Reviews, narrative and other documents for proposal

201003300: Estimate the Relative Reproductive Success of Hatchery and Natural Origin Steelhead in the Methow River Basin

(View full proposal and assessments at

ISRP final recommendation: Meets Scientific criteria? Response requested

The project proponents should submit a more developed proposal in the RME categorical review. A point-by-point response to the ISRP's concerns should accompany the proposal. The ISRP preliminary review requested a revised narrative with clarified and expanded information before making a final recommendation. The ISRP anticipated that the revised narrative would be accompanied by a “point-by-point” response to aid the ISRP in reaching a final recommendation. The ISRP acknowledges the receipt of a revised narrative, but the issues of concern to the ISRP are not satisfactorily addressed, nor sufficiently clarified. This leaves ambiguity with incomplete resolution of the issues. Nonetheless, the ISRP believes this project is worth pursuing. The first issue the ISRP raised was the contrast used to estimate Relative Reproductive Success (RSS). In the original narrative, on page 13 and 14 the proponent states, “We will compare four categories of parent-pairs: HfHm, HfNm, NfHm, and NfNm and calculate RRS at three different life stages (parr, smolt, adult).” On page 21 Task 4 Data Analysis Relative reproductive success analysis: proponent states “We will compare four categories of parent-pairs:...., using equation 14 from Araki and Blouin 2005.” The ISRP interpreted the statement in the proposal to indicate that the progeny that would be used in the analysis of relative reproductive success would be those in which both parents were identified. It also gave the ISRP the impression that somehow there was going to be a simultaneous analysis of relative production from four types of mating. The Araki papers, and others, use parentage assignments from progeny where only one of the two parents has been identified, in addition to those progeny where both parents are identified. Using only triplets – both parents and offspring – reduces the data pool used to estimate RRS. The ISRP is requesting clarification on this point and suggesting that a less restrictive inclusion of progeny for the RRS contrast will improve the power of the analysis Regarding the specific RRS contrast, the proponent indicates they will use Equation 14 from Araki and Blouin (2005) to produce an unbiased estimate of relative reproductive success between groups. Once again the ISRP is left with the impression that 4 groups will be contrasted in the F1 generation and 16 in the F2 (carryover analysis). Equation 14 in Araki and Blouin (2005) is for a contrast of two groups. Araki applied this formula in Araki 2007 and in Araki 2009. In both of those publications the application was to paired comparisons of fish with different ancestry, calculated separately for each sex. In Araki (2007) they compared hatchery relative to wild production for male parents and for female parents without respect to the other parent. In each contrast they were comparing two groups: wild males versus hatchery males and then wild females versus hatchery females. In Table 5 in the 2007 publication they extended the analysis to contrasts of mating types, but again restricted the contrast to two groups. In this case a parental genotype (wild versus hatchery) mated to hatchery fish is contrasted to that parental genotype mated to wild fish. In the 2009 carry-over paper they contrast two groups, arranged by sex, differing in the ancestry of a single parent (Table 1). The informative F2 parents, had both parents identified, but the F3 progeny that constituted the data pool are not limited to those where both parents were assigned by pedigree analysis. The ISRP concern is that the method of estimation may not be applicable to establishing RRS for multiple groups; that this would require both parents and offspring to be identified in a triplet restricting the number of useful data points; and that the methods would be less powerful in establishing differences. A second issue is the parr and smolt sample size. The proponents anticipate they will be able to capture nearly all the anadromous adults that will contribute to spawning and propose to systematically sample 1,250 parr and 750 smolts (we presume annually for each group of spawning parents). This raises questions for the ISRP of whether this is a sufficient number given genotyping errors, parentage assignment errors, and individuals that have no assigned parents (common in steelhead investigations). The ISRP believes that a thorough evaluation of the sample sizes required for detecting pre-defined (5, 10, 20%) differences in RSS using error rates derived from existing published reports should be presented in the proposal to justify the sampling plan. Third, the suggestion for AHA modeling, and executing the investigation as an evaluation of the AHA assumptions, was not considered. Since supplementation of Upper Columbia steelhead is being proposed in the Methow and Okanogan subbasins and will use AHA in developing the program goals, the ISRP would like to see this addition considered by the proponents. Finally, the carryover experiment may also be more difficult to interpret because the history of the stock under investigation is not well defined. In the case of the Hood River steelhead, the hatchery stocks were derived from wild fish with little history of hatchery steelhead introgression. In the case of the Methow River, the steelhead in the wild and in the hatchery may be at genetic equilibrium and highly influenced by domestication selection. Under these circumstances the interpretation of any carryover effect will not be unambiguous. This should be discussed in the proposal narrative. Araki, H., and M. S. Blouin. 2005. Unbiased estimation of relative reproductive success of different groups: evaluation and correction of bias caused by parentage assignment errors. Molecular Ecology 14:4097-4109. Araki, H., W. Ardren, E. Olsen, B. Cooper and M. Blouin. 2007. Reproductive success of captive-bred steelhead trout in the wild: evaluation of three hatchery programs in the Hood River. Conservation Biology 21:181-190. Araki, H., B. Cooper and M. Blouin 2009. Carry-over effect of captive breeding reduces reproductive fitness of wild-born descendants in the wild (Biology Letters doi:10.1098/rsbl.2009.0315)

from Apr 2010 ISRP 2010-10 report

Narrative and other documents

201003300 Methow Steelhead Relative Reproductive Success Study.doc (narrative)
Methow Steelhead Relative Reproductive Success Study 01-20-10.doc (narrative)

ISRP preliminary recommendation: Meets scientific criteria? Response requested

The study of relative reproductive success of hatchery and natural steelhead in the Twisp River proposed is needed. The ISRP believes investigation of natural production by spawning hatchery steelhead in the tributaries above Wells Dam is essential for understanding the status and viability of the natural population. The proposal included three primary objectives: 1. in a first generation compare the relative production from hatchery and natural fish spawning in the Twisp River, a tributary to the Methow River; 2. evaluate potential biological attributes of the fish and environmental attributes of the spawning site and time that might account for differences in the performance of hatchery and natural steelhead; and 3. in a second generation compare the success of natural spawning adults that had zero, one, or two hatchery-origin parents in the previous generation. The ISRP raises questions about the field and analytical methods in section 3 below. A response is requested in the form of a revised proposal narrative that elaborates on the analysis anticipated for each objective. This investigation also becomes a test of the AHA model. AHA should be run on this population (if not done already by the HSRG) and this project used to test the assumptions in AHA. The ISRP is interested in how the environment—tributary habitat capacity, interannual variation—might affect the outcome. Could different environmental conditions be added to the study? This would add a dimension to objective 2 - correlation analysis. 1. Technical Justification, Program Significance and Consistency, and Project Relationships The proponent proposes to examine Relative Reproductive Success (RRS) for Twisp River (Methow River subbasin) summer steelhead. The steelhead run is part of the upper Columbia River basin Evolutionarily Significant Unit (ESU) and is listed for Endangered Species Act (ESA) protections. This project is similar to ongoing RRS investigations in the Hood River, Oregon, that have provided evidence that multi-generation hatchery stocks of steelhead are less productive when spawning naturally than non-captive fish, that a single generation in the hatchery results in depressed performance in the wild, and that hatchery effects on natural production persist in wild-born individuals with hatchery-born parents. The objectives, rationale, and approach are clearly presented and suggest a project that will provide another data set for comparing hatchery and natural steelhead reproductive performance that will complement the Hood River investigations. Until initial evaluations of progeny production from natural and hatchery steelhead are completed it will not be known whether the Twisp River “case” is biologically similar to the Hood River “case.” In the Hood River many of the interesting results that have been published are based on comparing recently established hatchery stocks with natural fish. The hatchery stocks have been established from the local natural stock. In the Methow subbasin, the hatchery fish are a long-established (1969) composite stock with broodfish collected at Wells Dam and progeny historically scatter-planted throughout the Methow and Okanogan subbasins. Recently the juveniles released from the hatchery program have been hatchery x wild crosses. The proposal does not present information on the relationship of the natural and hatchery steelhead, but it is possible that the natural fish are descendents of wild-born hatchery fish. This possibility is important to consider when interpreting the results of the investigation. For example, hatchery- and natural-origin coho salmon in Minter Creek, Washington have indistinguishable reproductive performance in the natural stream, and this is attributed to 60 years of hatchery production with the majority of natural spawning by hatchery-origin adults (Ford et al. 2006). It is noteworthy that in the Minter Creek coho situation the production of smolts has decreased from levels in the 1940s and run- and spawn-timing are earlier. Analysis suggests that optimum run-time is later than the present timing (Ford et al. 2006). The important point is that likelihood of substantial past crossing of wild and hatchery fish will complicate using a difference in relative reproductive success between the hatchery- and natural-origin steelhead as a valid basis for drawing biological conclusions and useful management implications. Indeed, if the high proportion of hatchery-origin steelhead present in the past were reproductively successful at reasonable rates, smolt yields would have been much higher. Even with these caveats, the investigation is important and will contribute to our understanding of the population status of upper Columbia River steelhead. 2. Project History and Results This is a new project. Proponents indicate that methods to collect tissue samples, genotype fish, and operate the Twisp weir and juvenile trap have been tested. 3. Objectives, Work Elements, and Methods The general outline of the proposed investigation employs established protocols for parentage assignment and assessment of relative reproductive performance of different categories of individuals. Most, or all, of the potential parents will be captured and genotyped; juveniles will be sampled, genotyped, and assigned to parents. The number of progeny produced by different categories of parents will be compared to establish their relative reproductive performance. The ISRP has several concerns about individual methods that need to be addressed before initiating the investigations. Reliance on rotary smolt traps for smolt capture may not provide sufficient sample size to confidently determine the relative reproductive success of wild versus hatchery recruitment to the smolt stage—the key response variable. A full smolt enumeration and sub-sample routine should be explored and employed if feasible. Sample size requirements to detect differences in reproductive performance should be established a priori. This should consider the power and minimum effect size that is likely to be detectable. For objective 1 and 3 the proponent outlines a comparison of production from parent pairs (4 for objective 1 and 16 for objective 3). In most investigations of RRS the contrast is among 4 categories – hatchery males and females and natural males and females. Additionally, the Hood River investigators have completed and published an evaluation of “carryover effects” identical to that proposed in objective 3 (Araki et al. 2009). This study was not listed in the literature citations. The ISRP urges that a compatible study design be employed in the Twisp, so this study can serve as a replication/comparison. The ISRP believes the proponents need to revisit the analysis design and ensure it is using contrasts compatible with other Pacific salmon and steelhead RRS investigations. It is not clear to the ISRP that the assumptions for testing random mating will be met. This should be addressed in a response. For objective 2 - determine the degree to which differences in fitness between hatchery and natural steelhead can be explained by measurable biological or life-history traits that differ between hatchery and natural fish the analytical approach to evaluating selection appears appropriate (using the methods from Lande and Arnold 1983), but the interpretation of whether the differences between hatchery and natural fish are genetic (from domestication selection) or from environmental effects of hatchery rearing is not clear. On page 8 the proponents conclude they will be able to determine not only if hatchery steelhead have lower relative reproductive success than natural steelhead, but also why. It is not evident that the design of the investigation can lead to interpretations of causation. In particular, on page 21 final paragraph the proponents state “If there are differences in relative reproductive success between hatchery- and natural-origin spawners, it is possible that these differences are more a function of biological factors that are correlated with the origin of the spawners rather than any direct hatchery effect.” It is not clear to the ISRP what is intended by this distinction – which is the genetic effect, which is the environmental effect? And how will the design not confound these effects? This should be addressed in a response. For objective 3, if the natural-origin steelhead in the Twisp are functionally the wild-born descendents of Wells hatchery steelhead, and the two components (hatchery and wild) are at genetic equilibrium because of past interbreeding, then one generation of wild parents may not yield an important production distinction between categories (wild with hatchery parents versus wild with wild parents). Both categories could have low productivity. The ISRP is under the impression that a longer term investigation of re-adaptation is underway with coho salmon at Minter Creek. The status of that investigation and approach should be confirmed. It would be worthwhile to have a longer term investigation of the re-adaptation of steelhead. This component should be added to the plan. Araki, H., B. Cooper and M. Blouin 2009. Carry-over effect of captive breeding reduces reproductive fitness of wild-born descendants in the wild (Biology Letters doi:10.1098/rsbl.2009.0315) Ford, MJ, H. Fuss, B. Boelts, E. LaHood, J. Hard, J. Miller. 2006. Changes in run timing and natural smolt production in a naturally spawning coho salmon (Oncorhynchus kisutch) stream after 60 years of intensive hatchery supplementation. Canadian Journal of Fisheries and Aquatic Sciences 63:2343-2355.

from Feb 2010 ISRP 2010-7 report