Selectivity Selectivity is a function of fish harvesting

  • Slides: 9
Download presentation
Selectivity

Selectivity

Selectivity is a function of fish harvesting gear • Selectivity is used to model

Selectivity is a function of fish harvesting gear • Selectivity is used to model the vulnerability of fish to the gear as well as the availability – “retention” • Selectivity can be a function of length, age, or both. Selectivity may also differ between the sexes, change over time or space. • What is the difference in catchability and selectivity? • Selectivity is one of the main components of a stock assessment model related to the fishing process

Temporal variability • Selectivity may change over time due to modifications to gear, changes

Temporal variability • Selectivity may change over time due to modifications to gear, changes in the season or area fished, temporal variability in growth, or other factors. • Therefore, modeling these changes may be important in the stock assessment. • In general, fisheries are defined so that their selectivity does not change over time. • For example, if a fisheries selectivity does change for some reason (e. g. mesh size changes) the fishery is often divided into two fisheries with different selectivity curves estimated – “Blocks”

Length based versus age based selectivity • Selectivity can be modeled as a function

Length based versus age based selectivity • Selectivity can be modeled as a function of length, age, or both. • Selectivity of nets generally has a length based component. • Small fish go through the mesh and large fish may be too large to be meshed (e. g. in set nets). • In other cases, fish may migrate once they reach a certain age and selectivity for fisheries in different areas is age based. • In most cases it is difficult to know which process is more important, age or length. • When length is a fixed function of age, it may not matter which is used. • However, if growth or the variation of length-at-age is being estimated, it may be important to use the correct basis for modeling selectivity.

Broad classification • There is a very large selection of models • Broadly: •

Broad classification • There is a very large selection of models • Broadly: • A sigmoid curve, increasing from some positive value less than one to one as a function of fish size. • A dome-shaped curve, increasing from some positive value less than one to one, then decreasing again as a function of fish size.

Asymptotic selectivity • Historically, selectivity was often assumed to be asymptotic. However, as stock

Asymptotic selectivity • Historically, selectivity was often assumed to be asymptotic. However, as stock assessment models started to become more complex and model different gear types and areas as different fisheries, it become apparent that many of the fisheries had dome shape selectivity curves. • However, if all fisheries are dome shape, the descending limbs of the selectivity curves are confounded with natural mortality, catchability, and other model parameters. • It is therefore often considered necessary to assume that one of the fisheries has an asymptotic selectivity to stabilize the parameter estimation. • Using an asymptotic selectivity when the actual selectivity is dome shaped will tend to cause estimates of fishing mortality to be biased high. In some fisheries, tagging data suggests that selectivity for all fisheries is dome shape. However, this is confounded with increasing natural mortality with age.

Logistic Cumulative Distribution Function • The size selection characteristics of trawl codend meshes and

Logistic Cumulative Distribution Function • The size selection characteristics of trawl codend meshes and some hooks can be represented by a logistic cumulative distribution function (LCDF) of the form:

Normal Probability Distribution Function • The size selection characteristics of gillnets and some traps

Normal Probability Distribution Function • The size selection characteristics of gillnets and some traps are represented by a truncated, scaled normal probability distribution function (NPDF):

Putting it all together: Thorson and Prager

Putting it all together: Thorson and Prager