Linking marine fisheries to environmental objectives: a case study on seafloor integrity under European maritime policies

environmental science & policy 14 (2011) 289–300

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/envsci

Linking marine fisheries to environmental objectives: a case study on seafloor integrity under European maritime policies Heino O. Fock *, Matthias Kloppmann, Vanessa Stelzenmu¨ller Johann Heinrich von Thu¨nen-Institute, Institute of Sea Fisheries, Palmaille 9, D-22767 Hamburg, Germany

article info Published on line 15 December 2010 Keywords: Marine Strategy Framework Directive Ecological risk assessment Pressure-state-response models Natura 2000 North Sea

abstract Fisheries is regarded a significant impact to the marine environment, and the management of fisheries under maritime environmental policies will be an important task for the future. A relative ecological risk model is applied to define risk components of gain and loss in relationship to 7 demersal fishing me´tiers for the seafloor ecosystem in the German EEZ. Four scenarios are evaluated against the policy goals from European maritime policies. It is shown that two measures combined in an integrative assessment, i.e. effort reduction to MSY and areal closures, are likely to meet requirements from 3 environmental policies, i.e. the Marine Strategy Framework Directive, the Habitats Directive, and the Common Fisheries Policy. Sustainability in terms of maximum sustainable yield for fisheries is likely to provide only partial improvement of the environmental status of the marine ecosystem. The implementation into the pressure-state-response framework of environmental management is discussed. # 2010 Elsevier Ltd. All rights reserved.

1.

Introduction

‘Seafloor integrity’ is one of the marine ecosystem descriptors proposed by the Marine Strategy Framework Directive (MSFD2008/56/EC). It comprises both the physical structure and the biotic composition of the benthic community, and the characteristic functioning of the ecosystem component (Cardoso et al., 2010). Improving the ecological status of the marine environment is a major goal of modern maritime policies. MSFD aims at maintaining or restoring ‘good environmental status’ for the seafloor, where fisheries are regarded as a major impact on benthic ecosystems (Kaiser et al., 2002; Pedersen et al., 2009a). Management is based on assessments of the ecological quality of the marine environment, and of the human activities in the marine environment. Whereas methods for evaluating and managing the effects of hazardous substances as a ‘classical’ problem are well established, the way in which effects from nonconventional pressures (i.e. fisheries, hydromorphological change) are

assessed is far from clear, as there is no clear way to undertake integrated assessments of multiple pressures and to account for multiple management objectives (Apitz et al., 2006). In recent years, three main assessment methodologies have evolved, i.e. pressure-state-response (PSR) models aiming at indicator-based management concepts (Greenstreet et al., 2009; Link et al., 2010; Rochet and Rice, 2005), processbased ecological risk assessment (ERA) models able to treat uncertainty in data and processes (Fock, 2011; Hayes and Landis, 2004; Landis and Wiegers, 1997), and score-based impact or vulnerability models preferably useful for broad scale assessments due to the wide range of impacts analyzed and the many ecosystem components covered (Ban et al., 2010; Halpern et al., 2008; Stelzenmu¨ller et al., 2010b). Based on the OECD model for sustainability indicators (OECD, 1993), PSR models have become highly influential in developing policies. In its extended form (DPSIR) PSR is stateof-the-art for integrated marine assessments in Europe (EEA, 2009). Key concept of PSR models is the description of the

* Corresponding author. Tel.: +49 40 38905 169; fax: +49 40 38905 263. E-mail address: [email protected] (H.O. Fock). 1462-9011/$ – see front matter # 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.envsci.2010.11.005

290

environmental science & policy 14 (2011) 289–300

Table 1 – Ecological indicators to indicate the good environmental status of the seafloor, i.e. good seafloor integritya. Indicators proposed for seafloor integrity Type, abundance, biomass and areal extent of relevant biogenic substrate (6.1.1) Extent of the seabed significantly affected by human activities (such as dredging, trawling or other alterations which may influence the substrate) for the different substrate types (6.1.2) Presence of particularly sensitive or tolerant species (6.2.1) Multi-metric indexes assessing functionality of the benthic system, such as such as the proportion of opportunistic to sensitive species (6.2.2) Proportion of number or biomass of individuals above some specified length/size (6.2.3). Parameters (slope and intercept) of the size spectrum of the aggregate size composition data (6.2.4). a

European Commission Decision (2010/477/EU) on criteria and methodological standards on good environmental status of marine waters.

environmental state evidenced by means of an indicator value. PSR rationale has tailored recent maritime legislation in Europe, i.e. MSFD and Water Framework Directive (WFD, 2000/ 60/EC), in that policy performance is evaluated through a set of traceable indicators each assigned to a specific ecosystem descriptor. However, the link between indicator and pressure may not be defined in all its intricacies or be even indirect, and ‘decoupling’ indicators may be applied if state and pressure trends do not correspond any longer (OECD, 2003). Thus, indicators are not applicable to integrating assessments for more than one pressure (Table 1). Score-based impact assessments are destined to undertake integrative large-scale assessments, given that the score-based characterization of impacts aims at delivering commensurable scales for all pressures. In turn, state of the ecosystem as independently obtained target measure is not an essential element of impact assessments, although in some cases ecosystem state is directly derived from the impact however not as independent measure (e.g. in HELCOM, 2010). Often, the link between pressure and ecosystem is established through matrices (e.g. Robinson et al., 2008) based on the concept of component interaction matrices (e.g. Shopley et al., 1990). In data rich environments and where high resolution of impacts is requested, ecological risk assessments (ERA) combine the merits of large-scale analyses with the modeling of stressor-component interaction processes such as mortality. Through its conceptual working steps, it is a systematic means by which risks may be understood and their estimation may be improved (Fock, 2011; Graham et al., 1992; Harwell et al., 1992; U.S. Environmental Protection Agency, 1998) and can solve complex ecological problems (Lackey, 1998). As relative ERA, cumulative impacts from different pressures can be analyzed and compared across a range of ecosystem components (Fock, 2011). The concept of risk has two sides, in that on the one hand threats (‘downside risk’) and on the other hand opportunities as positive consequences (‘upside risk’) can be imaged, both with their associated uncertainties (Chapman, 1997). For WFD purposes and thus not yet assigned to marine offshore waters, mainly indicator-based methodologies have been applied for assessments of benthic environments, either with a focus on integrating pressures (Aubry and Elliott, 2006) or on the state of the benthic environment (Borja et al., 2009a; Magni et al., 2005). As a novel approach, the application of both up- and downside risk is exemplified for the benthic ecosystem component in the German EEZ of the North Sea in relation to different fishing me´tiers. The concept of negative risk as a measure of impact on an ecosystem component as

part of the impact assessment (Fock, 2011) is complemented with a further measure of effect on the state of the ecosystem component through a gain function. Prior to the risk assessment, the mapping of fisheries (Fock, 2008) and the identification of conservation issues in relation to fisheries (Pedersen et al., 2009a) were undertaken. Environmental objectives relevant to European maritime environmental policies under MSFD or the Habitats Directive (HD-92/43/ EEC) are addressed, and it is demonstrated how multiple pressures and multiple objectives can be integrated into one assessment protocol. Links to the PSR methodology are outlined and prospective development of this standard methodology with regard to risk assessment models and Bayesian networks are discussed.

2. The concept of loss and gain in defining objectives A link between human activity and ecosystem component can be defined as that a human activity (e.g. fisheries) exerts several pressures (e.g. abrasion, extraction of biomass, . . .), which affect ecosystem components in different ways. Ecosystem component and pressure are quantitatively defined by their state (quantity, extension). Ideally, a state is discrete and measurable, it is sensitive to changes in an ecosystem and its response is specific to certain pressures (Link et al., 2010). For fisheries as a source of pressure, a suite of state indicators is available (e.g. measures of fishing effort Piet et al., 2007). The state of ecosystem components (e.g. benthos, birds) is defined in terms of certain endpoints (biomass per unit, abundance, diversity, etc). In both the ERA and the PSR framework, the link between pressure and ecosystem component is formalized in a conceptual modeling step (Fig. 1A) (U.S. Environmental Protection Agency, 1998), however this link is treated differently. In the PSR framework, the state of the indicator is a direct consequence of the pressure (Fig. 1B), which is the basis for the indicator-based rationale. This concept bears a number of caveats. First, the conciseness of the link itself depends on the adequate representation of underlying processes, the adequate selection of the indicator and the degree of resolution and aggregation the state indicators have (see hierarchy of indicators in Piet et al., 2007) considering that univariate numerical state variables might not be able to reflect actual complexities in ecosystems (Rees, 2009). Indices require careful validation and selection from the suite of available

environmental science & policy 14 (2011) 289–300

291

Fig. 1 – (A) Procedural steps for ecological risk assessment (U.S. Environmental Protection Agency, 1998). (B) Formalizing PSR assessments (left) and the relative ecological risk assessment referring to the steps problem formulation, analysis and characterization. Some PSR models approach risk models, so there is a transition from left to right. Note, that ecological state is not an integral part of the risk model, but for the PSR models.

methods, and generalizations for benthic components are yet not widely accepted (Borja et al., 2009b). Second, due to delayed responses like hysteresis or resilience there may be no immediate reaction of the state to a change in the pressure. Long-term effects as known for benthic impacts of fisheries show that such instantaneous reaction is rather unlikely (Tillin et al., 2006). Stochastic events may prevent the recovery of the system or attaining a predicted state. Long-term natural dynamics such as climate forcing add further variability and may lead to regime shifts, and thus weaken the established link between pressure and indicator (Frank et al., 2007; Link, 2002; Mangel, 2006). This leads to the primacy of human impact management over ecosystem management (Frid et al., 2006; Lotze, 2004). Practically, needing to define conservation targets for Natura 2000 sites under the European Habitats Directive the respective ICES working group recognized that it is straightforward to improve site management through the absence of human pressure if ecosystem reference conditions are unknown (ICES, 2008). Third, a lack of data or a weakly defined link between pressure and indicator may hinder to precisely define reference conditions for state indicators (Cardoso et al., 2010). Reference conditions may be rather contemporary and reflect only a little of the desired background level (Rees, 2009).

Hence, the state of the indicator has limited operational properties. Consequently, the operational property to use should be fully reactive to the pressure as part of the activity, but also linked to the ecosystem indicator. Risk models solve this by means of an additional step that quantifies the potential interference between pressure and endpoint (i.e. indicator) in terms of up- and downside risk. Whereas the latter is assigned to the pressure and can be managed in terms of human impact management, the former delivers some prospect on the state of the ecosystem component itself. Opposite to state indicators, risk is probabilistic and multivariate (Fock, 2011). Risk is determined by the exposure likelihood or encounter probability taking into account quantitative state properties of both the pressure and the endpoint, and the strength of the consequence in terms of frequency, duration and impact rates. PSR models at their highest level resemble to some degree risk models (Piet et al., 2007), but as mentioned above, integrative capacities remain limited. Both up- and downside risk explicitly are subject to uncertainty, i.e. uncertainty due to extrapolation and/ or limited information which can be removed with more data, and uncertainty due to randomness of the system itself (Landis, 2002; Mangel, 2006). To incorporate uncertainty into ecosystembased management advice is a key factor for future fisheries management (Frid et al., 2006).

292

environmental science & policy 14 (2011) 289–300

As output from the risk model approach, up- and downside risk and further derived properties, can be adopted to evaluate management options, whereas the state of the ecosystem component is not an integral part of the risk model and therefore outside the risk model framework (Fig. 1B). Downand upside risk are defined through loss and gain functions, respectively. The relationship between gain or upside risk and the state of the indicator may be seen as a time function, i.e. the classical succession-time trajectory of an ecosystem component, which also takes into account its natural longterm sometimes irreversible dynamics. Process-based risk models are data hungry and require extensive process studies (e.g. Hiddink et al., 2007; Piet et al., 2000) and knowledge on spatial distributions of involved features.

3.

Materials and methods

3.1.

DATA

In fisheries management often a me´tier-concept is applied, in which a fishery is defined through gear, mesh size, target species, and vessel categories such as engine power and vessel size. The current EU Council Regulation (EU) 23/2010 (Annex IIa) employs 8 demersal me´tiers to allocate fishing opportunities in certain areas based on the definitions from the Cod Recovery Plan ((EC) 1342/2008 Annex I). In this study, seven gear types aggregated into five bottom contacting gear types are analyzed based on their 2006-distribution in the German EEZ (Fock, 2008): all large beam trawlers (>300 Hp) are assigned to 80–99 mm mesh size category, all small beam trawlers (<300 Hp) also to 80–99 mm, shrimp beam trawlers (<300 Hp) to 16–32 mm, and seiners and otter board trawlers both to 80– 99 mm. Albeit effects from vessels deploying larger mesh sizes are overrated with this approach (i.e. have less by-catch), this provides a more complete picture of fisheries impacts on benthos than in a previous study with only three gear types exemplified (Fock, 2011) and can be easily reassigned to demersal me´tiers as applied by the 2010 EU effort management regulation. In particular, based on results from Berghahn and Vorberg (1998), the abundant shrimp fisheries are included as fisheries me´tier in this study. Gear types are very different with respect to bottom contacting components and gear weight (Supplementary Material Table S1). Heavy rigged large beam trawlers likely equipped with chain mats instead of tickler chains (TBBHG in Fock, 2008) are summarized under large beam trawlers, since there is almost no difference in bycatch statistics for these two gears (Lindeboom and Groot, 1998). Unspecified trawling (TX in Fock, 2008) was summarized under otter board trawling, since it is likely that much of the unknown effort is attributed to effort from non-EU countries, e.g. Norway, for which neither logbook nor EU vessel register information is available, but is clear that these vessels are not beam trawlers. Distribution maps, process data and references on benthic assemblages and fisheries are in the Supplementary Materials S1–S4. The analysis was limited by the available distribution of benthic communities, thus the coastal zone with the Wadden Sea proper is not included.

3.2.

Formulating loss and gain as continuous measures

Risk Z comprises two terms, i.e. the likelihood or exposure (E) and the consequence (C) of an event (see Fox, 2006), Z ¼ E  C:

(1)

For upside risk, we define the consequence as a gain function G, and correspondingly for a downside risk as loss function L: Zþ ¼ E  G

(2)

and Z ¼ E  L:

(3)

In risk models, up- and downside risks are the operational targets in a management process, with a human activity on the one side and the ecosystem component on the other side. In the relative ERA framework, the loss function of a pressure with regard to an ecosystem component is defined by a relative term with mortality potential M over recovery potential R (Fock, 2011), L ¼ ck ð1  eðM=RÞ Þ ¼ ck ð1  eð f ðm;Fm Þ= f ðr;Fr Þ Þ;

(4)

with Mi j ¼ 1  ð1  m jk ÞFik

and Ri j ¼ 1  ð1  0:9m jk ÞFir ;

(5a,b)

where M follows the definition from Piet et al. (2000), mjk is the initial decline in %-abundance after one passage/single application of a bottom trawled fishing gear k for a given sediment type/assemblage j in area unit i (Supplementary Material Table S3), and ck is weighting factor for fisheries k based on benthic bycatch (Table S1). Recovery is interpreted as a function of recovery time as frequency recoveries per year. The factor 0.9 refers to a 90% recovery level in relation to Fr due to available data. Fr is given by sediment type (Table S3) (Hiddink et al., 2006). The gain function summarizes effects for ecosystem component j by a relative term of recovery potential over mortality potential for all impacts k in area unit i, weighted by the impact factor ck for each impact k, i.e. 0

0

G ¼ ð1  eðR =M Þ Þ ¼ ð1  eð f ðc;r;Fr Þ= f g ðc;m;Fm Þ Þ

(6)

with P

k ck ð1  ð1  0:9m jk ÞFi jr nik X ¼ ck ð1  ð1  m jk ÞFi jk

R0i j ¼

and M0i j (7a,b)

k

where njk is the number of pressures k in area unit i. The exposure term is defined differently for up- and downside risk. In both cases, the exposure is a function of the distribution of the ecosystem component as an entity. For downside risk, E is defined as overlap term between pressure and ecosystem component j (Fock, 2011), for upside risk, it is the proportion pi j ¼

ni j Nj

(8)

of the ecosystem component present in an area unit i with an abundance nij and the total abundance Nj in the entire inves-

293

environmental science & policy 14 (2011) 289–300

i

3.3.

i

The disturbance indicator as discrete measure

As derivate of the upside risk on population level, a local disturbance indicator on the level of the area unit i is defined based on the relationship between R and M. The disturbance indicator I is Ii ¼ P

Ri k Mik

;

(10)

after Hiddink et al. (2006). This indicator is unweighted by the impact factor c and thus is more conservative than upside risk. However, due to its local dimension it has no meaning on the population level. Of particular interest are cases with which Ii  1, so that these cases can be treated as discrete measure and the probabilities of reaching a value >1 can be expressed in terms of 90%, 50% and less. Respectively, means from 0/1 coded Ivalues were calculated. These discrete probabilities can be used as input for other assessment frameworks, for instance based on Bayesian networks and multiway contingency tables used within PSR frameworks.

3.4.

Process parameterization

The risk model is sensitive to parameterization in three groups of parameters (Fock, 2011), i.e. the factor determining the exposure in the selected area as part of the biogeographical range of an ecosystem component, the respective mortality and recovery processes, and the determination of the impact factor for the different pressures. In a first approach it was assumed, that 50% of the full distributional range of benthic assemblages was represented for by the German EEZ (Fock, 2011). Here, this figure is revised to 30% (=0.3) based on recently published distributional ranges for assemblages in the North Sea (Reiss et al., 2010). Mortality rates indicate the decrease in numbers after one single passage of a gear. Recovery can be treated in two ways, either with respect to biomass or to abundance (Hiddink et al., 2006). Opportunistic species which return first after disturbance, have faster growth and earlier maturation and normally lower biomass than long-lived species, which as hard-bodied species take more than 1 year to maturity (Gosselin and Qian, 1997). Thus, the return to the level of previous abundance at a site is obtained much faster, often within a half year, than the return to a certain

level of biomass. Whereas the abundance approach is more optimistic (see application in Fock, 2011), the biomass approach is likely more conservative and applied in this study. The respective parameters depend on life stage, sediment type and gear type (Table S3), and average rates by sediment type are applied as surrogates of the typical composition of assemblages. The expert group on MSFD descriptors (Cardoso et al., 2010) considered four specifically susceptible sediment types and their corresponding biotic communities, i.e. soft sediments, gravel, hard substrates and biogenic substrates. Biogenic substrates are not considered in this analysis, but however may occur in the Wadden Sea proper. Hard substrates appear around the island of Helgoland, however, due to uncertainties for calculating fishing effort around this island four adjacent area units were blanked and not included (see Fig. 3). In turn, soft substrates were treated in a more diverse way as sand, mud and muddy sand. For bottom contacting gears, the impact factor was determined by the amount of by-catch in bottom trawling, scaled to 1 for the maximum value. The number of sessile specimens dismantled and hard-bodied specimens damaged should be proportional to the by-catch. This concept was based on findings from the IMPACT project (Lindeboom and Groot, 1998) and is supported by a gear-specific analysis of trawling impacts from Kaiser et al. (2006) for otter and beam trawls.

3.5.

Limit reference values for risk

A challenge for management is to determine key limits for human activities that can be sustained without compromising the functioning of the ecosystem (Frid et al., 2006). Within the relative ERA framework, three evaluation tools are available: the locally defined disturbance indicator I, and on the population level up- and downside risk. Limit reference values for up- and downside risk can be derived analytically when with full overlap the ratio for loss and gain function reaches unity in the power term, so that R

Disturbance indicator I (average)

tigation area (here: EEZ). The proportion in area unit i is again weighted by the biogeographical proportion P of the population present in the investigation area, i.e. if only 20% of a stock is present in an area, the exposure inside the area is weighted by 0.2. The biogeographical proportion accounts for differences in distribution ranges. Thus, species with only a local population entirely subject to a local pressure are weighted differently as compared to ubiquitous species, for which a local pressure is less significant. The upside risk for an ecosystem component j in all area units i is X X Zþ ð pi j  Gi j Þ (9) Zþj ¼ ij ¼ Pj

1 0.8 0.6 0.4 0.2 0

0.001

0.01

0.1

1

10

100

Sum of trawling frequencies Fig. 2 – Distribution of the disturbance indicator I under different trawling frequencies. Trawling frequencies as sum of all frequencies for small and large beam trawlers, shrimp trawlers, otter board trawlers and seiners. Trawling frequencies in yS1 taken from the 2006-scenario without randomization.

294

environmental science & policy 14 (2011) 289–300

environmental science & policy 14 (2011) 289–300

and M and R0 and M0 are equal and the respective absolute values for Z+ and Z are 0.632. This allows for a priori setting of acceptance criteria (Borja et al., 2009b). On the population level, for downside risk a definitely negative effect must be expected if Z increases further beyond the limit reference value. In turn, for upside risk a potential to fully recover may be assumed at Z+ > 0.632. This follows the rationale outlined by Hiddink et al. (2006), where a biomass ratio of 0.9 or better between present state and pristine state has been taken as threshold to indicate an ecologically unimpacted site. For the disturbance indicator I, this means a locally defined limit reference value = 1, and the probability of 50% and 90% of reaching this value is indicated (see Fig. 3B), together with its associated uncertainty. Hiddink et al. (2006) explain why the 90% recovery level is adopted as reference level, i.e. that low frequencies of trawling are still regarded sustainable. For example with a trawling frequency of 0.25 y1 the whole area is only trawled once every four years, or 3/4 of the area remains permanently untrawled, while 1/4 is trawled once per year. For low trawling frequencies the indicator I approaches unity, so that I is able to indicate area units with benthic communities likely being able to sustain themselves (Fig. 2). Different management options and the status quo of an ecosystem can be assessed against the limit reference values for up- and downside risk.

3.6.

Policy requirements and scenarios

MSFD has been regarded weak in that there is no strong integration of MSFD in terms of protocols or instruments into sectoral policies such as CFP and programs under international conventions for the protection of the sea (Salomon, 2009). Only first steps are undertaken to reconcile MSFD and WFD (Borja et al., 2010). Here, we integrate fisheries and environmental goals from MSFD, the Habitats Directive, and the Common Fisheries Policy in a risk assessment framework for fisheries management. For, fisheries management problems, solutions can be described formally as maximization of benefit Q based on decisions Xn affecting multiple pressures, given a set of ecological or management constraints Ym representing policy goals (Lackey, 1998): Q ¼ max f ðX1 ; . . . ; Xn jY1 ; . . . ; Y m Þ

(11)

The above interpretation of sustainability for benthic assemblages is employed to develop an indicator based on risk assessment to meet the policy goals of favorable conservation status as defined in the Habitats Directive (HD-92/43/EEC) and good environmental status by MSFD. HD defines that ‘‘the conservation status will be taken as ‘favorable’ when . . . population dynamics data on the species

295

concerned indicate that it is maintaining itself on a long-term basis as a viable component of its natural habitats, . . .’’. MSFD defines, ‘‘‘good environmental status’ means the environmental status of marine waters where . . . the use of the marine environment is at a level that is sustainable, thus safeguarding the potential for uses and activities by current and future generations, i.e.: (a) the structure, functions and processes of the constituent marine ecosystems, together with the associated physiographic, geographic, geological and climatic factors, allow those ecosystems to function fully and to maintain their resilience to human-induced environmental change.’’ Maintaining the function is reached at R > M, which is incorporated in the definition of the limit reference value for gain and the disturbance indicator I. In line with this interpretation, the expert group on developing indicators for the MSFD descriptors (Cardoso et al., 2010) states that there may be uses that do not cause serious adverse impacts, and the pressures from these uses cause so small perturbations that recovery is rapid. The goal delivered by the European fisheries policy is to obtain sustainable fisheries under a maximum-sustainableyield (MSY) regime (COM(2009)163). Four different scenarios are investigated, i.e. the status quo scenario in terms of the 2006 distribution of fishing effort (Fock, 2008), scenario I as the closure of marine protected sites for large and small beam trawlers to satisfy HD requirements, scenario II as a fishing effort reduction scenario to meet requirements from the MSY limit for plaice as one of the key target species in the area (Pedersen et al., 2009b). Scenario III is a combination of scenarios I and II, i.e. effort reduction plus area closures. Marine protected areas under the European HD, i.e. the Natura 2000 network, is one of the measures to meet the goals of the EU Common Fisheries Policy to develop environmentally sustainable fisheries (SEC(2008)449), and MSY is the corresponding reference level to achieve sustainable exploitation of fish stocks (COM(2009)163). The network in the German EEZ in the North Sea comprises three areas (Fig. 3A1). For these benthic habitats effects from towed bottom gears have been identified as potential conflict in accordance to the conservation objectives as defined by the HD (Pedersen et al., 2009a). In this study, the gear types with the greatest impact factors, i.e. small and large beam trawlers (see Table S1), are excluded from these areas and effort is proportionally reallocated to the remaining area. Fishing effort at MSY level of a stock can be deduced from yield-per-recruit curves. Here, we assume that for the most important stocks of the area, i.e. sole and plaice, MSY fishing effort is ca. 0.4 of the effort undertaken in 2006 (see www.ices.dk for corresponding stock assessment of plaice in the North Sea). This factor is applied to all beam trawlers except for shrimp trawlers as main fleets targeting flatfish (Beare et al., 2010).

Fig. 3 – Distribution of the disturbance indicator I indicating the likelihood for selfsustaining benthic assemblages under three scenarios. (A) Status quo scenario from year 2006, (B) scenario I – closure of Natura 2000 areas for small and large beam trawling, (C) scenario II – reduction of beam trawl fishing effort to MSY levels for plaice (Pleuronectes platessa). EEZ boundaries, Natura 2000 areas and plaice box boundaries indicated in panel A. Area unit size is 1/100 ICES square or 0.18lon T 0.058lat. Four area units around the island of Helgoland are blanked and not included in the analysis (app. 54.18N 7.98E).

296

environmental science & policy 14 (2011) 289–300

For each of the scenarios, 5000 Monte Carlo simulations were run to calculate average values by area and community type (for details see Fock, 2011). To generate a spatial distribution for the disturbance indicator I, further 200 MC runs were undertaken.

Results

0.08

Doggerbank

0 0.08

SOR

0 0.08

Borkum RG

I

II

combined I+II

0 1

0.8

remaining area

0.6

0.4

0.2

TBB1+2 TBBL1+2

TBBmsy TBBLmsy

TBB1 TBBL1

0

OTB2006 SX2006 Shrimp2006 TBB2006 TBBL2006

The distribution of the disturbance indicator I against trawling frequency F shows that trawling is exerted locally in different frequencies, and that reaching an indicator value of 1 is possible (Fig. 2). A high likelihood of locally self-sustaining assemblages (indicator I = 1) can be found within a range from 0.1 y1 to about 1 y1 for all gear types combined, dependent on benthic community type. Accordingly, the distribution of the disturbance indicator is variable with respect to main fishing grounds for each me´tier and the distribution of benthic communities (Fig. 3A1). Relatively good conditions in terms of a likelihood >50% appear for the Sylt outer reef area, a southern coastal zone including the Borkum reef ground and areas adjacent to the Dogger bank MPA. The preponderance of coastal areas is partly due to the exclusion of larger beam trawlers from the coastal areas due to fisheries regulations (plaice box, see Beare et al., 2010), and partly due to the high energy environments of the tidally influenced sandy bottom habitats and tidal gullies inhabited by Common shrimp (Crangon crangon), which are less impacted than soft bottom habitats. High uncertainty associated with this assessment is assigned to the main flatfish fishing area in the middle of the German EEZ, whereas for the Sylt outer reef area and the Borkum reef ground relatively high certainty is obtained (Fig. 3A2). Further high uncertainty is found for the Dogger bank area. However, here a considerable fraction of beam trawling is carried out with gears with 100–120 mm mesh sizes, so that uncertainty is likely overestimated. Under the two reduction scenarios (Fig. 3B, C), the areas with likely sustainable benthic stocks are enlarged. Under the closure scenario for Natura 2000 areas, a clear improvement occurs for the Sylt outer reef area, whereas little improvement is observed for the two other areas (Fig. 3B). The improvement for the Sylt outer reef area is due to abandoning the effect of beam trawlers in that area, whereas beam trawling is only of little importance for the other two areas. In turn, the overall reduction of beam trawl effort according to the MSY level reduces pressure in the entire area, so that a larger area is positively affected. The relatively high importance of beam trawling is evident in the distribution of downside risk by area (Fig. 4). Large beam trawlers can be identified as major risk component in all subareas, so that a change in large beam trawlers’ effort will cause the most significant change in the impact for benthic assemblages. As seen in Fig. 3B, the reduction of beam trawl effort for the Dogger bank led to a moderate increase for the disturbance indicator, which is due to the relatively high risk component assigned to otter trawl fishing in this area (Fig. 4, Dogger bank). Risk from shrimp trawling is relatively small, and only in the ‘remaining area’ an impact twice as high as compared to otter board trawling and small beam trawls is

Cumulative downside risk score Z - by area

4.

Year 2006

Fig. 4 – Distribution of cumulative downside risk by area for three different scenarios: status quo – year 2006 (left), I – Natura 2000 closures for small and large beam trawlers (middle), II – small and large beam trawl effort reduction according to plaice MSY (right), and I and II combined. Scores for OTB, SX, and shrimp trawlers remain unchanged for scenarios I, II, and I + II. Doggerbank – Dogger bank; SOR – Sylt outer reef; Borkum RG – Borkum reef ground. Means by area and SE from 5000 MC runs.

indicated, which are of low values themselves. Seiners are of very low effect. The limit reference value for cumulative downside risk is taken as 0.632  0.3, under the assumption that pressures and communities are evenly distributed within the biogeographical range of the assemblages. The factor 0.3 was the weighting factor for the biogeographical range of the benthic assemblages. The choice of the weighting factor becomes crucial if two or more different EEZs or different ecosystem components with different distributional ranges and effort are to be compared. Under all four scenarios, the sum of cumulative downside risk scores for bottom trawled gears excluding large beam trawlers exceeds the limit reference value thus exerting

297

Combined I + II

0.3 0.2 0.1 0

II: Flatfish MSY

0.3 0.2 0.1 0

I: N2000 clos.

0.3 0.2 0.1 0 0.3

y2006

0.2 0.1

Reef

Mixed

Helg. Deep

Gona-Spisu .93

Mac-balt

Nuc-nit

Gona-Spisu 1

Tell-fab-Gona

Amph-fil

Bath_Amph-fil

0 c North Sea

Cumulative upside risk score Z + by community type

environmental science & policy 14 (2011) 289–300

Fig. 5 – Distribution of upside risk by benthic assemblage und der three scenarios. Assemblage names in legend of Table S2. Limit reference value indicated as hatched line. Means by assemblage and SE from 5000 MC runs.

a definitely negative impact on benthic communities. The cumulative risk for large beamers taken alone is always much higher. Resolving impacts to upside risk by benthic community type allows a differentiated approach to evaluate the scenarios (Fig. 5). In the status quo scenario, only the central North Sea assemblage achieves a value significantly above the limit reference level, i.e. the reference level is not within 1 sample SE. All other assemblages do not reach the limit reference value. In scenario I, closing the Natura 2000 sites for large and small beam trawlers improves the situation significantly for 1 more assemblage, i.e. the Helgoland Deep, whereas an improvement is indicated for 4 assemblages. This includes the assemblages characteristic for the Natura 2000 sites, i.e. Bathyporeis-Amphiura filiformis, reefs and the Tellina fabula-Gonadiella assemblage. Under scenario II, 2 more assemblages are significantly improved, i.e. the Tellina fabula-Gonadiella and the Gonadiella-Spisula assemblages. Also considering non-significant improvements, under the MSY regime associations characterized by mussels, i.e. Tellina fabula, Macoma baltica and Spisula spisula, take benefit from reduced fishing effort. Hard-bodies species are considered highly vulnerable to bottom trawling and thus improvements for this type of species are a valuable indicator. However, the best result is obtained combining both reduction regimes. This will likely increase the upside risk above the limit reference value for 8 out 11 community types, of which the improvement is significant in 5 cases. This shows, how upside risk as measure can be used to develop management strategies.

5.

Discussion

5.1.

Integrative assessments

This study shows, how a process orientated relative ecological risk model provides a background for developing integrative management options with respect to specific gears, multiple goals and different types of fisheries. This approach is completely compliant to current EU maritime policies, in that it applies the me´tier resolution of the CFP management effort regime, and in that it operates objectives from three maritime policies, the MSY CFP goal, the FCS HD goal, and the good environmental status MSFD goal. It is demonstrated, that in an integrative assessment, applying two measures can be sufficient to meet requirements from 3 different policies. It was shown, that sustainability in one policy field, the MSY goal in scenario II, is not fully compatible with requirements from conservation goals under MSFD and HD, and that combined measures are needed. The disturbance indicator could be applied to locally indicate the performance of management options in specially defined areas, e.g. marine protected areas, since it shows both the level of gain and the uncertainty associated with it. The concept of relative ecological risk is not restricted to fisheries, but may be applied to all types of impact for which a reasonable parameterization is available (examples for birds and mammals in Fock, 2011), which is usually the case for data-rich assessment environments (e.g. Wiegers et al., 1998). The integrative approach to three policy goals requires a consistent definition of the terms applied, in this case the term

298

environmental science & policy 14 (2011) 289–300

‘sustainability’ from the model was considered consistent with the definitions from MSFD and HD, so that a transposition from model to policy requirements was possible. It is further indispensable to have all spatial uses included. In this study, only fisheries and MPAs are considered, however, with the boost in offshore engineering likely more competitors for space and likely more impacts become important (Berkenhagen et al., 2010; Stelzenmu¨ller et al., 2010b).

5.2.

Integration into the PSR framework

Treating fisheries in a data-poor environment means, that a different kind of parameterization is needed. For the disturbance indicator I an alternative parameterization is discussed, in that the likelihood of reaching state = 1 is discussed in terms of the levels of <50%, >50% and >90%. This discrete distribution of states can be incorporated into discrete models such as Bayesian networks, and the necessary distribution of state = 1 values can be derived from ecological risk models. In turn, Bayesian belief networks could be preferred tools to operate PSR and DPSIR frameworks (for DPSIR see Gabrielson and Bosch, 2003). These networks often operate categorized data and apply more generic links between components (Stelzenmu¨ller et al., 2010a), a characteristic feature for datapoor environments. In turn, applying the risk concept for PSR frameworks means to substitute the state of the indicator by adequately defined risk parameters, referring to the arrow in Fig. 1B from left to right in the opposite direction. Seafloor integrity as well as benthic habitats are important conservation targets under the European maritime policies. So far, mainly based on the PSR methodology, a series of state indicators has been proposed to develop management options within the MSFD framework (Table 1). None of them is operational in a way that limit reference values are easy at hand. In an expert report on the development of these state indicators, for seafloor integrity the expert group states that ‘‘Reference levels for extent of substrate types and abundance of species associated with specific substrates need to be evaluated relative to local historical baselines, which are often not quantified’’ (Cardoso et al., 2010). As a solution to overcome the shortcomings of ecological state indicators, the expert report in turn proposes to replace the ecological indicator in some cases with a pressure indicator (e.g. fishing effort), in that ‘‘Pressure indicators are likely to be more cost effective and sensitive than many direct indicators of substrate features.’’ This is a step behind from the ecological indicator concept and risk modeling in that only pressure is considered. Further, it does not link the pressure to a level that ensures sustainable development of the ecosystem component under consideration which is definitely required to meet the policy standard from the MSFD. The missing link between pressure and ecological state is a clear disadvantage of the PSR concept and in turn underlines the value of ecological risk models in such context. The choice management ought to seek is to enable opportunities to develop, and in that upside risk is probabilistic, i.e. it is a chance, it is the exact measure to quantify this step. As mentioned before, the ecosystem component will respond

according to its natural dynamics and reach a certain state after a given time period, but the definition of upside risk itself is not fixed to a certain state value.

5.3.

Application of the risk model

Downside risk is a means to identify the main sources of environmental pressure and upside risk to define the extent of possible management scenarios necessary to likely reach a certain environmental status. Both are straightforward outputs from the risk model. MC simulations provide error ranges associated with the processes and the data. As further spatial indicator the disturbance indicator is developed, which is very similar to the indicator developed by Hiddink et al. (2006). But as mentioned above, the ecological risk model is applicable to more than one ecosystem component and impact, proving its general applicability for maritime policy purposes.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.envsci.2010.11.005.

references

Apitz, S.E., Elliott, M., Fountain, M., Galloway, T.S., 2006. European environmental management: moving to an ecosystem approach. Integr. Environ. Assess. Manage. 2, 80–85. Aubry, A., Elliott, M., 2006. The use of environmental integrative indicators to assess seabed disturbance in estuaries and coasts: Application to the Humber Estuary, UK. Mar. Pollut. Bull. 53, 175–186. Ban, N.C., Alidina, H., Adron, J.A., 2010. Cumulative impact mapping: advances, relevance and limitations to marine management and conservation, using Canada’s Pacific waters as a case study. Mar. Policy 34, 876–886. Beare, D., Rijnsdorp, A., Kooten, T.V., Fock, H.O., Schro¨der, A., Kloppmann, M., Witbaard, R., Meesters, E., Schulze, T., Blaesbjerg, M., Damm, U., Quirijns, F., 2010. Study for the Revision of the Plaice Box – Draft Final Report, C002/10. IMARES. Berghahn, R., Vorberg, R., 1998. Shrimp Fisheries and Nature Conservation in the National Park Wadden Sea of Schleswig-Holstein 82–97. Federal Environmental Agency, Berlin. Berkenhagen, J., Do¨ring, R., Fock, H.O., Kloppmann, M.H.F., Pedersen, S.A., Schulze, T., 2010. Decision bias in marine spatial planning of offshore wind farms: problems of singular versus cumulative assessments of economic impacts on fisheries. Mar. Policy 34, 733–736. Borja, A., Elliott, M., Carstensen, J., Heiskanen, A.-S., van de Bund, W., 2010. Marine management – towards an integrated implementation of the European Marine Strategy and the Water Framework Directives. Mar. Pollut. Bull., doi:10.1016/j.marpolbul.2010.09.026. Borja, A., Muxika, I., Rodriguez, J.G., 2009a. Paradigmatic responses of marine benthic communities to different anthropogenic pressures, using M-AMBI, within the European Water Framework Directive. Mar. Ecol. 30, 214–227.

environmental science & policy 14 (2011) 289–300

Borja, A., Ranasinghe, A., Weisberg, S.B., 2009b. Assessing ecological integrity in marine waters, using multiple indices and ecosystem components: challenges for the future. Mar. Pollut. Bull. 59, 1–4. Cardoso, A.C., Cochrane, S., Doerner, H., Ferreira, J.G., Galgani, F., Hagebro, C., Hanke, G., Hoeppffner, N., Keizer, P.D., Law, R., Olenin, S., Piet, G.J., Rice, J., Rogers, S.I., Swartenbroux, F., Tasker, M.L., van de Bund, W., 2010. Scientific Support to the European Commission on the Marine Strategy Framework Directive – Management Group Report, draft. JRC-ICES. Chapman, C., 1997. Project risk analysis and management – PRAM the generic process. Int. J. Proj. Manage. 15, 273–281. EEA, 2009. EEA Strategy 2009–2013 European Environment Agency, Copenhagen. Fock, H.O., 2008. Fisheries in the context of marine spatial planning: defining principal areas for fisheries in the German EEZ. Mar. Policy 32, 728–739. Fock, H.O., 2011. Integrating multiple pressures at different spatial and temporal scales: A concept for relative ecological risk assessment in the European marine environment. Hum. Ecol. Risk Assess. 17 , doi:10.1080/ 10807039.2011.538634. Fox, D.R., 2006. Statistical issues in ecological risk assessment. Hum. Ecol. Risk Assess. 12, 120–129. Frank, K.T., Petrie, B., Shackell, N.L., 2007. The ups and downs of trophic control in continental shelf ecosystems. Trends Ecol. Evol. 22, 236–242. Frid, C.L.J., Paramor, O.A.L., Scott, C., 2006. Ecosystem-based management of fisheries: is science limiting? ICES J. Mar. Sci. 63, 1567–1572. Gabrielson, P., Bosch, P., 2003. Environmental Indicators: Typology and Use in Reporting. European Environment Agency, Copenhagen. Gosselin, L.A., Qian, P.-Y., 1997. Juvenile mortality in benthic marine invertebrates. Mar. Ecol. Prog. Ser. 146, 265–282. Graham, R.L., Hunsaker, C.T., O’Neill, R.V., Jackson, B.L., 1992. Ecological Risk Assessment at the Regional Scale. Ecol. Appl. 1, 196–206. Greenstreet, S.P.R., Holland, G.J., Fraser, T.W.K., Allen, V.J., 2009. Modelling demersal fishing effort based on landings and days absence from port, to generate indicators of ‘‘activity’’. ICES J. Mar. Sci. 66, 886–901. Halpern, B.S., Walbridge, S., Selkoe, K.A., Kappel, C.V., Micheli, F., D’Agrosa, C., Bruno, J.F., Casey, K.S., Ebert, C., Fox, H.E., Fujita, R., Heinemann, D., Lenihan, H.S., Madin, E.M.P., Perry, M.T., Selig, E.R., Spalding, M., Steneck, R., Watson, R., 2008. A global map of human impact on marine ecosystems. Science 319, 948–952. Harwell, M.A., Cooper, W., Flaak, R., 1992. Prioritizing ecological and human welfare risks from environmental stresses. Environ. Manage. 16, 451–464. Hayes, E.H., Landis, W.G., 2004. Regional ecological risk assessment of a near shore marine environment: Cherry Point, WA. Hum. Ecol. Risk Assess. 10, 299–325. HELCOM, 2010. Ecosystem Health of the Baltic Sea 2003–2007: HELCOM Initial Holistic Assessment, 122. HELCOM, Helsinki. Hiddink, J.G., Jennings, S., Kaiser, M.J., 2006. Indicators of the ecological impact of bottom-trawl disturbance on seabed communities. Ecosystems 9, 1190–1199. Hiddink, J.G., Jennings, S., Kaiser, M.J., 2007. Assessing and predicting the relative ecological impacts of disturbance on habitats with different sensitivities. J. Appl. Ecol. 44, 405–413. ICES, 2008. Report of the Workshop on dealing with Natura 2000 and Related Requests (WKN2K), ICES CM 2008/ACOM:46. ICES, Copenhagen. Kaiser, M.J., Clarke, K.R., Hinz, H., Austen, M.C.V., Somerfield, P.J., Karakassis, I., 2006. Global analysis of response and recovery of benthic biota to fishing. Mar. Ecol. Prog. Ser. 311, 1–14.

299

Kaiser, M.J., Collie, J.S., Hall, S.J., Jennings, S., Poiner, I.R., 2002. Modification of marine habitats by trawling activities: prognosis and solutions. Fish Fisheries 3, 114–136. Lackey, R.T., 1998. Fisheries management: integrating societal preference, decision analysis, and ecological risk assessment. Environ. Sci. Policy 1, 329–335. Landis, W.G., 2002. Uncertainty in the extrapolation from individual effects to impacts upon landscapes. Hum. Ecol. Risk Assess. 8, 193–204. Landis, W.G., Wiegers, J.A., 1997. Design considerations and a suggested approach for regional and comparative ecological risk assessment. Hum. Ecol. Risk Assess. 3, 287–297. Lindeboom, H.J., Groot, S.J.d. (Eds.) (1998) Impact II – The Effects of Different Types of Fisheries on the North Sea and Irish Sea Benthic Ecosystems. Netherlands Institute for Sea Research, AB Den Burg, The Netherlands, p. 404. Link, J.S., 2002. What does ecosystem-based fisheries management mean? Fish Manage. 27, 18–21. Link, J.S., Yemans, D., Shannon, L.J., Coll, M., Shin, Y.-J., Hill, L., Borges, M.D.F., 2010. Relating marine ecosystem indicators to fishing and environmental drivers: an elucidation of contrasting responses. ICES J. Mar. Sci. 67, 787–795. Lotze, H.K., 2004. Repetitive history of resource depletion and mismanagement: the need for a shift in perspective. Mar. Ecol. Prog. Ser. 274, 282–285. Magni, P., Hyland, J., Manzella, G., Rumohr, H., Viaroli, P., Zenetos, A., 2005. In: Proceedings of the Workshop ‘‘Indicators of Stress in the Marine Benthos’’, 195. UNESCO, Paris. Mangel, M., 2006. Commentary: accounting for uncertainty in marine reserve design. Ecol. Lett. 9, 11–12. OECD, 1993. OECD Core Set of Indicators for Environmental Performance reviews, OCDE/GD(93)179. OECD, Paris. OECD, 2003. OECD Environmental Indicators – Development, Measurement. OECD, Paris. Pedersen, S.A., Fock, H.O., Krause, J., Pusch, C., Sell, A., Bo¨ttcher, U., Rogers, S., Sko¨ld, M., Skov, H., Podolska, M., Piet, G., Rice, J., 2009a. Natura 2000 sites and Fisheries in German Offshore Waters. ICES J. Mar. Sci. 66, 155–169. Pedersen, S.A., Fock, H.O., Sell, A., 2009b. Mapping fisheries in the German exclusive economic zone with special reference to offshore Natura 2000 sites. Mar. Policy 33, 571–590. Piet, G.J., Quirijns, F.J., Robinson, L., Greenstreet, S.P.R., 2007. Potential pressure indicators for fishing, and their data requirements. ICES J. Mar. Sci. 64, 110–121. Piet, G.J., Rijnsdorp, A.D., Bergman, M.L.N., van Santbrink, J.W., Craeymeersch, J., Buijs, J., 2000. A quantitative evaluation of the impact of beam trawling on benthic fauna in the southern North Sea. ICES J. Mar. Sci. 57, 1332–1339. Rees, H.L., 2009. Role of benthic indicators in regulating human activities at sea. Environ. Sci. Policy 9, 496–508. Reiss, H., Degraer, S., Duineveld, G.C.A., Kro¨ncke, I., Aldridge, J., Craeymeersch, J.A., Eggleton, J.D., Hillewaert, H., Lavaleye, M.S.S., Moll, A., Pohlmann, T., Rachor, E., Robertson, M., Berghe, E.V., Hoey, G., Rees, V.H.L., 2010. Spatial patterns of infauna, epifauna, and demersal fish communities in the North Sea. ICES J. Mar. Sci. 67, 278–323. Robinson, L.A., Rogers, S., Frid, C.L.J., 2008. A marine assessment and monitoring framework for application by UKMMAS and OSPAR-Assessment of Pressures, F90-01-1075. Univ. of Liverpool/CEFAS. Rochet, M.-J., Rice, J.C., 2005. Do explicit criteria help in selecting indicators for ecosystem-based fisheries management? ICES J. Mar. Sci. 62, 528–539. Salomon, M., 2009. Recent European initiatives in marine protection policy: towards lasting protection for Europe’s seas? Environ. Sci. Policy 12, 359–366. Shopley, J., Sowman, M., Fuggle, R., 1990. Extending the capability of the component interaction matrix as a

300

environmental science & policy 14 (2011) 289–300

technique for addressing secondary impacts in environmental assessment. J. Environ. Manage 31, 197–213. Stelzenmu¨ller, V., Lee, J., Garnacho, E., Rogers, S.I., 2010a. Assessment of a Bayesian Belief Network-GIS framework as a practical tool to support marine spatial planning. Mar. Pollut. Bull. 60, 1743–1754. Stelzenmu¨ller, V., Lee, J., South, A., Rogers, S.I., 2010b. Quantifying cumulative impacts of human pressures on the marine environment: a geostatistical framework. Mar. Ecol. Prog. Ser. 398, 19–32. Tillin, H.M., Hiddink, J.G., Jennings, S., Kaiser, M.J., 2006. Chronic bottom trawling alters the functional composition of benthic invertebrate communities on a sea-basin scale. Mar. Ecol. Prog. Ser. 318, 31–45. U.S. Environmental Protection Agency, 1998. Guidelines for Ecological Risk Assessment, EPA/630/R-95/002F. Risk Assessment Forum, Washington, DC. Wiegers, J.K., Feder, H.M., Mortensen, L.S., Shaw, D.G., Wilson, V.J., Landis, W.G., 1998. A regional multiple-stressor rankbased ecological risk assessment for the Fjord of Valdez, Alaska. Hum. Ecol. Risk Assess. 4, 1125–1173. Heino O. Fock joined the Institute of Sea Fisheries, Hamburg, in 2005 after working as consultant for the Census of Marine

Life-project, the Common Wadden Sea Secretariat, the German Federal Agency for Hydrology and others. 2000 to 2002, he served as Postdoc at the Alfred-Wegener Institute in the field of seamount ecology. His main experience is the ecology of demersal fish stocks in the North Atlantic and fisheries effects on marine ecosystems. Matthias Kloppmann works in the Institute of Sea Fisheries, Hamburg, since 2002. Before that he was staff member at the Insitute of Hydrobiology and Fisheries of the University of Hamburg (1999-2001) and the Biologische Anstalt Helgoland, Hamburg (1994-1997) after being granted of a doctorate in 1994 with a thesis on ‘‘Small scale distribution of fish larvae off the island of Helgoland’’. His main expertise lies in the field of fisheries ecology in the North Sea. Vanessa Stelzenmu¨ller joined the Institute of Sea Fisheries, Hamburg in 2010 after working at the Centre of Environment, Fisheries and Aquaculture in Lowestoft/UK for three years. Before that, she served as Postdoc at the Instituto de Ciencias del Mar (ICM) in Barcelona in 2005-2007. She was granted of a doctorate in 2005 at the University of Oldenburg, Germnay. Her main expertise lies in the field of marine spatial planning and georeferenced impact assessments.