Bias in the Estimation of Impacts of Simultaneous Mark-Selective and Nonselective Fisheries on Ocean Salmon

Abstract

The management models currently used in the Pacific Fishery Management Council's preseason planning process to project mortalities during proposed fisheries for coho salmon Oncorhynchus kisutch and Chinook salmon O. tshawytscha produce biased estimates of the mortalities in mark-selective fisheries and concurrent nonselective fisheries. The bias is caused by (1) unmarked fish encountering fishing gear more than once because of the required release of all unmarked fish and (2) the change in the unmarked–marked ratio resulting from the selective removal of marked salmon by the mark-selective fishery. Biased estimates of mortalities in nonselective fisheries operating during the same time and area as mark-selective fisheries occur because of the increasing unmarked–marked ratio in the common pool of salmon being fished. This bias is an increasing function of the stock exploitation rate on marked fish. The expected exploitation rates must also take into account mark-recognition errors whereby marked fish are released by mistake and unmarked fish are landed by mistake. The adjustment for mark-recognition error is a function of the unmarked–marked ratio.

Received January 18, 2011; accepted June 13, 2011

ACKNOWLEDGMENTS

Paul McElhany (National Oceanic and Atmospheric Administration–Fisheries) suggested using a power function to model the release mortality rate with each successive encounter. Andy Rankis (Northwest Indian Fisheries Commission), Angelika Hagen-Breaux (Washington Department of Fish and Wildlife), and several anonymous reviewers provided helpful suggestions for the analysis and manuscript. The findings and conclusions in this manuscript are those of the authors and do not necessarily represent the views of the U.S. Fish and Wildlife Service or the Northwest Indian Fisheries Commission.