Evaluation Of The Use Of Segmented Regression Through Simulation For A Characterisation Of The North Sea Cod (Gadus Morhua L.) Stock, In Order To Determine The Properties Of Blim (The Biomass At Which Recruitment Is Impaired)

No abstracts are to be cited without prior reference to the author.

A new objective technique to determine the level of biomass (Blim) at which recruitment is impaired based upon a segmented (or piecewise linear) regression has previously been proposed and developed by O’Brien & Maxwell. The formal statistical details of the approach are presented; together with hypothesis tests and goodness-of-fit statistics, and the approach evaluated for a characterisation of the North Sea cod (Gadus morhua L.) stock. This paper computes the values of Blim obtained by segmented regression for a simulated stock based upon North Sea cod. The properties of the estimators are explored for a variety of hypotheses about stock dynamics (e.g. stock-recruitment) and data quality (e.g. discarding and mis-reporting). In particular, the robustness of the method to underlying assumptions is evaluated.