No abstract

Phytoplankton primary production is at the base of the marine food web; changes in primary production have direct or indirect effects on higher trophic levels, from zooplankton organisms to marine mammals and seabirds. Here, we present a new time-series on gross primary production in the North Sea, from 1988 to 2013, estimated using in situ measurements of chlorophyll and underwater light. This shows that recent decades have seen a significant decline in primary production in the North Sea. Moreover, primary production differs in magnitude between six hydrodynamic regions within the North Sea. Sea surface warming and reduced riverine nutrient inputs are found to be likely contributors to the declining levels of primary production. In turn, significant correlations are found between observed changes in primary production and the dynamics of higher trophic levels including (small) copepods and a standardized index of fish recruitment, averaged over seven stocks of high commercial significance in the North Sea. Given positive (bottom-up) associations between primary production, zooplankton abundance and fish stock recruitment, this study provides strong evidence that if the decline in primary production continues, knock-on effects upon the productivity of fisheries are to be expected unless these fisheries are managed effectively and cautiously.

During the 1980s, a rapid increase in the Phytoplankton Colour Index (PCI), a semiquantitative visual estimate of algal biomass, was observed in the North Sea as part of a regionwide regime shift. Two new data sets created from the relationship between the PCI and SeaWiFS chlorophyll a (Chl a) quantify differences in the previous and current regimes for both the anthropogenically affected coastal North Sea and the comparatively unaffected open North Sea. The new regime maintains a 13% higher Chl a concentration in the open North Sea and a 21% higher concentration in coastal North Sea waters. However, the current regime has lower total nitrogen and total phosphorus concentrations than the previous regime, although the molar N : P ratio in coastal waters is now well above the Redfield ratio and continually increasing. Besides becoming warmer, North Sea waters are also becoming clearer (i.e., less turbid), thereby allowing the normally light-limited coastal phytoplankton to more effectively utilize lower concentrations of nutrients. Linear regression analyses indicate that winter Secchi depth and sea surface temperature are the most important predictors of coastal Chl a, while Atlantic inflow is the best predictor of open Chl a; nutrient concentrations are not a significant predictor in either model. Thus, despite decreasing nutrient concentrations, Chl a continues to increase, suggesting that climatic variability and water transparency may be more important than nutrient concentrations to phytoplankton production at the scale of this study.

Although the terms 'health' and 'healthy' are often applied to marine ecosystems and communicate information about holistic condition (e.g. as required by the Ecosystem Approach), their meaning is unclear. Ecosystems have been understood in various ways, from non-interacting populations of species to complex integrated systems. Health has been seen as a metaphor, an indicator that aggregates over system components, or a non-localized emergent system property. After a review, we define good ecosystem health as: 'the condition of a system that is self-maintaining, vigorous, resilient to externally imposed pressures, and able to sustain services to humans. It contains healthy organisms and populations, and adequate functional diversity and functional response diversity. All expected trophic levels are present and well interconnected, and there is good spatial connectivity amongst subsystems.' We equate this condition with good ecological or environmental status, e.g. as referred to by recent EU Directives. Resilience is central to health, but difficult to measure directly. Ecosystems under anthropogenic pressure are at risk of losing resilience, and thus of suffering regime shifts and loss of services. For monitoring whole ecosystems, we propose an approach based on 'trajectories in ecosystem state space', illustrated with time-series from the northwestern North Sea. Change is visualized as Euclidian distance from an arbitrary reference state. Variability about a trend in distance is used as a proxy for inverse resilience. We identify the need for institutional support for long time-series to underpin this approach, and for research to establish state space co-ordinates for systems in good health.