French Guiana is an overseas department of France, bordered by Brazil in the south and east and by Surinam in the west, with a land area of 83,534 km² and a coastline of 320 km in length. Its population of around 275,000 inhabitants in 2016 is mainly concentrated within 10–30 km-wide coastal strip (Zouari, 2015). The main city is Cayenne where the main transport infrastructures of the territory are located (international airport, main commercial port) and which shelters half of the Guianese population. The two other main populated areas are Kourou where the Guianese Spatial Center is located and the estuary of Maroni river that is characterized by the highest population growth.

It is located in a equatorial climate and has all the necessary characteristics for mangrove colonization and growth, with air temperature fluctuating between 26 and 30°C and rainfall ranging from 2,500 to 3,000 mm.yr^–1 (Marchand, 2017; Walcker et al., 2018). Mangroves in French Guiana occupy almost 75% of the coastline (Walcker et al., 2015).

French Guiana’s coastline is characterized by the dynamics of its coastline, that is deeply affected by the Amazon River. Sediments eroded in the Andes are transported along the Amazonian basin down to the river mouth (Martinez et al., 2009). There, under the influence of tides, waves and currents, accumulated sediments form individual mud-banks typically extending 10–60 km along the coastline, 20–30 km offshore and 5 m thick, that move along the coast toward the Orinoco river (Venezuela) at a rate of 1.5–3.5 km/yr (Gardel and Gratiot, 2005). This dynamic geomorphology offers uncommon conditions for coastal mangroves (Walcker et al., 2018). During their formation, mud-banks are colonized by propagules carried by the tides; with the arrival of new seeds the mangrove forest accumulates in successive strips of even-aged stands. During the erosion phase, mangroves are swept away starting with the youngest forest stands, until the erosion stops or until the mangrove locally disappears with its substrate. Coastal mangroves, exposed to mud-banks migration, are mostly dominated by Avicennia germinens that are more effective in rapidly colonizing and developing on such an unstable and stressful substrate (Fromard et al., 2004).

Mangrove succession in French Guiana is also well described (Fromard et al., 1998) starting with pioneer mangroves that accumulate biomass while maturing. Mangroves are dominated by Laguncularia racemosa and A. germinens, and depending on environmental conditions, sediment can accumulate and the landward mangrove forest can turn into savannah. Under stronger riverine influence, mangrove species association also include Rhizophora spp. and can evolve into marshy forest, depending on sedimentary conditions. Excess sedimentation can also suffocate mangrove trees, leading to dead mangrove areas, that can enter a new cycle of colonization, if conditions are favorable. Such dead mangrove areas are a characteristic feature of the Guianese coast (Fromard et al., 1998). In a nutshell, mangrove in French Guiana are not homogenous and an important distinction exists, between coastal mangroves that are exposed to mud-bank migration and estuarine mangroves located under riverine influence.

Table 2 presents the graph theory indices obtained for the 19 FCMs collected after the individual interviews. Interviewees identified a total of 257 variables to characterize French Guiana’s mangrove systems, with a mean number (± SD) of 13.53 (± 6.26) variables per map and 15.26 (± 8.39) connections. Note that 5 additional variables were mentioned without any connection to other variables; they are excluded from the following calculations.

A visual analysis of the indices suggests that economic actors identified fewer variables and connections. Moreover, in comparison to other categories, they identified a very low number of transmitters suggesting a high level of complexity with few controlling forcing functions. We ran a one-sided ANOVA on the indices that showed only one statistical difference amongst our stakeholder groups regarding the number of receivers (F = 33,156; df = 18; p = 0.049): economic actors and scientists perceive a below-average number of receivers showing a low number of “outcomes” from the system.

A first homogenization of variables allowed identifying several categories in which they could be grouped. Many variables have a very low number of records: 46 variables were mentioned only once and 20 only twice or thrice. 21 concepts were mentioned between four and eight times. Finally, five variables were mentioned by more than half of respondents, namely “Nurseries” (10 times), “Coastal protection” (11), “Biodiversity” (12), “Fishing” (13), and “Mangrove” (19).

Before presentation of the results during the workshops we condensed the number of variables from 92 to 29. After aggregation of individual maps among stakeholders we obtained four stakeholder FCMs with a mean number (± SD) of 21.5 (± 4.7) variables and 34 (± 13.5) connections.

A descriptive analysis of the graph indices of the different stakeholders FCMs (Table 3) shows that the FCM from economic actors presents the highest level of complexity, i.e., stakeholders perceive more outcomes from the system than options to intervene on it to make it change (given the number of identified transmitters). FCMs from conservationists present the highest density meaning that their perception of the system is the most interconnected with many links between variables. The study of the FCMs from scientists shows a more concentrated vision of the system with fewer variables (17) and a relatively high level of complexity. Finally, managers have the most extended perception of the system with many variables and a great majority of ordinary variables.

After the workshops, the mean number of variables across FCMs increased to 24.3 (± 3.3) and the mean number of connections also increased to 40.3 (± 11.7). Standard deviation decreased for all the graph indices, except for the number of receivers.

In this section, we present the community FCMs obtained from the aggregation of our four expert groups FCMs after the workshops. The aggregated community map is presented in Figure 1 and the corresponding graph theory indices are presented in Table 3.

The high number of ordinary variables indicates numerous interactions between system components. This outcome corroborates the important number of connections identified, suggesting perceptions of a strongly integrated and interdependent mangrove socio-ecosystem. That is, the system’s dynamics result from mutual influences between the French Guiana society and the mangrove ecosystem. The low hierarchical index (close to 0) indicates a relatively “democratic system” with a high level of integration and dependence between the different components of the system. A system perceived as democratic is a sign that experts consider there are various ways to change it. Yet, the rather low density index reports a sparse map, reflecting that few management options are identified as capable of influencing the system’s dynamics.

Table 4 ranks the community map variables by order of centrality. Variables with highest centrality scores are influential within the system: i.e., they can either be highly connected to other variables or display few connections with a high weight.

The most influential variables are associated to the biological compartments of the mangrove ecosystem, namely “Mangrove,” “Mudflat/Mud-banks,” and “Biodiversity,” followed by “Nursery,” “Accessibility issues” and to a lesser extent, “Recreational activities.” Together, these elements come out as key influential dimensions of the socio-ecosystem’s functioning. In accordance, the anthropic activities “Tourism” and “Fishing” come next. Tourism is most directly supported by biodiversity and as such, increases demand for recreational services, themselves relying on biodiversity, while fishing rests on mangrove supply of fishery resources, through its habitat support function. In turn, all three anthropic activity types are concerned by accessibility issues arising from mud-banks and mangrove forest. These latest components can be detrimental to fishing vessels’ access needs to the coastline and/or to the amenity value of beach landscapes sought by (certain) tourists. The influence of mangrove on recreational activities is more ambiguous, since these activities are supported by mangrove forest (for kayaking, bird watching, hunting, hinking, etc.), while they can also be negatively affected by the limited access forest and mud-banks impose on coastal areas (preventing watching the laying of turtles on the beach for instance). Pollution is the environmental impact factor displaying the highest centrality score. It is harmful to several benefits derived from the mangrove ecosystem, via its negative impacts on biodiversity and ecological habitats.

The centrality of the subset of variables presented above may owe to the fact that they are mentioned by four stakeholders’ groups we surveyed who consider them important to the functioning of the system (Table 4). Indeed, looking at the very core of the representation shared by the expert community we interviewed (Supplementary Figure 1), it is clear that the mutual interdependence between mud-banks and mangrove forest is acknowledged by all experts, so as the role of the mangrove ecosystem in supporting biodiversity and the problems it causes in terms of accessibility. In relation to society’s demands, only fishing activities stand out, through the channel mangroves foster “Nurseries” (support service) and “Fishery resources” (provisioning service).

Other variables emerge in Supplementary Figure 1 as being mentioned by every category of stakeholders. Because of its relatively high outdegree score but low indegree score (id = 0.25 vs. od = 1.75), “Legislation” does not count as “very” central to the system’s dynamics: it is perceived as a lever for mangroves conservation with direct and positive relationship with mangroves and biodiversity. On the other hand, climate regulation is a mangrove ecosystem service with a higher indegree score than outdegree score (id = 1 vs. od = 0.5). While expert groups are not unanimous and clear on the consequences of climate change on the mangrove socio-ecosystem, all agree on the fact that mangrove ecosystems play a role in climate regulation, in particular with respect to the carbon cycle. The variable “Insects issues” presents a similar behavior (id = 1 vs. od = 0.5) showing that all stakeholders have this issue in mind notably regarding the Yellowtail Moth (Hylesia metabus) that can provoke severe allergic reactions. Other ES are mentioned by all stakeholders but their connection with other variables of the system is less unanimous (namely “Coastal protection,” “Landscape”). Finally, all stakeholders recognize the role of “Agriculture,” “Pollution,” and “Land-use planning” in the system but there is no consensus regarding the nature of this role. For example, in the case of “Agriculture”: conservationists assumed an impact of agriculture on mangroves through territory planning, and an unclear effect of agriculture on mangroves through pollution carried by rainfall runoff; managers perceived a negative impact of agriculture via deforestation; scientists through land conversion; while the overall interaction was unclear for economic actors.

All expert groups underline the role of mangroves in sustaining coastal fisheries. Indeed, the most unanimous relationship in the community FCM (Figure 1) is the inseparable link between mud-banks and mangroves, owing to “Nursery” functions necessary to maintain “Fishery resources.” “Fishing” is thus perceived as the most emblematic activity depending on the mangrove ecosystem within the French Guiana society. In 2019, 2,820 tons of fish were landed in French Guiana, worth €5.4 million (≈ US$ 6.1 million) (IFREMER, 2020). Acoupas (Cynoscion spp.) and Crucifix sea catfish (Arius proops) that represent the majority of catches of coastal fisheries (76% of total landed volume in 2019; IFREMER, 2020), spend part of their life cycle in mangrove (Rojas-Beltran, 1986; Rousseau et al., 2018). Also, mangrove fluctuation has shown to directly impact shrimp fisheries, that represented an important source of export for French Guiana (Diop et al., 2018). These numbers do not include illegal and subsistence fishing that could represent around 60% of total catches (Levrel, 2012). Mangrove conservation is thus critical for the sustainability of fishing practices, especially in the context of climate change that could lead to a collapse of both biomass of targeted species and fishing activities (Gomes et al., 2021). Meanwhile, fishing was reported by managers and scientists as an extraction activity that puts pressure on the natural environment. Fisheries can modify mangroves fish assemblages and impact their sustainability (Reis-Filho et al., 2019). However, empirically, it is currently unclear whether fishing has any impact on the quality of mangroves or the delivery of other services.

The importance of mangrove in “Climate regulation” is also recognized by all experts. Mangroves are very productive ecosystems that capture carbon from the atmosphere to develop. This carbon is then trapped and stored into the soil (Hamilton and Friess, 2018; Richards et al., 2020). In the case of French Guiana, the migration of mud-banks brings uncertainty to this role. During an accretion phase, the mud-bank is colonized by mangrove and starts to accumulate carbon (Marchand, 2017). When it enters an erosion phase, mangrove is destroyed and important quantities of organic matter are exported to coastal and offshore waters (Mongruel et al., 2018). Nevertheless, the total mangrove carbon stock at the scale of French Guiana is estimated to 23.06 ± 5.03 TgC (Walcker et al., 2018). Mangroves destruction would liberate this sequestrated carbon (Hamilton and Friess, 2018; Richards et al., 2020). As a signing party of the Paris Agreement on climate change, France has a duty in conserving this carbon stock.

There is a common agreement among our experts that mangroves have a potential role to play in the development of “Tourism” and “Recreational activities” (Supplementary Figure 1). In their review, Himes-Cornell et al. (2018) found that almost all the studies they selected provided economic values for recreation and tourism services by mangroves, confirming this role at the global level. However, we found no studies in French Guiana and experts during workshops underlined that eco-tourism in mangroves remains relatively marginal. This reflects the perception that mangrove and its biodiversity richness are identified as a potential for the development of the territory (WWF, 2017). Indeed, while only 21% of tourists come to French Guiana for leisure and 28% to discover the forest or the coast (CTG, 2016), there is a political will to increase those numbers relying on the development of eco-tourism (CTG, 2013). The final map provides the vision of a positive effect of mangroves on tourism and recreation that suggest that a development of these activities compatible with mangrove conservation is possible. However, the link between mangroves and the development of tourism and recreation is ambiguous (Figure 1). Firstly, because Recreation and Tourism encompass various practices that may be positively or negatively affected by mangroves. Certain activities (e.g., beach activities) can be negatively affected whether because the mangroves prevent access to areas of interest or because they modify the seascape. On the other hand, mangroves can also be attractive for eco-tourism and nature-related activities (e.g., faunistic observation). Secondly, because mangrove is closely related to mud-banks that are perceived as a constraint for the development of these activities. The migration of mud-banks can bring mangroves where they will be considered as a discomfort (e.g., on cities seafronts). In addition, there is currently a lack of infrastructure in French Guiana (e.g., there are only two marinas in French Guiana), developing tourism and recreation will imply new development projects.

Mangrove socio-ecosystems are also involved in the construction of coastal “Landscape” but the nature of this role is not obvious. If some experts consider that mangroves can have a positive impact on landscape, this is also negatively impacted by mud-banks and accessibility issues. Experts underline that urban seafronts are deserted when mangroves obstruct the seaview, while mangrove landscapes are appreciated in rural areas where they contribute to the identity of French Guiana. Aesthetic values of mangrove are understudied (Himes-Cornell et al., 2018). The question of the perception of the aesthetic value of a landscape is complex as it results from the link between the intrinsic characteristics of an object and its perception by an observer which is influenced by human nature, education and society (Tribot et al., 2018). According, to Tribot et al. (2018) there is a disconnection between the landscape aesthetic and the ecological value of ecosystems: mangroves are more likely to trigger negative perceptions than agricultural landscapes. The authors propose a virtuous loop in which, knowledge and experience on the functioning of ecosystems could increase mangroves’ aesthetic value that would be more likely to be protected.

The last ES mentioned by every group of experts is “Coastal protection.” The role of mangroves in providing coastal protection in French Guiana is closely associated to the dynamics of mud-banks as an alternation of “bank” and “inter-bank” phases. In bank areas, ocean energy is at first dissipated by the mud-banks, closer to the shore, the remaining energy is gently dissipated by mangroves (Anthony and Gratiot, 2012). Moreover, mangroves favor sedimentation (Furukawa et al., 1997) and enhance the resistance of the substrate during erosive inter-bank phases (Fiot and Gratiot, 2006). This process has been effective over the last 5,000 years, resulting in a net coastal progradation (Anthony et al., 2010). In addition, in inter-bank areas where higher energy waves result in mangrove destruction, mangrove trees still dissipate wave energy and thus contribute positively to coastal protection (Anthony and Gratiot, 2012). In French Guiana, coastal erosion and coastal flood are two issues well identified by the natural risk prevention plan (DEAL, 2015), which should advocate for mangrove conservation.

The experts also all pointed two disservices associated to the functioning of the mangrove ecosystem, namely “Accessibility issues” and “Insect Pests.” “Accessibility issues” reflect the fact that some coastal facilities (e.g., slipways, fishing docks, touristic seaside infrastructures) may get directly obstructed by mangroves or see their value decrease because of the negative perception of mangrove in the vicinity. “Insect Pests” relate to the fact that mangroves shelter species that can impact public health, notably the moth Hylesia metabus that cause skin rashes (Jourdain et al., 2012). According to stakeholders, mangrove deforestation near residential areas occurred for sanitary reason in the recent past. This study shows that mangrove conservation can be exposed to a trade-off with mangrove destruction aimed at reducing such disservices. This trade-off is generally under-estimated in ES assessments of mangroves and comprehensive framework should be implemented (e.g., Knight et al., 2017).

The services mentioned by only some groups were listed (e.g., Knowledge production, Heritage value) and can be used in future discussions with local focus groups to further establish whether these warrant additional investigation, given their perceived importance by French Guianese residents and economic actors.

In order of importance given the respondent’s perception (Table 4), the first perceived threat on mangroves in French Guiana is pollution, as mangroves are affected by wastewater near urban areas. Close to Cayenne, the impacts of pollution are visible on the microbial taxa of the mangrove (Fiard et al., 2022). Water pollution is mitigated by the purification ES from the mangrove that benefit society. Mangroves and their associated ecosystems act as natural sinks that trap all kind of anthropogenic pollutants (e.g., heavy metals, MacFarlane et al., 2007; Kulkarni et al., 2018; organic matter, Xiao et al., 2018; Adame et al., 2019). However, the sustainability of this purification role is questionable as high loads of pollutants can considerably modify mangrove ecosystems. Eutrophication in mangrove favor growth of shoots over roots and decrease their resilience (Lovelock et al., 2009), it also modifies the phytoplankton communities (Manna et al., 2010). There is a lack of information to conclude on the threat that chemicals represent for mangroves. Still, some studies has shown the impacts of chemicals on mangrove trees (particularly in early life stages; Lewis et al., 2011) and on the associated trophic network (Kulkarni et al., 2018). The question of water quality in French Guiana is the subject of a dedicated strategy established in 2017 that provides for public support dedicated to wastewater treatment (DEAL, 2017). If this plan is implemented primarily for sanitary reason, the improvement of water quality should benefit to mangrove ecosystems.

The second main threat raising concern among stakeholders is urbanization. Land-use planning, i.e., the extension of human infrastructures to face the increasing needs of society is positively related to this pressure on mangroves in the FCM. As population is concentrated on a 10–30 km wide strip along the littoral (Zouari, 2015), population growth may thus affect primarily this area and thus negatively affect the mangrove. Estuarine mangroves, which thrive on rivers, are considered as more subject to pressure from territorial development near urban areas. Indeed, they offer stable land that can support infrastructure, as is the case of a recent project of power plant construction near Cayenne, that has been criticized for its impact on mangroves and insufficient mitigation requirements (Autorité Environnementale, 2019). Coastal mangroves would be less affected, as they are under influence of mud-banks migration that make the coastline very unstable. The temptation to stabilize this dynamic system for infrastructure, urban and economic development may be associated with high risks (Jolivet, 2019). In neighboring Guyana, mangroves have been replaced, to make space for agriculture and aquaculture, and coastal protection is now provided by coastal dikes. This has considerably modified the sedimentary dynamics and the country is now exposed to erosion that can only be countered by expensive engineering solutions (Anthony and Gratiot, 2012). The integration of mangrove variability in land-use planning is thus necessary, in order to integrate its positive and negative effects in the best possible ways.

Agriculture is actually booming with an increase by nearly 38% of cultivated area between 2010 and 2019, in order to cater for the growing needs linked to the territory’s increasing population. To maintain food self-sufficiency, 1,000 hectares should be turned into agricultural land every year (CEREMA, 2016). Total used agricultural land was 33 800 ha in 2019, with more than half concentrated in the west part of the country (DEAAF, 2020). However, there was no consensus on the impacts of such development among our stakeholder groups calling for more investigation. Farming in French Guiana combines traditional manual itinerant agriculture, breeding and mechanized agriculture for commercial purposes. Sorting out the incidence of these different practices and their expansion on mangroves’ ecological states calls for further investigation as was pointed out by most of our expert groups.

The third threat on mangrove ecosystems identified by stakeholders is climate change. According to stakeholders, the impact of climate change does not directly affect mangroves but rather elements of the ecosystem, namely “Mud-banks,” “Biodiversity,” and “Fishery resources” (Figure 1). For most stakeholders (excepted one scientist), climate change was considered as a broad variable. However, it is difficult to summarize Climate change interactions with mangroves in a single variable as this corresponds to multiple factors of environmental changes (e.g., sea level rise, drought, change in salinity) that can have different—if not reverse—impacts on mangroves, depending on the context (Alongi, 2015; Lee et al., 2021). For example, in French Guiana, climate change intensifies the swell regime, which accelerates mud bank movements that should impact coastal mangroves. Further investigation is needed in this regard. As for perceived negative effects of climate change on biodiversity, these seemed to corroborate established knowledge on ocean warming and acidification risks, rather than being derived from local evidence.

Fourthly, deforestation has a very low importance in the community map, compared to its importance at the international scale (e.g., Richards and Friess, 2016). This may be explained by the absence of some extraction activities generally associated with mangroves. First, wood harvesting for construction, manufactured goods, firewood or coal is mostly absent in French Guiana. The few extractions of mangrove wood that serve handicraft productions or smoking processes are negligible in the global functioning of the socio-ecosystem as shown by their absence from the FCMs. Second, the expansion of aquaculture remains limited and takes place mainly in freshwater from coastal plains. A small amount of mangrove oyster farming currently takes place in Montsinéry (upstream from Cayenne). The mention of aquaculture in the FCM reflects the several scientific and governmental programs that aim at developing the sectors’ potential. Given the threat that aquaculture has applied on mangroves worldwide (Naylor et al., 2000), this may require special attention.

Facing those threats, one main lever of action is identified by stakeholders, namely “Legislation.” Legislation is supposed to be able to act directly on the enhancement of state of the ecosystem (“Mangrove” and “Biodiversity”) or indirectly in reducing “Land-use planning” (Figure 1). A single positive feedback loop is identified from the ES of “Knowledge production,” with a positive effect on mangrove. However, stakeholders perceive no variables with the opportunity to reduce the identified threats. This result is surprising as it reflects a command-and-control perception of conservation, rather than a vision developed following socio-ecosystems management principles (Ostrom, 2009). Indeed, solutions based on more flexible institutional arrangement can also increase the ecological outcomes of conservation, as well as its social and economic benefits (Scemama and Levrel, 2019; Bellanger et al., 2021). Nevertheless, imagining effective solutions for mangrove conservation needs to take good consideration of the multiple interactions between mangroves and societies, the associated positive and negative incentives for mangrove conservation, and the numerous sources of variability and uncertainty. In particular, there is a need for innovative solutions to better integrate the dynamism of the coastline in the future development of the territory.

FCMs were created by merging variables and connections raised by all experts in each group. Such merging can lead to overly complex system representations that potentially include artifacts associated with individual misconceptions or biases merged into group responses. The combination of individual interviews and workshops allowed to validate and increase the confidence in the credibility and relevance of the results (Teixeira et al., 2018). Moreover, the organization of the workshops allowed to increase consensus regarding the variables and the connections between them, as we can see with the reduction of standard deviation between the stakeholders FCMs. Finally, discussions during the workshops provided qualitative arguments that help understanding the FCMs configuration in light of the variability observed in the territory under consideration. Indeed, the final FCM summarizes the perception of the socio-ecosystem at the scale of the French Guiana. It does not explicitly represent the many sources of variability such as the differences between coastal and estuarine mangrove or the temporal variability associated to the migration of mud-banks, while the functioning of mangrove ecosystems and their ES is closely related to their biogeographic and geomorphological characteristics (Lee et al., 2014). However, the process of generating the map enabled identifying these differences.

Another risk with merging variables under broader concepts is to lose information (see e.g., the discussion on climate change or recreation and tourism). As a result, FCMs may fail to reflect some local variations. The final community map provides the perception of the expert groups consulted regarding such local variation and how they should adequately be captured, at the scale of the entire French Guiana. As a result, it provides a relevant overview of the perception of the functioning of the mangrove socio-ecosystem of French Guiana, and of the actual state of knowledge on this system. The organization of workshops with stakeholders allowed to collect qualitative material that can help to appreciate variability. It would be interesting to realize similar exercises on more restricted areas to focus on local issues.

The use of expert knowledge in ES assessment is considered one of the most popular ES assessment techniques today (Jacobs et al., 2015; Campagne and Roche, 2018). Indeed, it is particularly adapted to face the uncertainty-urgency dilemma that characterizes biodiversity conservation. As such, expert consultation results fit within a post-normal framework for ES assessment (Ainscough et al., 2018). In such a framework, expert knowledge is used to overcome uncertainty issues that can hinder conservation decision-making, to the benefit of a status quo detrimental to biodiversity and ES protection. Moreover, this is particularly interesting where the scientific evidence is not sufficient to support a comprehensive ES assessment as underlined by Mongruel et al. (2018). In such comprehensive assessments, economic analysis generally relies on the use of benefit transfer (e.g., Giry et al., 2017; Trégarot et al., 2021), using values commonly associated to mangroves in the literature and applying them to the studied territory. Our approach enables capturing the originality of the mangroves of French Guiana regarding the ES provided, in comparison to mangroves at the global scale, and shows that the use of benefit transfer without better knowledge of these key ES and disservices along with their variability would be hazardous at best. Based on our results, future research needs regarding mangrove ES in French Guiana, and their interactions with mangrove conservation policy can also be identified.