Workpackages

To determine the statistical resolution and power of combined techniques.
To describe the inter-relationships of herring stocks west of the British Isles using the biological information gathered in WP’s 02 - 06.
To use the group membership data to evaluate catch data to identify current patterns of mixing and spatial shifting of herring stocks in western European waters.

Data resulting from all analyses in workpackages (WP) 01 – 06 will be passed on to Partner 2 and entered into a database. Data will initially be analysed using multinomial stepwise logistic regression (MSLR). This multivariate statistical technique is used to predict a discrete outcome, as group membership from the mix of variables (continuous, discrete and nominal) produced by the different WESTHER WP’s. MSLR analyses will be performed using a stepwise inclusion of data produced by each component WESTHER WP. The MSLR statistics will be tested against other predictive modelling techniques such as neural networking (Swingler 1996). Developing techniques will be incorporated when appropriate. The methods will be tested and refined using data from the first year to be readily available for the evaluation of both years’ data at the end of the project.
The stock separation data will be amalgamated with data on catches by area and the standard stock assessment data (see ICES 2001) to examine the interaction between stock dynamics and spatial heterogeneity in stock and catch distribution.

(C) EU-project WESTHER 2003 | WESTHER home | mail