Fish Surveys Not Accurate--NRC

To protect commercially important fishes from being overharvested, regulators need better ways to count them, says a report to be released tomorrow by the National Research Council. The NRC's blue-ribbon panel found that current assessment methods consistently overestimate declining fish populations.

Currently, the U.S. National Marine Fisheries Service (NMFS) estimates that at least a third of the nation's 275 commercial fish stocks are overharvested. Last year, NMFS officials asked scientists to see whether current counting methods, which extrapolate fish populations from scientific and commercial hauls, are partly to blame.

As part of their review, the researchers examined the accuracy of stock assessment models by creating hypothetical fish populations, then simulating sampling. When the output of the models was compared to the hypothetical populations, the researchers found that the models consistently misrepresented them--often overestimating the size of shrinking fish stocks by more than 25%, while underestimating the number of fish in recovering populations. These findings "should be sobering to scientists, managers, and [fishers]," concludes the panel, which was headed by Richard Deriso of the University of Washington, Seattle, and Terrance Quinn of the University of Alaska, Fairbanks.

To improve stock assessments, the committee recommends that regulators pay more attention to fishing methods--which vary in their impact on different species--when predicting populations. They should also develop more sophisticated models that better incorporate statistical uncertainties. Finally, the committee proposed that independent scientists more regularly review population estimates made by NMFS, which both fishers and environmentalists say have become colored by politics.

The recommendations are "good, wise, useful and important," says Kimberly Davis of the Center for Marine Conservation in St. Petersburg, Florida. She cautions, however, that "it won't matter how good the science is if it's not used in management decisions--and that has been the problem in the past."