Premium
An Assessment Model for a Standard Gill Net Incorporating Direct and Indirect Selectivity Applied to Walleye
Author(s) -
Radomski Paul,
Anderson Charles S.,
Bruesewitz Richard E.,
Carlson Andrew J.,
Borkholder Brian D.
Publication year - 2020
Publication title -
north american journal of fisheries management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.587
H-Index - 72
eISSN - 1548-8675
pISSN - 0275-5947
DOI - 10.1002/nafm.10384
Subject(s) - selectivity , fishing , net (polyhedron) , statistics , standard deviation , log normal distribution , mathematics , fishery , environmental science , biology , geometry , biochemistry , catalysis
Fisheries managers regularly use catch‐per‐effort ( CPE ) statistics and length‐frequency distributions from standard gill‐net surveys to inform future management activities. The data from these surveys is implicitly tied to the net's selectivity, and caution should be used when not accounting for selectivity. We combined indirect and direct methods for estimating the parameters of absolute selectivity by using data from Minnesota's standard gill‐net surveys for Walleye Sander vitreus , enabling the estimation of density from CPE . The indirect piece used a state‐wide gill‐net database and generalized linear modeling to identify a set of possible shapes for the selectivity curves. The direct piece added information based on 94 mark–recapture experiments that were paired with standard gill‐net surveys. The resulting statistical assessment model used a bi‐lognormal‐shaped selectivity curve that was geometrically similar for each mesh, and it estimated fishing intensity that differed substantially with mesh size. Applications of the selectivity model allow the estimation of Walleye density as CPE divided by absolute selectivity by length‐bin. The estimated absolute selectivity for the Minnesota standard gill net increased with increasing Walleye length to a peak of 0.76 ha/net at 535 mm and then decreased to about 0.26 at about 800 mm (1.0 and 0.34 relative selectivities, respectively). Fishing intensity increased with mesh size, so the selectivity curves for the individual meshes and that for the standard gill net (5 meshes in series) differed substantially from the curves that have resulted from indirect selectivity models that have ignored size‐dependent encounter probabilities and assumed equal contact probabilities for all meshes. We caution investigators that adjusting gill‐net catch data with indirect selectivity models that assume equal contact probabilities for each mesh may introduce considerable bias to estimates of population abundance and length distributions.