The first step of the analysis is to develop an understanding of the key elements and needs of an MSP process as well as social-ecological analytical frameworks, which shape impact analysis ( Figure 1 , step one). The second step is to analyze two existing approaches for i) quantification of ecosystem service supply following a linkage framework approach and ii) assessment of cumulative impacts based on the widely used Halpern et al. (2008) approach, as well as the tools that facilitate them ( Figure 1 , step two). Based on this synopsis a conceptual framework for an integrated, multi-map toolset for the analysis of relative cumulative impacts on ecosystem service supply is proposed ( Figure 1 , step three) and illustrated with an example ( Figure 1 , step four).

The MSP process has some defined phases or stages, some of which are focused on setting the objectives, others on analyzing the possible outcomes as well as implementing and monitoring the plans while continuously involving stakeholders and drawing on various types of knowledge, data, and information (Giacometti et al., 2020). MSP is place-based, and places serve different purposes and mean different things to different actors, thus, negotiation, discussion, and common objective setting is a key part and principle of MSP (Ehler et al., 2019). In addition to organizing the planning process and stakeholder engagement, MSP planners are also tasked with analysis of current conditions and future states of the ecosystems and society, which involves stock-taking, what-if scenarios, and trade-off analysis (Douvere and Ehler, 2011; Gissi et al., 2019; von Thenen et al., 2020a). Planners then prepare, approve, and implement the plan and monitor, evaluate, and adapt it, which then leads to the next planning cycle (Douvere and Ehler, 2011). MSP is a continuous, adaptive, and incremental process, and alternative routes to reach the objectives are explored during the plan development, the results of which inform measures and actions (Douvere and Ehler, 2011; Piet et al., 2020).

Scenario building is a way of exploring causal processes and evaluating potential outcomes against goals and objectives of the plan (McGowan et al., 2019). Scientific knowledge about the ecosystem, its dynamics, and an understanding of the functional relationship between people and the ecosystem is essential for scenario analysis and the implementation of ecosystem-based MSP (Bednarek et al., 2018; UNESCO-IOC/European Commission, 2021), that is, decision-making whereby resource demand and consumption, as well as economic development, is maintained within environmental thresholds and supports achieving good environmental status (Frazão Santos et al., 2014; HELCOM, 2020).

Understanding the complex interdependencies between ecosystems and human societies is important for many sustainability problems (De Vos et al., 2019; Schlüter et al., 2012; Schlüter et al., 2019). Such systems thinking – an issue-oriented approach to problem solving – recognizes the complexity of social, ecological, and economic phenomena (Davies and Stroink, 2015, Hopkins et al, 2011). Hereby, these phenomena, or sustainability problems, are not addressed in isolation but are viewed as part of social-ecological systems (SES), in which people and their environment are intrinsically interconnected. Various analytical frameworks have been developed to breakdown these complex systems into bitesize pieces in order to facilitate analysis, communication, and problem solving (Patrício et al., 2016; Potchin-Young et al., 2018; Delacámara et al., 2020). The Driver-Pressures-State-Impact-Response (DPSIR) was one of the first frameworks to identify the indicators for and describe pressures and impacts of human activities on the environment to inform decision-making (Binder et al., 2013; Gómez et al., 2016). DPSIR was introduced in the late nineties and since then has become one of the most widely used and adapted analytical frameworks in the context of marine environmental risk assessment and management (Patrício et al., 2016). According to the Patrício et al. (2016) study of the evolution of DPSIR-type frameworks, the most recent and elaborate adaptation of DPSIR for marine ecosystem-based management is the Drivers-Activities-Pressures-State-Impact (on human Welfare)-Response (as Measures) known as the DAPSI(W)R(M) framework. It defines drivers as the demand from society to fulfil basic human needs that result in human activities, which create pressure mechanisms through which the state of the ecosystem is changed (Patrício et al., 2016). A single activity can generate a whole range of pressures, which generate stressors resulting in cumulative impacts – impacts greater than the sum of its individual parts – affecting the ecosystems state immediately or over a longer period of time (Reckermann et al., 2022; Directive 2008 /56/EC). Changes in the state of the ecosystem include changes in species abundance or habitat composition, which in turn impact human welfare, for instance results in a decrease in food supply or natural flood defense (Patrício et al., 2016). Changes in the ecosystem state and impacts on human welfare are met with responses in the form of management measures, that is, actions targeted at controlling drivers, activities, and impacts on human welfare of the DAPSI(W)R(M) framework (Elliott and O’Higgins, 2020).

The ecosystem services (ES) approach shifts the focus from the effects of human activities on ecosystems to how we benefit from ecosystems, and aims to quantify these benefits (Sharpe et al., 2020; Elliott and O’Higgins, 2020). One of the most widely used conceptualizations of ES is the ES cascade, which represents human-environment interaction as a continuous process of ecosystem service supply and societal demand for services (Potchin-Young et al., 2018). Ecosystem services have generally been grouped into three types – provisioning, regulating and maintenance, and cultural – all of which play different yet incremental roles in maintaining our material well-being, physical and mental health, and other key factors determining our life quality and livelihoods (Alcamo et al., 2003; Potschin-Young et al., 2018). The ecosystem services element of an ecosystem services cascade has been seen as synonymous with “Impacts” in the DAPSI(W)R(M) framework, which has resulted in the substitution of the term “Impacts” to “Ecosystem Services”, and the development of ecosystem-based management-focused analytical frameworks such as the butterfly model (for more detail on the butterfly model see Atkins et al., 2011; Gómez et al., 2016, Elliott and O’Higgins, 2020).

MSP-SA introduced by Frederiksen et al. (2021) is another example of an analytical framework, which has adapted elements of DPSIR-type models and the ecosystem services cascade. It was designed to draw attention to economic, environmental, and the considerably under-considered social consequences of MSP (Frederiksen et al., 2021). In order to broaden the scope of impact analysis, MSP-SA divides “Impacts” of the DPSIR framework into three types corresponding with some of the elements of the ES cascade: ecosystem capacity; ecosystem services; and benefits, disbenefits, and the distribution of benefits and beneficiaries over time and space (ibid., Figure 2 ).

MYTILUS tool scenario functionalities were used to assess cumulative impacts for the Gulf of Riga, a basin in the eastern Baltic Sea. This was seen as an appropriate test case site because MYTILUS has a ready-made dataset sourced from HELCOM (as described in section 2.3.2) equipped to facilitate cumulative impact assessments for the Baltic Sea region. A key advantage of using MYTILUS is that the outputs can be exported in raster format for further geospatial analysis outside the MYTILUS environment, for instance using freely available QGIS software. The Gulf of Riga was also chosen as a test case because an ecosystem service supply assessment has already been conducted for the basin and the results of the assessment – a linkage diagram and numerical relative description of ecosystem component contribution in service supply and ecosystem functioning – are readily available in Armoškaite et al. (2020).

The assessment of “reference” and “new” scenario additive cumulative impacts and mapping of relative ecosystem service supply were first done separately as explained below, and then the results were combined to reveal the relative change in ecosystem service supply as a result of change in cumulative impact index scores.

Cumulative impacts and pressures analysis of the Gulf of Riga via MYTILUS suggests that the Gulf of Riga ecosystem is under several types of pressure ( Figure 4 ). Figure 4 suggests that the first and second largest pressures in the basin are phosphorus and nitrogen inputs, which are related to terrestrial activities. Third is hazardous substances, followed by introduction of non-indigenous species, fishing of herring, and disturbance of species. Fishing of herring was seen as a suitable variable for the test case as it is a pressure associated with a single source and therefore can be controlled through management of one activity.

Spatial decision-support tools can provide indispensable support to planners, scientific experts, and other actors involved in the MSP process (Stelzenmüller et al., 2013; Depellegrin et al., 2021; Keenan and Jankowski, 2019). Once made operational, the multi-map toolset presented in this paper would provide the possibility to conduct geospatially explicit assessments of ecosystem service supply and cumulative impacts in a single environment. It would be a platform for visual analysis and exploration of the different outputs foreseen and presented in the conceptual framework. For instance, cumulative impact and ecosystem service supply maps for a specific area could be compared to identify locations under high/low levels of stress due to cumulative anthropogenic impacts as well as locations low/high service supply. This paper builds on the Hammar et al., 2020 approach applied to integrate cumulative impact assessments in Swedish MSP by taking the relative cumulative impact index change values and applying them in geospatial ecosystem service analysis. The toolset proposed in this paper shall enable the user to assess cumulative impacts on ecosystem service supply providing an insight into impacts on the ecosystem state (ecosystem components and functioning) as well as impacts on human wellbeing through changes in multiple service supply, thereby providing a more complete representation of the “supply” side of the social-ecological systems butterfly model.

The outputs of an ecosystem service supply assessment following Armoškaite et al. (2020), that is, quantification of service supply in relative terms, mapping and display of results in linkage web diagrams, have already informed an ecosystem services and benefit valuation study (see Pakalniete et al., 2021), evidently facilitating communication across disciplines and paving a way for a more harmonized link between service “supply” and “demand” analysis. This suggests that the outputs of the toolset will also facilitate cumulative impacts on ecosystems and change in service supply and lay the foundations for further benefits, and potentially beneficiaries and benefit distribution analysis (Frederiksen et al., 2021).

Whereas before change in relative ecosystem service supply has been assessed based on expert knowledge (Armoškaite et al., 2020) or geodata derived from field observations (Armoškaite et al., 2021), the approach for linking cumulative impact assessments results and service supply presented in this paper provides an alternative method – utilizing cumulative impact assessment results. This means that the toolset could cater for a wider range of contexts with respect to data availability as well as study objectives, including more explorative scenario analysis based on arbitrary estimates of ecosystem change and analysis rooted in evidence, and more thorough understanding of the interaction between drivers of change and ecosystems.

Studies of MSP-related decision-support tools have shown that very few tools support several functionalities and respond to the integrated, multi-objective nature of MSP tasks (Pinarbasi et al., 2017). The outcomes of the tool – the maps of relative changes in cumulative impacts and ecosystem service supply – could be used for further, MSP-related, analyses. In order to decide where to place or restrict certain human activities, the CIA-ES maps could be overlaid with other spatial or environmental requirements for the respective human activity. For the spatial and environmental requirements, tools such as SPACEA (von Thenen et al., 2020b) could be applied and the outcomes – also raster data files – combined or visually compared with the CIA-ES maps. This approach would highlight potential trade-offs between optimal locations for human activities and areas that are already highly impacted by other pressures or feature high ecosystem service supply. In the long-term, the functionalities of SPACEA could be integrated into the toolset to carry out the analyses in a one-tool environment.

The toolset may also contribute to holistic assessments of the state of marine areas as it provides both an idea on the cumulative impacts on the ecosystem and the implications for human well-being through the linked ecosystem services. One example of a holistic assessment of a marine area is carried out by HELCOM in the Baltic Sea. These assessments already include cumulative impacts (HELCOM, 2018). The Baltic Sea Impact Index (BSII) developed by HELCOM for this purpose is based on Halpern et al. (2008), just like MYTILUS, and MYTILUS in fact uses the sensitivity matrix from HELCOM (see above). However, HELCOM’S BSII uses a sum impact approach, whereas MYTILUS and the proposed toolset additionally uses mean impact index as well as a maximum pressure index. A further development of functionalities should include synergistic as well as antagonistic impacts, something the current version of HELCOM’s BSII does not take into account (HELCOM, 2018). HELCOM is currently in the third round of the holistic assessment of the state of the Baltic Sea. For this third assessment, ecosystem services will also be assessed, including maps of ecosystem supply areas taking into account the distribution of ecosystem components and a marine ecosystem accounting approach (HELCOM, 2021). The underlying ES matrix from HELCOM, linking the ecosystem components to the specific ES, makes use of a binary scale (HELCOM, 2021). The proposed toolset here has a different approach (following Armoškaite et al., 2020) and uses expert elicitation to specify how much the ecosystem components contribute to specific ES. As the toolset additionally makes the link to the cumulative impact assessments, it is very well suited to support HELCOM’s assessments of the state of the Baltic Sea. It can also support assessments of other sea basins or areas, as the toolset’s functionalities are generic. The input data can be changed to fit other regional contexts. However, the sensitivity scores and the contribution of ecosystem components to ES would need to be adapted as they are tailored to the Baltic Sea at the moment.

In a participatory decision-making context, maps are known to increase levels of engagement among participants as well as a sense of common interests and shared spaces (Bonnevie et al., 2023; Koski et al., 2021). Therefore, on top of supporting analysis of the ways pressures and activities may be impacting ecosystems and in turn human wellbeing, the toolset also may be used to facilitate a dialogue among researchers, decision-makers, and stakeholders through multi-map comparison and scenario or trade-off analysis. A multi-map toolset may also be advantageous or of use among students and MSP researchers for capacity building in an active learning environment (Bonnevie et al., 2022).

One key challenge the test case highlights is that the method for assessing change in service supply proposed in this paper provides a linear, simplified representation of the interactions between the drivers of change, the ecosystem, and service supply. In the test case, the assessment results do not indicate service supply loss, which would be expected in the real world given that “fish for food – herring” is one of the services making up the total supply. On the one hand, this is a result of the simplification of causal relationships and links between elements of the system, which is a limitation that comes with linear representations and non-model-based methods of social-ecological systems analysis. On the other hand, this also highlights the significance of widespread challenges related to definitions and distinguishing the differences between “activities”, “pressures”, and “services” or “services” and “ecosystem functions”, and “services” and “benefits”, which need to be addressed among other reasons to reduce likelihood of double counting in monetary valuation of services and benefits (Hyytiäinen et al., 2014). To manage these challenges, it is important not to lose sight of initial pressure (whether anthropogenic or climate (that can be viewed as natural or anthropogenic) or natural (e.g., naturally occurring introduction of invasive species)) that has generated a cascade of events. Furthermore, it is important to approach classification of the different elements of the social-ecological system systematically, and with reference to standardized classification systems (e.g., CICES) and regulatory frameworks and strategies (e.g., the MSFD).

In the test case presented in this paper, the values describing ecosystem component contribution in service supply are aggregated to estimate change in total per-cell supply. Analysis of change in the supply of a specific service of interest is also possible but requires an additional step - identification of the ecosystem components maintaining that specific ecosystem service. This entails identification of the linkage chains connecting all ecosystem components and functions related to the services of interest, followed by mapping of the select component contribution values, and application of impact index values through geospatial analysis as described in section 3.1.3.

Social-ecological systems analysis, including ecosystem services assessments, is a selective, place-based, time-sensitive process. The focus of the test case, and the pressures and services assessed, are those that were identified through HELCOM and the expert elicitation processes of Armoškaite et al. (2020). The toolset presented in this paper will accommodate analysis of cumulative impacts and service supply of any location, leaving the scope of the assessment to be determined by the objectives of the assessment as well as data availability (i.e., geodata availability, quality, and availability and extent of expert knowledge).

It is important to emphasize that, while qualitative and semi-quantitative representations of social-ecological systems are a useful source of evidence to guide the decision-making process by highlighting problematic areas for further management (Piet et al., 2020), unknown feedback and regime shifts make scenario analysis uncertain and multiple sources of evidence should be considered (Hammar et al., 2020). Therefore, all outputs featured in the toolset and the results of the test case need to be interpreted with a level of caution and would need to be used in conjunction with other evidence-bases and should not support decision-making alone.

Once developed, the toolset functionalities will need to be tested in a real-world planning context and will require fine-tuning and further development. Studies of MSP processes and the supporting role of tools have found that “whether a tool can successfully contribute to improving the overall outcomes of the MSP process at least partially depends on how it is linked to the MSP process” (Gee et al., 2019). Therefore, when developing and testing this toolset, it is important to consider practical (i.e. application in MSP, tasks in MSP process, uncertainty and output communication, and stakeholder use), conceptual (including objectives, functionalities, developers, and users), and technical (such as the programming language, software, input data, and source code availability) aspects (Depellegrin et al., 2021) as well as co-develop the toolset with decision-makers, planners, experts, and scientists from social and natural sciences as well as stakeholders from various maritime sectors.