Figure 1.
Bathymetric relief map of the study area.
Features mentioned in the text are labelled.
Figure 2.
Distribution of the blue whiting spawning stock from a fisheries acoustic survey.
The acoustic intensity of blue whiting (sA, which is directly related to abundance) from the International Blue Whiting Spawning Stock Survey (IBWSS) is shown for the 2013 acoustic survey [52]. Isobaths are plotted as grey lines. International maritime boundaries are plotted as red dotted lines. Note the continuous distribution along the shelf edge and limited southern extension of the survey.
Table 1.
Model fitting results.
Figure 3.
Spatial distribution of CPR samples.
Grey points are locations where CPR samples have been checked for fish larvae. Red circles are where these samples were found to contain blue whiting. The blue box denotes the spatial region of interest used in further model-based analyses.
Figure 4.
Temporal distribution of CPR samples.
Temporal distribution of samples checked for blue whiting larvae obtained from the CPR in the region of interest outlined in Figure 3. a) Sampling frequency in each year b) Presence frequency in each year c) Sampling frequency as a function of date in the year d) Presence frequency as a function of date in the year. In a) and b), each bar corresponds to a single year, whilst in c) and d) it corresponds to a day of year.
Figure 5.
Distribution of larval abundances reported in the CPR.
The relative proportion of each non-zero abundance category reported (bars) and the cumulative proportion (line) are show. Cumulative proportion is defined here as the proportion of presences with an abundance less than or equal to the given category. Note that the abundances are the abundance categories reported by the CPR survey [24]. Observations of zero larvae (absence) are omitted from this distribution.
Figure 6.
Spatial larval-presence probability distribution.
Results predicted from Model 10 ( = east * north * doy + s(year)
comp) are plotted as a probability density function for each population (i.e. the spatial integral over the domain of each of the two populations is 1). The black horizontal line indicates the location of the arbitrary division between a northern and southern population at 53 °N. Note abundances cannot be compared between the domains, as each domain is normalised to give an integral of 1. Isobaths are draw at 200 m (thin line) and 1000 m (thicker line) depths for reference. Map projection is UTM Zone 28.
Figure 7.
Timing of peak probability of occurrence.
Results predicted from Model 10 ( = east * north * doy + s(year)
comp). The day of year (colour scale) when the local maximum in probability of larval presence occurs is plotted as a function of space. The black horizontal line indicates the location of the arbitrary division between a northern and southern population (at 53 °N). The spatial distribution in Figure 6 is used to mask the output so that only the core 75% of the larval distribution in each region is plotted: regions where there are few larvae, and the estimated timing of spawning is therefore imprecise, are thus omitted. Isobaths are draw at 200 m (thin line) and 1000 m (thicker line) depths for reference. Map projection is UTM Zone 28.
Figure 8.
Zonally integrated larval-presence probability distribution.
Results from Model 10 ( = east * north * doy + s(year)
comp), plotting the probability distribution of larval-presence as a function of latitude and day of year. The probability of larval-presence is expressed as a density function for each population (i.e. the integral over each of the two populations is 1). The black horizontal line indicates the location of a hypothesised division between a northern and southern spawning population (at 53 °N). Note that because this model allows the relative abundances of the two populations to vary from year to year, abundances cannot be compared between the domains. The projected UTM coordinates used in the fitted model have been reprojected back to longitude here for ease of interpretation.
Figure 9.
Annually integrated larval occurrence-probability.
Results from Model 10 ( = east * north * doy + s(year)
comp). The probability of observing larvae integrated over the spatial domain and day of year is a measure of larval abundance in that year and is plotted as a function of the year for the northern (red) and southern (blue) populations, with the associated 67% (i.e. corresponding to 1 standard deviation) confidence intervals. The units of larval abundance plotted here are arbitrary but scale linearly.