Marine environmental managers primarily aim to protect and maintain natural structure and functioning while simultaneously ensuring that ecosystems provide services, which in turn deliver benefits for society (Atkins et al., 2011; Elliott, 2011). In the management of human activities in the marine environment, it is axiomatic that a regulatory body (i.e., an environmental protection agency, natural conservation body, fisheries body, or marine licensing body), does not have to prove that an activity or its developer (the “user,” “polluter”—those undertaking the activity, such as a dredging company, industrial plan, or wind farm operator) causes an adverse impact (Gray and Elliott, 2009). In contrast, the developer must prove they will not cause an impact, hence creating the scientific and statistical challenge of “proving the negative.” A second key feature, “the precautionary principle” (PP), assumes a deleterious effect resulting from a given activity in the system unless proven otherwise (O'Riordan and Jordan, 1995). However, detractors criticize the vague definition of PP, and balancing scientific uncertainty and appropriate management measures remains a challenge (Steel, 2014).

The third key feature states that any developer wishing to use the marine system must obtain permission from a regulatory body, hence the importance of sufficient administrative bodies (Boyes and Elliott, 2014, 2015; Elliott, 2014); this encompasses the whole of marine governance, defined as the net result of policies, politics, legislation, and administration (Barnard and Elliott, 2015). The fourth feature, the “polluter pays principle,” requires a developer to pay for the costs associated with that use: the licensing of the activity, the monitoring, remediation and mitigation of any damage to the system and, if necessary, compensation. The latter requires integrating natural and economic sciences to enable sustainability within and across generations and it may require developers to compensate affected users, the affected resource (e.g., restocking affected fish), or the affected environment (e.g., by creating new environment; Elliott et al., 2016). However, all of these central features relate to how users use an area of the sea (e.g., dredging, wind farm, fishing, etc.) but superimposing a wider suite of natural and human influences, such as climate change, on all of these activities (Elliott et al., 2015). This complexity demands, as the fifth feature, assessing the anthropogenic change or pressure in question (a “signal”) against a background of inherent variability and natural change or wider influences, i.e., the changes emanating externally to the area being managed (the “noise”; Gray and Elliott, 2009; Elliott, 2011). Finally, a sixth key feature requires quantitative and legally defendable detection of such change with a direct feedback into management.

These key features require a defendable, holistic, underlying framework, accepted, and communicable to marine managers and wider users. That framework must link causes of potential and actual changes to the marine environment, the types of changes experienced and societal responses to mediating or removing the drivers of change or at least accepting change for the benefits provided. Even in the recent past, stakeholders frequently used the DPSIR (Drivers, Pressures, State change, Impact and Response) interlinking framework (e.g., Atkins et al., 2011; Smith et al., 2014), without clearly defining each element. Hence, the wide use of DPSIR model (Gari et al., 2015; Lewison et al., 2016; Patrício et al., 2016a) not only introduced many variants and perpetuated confusion but also made it not-fit-for-purpose in providing management guidance.

Previous studies document the evolution of the DPSIR approach (Smith et al., 2014, 2016), and here we summarize and focus on the evolution from DPSIR to the most recent derivative DAPSI(W)R(M) (Patrício et al., 2016a; Scharin et al., 2016; Burdon et al., in press). This modified approach adds Activities, and relates the Impact to human Welfare and the Responses to the use of Measures (the term preferred by EU Directives). Drivers describe underlying basic human needs, such as for food, security, space, and well-being, which require Activities (fishing, building wind farms, creating navigation routes). These activities then create Pressures, such as scraping the seabed with bottom trawls or building infrastructure that removes space. Pressures are the mechanisms that change the system, potentially causing concern. Those changes encompass both the natural system, including its structure and functioning (the “State change”; Strong et al., 2015), and the human system [the Impact (on human Welfare)]. The term Welfare is used sensu stricto to include economic welfare and human and societal well-being (Oxford English Dictionary).

Furthermore, all of the activities and external changes could potentially adversely affect that main aim (the protection of the social and ecological systems), and may thus be considered hazards. If these hazards damage parts of the socio-ecological system we value, they may be termed risks, thus providing a hazard and risk typology used in the DEVOTES project (Elliott et al., 2014). Smith et al. (2016) illustrated DPSIR, using fishing activity and the pressure of trawling from abrasion on the seabed and its impacts on particular components as an example. The challenges were addressed in moving from conceptual models to actual assessments including: assessment methodologies (interactive matrices, Bayesian Belief Networks, ecosystem modeling, the Bow Tie approach, assessment tools), data availability, confidence, scaling, cumulative impacts, and multiple simultaneous pressures, which more often occur in multi-use and multi-user areas (Smith et al., 2016).

Society and environmental managers need to know not only the current status of a marine system, but also whether it has been altered, the cause of that alteration, its significance, and what can be done to reverse that change. Therefore, this requirement creates the need to consider how Pressures result in State change, in the natural system, and a societally relevant Impact of sea use (including the assessment of cumulative pressures and impacts, as shown by Korpinen and Andersen, 2016); hence the need to consider not just Welfare (sensu DPSWR in Cooper, 2013) but the Impact (on human Welfare). This need explicitly includes an economic approach and a human health and well-being approach to human-induced changes. Furthermore, while that State change may often relate to the physico-chemical and ecological structure of the marine system, it increasingly requires users to consider the ecological functioning (Strong et al., 2015) especially given that many MSFD descriptors relate to functioning aspects. This “biodiversity-ecosystem functioning debate,” regarding the effect of functioning on biodiversity and vice versa, is an important and developing field (Zeppilli et al., 2016).

The detection or prediction of changes to the natural state and impacts on human welfare require action to minimize, mitigate, compensate, remove, or even accept changes through societal Responses (the R in DPSIR). However, based on terminology used in the EU Directives, environmental managers now refer to those Responses as Measures [hence Responses -using Measures- in DAPSI(W)R(M); Scharin et al., 2016]. During the past decade, management has recognized the need to include all measures which therefore, as referred to as the Programme of Measures in the MSFD, should consider aspects of ecology, technology, economy, legislation, and administration. They should also satisfy societal, cultural and moral imperatives while communicating decisions to stakeholders; hence the so-called “10 tenets” for sustainable and successful marine management (Elliott, 2013; Barnard and Elliott, 2015).

The prevailing governance system provides a central control on adverse effects of human activities. The EU arguably represents the pre-eminent proponent of marine environmental legislation and other aspects of governance (Boyes and Elliott, 2014), but the complexity of the marine system, the need for transboundary action and the joint implementation of different systems have produced anomalies, confusion, and a need for an inter-governmental transboundary approach (Cavallo et al., 2016).

Most of the above framework relates to activities and pressures emanating from within a system such as a sea region, under management, for example the Baltic or North Seas (Andersen et al., 2015; Scharin et al., 2016). These may be termed endogenic managed pressures in which the causes and consequences in the region are managed (Elliott, 2011) and under legislative control (Boyes and Elliott, 2014). Exogenic unmanaged pressures (i.e., those aspects emanating from outside a managed system; for example global climate change Elliott et al., 2015) represent the major current challenge; environmental managers cannot control the causes but must respond to the consequences. Climate change offers a primary example, in which human impacts (e.g., ocean acidification, increase in alien species, sea-level rise, temperature regime change; Danovaro et al., 2013; Katsanevakis et al., 2014, 2016) add to internal pressures in an area. Climate change therefore shifts baselines, complicating evaluation change associated with internal activities in a region, but also potentially nullifying the use of quantitative indicators or at least requiring the target values of those indicators to be continually revised. A Member State not meeting legislative controls, such as directives, may therefore cite climate change as a modifying factor but one outside of its control (Elliott et al., 2015). Targets that cannot be reached due to changes caused by climate change effects are not manageable and need to be revised as a part of the 6 years management cycle.

Successful ocean use management relies on adequate and comprehensive monitoring, and identifying appropriate measurements of change. Management response requires a clear understanding of underlying causes and effects of change in the marine environment and their consequences. Hence, the use of conceptual models linking the marine drivers, activities, and pressures can provide that solid foundation to link to state changes, impacts on societal welfare, and the resulting management responses using programmes of measures. Similarly, management relies on the ability to predict and detect future responses of the system to changes with sufficient certainty; prediction requires conceptual, empirical, and deterministic models, whereas detection implies the presence of robust monitoring systems at appropriate spatial and temporal scales. However, the “paradox of environmental assessment” sets the backdrop for this framework whereby increasing national and European legislation (such as the MSFD) requires more understanding and better monitoring but monitoring organizations face reduced budgets (Borja and Elliott, 2013). Therefore, by expanding the concept of DPSIR into DAPSI(W)R(M), understanding the gaps and the Strengths, Weaknesses, Opportunities, and Threats in monitoring, and exploring how climate change could affect GES, DEVOTES has included human welfare in the modified approach, emphasizing the importance for future policy and management measures. Hence, an adequate assessment of marine status can only be achieved through fit-for-purpose monitoring based on sound scientific knowledge.

The multifaceted concept of biodiversity encompasses everything from the genetic composition of species to the organization of habitats and ecosystems (CBD, 1992). Despite the widely recognized need to maintain biodiversity, its many interpretations make difficult any comprehensive evaluation and therefore it is necessary to use indicators, or simplified measures, that reflect or synthesize the status of important aspects of ecosystem structure or function. Marine assessments depend upon indicators to detect and evaluate changes in environmental status driven by either natural or human pressures, often in the context of implementing management targets for environmental objectives and measures. Therefore, scientists and managers worldwide seek accurate and reliable indicators that represent all relevant aspects of marine biodiversity either as individual aspects or as surrogates (proxies) for series of changes (for example the use of the breeding health of piscivorous seabirds as a proxy for the whole marine trophic system).

Although, many nations worldwide recognize the need for an ecosystem approach to ocean management, the EU has led in developing specific metrics toward that objective. The European Commission (2010) Decision specifies criteria and methodological standards to evaluate environmental status of marine waters, based upon a set of 56 MSFD indicators. Some indicators used in the assessment of coastal ecosystems under the Water Framework Directive (WFD; European Commission, 2000; Birk et al., 2012) also apply to the MSFD assessment beyond the narrow coastal strip where MSFD and WFD overlap (Borja et al., 2010; Boyes et al., 2016). In practice, during the first phase of the MSFD implementation, EU Member States used different methodological approaches to determine and assess ecosystem status (European Commission, 2014; Palialexis et al., 2014). Data availability, regional specificities, and potentially different interpretations of the EU Commission Decision led to discrepancies within methodologies reported by Member States, increasing the potential for non-harmonized approaches to status determination. Managers require further guidance on criteria for “good” indicators, and assessment of status (Patrício et al., 2014), and such a plan is currently being developed by the EU and its Member States, ICES (International Council for the Exploration of the Sea), EEA (European Environment Agency), and RSCs (The Regional Sea Conventions). Concurrently, the RSCs are developing indicators for holistic marine assessments (e.g., HELCOM, 2013; OSPAR, 2015; UNEP, 2016).

Increasing legal challenges of marine and coastal management, both to the EU Member State implementation of Directives and industry compliance with national laws, which hinge upon detecting and demonstrating marine environmental change (Elliott et al., 2015), increases the need for scientifically defensible indicators. Those indicators must be comprehensive, either in covering all relevant aspects of the marine system or as conceptually defensible surrogates that represent a well-defined and well-accepted causal link (e.g., the health of breeding populations of top seabird and fish predators being dependent on the health of seabed populations).

We tested and refined 13 available biodiversity indicators, developed 16 new options for assessment, particularly for biological descriptors (considering species, habitat and ecosystem levels), identified gaps for future research, developed indicator performance criteria, and provided a user-friendly tool to select and rank indicators (Table 1). These publicly-available contributions (Berg et al., 2016), support the second phase of the MSFD implementation and assist marine management in Europe and elsewhere.

Understanding how changes in biodiversity link to food-web functioning, anthropogenic pressures, and climate changes requires novel, integrative modeling tools. Similarly, scaling determining change from small to large areas and from the present to future, also requires such modeling approaches. Once validated, modeling tools can elucidate expected risks and rewards for a range of management options, aimed at achieving or maintaining GES. The evidence base from such scenario testing thus provides a suitable platform to enable informed decision-making. Prior to 2012, the proposals for using models in the MSFD implementation were very limited (Cardoso et al., 2010) but now, in the context of using models in assessments, Pinnegar et al. (2014) for example have demonstrated the value of food-web models in assessing potential responses of ecosystems to invasions.

We assessed the capabilities of state-of-art models to provide information about current and candidate indicators outlined in the MSFD, particularly on biological diversity, food-webs, non-indigenous species, and seafloor integrity descriptors (Piroddi et al., 2015; Tedesco et al., 2016). We demonstrated that models could explain food-webs and biological diversity, but poorly-addressed non-indigenous (alien) species, habitats and seafloor integrity (Lynam et al., 2016).

Marine research and assessment require modeling studies that can support the use of indicators in fully encompass the functional linkages between ecosystem components and overwhelming pressures on the marine environment, such as climate change and ocean acidification. Such modeling provides the evidence for setting realistic targets and thus supporting better long-term marine planning. We used case studies to illustrate that modeling can assist in MSFD implementation, contributing to each step of the assessment and management cycle. Modeling can help to develop and refine novel indicators to support indicator-based assessment of GES. Modeling can incorporate indicator trends and responses, incorporating prevailing climatic conditions and anthropogenic pressures and, in this way, support the review of objectives, targets and indicators. Moreover, modeling can both inform adaptive monitoring programmes and be used in scenario testing to inform management decisions.

Methods traditionally used in marine monitoring to investigate spatial and temporal variation in abiotic and biotic variables are time-consuming, costly and often limited in resolution (de Jonge et al., 2006; Borja and Elliott, 2013; Carstensen, 2014; Fraschetti et al., 2016). These constraints can severely limit our capacity to detect spatial and temporal changes in marine environmental health. In addition, most countries lack the tools to expand marine monitoring to the deep sea (Ramirez-Llodra et al., 2011), severely constraining the expected implementation of the MSFD in the open ocean and deep sea (Zeppilli et al., 2016). Moreover, marine monitoring methods currently limit analyses of some descriptors. For example, detecting cryptic and/or alien species (including those causing harmful algal blooms) will benefit from molecular approaches (Bourlat et al., 2013).

Activities that smoother, abrade or permanently-remove seabed habitat represent the greatest threats to seafloor integrity (Rice et al., 2012). Previous studies used benthic faunal analysis to indicate general seafloor integrity (Pearson and Rosenberg, 1978), drawing on an extensive catalog of methods and approaches for such a fundamental change (Gray and Elliott, 2009), but increasingly together with various visual assessment tools (Solan et al., 2003). Specific benthic faunal indicators exist for trawl abrasion (Jorgensen et al., 2016) but deriving these indicators is time-consuming and expensive to implement. Video inspection of seafloor smothering using Remotely Operated Vehicles (ROV), such as from seabed drilling activities, can visually map the environmental footprint (Gates and Jones, 2012), but we lack data to validate the uncertainty of the method compared to conventional biological sample collection.

The implementation of the assessments of marine environmental status required by the MSFD thus requires development and/or testing of innovative monitoring systems. Despite creating recent methodologies/technologies in DEVOTES, these are not yet used in routine monitoring of the MSFD descriptors. We encourage this through our summary analysis encompassing a catalog of monitoring networks and a wide array of potential tools, including: (i) molecular approaches (e.g., barcoding and metagenomic tools), (ii) remote sensing/acoustic methods, and (iii) in situ monitoring techniques.

We have produced a catalog with the biodiversity monitoring networks, currently available in European Seas, with the aim to: (i) present a critical overview of the monitoring activities in Europe (i.e., the amount and reason for ongoing monitoring, whether it fulfills its objectives and to what pressures it is links), (ii) identify areas where no monitoring occurs, and (iii) recommend the further development and improvements for optimizing marine biodiversity monitoring in the context of the MSFD. Since the publication of the catalog (Patrício et al., 2014), new material has been added so that it currently identifies 865 monitoring activities corresponding to 298 monitoring programmes. A gap and SWOT (Strengths, Weaknesses, Opportunities, and Threat) analysis of the catalog (Patrício et al., 2016b), highlights uneven distributions of monitoring across regional seas (i.e., more monitoring activities in the North Eastern Atlantic and Mediterranean). Specifically, we note uneven monitoring effort between descriptors (e.g., more monitoring for Descriptor 1 on Biological diversity and Descriptor 4 on Food webs), between biological components (e.g., monitoring emphasis on fish and phytoplankton) and between pressures (e.g., high level of monitoring of organic matter enrichment across all regional sea). In addition, we consider whether monitoring networks are fit-for-purpose or sufficient for adequate implementation of the MSFD within the context of the need for better coordination, harmonization of methodologies, and cost-effectiveness considerations (Patrício et al., 2016b). This allowed us to explore different innovative monitoring approaches. Below we discuss these new approaches in terms of their potential applications to some of the 11 descriptors of the MSFD investigated by DEVOTES, in order to evaluate their broader applicability to future marine environmental monitoring.

As indicated above, we currently face a “paradox of environmental assessment”—with increasing monitoring requirements set against a backdrop of decreasing budgets. This paradox ensures the need for more cost-efficient and effective monitoring, and may eventually produce cheaper traditional monitoring, especially where monitoring requirements span large areas, as in the MSFD. The paradox requires wide-scale and rapid surveillance techniques, including innovative tools such as genomic approaches, remote sensing and acoustic sensors.

The European Commission (2010) identified 56 indicators to consider when evaluating environmental status, but at least an order of magnitude more indicators already exist (Berg et al., 2015). Despite this, many of these indicators are variants on similar themes and hence measure related attributes, and are often geographical derivations, for example the health of seabed communities. The relevance and availability of indicators vary substantially among regional seas and their subdivisions; however, the MSFD provides no guidance on integration principles, despite multiple approaches to aggregating indicators whose selection may produce highly diverging results (Borja et al., 2014). These choices challenge the scientific community to develop harmonized approaches for integrating these indicators to compare across different assessment areas.

The Ocean Health Index (OHI; Halpern et al., 2012) was developed to assess the consequences of human impacts as well as societal benefits by calculating a weighted average of scores for pressure, status and resilience goals in different areas globally. Borja et al. (2011) were the first to address specifically the challenges of the MSFD, using weighting averaging principles for integrating indicator information. The MARMONI (Innovative approaches for MARine biodiversity MONItoring and assessment of conservation status of nature values in the Baltic Sea) assessment tool (Martin et al., 2015) then used an aggregation principle based on the hierarchical structure laid out by the European Commission (2010), rather than using aggregation approaches based on the structures of marine ecosystems. Nevertheless, all assessment methods standardize indicators to a common scale prior to aggregation (Borja et al., 2016). This standardization relies upon defining of targets or reference states, which MSFD describes as targets for GES (Borja et al., 2013). The OHI uses the relative deviation from a reference state, whereas the MARMONI tool uses a binary scoring system to determine whether GES has been achieved (score of 100) or not (score of 0). However, these standardization approaches do not always achieve translating indicator values to a common scale. A relative deviation from a reference state of 50% could indicate a minor human disturbance for one indicator but a major human disturbance for another. Similarly, a binary standardization approach does not differentiate between whether minimal attainment or high status level of GES was achieved.

We developed and released software for NEAT (Nested Environmental status Assessment Tool; freely available at: www.devotes-project.eu/neat), to overcome some of the deficiencies of current integrated assessment tools (e.g., aggregation of multiple indicators at multiple temporal and spatial scales; absence of uncertainty determination, etc.) NEAT is loosely based on previous tools (Andersen et al., 2014, 2016) and translates indicator values to a common scale ranging from 0 (worst possible status) to 1 (best possible status), with 0.6 defining GES or the good-moderate boundary according to the WFD. Similarly, NEAT also allows users to set boundaries representing high-good status (value of 0.8), moderate-poor status (value of 0.4), and poor-bad status (value of 0.2). It also employs stepwise linear interpolation between these fixed points to produce transformations with a high degree of flexibility spanning the entire scale (0–1) and in which 0.6 always represent GES. In comparison with the OHI and MARMONI tool, this transformation produces a more comparable scale for integrating standardized indicator values. NEAT also employs weighted averaging of standardized indicators, but bases averaging on ecosystem features to represent the whole ecosystem. The approach primarily divides the entire ecosystem into multiple Spatial Assessment Units (SAU) that are nested to define a hierarchy of SAUs. Habitat information and relevant indicators according to organism groups are used to describe the environmental status which the given habitat may enter at different levels of the hierarchy, depending on the spatial representativity of the indicator and organism. First, averaging aggregate indicators at the organism level to produce a more even representation of relevant organism groups, i.e., to avoid an assessment biased by many indicators for the same organism group, before aggregating across habitats and SAU (Clark et al., 2011). Spatial information of the different SAUs, if provided, is used for weighting and habitats can be prioritized to weight complex habitats such as vegetated sea bottoms more heavily than deep, muddy sediments. In addition, NEAT indicators are associated with the different MSFD descriptors, supporting assessments based on various descriptor combinations (essentially from one to all).

Application of NEAT to 10 case studies across European marine waters with very different challenges, environmental conditions, and scales (Uusitalo et al., 2016b) highlights its flexibility adapting to these very different cases. This also highlighted the need for careful evaluation of the indicator set, their GES boundaries, and the selection of the SAUs, all of which can increase the accuracy of the GES assessment.

Finally, NEAT includes an uncertainty assessment at all levels of integration based on the propagation of errors (uncertainties) associated with the provided indicator information (Uusitalo et al., 2015). Therefore, assessing the confidence in the integrated assessment requires including an indicator value with an estimate of the standard error of that indicator value. Noting that few studies report or even determine the standard error of an indicator value, Carstensen and Lindegarth (2016) provide a framework for quantifying indicator uncertainty to enable such calculations. Knowing the distributions of the indicator estimates enables the calculation of the distribution of the standardized indicators as well as their aggregated values.

A true ecosystem approach for ocean use management requires an integrative assessment of marine water status. In this way, NEAT provides a second-generation, integrated assessment tool that builds on the hierarchical structure of marine ecosystems and the organisms inhabiting different compartments within this structure, thereby improving upon previous tools. Such a hierarchical approach allows users to interrogate the results to understand the reasons for the failure or success at achieving GES. However, the integrated assessment is only as good as indicator information allows, and missing or omitting information on specific groups (e.g., biological components or descriptors relevant to the assessed area) can bias the assessment results. Therefore, managers should produce guidelines stipulating indicator minimum requirements [e.g., type, coverage (ecosystem components, area, etc.), number] and the integrated assessment tool should clearly indicate if there is non-compliance with such guidelines. Moreover, because NEAT includes a comprehensive uncertainty assessment, researchers should incorporate this information as part of their interpretation of outcomes and decision support, thus needing guidelines for confidence levels of decisions.

In conclusion, environmental managers must assess the status of marine waters, not only to comply with current legislation (i.e., MSFD, WFD), but also to determine how far from targets marine ecosystems may be. Such information will allow managers to make informed decisions on sustainable resource use and the adequate restoration of degraded systems.

Inevitably, new legislative framework directives bring about unforeseen challenges to the different stakeholders who need to be involved in their implementation, particularly when first applied. Managers already apply the MSFD, which is itself complex, to complex, heterogeneous, and dynamic environments. Furthermore, initiation of the MSFD coincided with a period of a growing, global economic crisis. The numerous objectives can potentially conflict with one another from the perspective of different government departments within the Member States and also between Member States sharing a regional sea. The MSFD legal status and implementation deadlines demand that scientists and decision makers ensure a collaborative and multidisciplinary approach to deliver multi-sectoral objectives that test the abilities of existing institutions. The rapid identification of the issues, and the problems that they can create, can help those responsible for MSFD implementation to consider best how to address such issues and ensure that the MSFD can provide the intended sustainable environmental benefits.

The introduction of complex and integrative environmental legislation such as the MSFD also inevitably incurs additional costs, such as establishing new monitoring and improving existing monitoring of multiple indicators across European seas. This demand can be economically challenging. Policy makers and regulators in all EU countries are obliged to manage their resources carefully and hence they will seek to comply with the MSFD in the most cost-effective way. Yet they have many choices on which types of monitoring to apply as they select the approaches that best comply with the legislative needs within the limits of their budgets (Veidemane and Pakalniete, 2015). Although the MSFD does not require consideration of the socio-economic aspects of monitoring, Borja and Elliott (2013) noted that limited financial resources represent the most significant threat to ensuring adequate monitoring.

Furthermore, the law requires that EU countries determine whether they need new management measures and monitoring schemes to enable them to achieve GES and, if so, to implement them. Here, the socio-economic analysis of the use of marine waters, the cost of present-date degradation of the marine environment, and the cost-benefit analysis of implementing monitoring and new management measures required under the MSFD could motivate Member States to achieve GES. However, whilst a dominant tool of all governments, economic analysis approaches to achieve such analyses specifically for the MSFD in relation to the marine environment and its management were not developed at the start of the MSFD process.

Ensuring that Member States implement sufficient and non-overlapping measures to achieve GES will require the continuous review of legislation and policy, and the assessment of its implementation (Boyes et al., 2016). Furthermore, effective stakeholder engagement is likely to facilitate the acceptance of the measures and associated costs. The effective application of the MSFD requires knowledge and databases but these currently are limited by economics (Borja and Elliott, 2013). It remains to be seen how the different Member States identify the additional measures needed to improve the marine environment toward GES and close the gap between current status and GES in 2020. However, Member States must use existing budgets carefully to avoid further economic hardship resulting from financial penalties due to legal infraction proceedings in the European Court. For example, reduced funding for monitoring, if not guided toward more effective monitoring tools (see Section Monitoring Networks in European Regional Seas: Is Traditional Monitoring Sufficient to Assess the Status of Marine Ecosystems?), can reduce the quality of monitoring (e.g., by reducing spatial and/or temporal coverage). Such reduction can ultimately entail a greater cost than investment in monitoring as inaccurate evaluation could increase the risk of decision-making errors, potentially resulting in reduced ecosystem services and a devaluation of ecosystem benefits. Political decision makers may consider other aspects of monitoring as societally important, such as maintaining a bank of knowledge, technological development, professional skill and experience development and enhancing public engagement. Tools for determining the cost-effectiveness of monitoring and of management measures, as well as use of the ecosystem service approach to determine ecosystem benefits in cost-benefit analysis can support decisions on activities undertaken to comply with the MSFD. However, many member states implementing the MSFD lack both the data required to underpin rigorous economic analysis of costs of monitoring and the valuation data for assessing changes in ecosystem benefits from improvements in ecosystem services. Furthermore, the relationship between MSFD indicators and ecosystem services still requires better understanding and the implementation of the MSFD still urgently requires such data and information.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.