The development of indicators for the Marine Strategy Framework Directive (MSFD) is particularly challenging for food-webs. Ideally, these indicators should signal the state of marine systems, despite that these are highly dynamic with complex and often non-linear interactions. Marine regions characterized by strong environmental gradients, such as the Baltic Sea, imply additional complexity, as the relevance of indicators needs to be known for the entire range of environmental conditions. Here, a robust testing of these indicators is necessary, accounting for their responsiveness to multiple pressures and the linkages between indicators caused by species interactions. In this project, we applied different approaches to robustly test pelagic food-web indicators across Baltic Sea offshore regions (Bornholm basin, Gotland basin, Bothnian Sea), organized into three work packages.
In work package 1, we developed a framework based on models in which single indicators were linked to single pressures (such as nutrient concentration and fishing) to validate the performance of individual indicators in terms of their sensitivity, robustness, and specificity. This framework allowed us to evaluate and compare six zooplankton-based and seven fish-based indicators suggested as pelagic food-web indicators across two central Baltic Sea basins (Bornholm and Gotland Basin) and the northern Bothnian Sea (only zooplankton-based indicators evaluated in this basin, due to lack of long time-series of pelagic survey data on fish). Those indicators that showed a better performance, including the two HELCOM core zooplankton-based indicators Total Zooplankton Abundance (TZA) and Mean Size (MS), were further analysed for potential coupling among indicators.
Two different novel indicator-testing approaches were developed to assess and account for the linkages among indicators arising from species interactions, and to identify the relationships between indicators and pressures while accounting for such linkages among indicators. These two methods were then applied to food-web indicators across two trophic levels (work package 2; size- and abundance-based indicators of piscivorous and planktivorous fish) and three trophic levels (work package 3; different fish- and zooplankton based indicators) in the Bornholm and Gotland basins.
To explain temporal variation of abundance- and size-based food-web indicators based on the predator species cod (Gadus morhua) and its two main prey sprat (Sprattus sprattus) and herring (Clupea harengus) in the Bornholm Basin, weshowed that linkages among indicators corresponding to predator-prey interactions and intra-specific density-dependence were essential. Long-term variation in the abundance-based indicators (tested for 1979-2012) was better explained than in the size-based indicators (1984-2012). We could not find any statistically validatedmodel for the time series on these two sets of indicators from the Gotland Basin. The tests of indicators across three trophic levels based on a different statistical method also showed that indicators at different trophic levels often influence each other. Robust models explaining temporal variation in the indicators were found for zooplankton-based indicators (in both basins) and for the piscivore-based indicator (cod abundance in the Bornholm Basin), whereas variation in size-based fish indicators were less well explained also when using this statistical method.
In both types of analyses in work packages 2 and 3, indicators responded to multiple pressures rather than single ones, including climate-linked variables, fishing and nutrient concentrations or primary production indicators. When including only fish-based indicators in the two-trophic level approach, salinity, temperature, oxygen conditions as well as nutrients were found essential in explaining the temporal development of the food-web indicators. In the threetrophic level approach, which included zooplankton-based indicators, climate variables had more often a stronger effect on these indicators than fishing and primary production. This highlights the need to account for climate change when interpreting indicators and evaluating management efforts. In addition, indicator responses to these external pressures were often not linear and even depended on the condition of another pressure variable. This illustrates the usefulness of advanced modelling tools to fully capture relevant dynamics in systems characterized by different regimes.
The project provides the first steps towards an assessment of the current food-web status based on basin-specific sets of well performing fish- and zooplanktonbased food-web indicators. Our results highlight that (1) linkages among indicators that arise from interactions between species at different trophic levels need to be accounted for in further development of food-web indicators, (2) pelagic food-web indicators representing zooplankton, planktivorous fish, and piscivorous fish all respond to multiple pressures, (3) variation in variables linked to fishing, eutrophication (primary production/nutrients) and climate change all explained indicator historical development (1979/1984-2012). Despite a few differences in the indicator-specific pressures identified as significant between the three approaches, we have confidence in the indicator-pressure relationships identified by at least two approaches. We consider these pressures as key drivers. However, not all indicators showed a statistically robust relationship to the tested pressures – either within one approach or across approaches. Consequently only a subset of indicators, given the underlying data and sampling scheme as basis, is recommended suitable for risk analysis or management strategy evaluations of pelagic food-webs in the studied areas, as required under an integrated food-web assessment programme within MSFD:
• Total zooplankton abundance (TZA), ratio cladocerans to copepods (RCC), Sprat, Herring, and Cod for the Bornholm basin
• Mean zooplankton size (MS), ratio cladocerans to copepods (RCC), Sprat, erring, and Cod for the Gotland basin
• Total zooplankton abundance (TZA), mean zooplankton size (MS), and opepod biomass (Cops) for the Bothnian Sea