Estimation of Gill-Net Selectivity for Lake Trout and Application to Demographic Parameter Estimates

Tuesday, August 19, 2014: 8:40 AM
2105 (Centre des congrès de Québec // Québec City Convention Centre)
Elizabeth L. Ng , Department of Fish and Wildlife Sciences, University of Idaho, Moscow, ID
Michael C. Quist , Department of Fish and Wildlife Sciences, University of Idaho, U.S. Geological Survey, Idaho Cooperative Fish and Wildlife Research Unit, Moscow, ID
James P. Fredericks , Idaho Department of Fish and Game, Coeur d'Alene, ID
Accurate estimates of catch rates and population length distributions are important for many monitoring and assessment techniques.  However, sampling gears such as gill nets are notoriously size selective.  Fishing multiple mesh sizes can reduce bias, but it is often useful to correct naive length distributions for gear selectivity.   In Priest Lake, Idaho, the first fishery-independent study of the Lake Trout Salvelinus namaycush population was conducted from March to May 2013.  Eight sizes of monofilament gill nets were used to sample 3,096 Lake Trout.  Selectivity (i.e., retention probability) was estimated indirectly for each mesh size by fitting a response surface to catch data using generalized linear models.  The unadjusted length distribution was then divided by the estimated selectivity for each length class to produce an adjusted length distribution.  The adjusted and unadjusted length distributions were then used to estimate annual mortality rates using catch curves.  These estimates were not significantly different, indicating the robustness of catch curves to variability in selectivity.  Although some estimates of demographic parameters are robust to unequal selectivity, adjustments should be useful for other methods that seek to explicitly estimate size-class abundance.  Estimates of selectivity can also be useful in the design of monitoring studies.