Fig 1.
Locations of sampling stations where biological data were collected from 1970–2015. There are 13 areas that include northern sections of the North Sea (red), west coast of Norway (cyan), east coast of Norway (purple), Skagerrak (blue), Kattegat (gold), western Baltic (pink) and adjacent fjords or lakes (also color-coded). Black dots indicate stations where the number of vertebrae were counted. Thick black stippled line = 12 NM zone. Mean position of hydrographic samples (green triangles) are shown for the North Sea, west coast, east coast, Skagerrak and western Baltic.
Fig 2.
Distribution of biological characteristics.
Histograms showing the total length, age and number of vertebrae (VS) by area (data from 1970–2015 pooled) including the mean value for all characters. The range of total length, age and VS was scaled to the amount of herring so very small values are not visualized.
Table 1.
Estimated parameters ± standard error for the von Bertalanffy growth model and the generalized linear model quantifying the length-at-first-maturity for each area.
Length where 50% (L50) and 95% (L95) of herring were mature and numbers of observation (N) with valid data are given for each area. Growth data from Lysefjorden was not sufficient to estimate growth model parameters. Due to missing of mature herring (Lysefjorden) and immature herring (Landvikvannet and Limfjorden), no maturity probabilities could be calculated.
Table 2.
Estimates of GAMs for the effects on vertebral counts and the monthly spawning dynamics in terms of the proportion of pre-, post- and spawning herring.
Modeled means together with each covariate’s degrees of freedom (Ref.df, explaining the oscillations of the modeled trend, where 1 would be linear) and p-value for each area, and the deviance explained (R2) by the model, are shown. For temperature and salinity, estimates were given before the variables were dropped from the model.
Fig 3.
Fitted results for the GAMs on the proportion of pre-spawning (A), spawning (B) and post-spawning (C) herring per month. Solid lines indicate the fitted values, dashed lines the 95% confidence intervals and points the observed values for the five major areas.
Fig 4.
Spatial and temporal dynamics of mean vertebral counts.
Mean vertebral counts (colored) for each geographic square (1° Longitude, 0.5° Latitude) and Landvikvannet (circle) per quarter (Q) of the year. ICES ‘transfer area’ is indicated by stippled line.
Fig 5.
Variation and dynamic of mean vertebral counts.
Fitted results for the GAM on mean vertebral counts (VS) in relation to year classes (A), quarters of the year (B) and stages of maturity (C). Solid lines indicate the fitted values, dashed lines the 95% confidence intervals and points the observed values for five major areas.