A CATCH-LENGTH ANALYSIS FOR CRAB POPULATIONS

A catch-length model was constructed to estimate abundance of crab pop ulations for which no survey data are available. The model incorporate s stochastic growth by length and gradual recruitment over length and assumes constant catchability. Required data include catch by length a nd shell condition, fishing effort, growth increment per molt by lengt h, and annual natural mortality rate. Model application to red king cr ab populations in Bristol Bay and off Kodiak Island, Alaska, generally provides accurate estimates of trends of relative population abundanc es. The accuracy of absolute abundance estimates depends on knowledge of natural mortality. The model provided a good fit to the catch by le ngth and shell condition for both populations. In comparison to popula tion abundances estimated directly from surveys, the catch-length mode l performed best With instantaneous natural mortality set equal to 0.4 and with fishing effort and catch-length/shell composition weighted e qually in the calculation of residual sum of squares.