Ecosystem regime state in the Baltic Proper and the Gulf of Riga

	Authors: Christian Möllmann1, Bärbel Müller-Karulis2, Rabea Diekmann1, Juha Flinkman3

Key message:

• Three distinct regimes in Baltic ecosystem structure were detected from 1974 – 2004 (Fig. 1)
• Regime shifts – major changes in pelagic ecosystem structure – occurred at the end of the 1980s and during the early 1990s
• Pelagic food web changes were climate driven and further accelerated by fishing pressure and internal processes

Figure 1: Ecosystem regime shifts in the Baltic Proper (left panel) and the Gulf of Riga (right panel), depicted by principal component scores of 75 and 31 indicator time series for the respective areas. Mathematically, principal components (PCs) are independent linear combinations of the original variables, ordered by the variation in the data they explain. The first two PCs, on which the years are projected in the figure, therefore summarize the most prominent structures in the dataset. Circles mark clusters of similar consecutive years, identifying three ecosystem regimes. Click images to enlarge.

Relevance of the indicator for describing developments in the environment

The indicator is based on multivariate analysis of biotic and abiotic time series. It describes major shifts in ecosystem structure and associates them with fundamental driving factors. Major drivers identified were climate-induced changes in temperature and salinity, fishing pressure, as well as variations in bottom water oxygen conditions, which are both climate and eutrophication related. The analysis also shows key interactions between ecosystem components (e.g. resource competition and limitation).

Policy relevance and policy references

The Baltic Sea is subject to pronounced climate-driven variations in temperature and salinity levels, which have a distinct impact upon ecosystem structure. Changes in temperature and salinity directly affect the composition of zooplankton and fish communities. Indirect effects are caused by the impact of the saline inflow regime on deep water oxygen conditions, which further influence not only fish recruitment, but also nutrient levels in the Baltic Proper.

Management has – if at all - only little impact on climate driven processes in the Baltic ecosystem. Target levels for Baltic ecosystem components should therefore account for the effect of large scale climatic variations. Also strong interactions between ecosystem components, for example the interrelation between cod, herring and sprat stocks, should be taken into consideration within an adaptive ecosystem based management framework for the Baltic Sea.

Assessment

Baltic Proper

The ecosystem of the Baltic Proper changed from a low temperature, high saline state with frequent inflows during the 1970s and 1980s through a freshening and warming stage with extensive bottom water anoxia and high nutrient concentrations in the beginning of the 1990s to a warm, low saline state with stable nutrient concentrations (Fig. 2). Associated were drastic changes in fish and zooplankton communities, driving the Baltic from a cod/herring to a sprat dominated system. Biomass of stenohaline species in the zooplankton community (Pseudocalanus acuspes) decreased, while the share of smaller, temperature controlled taxa (Acartia, Temora) grew.

The changes in the fish community are the combined results of direct and indirect effects of climate on recruitment, as well as of low cod predation pressure. Decreasing salinities and accompanying low oxygen conditions caused recruitment failure of Eastern Baltic cod, largely explaining the collapse of the stock, which was further amplified by high fishing pressure (Köster et al. 2005). Processes leading to recruitment failure were high egg mortality due to low oxygen contents (Köster et al. 2003) and low Pseudocalanus acuspes availability, the preferred larval food (Hinrichsen et al. 2002). In contrast, sprat profited from increased temperatures, which positively affected egg survival (Köster et al. 2003, Baumann et al. 2006) and led to an increase in Acartia spp. biomass, the major food source of sprat larvae (Voss et al. 2003).

Over the time scale investigated, fluctuations in nutrient concentrations were largely driven by the duration of stagnation periods, during which NH₄ and PO₄ accumulate in the bottom water. Winter DIN and DIP pools in the surface layer loosely followed the bottom water dynamics.

Linkages between phytoplankton and the large scale changes in the Baltic Proper ecosystem were only weakly expressed. Spring dinoflagellate biomass increased with temperature, which Wasmund et al. (1988) explain by reduced deep water mixing during warm winters. The apparent lack of regime shifts in the phytoplankton community was potentially also caused by the shortness of the available time-series.

Figure 2: “Traffic light” plot for the Baltic Proper; time-series transformed into quintiles and sorted according to PC1; factor loadings for PC1 and PC2 are listed next to variable abbreviation. Click image to enlarge!

Gulf of Riga

The Gulf of Riga (Fig. 3) underwent similar changes in temperature as the Baltic Proper, but due to the absence of high saline bottom water, the effect on salinity conditions was smaller. Also nutrient loads followed a climate driven signal. Similar to the Baltic Proper, zooplankton development in spring benefited from higher temperatures. In the Gulf of Riga this led to increased herring recruitment and growth of the herring stock, which in turn caused a cascading decrease of summer zooplankton biomass (see also Kornilovs et al. 2004). In contrast to the Baltic Proper, summer phytoplankton concentrations and winter nutrient (DIP) pools were tightly related. The increase in winter DIP, decoupled from the load signal, is probably an effect of the long residence time of DIP (Savchuk 2002) in the Gulf of Riga. Internal loading from the bottom sediments, which have accumulated the surplus of past riverine inputs, therefore dominated the DIP dynamics in the Gulf. 

Figure 3: “Traffic light” plot for the Gulf of Riga; time-series transformed into quintiles and sorted according to PC1; factor loading for PC1 and PC2 next to variable abbreviation. Click image to enlarge!

Data

Metadata

This indicator fact sheets summarizes the outcome of the ICES/HELCOM/BSRP “Workshop on Developing a Framework for an Integrated Assessment for the Baltic Sea [WKIAB]”, held in Tvärminne (Finland), March 1 – 4, 2006. During the workshop, time series characterizing key components of the climatic, hydrographic, nutrient, and trophic system of the Baltic Sea were analyzed by principal component analysis to identify distinct temporal ecosystem states, their main properties and underlying driving factors. Detailed information on the workshop outcome can be found in ICES (2006).

Technical information

1. Data source 

Table 1. Time-series used in the meta-analysis of the Baltic Proper.

§ data series with gaps

* fish data from the following ICES assessment units: Sprat Subdivisions 22 to 32, Subdivisions 25 to 29 and 32 excl. Gulf of Riga, Flounder Subdivisions 24 & 25, Cod Subdivisions 25 to 32.

⁺ SST – Sea Surface Temperature, SSS – Sea Surface Salinity

^ø Zooplankton data are from the Gdansk Deep and Gotland Basin

Table 2. Time-series used in the meta-analysis of the Gulf of Riga.

Abbreviations: ICES – from ICES data centre; BED – from Baltic Environment Database, Stockholm University, LATFRA – Latvian Fish Resources Agency, SBFÖ – Swedish Board for Fisheries Öregrund; IFM – Leibniz Institute for Marine Science Kiel, IOW – Leibnitz Institute for Baltic Sea Research; SMHI – Swedish Meterological and Hydrological Institute, SMNH – Swedish Museum for Natural History; IAE – Institute of Aquatic Ecology, University of Latvia, FIMR – Finnish Institute of Marine Research, HELCOM – HELCOM load database, IMWM - Institute of Ecology and Water Management, Gdynia.

2. Description of data

Nutrient, hydrographic and phytoplankton data were taken from various national marine monitoring programs within HELCOM COMBINE, collected both in the ICES marine monitoring database, as well as by the Baltic Environment Database at Stockholm University. Zooplankton data was contributed from the Latvian Fisheries Resource Agency monitoring programme. Stocks, landing and recruitment of commercial fish species were used as published by the ICES fish stock assessments (ICES 2005). Tables 1 and 2 contain a detailed description of the data sources of the indicator time series for the Baltic Proper and the Gulf of Riga, respectively.

3. Temporal coverage

Most indicator time series used in the analysis covered the time period 1974 to 2004, but phytoplankton and chlorophyll a data for the Baltic Proper were available only from 1979. Tables 1 and 2 list the temporal coverage of each indicator time series for the Baltic Proper and the Gulf of Riga, respectively.

4. Methodology and frequency of data collection

Nutrient, hydrographic and phytoplankton time series were mostly collected according to the methods of the HELCOM COMBINE manual. Zooplankton data were derived from seasonal surveys by the Latvian Fisheries Resource Agency using a Judai-net (Möllmann et al. 2000).

5. Methodology of data manipulation

Datasets were analyzed by Principal Component Analysis (PCA). Missing values in the dataset were replaced by the mean values of the respective variable. To improve linearity between variables and reduce the relationship between the mean and the variance some of the variables were ln(x+1) transformed. Subsequently a standardized PCA based on the correlation matrix was performed on the transformed values. The occurrence of regime shifts - rapid changes in the datasets from one state to another - was subsequently examined by chronological clustering, using intermediate linkage clustering (Legendre et al. 1985, Legendre and Legendre 1998).

The ecosystem changes were further presented in “traffic-light plots” (Choi et al. 2005). The plots present the raw values of each variable, categorized into quintiles, which are assigned a specific colour. To detect systematic patterns in the time series, the variables are sorted according to their loadings along the first principal component.

Quality of information

1. Strength and weakness (at data level)

The indicator summarizes changes in time series characterizing a wide spectrum of ecosystem components and potential driving factors, ranging from climate and hydrography to nutrients, lower and upper trophic level species biomass. The analysis also includes direct anthropogenic forcing as for example nutrient loads and fishing pressure.

2. Reliability, accuracy, robustness, uncertainty (at data level)

The reliability of the regime shifts detected depends on the ability of the indicator time series used to depict key processes in the Baltic ecosystem. Further, the robustness of the analysis depends on the length of the time series in comparison to the time scale of processes, by which they are influenced, i.e. systematic patterns can only be detected if they occur within the ~ 30 year timeframe covered by the analysis. Further, the data analysis highlights consistent patterns in indicator time series, involving a large number of indicators but reveals processes impacting on single or few ecosystem components only in higher order principal components.

3. Further work required (for data level and indicator level)

Further work is required to refine the indicators used in the analysis. The raw data provided by the Baltic marine monitoring programs should be routinely aggregated into indicator time series that characterize key processes in the Baltic ecosystem, involving scientific expert groups. Future analysis should also expand the spatial coverage to further subsystems of the Baltic Sea.

References

Baumann, H., Hinrichsen, H.-H., Malzahn, A., Möllmann, C., Köster, F.W. and Temming, A. 2006. Sprat recruitment in the Baltic Sea: the importance of temperature and transport variability during the late larval and early juvenile stages. Can. J. Fish. Aquat. Sci. in press.

Choi, J. S., Frank, K. T., Petrie, B. D. and Leggett, W. C. 2005. Integrated ecosystem assessment of a large marine ecosystem: a case study of the devolution of the Eastern Scotian Shelf, Canada. Oceanography and Marine Biology: an Annual Review, 43: 47–67.

Hinrichsen, H.H., Möllmann, C., Voss, R., Köster, F.W. and Kornilovs, G. 2002. Bio-physical modelling of larval Baltic cod (Gadus morhua) survival and growth. Can. J. Fish. Aquat. Sci., 59: 1958-1873.

ICES. 2006. Report of the ICES/BSRP/HELCOM Workshop on Developing a Framework for Integrated Assessment for the Baltic Sea (WKIAB), 1-4 March 2006, Tvärminne, Finland. ICES CM 2006/BCC:09. 57 pp.

ICES. 2005. Report of the Baltic Fisheries Assessment Working Group (WGBFAS). ICES CM 2005/ACFM:19.

Köster, F.W., Möllmann, C., Neuenfeldt, S., Vinther, M., St. John, M.A., Tomkiewicz, J., Voss, R., Hinrichsen, H.H., Kraus, G. and Schnack, D. 2003. Fish stock development in the Central Baltic Sea (1976-2000) in relation to variability in the physical environment. ICES Mar. Sci. Symp., 219: 294-306.

Köster, F.W., Möllmann, C., Hinrichsen, H.-H., Tomkiewicz, J., Wieland, K., Kraus, G., Voss, R., MacKenzie, B.R., Schnack, D., Makarchouk, A., Plikshs, M. and Beyer J.E. 2005. Baltic cod recruitment – the impact of climate and species interaction. ICES J. Mar. Sci., 62: 1408-1425.

Laznik, M., Stalnacke, P., Grimvall, A., Wittgren, H.B. 1999. Riverine input of nutrients to the Gulf of Riga – temporal and spatial variation. Journal of Marine Systems 23:11-25.

Legendre, P, and Legendre, L. .1998.. Numerical Ecology, 2nd English Edition. Elsevier Science B.V., Amsterdam, 1998, XV + 853pp.

Legendre, P., Dallot, S., and Legendre, L. 1985. Succession of species within a community: Chronological clustering, with applications to marine and freshwater zooplankton. Am. Nat., 125: 257-288.

Möllmann, C., Kornilovs, G. and Sidrevics, L. 2000. Long-term dynamics of main mesozooplankton species in the Central Baltic Sea. J. Plankton Res., 22: 2015–2038.

Voss, R., Köster, F.W. and Dickmann, M. 2003. Comparing the feeding habits of co-occurring sprat (Sprattus sprattus) and cod (Gadus morhua) larvae in the Bornholm Basin, Baltic Sea. Fish. Res., 63: 97-111.

Wasmund, N. and Uhlig, S. 2003. Phytoplankton trends in the Baltic Sea. ICES J. Mar. Sci., 60 : 177-186.

Wasmund, N., Nausch, G., and Matthäus, W. 1998. Phytoplankton spring blooms in the southern Baltic Sea - spatio-temporal development and long-term trends. J. Plankt. Res., 20: 1099-1117.

¹ Hamburg University

² Latvian Institute of Aquatic Ecology

³ Finnish Institute of Marine Research

For reference purposes, please cite this indicator fact sheet as follows:
[Author’s name(s)], [Year]. [Indicator Fact Sheet title]. HELCOM Indicator Fact Sheets 2006. Online. [Date Viewed], http://www.helcom.fi/environment2/ifs/en_GB/cover/.

Last updated: 21.9.2006