Figure 1.
Investigation area and coverage of the stratified random and standardized ICES BITS trawl survey with new survey design according to Nielsen et al. [22] and Lewy et al. [23].
The stratified random haul locations are black dots (upper panel) and the additional coverage for the extended BITS survey in the Fehmarn Belt Area of the Western Baltic Sea with haul locations are indicated by black dots and associated hydrographical CTD stations as light dots (lower panel), exemplified for the quarter 4 2010.
Figure 2.
Comparison between fish size correlation matrix from different single-species model specifications for cod year 2009 quarter 4.
SS1, Unconstrained free size correlation structure given as a positive definite correlation matrix (a); SS2, natural untransformed scale (b); SS3, log scaled (c); SS4, logistic scaled (d).
Table 1.
Overview of the models considered and tested, where single, multispecies and multiyear models are covered. In addition, the hierarchical structure of the model testing is indicated.
Figure 3.
Model comparison for cod year 2010 quarter 4.
SS1, unconstrained (a); SS2, natural scale (b); SS3, log scaled (c); SS4, logistic scaled (d).
Table 2.
Model comparisons with unconstrained size correlation versus natural, log and logistic scaled and model reductions according to either species or year.
Figure 4.
Estimated correlation (y-axis) for different size correlation functions, Eq. (5), of cod and whiting (single species runs) using the log transform model parameters given by Table 3.
The vertical bars indicate the 95% confidence intervals.
Table 3.
Summary of single species runs: Parameter estimates of log-transform model and size-correlation (first 5 parameters) plus spatial covariance parameters (final 9).
Figure 5.
Multispecies models for cod and whiting year 2009 quarter 4.
MS1, unconstrained species correlation with separate parametric size correlation for each species (a); MS2, unconstrained species correlation with common parametric size correlation for both species (b); MS3, separable species-size correlation (c); MS4, no species correlation (d).
Figure 6.
Multiyear models for cod year 2009 and 2010 quarter 4.
Unconstrained year correlation with separate parametric size correlation for each year (MY1, a), Unconstrained year correlation with common parametric size correlation for both years (MY2, b), separable year-size correlation (MY3, c), No year correlation (independence) (MY4, c).
Figure 7.
Maps of relative whiting abundance 2009/Q4 based on whiting observations (left column) versus the same maps based on cod observations (right column) utilizing MS1 model of Table 2.
The three row panels indicate three whiting size groups in cm.
Figure 8.
Predictive power of the models illustrated by maps of relative cod abundance 2010/Quarter 4 based on cod observations that year (left column) versus the same maps based on cod observations previous year (right column) utilizing model MY3.
The three row panels indicate three cod size groups in cm.