Comments to ODFW on analysis of steelhead status

5 November 1997

Mr. Mark Chilcote
Natural Production Program Leader
Oregon Department of Fish and Wildlife
P.O. Box 59
Portland, OR 97207

Dear Mr. Chilcote:

The Oregon Chapter of the American Fisheries Society (ORAFS) appreciates the opportunity to review and comment on the document “Conservation Status of Steelhead in Oregon” (Peer Review Draft, 9 September 1997). As you know, ORAFS is a volunteer organization of some 500 professionals representing a diverse mix of scientists in federal, state and tribal agencies, higher education, and the private sector. A principal part of the Chapter mission is promoting the application of sound science in the conservation and management of fisheries and aquatic resources.

The steelhead conservation status document clearly represents a substantial investment of time and effort by you and your colleagues. Since we received the document on 10 October, with the request that comments be provided to you 7 November, we are unable to approach the level of effort that went into this assessment in our review. In addition, the document does not provide the data used in reaching its conclusions in each of the several assessment areas. For these reasons, our comments are general and are intended to point out areas that ODFW may wish to consider in preparing the next version of this document, and in its subsequent discussions with National Marine Fisheries Service scientists.

As we understand it, the assessment developed spawner-recruit models for 26 Oregon steelhead populations. The time series of data used to develop these spawner-recruit models are neither described nor presented. One would expect that the parameters estimated from such an exercise to be sensitive to 1) the number of years of data used, and 2) whether the numbers of spawners and subsequent recruits cover the spectrum of population sizes over which density dependent interactions would be expected to manifest themselves. Further, spawner-recruit analyses and the derivation of model parameters inherently assume that all other relevant parameters are unchanged. Misleading results are likely if there have been trends in abundance or changes in survival (marine or freshwater) during the time series employed.

The document does not include any discussion of freshwater spawning and rearing habitat. Because steelhead typically spend as much as half or more of their lives in freshwater, the availability, abundance and distribution of appropriate spawning and rearing habitat are essential components of steelhead survival through early life history stages, and hence of the estimated parameters in a stock-recruit analysis. The document would be improved if the analyses addressed the trends in habitat availability, abundance and distribution. It would also be improved if it addressed the potential for downward bias in risk estimates if the data used were derived from populations whose habitat conditions are better than average for steelhead as a whole. Declining alpha values (in a stock-recruit equation) would be an expected consequence of habitat degradation. Therefore, the Department may wish to present a case for habitat stability in order to support use of a spawner-recruit approach to modeling long-term viability. An possible alternative modeling approach would integrate habitat trends into a population-habitat viability analysis which would bring population dynamics of each modeled population into an ecological context.

The assessment makes use of the available spawner-recruit data to derive estimates of equilibrium population sizes. As with the parameters of a stock-recruit equation, equilibrium population sizes will also be dependent upon habitat quality and quantity, as well as abundance trends and changes in survival as noted above. The significance of the minimum population abundance parameter relative to the calculated natural equilibrium level cannot be adequately evaluated without reference to habitat.

The assessment uses a shrinking distribution of populations indicator that emphasizes extinction of populations rather than shrinking ranges within populations. Except for the Metolius and Crooked River populations of the mid Columbia ESU extirpated by passage blockage, the assessment document generally addresses this factor with a statement that none of the populations in the given ESU are believed to have gone extinct within the past 100 years, therefore the shrinking populations indicator score is 0. Since this indicator receives equal weight with each of the other four indicators, it acts to compensate for poor scores on the extinction (EAM) or trend indicators. We suggest that the rationale for its use and weighting receive more thorough treatment.

One final area that ODFW may wish to discuss in further depth in this paper is the use of a simple arithmetic average of five indicators to assess risk to an ESU. Acceptable levels of risk involve policy determinations as well as scientific analysis and guidance. As we mention above, some of these indicators may be highly sensitive to the data used, and the conditions in which the focal populations live. More importantly, it is possible that some of the indicators may be redundant, since they are derived from the same data (e.g., quantitative extinction model and minimum population abundance indicators). Finally, the assessment warrants a discussion of the weighting of these indicators, particularly in those instances where the trend in abundance is sharply down (e.g. Klamath Mtn. province, Oregon Coast, upper Willamette, lower Columbia, and mid-Columbia ESUs) yet the overall score is strongly moderated by other indicators which may be less reliable or sensitive to the data selected.

In summary, we offer the following suggestions to improve and support the assessment presented:

1. an evaluation of steelhead spawning and rearing habitat, and trends in habitat quality and quantity for each population or population group for which a spawner-recruit analysis is conducted,

2. explicit consideration of the confidence in derived alpha and beta parameters from stock-recruit modeling in light of changing habitat conditions, and freshwater and ocean survival. Is the modeling approach chosen robust to such changes?

3. an assessment of the “representativeness” of the populations chosen for modeling,

4. a thorough discussion of the rationale for averaging across populations within ESUs, and averaging equally weighted ordinal scores in arriving at an overall “score.” This discussion should also address whether some of the indicators scored are in effect “double-counted.”

We appreciate the opportunity to comment on the draft steelhead assessment. The short time provided by ODFW for review, combined with the limited time and expertise of our volunteer reviewers necessarily limit the scope and depth of our comments. Nevertheless, we hope that our comments are useful in revising this draft and in subsequent discussions with NMFS on the status of steelhead in Oregon.


Thomas Backman, Past President for the Executive Committee

c. R. Waples, NMFS Seattle

G. Griffin, NMFS Portland