ICES WGIAB REPORT 2010. Report of the ICES/HELCOM Working Group on Integrated Assessments of the Baltic Sea (WGIAB)

ICES WGIAB REPORT 2010 SCICOM STEERING GROUP ON REGIONAL SEA PROGRAMMES ICES CM 2010/SSGRSP:02 REF. SSGRSP, SCICOM Report of the ICES/HELCOM Working Group on Integrated Assessments of the Baltic Sea (WGIAB) 19 23 April 2010 ICES Headquarters, Copenhagen, Denmark

International Council for the Exploration of the Sea Conseil International pour l Exploration de la Mer H. C. Andersens Boulevard 44 46 DK-1553 Copenhagen V Denmark Telephone (+45) 33 38 67 00 Telefax (+45) 33 93 42 15 www.ices.dk info@ices.dk Recommended format for purposes of citation: ICES. 2010. Report of the ICES/HELCOM Working Group on Integrated Assessments of the Baltic Sea (WGIAB), 19 23 April 2010, ICES Headquarters, Copenhagen, Denmark. ICES CM 2010/SSGRSP:02. 94 pp. For permission to reproduce material from this publication, please apply to the General Secretary. The document is a report of an Expert Group under the auspices of the International Council for the Exploration of the Sea and does not necessarily represent the views of the Council. 2010 International Council for the Exploration of the Sea

ICES WGIAB REPORT 2010 i Contents Executive summary... 1 1 Opening of the meeting... 3 2 Adoption of the agenda... 3 3 Introduction... 3 4 Recent changes in the Baltic ecosystem update of the Integrated Status and Trend Assessments (ToR a)... 5 4.1 Methods for Integrated Status and Trend Assessments... 5 4.2 Summary of the results of the Integrated Status and Trend Assessments... 6 5 Indicator systems for ecosystem-based fisheries advice and management (ToRs d & e)... 9 5.1 Introduction... 9 5.2 Review of policy developments, with MSFD in focus... 9 5.3 HELCOM BSAP & EU MSFD indicators and WGIAB analyses... 11 5.4 Fuzzy-logic based indicator systems for the Baltic Sea ecosystem... 12 5.4.1 Ecosystem Based Assessment of Eastern Baltic Cod Developing a knowledge-based approach... 12 5.4.2 An indicator-based framework to advance ecosystem approach to fisheries in the Baltic Sea: recognizing potentials and limitations of fisheries management... 14 5.4.3 Assessment of the environment of the Baltic Proper: application of knowledge-base systems... 15 6 Biological Ensemble Modelling (ToR c)... 15 6.1 Introduction... 15 6.2 Modelling indicators... 15 6.3 Coupling ecological and economic models... 21 6.3.1 Introduction... 21 6.3.2 Background... 21 6.3.3 Ecological simulations of landings... 22 6.3.4 Economic-ecological multi-species modelling within WGIAB & EU 7 th FP project FACTS... 22 6.3.5 Impact of using different cost functions... 23 6.3.6 Production function for the German Saithe fishery... 25 6.3.7 Assessing the spatial dynamics of Small-Scale Coastal Fisheries of the Baltic Sea- A Geographic Information System based approach... 26 6.4 Further development of BEMA: a potential strategy for the use of ecosystem modelling within the Baltic Sea assessment and ecosystem-based advice... 26

ii ICES WGIAB REPORT 2010 7 Development of ecosystem-based advice for the Baltic Sea (joint meeting with WGBFAS contribution to the workplan of SSGRP) (ToR f)... 28 8 Implications of WGHAME results for WGIAB (ToR b)... 30 9 Summary of other presentations given at the meeting... 31 9.1 Change in community structure of coastal fish and zoobenthos assemblages - local, regional or global drivers?... 31 9.2 Changes in the Limfjord ecosystem: has a regime shift taken place in recent times?... 31 9.3 HELCOM Monitoring, Assessment and Indicator development... 32 9.3.1 References... 33 10 References... 34 Annex 1: List of participants... 38 Annex 2: Agenda... 41 Annex 3: WGIAB terms of reference for the next meeting... 44 Annex 4: Integrated Status and Trend Assessments per sub-basin... 46

ICES WGIAB REPORT 2010 1 Executive summary WGIAB was setup in 2007 as a forum for developing and combining ecosystem-based management efforts for the Baltic Sea. The general approach of WGIAB is to assess the state and development of the different Baltic Sea sub-ecosystems considering all trophic levels and the impact of climate, fisheries and eutrophication. WGIAB therefore is intended to serve as a counterpart and support for the ICES Baltic Fisheries Assessment Working Group (WGBFAS), but also to support related HELCOM assessment efforts such as HELCOM BIO and HELCOM FISH. WGIAB has given itself 3 main tasks: 1 ) to conduct holistic ecosystem assessments based on large multivariate datasets; 2 ) to consider the use of ecosystem modelling in the assessment framework; and 3 ) to develop adaptive management strategies for the different Baltic Sea ecosystems. During the meetings in 2007 and 2008 WGIAB concentrated on collection and analyses of large multivariate datasets. This effort resulted in ecosystem assessment for 7 subsystems, i.e. the Sound, Central Baltic Sea, Gulf of Riga, Gulf of Finland, Bothnian Sea, Bothnian Bay and a coastal ecosystem (ICES, 2008a). These ecosystem assessments (above Task 1) demonstrated dramatic changes (i.e. regime shifts) during the last 3 decades on all trophic levels of the ecosystems related to climate variability and human exploitation (Möllmann et al. 2006, Möllmann et al. 2009, Lindegren et al. 2010, Blenckner et al. in prep.). A major product of these analyses is an ICES Cooperative Research Report on Integrated ecosystem assessments of seven Baltic Sea areas covering the last three decades (Diekmann and Möllmann, 2010) which will be published during the next month. During the 2010 meeting time series of three out of the seven Baltic sub-systems were updated, including the Bothnian Sea (BoS), the Gulf of Riga (GoR), and the Central Baltic Sea (CBS). The BoS data set was thereby largely revised. The CBS was analysed in four different ways: First, the dataseries from previous years were simply updated until 2008. Then, an alternative analysis was conducted accounting for newly available zooplankton data. Finally, to account for the observed changes in the spatial distribution of the sprat stock, separate analyses were conducted for the Bornholm basin (SD 25) and the Gotland basin (SD 28). Eventually the Kattegat was included as a new sub-system to the datasets and ecosystem analyses conducted by WGIAB. The analyses of the now 8 sub-systems analysed confirmed pronounced structural changes in the last two to three decades, related to climate, fisheries and eutrophication. Another major activity of WGIAB (in relation to above Task 2) is the Biological Ensemble Modelling Approach (BEMA) which was started during the 2009 meeting. WGIAB performed comparative analyses of a set of cod population dynamics, multispecies and food web models following the ensemble approach in climate research. Different models are forced with the same scenarios (e.g. fishing and future climate) and their projections are collected in an ensemble. By this WGIAB evaluated alternative fisheries management scenarios for cod and sprat under alternative scenarios of further climate change. The main aim of the modelling in the 2010 WGIAB meeting was to further BEMA as a tool for the final step in an IEA: management strategy evaluation on ecosystem level. This entailed developing (i) simulations of indicator

2 ICES WGIAB REPORT 2010 beyond those covered in the 2009 meeting, (ii) coupling of economic analyses to the ecological simulations, and (iii) the deterministic BEMA into stochastic simulations providing probabilities of alternative ecological outcomes. While continuing and further developing data-based ecosystem analyses and BEMA, beginning with the 2010 meeting WGIAB started to tackle above Task 3). WGIAB will thereby follow the international development towards a broader understanding of Integrated Ecosystem Assessments (IEAs) developed by NOAA (Levin et al. 2009, Tallis et al. 2010). The central WGIAB-activities, i.e. the multivariate analyses on ecosystem status and trends as well as BEMA, already cover some aspects of the IEAprocess. During the 2010 meeting WGIAB conducted considerable work on developing indicators and indicator systems for ecosystem-based management. This included (i) reviewing the indicator systems of the HELCOM Baltic Sea Action Plan and the EU Marine Strategy Directive and relating them to WGIAB models and data, (ii) developing indicators systems to support ecosystem-based assessment of the Baltic fish stocks and the ecosyste, and (iii) projecting simple indicators using the BEMA as well as network indices using individual food-web models. A further activity in 2010 was a back-to-back meeting with WGBFAS with the aim to discuss ways to include environmental information generated by WGIAB (and other Baltic expert groups) into fish stock advice and management. By presentations, exemplifying ways towards more ecosystem-based approaches, the discussion between both groups on ways to integrate environmental information into the assessment routine was initiated. In 2011 WGIAB intends to continue its work along the above described lines, conducting (i) Integrated Status and Trends Assessments for the different Baltic Sea subsystems, (ii) continue developing the Biological Ensemble Modelling (BEMA), (iii) continue the work on indicator selection, testing, and target level evaluation, (iv) further develop and promote ecosystem-based advice for Baltic Sea fish stocks. A further task for 2011 (and potentially intersessional work) is a WGIAB contribution to the suggested Baltic Sea ecosystem observing system.

ICES WGIAB REPORT 2010 3 1 Opening of the meeting The Co-Chairs Christian Möllmann (CM), Anna Gårdmark and Thorsten Blenckner welcomed the participants (Annex 1) of the meeting. CM introduced the history and the goals of WGIAB, the state of the different tasks to be conducted by the group and the purpose of this meeting (see section 3). The meeting has been given the following Terms of References: The ICES/HELCOM Working Group on Integrated Assessments of the Baltic Sea (WGIAB), chaired by Christian Möllmann, Germany; Anna Gårdmark, Sweden; and Thorsten Blenckner, Sweden, will meet at the ICES Headquarters, Copenhagen, Denmark, 19 23 April 2010 to: a) Update the Integrated Ecosystem Assessments (IEAs) for the different Baltic Sea subsystems, conducting an IEA for the Western Baltic, as well as a spatially disaggregated IEA for the Central Baltic to investigate the effects of spatial variability in the relative distribution of the cod and clupeid stocks; b) Review the report of WGHAME and discuss implications for the work of WGIAB; c) Continue the Biological Ensemble Modelling (BEMA) and finalize a proposal for a strategy of the use of ecosystem modelling within the Baltic Sea assessment and ecosystem-based advice; d) Develop subsystem specific indicator systems for ecosystem-based fisheries advice and management based on the results of IEAs and BEMA; e) Evaluate target levels of the HELCOM Baltic Sea Action Plan and the HEL- COM BIO set of indicators and assessment system in relation to the WGIAB data analyses; f) Discuss a strategy on delivering ecosystem-based advice for the Baltic Sea in a joint meeting with WGBFAS and as a contribution to the workplan of SSGRP. WGIAB will report by 12 May 2010 (via SSGRSP) for the attention of SCICOM. The participation in WGIAB increased considerably during its life-time (12 participants in 2007, 23 in 2008, 28 in 2009) with participants from 8 countries. The initial confirmed membership for the 2010 meeting was c. 35. Due to the problems with the air traffic due to the volcanic activity on Iceland about 1/3 of the participants had to cancel their participation which affected parts of the intended work (see sections 3, 4, 6 and 7). Eventually 23 participants from 4 countries participated in the 2010 WGIAB meeting (see Annex 1) 2 Adoption of the agenda CM introduced the agenda which was shortly discussed, adjusted and finally adopted by the participants. 3 Introduction WGIAB was setup in 2007 as a forum for developing and combining ecosystem-based management efforts for the Baltic Sea. The general approach of WGIAB is to assess the state and development of the different Baltic Sea sub-ecosystems considering all

4 ICES WGIAB REPORT 2010 trophic levels and the impact of climate, fisheries and eutrophication. WGIAB therefore is intended to serve as a counterpart and support for the ICES Baltic Fisheries Assessment Working Group (WGBFAS), but also to support related HELCOM assessment efforts such as HELCOM BIO and HELCOM FISH. WGIAB has given itself 3 main tasks: 1 ) to conduct holistic ecosystem assessments based on large multivariate datasets; 2 ) to consider the use of ecosystem modelling in the assessment framework; and 3 ) to develop adaptive management strategies for the different Baltic Sea ecosystems. During the meetings in 2007 and 2008 WGIAB concentrated on collection and analyses of large multivariate datasets. This effort resulted in ecosystem assessment for 7 subsystems, i.e. the Sound, Central Baltic Sea, Gulf of Riga, Gulf of Finland, Bothnian Sea, Bothnian Bay and a coastal ecosystem (ICES, 2008a). These ecosystem assessments (above Task 1) demonstrated dramatic changes (i.e. regime shifts) during the last 3 decades on all trophic levels of the ecosystems related to climate variability and human exploitation (Möllmann et al. 2006, Möllmann et al. 2009, Lindegren et al. 2010, Blenckner et al. in prep.). A major product of these analyses is an ICES Cooperative Research Report on Integrated ecosystem assessments of seven Baltic Sea areas covering the last three decades (Diekmann and Möllmann, 2010) which will be published during the next month. Another major activity of WGIAB (in relation to above Task 2) is the Biological Ensemble Modelling Approach (BEMA) started during the 2009 meeting. WGIAB performed comparative analyses of a set of cod population dynamics, multi-species and food web models following the ensemble approach in climate research. Different models are forced with the same scenarios (e.g. of future climate development) and their projections are collected in an ensemble. By this WGIAB evaluated alternative fisheries management scenarios for cod and sprat under alternative scenarios of future climate change. Results of the modelling exercise were presented on 3 major conferences: 1) ICES-GLOBEC Symposium on Marine Ecosystems: from function to prediction (22 26 June 2009, Victoria, British Columbia, Canada) by Christian Möllmann; 2) ICES/PICES/UNCOVER Symposium 2009 on Rebuilding Depleted Fish Stocks Biology, Ecology, Social Science and Management Strategies (3 6 November 2009, Warnemünde/Rostock, Germany) by Anna Gårdmark; and 3) ICES/PICES/FAO Symposium on Climate Change Effects on Fish and Fisheries: Forecasting Impacts, Assessing Ecosystem Responses, and Evaluating Management Strategies (26 29 April 2010, Sendai, Japan) by Martin Lindegren. While continuing and further developing data-based ecosystem analyses (see section 4) and BEMA (see section 6), starting with the 2010 meeting WGIAB intends to tackle above Task 3). WGIAB will thereby follow the international development towards a broader understanding of Integrated Ecosystem Assessments (IEAs) developed by NOAA (Levin et al. 2009, Tallis et al. 2010) The suggested IEA process covers five steps (modified from Koslow et al. 2009): 1 ) An initial scoping will identify management objectives, ecosystem attributes of concern, and relevant ecosystem stressors. 2 ) Indicators that reflect ecosystem attributes and stressors specified in the scoping process will be developed and tested. These must be linked objectively to decision criteria.

ICES WGIAB REPORT 2010 5 3 ) A hierarchical risk analysis will explore the susceptibility of an indicator to natural or human threats, as well as its resilience, the ability of the indicator to return to its previous state after being perturbed. 4 ) Results from the risk analysis for each ecosystem indicator are integrated in an ecosystem assessment, which quantifies the overall status of the ecosystem relative to historical status and prescribed targets. 5 ) The final phase of the IEA is an evaluation of the potential of different management options to influence ecosystem status, using ecosystem models and a formal Management Strategy Evaluation (MSE) Some of the recent WGIAB activities already cover some aspects of the IEA-process, i.e. the multivariate analyses on ecosystem status and trends can contribute data and knowledge to step 1, 2 and 4, while the BEMA is a prerequisite for steps 2, 3 and 5. Originally a main aim of the 2010 meeting was to discuss and implement risk analysis (step 3) using the models developed within WGIAB. The invited expert however could not join the meeting due to the air traffic problems, but provided written input (see section 6). During the 2010 meeting WGIAB conducted considerable work on developing indicators and indicator systems for ecosystem-based management (see section 5). This included (i) reviewing the indicator systems of the HELCOM Baltic Sea Action Plan and the EU Marine Strategy Directive and relating them to WGIAB work, (ii) developing indicators systems to support ecosystem-based assessment of the Baltic fish stocks and the ecosystem (see also section 7), and (iii) projecting simple indicators using the BEMA as well as network indices using individual food-web models (see section 6). In 2011 WGIAB intends to continue its work along the above described lines, conducting (i) Integrated Status and Trends Assessments for the different Baltic Sea subsystems, (ii) continue developing the Biological Ensemble Modelling (BEMA), (iii) continue the work on indicator selection, testing, and target level evaluation, (iv) further develop and promote ecosystem-based advice for Baltic Sea fish stocks (see Annex 3). A further task for 2011 (and potentially inter-sessional work) is a WGIAB contribution to the suggested Baltic Sea ecosystem observing system. 4 Recent changes in the Baltic ecosystem update of the Integrated Status and Trend Assessments (ToR a) 4.1 Methods for Integrated Status and Trend Assessments During the WGIAB meeting in 2008, integrated analyses of ecosystem state and trends were conducted for 7 subregions of the Baltic Sea (ICES, 2008a): 1) The Central Baltic Sea (CBS), encompassing the 3 deep basins, Bornholm Basin, Gdansk Deep and Gotland Basin; 2) the Sound (ÖS); 3) the Gulf of Riga (GoR); 4) the Bothnian Sea (BoS); 5) the Bothnian Bay (BOB); 6) a coastal site in Sweden (COAST); and 7) the Gulf of Finland (GoF). For each area a multitude of time-series were collected providing information about climate, hydrography, nutrients, phytoplankton, zooplankton, fish and fisheries, plus, where available, data for benthos and top predators like seals. The type and number of variables available for each system differ, but were balanced according to drivers and response variables and included to the best knowledge of the group all available key components describing each food web. A description of the time series and data sources is given in the Annex of ICES (2008a).

6 ICES WGIAB REPORT 2010 Generally the intention of WGIAB is to update the different ecosystem analyses regularly depending on data availability and system coordinators time-schedule. During the 2009 meeting, WGIAB managed to update and analyse the data series of four subsystems, i.e. CBS, GoR, GoF and COAST (for info on sub-systems see ICES, 2008a and Annex 4). During this year s meeting updates were performed for the CBS, GoR and the BoS (see below and Annex 4). Like in previous years we followed the same, although slightly reduced analytical strategy by: 1 ) creating a traffic light plot (Link et al., 2002) to ascertain the status and temporal development of the system and its variables. Quintiles of metrics are colour-coded and variables are sorted according to their subsequently derived PC1 loadings; 2 ) using Principal Component Analysis (PCA) based on the correlation matrix of all variables and taking the PC-scores of the first and second axis to visualise the time-trajectory of the system; and 3 ) applying Chronological Clustering (Legendre et al., 1985) to detect sudden changes in the multivariate dataset. The method was used on the normalised data with the Euclidean distance measure, a connectedness level of 0.5, and different α values. To facilitate the analytical procedure and make the methods available to other WGIAB members an R-code was written inter-sessionally, based on the freely available R and related packages as language for statistical computing (R Development Core Team, 2007). The results and graphs, with the exception of Chronological Clustering, which was performed with the help of the computer program BRODGAR (www.brodgar.com), were produced by the respective codes. Further, the analyses for the CBS were refined during the 2010 meeting. In addition to the regular update (until 2008, Analysis CBS-1; see Annex 4), alternative analyses were conducted accounting for the separate time series collected within the Bornholm Basin (SD 25) and the Gotland Basin (SD 28) (Analysis CBS-2, CBS-3 a & b; see Annex 4). Another aim of WGIAB is to include more systems (especially west of the Bornholm Basin) into the ecosystem analyses. During this year s meeting a dataset could be assembled for the Kattegat and first analyses were performed. An inventory for more western Baltic system analyses has been continued, and respective analyses are envisaged for the 2011 meeting. The next section provides a short overview on the results of the ecosystem analyses conducted during this year s meeting. Detailed descriptions of the analyses are provided in Annex 4. 4.2 Summary of the results of the Integrated Status and Trend Assessments Time series of three out of seven Baltic sub-systems were updated this year, including the Bothnian Sea (BoS), the Gulf of Riga (GoR), and the Central Baltic Sea (CBS). The variable selections from previous years were retained as far as possible (but see below for BoS). However, data exploration was redone for each variable and some variables were not transformed for the PCA analysis as they had been last year. Furthermore, the Baltic Sea Index (BSI, Lehmann et al., 2002) used in previous years was replaced by winter NAO indices because of its better availability.

ICES WGIAB REPORT 2010 7 The data collected for BoS were revised. The analyses were based on a shortening of the time-series because of large amount of missing data in the first years. To improve the quality of the data further, variables on hydrography, nutrients and plankton were recompiled using a stricter station selection. Further, in the data revision all coastal variables and 8 off-shore variables were excluded and data the abundance of the polychaetae Marenzelleria viridis, sprat biomass, concentrations of oxygen (bottom) and NAO were added in this year s analyses. In addition to the standard analyses, the sensitivity of the results to the inclusion/exclusion of these additional variables was assessed. This year the CBS was analysed in four different ways: First, the dataseries from previous years were simply updated until 2008 (Analysis CBS-1; see Annex 4). Then, an alternative analysis was conducted accounting for newly available zooplankton data. The original data set contained zooplankton time-series only from the Gotland Basin (SD 28), while now a time-series from the Bornholm Basin (SD 25) were available. Hence the new analysis (Analysis CBS-2) included these data now matching with the spatial resolution of hydrographic and phytoplankton variables in the dataset. Finally, to account for the observed changes in the spatial distribution of the sprat stock, separate analyses (Analysis CBS-3 a & b) were conducted for SD s 25 & 28 using also spatially disaggregated data for sprat, herring and cod that were derived during the meeting (see Annex 4, Analysis CBS-3 for details). For the first time, time series from the Kattegat area were analysed this year. Even though a more in-depth analysis is necessary in the future the results presented in Annex 4 give a first overview about the temporal development of this transition area between the North Sea and the Baltic Sea. Table 4.2.1 summarizes the abrupt changes in the multivariate time series observed in the ecosystem analyses conducted by WGIAB since 2008. In total 8 Baltic sub-systems have been investigated, not considering the spatial spilt of the CBS into SD25 and SD 28. Overall all sub-systems displayed pronounced structural changes in the last two to three decades, related to climate, fisheries and eutrophication. Regime shifts were identified in all multivariate datasets. The major period of reorganisation in all Baltic systems was at the end of the 1980s (between 1987 and 1989), when the strongest and most persistent changes were observed. Although the main drivers of this change were different between sub-systems, sudden increases in temperature and decreases in salinity were observed throughout. Both these variables are influenced by largescale atmospheric processes illustrated by the Baltic Sea Index (BSI), a regional homologue to the North Atlantic Oscillation index (NAO) (Lehmann et al., 2002). The change from a generally negative to a positive BSI in the late 1980s was associated with more frequent westerly winds, warmer winter and eventually a warmer climate over the area. Further, the absence of major inflow events has been hypothesized to be related to the high NAO period (Hänninen et al., 2000). An indication of this is that only two major inflows to the Baltic Sea have been recorded during the high BSIperiod since the late 1980s. Several regions underwent a structural change also during the middle of 1990s, probably related to the major inflow in 1993. Weak shifts have been observed for the 1970s in CBS and COAST, when using the longest time series. However, because of many missing values in the 1970s this observations needs to be interpreted with caution. In contrast, increasing support is shown for a reorganization at the beginning of the present century, with shifts detected in CBS, COAST, GoF and BoS.

8 ICES WGIAB REPORT 2010 The new analytical runs performed this year and in 2009 showed that the major shift identified at the end of the 1980s persisted throughout. The same holds for all other shifts identified for the Gulf of Riga and the Gulf of Finland. However, the analysis of the Central Baltic Sea and the Swedish coast datasets gave slightly different results. In the Central Baltic Sea the shift in the mid-1990s is not always detectable (when conducting the spatial resolution: only in CBS-28), whereas a reorganisation of the system is already indicated in the first half of the 1980s. This earlier abrupt change is no longer identifiable if the dataset is shortened, i.e. the time-series do not start in 1974 but in 1979. Because of missing phytoplankton data from 1974 1978 and the necessary exchange of missing values with the averages of the four nearest data points, the outcome of the first analysis is definitely influenced, and this can indirectly cause a shift in the multivariate dataset. The results of the previous and updated analysis of the Swedish coast dataset did also differ: The first shift in 1976 1977 is no longer detectable. Furthermore, an additional shift is identified in the late 1990s. In the newly available time series from the Kattegat, starting in 1982, a regime shift was observed in 1988 1989 but not in more recent years. Table 4.2.1. Summary of the WGIAB-ecosystem analyses since 2008. Given are years with sudden changes in the multivariate dataset ( Shift ) detected by Chronological Clustering; Subsystem abbreviations: K Kattegat, TS The Sound, CBS Central Baltic, CBS-25 Subdivision 25, CBS- 28 Sub-division 28, COAST Swedish coastal system, GOR Gulf of Riga, GOF Gulf of Finland, BOS Bothnian Sea and BOB Bothnian Bay. System K TS CBS CBS-25 CBS-28 COAST Meeting year 2010 2008 2008 2009 2009 2010 (CBS-1) 2010 (CBS-2) 2010 2010 2008 2009 Period studied 1982-2008 1979-2005 1974-2006 1974-2007 1979-2007 1975-2008 1975-2008* 1976-2008 1976-2008 1971-2006 1971-2008 Shift 1 1976/77 1976/77 Shift 2 1984/85 1980/81 1980/81 1980/81 1984/85 Shift 3 1988/89 1987/88 1987/88 1987/88 1987/88 1987/88 1987/88 1987/88 1987/88 1988/89 Shift 4 1994/95 1993/94 1993/94 1994/95 1993/94 1997/97 Shift 5 2004/05 2002/03 2003/04 2004/05 * additional zooplankton data in SD 25

ICES WGIAB REPORT 2010 9 Table 4.2.1. continued System GOR GOF BOS BOB Meeting year 2008 2009 2010 2008 2009 2008 2010 2008 Period studied 1973-2006 1973-2007 1973-2008 1979-2007 1979-2008 1979-2006 1983-2008 1979-2006 Shift 1 Shift 2 1982/83 Shift 3 1988/89 1988/89 1988/89 1988/89 1988/89 1988/89 1988/89 1987/88 Shift 4 1997/98 1997/98 1997/98 1995/96 1995/96 1993/94 Shift 5 2002/03 2002/03 1999/00 5 Indicator systems for ecosystem-based fisheries advice and management (ToRs d & e) 5.1 Introduction During the 2010 meeting WGIAB conducted considerable work on developing indicator an indicator systems for ecosystem-based management. This included (i) reviewing the indicator systems of the HELCOM Baltic Sea Action Plan (BSAP) and the EU Marine Strategy Framework Directive (MSFD) and relating them to WGIAB work (see below sections 5.2 & 5.3), (ii) developing indicators systems to support ecosystem-based assessment of the Baltic fish stocks and the ecosystem (see below sections 5.4 and 7), and (iii) projecting indicators derived by an ECOSIM foodweb model (see section 6). The main of these exercise were to contribute to the expertise and databases hold by WGIAB to the developments within BSAP and MSFD as well as to further develop ecosystem-based fisheries management advice and a full IEA for the Baltic Sea. 5.2 Review of policy developments, with MSFD in focus As a first step WGIAB reviewed the recent developments related to the HELCOM BSAP and the EU MSFD. Both represent recent policy developments relevant for the ongoing and future work of WGIAB. The HELCOM Baltic Sea Action Plan (HEL- COM, 2007) is a multilateral Ministerial Declaration in which the HELCOM contracting parties, coastal country governments and the European Commission, commit to actions to achieve a number of agreed ecological objectives and, eventually, a Baltic Sea in good environmental status by 2021. The BSAP is explicitly based on the Ecosystem Approach and includes a number of initial targets pertaining to HELCOM s four main themes, i.e., eutrophication, hazardous substances, biodiversity and nature conservation and maritime activities, as well as indicators to measure progress toward the commitments. Ongoing activities to implement the BASP, e.g. on the preliminary indicators, is mainly carried out in the HELCOM working groups responsible for the implementation of the BSAP. For example, it is the task of the HELCOM WG MONAS to develop a core set of HELCOM indicators that support the elaboration of indicatorbased thematic assessments, holistic assessments, and implementation of the HEL- COM Baltic Sea Action Plan (BSAP), by allowing the evaluation of progress towards

10 ICES WGIAB REPORT 2010 HELCOM Ecological Objectives agreed upon in the BSAP (http://www.helcom.fi/groups/monas/en_gb/monas_main/). The MSFD (EC 2008) was adopted and entered into force in 2008. It sets up a comprehensive list of ecological descriptors and characteristics, pressures and impacts that are to be used to assess the environmental status of European marine waters, and to elaborate marine strategies, including programmes of measures to achieve or maintain Good Environmental Status (GES) in those waters by 2020. The Baltic Sea (excluding the Kattegat) is one of the Marine regions defined by the MSFD that Member States (MS), when implementing their obligations in the waters of their jurisdiction, should take due account to. MS shall cooperate to ensure that measures required are coherent and coordinated across the marine region. Therefore, MS shall also use existing regional institutional cooperation structures, like HELCOM, to achieve this coordination. To achieve GES, MS have to: assess the state of their marine waters, jointly determine what GES means for these waters in each marine region, establish a series of environmental targets, establish and implement a monitoring programme to measure progress, devise programmes of measures in order to achieve a high level of protection of the marine environment (=GES) The MSFD on the EU level formulates only general aims and principles, specific measures are not to be decided. The MSFD will have to be made operational and implemented by the MS at the level of the marine regions. However, guidelines and deadlines towards the development of marine strategies by MS are laid out. Thus, MS are to be provided with criteria and methodological standards as to ensure consistency and to allow comparison between marine regions of the extent to which GES is being achieved. At its last meeting, WGECO (ICES, 2010), had as one Term of Reference to assess the development of integrated ecosystem assessments, in particular focussing on how assessment will be used for the MSFD(...). WGECO (ICES, 2010) concluded: ( ) for implementing the marine strategy framework directive, no integrated ecosystem assessment (IEA) is formally required. However, the necessary assessments are expected to include an explicit description of the relationships between pressure and state, multiple impacts and socio-economic aspects. These assessments should also be able to provide the basis for developing marine strategies including programmes of measures. Assessments sufficient to meet these needs will have to have many of the characteristics reviewed for sound integrated ecosystem assessments. The guidance for conducting sound IEAs, particularly the detailed guidance on how to choose indicators, will be invaluable guidance for doing these tasks in support of the MSFD. Moreover, by following a common approach ( ), the possibility for consistency and comparability of assessments across marine regions is increased greatly. WGECO developed a framework to address the development of IEAs in a consistent manner. This Report also considers how the diversity of marine ecosystems, uses, socioeconomic settings and availability of data across marine regions, means that rigid methodological guidance on setting assessment benchmarks cannot be expected to be an appropriate strategy to achieve consistency among assessments. Rather, the con-

ICES WGIAB REPORT 2010 11 sistency is achieved by the functional equivalence of the elements assessed, indicators chosen and reference levels established. In the WGECO (ICES, 2010) report, it is explained: how ecosystem elements and indicators can be considered functionally equivalent when they are appropriate for measuring status of a pressure, structural or functional property or process that is of similar ecological significance across ecosystems, even if the exact indicators or properties differ across ecosystems, and how reference levels can be considered functionally equivalent if they reflect the same level of sustainability, or risk of serious harm, across ecosystems, even if the value of the indicator that reflects this level varies across ecosystems. WGECO (ICES, 2010) developed a framework through which Member States can assess good environmental status (GES). It lays out six necessary steps (i) how to evaluate the list of ecosystem components required, (ii) how to evaluate the list of pressures and drivers required, (iii) how to identify the key interactions between ecosystem components and pressures, (iv) how to select indicators for those key interactions identified in (iii), (v) how to set reference points for these indicators, and (vi) how to combine information across indicators at various levels of integration. It also includes two additional additional pieces of technical guidance. The first is an approach for setting ecologically consistent reference levels for pristine conditions, in the few Descriptors that the MSFD implies should not be impacted as opposed to being used sustainably. The second is for step (vi), with guidance both on analytical/technical aspects of integration of information across indicators, and on aspects of the scientific processes necessary and appropriate for such integration. WGIAB decided to contribute to the further identification of indicators for the Baltic, based on its broad experience and in cooperation with relevant HELCOM working groups, at its next meeting in 2011. During this year's meeting a first evaluation on how the databases and models developed and hold by WGIAB for the Central Baltic Sea can be used to define final Baltic-specific suites of indicators (see section 5.3). 5.3 HELCOM BSAP & EU MSFD indicators and WGIAB analyses To contribute to the indicator development for the Baltic Sea, WGIAB investigated how the datasets developed and hold by the group can contribute and help specify the HELCOM and EU MSFD indicators systems. Therefore candidate HELCOM and EU MSFD indicators were matched with the time series of ecosystem data available to WGIAB. The product of this exercise is a MS EXCEL which cannot be presented in this report due to its size, but is available through the co-chairs of WGIAB. It was furthermore assessed whether an indicator could be calculated from the information available to WGIAB. In addition the indicators were matched with the 8 models used in the Biological Ensemble Modelling approach developed in WGIAB (see section 6). Here as well, it was assessed whether an indicator could be calculated from a model. It is important to note that the focus of this exercise was on the individual indicator time series and on the model outputs. That means that the cells of the table that contain data required to parameterize a (ecosystem-) model are not ticked when an indicator can be calculated from the model output, because this would result in an incomplete list of data needed to parameterize a model. Analogous to that, any indicator that requires several time series to be calculated, like a diversity index or a mean size of fish in a fish assemblage, does not lead to a ticked cell in a WGIAB time

12 ICES WGIAB REPORT 2010 series with information on single species or species groups. These ecosystem-, or community level indicators require the combined information from ecosystem surveys and their calculation is naturally located in the respective ICES groups planning this surveys like WGBIFS. The same holds true for other areas of expertise like all cod fisheries related indicators which have their home in WGBFAS and the contaminant/hazardous substance related indicators in their respective groups. Based on this exercise, WGIAB intends based on its expertise, data bases and models to further develop appropriate suites of indicators for the different sub-systems of the Baltic Sea. This will include recommendations which indicators should be measured and how. This will eventually contribute to a recommendation for the suggested Baltic Sea ecosystem observing system. 5.4 Fuzzy-logic based indicator systems for the Baltic Sea ecosystem To complement the present advice for Baltic fish stocks with environmental information, WGIAB reviewed three approaches using knowledge-based fuzzy-logic approaches. The three approaches are incremental with respect to the level of the target for ecosystem-based advice and management. The first approach is presently developed specifically for Eastern Baltic cod, while the second considers all the three main fish stocks in the Central Baltic, i.e. cod, herring and sprat. The third approach considers in addition to fisheries also other sectors affecting the Baltic Sea, hence represents an approach towards a full Ecosystem Approach. All approaches are shortly outlined below. 5.4.1 Ecosystem Based Assessment of Eastern Baltic Cod Developing a knowledge-based approach (Jens Floeter, Rabea Diekmann, Michele Casini & Christian Möllmann) The ecosystem approach to fisheries (EAF) has often been defined (e.g. FAO 2003), conceptually discussed (Garcia et al. 2003), categorized (e.g. Morishita 2008) and also implemented in international regulations like the European CFP. However, one of the main tasks is to translate this generic EAF framework into an operational one at regional (e.g. ecosystem or fishery) level by adapting the vague principles and conceptual goals to the reality of specific situations (Garcia & Cochrane 2005). The key elements in supporting the EAF are assessment, monitoring, and scientific research. These elements provide a sound basis for identifying ecological and associated operational objectives, selecting indicators, and identifying reference points (ICES, 2005). Furthermore, Evaluating the Ecosystem Status is one of the preconditions for management within the EAF and requires the description of the ecosystem, based on knowledge of ecosystem structure, function, and environmental quality (ICES 2005, Diekmann & Möllmann 2010). The goal of this approach is to apply the conceptual principles of the EAF to a specific stock assessment and thereby create a prototype of an Ecosystem Based Assessment for a single fish stock. As an example the Eastern Baltic Cod stock was chosen and the explicit task was to couple the XSA assessment results (ICES, 2009b) and the broader ecosystem knowledge of ICES WGIAB. Thereby any interpretation of the broader ecosystem knowledge was done with the focus on the Eastern Baltic Cod stock. This study is work-in-progress and large parts exist only on a conceptual level. The study was based on selected time series of ecosystem data assembled by WGIAB and as a method we used the software NetWeaver2 (Rules of Thumb, inc.), because it is

ICES WGIAB REPORT 2010 13 capable of using incomplete and qualitative data in its knowledge base development system with object-oriented fuzzy-logic networks. The approach was to first build an indicator based qualitative recruitment model. This model was set up with the goal to make a qualitative prediction of cod recruitment (age 2, XSA, 1976 2008) from environmental indicators and stock size. The following indicators were included in the model: Cod SSB in SD25-29+32from XSA, Cod Reproductive Volume, depth of 11-psu isoline in the Gotland basin, Biomass of Pseudocalanus acuspes, mean weight at age 5 of Eastern Baltic Cod, Sprat SSB, and the PC-1 Indicator from the IA of the central Baltic Sea conducted at the WGIAB meeting in 2009 (ICES, 2009a). Qualitative categorical linkages between the indicator values and the cod recruitment levels (shifted 2 years back to the year of birth of the cod cohort) have been defined and transformed into fuzzy relations. Cod recruitment levels were defined as below median, above median and above the 75% percentile. Then a number was calculated from the combination of the indicators for each year. Next, a qualitative categorical linkage was obtained between this number and the recruitment of cod in every year. This number then represents the status of the ecosystem in terms of being favourable or not for cod recruitment. When comparing the categorised XSA recruitment with the recruitment categorised by the number, a success of ca. 75% was obtained for the time series from 1976 2008. Trials with different weighting factors for the indicators showed that this success rate can be increased. After having established a qualitative cod recruitment model and calculating a number that represents the status of the ecosystem in terms of being favourable or not for cod recruitment in every year the next step is to extend this model to include other areas of knowledge important for an Ecosystem Based Assessment of Eastern Baltic Cod. This first includes a set of indicators which reflect the quality of the fish stock assessment carried out by WGBFAS. These indicators will be quantitative and qualitative and require expert knowledge from WGBFAS. A close cooperation has been agreed already and the approach will be further developed by both WGBFAS and WGIAB. Second, it is planned to include a set of indicator that stand for the performance of the EU management plan of the Eastern Baltic Cod stock, e.g., an indicator for the trend (last 5 years) in F towards the target F, or an indicator whether F is already < 0.3. Third, it is under discussion to include other areas affected by the cod management system, like other stocks impacted by cod or economic sustainability as well as social well being. The next steps will contain the following tasks: (i) further refine and define the suite of indicators and their parameterisation, (ii) define weighting factors for indicators, also with the help from statistical model outputs. Once the final suite of indicators of all areas has been included in the knowledge based approach, we will be able to calculate a new number for each year, representing the: status of the ecosystem in terms of being favourable or not for cod recruitment performance of the EU Fisheries Management Plan uncertainty in the stock assessment (status of the remaining parts of the ecosystem affected by cod) (economic sustainability) (social well being)

14 ICES WGIAB REPORT 2010 Thus, we can think of the new number being an Ecosystem Based Assessment Score for the Eastern Baltic Cod stock assessment and management. The ultimate goal would be to make operational use of this EBA-Score and connect it to some management action. For example if fishing mortality F is already below 0.3, the EU Management Plan would allow an increase in the TAC. Then three conceptual scenarios could be outlined: i ) ii ) iii ) the EBA-Score is high and supports the recommendation of an increase in TAC, the EBA-Score is low, meaning that the status of the ecosystem is not favourable for high cod recruitment and / or the uncertainty in the assessment is currently high. This would not lead to a support of the recommendation of an increase in TAC. the EBA-Score is average and does neither support or oppose the management action suggested by the EU Management Plan. The specific relation between the EBA-Score and the management action triggered would certainly need to be defined by all stakeholders involved! The road to make this kind of approach operational will certainly be longer and the approach needs to be tested for a couple of years within the real assessment framework to assess its performance and usability. Eventually, the concept of the approach needs to be further discussed, especially the danger to be double-precautious, as the goal is to use ecosystem knowledge as additional information and not to duplicate precautious buffers already included in the reference points used in the current stock assessment framework. 5.4.2 An indicator-based framework to advance ecosystem approach to fisheries in the Baltic Sea: recognizing potentials and limitations of fisheries management (Margit Eero, Astrid Jarre, Henn Ojaveer, Friedrich W. Köster, Martin Lindegren & Maciej Tomasz Tomczak) (see also section 7) An ecosystem-based approach to fisheries management requires a holistic framework capable of integrating over a wider knowledge-base than previously considered within single-species management practices. Moreover, communicating the complexities and interactions of ecological-social systems to a wide range of stakeholders is becoming increasingly important. Indicators and knowledge-based systems can assist in advancing these processes by combining widely different types of information into a single coherent framework. This paper provides a first step towards promoting the development and implementation of an indicator-based framework for the Baltic Sea, using the methodology of knowledge-based systems. We create an ecosystem-based framework to combine ecological/biological, environmental and fisheries indicators related to the recruitment, growth and survival of the three main commercially exploited fish species in the central Baltic Sea. We demonstrate the benefits of this approach, which include (i) tracking and visualising the performance of underlying forcing factors of developments in fish stocks; and (ii) demonstrating the potential and limitations of fisheries management to regulate fish stocks under different ecological/environmental conditions.

ICES WGIAB REPORT 2010 15 5.4.3 Assessment of the environment of the Baltic Proper: application of knowledge-base systems (Margit Eero & Henn Ojaveer) The HELCOM Baltic Sea Action Plan (BSAP) aims at achieving good environmental status (GES) of the Baltic Sea by 2021. The four adopted strategic goals under the BSAP, describing the desired state of the marine environment, are: Baltic Sea unaffected by eutrophication, Marine life undisturbed by hazardous substances, Maritime activities carried out in an environmentally friendly way and Favourable conservation status of Baltic Sea biodiversity. We have carried out preliminary evaluation of the status of the environment of one of the major sub-basins of the Baltic Sea Baltic Proper in relation to the agreed BSAP targets. In doing this, we have utilized altogether 145 state- and pressure indicators, several of which date back to the early 1970s. Our approach is based on the methodology of knowledge-based systems, which allows for combining widely different types of information into a single coherent framework and facilitates communication of the obtained results outside scientific community. We focus particular attention on the sensitivity of the applied assessment method to some critical steps, such as defining reference values, and procedures of aggregation of different indicator information. Accordingly, we present preliminary results of the estimates of the current status in relation to GES targets, together with long-term trends in the environment (incl. biodiversity, eutrophication, pollution) and human pressures, and the robustness of these two approaches is discussed. 6 Biological Ensemble Modelling (ToR c) 6.1 Introduction The main aim of the modelling in WGIAB 2010 was to further the biological ensemble modelling approach (BEMA) initiated during WGIAB 2009 as a tool for the final step in an IEA: management strategy evaluation on ecosystem level (see section 3). This entailed developing (i) simulations of indicator beyond those covered in the 2009 meeting, (ii) coupling of economic analyses to the ecological simulations, and (iii) the deterministic BEMA into stochastic simulations providing probabilities of alternative ecological outcomes. Because the invited expert on risk analytical modelling could not take part in the meeting because of the air-traffic closure, plenary discussions on (iii) was replace by written input (see section 6.4) and intersessional developments of the BEMA, and will be taken up in 2011 (see proposed ToR b in Annex 3). The 2010 meeting instead focussed on ensemble modelling of indicators (section 6.2) as a basis for step 3 in an IEA and on comparing methods of ecological-economic modelling (section 6.3). 6.2 Modelling indicators Two different types of indicators were addressed using the BEMA: simple indicators of particular food-web components (e.g., cod SSB, sprat SSB, zooplankton species biomass) and indicators of food-web properties (a.k.a. network indices). As the ensemble of models ranges from single-species biomass models to age-structured foodweb models (see Table 6.2.1), the ability of the models to simulate the different indicators varies. As a first step, an inventory was made of which indicators of the HEL- COM Baltic Sea Action Plan and the EU Marine Strategy Directive indicator systems that could be simulated in the respective models (summarised in an extensive electronic table, available upon request from Christian Möllmann). Secondly, inter-

16 ICES WGIAB REPORT 2010 sessional work on the BEMA was presented by Anna Gårdmark (Gårdmark et al., in preparation). This work has focussed on further developing the BEMA analyses of cod SSB, the only ecological indicator addressed by all eight models in the ensemble (see section 6.4 for more details). Thirdly, simple indicators and network indices were simulated using one or more models in the ensemble. Table 6.2.1. Models used in the intersessional work on BEMA (for more details see ICES 2009a, Gårdmark et al., in prep.). Models applied for indicator simulations during WGIAB 2010 are indicated in bold. No. Name Reference Brief description 1 Stochastic Cod model Wikström et al, in prep. Auto-regressive (AR(1)) model of total cod biomass. 2 MCMC cod long-term projection model Aro, E. ICES 2008b. Modified ICES medium-term projection model, age-structured cod 3 Cod mini-model Müller-Karulis, in prep. Age-structured cod model, similar to medium-term prediction models for Baltic herring stocks 4 Dynamic cod-herringsprat model 5 SMS (stochastic multispecies model) 6 Stage-structured multispecies model Heikinheimo, in prep. Neuenfeldt et al. in prep.; Lewy and Vinther (2004) Van Leeuwen et al. (2008) Age-structured cod, sprat, herring model including cod predation, modified from MSVPA Age-structured cod, sprat, herring model including cod predation and cannibalism, with size-based diet parametrisation Size-structured cod and sprat, with 2 zooplankton and 1 zoobenthic resources, including cod predation and resourcedependent growth of cod and sprat 7 BALMAR Lindegren et al. (2009) Multi-variate autoregressive (MAR(1)) model of total biomass of cod, sprat, herring, Pseudocalanus, including cod predation, negative effect of sprat on herring and cod 8 Baltic NEST EwE food-web model Tomczak et al., in prep. Ecopath/Ecosim model of age-structured cod, sprat and herring, and total biomass of foodweb components on 7 trophic levels (incl. plankton groups, benthic groups and seals). The only model in the ensemble able to simulate direct climate effects on different zooplankton species/groups, the Ecopath/Ecosim food-web model by BNI ( model 8, Table 6.2.1), was used to simulate the effect of different cod fishing scenario on Acartia spp. and Pseudocalanus elongates biomasses under two different scenarios on further climate change. The intention was to investigate possibilities of estimation of ecosystem indicators. Simulations from year 2006 to 2100 with different climate and fisheries management scenarios have been performed. Environmental changes scenarios with either no further climate change, or with decreasing salinity and increasing temperature as in WGIAB 2009 (ICES, 2009a) were applied, on top of the cod fisheries management scenarios as: business as usual (BAU, F=0.8), cod recovery plan (REC, F=0.3) and ban of Cod fisheries (BAN, F=0). Fishing morality for other species were kept constant. In the Baltic Sea, the most pronounced changes have occurred in the zooplankton and fish trophic levels. The zooplankton dominance changed between Pseudocalanus elongates and Acartia spp. (Möllmann et al., 2003), while changes in hydrography and