the film forum library tutorial contact

Researcher Disputes Accuracy of NMFS Extinction Analysis by Bill Rudolph NW Fishletter, July 19, 2002

The Canadian Journal of Fisheries and Aquatic Sciences has published a paper (only available online at present) by Seattle fisheries consultant Rich Hinrichsen that questions the accuracy of the NMFS model used to determine growth rates of salmon populations. The peer-reviewed paper reports on a study funded by BPA that examines four alternative estimators of stochastic growth rates for salmon populations.

"Like the approaches used in this paper," says Hinrichsen, "the Kareiva et al. (2000) analysis [used by NMFS] was density-independent; however, it used data from brood years 1990-1999 only, ignoring years of high productivity in the early 1980s. The severe population decline indicated by Kareiva et al. (2000) matrix analysis is not evident when the spawner data of 1980-1999 are used instead."

Hinrichsen's analysis of several index stocks for Snake River spring chinook showed all but one to actually be improving, and that one (Marsh Creek) was not declining either, but staying the same.

He said the ocean regime shifts that occur every 30 years or so can have large effects on salmon numbers, but modeling these shifts and their uncertain effects on fish populations is difficult. Hinrichsen also pointed out that effects of hatchery-reared spawners is another important unknown for estimation growth rates.

Hinrichsen said the single growth rate estimator used by NMFS to examine extinction risk of the 12 evolutionarily significant units listed as threatened or endangered under the ESA is the first time an assessment of the stocks was analyzed in a standardized way. "However," he says, " this type of analysis has come at a price."

By ignoring populations rich in age-structured data, the paper says loss of accuracy can be great. "In fact," says the researcher, "because of the large errors of estimating stochastic growth rate, the primary goal of NMFS in their broad analysis, identifying the stocks with greatest population decline, may be thwarted. A one-size-fits-all approach is not needed to compare risk."

Related Sites:
Canadian Journal of Fisheries and Aquatic Sciences