MEPS 159:151-164 (1997)  -  doi:10.3354/meps159151

Generalized Additive Models (GAMs) were used to test the hypothesis that trends in herring abundance are related to the location of ocean fronts and the temperature based variability of the northern North Sea ecosystem. Data from July of 4 years (1992 to 1995), collected during the ICES coordinated herring acoustic surveys (ICES Division IVa), were examined. It was found that geographic position and water mass characteristics influence the temporal distribution patterns of herring. The 2-stage modelling approach undertaken in the present study conveyed additional insights into the ecological behaviour of herring in the area. Temperature and depth of the thermocline, sea surface temperature, temperature difference between sea surface and seabed, and spatial location appeared to be key factors that modulate both presence and relative abundance of herring within the northern North Sea. Herring encounter was more probable in areas with cooler surface waters in the south than in the north, these areas having deeper thermoclines and temperatures at 60 m around 10°C. Results indicated that areas with higher probability of finding herring present were also located in well-mixed waters and transition zones between frontal and stratified waters. The largest herring aggregations were consistently observed in the same areas. Herring appeared to avoid the cold bottom waters of the North Sea during the summer, probably due to the relatively poor food resources there. The pronounced similarity and stability in the multiyear relationship between prespawning herring abundance, its spatial distribution and ocean environmental conditions, with and without the zero observations, support the hypothesis that the observed relationships are authentic and characteristic of the stock.

Herring (Clupea harengus) · Abundance · Environment · North Sea · Distribution · Generalized additive models (GAMs)