Reliable Measurements of Codend Selectivity Even When the Mesh Opening of the Reference Codend Is Not Sufficiently Small
9 Pages Posted: 11 Oct 2013 Last revised: 17 Jun 2015
Date Written: October 10, 2013
The selection properties of gears are mainly determined by the selection properties of the codend, frequently described by the logistic function. This function reflects the retention probability of fish in the gear as a function of length caught from the length distribution of fish in that area. Certain conditions must be met for the measurements of the selection properties of trawls. According to the approach of Millar and Walsh (1992) the selection properties for the reference-trawl codend have to be chosen in a way that the frequencies of fish caught by the reference gear have a retention probability pr =1 for the length range within the test gear catches fish. We need small meshes in the reference codend. But small meshes in the reference trawl can reduce dramatically the flow rate in the codend (Thiele et al., 1997). This may change the fish behavior and may lead to measurement errors. A compromise must be found. In order to devastate not too much young fish, in practice the meshes of the reference-trawl codend are usually chosen as large as possible. Here, errors may occur, especially since the requirements for the selection properties of the reference-trawl codend depend on the results of the measurements which are obtained later. Applying the traditional approach but using too large meshes in reference-trawl codend can cause measurement errors. In the chosen example we get the largest measuring error for the selection-range factor SR with 36 % but for all other parameters lower errors. The new approach addresses this issue and avoids the occurrence of these errors.
Keywords: Logistic function, Log-Likelihood function, Mesh opening, 50% retention length, Retention probability, Selection length factor, Selection range factor
JEL Classification: Q22
Suggested Citation: Suggested Citation