A Simple Stratified Design for Mark-Recapture Estimation of Salmon Smolt Abundance

Abstract

We describe a mark–recapture (M–R) technique in which a stratified design and sampling at 1 or 2 stream locations are used to estimate the abundance of a migrating salmon smolt population. The method consists of counting smolts captured at a designated downstream site and releasing marked smolts back into the population at an upstream site. Marked smolts subsequently recovered at the downstream site are counted to estimate capture probability (trap efficiency), which is used to estimate smolt abundance for a segment of the population. This procedure is temporally stratified such that each trap efficiency trial is discretely paired with one capture period; this can typically be accomplished by releasing marked smolts at relatively short intervals (a few days) with little chance of recaptured fish occurring in later strata. This approach accounts for potential temporal changes in capture probability under a fairly modest assumption of stratum consistency. The method simplifies the generalized 2-sample stratified design and provides some important advantages: (1) because marking occurs in discrete intervals, personnel costs are substantially reduced; (2) because each release of marked smolts corresponds to one capture period, only one type of mark is needed, which greatly simplifies marking procedures and recapture tallying; and (3) when only one capture site is used, material costs are reduced by about half. We present approximately unbiased abundance and variance estimators of the total smolt population and develop a method of estimating the number of smolts to be marked. A parametric bootstrap technique for quantifying precision is also developed. An example of the method is given using the 1997 sockeye salmon Oncorhynchus nerka smolts migrating from Akalura Lake, Kodiak Island, Alaska. The Akalura Lake study included a weir count of smolts, which we used to evaluate the accuracy of the M–R estimate. Authors: STAN R. CARLSON is a biometrician for the Alaska Department of Fish and Game, Division of Commercial Fisheries, 34828 Kalifornsky Beach Road, Soldotna, Alaska 99669. LEWIS G. COGGINS JR., former fishery biologist for the Alaska Department of Fish and Game, Commercial Fisheries Division, Kodiak, is a research fisheries biologist with the Alaska Biological Science Center, 1011 E. Tudor Road, Anchorage, Alaska 99503. CHARLES O. SWANTON, former area fishery biologist for the Alaska Department of Fish and Game, Division of Commercial Fisheries, Kodiak, is now a regional management biologist for the Alaska Department of Fish and Game, Division of Sport Fisheries, 300 College Road, Fairbanks, Alaska 99701. Acknowledgments: Dana Schmidt, former principal limnologist for the Alaska Department of Fish and Game, encouraged the completion of this work. Gary Kyle, Gary Fandrei, Bruce King, Patrick Shields, and Gary Todd — helpful discussions. Project Sponsorship: This research was partially funded from Exxon Valdez Oil Spill Restoration projects 258 and 259 with the balance funded by the Alaska Department of Fish and Game. 88 Alaska Fishery Research Bulletin 5(2):88–102. 1998. Copyright © 1998 by the Alaska Department of Fish and Game