m t~ h~ fISh ...... : : :::::: : :: :::::: : :: ::: ::::: ::: ::::::: . =~ - 6^ {n 1~ ~ .~ tn .~1 ln S~e 1~s =9 - ~. ~ : ~ ~ ~ ::: ::: : ........ - am ~.~.~.~.~ - . a~ ~t fISh am ...: Ems t.~.~.~:= sm~' ~ ~ ~ . $ . ~...... ............................................ ..... .. ................. .......... . ~., ~.~: En' 0"~:~"" X :: . ........................... .''.'. . .' - , to.i '=.~g'""~""~" '~%' '~"'the "'~""""'~"''~''t~ ~'"'~'.''2'm"''e''~ '' ' ' ' :::::S::: ' '' ' ' '' 'I'"' ~'~' , . ....... . ~ ~Am wIdI ~ . ~. 9~. :::: :: ::: ::::: :: :: ::::: :.. ax tn the ~ Riser bow: mIlan~ on ~ mu: ~ ^~k {n .. .. :: : : : x ...~..~.. ~ 75 m {~ .. ~ Em. ~ ...~..~.~...~..4 ~ ... at... .... ...... :: Em::: :::: :: ::: :: :::: :. :: ::::::: :: :: :: :: :::: :::: :: ::::::::: :::: :: ::: m~ ~ al .~) : ~ . . it ~...~= ~ =~ ~ns 1n ~ ~at ~s wer~ ~ - Q..d, A ' ' ~ ' "' " ' ' 'A . . .............................. ....... .......... . ................... . . . .......... ............. ''''""'''""""""""""""'sx"""""'Pmm"""1'~""~"'the'"'1"~:''''m ~., ~O ~It ~ , .. . . . .... ..... .... .......................................... . ~. ~a '""~'."'"''~''"''''m"'"'2'""""""'"""'' ''' I' ................................................................................................................................... ...... ......... ~ ~ :: : By: : : ::: : :::.: : : : ::::: : :::: ::: ::::: ::::: : : :: :: :: ::::::::::: ::: ::: ::::::: : :: :::: ::: :::: : . ~ . H~ ~h ~ ~ ~...~0 ... ~ : ::: : :: :: : . ::::: : :: : : 0~.~.~. . ~ ~ ~ ~ st - Ad. .. Em. x : : : :: w=bx~n a~ ~e =~ P~r (~ 1 9~) ............ ....................................... .. .... . . ................................................................................ "'.'..'.'.'..2...'..'.''"""""""'lx"""'T~''~'~"~:~"'spn~"'Cht~"'cn"~"'R~''Stw(''~ ..... ................................................................. . ~ ~. gI ~w . e n ~ .... ........ ..... . Amp:. T.~h ~ 19~ a~ 1991 wane ... ~ ~e :: : :: :: ~ ~ ~m ~ 0 ~ ~ ~ s 0 ~ ~ ~ ~ ~r x: : : : .. ~ ... ..... Exam ~ wIU~ c~y ~e ~ Oh :::: : :: :: ::: . ::: ::: ::::: ::::::::::: ::: :: ::::: ::::: :::: ::::::: ::: ::::::: :::::: Ace. x x x~" ~8 0p ~ ~ $~l p~= , . . . . ..... ............................................... . . $ ~..... t ~. xn ~x ~ ..... ....................... ~ - r (~ Mana~= nc't . ~. ~,l.~. ~ . ~.................. ........ ............. . ..... ..................... "2g'~ "''"'"'~.'."'''."'""'~."'' "I.""""""""""""" ' '' .. . . . . annoys ln .~ 1~1 . .~ .~f ~ 1~ ~ .. .. ~ ............... ..... .. . . ::: : ~::::: :: :: :. : ::: : 303

OCR for page 302
304 UPSTREAM: SALMON AND SOCIETY IN THE PA CIFIC NORTHWEST nearly impossible for us to differentiate impacts of hatchery programs from im- pacts of other human interventions or of natural environmental changes. Since its inception in 1977, even the ambitious Salmon Enhancement Program of British Columbia failed to collect data needed to evaluate its benefits and risks (Hilborn and Winton 1993, Winton and Hilborn 1994~. Over the last century, the Pacific Northwest culture passed up the chance to learn adaptively about artificial propa- gation of anadromous Pacific salmon. In this chapter, we discuss the effectiveness and effects of hatcheries. In retrospect, it is clear that hatcheries have caused biological and social problems. For example, hatcheries have contributed to the more than 90% reduction in spawning densities of wild coho salmon in the lower Columbia River over the past 30 years (Flag" et al. 1995~. Hatcheries were not part of an adaptive man- agement program, so inadequate evidence was gathered to assess their impacts. As in so many other aspects of the salmon problem, there is a dearth of good scientific information about hatcheries' effects. In this chapter, we base our discussions of hatcheries' effects on a synthesis of biological principles and empirical evidence. Generally, the problems stem from the predominant goals of hatchery opera- tions increasing run sizes to mitigate other human-induced mortality-and from insufficient incorporation of basic genetic, evolutionary, and ecological prin- ciples into hatchery planning, operation, and monitoring. In addition to the direct effects of hatcheries, management practices often associated with hatchery opera- tions but not required by them have also had adverse effects on wild populations. The practices include outplantings (population transfers), overfishing of wild fish in mixed fisheries on hatchery and wild fish, broodstock extraction, and environ- mental manipulations (examples include the building of weirs and removal of salmon carcasses from streams) associated with the management of hatchery fish in the wild (Campton 1995~. Hatcheries have had some successes, but the ideas behind their operation and the ways that they are managed and operated will need to change if they are to play an important role in maintaining the long-term sustainability of naturally reproducing salmon. PROBLEMS ASSOCIATED WITH HATCHERY PRACTICES Traditional approaches of hatchery programs have imposed different types of biological problems on salmon populations, including demographic risks; ge- netic and evolutionary risks; problems due to the behavior, health status, or physiology of hatchery fish; and ecological problems. One or more of those problems might have affected either the populations that a hatchery program aimed to rebuild, other populations with which they interact, or both (Allendorf and Ryman 1987, Gharrett and Smoker 1991, Kapuscinski 1991, Riddell 1993b, Allendorf and Waples in press). The extent of the effects of any particular

OCR for page 302
HATCHERIES 305 hatchery program is hard to document because of a lack of appropriate monitor- ing over appropriate periods. Given the Pacific Northwest's great diversity of hatchery operations and ecological contexts in which the hatcheries have oper- ated, there probably has been much variability in the degree to which any of the problems has occurred (see Box 12-2~. Growing scientific evidence supports the notion that hatchery-caused problems cannot be ignored without further threaten- ing the future of depleted salmon populations. Demographic Risks Large-scale releases of hatchery fish have greatly exacerbated the mixed- population fishery problem. Less productive populations in the mixture, often the naturally reproducing ones, are overfished as a consequence of relatively high exploitation rates, which are set in response to the relatively large contribution of hatchery fish to the mixture. Purely because of the demographic problem of excessive mortality of sexually mature adults, wild populations in the mixture are eventually driven to extinction as their escapements drop below replacement levels. Chapter 11 discusses overfishing and mixed-population fishing in greater detail. Genetic and Evolutionary Risks Current approaches to artificial propagation can impose four types of genetic risks on hatchery fish and interacting populations (Busack and Currens 1995~: extinction due to such genetic problems as inbreeding depression, loss of popula- tion identity (between-population genetic variation), loss of within-population genetic variation, and domestication selection. Redesign of hatchery programs can greatly reduce imposition of these risks, particularly the loss of genetic varia- tion between and within populations (Campton 1995~. Population Identity and Within-Population Variability Numerous studies involving comparisons of hatchery and wild populations have yielded strong evidence that traditional hatchery practices have caused losses of genetic variability between and within anadromous-salmon populations (e.g., Allendorf and Ryman 1987, Riddell 1993a, Allendorf and Waples in press). Only isolated sets of data are available that can be used to estimate roughly the prominence of hatchery-released fish for particular species in particular loca- tions. Box 12-1 contains some examples of these. Release of hatchery fish propagated from nonindigenous broodstock without consideration of existence of local breeding populations and metapopulations (see Chapter 6) was common in the past and was a principal mechanism of loss of genetic variability between populations. Most hatchery programs now avoid this practice. Straying of hatch

OCR for page 302
306 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST ery fish, another mechanism of loss of between-population variation, is more difficult to prevent. Waples (1991) reviewed evidence suggesting that hatchery fish stray at a higher rate than naturally reproduced fish, but there are few data sets on straying of wild salmon (Quinn 19931. Within-population genetic diversity will be eroded if the effective population size (No) of the hatchery population-a number usually smaller than the number

OCR for page 302
HATCHERIES 307 of adults actually mated in the hatchery is depressed because of insufficient numbers of broodstock and inappropriate mating protocols (Simon 1991, Kapuscinski and Miller 1993~. Traditional hatchery broodstock practices have often led to an N.: much smaller than the census population size (Simon et al. 1986, Bartley et al. 1992b). In some hatchery broodstocks, deleterious changes suggestive of inbreeding depression, such as increased incidence of abnormal vertebral columns and missing fins, have been associated with declines in within- population genetic diversity (e.g., Allendorf and Leary 1988~. Within-population genetic diversity can also be eroded if artificial selection against a given heritable trait recurs breeding season after breeding season as a consequence of some hatchery practices (as explained in the next section). Genes for quantitative traits (i.e., such continuously varying traits as weight, growth rate, and fecundity) usually occur in blocks of polygenes, whose products interact to affect the physi- ology and fitness of individual fish. Recombination among such polygenes is an important basis of adaptation via natural selection. If repeated artificial selection occurs in a hatchery program, genetic diversity of these polygenes will decline, closing off options for evolution and jeopardizing long-term persistence of the population. For example, culling of smaller fish imposes a selection differential against length, weight, or other size-related traits (see Figure 10 in Kapuscinski and Jacobson 19871. Alleles that influence the size traits are removed from the population (Lacy 1987, Falconer 19891. In addition, alleles that are linked to the alleles for size traits but whose phenotypic effects and fitness value are usually unknown are removed from the population. Domestication If fish become domesticated by genetic adaptation to a hatchery, they will have a commensurate decline of fitness in natural environments. Domestica- tion selection imposed by human actions can be imposed in two ways. One potential source is nonrandom collection of hatchery broodstock over the dura- tion of a spawning run. Genetic response to this source of selection is well documented. For example, Steward and Bjornn (1990) reviewed numerous ob- servations of hatchery stocks shifting to early run timing (a genetic response) after consecutive collections of hatchery gametes primarily from early-returning adults. A second source of domestication selection is altered selection pressures due to differences between the natural environment and the hatchery culture resulting from physical conditions or operational practices in the hatchery. Effects of hatchery practices. Historically common hatchery practices, such as size-grading of smelts, differentially affect survival of hatchery fish in natural environments. The risks of such selective practices are often recognized today, and these practices seem to have been stopped or much reduced. Explicit state- ments about the undesirability of such practices should be included in hatchery

OCR for page 302
308 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST policies and guidelines for all hatchery programs that affect wild populations (Kapuscinski 1991, Kapuscinski and Miller 1993~. Disruption of natural sexual selection by artificial matings. The practice of making artificial matings-now the dominant hatchery method is a serious concern because it disrupts natural patterns of sexual selection with negative implications for fitness of hatchery fish in natural environments. Sexual selec- tion is an important part of the evolution of most vertebrate species. The salmon case is especially striking, in that millions of "trial balloons" (fertilized eggs) for the next generation are produced all at once by the spawners of any generation, followed by the death of the fish that produced them. Sexual reproduction in salmon, as in most other forms of life, is the exclusive gateway to the genetic future of each of the many populations. The relatively small number of breeding salmon that arrive at a home site to spawn are true survivors. In some manner, throughout their life cycle, they have survived many challenges: predators, scarcity of food, accidents, diseases, and pollutants that might have weakened or killed them. Under natural conditions, sexual selection would normally occur within these few survivors. Sexual selection favors the pairing off and differential reproduction of some individuals at the expense of others. The elaborate structures and colors of spawning salmon-especially males, with humped backs, enlarged and hooked jaws, enlarged canine teeth, and brilliant reds and greens are used only in the mating process to increase the probability of mating (Quinn and Foote 19941. Breeding competition in Pacific salmon is especially concentrated and intense because most breed only once during one spawning season and then die, allocating all their breeding effort to one massive bout of reproduction. The most effective breeders in general are few, implying that this process of sexual selection is extremely intense and important evolutionarily. In hatcheries, the whole process is bypassed: humans select adult males and females, strip their eggs and sperm, and then rear the fry. The human breeders have no way of identifying the genetic relatedness of spawners or fish that would be the best natural breeders. Although not all the effects of this inadvertent interference with natural selection are precisely known, it is almost certain that one result is loss of general vigor, adaptation to local environments, and evolutionary fitness. Ex- trapolating from other examples of human interference in breeding processes in domesticated animals leads to the conclusion that the species' ability to survive unaided in natural conditions must be diminished. For example, substantial differences in potency among male pink salmon have been experimentally dem- onstrated (Gharrett and Shirley 19851. Such differences if they have a genetic basis would provide a mechanism for hatchery processes that bypass natural mating processes to produce genetic changes in breeding ability. Indeed, some direct evidence shows that genetic differences in reproductive behavior have

OCR for page 302
HATCHERIES 309 resulted from the bypassing of sexual selection in hatcheries (Fleming and Gross 1993, 1994). Probability of selection due to physical hatchery conditions. To date, physi- cal and biological conditions in hatchery facilities have differed greatly from those in natural environments. For example, hatchery rearing systems have lacked diversity in habitat structure, cover, diversity in flow and temperature regimes, and exposure of fish to natural prey and predators. The physical conditions of the hatchery rearing environment have the potential to alter selection pressures in one generation, lead to a genetic response (adaptation to the hatchery environment' in descendants, and thus decrease fitness in natural environments. However, the degree to which the potential has been realized is not clear from review of the published literature (Kapuscinski and Miller 19931. Rapid domestication has been demonstrated in captive broodstocks that have been reared in captivity for the entire life cycle over several or many generations (Vincent 1960, Flick and Webster 1962, Flick and Webster 1964, Mason et al. 1967, Moyle 1969, Johnson and Abrahams 19911. In that situation, genetic adaptation to the hatchery envi- ronment is greatly enhanced because of the lack of exposure of the population to natural selection in natural environments. The risk of domestication is less certain in most Pacific salmon hatchery programs, where only part of a fish's life cycle is exposed to hatchery selection pressures and the rest is exposed to a complex suite of selection pressures in freshwater and marine environments, including selection pressure by natural en- vironmental conditions, such as predation in the ocean, and by human-induced environmental conditions, such as dam passage, stream habitat alterations, al- tered stream fish communities (e.g., Reznick et al. 1990), and possible fishing selection pressure (e.g., Policansky 19934. Interactions among the various sources of natural and anthropogenic selection can be complex and nonlinear and can yield surprising and counterintuitive results. That is even true for a single anthro- pogenic source of genetic selection: size selectivity of fishing gear. Miller and Kapuscinski (1994) showed that genetic selection on size traits can vary substan- tially between fishing seasons. No selection, selection against large size, or selection against small size can occur in a given fishing season, depending on the interaction between the relatively constant selectivity curve of the fishing gear and the more variable characteristics of the fish population, such as sex ratios and the frequency distribution of different age classes (because of correlation with size classes). To determine the relative importance of hatchery environments that cause genetic change, compared with other sources of genetic change, and to guide hatchery reforms, the degree and severity of selection due to hatchery conditions would need to be established through research. Insufficient understanding of the potential of hatchery environments to cause domestication with consequently low fitness in the wild does not justify ignoring the issue in the formulation of rehabilitation strategies. Indeed, the risk of do

OCR for page 302
310 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST mastication should be explicitly addressed in any rehabilitation strategies that include releases of hatchery fish. Furthermore, questions about domestication should be directly addressed by a research program. Behavior The behavior of hatchery-propagated fish differs from that of their wild counterparts. These changes can result from learning or from genetic differences. The hatchery and stream environments differ in many respects, and the foraging, social, and predator-avoidance behavior patterns of fish reared in the different environments are often different (e.g., Moyle 1969, reviewed by Suboski and Templeton 1989, Mesa 1991~. The selection regimes posed by hatchery versus stream environments also differ. For example, the behavior of coho salmon surviving a year in the river and in the hatchery will probably differ. Fewer of the hatchery fish will survive the rigors of the stream, and their culling will almost certainly not be random with respect to behavioral traits. The lower mortality in the hatchery might reduce selection pressure against some behavioral traits, but the hatchery environment might also select against different behavioral traits. Hatchery fish could be more aggressive than wild salmon (e.g., Fenderson et al. 1968), and recent studies have shown that both aggression (Riddell and Swain 1991) and boldness when foraging (Johnson and Abrahams 1991) have heritable components. Thus, hatchery-reared salmon are often more aggressive (and larger) than their wild counterparts for genetic or environmental reasons. Paradoxically, they also generally experience much higher mortality once released from the hatchery (i.e., lower survival from release as smelts to return as adults) than wild fish. Efforts to use hatchery-produced underyearling coho salmon to rebuild popula- tions in Oregon had the worst possible result (Nickelson et al. 19861. The hatch- ery fish displaced the wild coho that were in the streams and the authors believed that this was the result of competition between the larger hatchery presmolts and the smaller wild juveniles. Other studies have shown fish size to be important in determining the outcome of competitive interactions (Chapman 1962, Mason and Chapman 1965, Chandler and Bjornn 1988~. In the Nickelson et al. study (1986), the hatchery-adult returns showed a large shift from late to early spawners; but early hatchery spawners contributed very little successful reproduction, so fewer offspring were in the stocked stream than would have been produced if the initial stocking of larger hatchery fish had not occurred. The lesson from that study is that any efforts to use hatchery fish to rebuild a naturally reproducing population should consider the ecological implications of and perhaps seek to avoid- changes in time of spawning and size at release for hatchery and naturally repro- duced fish.

OCR for page 302
HATCHERIES 311 Fish Health Although there is much information on the incidence of disease and effects on salmon and steelhead in hatcheries, the effects of disease on released hatchery fish and their interaction with naturally reproduced fish are not well understood (reviewed by Steward and Bjornn 19901. Disease is thought to be directly or indirectly responsible for substantial post-release mortality of hatchery fish. Disease outbreaks are relatively common in hatcheries and are managed with one or more of the following methods: disinfection of influent water if the water source is surface water and if finances permit, separation of carrier fish if they can be identified, therapy, and, if the disease is widespread in the hatchery stock, destruction of all the fish. Steward and Bjornn (1990) suggested that widespread use of chemotherapy to control disease in hatcheries could result in the develop- ment of new, drug-resistant strains of viral and bacterial pathogens. Research and development are being actively pursued to improve the sensitivity and cost effectiveness of screening for carrier and diseased fish and to develop vaccines for immunizations against major diseases for which therapy is not effective, such as bacterial kidney disease and infectious hematopoietic necrosis virus. In spite of comparatively high incidence among some hatchery-fish popula- tions, there is little evidence of transmission of disease from infected hatchery fish to naturally reproduced fish (reviewed by Steward and Bjornn 1990~. How- ever, there has not been much research on this question, and most disease-related losses in natural environments would probably go undetected. The ability of hatchery fish to transmit disease probably depends on the ecological characteris- tics that influence the spread and pathology of the particular disease, the environ- mental conditions of the particular site, and the abundance and distribution of the released hatchery fish. Disease concerns interact with genetic factors in some important ways. Inat- tention to the genetic risks of loss of between- and within-population diversity can seriously compromise inherited modes of combating disease. Substantial evidence has accumulated that parasites (including pathogenic viruses, bacteria, protozoans, helminths, and arthropods) play a critical role in the adaptive evolu- tion and persistence of natural populations of vertebrate species (O'Brien and Evermann 19881. A wide variety of genes that encode or regulate mechanisms for the host to combat infectious disease have recently been discovered in numer- ous vertebrates, including fish (Stet et al. 1990, Stet and Egberts 19911. These genes are polymorphic within species and populations. Dramatically increased vulnerability to infectious disease has been demonstrated in some natural popula- tions, such as the African cheetah, which have experienced demographic declines and dramatic reductions in overall genetic variability due to human activities. The implication for rehabilitation of anadromous salmon is clear: it is critically important to maintain the remaining genetic diversity within and between popu- lations to conserve diversity for genes that are involved in disease defense.

OCR for page 302
312 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST Differential susceptibility to the myxosporean parasite Ceratomyxa Shasta has been shown both between species and between populations of the same species of Pacific salmon. Fish originating from Columbia River basin tributar- ies where the parasite is endemic were less susceptible than fish from river basins that are free of the parasite (Zinn et al. 1977, Buchanan et al.19831. Furthermore, cross-breeding experiments demonstrated that interpopulation differences are genetically based (Hemmingsen et al. 19863. The results indicate that Columbia basin populations have evolved resistance in response to the natural selection imposed by the endemic C. Shasta. Differences between populations, correlated with different transferrin genotypes, have also been found in resistance to bacte- rial kidney disease and vibriosis (Winter et al. 1980, Withler and Evelyn 1990~. Other types of evolutionary adaptation to disease agents and other environmental stressors are probably present in the relatively few salmon populations that per- sist in their native watersheds (e.g., Bower and Margolis 19841. Conservation of the resulting spatial pattern of adaptations within different salmon populations native to different watersheds is essential for rehabilitation of salmon populations and, if rehabilitation is successful, for their long-term sustainability. Physiology The physiological state of hatchery fish often is suboptimal, even though they are sometimes larger than naturally produced progeny at the time of release, and this might lead to undesirable impacts on their health, on poststocking sur- vival, and interactions with naturally reproduced fish. Physiological stress in hatchery fish due to crowded rearing conditions, handling, and transportation probably increases their postrelease mortality (reviewed by Steward and Bjornn 1990:451. The contribution of environmental stress to reducing the immune response of salmon is well documented (reviewed by Steward and Bjornn 1990:63~. Incomplete smoltification of some hatchery fish is also a major cause of concern (e.g., Shrimpton and Randall 1992~. Research in the Columbia River basin suggests that poorly smelled hatchery fish lack strong downstream migra- tion behavior, so they reside longer in stream habitats, while fully smelted fish tend to migrate downstream with little or no delay (Bradford and Schreck 1989, 1990; Snelling et al. 1991; Snelling and Schreck 1992J. In addition to probably reducing adult return rates, longer stream residence is undesirable because it could increase opportunities for hatchery fish to prey on or compete with natu- rally reproduced fish, particularly if the hatchery fish are larger. The cause of incomplete smoltification of hatchery fish is not well understood but is believed to be related to environmental factors that differ from stream conditions, such as the lack of natural seasonal oscillations in water temperature if the hatchery water supply is a well (Shrimpton and Randall 19921. Research in an adaptive-manage

OCR for page 302
HATCHERIES 313 ment context (Walters 1986) is needed to design better hatchery rearing environ- ments so that smoltification problems are prevented. Ecological Problems Hatchery programs have given virtually no attention to the ecological con- text into which fish are released. Yet some ecological factors, such as carrying capacity of proximate and distant stream environments for various juvenile life stages and density-dependent interactions within and between species, can exert substantial control over the fate and impacts of released fish. The ability of hatchery fish to survive and be integrated into natural habitats without imposing unnatural changes on wild populations of salmon and other species depends on the numbers and sizes of fish released, their physiological state, their behavior and health status, and the locations and timing of release. Another ecological effect to be concerned about is that the use of hatcheries leads to a lack of salmon carcasses in the rivers. In natural situations, the car- casses of the spawning salmon that die provide important nutrients to the river ecosystems (Cederholm et al. 1989, Kline et al. 1990, Kline et al. 1993), includ- ing such terrestrial animals as bears and eagles. In addition, at many hatcheries managers prevent upstream migration of naturally reproducing adults although some agencies are beginning to recognize this and will, it is hoped, change it. The carcasses of fish that have been intercepted for breeding purposes are not returned to the river, and this also constitutes a substantial loss of nutrients in the ecosystem. The declines in adult carcasses might have reduced the nutritive capacity of stream environments and triggered low survival rates of naturally reproduced fry and juveniles because of lack of food that would normally result from carcass decomposition. That has been hypothesized as a partial explanation for the depressed natural reproduction of anadromous salmon in the middle fork of the Salmon River in spite of the high quality of the physical habitat in this wilderness area (Scientific Review Group 1993~. In addition, a dramatic reduc- tion in the number of reads reduces the number of disturbed reads from which released eggs drift downstream. Such drifting eggs are a source of food for resident trout, char, and sculpins. Finally, there is a major concern about the effect of hatcheries on the carry- ing capacity of the rivers and the oceans. In some cases, enormous numbers of fish are released by hatcheries; about 5.5 billion smelts of Pacific salmon of all species are released annually from hatcheries around the Pacific rim from Cali- fornia to Russia and Japan; one chinook hatchery in Canada (Robertson Creek) contributed 366,000 chinook to coastwide catches in 1991 alone (Riddell 1993a). Other observations adding to this concern include decreasing body size at matu- rity and increasing age at maturity of Japanese chum as total returns have in- creased, suggesting density-dependent rearing limitations in the oceanic environ- ment (Kaeriyama 1989 cited by Riddell 1993a); reduced catch of chinook salmon

OCR for page 302
314 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST in the Strait of Georgia, British Columbia when hatchery releases exceeded 8.3 million fish per year (Riddell 1993a); and the suggestion that interannual vari- ability in salmon abundance might increase as releases of hatchery fish increase (McCarl and Rettig 1983, Fagen and Smoker 19891. ROLES OF HATCHERIES IN THE FUTURE OF SALMON What roles do hatchery programs have in the rehabilitation of Pacific anadro- mous salmon? Answering that question implies agreement about the values human place on features of the environments in which Pacific salmon and steel- head occur. Hatcheries are only one human tool among many human and natural tools available to meet various objectives. Just as a carpenter would not invest in expensive tools before knowing what kind of construction was required, fishery managers should not reach for the hatchery tool or any other tool of human intervention before formulating explicit objectives for specific salmon popula- tions and their ecosystems. Natural aquatic biotic systems are intrinsically dynamic. To rehabilitate salmon populations successfully, human actions must accommodate natural types and rates of change. Emerging findings and ideas in ecology indicate that the long-term persistence and resiliency of natural systems is tied to their dynamic variability. Even if there were no human disruptions of habitat, the abundance of Oregon coastal coho, for example, would still be influenced by natural fluctua- tions in ocean productivity, with the entire period of fluctuation from low to high productivity lasting around 40 years (reviewed by Lawson 19931. Any human institutions involved in rehabilitation of coastal Oregon coho salmon, whether or not hatchery programs are a component, should take such variations into account. Thus, a period of many years probably around 40 is needed for completing a single cycle of adaptive management (Lawson 19931. It is important also not to be blinded by a "salmonocentric" perspective. Although it is critically important to address the needs of all life-history stages of salmon when designing rehabilitation strategies, too narrow a focus on salmon might prevent the attainment of a goal, even if the goal itself is focused on salmon. Rehabilitation strategies designed for freshwater life stages of anadro- mous salmon must include conservation and, when needed, rehabilitation of im- portant ecosystem linkages. Examples of linkages that might be essential for rehabilitation and long-term sustainability of salmon are food-web interactions among salmon, their predators, and their prey; nutrient cycles, such as the contri- bution of nutrients by riparian vegetation of a stream; and the influence of diver- sity in physical structures, such as woody debris, on natural functioning of an entire stream ecosystem and thus on the quality of salmon habitats (chapters 7 and 81. If there is no societal commitment to prevent and reverse human-induced damage to freshwater habitats, constraints arising from damaged ecosystems are

OCR for page 302
HATCHERIES 315 likely to prevent ~ hatchery from increasing the abundance of adult returns. In the worst case, a hatchery used under such severe conditions could accelerate rather than reverse the decline of populations. Consider, for example, plans for a new hatchery program to rebuild depleted salmon populations in a river basin. Plans are often based on the expectation that more than 0.1% of all released smelts will survive to return as spawning adults (a 0.1% survival rate is the approximate breakeven point at which each adult spawned in the hatchery will be replaced by one adult return). Recently, serious concern has been raised that survival rates of smolts are too low within some habitat-altered basins between the upstream sites of release and the river mouths to ever reach the expected 0.1% return rate (e.g., Fast et al. 1991, Currens 19931. If the return rates were lower than 0.~%, the hatchery could "mine" remnant populations of natural spawners, perhaps leading to a decline in their numbers. Regarding natural large- scale controls, natural productivity cycles of the Pacific Ocean will control the maximum possible abundance of hatchery returns by forcing a low value during periods of low productivity and allowing increased values during periods of increased productivity. The role of hatcheries is affected by society's choice between continual high human inputs to salmon ecosystems and rehabilitation of their natural regenera- tive capacity. Although many people now agree on the general goal of rebuilding the abundance of anadromous salmon populations, current debate is vigorous about the best actions for reaching this goal. There are essentially two options. The first is substitution society commits an indefinite and relatively high amount of human input In an attempt to make up for natural regenerative processes that are damaged in salmon ecosystems. As soon as human societies start to attempt to replace natural processes, they are committing themselves to indefinite expen- ditures. This option therefore involves high costs to society in the form of money, goods, services, and institutional structures. Costs will increase as the ability of human actions to make up for damaged natural regenerative processes decreases and falls short of expectations. Increased control by humans does not protect against surprises in the behavior of a managed system, such as declines in salmon abundance. Such surprises remain possible because of natural phenom- ena (e.g., oceanic productivity cycles) at a spatial or temporal scale beyond the scale of human controls. The other option is rehabilitation of the natural regenerative capacity of salmon ecosystems (Christie et al. 19871. This option reduces costs to society but requires a long-term commitment to obtain positive results. It can also require revision of fishery-management goals to reflect more realistic expectations about issues, such as increases in fish abundance that are possible under current versus revised land use and fluctuations in catch levels due to large-scale environmental controls that cause variability in fish abundance. Because this option emphasizes repair of environmental functions and processes, including redirection of human

OCR for page 302
316 UPSTREAM: SALMON AND SOCIETYIN THE PACIFIC NORTHWEST technologies, it is more likely to rebuild salmon populations to a sustainable status. Choosing between those two broad options for human intervention requires a policy decision. Making such a decision in an informed and democratic manner is extremely difficult, and our nation has an uneven record in addressing imper- iled natural resources. Such a decision greatly influences what human interven- tions will be carried out including how different technologies will be used in the name of rebuilding anadromous Pacific salmon populations. Intervention with the hatchery tool will differ in magnitude between the two options. Major hatchery reforms are imperative with either option. In general, this committee favors the rehabilitation option as more likely to be successful over the long term. It is also more consistent with the spirit of several current laws, including the Endangered Species Act. Under the option of rehabilitation, the role of hatcheries would be much more limited and refined than their historical role because rehabilitation of the natural regenerative capacity of an ecosystem requires congruence of each hu- man intervention with natural structures and processes of genetics, evolution, and ecology. An important building block of natural structures is the pattern of genetic diversity between and within anadromous salmon populations, and an important building block of natural processes is the evolution of salmon popula- tions. Therefore, rehabilitation implies a genetic-conservation goal, as described below. Hatcheries in the Rehabilitation Option Two guiding principles should be applied to all uses of hatcheries in light of the goal of rehabilitation. First, a hatchery program should be only one compo- nent of a comprehensive rehabilitation strategy designed to remove or substan- tially reduce the human-induced causes of decline of anadromous-salmon popu- lations. Because human causes of decline differ across the regions of the Pacific Northwest, such comprehensive strategies should be developed by region (see Box 12-31. The second guiding principle is that all hatchery programs should adopt the genetic-conservation goal of maintaining the genetic resources that exist in natu- rally spawning and hatchery populations. These remaining genetic resources are the building blocks of the natural "shifting variability system" (Carson 1983) needed for long-term evolution and persistence of salmon populations. All agen- cies involved in management or rehabilitation of anadromous salmon must rec- ognize that achievement of population-rebuilding goals will be jeopardized with- out concurrent adoption of a genetic-conservation goal (Riggs 1990, Gharrett and Smoker 1993, Riddell 1993b, Allendorf and Waples in press). As one positive step in this direction, the Northwest Power Planning Council recently revised its 1987 goal of doubling returns in the Columbia River basin to include prevention

OCR for page 302
HATCHERIES

OCR for page 302
318 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST of "further loss of biological diversity [in] anadromous and resident fish popula- tions" (;~PPC 1992a: 181. One way that hatcheries could contribute to a comprehensive rehabilitation program is outlined below. A second scenario-use of hatcheries for catch augmentation is not a method of biological rehabilitation, but using hatcheries in this way, while monitoring their performance in a sensibly comprehensive adaptive-management protocol, could reduce or avoid further biological damage. Te~nporary Hatcheries A temporary hatchery can be used to prevent extinction of a severely de- pleted population or rebuild a depleted population to self-sustaining status while the human causes of its decline are being rehabilitated to the extent feasible (e.g., while freshwater habitat damage is being repaired). Such temporary use supports rehabilitation in two ways. First, human costs of control are reduced because hatchery expenses and hatchery-associated genetic and ecological risks can be terminated when the populations and their habitats are restored so that they can sustain themselves. Second, re-establishment of a self-sustaining anadromous salmon population is a good indicator of improved natural functioning of the overall watershed ecosystem and increases confidence that there will be long- term persistence of various goods and services that humans desire from salmon ecosystems. For example, re-establishment of self-sustaining false trout popula- tions has been treated as such an ecosystem indicator In the Great Lakes (Edwards et al. 1990~. The hatchery facilities required for this scenario, if they are tempo- rary, would be less expensive than current, permanent hatcheries. Temporary hatcheries could be extended in principle if habitat damage or loss cannot be reversed in a reasonable period and people are not prepared to give up the salmon populations associated with the habitats. To be compatible with the goal of rehabilitation, this extension requires that negative impacts of hatch- ery fish on other naturally spawning populations be kept negligible. To be truly compatible with the rehabilitation goal, this type of longer-term holding action should be considered only as a complement to comprehensive efforts to rehabili- tate damaged habitat. If a temporary hatchery is extended as a substitute for serious efforts to reverse initial causes of decline (as was the case in historical mitigation programs), decision-makers will be accepting a future of indefinite and high human input; such a future should be accepted consciously and openly, with full disclosure to stakeholders and thorough weighing of costs and benefits. For instance, hatcheries planned for the Yakima basin (Fast et al. 1991) fit this scenario only if plans also include major efforts to rehabilitate freshwater habitat within the Yakima and its tributaries; recent concerns about high -~n-basin mortal- ity of released smelts underscore the importance of this point. Otherwise, hatch

OCR for page 302
HATCHERIES 319 cries in the Yakima will constitute implementation of the goal of indefinite and high human input. Catch-Augmentation Hatchery The decline of Pacific salmon constitutes not only a decline in the biological productive capacity of the Pacific Northwest, but also a decline in human fishing activity. Other human activities have precluded the historically abundant tradi- tional, recreational, and commercial catch of salmon. To restore or rehabilitate those human activities is a valid goal, and it might involve use of a long-term hatchery program to augment catch of some populations above levels that are now supportable by the quantity and quality of available freshwater habitat. This is perhaps the most controversial hatchery scenario in the eyes of many people who advocate environmental protection and genetic conservation. But in the Columbia River basin, some American Indian tribes are concerned that even if upriver populations are successfully rehabilitated to self-sustainability, numbers of returning adults will be so low without augmentation of the runs as to allow little or no catch in treaty-protected freshwater fisheries. To prevent genetic and ecological risks to long-term sustainability of nonaugmented populations and interacting species, this scenario requires feasible methods of separating hatchery fish from naturally spawning fish in freshwater habitats and selective catch of hatchery fish in the fishery (see Chapter 111. These conditions will be hard to meet in many cases and would require changes in fishing methods and manage- ment. Such changes require targeted research to develop simpler techniques for identification of hatchery fish and testing of alternative fishing methods for selec- tive catch of hatchery fish, as proposed in Section 5 of the Strategy for Salmon (NPPC 1992b). Regardless of the choice of overall goal (rehabilitation versus high human input) and subordinate hatchery objectives, major hatchery reforms are needed. They are outlined in the recommendations at the end of this chapter. CONCLUSIONS Despite some successes, hatchery programs have been partly or entirely responsible for detrimental effects on some wild runs of salmon. The loest- documented detrimental effects are loss of wild populations because hatchery fish swamp a mixed fishery, which encourages fishery managers to set exploita- tion rates that cause overfishing of natural populations, and loss of natural pat- terns of genetic variability between and within populations. There have been virtually no efforts to test for ecological effects of hatchery programs although scientific knowledge about salmon ecology strongly suggests that detrimental effects are possible. Hatchery use has not favored conservation of biological diversity. As a

OCR for page 302
320 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST result, institutional arrangements in management agencies have not encouraged collaboration between hatchery managers and fisheries and land managers to attempt to protect or enhance the natural functioning of watershed ecosystems, and that natural functioning has deteriorated. Hatcheries can be useful as part of an integrated, comprehensive approach to restoring sustainable runs of salmon, but they are not a panacea and should be used only sparingly and thoughtfully. The goals, specific objectives, and methods of past hatchery programs were not critically reviewed for scientific validity and practical feasibility. That led to the unrealistic expectation that releases of fish from hatcheries would always lead to increases in salmon abundance. In some cases, there also was the expectation that hatcheries would stabilize fluctuations in salmon abundance. By raising unrealistic expectations (Gale 1987), hatchery programs have discour- aged societal acceptance and understanding of variability in salmon abundance as a natural, complex phenomenon influenced by oceanic and other factors that are well beyond human control. The need to make societal expectations more realistic is heightened by recent oceanographic research that suggested that former high survival rates of hatchery fish were influenced more by ocean conditions than by hatchery activities themselves (see Chapter 21. Overreliance on hatcheries has also discouraged development of institu- tional arrangements and behaviors that would accommodate natural large- scale fluctuations in salmon abundance. For instance, heavy emphasis on hatcheries as the means of mitigating the effects of dams and overfishing has drained financial and other social resources that could be available for redressing "proximal" human causes of decline (as would be emphasized if the goal were rehabilitation). Hatchery programs have lacked proper monitoring and evaluation (i.e., there has been no adaptive management). The typical measure of success of a hatchery program has been the number of fish released rather than the number of returning adults, much less the health of nearby wild populations. In addition, cumulative effects of all the hatchery programs in a given region have not been evaluated. Potential effects on the genetic diversity and ecology of wild popula- tions have been ignored, although this is beginning to change. For example, a new Integrated Hatchery Operations Team (IHOT) in the Columbia River basin is an interagency effort involving 11 fisheries comanagement entities and seven cooperating entities. As a promising first step towards improved coordination among and monitoring of different hatchery programs, the IHOT recently issued and the Northwest Power Planning Council approved five general policies. They covered hatchery coordination, hatchery performance standards, fish health, eco- logical interactions, and genetics. There is also an implementation plan that includes independent audits of all 90 anadromous salmon hatchery facilities in the basin (Integrated Hatchery Operations Team 19951. There has also been

OCR for page 302
HATCHERIES 321 insufficient attention to potential effects on long-term sustainability of both hatch- ery-propagated and naturally reproducing populations. RECOMMENDATIONS The approach to hatchery operations should be changed in accordance with the goal of rehabilitation and the ecological and genetic ideas that inform that goal. Whenever hatcheries are used, great care should be taken to minimize their known and potential adverse effects on genetic structure of stocks and on the ecological capacities of streams and the ocean. For example, it is important to prevent blocking wild fish from migrating upstream of hatcheries because this is likely to compromise recovery and sustainability of wild popula- tions. In streams where only hatchery fish are found, managers should test the assumption that blockage is necessary to avoid disease incidence in hatcheries; this should be done by conducting different adaptive management experiments at different hatcheries that use this practice. More distant, upstream wild popula- tions should be conserved even if their presence complicates efforts to keep hatchery fish separated from wild fish. The aim of temporary hatcheries should be to assist recovery and opportunity for genetic expression of wild populations, not to maximize fishing opportunities. Augmentation hatcheries do aim to in- crease fishing opportunities, but they must do so in ways that will not jeopardize the genetic makeup and long-term persistence of nonaugmented populations. Adoption of this recommendation will almost certainly result in a significant reduction in total output of salmon hatcheries. Some ways of implementing this recommendation follow. The term "supplementation" as a goal of hatchery programs should be abandoned. The ambiguity of this term has generated confusion about appropri- ate roles of hatcheries, as illustrated by its many and often incompatible defini- tions in the published and "gray" literature. Instead, precise thinking and terms are needed to define and select ecosystem-level goals (e.g., rehabilitation of natural regenerative processes versus high human controls that will set the con- text for any proposed hatchery program. Goals for each hatchery program are needed so that they are compatible with the ecosystem-level goal. Given the ecosystem-level goal of rehabilitation, a compatible hatchery goal might be to assist rehabilitation of a given population to self-sustainability at the carrying . ~ . . ~ . capacity of its suite ot environments. Hatcheries should be dismantled, revised, or reprogrammed if they inter- fere with a comprehensive rehabilitation strategy designed to rebuild wild popu- lations of anadromous salmon to sustainability. It makes the most sense to test the ability of hatcheries to rehabilitate populations whose natural regenerative potential is severely constrained by both short-term and long-term limitations on rehabilitation of freshwater-habitats. Hatcheries should be excluded from re .

OCR for page 302
322 UPSTREAM: SALMON AND SOCIETY IN THE PACIFIC NORTHWEST "ions where the prognosis for freshwater habitat rehabilitation is much higher, as is the case in many watersheds of the Oregon coast (for example Cummins Creek on the central Oregon coast. This recommendation would allow continuation of, for example, the hatchery-supported fishery on chinook at Willamette Falls be- cause it is disrupting neither wild populations or rehabilitation efforts. Hatcheries should be rigorously auditedfor their ability to prevent demo- graphic, genetic, fish-health, behavioral, physiological, and ecological prob- lems. Any hatchery that "mines" broodstock from wild Natural) spawning popu- lations should be a candidate for immediate closure or conversion to research. Diseased broodstock should be rigorously culled to minimize disease in progeny. All hatchery programs should adopt a genetic-conservation goal of main- taining genetic diversity that exists between and within hatchery and natu- rally spawning populations. All agencies involved in management of anadro- mous salmon should recognize that achievement of population-rebuilding goals will be jeopardized without concurrent adoption of a genetic-conservation goal. Intentional artificial selection should be discouraged because of four major concerns. First, intentional selection might reduce within-population ge- netic variation, which could be decreased already as a result of previously im- posed artificial selection, hatchery founder effects, and mating practices. Sec- ond, our ability to accomplish the desired selection is uncertain, given the difficulty of determining the intensity of selective forces, the traits on which they are acting, and the uncertainty about how natural selection can reinforce or op- pose the imposed artificial selection. Third, counterselection for a given pheno- typic distribution, such as later run timing, is highly unlikely to accomplish the goal of re-establishing the original genetic structure. Fourth, our understanding of natural populations is insufficient to predict which phenotypic distributions and underlying genotypes are necessary for optimal fitness over the long term in the environments of anadromous salmon particularly because of their variability and different degrees of human-induced alterations. Genetic and ecological guidelines, based on the most up-to-date informa- tion base, are needed for all aspects of hatchery operations. Scientific peer review and appropriate revision of draft guidelines must occur before they are implemented and evaluated. Incentives are needed for uniform and coordinated application of scientifically sound genetic guidelines in anadromous salmon hatcheries in the five Pacific Northwest regions. Hatchery programs should avoid intentional transplantation of fish and unnatural patterns of straying by adult returns. This is necessary to prevent disruption of differences in evolved adaptations that exist among populations indigenous to different watersheds. . All hatchery fish should receive identifiable marks. Visible marks, such

OCR for page 302
HATCHERIES 323 as clipped fins, have advantages, but other methods, such as passive integrated transponder (PIT) tags and coded-wire tags, are also useful. Marking hatchery fish is important so that managers can distinguish between hatchery and wild runs. Implementation of this recommendation requires careful consideration of possible increases in mortality caused by different marking methods. Support is needed for research in and development of biological and other marks that will minimize handling-related mortality. Research is also needed to find selectively neutral genetic marks that are peculiar to hatchery fish and are clearly inherited by their naturally reproduced descendants but do not influence the fitness of individuals. Recent advances in application of molecular genetic markers to hatchery-bred fish suggest that it is feasible to develop such marks (Doyle et al. 1995). Decision-making about uses of hatcheries should occur within the con- text of fully implemented adaptive-management programs that focus on watershed management, not just on the fish themselves. Coordination should be improved among all hatcheries in release timing, scale of releases, operating practices, and monitoring and evaluation of individual and cumulative hatchery impacts, including a coastwide database on hatchery-wild fish proportions and numbers. Given the differences among the five regions in the status of environ- mental conditions that affect Pacific salmon populations, a different suite of interventions under adaptive management should be assembled for each region, and each intervention in a region should be treated as an experiment. Hatcheries would be an experimental treatment within adaptive management in some re- gions but not in others.