U.S. Dept Commerce/NOAA/NMFS/NWFSC/Tech Memos
NOAA-NMFS-NWFSC TM-29: Estuarine and Ocean Survival of Northeastern Pacific Salmon


CRITERIA FOR EVALUATING THE SURVIVAL VALUE OF
ESTUARIES FOR SALMONIDS

C. D. Levings1 and D. Bouillon2

1Department of Fisheries and Oceans (DFO)
West Vancouver Laboratory
4160 Marine Drive
West Vancouver, BC, Canada V7V 1N6

2ESSA Technologies Ltd., Suite 300
1765 W 8th Avenue
Vancouver, BC, Canada V6J 5C6

It is well known that estuaries have been affected by human use of the coastal zone and in the Northeast Pacific between southwestern British Columbia and California most of these habitats are either degraded or threatened (Levings 1994a). In the life cycle of a salmon, the estuary is also the animal's last "address" where the natural landscape can be changed by the direct activities of industrial activity or urbanization such as filling and direct habitat change. After leaving the estuary, the juvenile salmon's address in the pelagic zone is susceptible to the biophysical attributes of water quality (temperature and salinity) but not to water quantity or direct loss of habitat. However, with the possible exception of greenhouse gases and their effect on global warming, pelagic habitats cannot be "managed." Since decisions about estuarine salmon habitat have significant societal impacts in addition to those dealing with salmonid productivity, it is important to verify the role of estuaries in salmon survival.

We describe a scheme developed by Fox (1991) which may help in the latter exercise, and in the process we briefly review some of the literature on salmon ecology. Fox recognized that concepts used by epidemiologists to test hypotheses on how diseases were spread could be applied to cause-effect relationships for pollution problems. By applying the null hypothesis, scientists are forced to consider how much information must be ignored to conclude that a causal relationship does not exist. In the present paper, we propose an extension to the issue of estuarine habitat management for maintenance of salmon production. Using seven criteria, the object of the scheme is to "draw together all the disparate threads of evidence and make them into a coherent whole, so that scientifically and socially defensible regulatory decisions can be made" (Fox 1991). The seven criteria are as follows: probability, time order or chronological relationship, strength of the association, specificity, consistency of the association upon replication, predictive performance, and coherence. Because the scheme is based on the functioning of natural systems, we have restricted our considerations to wild salmon since it is well known that hatchery reared fish show very different patterns of estuarine use relative to wild stocks (e.g., Myers 1978, Levings et al. 1986).

Probability

Rice (1993) used the kernel estimator methodology to determine cumulative probability ogives for the abundance of several fish species including Atlantic cod (Gadus morhua) and Atlantic salmon (Salmo salar). We have tested this nonparametric technique with beach-seine data for juvenile hatchery chinook salmon (Oncorhynchus tshawytscha) at the Campbell River estuary. We used salinity as a parameter and found the method shows some promise to give predictions about the salinity adaptations of young salmon (Fig. 1). The preliminary results showed that while there is 80% probability of catch per unit of effort (CPUE) of 15 at 0.5‰, the probability of this CPUE at a higher salinity (5‰) is only about 10%. This indicates that hatchery chinook were more abundant in the lower salinity habitats of the estuary, probably in response to osmoregulatory requirements.

Time Order or Chronological Relationship

Fry from ocean-type chinook, chum (O. keta), and certain sockeye (O. nerka) salmon stocks undergo a rearing migration from freshwater habitats in the upper estuary to brackish areas of intermediate salinity and finally into marine habitats as they complete smoltification (Levy and Northcote 1982, Levings et al. 1986, Birtwell et al. 1987). The residence time of the fish in particular salinity regimes is attuned to the species' development of osmoregulatory ability at certain sizes and ages. The movement through the estuary over the spring to summer period follows an adaptive chronology which bears out the survival value of estuarine residency.

Strength of the Association

It is a qualitative, but well-documented, observation that juvenile ocean-type chinook salmon appear to be strongly linked to estuarine habitats since they reside in these habitats for 1 to 2 months (e.g., Healey 1982, Levy and Northcote 1982). Many estuaries almost totally drain with falling tides and are characterized by strong currents (e.g., Campbell River estuary; > one m•s-1, Ages et al. 1990) and shallow brackish layers which a priori might lead to flushing the fry out of these habitats. However, they are adapted to remain and rear before smoltification and departure to the open sea.

Specificity

Specificity refers to the precision of the association between a cause (X) and an effect (Y). Does X lead only to Y (specificity of effect) or does only X lead to Y (specificity of cause) (Fox, 1991)? In our context the question could be reformulated to ask if the loss (X) of an estuary in the life history of a particular wild salmon stock leads to decreased survival (Y) (specific of effect). It is obvious that a very elaborate and detailed research program would be needed to identify the second part of the proposal, namely specificity of cause.

The initial observations of Reimers (1973) in Oregon estuaries concerning the relative survival of several chinook salmon life history types that showed differences in estuarine use led to several major research programs in estuaries of the Northeast Pacific. This question was asked in a series of habitat release experiments with juvenile chinook salmon at the Campbell River estuary, as described in Levings (1984), Macdonald et al. (1988), and Levings et al. (1989). Releases of specially marked fish were made to river, estuary, transition, and marine habitats in 1983, 1984, and 1985. About 35,000 fish were released into the four habitats each year. Results showed that survival, to catch and escapement, of the fish that transitted the estuary was higher than for those that did not (Table 1). A similar result was found with steelhead (O. mykiss) reared and released in the estuary of the Keogh River on Vancouver Island. Releases directly to ocean habitats also showed lower survival in the latter experiments. However, there were numerous complications to the interpretations of the results of both the chinook salmon experiments (Levings et al. 1989) and the steelhead releases (Ward and Slaney 1990), including size effects and differences in temperature and ocean conditions that the three brood years were exposed to.

Consistency of the Association upon Replication

Levings (1984) summarized data to show that densities of juvenile salmon in estuaries were within a relatively narrow range, which might suggest some consistency in the number of juvenile salmon that the estuaries could support. To further assess the use of this concept to ecological data on the importance of estuaries to juvenile salmon, we assembled data on instantaneous growth in weight for chinook and chum salmon in six estuaries off the Strait of Georgia (Fig. 2). The six estuaries might be considered replicates of the habitats and ecosystems we are concerned with. Instantaneous growth rates for chinook salmon ranged between 0.02% and 0.04% and for chum salmon between 0.01% and 0.03%. These data show that growth in the various estuaries did not vary randomly over the landscapes sampled. These findings suggest that these wild salmonids are adapted to estuarine rearing and imply that growth was not being impaired, relative to comparisons possible with this particular data set.

Predictive Performance

In a quantitative or statistical sense (as per Fox 1991), predictive performance may be interpreted as the size of effect produced by the presumptive cause. For example, Morley et al. (1996) used response surface analysis to investigate relationships between size and time of release of hatchery reared chinook salmon. Experiments were conducted in 1982 and 1983, with releases in May, June, and July. Results showed that fish released in May had best survival and that time effects were apparently more important than those for size.

Coherence

Sibert and Kask (1978) analyzed the correlations between the diet of juvenile chinook salmon at the Nanaimo, Cowichan, Fraser, and Campbell River estuaries, as measured by conventional taxonomic methods. They found significant correlations in a minority of cases, indicating low coherence in fish diets at the four estuaries. It is clear that factors such as prey size, color, behavior, and perhaps nutrition may be important factors influencing prey selection from the suite of available organisms (Levings 1994b, Higgs et al. 1995). However, coherence in food webs may also be assessed by methods such as radiotracer experiments or stable isotope concentrations at various trophic levels. Over the past two decades there have been numerous papers which show that juvenile salmonids are linked to the detritus-based food webs of estuaries (e.g., Sibert et al. 1977, Simenstad and Wissmar 1985, Levings 1994c). These reports support the idea that heterotrophy and detrital production are responsible for coherent energy flow patterns in Northeast Pacific estuaries.

Conclusions

Several experiments have investigated the causal relationship between estuaries and survival of juvenile salmonids since the 1983 Workshop (Levings 1984), but it is clear that much research has yet to be done in this particular habitat. The existing data, tested using a causal inference scheme, supports the rejection of the null hypothesis that estuaries are not important for salmon survival. Salmon ecologists from outside the Northeast Pacific, notably Thorpe (1994), agree with this conclusion based on evolutionary considerations. Thorpe concluded that wild pink (O. gorbuscha), chum, some sockeye, and ocean-type chinook salmon are adapted to rear in estuaries and thus their natural habitat characteristics have survival value relative to fresh water. Evidence for the significance of smolt rearing in estuaries by coho, most sockeye, stream-type chinook, masu (O. masou), and Atlantic salmon, cutthroat and brown trout (Salmo trutta), and steelhead is more equivocal.

One of the problems of resolving the question of the importance of estuaries for salmon in the Northeast Pacific is the fact that most of the data are from south of 50°N latitude where stocks are dominated by hatchery fish and habitats are disrupted. Thus many of the data are not applicable to more northern areas where a major concerted effort must be maintained to preserve wild stocks and natural habitat. In some designated watershed-estuary ecosystems, production of salmon may have to be sacrificed to conservation if we are to appreciate and understand the role of estuaries in salmon ecology (Levings 1993).

Acknowledgments

Thanks are owed to Derek Nishimura for assistance with computing probability density functions. Some of the work reported herein was supported by the DFO Habitat Action Plan, Environmental Analysis component.

Citations

Ages, A. B., F. A. Dobbs, and C. D. McAllister. 1990. The salinity intrusion in the Campbell River estuary: Salinity, temperature and current observations, 1984-1986. Can. Data Rep. Hydrog. Ocean Sci. 83(1):1-243 and 83(2):1-191.

Birtwell, I. K., M. D. Nassichuk, and H. Buene. 1987. Underyearling sockeye salmon (Oncorhynchus nerka) in the estuary of the Fraser River. In H. D. L. Smith, L. Margolis and C. C. Wood (editors), Sockeye salmon (Oncorhynchus nerka) population biology and future management. Can. Spec. Publ. Fish. Aquat. Sci. 96:25-35.

Bouillon, D. 1996. A comparison of growth and bioenergetic consumption of juvenile salmon in several estuaries of the Strait of Georgia, 68 p. (Available from DFO, West Vancouver Laboratory, West Vancouver, BC, Canada.)

Fox, G. A. 1991. Practical causal inference for ecoepidemiologists. J. Toxicol. Environ. Health 33:359-373.

Healey, M. C. 1982. Juvenile Pacific salmon in estuaries: The life support system. In V. S. Kennedy (editor), Estuarine comparisons, p. 315-342. Academic Press, New York.

Higgs, D. A., J. S. Macdonald, C. D. Levings, and B. Dosanjh. 1995. Nutrition and feeding habits of Pacific salmon (Oncorhynchus spp.) in relation to life history stage. In R. Brett, W.C. Clarke, K. Groot, and L. Margolis (editors), Physiological ecology of Pacific salmon, p. 161 to 315. Univer. B.C. Press, Vancouver, B.C.

Levings, C. D. 1984. Commentary: Progress in attempts to test the null hypothesis that juvenile salmonids aren't dependent on estuaries. In W. G. Pearcy (editor), The influence of ocean conditions on the production of salmonids in the North Pacific, p. 287-296. Oreg. State Univ., Sea Grant Coll. Prog. ORESU-W-83-001.

Levings, C. D. 1993. Requirements for genetic data on adaptations to environment and habitats of salmonids. In J. G. Cloud and G. H. Thorgaard (editors), Proceedings of NATO Advanced Study Institute: Genetic Conservation of Salmonid Fishes, Moscow, Idaho, June 23-29, 1992, p. 49-65. Plenum Press, New York.

Levings, C. D. 1994a. Science and management needed to maintain salmon production in estuaries of the Northeast Pacific. In K. Dyer and R. J. Orth (editors), Proceedings Joint Symposium of Estuarine Research Federation and Estuarine and Coastal Sciences Association, September 14-18 1992, Plymouth, England, p. 417-421. Olsen and Olsen, Denmark.

Levings, C. D. 1994b. Feeding behaviour of juvenile salmon and significance of habitat during estuary and early sea phase. Nordic J. Freshw. Res. 69:7-16.

Levings, C. D. 1994c. Life on the edge: Structural and functional aspects of chinook and coho rearing habitats on the margins of the lower Fraser River. In Salmon ecosystem restoration: myth and reality, Proceedings of 1994 Northeast Pacific Chinook and Coho Salmon Workshop, Oregon Chapter, American Fisheries Society, Corvallis, OR, p. 139-147.

Levings, C. D., McAllister, C. D., and B. D. Chang. 1986. Differential use of the Campbell River estuary, British Columbia, by wild and hatchery reared juvenile chinook salmon (Oncorhynchus tshawytscha). Can. J. Fish. Aquat. Sci. 43:1386-1397.

Levings, C. D., C. D. McAllister, J. S. Macdonald, T. J. Brown, M. S. Kotyk, and B. A. Kask. 1989. Chinook salmon (Oncorhynchus tshawytscha) and estuarine habitat: A transfer experiment can help evaluate estuary dependency. Can. Spec. Publ. Fish. Aquat. Sci. 105:116-122.

Levy, D. A., and T. G. Northcote. 1982. Juvenile salmon residency in a marsh area of the Fraser River estuary. Can. J. Fish. Aquat. Sci. 39:270-276.

Macdonald, J. S., C. D. Levings, C. D. McAllister, U. H. M. Fagerlund, and J. R. McBride. 1988. A field experiment to test the importance of estuaries for chinook salmon (Oncorhynchus tshawytscha) survival: short term results. Can. J. Fish. Aquat. Sci. 45:1366-1377.

Morley, R. B., A. Y. Federenko, H. T. Bilton, and S. J. Lehmann. 1966. The effects of time and size at release on returns at maturity of chinook salmon from Quinsam Hatchery, B.C., 1982 and 1983 releases. Can. Tech. Rep. Fish. Aquat. Sci. No. 2105. 62 p.

Myers, K. W. 1978. Comparative analysis of stomach contents of cultured and wild juvenile salmonids in Yaquina Bay, Oregon. In S. J. Lipovsky and C. A. Simenstad (editors), Fish food habits studies, p. 155-162. Proceedings of the Second Pacific Northwest Technical Workshop, October 10-13, 1978, Washington Sea Grant, Univ. Washington, Seattle.

Reimers, P. E. 1973. The length of residence of juvenile fall chinook salmon in Sixes River, Oregon. Res. Briefs Fish. Comm. Oregon 4(2):1-43.

Rice, J. R. 1993. Forecasting abundance from habitat measures using non-parametric density estimate methods. Can. J. Fish. Aquat. Sci. 50:1690-1698.

Sibert, J. R., T. J. Brown, M. C. Healey, B. A. Kask, and R. J. Naiman. 1977. Detritus-based food webs: exploitation by juvenile chum salmon (Oncorhynchus keta). Science 196:649-650.

Sibert, J., and B. A. Kask. 1978. Do fish have diets? In B. G. Shepherd and R. M. J. Ginetz (rapporteurs), Proceedings of the 1977 Northeast Pacific Chinook and Coho Salmon Workshop, p. 48-56. Fisheries and Marine Service Tech. Rep. 759.

Simenstad, C. A., and R. C. Wissmar. 1985. d 13C evidence of the origins of organic carbon in estuarine and nearshore food webs. Marine Ecology (Progress Series) 22:141-152.

Thorpe, J. E. 1994. Salmonid fishes and the estuarine environment. Estuaries 17(1A):76-93.

Ward, B. R., and P. A. Slaney. 1990. Returns of pen-reared steelhead from riverine, estuarine, and marine releases. Trans. Am. Fish. Soc. 119:492-299.



Table 1. Total number of tags recovered from experimental releases of juvenile chinook salmon into four habitats at Campbell River estuary (unpublished DFO data subject to verification, compiled to 1991).


Habitat/Year River Estuary Transition Marine Total

1983 136 161 45 25 367
           
1984 83 102 33 27 245
           
1985 150 137 140 137 564
           
Total 369 400 218 189 1,176



Go to next paper

Table of Contents