The Romanian coastline along the Black Sea extends from the northern border with Ukraine to the southern border with Bulgaria, spanning a length of 244 km, representing 6% of the total Black Sea coast. The Black Sea receives a significant influx of freshwater annually from various rivers, including the Danube, Dniester, Dnieper, Don, and Bug. The Danube, Europe’s second-largest river basin after the Volga, contributes the most to the Black Sea, accounting for 55% of the freshwater input. The Danube has a total river basin area of 801,463 km², covering approximately 33% of the Black Sea basin and traversing the territories of 19 countries ICPDR, 2005). Originating in the Black Forest Mountains of Germany, the Danube flows 2,857 km before reaching the Black Sea, with an average flow of 194 km³/year out of the total freshwater input of 350 km³/year ICPDR, 2005; Mee, 1992) (Panin and Jipa, 2002) This substantial contribution significantly influences the hydrography, chemistry, and biology of the entire north-western region of the Black Sea (Mee, 1992).

The continental shelf in the North-Western sector of the Black Sea has the largest expansion in the entire basin, with the 100 m isobath moving up to 180 km away from the shore. This expansion is due to the substantial sediment brought by the hydrographic network and the basin’s configuration. Near the Romanian shore, the continental shelf narrows from north to south, placing the 100 m isobath at 180–200 km in the northern sector and 100–110 km in the southern sector (ESPON, 2013). Romania’s maritime zones, including territorial waters, contiguous zone, and Exclusive Economic Zone, cover an area of approximately 29,600 km². These zones overlap with the continental shelf (0–200 m depth) by about 75%. The maximum depth in the Romanian sector reaches 1500–1700 m (Monitorul Oficial, 2023).

The ongoing study incorporated diverse data sources, including information on pressures from activities occurring on land, at sea, and offshore involving the release or emission of pollutants into the atmosphere, water, or soil. While some data sets are accessible to the public (https://sdi.eea.europa.eu/), others were acquired from the National Romanian Water Authority ( Table 1 ; Supplementary Table 1 ). Discharge data were used for activity characterization (step 2) and pressure assessment (step 3).

The methodology described in Figure 1 has the following steps:

1. Inventory of the human activities and utilization of marine ecosystem - the inventory process undertaken within the Romanian Exclusive Economic Zone (EEZ) ( Figure 2 ) consisted of identifying activities, identification of significant pressures (according to WFD) and key pressures. The process was significantly enriched by actively involving stakeholders, mainly Romanian Water Authorities. The method ensured a thorough understanding of the Romanian EEZ, laying a solid foundation for the next steps. An Excel spreadsheet that includes all existing activities (N=40) and incorporates spatial data for those with direct discharge into the Black Sea (N=8) was created. Additional information about the coastal defence and flood protection sector - beach nourishment works was collected from the National Romanian Water Authority (https://dobrogea-litoral.rowater.ro/?page_id=551). The format was designed by the project partners (National Institute for Marine Research and Development, National Water Authority, Local Water Authority and Ministry of Environment, Water and Forests throughout three virtual workshops (Zoom) held on April 6, April 29, and May 14, 2020.

2. Characterization of identified pollution sources uses and activities that may generate pressures and associated indicators/data – this step was focused on analyzing the types of emissions from identified activities. This step involved quantifying emissions to understand their impact on the marine environment ensuring each activity was accompanied by detailed emissions data, where available ( Table 1 ; Supplementary Table 1 ). In the inventory, the sources of pollution are classified as potentially significant pressures (by the Local Water Authority) by applying the set of criteria presented below:

3. Pressures assessment – the assessment involved comparing each discharge within the Romanian EEZ against environmental legislation and analyzing the ecological status data from the WISE Freshwater portal. This step aimed to identify discrepancies between emissions and legal standards, finding areas where activities might pose significant environmental risks. The pressures on the marine environment were classified and assessed according to Annex III (Table 2A) of Commission Directive 2017/845 (EC, 2017b).

4. Impact analysis – ecosystem components status against targets, limits, and legislation.– this step involves a detailed analysis of various components of the marine ecosystem status, both from pelagic and benthic habitats – phytoplankton, zooplankton, phytobenthos, macrozoobenthos and abiotic conditions from seawater and marine sediments. The ecosystem components were analysed according to ANNEX III Indicative lists of ecosystem elements, anthropogenic pressures and human activities relevant to the marine waters, Table 1 Structure, functions and processes of marine ecosystems of Commission Directive 2017/845 (EC, 2017b). This step is crucial for bridging the gap between recognising environmental pressures and comprehensively evaluating their actual and potential effects on marine biodiversity, habitat integrity, and ecosystem functions.

5. Evaluation of pressure – impact casual relationship - matrix of habitats sensitivity - The evaluation of the connections between pressures and the ecosystem components relies on the ecosystem components’ sensitivity matrix, which provides a systematic framework for evaluating the sensitivity of various marine ecosystem components to different stressors or disturbances ( Figure 3 ). The matrix takes the form of a table with rows designated for the pressures as outlined in Annex III (Table 2A) of Commission Directive 2017/845 (EC, 2017b) and columns corresponding to different habitat types and species from the studied area. Thus, each cell contains the product of ranks as described in (Halpern et al., 2007) ( Supplementary Table 2 ) for features (surface, frequency of apparition, functional impact, resistance, and recovery) within each cell at the intersection of a specific pressure and ecosystem component ( Supplementary Table 3 ). The values were standardised to fall within a range of 0 to 1.

6. Prioritisation of pressures - Semi-quantitative modelling was done with Mental Modeler software to obtain fuzzy cognitive maps (FCM) and metrics about the correlation between activities, pressures, and ecosystem components. Mental Modeler is a decision support software that helps experts understand the impact associated with environmental change and develop strategies to mitigate unwanted outcomes by capturing, communicating, and representing knowledge (Gray et al., 2014). Fuzzy-Logic Cognitive Mapping (FCM) represents a structured form of concept mapping that allows for the creation of qualitative fixed models, which are then transformed into partially quantitative dynamic models (Gray et al., 2013). By building the FCM, Mental Modeler allowed us to develop the semi-quantitative Activities – Pressures – Impact model that defines essential components and the strength of relationships between them and may run scenarios which determine how the system reacts under certain conditions. FCM was built by graphically representing knowledge by defining three characteristics of the system:

Thus, weight factors were set for activity-pressure links at 0.8 for “high probabilities”, 0.5 for “moderate probabilities”, and 0.2 for “low probabilities”. In contrast, pressure-impact links were weighted with values calculated in the habitats’ sensitivity matrix.

7. Cumulative impacts assessment - The cumulative impacts assessment process involves the use of a sensitivity matrix to methodically assess the impact of various pressures on marine ecosystem components along the Romanian Black Sea coast. By summing the scores from the sensitivity matrix for each habitat, a cumulative impact score is derived, offering a comprehensive measure of how these pressures collectively affect different marine habitats.

8. Scenarios development - The Shared Socioeconomic Pathways (SSPs) encompass climate change projections outlining anticipated global socioeconomic shifts until 2100 (Riahi et al., 2017). These pathways are the foundation for formulating greenhouse gas emissions scenarios under diverse climate policies (Pinnegar et al., 2021). Utilised prominently in the IPCC Sixth Assessment Report on climate change in 2021, the SSPs offer descriptive narratives depicting alternative socioeconomic trajectories (O’Neill et al., 2020; Rozenberg et al., 2014; Nilsson et al., 2017). These storylines articulate a qualitative understanding of the logical connections among various narrative elements. Employing various Integrated Assessment Models (IAMs), the SSPs can be quantified to investigate potential future trajectories encompassing socioeconomic and climate dimensions (Kok et al., 2019; Nilsson et al., 2017). Out of the initial five scenarios, we investigated three: SSP1, known as “Taking the Green Road”; SSP2, named “Middle of the Road,” and SSP5, “Taking the Highway”. We defined the narratives of each scenario according to the literature. Thus, the story of SSP1, “Taking the Green Road” for the coastal Black Sea region centers on a future in which sustainable development is prioritized. In this scenario, there is a concerted effort towards environmental preservation, renewable energy adoption, and a strong focus on mitigating the impacts of climate change (Riahi et al., 2017; Nilsson et al., 2017), and stringent environmental regulations and policies to preserve the marine ecosystem. It could involve initiatives to drastic pollution reduction and protect biodiversity, and promote eco-friendly tourism practices. Coastal communities might have extended and affordable investments in renewable energy sources, such as wind or solar power, reducing reliance on fossil fuels and minimising carbon emissions. Additionally, SSP1 strongly enhances the international collaboration among Black Sea countries to actively address common environmental challenges, fostering cooperation for marine conservation and sustainable fisheries management. The narrative would likely portray a region where economic growth is balanced with environmental conservation, ensuring a resilient and thriving coastal ecosystem for future generations. The scenario looks like the Black Sea’s riparian countries embrace the Green Deal initiatives – Fit for 55, European climate law, EU strategy on adaptation to climate change, EU biodiversity strategy for 2030, Farm to fork strategy, European industrial strategy, Circular economy action plan, Batteries and waste batteries, A just transition, Clean, affordable, and secure energy, EU chemicals strategy for sustainability and Forest strategy and deforestation (Ciot, 2022; Vela Almeida et al., 2023).

In the narrative of SSP2, “Middle of the Road” (Riahi et al., 2017; Nilsson et al., 2017), the Black Sea region follows a trajectory characterised by moderate changes and a balance between economic progress and environmental concerns. In this scenario, the approach to coastal development might involve a mix of sustainable practices and some degree of conventional methods. Efforts to address environmental concerns could exist, but they might not take centre stage compared to SSP1. There might be moderate regulations on industries and tourism activities along the Black Sea coastline, aiming for a balance between growth and conservation. Coastal communities may see some adoption of renewable energy sources, but the region might still rely significantly on traditional energy sources. This scenario could depict collaborative efforts among Black Sea countries, although they might not be as ambitious or comprehensive as those in SSP1. Overall, the narrative of SSP2 for the coastal Black Sea suggests a path where environmental considerations coexist with economic development, but without the same heightened emphasis on sustainability as in SSP1.

In the narrative of SSP5, “Fossil-fuelled Development” or “Taking the Highway” (Nilsson et al., 2017; Riahi et al., 2017), the coastal Black Sea region undergoes a trajectory marked by prioritising rapid economic growth and energy-intensive development, often at the expense of environmental concerns. This scenario might showcase extensive industrialisation along the Black Sea coastline, heavily relying on fossil fuels for energy generation and industrial activities. Immediate financial benefits may take precedence over environmental laws and conservation initiatives, which can result in more pollution and the deterioration of coastal habitats. Coastal communities may witness significant infrastructure development for industries and transportation, potentially impacting natural habitats and marine biodiversity. Renewable energy adoption might be limited, with the region largely dependent on conventional and less sustainable energy sources. Individual economic and territorial ambitions might overshadow collaboration among Black Sea countries, potentially leading to conflicts over resource utilisation and environmental degradation. The narrative of SSP5 could portray a future where short-term economic gains outweigh long-term sustainability, potentially posing challenges to the health and preservation of the coastal Black Sea environment.

In the context of addressing climate change across all three scenarios (SSP1, SSP2, and SSP5), consider the guiding pathway RCP2.6, aligning with the ambitious objective of limiting global warming to below 2.0°C until 2050, as set forth by the Paris Agreement (IPCC, 2014; UN, 2015).

The information about the marine ecosystem was collected through six expeditions (in 2017 – March, July, and November; in 2018 – July and September; and in 2019 – August) carried out with R/V “Steaua de Mare”. During these expeditions, we conducted the sampling of water, sediment, and biota from a network comprising 45 stations located on the Romanian shelf and covering three of the marine reporting units (MRU) according to the Marine Strategy Framework Directive (MSFD) (European Union, 2008) (Figure 2). The samples were analysed within NIMRD (National Institute for Marine Research and Development) laboratories, using specific methods for each parameter.

The pollution sources inventory analysis reveals 40 activities regarding the physical restructuring of rivers, coastline, or seabed (water management), extraction of living and non-living resources, production of energy, urban (direct discharging of wastewater from treatment plants) and industrial uses (refinery, ports, shipyards), tourism and leisure. Of these, 12 are coastal villages without sewage and wastewater treatment facilities. All sources are situated along the central-southern Romanian Black Sea coast, primarily in Constanta County. This region boasts roughly 70 kilometres of coastline, with 40 kilometres for tourism (beaches). Approximately 400,000 residents inhabit the broader coastal area covered by protected zones, attracting over 1.3 million visitors annually (Nuno, 2022).

The characterisation of pollution sources relies on a dual assessment: tracking sources by emissions and categorising them as insignificant or significant in terms of their pollution potential, utilising the methodology outlined by the Water Framework Directive (WFD) (EU, 2014). Thus, in this area, there were identified as potential significant sources of pollution seven big human agglomerations (>10,000 p.e.) connected to the sewage system and four wastewater treatment plants (WWTP) with direct discharge into the Black Sea, eight human agglomerations with 2,000–10,000 p.e. from which only four are connected to the sewage system and one WWTP and five industrial sources (ports and shipyards, refinery, oil and gas exploitation rigs). Four agglomerations with 2,000–10,000 p.e. and eight with less than 2,000 p.e. with agriculture as a primary activity are not connected to any sewage system or WWTP (River Basin Management Plan 2016–2011 available at https://dobrogea-litoral.rowater.ro/?page_id=469). Additionally, temporary activities include coastal protection and beach rehabilitation, natural resource exploration (oil and gas), seasonal population increase due to intensified summer tourism (14 resorts), and, more recently, the side effects of the war in Ukraine (e.g., an increase in the number of ships in the Romanian area) (OECD, 2022). This assessment does not take into consideration the Danube’s input.

Based on the available emissions data, the primarily monitored pressures encompass substances, litter, and energy themes. These pressures predominantly involve the introduction of nutrients and organic matter, along with releasing substances (such as heavy metals and hydrocarbons) from specific point sources. The most recent WFD assessment within the 2^nd River Basin Management Plan indicated moderate to poor status in transitional and coastal waters. Specifically, according to WISE Freshwater information system for Europe, coastal waters deteriorated compared to the previous RBMP cycle).

The matrix of the ecosystem components’ sensitivity is a valuable tool for conservation, policy development, and scientific research ( Figure 4 ). It aids in making informed decisions about the management and protection of species and habitats, ensuring the sustainability of ecosystems in the face of various environmental challenges (Goodsir et al., 2015; Korpinen et al., 2019; Quemmerais-Amice et al., 2020). Thus, by categorising habitats based on their sensitivity, the matrix helps prioritise management efforts. Consequently, the ecosystem components identified as highly sensitive may require more immediate and targeted conservation measures to prevent degradation or loss. In the development of the Matrix of the ecosystem components’ sensitivity, was adopted the classifications for activities and pressures by referencing previous categorizations established in the Marine Strategy Framework Directive and by the designation of seabed conservation features, as mandated by the Habitats Directive 92/43/EEC. Scores are computed by integrating data from surveys and expert judgment, as highlighted in the degree of certainty feature ( Supplementary Tables 2 , 3 ) focusing on criteria such as the extent (surface), the functional impact, the habitat’s ability to regenerate, its tolerance to alterations in environmental conditions, and the reversibility of impact. High scores indicate habitats that are highly sensitive and potentially less capable of recovering from disturbance, signifying a need for prioritized conservation efforts. Conversely, lower scores are assigned to more resilient habitats, indicating a lesser immediate need for intervention. This scoring system enables a ranked prioritization of habitats based on their sensitivity, guiding more focused and effective management and conservation strategies.

Obtaining the semi-quantitative model and assessing as objectively as possible the causal relationships in which pressures on the marine ecosystem are involved allowed the establishment of the hierarchy of pressures based on the centrality score calculated by Mental Modeler software, considering all causal relationships (activities-pressures and pressures-impacts) and their intensity from the sensitivity matrix ( Figure 5 ). The hierarchy of pressures was achieved by the descending order of “centrality”, which represents the absolute value of the influence of pressure in the model. For calculating centrality, Mental Modeler measures the importance or influence of a node within a network by analysing its connections, pathways, and relationships with other nodes, aiding in understanding their impact or prominence in the overall system ( Figure 6 ). Thus, the higher this value, the greater the importance of pressure in the model. The model has 29 components – 7 drivers, 13 receivers, 9 ordinary connections, and 64 total connections with a density of 0.08, representing 2.21 connections/components ( Supplementary Figure 1 ).

The pelagic ecosystem is influenced by multiple pressures, with the introduction of invasive planktonic species constituting a 40.7% influence. Nutrient loading and the addition of other substances each have a 20.5% impact. The accumulation of litter and organic debris further exacerbates contamination levels, indicating a significant threat of biological invasions. Macrophytes face substantial threats from the addition of substances (80%) and nutrients (20%). Coastal environments, including the littoral rock (0–1m), littoral sediment (0–1m), mediolittoral coarse or medium clean sand (0–1m), and upper infralittoral rock (3–10m) habitats, experience pronounced effects (99%, 96%, 97%, and 92% respectively) from physical loss due to beach nourishment activities. The infralittoral rock habitat accumulates impacts from various stressors, predominantly from physical loss (67.2%), followed by the introduction of non-native species (14.9%) and impacts from bottom trawling fisheries (14.9%). These fisheries also affect the infralittoral coarse, mixed, sand, and mud sediment (1–20m) habitats, where they contribute to a 27.5% impact, alongside physical seabed disturbances (4.9%) and a significant threat of physical loss (67.6%). In the Mytilus galloprovincialis beds located in circalittoral mud and mixed sediments (30–60m), a combination of pressures significantly impacts the environment. Here, substance input is the predominant issue (35.6%), followed by the introduction of the non-indigenous Rapana whelk (23.7%), and equal inputs of litter and organic matter (11.9% each). Although the intensity of bottom trawling decreases offshore, its impact, including the extraction of wild species (8.5%) and physical seabed disturbances (8.5%), remains notable. The deep circalittoral shelly mud habitat, inhabited by Modiolula phaseolina (60–120m), primarily suffers from substance input (66.7%) and litter (33.3%). Bird populations are predominantly disturbed (93.3%) by various anthropogenic activities, which disrupt natural behaviours, coupled with a 6.67% impact from litter. Mammals are notably affected by accidental captures in fisheries (79.8%), along with substance input (6.7%) and litter (13.5%), highlighting the consequences of human exploitation and incidental harm. Similarly, fish populations are significantly impacted by direct extractions or mortality/injury (92.3%), with additional pressures from litter (3.3%) and substance input (3.3%), indicating the extent of anthropogenic stressors on these species ( Figure 7 ).

Knowing the sources, pressures and impacts on prevalent ecosystem components and their cumulative effect facilitates the development of scenarios, allowing researchers and policymakers to assess potential future outcomes based on different intervention strategies according to each narrative (OECD, 2016). This foresight is crucial for informed decision-making and long-term planning. In response to these scenarios, we have identified specific variables for each respective Key Performance Indicator (KPIs), which serve as guiding metrics to assess and track environmental impact and sustainability within each scenario ( Table 2 ). The percentage values associated with each scenario typically reflect the likelihood or desirability of that future pathway based on narratives, expert judgment and Romania’s profile (https://ssp-extensions.apps.ece.iiasa.ac.at/profile/Romania/development). Thus, KPIs in SSP1 are strongly reduced based on concerted efforts toward environmental preservation and climate change mitigation, alongside stringent policies, drive significant pollution reduction and biodiversity protection in the marine ecosystem. KPIs in SSP2 reflect a scenario where environmental considerations and economic development proceed together, though without the intense focus on sustainability found in SSP1. Finally, for SSP5, KPIs are aligned with prioritizing short-term economic gains over long-term sustainability, which could challenge the health and conservation of the coastal Black Sea environment.

Changes were made individually to each pressure ( Table 2 ) ( Figure 5 ), and the Mental Modeler produced corresponding values ( Figure 8 ) based on the initial model. The Mental Modeler scenario interface ( Supplementary Figures 2, 4 ) shows the relative change in the components included in the model, influenced by the defined relationships in the FCM for each selected scenario. Running scenarios involve changing variables from −1 (significant negative change) to +1 (significant positive change). As a result, a bar graph visually represented the relative system change, displaying how components could respond within the specified scenario ( Supplementary Figures 2, 4 ) (Gray et al., 2013).