Spatial distribution of herring in the orkney/shetland area (northern north sea): A geostatistical analysis

Abstract

Geostatistical methodology was used to analyse the structure and describe the spatial patterns of North Sea herring (Clupea harengus L.), using data from the 1992 ICES (Division IVa) acoustic survey. Three different scales of spatial structures were identified: an unresolved small-scale variability, which accounted for 48% of the total variance, and two structure components, the first being a meso-scale of ≈9 nmi (nautical miles) (≈30%) and the second a large-scale of ≈17 nmi (22%). Geostatistical analysis permitted the determination of spatial density gradients as well as patch sizes (range from 9 to 17 nmi). The use of the truncated data and the robust variogram on the raw data provided additional information for the structure. The utilization of this information in the variographic analysis resulted in better estimations. The best unbiased predictor was used to objectively map the herring population distribution by kriging. The kriging estimates were better with interpolation of a large number of points. Herring tended to aggregate mainly in meso-scale patches with a diameter of 9 nmi and to a lesser extent in large-scale patches of 17 nmi diameter. Environmental factors (depth, salinity and temperature) partly explained the spatial distribution of herring, despite the absence of a trend in the variogram structure. The study demonstrates the existence of spatial correlation and an objective way of optimal mapping of the population. Geostatistics provided additional information on herring spatial organization which is important to understand the behaviour of the species and to study its relationship with the environment.