Entrix, Inc.

January 2003

Review of Hedgecock et al. 2002 Documenting Biodiversity of Coastal Salmon (Oncorhynchus spp.) in Northern California

By RUTH SUNDERMEYER

Summary

This review of Hedgecock et al. 2002 includes comments on the report's scientific merit, methodology, fulfillment of contract terms and conditions, and literature review.

Hedgecock et al. 2002 provides extremely useful information for resource managers as they make decisions in resource and recovery planning. Furthermore, the study is written in such a way that should make it easier for managers and biologists who are not geneticists to not only understand the results of the study, but to understand how this information can and can not be used. It will be important for people with local knowledge within their watershed to help interpret the genetic information presented so that informed management decisions can be made. One of the strongest features of the study is the statistical tools developed to address factors that are important to consider when interpreting genetic results, in particular, correction of juvenile samples for sibling relationships. When working with threatened or endangered species, assumptions of hypothetical models or statistical analysis packages can be difficult to meet, and the researchers on this study have made important contributions to the resolution of some of these issues.

Fullfillment of Tasks in the Contract

The major objective of the contract, to describe the genetic diversity of the coho salmon populations along the central and northern coast of California, was fullfilled. Specific tasks that were fullfilled included I) to determine relatedness in samples comprised of juveniles, 2) to determine temporal genetic variation among year classes and 5) to relate the genetic diversity of California coho populations to environmental and biological factors being measured in the sampling process. Task 3) to estimate genetic divergence among and effective population sizes of spawning runs, was mostly completed, but effective population sizes of spawning runs were not determined for all populations for which this information might be useful. Task 4) to determine genetic change between historical and extant coho populations to assess influence of hatchery plantings and reductions in abundance, is a difficult task that may never be completely resolved. Furthermore, an historical collection free of hatchery influence would be difficult to compile, given the extensive and incompletely documented stocking history.

Tasks in the Chinook portion of the contract were fulfilled.

Specific Comments to Text

Page 3. Summary. end of 2nd paragraph.

"The congruence of genetic and geographic distance is surprising in light of the history of coho stock transfers within California and between California and other Pacific Coast states."

However,

"Stock transfers appear to have left no genetic mark on extant populations. Alternatively, or in addition to stock transfers, the diversifying effect of genetic drift within the relict coho populations of California may be keeping pace with whatever homogenization has been or is being affected by hatchery practices."

These two sentences sum up to 'We don't know the effect of stock transfers.' The analysis presented in this report does an excellent job of outlining existing population structure, but without an accurate, historical baseline and a more through review of past hatchery practices, analysis of effects of stock transfer remains speculative. It is hoped that resource managers familiar with local stocking histories and local ecological factors will be able to apply their knowledge to help refine the interpretation of the fmdings presented in this study.

Page 4 Last paragraph.

This is a good summary to help show how F ST can be interpreted. It might be helpful to add a point about local adaptation as a potential factor for diversity between populations. An artificially induced deficiency of heterozygotes is a different kind of management problem than genetic structure and amount of genetic diversity influenced by natural selection, which may have a direct effect on fitness of the population. Furthermore, population structure that is based on natural selection can change if the adaptive landscape changes, such as large changes in weather patterns associated with extremely wet years and El Nino events, and may also cause changes in H-W and FST. Although this is a difficult question to answer, conceptually it may be important to when trying to make management decisions based on genetic analysis. A through understanding of historical hatchery practices and local ecological data are essential, especially when one is trying to identifY remnant "natural" populations for protection.

It can be difficult to know at what scale management of differentiated population units should occur - one can micromanage populations on too fine a scale within a watershed, or manage on a scale that does not protect local adaptations.

Page 14.

"Wahlund effect in the original sample would be evidenced by non-significant departures from H- W within subsamples, but significant F ST among subsamples."

Are there no other possibilities besides an artificially induced Wahlund effect? Subsamples collected from different sites within a watershed - couldn't they be in H-W equilibrium and have significant F ST if admixture occurred with a change in the adaptive landscape other than artificial stocking?

It should be noted that the analysis on subgroups within samples based on information such as year class, size, and geographic information is information that is not routinely presented in studies of this kind, but can be very useful to interpret the population structures observed. It may also help identifY admixture due to hatchery influences.

Page 19 Waddell and Scott Creeks

Waddell and Scott creeks provide an interesting case study because there is a long-term juvenile abundance data set for Scott and Waddell creeks spanning almost a decade, as well as relevant ecological information and stocking history that help explain population trends (Smith 2002).

WADY99: "Samples originatingfrom RM 4.7 were heterogeneous to both RM 3.1 and 3.9 and were removed (WADY99up)."

SCY99: Removal of Upper Fork and RM 4.9 samples (SCY99up) yields a homogenous population (SCY99Iow) with a substantial number of siblings.

Big Creek hatchery (from the Scott Creek watershed) coho fry were planted in lower Waddell in 1996, (perhaps progeny of SCA95), but not in upper Waddell. Juvenile fish from the 1996 year class in Waddell would have spawned the 1999 year class, including WADY99up and

WADY991ow, while juvenile fish from the strong Scott Creek 1996 year class would have contributed to the SCA98 and SCY99 samples. This hatchery stocking might help explain why WADY9910w is more closely related to SCA9798 (FST 0.019), SCA95 (PST 0.014) and SCY9910w (FST 0.017) but W ADY99up is more distant to these same populations (FST 0.076, 0.074, and 0.041 respectively). It also appears that SCY99up is more closely related to SCY9910w (FST 0.024) than to WADY99up (FST 0.120).

The survival of naturally spawned juvenile fish is more certain in some sections of Waddell and Scott creeks than in others depending on winter storms and summer flow (Smith 2002) and it would be interesting to see if the population structure documented in this study persists. The 1993, 1996, 1999 and 2002 year class set (a set that spans the three year life history of coho) is currently the only viable set in Waddell (Smith 1999, 2002), which makes the W ADY99 populations (and 2002 juveniles) important. Smith 2002 describes the status of coho in several streams in the south of San Francisco group in detail.

Page 36 Temporal Variation

Although temporal samples are available for seven sites, additional information is needed to interpret some of these data.

In Scott Creek, presumably the SCA95 population contributed to the SCA98 population. In 1993, a strong juvenile year class was documented (Smith 2002) and precocial females from that year were raised in the hatchery to supplement weak year classes. This suggests that the Scott Creek populations examined in this study may be a better case study for artificially induced year-to-year variation than for natural temporal variation. Furthermore, the persistence of weak year classes could contribute to the genetic variability between year classes, and tests for inbreeding would be helpful.

It would be helpful to examine hatchery stocking and natural factors affecting population trends in other watersheds as well. For example, hatchery planting may contribute to homogeneity of samples between locations and years, such as is found in Lagunitas. The surprising heterogeneity between the Redwood Creek samples (RWMA97 and offspring RWMY98), as well as the fact that they are outliers is a mystery that draws curiosity.

Page 42 Discussion on Chinook results

"Chinook in the Russian River do appear to belong to a diverse set of coastal chinook populations."

Although it is useful to know that Russian River chinook are not closely related to Central Valley or the Eel River chinook, the relationship to other populations in this ESU can not be determined without comparison to additional data, such as hatchery populations from which extensive stocking in the Russian River has occurred. In any case, this may be a difficult question to answer, given the long history of stock transfers in this basin.

References

Smith, II 2002. Distribution and abundance of juvenile coho and steelhead in Gazos, Waddell and Scott creeks in 2002. Department of Biological Sciences, San Jose State University, San Jose, CA.

Smith, II 1999. Distribution and abundance of juvenile coho and steelhead in Gazos, Waddell and Scott creeks in 1999. Department of Biological Sciences, San Jose State University, San Jose, CA.