Cross-scale observations on distribution and behavioural dynamics of ocean feeding Norwegian spring-spawning herring (Clupea harengus L.)

Interpretation of the behavioural dynamics and distribution of herring requires explicit consideration of spatial and temporal scales since no single mechanism can explain the pattern on all scales. High resolution sonar and echosounders were used as our observation tools. The cluster ratio, mean Nearest Neighbour Distance (mean ISD): mean of the average Inter-school Distance (mean ISD) provides a descriptive index linking scales of distribution pattern within and between clusters of schools. It can be used to compare school clustering patterns for surveys made at different places and seasons. Two mesoscale patterns of school clustering were observed, as indicated by the distribution of nearest neighbour distances; one at 0.05–0.3 km revealed by sonar and another at 0.8–2.5 km revealed by echosounder. In general, schools were tightly aggregated, whilst school clusters were patchily distributed. Heightened feeding motivation explains the predominance of relatively small schools and their tendency to be found closer together and more clustered than large schools. Differences in distribution and density of food patches and predators may account to some degree for the observations (1) schools were distributed according to their size; both small and large schools had nearest neighbours of similar size more often than expected, (2) splitting and joining of schools were the most frequently recorded behavioural events. A considerable amount of variation in distribution and dynamics can be accounted for by changes in behaviour associated with diurnal vertical migrations. Some schools apparently did not perform diurnal vertical migrations. We suggest that when profitable, some herring may ‘‘dip in to the fridge’’ (feed in deep, cold water) during the day rather than pursue food on potentially high energy migrations. 1999 International Council for the Exploration of the Sea