Habitat-specific response of fish assemblages in a small fully protected urban MPA

Abstract

This study aims at assessing the reserve effect on fish assemblage in a small urban fully protected and highly enforced Marine Protected Area (MPA), Larvotto MPA (Monaco, Mediterranean Sea). The data about fish density, size, and biomass was collected by underwater visual census in the MPA and unprotected areas in two habitats, artificial rocky substrates and Posidonia oceanica meadows. On artificial rocky substrates, we recorded significantly higher fish biomass within the MPA compared to unprotected areas, while no significant difference was detected on Posidonia, with this suggesting the habitat-specific responses of fish assemblages to protection. Here we highlight the potential effectiveness of highly enforced small-sized urban MPAs, such as the Larvotto MPA, in generating ecological benefits, and speculate about their potential role in supporting networks of MPAs. This study supports increased attention to urban MPAs in conservation efforts, emphasizing the need for efficient management strategies in the face of ongoing coastal expansion.

Marine Protected Area, Mediterranean Sea, fish assemblages, visual census, urban MPA, seagrass meadows, rocky habitats

Introduction

Local extractive and non-extractive human uses (e.g. fishing, coastal development, recreational activities) and global threats (e.g. climate change) induce disturbances on marine ecosystems, potentially eroding biodiversity and disrupting related ecosystem services worldwide (Coll et al. 2010, Díaz et al. 2019, Halpern et al. 2019). In an attempt to prevent further biodiversity loss and to restore and preserve marine ecosystems, national and international policymakers have developed a set of legislative instruments/policies (e.g. Barcelona Convention, Marine Strategy Framework Directive, and EU Biodiversity Strategy for 2030) setting ambitious targets. To meet these targets and comply with international commitments, a portfolio of management tools is available to managers; among these, marine protected areas (MPAs) have been created as a tool to mitigate the impact of fishing and other human activities on marine ecosystems, with the purpose of increasing marine biodiversity (Gaines et al. 2010, Grorud-Colvert et al. 2021). The ecological benefits that MPAs can deliver are multiple and have been widely documented, especially on reef fish assemblages (Halpern 2003, Guidetti et al. 2014, Gill et al. 2017). Specifically, effective MPAs can increase density, size, and biomass of species targeted by fishing (with this process called ‘reserve effect’, Harmelin et al. 1995, Claudet et al. 2006, Sala et al. 2012, Giakoumi et al. 2017). However, the response of non-commercial (and non-targeted) species is less clear, and generally, no significant reserve effect is recorded (Guidetti et al. 2014). Despite this, these species could still show an effect due to the fact that they can be fished as by-catch, or that they are affected indirectly, e.g. via trophic interactions (Micheli et al. 2005, Badcock et al. 2010).

Fully protected zones, areas of MPAs where no extractive or destructive activities are allowed (Grorud-Colvert et al. 2021), have the potential to contribute to restoration of fish assemblage structure, with the return of large adults of low/medium and high-level predators, restoring ecosystems functioning through trophic interactions (Micheli et al. 2005, Aburto-Oropeza et al. 2011, Guidetti et al. 2014). In the Mediterranean Sea, although 6.01% is covered by MPAs, only 0.06% is fully protected (Claudet et al. 2020), a figure still far from the target of 10% strict protection and 30% protection established by the new European Biodiversity Strategy by 2030 (European Commission 2021). A large body of literature, both in the Mediterranean Sea and globally, has identified several factors to be key determinants of the ecological effectiveness of an MPA, such as level of protection, size, age, and enforcement (Claudet et al. 2008, Edgar et al. 2014, Grorud-Colvert et al. 2021). Most of the Mediterranean MPAs where a reserve effect has been observed are generally located in isolated regions with low human density and/or in small islands as the Scandola Nature Reserve (Francour 1994), the Port-Cros National Park (Astruch et al. 2018), the Medes Islands Marine Reserve (García-Rubies and Zabala i Limousin 1990), and Tavolara-Punta Coda Cavallo MPA (Guidetti et al. 2011). However, with the exponential increase of coastal development concurrently with the newly established goals of protection and restoration of biodiversity, the establishment of urban MPAs, ‘in or at the edge of larger population centres’ (sensu Trzyna 2014) will likely become more frequent (Gill et al. 2023). This poses a challenge since human-related activities, such as tourism, fishing, noise, and light pollution and city effluents can impair potential benefits associated to the presence of MPAs (Cinner et al. 2018). In the Mediterranean Sea, there are only few examples of MPAs close to urbanized areas, and fewer with demonstrated benefits (Harmelin et al. 1995, Trzyna 2014). Therefore, it is crucial to understand how fully protected MPAs can contribute as conservation areas in the context of coastal development, as it is already the case for terrestrial protected areas (Edmiston et al. 2014, Rodríguez-Rodríguez and Martínez-Vega 2018).

In the Principality of Monaco, 90% of the coastline and 90% of shallow habitats (up to 10 m depth) have been modified and/or artificialized (https://www.medam.org; Bevilacqua et al. 2021). Despite an urban development constrained by a limited territory located between sea and mountain, the Principality of Monaco has two fully protected MPAs, the Larvotto MPA created in 1976 with the main objective of protecting the only Posidonia oceanica meadow in Monaco, and the Spélugues MPA created in 1986 for the protection of a coralligenous drop-off. MPAs located close to large urban areas (18 774 inhabitants/km² in the Principality of Monaco; www.monacostatistics.mc) like the Larvotto MPA are rare and understudied. The scarcity of scientific data on their success, combined with social constraints like conflicts of use, stakeholder opposition, and economic pressures, likely contribute to their exclusion from conservation planning compared to other more isolated and less populated MPAs. Most of the studies on the reserve effect, are focused on rocky areas (e.g. Sala et al. 2012, Giakoumi et al. 2017), with a few exceptions (Valle and Bayle-Sempere 2009, Seytre and Francour 2013). However, given the prevalence and ecological significance of P. oceanica meadows in the Mediterranean Sea, it is important to extend the knowledge of reserve effects to encompass this or other important habitats, to ensure a more comprehensive understanding of MPAs effectiveness. The aim of this study was to assess the potential reserve effect of an urban MPA on fish assemblages (in terms of species richness, size, and biomass) in two key marine habitats of the Mediterranean Sea, rocky reefs and P. oceanica meadows.

Materials and methods

Study area

This study was carried out at the fully protected Larvotto MPA (43.743950°N, 7.434700°E), located in the Principality of Monaco. The MPA was established in 1976 and covers an area of 33.6 ha. It is characterized by a gentle-to-medium slope, decreasing from the water surface to ∼30 m over the only P. oceanica meadow of the Principality of Monaco (14 ha). The rocky reefs of the MPA are artificial dikes made of boulders. It is a no-take zone (professional and recreational fishing are forbidden), with navigation access restricted to MPA staff, law enforcement, and researchers. Leisure activities (e.g swimming, paddle) are allowed. Enforcement started and has been continuously high since the establishment of the MPA.

Sampling design and data collection

Sampling was carried out inside the MPA and in unprotected areas (to the west and east of the MPA) in two different habitats, P. oceanica meadows (P) and artificial rocky substrates (R). This study focuses on two locations in the MPA, Sporting (MPA 1; P, R) and Larvotto Central (MPA 2; P, R), two unprotected areas (UA) to the west, Saint-Jean-Cap-Ferrat (UA 1; P, R), Beaulieu-sur-Mer (UA 2; P, R), and two unprotected locations to the east, Menton (UA 3 and UA 4; R) and Roquebrune-Cap-Martin (UA 3 and UA 4; P) (Fig. 1). The two sites, MPA1P and MPA2P, belong to the same P. oceanica meadow (the only one in Larvotto MPA), and the unprotected sites UA3P and UA4P also belong to the same meadow. Unprotected areas (control sites) were randomly selected from a set of potential sites, each located at least 500 m from the MPA border to avoid any direct spillover effects from the MPA (Di Lorenzo et al. 2020). The sites were selected to have comparable features to those occurring within the MPA (in terms of habitats, depth, slope, etc.). Additionally, they were chosen to be interspersed on both sides of the MPA. Sampling was carried out along a time series during 10 sampling times (July and October 2017, July 2018, June 2019, June and September 2020, June and September 2021, and June and September 2022). At each location, two sites were selected and six replicates (i.e. visual census transects) were performed in shallow artificial rocky reefs (2−5 m) and P. oceanica meadows (7−10 m). Fish assemblages were assessed through underwater visual census using strip transects of 25 × 5 m (Harmelin-Vivien et al. 1985). In the first three sampling times, fish length was estimated using three size classes (small, medium, and large), corresponding to the division into size ranges based on their overall maximum total length (TL), a technique used until relatively recently (Guidetti et al. 2008, Seytre and Francour 2013). Maximum TL for each fish species was obtained from the Fishbase database (www.fishbase.org), where records for almost all fish species are available. For the following sampling times, a new protocol was employed, which has a higher resolution and is largely used in the Mediterranean (e.g. Di Franco et al. 2009, Sala et al. 2012, Di Franco et al. 2021). Total length was recorded within 5 cm size classes for large-sized fish (species maximum size >50 cm) and 2 cm size classes for other fish species. Data was collected by trained and experienced scientific divers. Length estimates of fish from surveys were converted to fish biomass by using the length–weight relationship W = a × L^b, where W is mass in grammes and L is total length in cm. Parameters a and b were obtained whenever possible from Fishbase (Froese and Pauly 2022) or literature.

Figure 1.

Location of the Larvotto MPA in the Principality of Monaco and the unprotected areas (UA) investigated in this study. P—P. oceanica meadows, R—artificial rocky substrate.

Fish taxa were pooled into trophic groups: planktivores, detritivores, herbivores, low/medium-level and high-level predators, based on Giakoumi et al. (2019).

Data analysis

The effect of protection per habitat on species richness, total fish biomass (all species pooled), and biomass of trophic groups was assessed through a four-way permutational univariate analysis of variance (PERMANOVA) (Anderson 2001). We used a four-way permutational multivariate analysis (PERMANOVA) to assess the response on whole fish assemblages (using ‘species × sample’ matrices; n = 43 species, n = 815 samples); therefore, accounting for the different responses of multiple species at once. ‘Protection’ (Pr) as a fixed factor (2 levels: MPA and unprotected areas), ‘Habitat’ (Ha) as a fixed factor (2 levels: P. oceanica meadows and artificial rocky substrates), ‘Time’ (Ti) was treated as a random factor (10 levels: T1—July 2017, T2—October 2017, T3—July 2018, T4—June 2019, T5—June 2020, T6—September 2020, T7—June 2021, T8—September 2021, T9—June 2022, T10—September 2022), and ‘Site’ (Si) random factor (4 levels) nested in Pr × Ha. The analysis was based on Euclidean distance (for univariate) and Bray−Curtis dissimilarities (for multivariate) of square root transformed data and was tested with 9999 random permutations. Pairwise tests were used whenever suited (i.e. when there was a significant interaction between fixed factors), to assess difference between protection levels and/or habitats. To visualize multivariate patterns, non-metric multidimensional scaling (nMDS) ordinations were obtained from Bray−Curtis dissimilarity matrices calculated from square root transformed data. nMDS are presented for the combined factor Protection × Habitat. Stress values are shown to indicate the goodness of representation.

To assess potential differences between protection levels in the size of relevant commercially important species targeted by fishing we used the two-sample Kolmogorov-Smirnov Test and we plotted fish size data into size-frequency distributions. The substantial number of individuals recorded in the artificial rocky habitats allowed the study of distribution classes for two seabream species (Diplodus sargus and D. vulgaris) and the Mugilidae family. Size class distribution could not be carried out in P. oceanica due to insufficient sample size.

The PRIMER 7 and PERMANOVA + (Plymouth Marine Laboratory) were used to perform univariate and multivariate analyses, while software R 3.4.3 (R Core Team 2023) was used to assess differences between size of targeted species.

Results

During this study, we recorded overall a total of 43 species (pooling the two habitats and both protected and unprotected conditions), including both commercial and non-commercial fish (Supplementary Table S1).

A significant interaction Pr × Ha was detected for species richness (pseudo-f: 3.5702, P = .0298; Supplementary Table S2). In artificial rocky substrates, species richness was significantly higher in the MPA than in unprotected areas (Fig. 2; Table 1) [t = 1.8389, P(Perm) = 0.0472; Supplementary Table S2]. On the contrary, across P. oceanica meadows, species richness was slightly higher in unprotected areas than in the MPA (Fig. 2; Table 1), although not significantly different [t = 0.78458, P(Perm) = 0.6896; Supplementary Table S3].

Figure 2.

Species richness inside the Larvotto MPA and unprotected areas across studied habitats, P. oceanica and artificial rocky substrate. The horizontal bar inside the box is the median; the limits of the box are the upper and lower quartiles and dots are outliers.

Table 1.

Summary of species richness (species per 125 m²), total biomass (gr. 125 m²), and biomass per trophic group (high-level predators, low- and medium-level predators, detritivores, herbivores, and planktivores) in artificial rocky substrates and P. oceanica meadows, both inside the Larvotto MPA and unprotected areas.

	Artificial rocky substrates		P. oceanica
	Inside MPA	Unprotected areas	Inside MPA	Unprotected areas
Total species richness	12.1 ± 0.27	11 ± 0.13	5.76 ± 0.12	5.34 ± 0.17
Total biomass	9924.3 ± 612.5	4223.2 ± 199.9	938.8 ± 105.7	1128.1 ± 87.5
High-level predators biomass	2025.1 ± 401.2	488.1 ± 222.5	650.8 ± 638.5	1359.2 ± 860.7
Low-and medium level predators	4121.6 ± 269.3	1379.2 ± 54.9	435.7 ± 66.5	327.8 ± 24.9
Detritivores	4798.3 ± 591.6	1858.6 ± 306.2	n.a	n.a
Herbivores	2075.6 ± 329.8	2327.9 ± 183.2	533.7 ± 112.5	976.4 ± 115.8
Planktivores	459 ± 89.9	304.1 ± 27.2	316.1 ± 39.5	286.6 ± 25.9

Values are presented as means ± standard errors. n.a—not applicable.

When considering total fish biomass, significant variability was detected for the interaction Pr × Ha (pseudo-f:  27.002, P = .0001; Supplementary Table S4). Total biomass in artificial rocky substrates was significantly higher in the MPA than in the unprotected areas (Fig. 3; Table 1) [t = 7.8454, P(Perm) = 0.0001; Supplementary Table S5]. This pattern was not consistent across P. oceanica habitat with no statistically significant differences found between the MPA and unprotected areas (Fig. 3; Table 1) [t = 0.43785, P(Perm) = 0.998; Supplementary Table S4].

Figure 3.

Total biomass of fish assemblages at P. oceanica and artificial rocky substrate in the Larvotto MPA and unprotected areas from July 2017 to September 2022. The horizontal bar inside the box is the median; the limits of the box are the upper and lower quartiles and dots are outliers.

Multivariate analyses on the biomass of fish assemblages showed an inconsistent effect of protection across the two habitats considered (Pr × Ha; pseudo-f: 2.9145, P = .0001; Supplementary Table S6). Pairwise tests showed that fish assemblages were significantly different in the MPA and unprotected areas, in artificial rocky substrates [t = 2.4617, P(Perm) = 0.0001; Supplementary Table S7; Fig. 4]. Fish assemblages in P. oceanica habitat were not significantly different between the MPA and the unprotected areas [t = 1.105, P(Perm) = 0.2136; Supplementary Table S7; Fig. 4].

Figure 4.

Fish assemblage structure assessed on square-root transformed biomass at both habitats, artificial rocky substrates (R), and P. oceanica meadows (P). Two-dimensional nMDS ordinations of centroids for the combined factors Protection × Habitat are shown. Bubble size indicated level biomass in the Larvotto MPA (blue) and unprotected areas (yellow).

Regarding biomass of trophic groups, different responses were detected for the interaction Pr × Ha. When considering high-level predators, the interaction Pr × Ha was significant (pseudo-f: 7.1148, P = .0026; Supplementary Table S8). In artificial rocky substrates, average biomass of high-level predators was significantly higher in the MPA than at unprotected areas (Fig. 5; Table 1) [t = 2.5754, P(Perm) = 0.0065; Supplementary Table S9]. In P. oceanica beds, on the contrary, high-level predators biomass was not significantly different in the MPA than at unprotected areas (Fig. 5; Table 1) [t = 0.93611, P(Perm) = 0.5433; Supplementary Table S9]. The interaction Pr × Ha for low/medium-level predators biomass was significant (pseudo-f: 16.876, P = .0001; Supplementary Table S10). In artificial rocky substrates, low/medium-level predators biomass was higher in the MPA than at unprotected areas (Fig. 5; Table 1) (t = 4.1238, P(Perm)=0.0002; Supplementary Table S11). In P. oceanica beds, low/medium-level predators biomass was not significantly different between the MPA and unprotected areas (Fig. 5; Table 1) [t = 0.72534, P(Perm) = 0.7856; Supplementary Table S11]. Considering the biomass of detritivores, interaction Pr × Ha was significant (pseudo-f: 19.655, P = .0001; Supplementary Table S12). The highest biomass found in artificial rocky substrates was that of detritivores, again, higher in the MPA than at unprotected areas (Fig. 5; Table 1) [t = 4.3963, P(Perm) = 0.0001; Supplementary Table S13]. No significant variability for the interaction Pr × Ha was detected for herbivores and planktivores (Fig. 5; Supplementary Tables S14−S15).

Figure 5.

Total biomass for the five trophic groups inside the Larvotto MPA and unprotected areas across studied habitats, P. oceanica and artificial rocky substrate.

Size frequency distribution of the white seabream D. sargus shows a significant difference among protection levels (D = 0.52332; P < .001; Fig. 6a), with the presence of larger individuals in the MPA (min–max: 4–38 cm), compared with unprotected areas (min–max: 6–36 cm), where >60% of individuals were observed between 10 and 14 cm. The size frequency distribution of the common two-banded sea bream, D. vulgaris shows differences between protection levels (D = 0.11338; P < .001; Fig. 6b). In the MPA, we observe the maximum size class (up to 24 cm), although >50% of individuals were observed between 10 and 12 cm. The size frequency distribution of the individuals from the Mugilidae family shows a higher frequency of large Mugilidae in the MPA (up to 56 cm), with >55% of individuals >36 cm compared with unprotected areas where 50% of individuals are comprised between 20 and 30 cm (maximum size in unprotected sites) (D = 0.762, P < .001; Fig. 6c).

Figure 6.

Size-class frequency distribution of the two important commercial species of the genus Diplodus, D. sargus (a) and D. vulgaris (b) and the family of Mugilidae (c) at artificial rocky substrate. Bars indicate the percentage of individuals for each size recorded in the Larvotto MPA (blue) and unprotected areas (yellow) between 2019 and 2022.

Discussion

This study highlights the potential ecological benefits delivered by highly enforced small urban MPAs. Our findings highlight the reserve effect generated by the Larvotto MPA and show evidence of a habitat-specific response of fish assemblages to protection. Specifically, reserve effect is observed in artificial rocky substrates but not in P. oceanica meadows.

During this study, we have observed the presence of commercial and non-commercial fish species, both inside the MPA and unprotected areas on P. oceanica meadows. Despite the presence of species targeted by fishing inside the MPA, we have not observed a direct benefit from protection. Effect of protection in P. oceanica fish assemblages have not been largely studied (here a few examples: Francour 1994, 2000, Valle and Bayle-Sempere 2009, Seytre and Francour 2013). In the Scandola Marine Reserve, Francour (1994, 2000), similar to our study, did not observe significant differences in density and biomass of fish assemblages between the fully protected zone and partially protected and unprotected areas, with the exception of species richness. However, Valle and Bayle-Sempere (2009) observed that a protection effect occurs in P. oceanica beds for several descriptors of fish assemblages, such as abundance and biomass in the Marine Reserve of Tabarca Island. The fact that we did not detect a ‘reserve effect’ in P. meadows could, as already suggested elsewhere (Francour 1994, Zubak et al. 2017), be the response to the presence of a high biomass of high-level predators in rocky substrates within the MPA, which could result in a higher predation within the P. meadows of the Larvotto MPA; therefore, masking a protection effect.

In artificial rocky substrates, protection increased total fish biomass by 2.3-fold within the fully protected area compared to unprotected areas. This pattern, where higher biomass is observed in fully protected areas compared to unprotected ones, is consistent with previous studies in other fully protected MPAs in the western Mediterranean Sea, with comparable values of total fish biomass (Sala et al. 2012, Guidetti et al. 2014, Giakoumi et al. 2019, Rojo et al. 2021). These results are unsurprising since full protection is one of the key features of MPAs delivering ecological benefits (Sala et al. 2012, Edgar et al. 2014, Sala and Giakoumi 2018). Size is another key feature in determining MPAs ecological effectiveness, with larger MPAs generally associated to more pronounced difference compared to unprotected areas (Claudet et al. 2008, Edgar et al. 2014). However, also some small-sized MPAs have been proven effective (Francour et al. 1994, Seytre and Francour 2013), and recently, Rojo et al. (2019) have shown that the combination of small-sized MPAs, like the Larvotto MPA with only 33 ha, with high and uninterrupted enforcement can lead to ecological benefits for fish populations. Smaller MPAs are potentially easier to monitor; therefore, enforcement can be more effective. The results of this study support previous findings suggesting that small MPAs with a high and continuous enforcement can lead to a high fish biomass. Furthermore, our findings extend this knowledge by demonstrating that this effect can occur even in the absence of natural coastal habitats. This is evidenced by the Larvotto MPA, which is dominated by artificial rocky substrates.

The structure of fish assemblages at artificial rocky substrates showed differences between the MPA and unprotected areas. Higher trophic levels (high-level predators and low- and middle-level predators) represent 50% of total biomass within the MPA, while at unprotected areas it represents only 33%, a response to protection already observed in other studies (La Mesa et al. 2011, Guidetti et al. 2014, Giakoumi et al. 2019). In this study, the biomass of species from higher trophic levels (e.g. groupers) in artificial rocky substrates is lower compared to other highly protected Mediterranean MPAs (Guidetti et al. 2014, Giakoumi et al. 2019). However, this could be explained by the shallow depth at which the underwater visual census was performed (this study: 2–5 m; Giakoumi et al. 2019; 6–10 m; Guidetti et al. 2014: 8-12 m), as these species are frequently found at higher depths (Harmelin and Harmelin-Vivien 1999, Koeck et al. 2014). In fully protected MPAs, ‘top-heavy trophic pyramids’ are expected, dominated by high-level predators (Trebilco et al. 2013, Guidetti et al. 2014). Here, although a higher biomass was observed in protected sites, no ‘top-heavy trophic pyramid’ was detected. This is due to the large biomass of detritivores, from the Mugilidae family, a surprising result if compared to the proportions of this trophic group in other highly protected MPAs (Guidetti et al. 2014, Giakoumi et al. 2019), where generally, the biomass of detritivores is absent or negligible and does not respond to protection (Guidetti and Sala 2007, Guidetti et al. 2008). Mugilidae are known to have broad ecological tolerance to various environmental conditions (Whifield et al. 2012) and can benefit in environments rich in particulate organic matter, detritus, benthic microalgae, and sediment (Guidetti et al. 2002, Laffaille et al. 2002, Tuya et al. 2006). In this study, in the fully protected zone, the percentage of detritivores in the artificial rocky substrate is higher than that of high-level predators (26.7 and 8.6% of total biomass, respectively). This could indicate that other local factors, independent of MPA enforcement and design, could be responsible for the structure of fish assemblages in the Larvotto MPA. The high biomass and high frequency of Mugilidae individuals of large size could be a response to the urban pressure undergoing around the Larvotto MPA. Indeed, habitat loss in Monaco, through frequent actions of coastal development and shore artificialization, could explain changes in food web components, as seen in other studies (Coll et al. 2010, Giakoumi et al. 2015, Holon et al. 2015, Bottin et al. 2022). In this context, the reduced fishing pressure on these species—which, despite being limited, have commercial value and are targeted by fishing—could explain the high biomass we recorded.

The Larvotto MPA exemplifies several characteristics that have been identified as determinants of MPA effectiveness: it is fully protected, highly enforced, and an old reserve. However, it is a small-sized reserve, and the effectiveness of such reserves is debated. Mixed evidence about the role of MPA size in determining ecological effectiveness exits, with some studies suggesting that larger MPAs have higher fish density and biomass and accrue higher benefits (Claudet et al. 2008, Edgar et al. 2014), while others report higher effectiveness in smaller MPAs (Giakoumi et al. 2017). The effect of MPA size on fish, is related to species home range, and therefore response can be specie-specific, and reserve effect could accrue provided that a relevant proportion of fish populations is protected by the MPA (Di Franco et al. 2018). Our findings align with existing literature on the positive effects of MPAs on fish assemblages, but with the added complexity of an urban setting. This study demonstrates that even within an urban context, a well-managed small MPA can yield significant local-scale benefits by preserving local diversity and maintaining high fish biomass. According to Edgar et al. (2014), a lack of isolation could result in poor performance for MPAs. However, proximity to urban settings could facilitate implementation and management, allowing for easier surveillance, could facilitate engagement of stakeholders and local communities, by increasing acceptance of no-fishing zones and increasing levels of compliance (Di Franco et al. 2016, 2020). Furthermore, proximity to the coastline, could facilitate connectivity to other areas, which could contribute to the design of networks of MPAs. In this context, by hosting high fish biomass, we can speculate that the Larvotto MPA could be a potential source of propagules (Di Franco et al. 2012, Marshall et al. 2019), which could be exported to unprotected areas or other MPAs. Urban MPAs could represent important nodes in a MPA network, supporting healthy populations and communities in areas that are generally non-protected, although heavily disturbed (by both extractive uses and coastal development), thereby potentially gaining significant benefits from effective protection. These urban MPAs could complement the ‘residual’ MPAs typically established in less disturbed areas (Devillers et al. 2014), contributing to the creation of an ecologically effective MPA network.

Coastal urban development threatens marine ecosystems and can be a major threat to MPAs, highlighting the need to fill the gap in the lack of studies focusing on urban MPAs, their effective management, and their potential contribution to mitigating the negative impacts of coastal development on coastal communities (Gill et al. 2023). The results of this study show that relevant ecological benefits can be delivered by a small highly enforced MPA even in an urban environment, thereby providing relevant information for policymakers in a context where increasing use of coastal areas need to be reconciled with the global conservation goals foreseen for the near future.

Acknowledgements

We would like to thank Stéphane Jamme, Heike Molenaar, Elisabeth Riera, and Virginie Raybaud for their help on field work, as well as all AMPN member participants. We thank Camille Devissi for her involvement in the organization and implementation of field campaigns. Finally, we thank the editor and the two anonymous reviewers for their useful comments on the manuscript, which have helped us improve the manuscript.

Author contributions

Conceptualization, P.F.; Methodology, P.V., A.D.F., A.P.; Data collection, J.G.D., P.F., A.P.; Formal analysis, P.V., A.D.F.; Writing-original draft, P.V., A.P.; Writing-review & editing, P.V., J.G.D., E.D.F., A.D.F., A.P.

Conflict of interest

The authors declare no competing interests.

Funding

This work was funded by the Prince Albert II of Monaco Foundation, whose support is gratefully acknowledged.

Data availability

Data supporting the findings of this study are available from the authors on reasonable request.

Author notes

The authors dedicate this article to the memory of their friend and colleague, Prof. Patrice Francour, who passed away in October 2019, and who initiated this research work for an MPA that was close to his heart.