Ecosystem-based assessment of a widespread Mediterranean marine habitat: The Coastal Detrital Bottoms, with a special focus on epibenthic assemblages

1 Introduction

Coastal detrital bottoms (CDB), sometimes also referred to as coastal (bio-)detritic bottoms in the literature, are sea bottoms composed of mineral and biogenic particles, characterized by a distinct structural complexity hosting sedimentary infauna together with epifauna and epiflora settled or not on coarser granules (Rees, 1999). They are considered as a distinctive biotope in the circalittoral zone of the continental shelf worldwide (Pérès, 1982; Konar et al., 2006; Amado-Filho and Pereira-Filho, 2012; La Rivière et al., 2021).

In the Mediterranean Sea, CDB are one of the most widespread ecosystems between 30 and 100 m depth, corresponding to the upper part of the circalittoral zone (Pérès and Picard, 1964; Pérès, 1967), and house a high alpha species diversity (Stamouli et al., 2022). Besides a mineral component (typically coarse sand with a variable proportion of finer particles), these bottoms contain significant quantities of recent organogenic and bioclastic sediments (remains from hard bottoms, shell debris, bryozoans or calcified macro-algae), either autogenous or imported from adjacent ecosystems (e.g., coralligenous reefs, photophilic rocky reefs, seagrass meadows). The structure and composition of CDB can therefore differ depending on adjacent ecosystems, light availability and hydrodynamic conditions (Pérès and Picard, 1964; Joher et al., 2015; Agnesi et al., 2020). The pelitic proportion (sediment grain< 63 µm) of the sediment is usually very low, but muddy CDB can be found (Bellan-Santini et al., 1994). Located within the same depth range as CDB, Muddy Detrital Bottoms (MDB) are characterized by a high sedimentation rate influenced by the terrigenous input from coastal rivers. MDB can be constituted by a compact mud including bioclastic shells, muddy sand, or sandy mud (Carpine, 1964; Pérès and Picard, 1964; Picard, 1965; Bellan-Santini et al., 1994; Michez et al., 2014; La Rivière et al., 2021). Deep Detrital Bottoms (DDB), generally between 80 and 250 m depth, correspond to fossilized detrital sediments from terrigenous (fluviatile) and bioclastic inputs corresponding to earlier environmental conditions and a lower sea level (Pérès and Picard, 1964; Bellan-Santini et al., 1994), covering the deepest part of the continental shelf. The sediment is a mix of pebbles, sand and mud but with a mud proportion significantly higher than in the CDB. Coarser sediment is made of calcareous remains from thanatocenoses of the Pleistocene ice ages (Laborel et al., 1961; Pérès and Picard, 1964; Bourcier and Zibrowius, 1973; Michez et al., 2011; Michez et al., 2014). Distinguishing the boundary between CDB and muddy or deep detrital ecosystems is sometime challenging.

Pérès and Picard (1964) based their description of CDB on the infauna, which contain the most characteristic species; however, the same authors and others (Costa, 1960; Joher et al., 2015; La Rivière et al., 2021) identified a number of CDB facies on the basis of epibenthic species, especially macroalgae, either erect or encrusting. The former include perennial and seasonal species, such as Cystoseira spp. C.Agardh, Ericaria spp. Stackhouse, Gongolaria spp. Boehmer, Osmundaria volubilis (L.) R.E.Norris, etc.; the latter are represented by the Corallinales Lithophyllum racemus (Lamarck) Foslie, Lithothamnion corallioides (P.Crouan & H.Crouan) P.Crouan & H.Crouan, L. crispatum Hauck, L. minervae Basso, L. valens Foslie, Neogoniolithon hauckii (Rothpletz) R.A.Townsend & Huisman, Spongites fruticulosa Kützing, the Peyssonneliales Peyssonnelia crispata Boudouresque & Denizot, P. heteromorpha (Zanardini) Athanasiadis, P. rosa-marina Boudouresque & Denizot, P. rubra (Greville) J.Agardh, P. squamaria (S.G.Gmelin) Decaisne ex J.Agardh, and the Sporolithales Sporolithon mediterraneum Heydrich, among others (Huvé, 1954; Huvé, 1956; Carpine, 1958; Jacquotte, 1962; Boudouresque and Denizot, 1973; Ballesteros, 1994). Free-living calcified macroalgae forming nodules of varied morphology, referred to as ‘rhodoliths’ (or more specifically ‘maerl’ in the case of branched growth forms), gave rise to a great interest among benthic ecologists (Sciberras et al., 2009; Basso et al., 2017; Deidun et al., 2022). Their conservation value is recognized under EU rules and international conventions (Barbera et al., 2003; Salomidi et al., 2012). Marion (1883); Meinesz et al. (1983); Holon and Harmelin (2014) and La Rivière et al. (2021) described a CDB facies with large calcified bryozoans, such as Pentapora fascialis (Pallas, 1766), Smittina cervicornis (Pallas, 1766), Turbicellepora avicularis (Hincks, 1860), Hornera spp. Lamouroux, 1821, etc., but the highest diversity of bryozoans on CDB is found among small species (Harmelin, 2017). By analogy with calcareous rhodophytes, free bryozoan nodules are called bryoliths (James et al., 2006; Lombardi et al., 2014). This bryozoan facies is classified as a priority habitat according to the Barcelona Convention (Bianchi, 2009a). Recently, coralliths formed by the scleractinian coral Cladocora caespitosa (L., 1767), also considered of conservation interest (Morri et al., 2000; Bianchi, 2009b), have been described from coarse sandy bottoms in the Balearic Islands (Kersting et al., 2017a; Kersting et al., 2017b). Several octocoral species, such as Eunicella singularis (Esper, 1791), E. verrucosa (Pallas, 1766), Leptogorgia sarmentosa (Esper, 1791), Alcyonium palmatum Pallas, 1766, Pennatula rubra (Ellis, 1764), Pteroeides griseum (Bohadsch, 1761) and others, may also form epibenthic facies on detrital seabeds (Morri et al., 1991). Fauna and flora associated with all these epibenthic facies are sometimes very similar to coralligenous assemblages (Ballesteros, 2006; Pergent et al., 2015). The EUNIS typology defines as coralligenous platforms bioconstructed horizontal formations developing within sedimentary beds subject to currents, at down to at least 100 metres depth in clear waters (Davies et al., 2004; Bajjouk et al., 2015). These formations are not usually built on rock substrata but result from the active development of constructional organisms (e.g., calcified algae, hard-skeleton invertebrates) from scattered elements on loose beds, such as shells, stones, and gravel (Pérès and Picard, 1964).

Important ecological functions and ecosystem services, up to now poorly assessed, are provided by CDB (Ballesteros, 1994; Barbera et al., 2003): (i) primary production (Basso, 2012; Basso et al., 2012); (ii) key habitat and trophic resources for invertebrates and fish; (iii) breeding, spawning and nursery grounds for fish and crustaceans including species of fishery interest, sometimes close to the deep limit of the Posidonia oceanica (L.) Delile meadow (e.g., Mullidae, Sparidae, Scorpaenidae, Palinurus elephas (Fabricius, 1787)) (Harmelin and Harmelin-Vivien, 1976; Verlaque, 1990; Kamenos et al., 2004; Soykan et al., 2010); (iv) fishing ground for trawlers and artisanal fishers (Stamouli et al., 2022); (v) long term carbon sink (Basso, 2012; Basso et al., 2012; Savini et al., 2012; Burrows et al., 2014; Watanabe et al., 2020); (vi) recycling of necromass (e.g., rhizomes and dead Posidonia oceanica leaves and drift macroalgae) (Boudouresque et al., 2016); (vii) exportation and settlement of larvae towards or from adjacent ecosystems (source-sink hypothesis; Levin and Dayton, 2009). In addition, CDB host numerous protected species and other species of conservation interest (Astruch et al., 2012; Joher et al., 2015; Astruch et al., 2019): e.g., the fan shell Pinna nobilis L. 1758 (annexe II Berne convention and UICN Red list: critically endangered), the red coralline alga Lithothamnion corallioides (Annexe 5 of the HD: species of interest with regulation of its exploitation), and several Fucales (Annexe 2 of the Barcelona convention), among others. Management goals for maerl beds are different in the Mediterranean compared to the rest of the world, where extraction activities are carried out (Foster et al., 1997; Steller et al., 2003; Foster et al., 2013; Amado-Filho et al., 2017; Harvey et al., 2017).

Nowadays, like most coastal biocenoses, CDB are facing the impact of human activities and climate change. They are threatened by numerous pressures (Dutertre et al., 2015; Demestre et al., 2017) such as trawling (Bordehore et al., 2000; Bordehore et al., 2003; Fragkopoulou et al., 2021), artisanal or recreational fishing (Rendina et al., 2020), dredging (De Grave and Whitaker, 1999; Bermejo et al., 2018; Bernard et al., 2019), terrigenous input, dumping of sediment and waste (Aliani et al., 1994; Cocito et al., 1994), fish farming (Sanz-Lázaro et al., 2011; Aguado-Giménez and Ruiz-Fernández, 2012), consequences of coastal development, sea water warming, the spread of non-indigenous species (Klein and Verlaque, 2009; Katsanevakis et al., 2014; Martin and Hall-Spencer, 2017; Hall-Spencer and Harvey, 2019) and potentially acidification (Noisette et al., 2013).

Until recently, studies on CDB in the Mediterranean Sea were mainly focused on the description of infauna (e.g., Pérès and Picard, 1964; Picard, 1965; Somaschini et al., 1998), rhodolith beds and associated communities (e.g., Ledoyer, 1966; Ballesteros, 1994; Basso et al., 2017), and taxonomy (e.g., Boudouresque and Denizot, 1973; Souto et al., 2010), without any overview at ecosystem scale (e.g., Gautier and Picard, 1957; Jacquotte, 1962; Boudouresque and Denizot, 1973; Augier and Boudouresque, 1975; Laborel et al., 1976; Augier and Boudouresque, 1978; Harmelin, 1978; Bourcier, 1982; Bourcier, 1985; Bourcier, 1988; Fornos et al., 1988; Joher et al., 2012; Joher et al., 2015).

Despite the high ecological and heritage value of CDB and the many threats facing it, this ecosystem was not considered as a habitat of European interest (Bensettiti et al., 2004) by the Habitats Directive of the European Union (Natura 2000, 92/43/CEE) while some facies of CDB are described in EUNIS typology (Gayet et al., 2018) and are included in the Barcelona convention. Recently, the Habitats Directive has taken a step forward, considering rhodolith beds as a reef habitat (1170) if they host communities characteristic of hard substrates (De Bettignies et al., 2021). In France, the extension of the overall surface areas of Marine Protected Area (MPA) (national parks, Natura 2000 sites, etc.) (Petit, 2019), since the 2010s, has incorporated large expanses of seabed occupied by CDB. In accordance with these new management priorities, it has become a challenge to better take into account this widespread ecosystem.

Even with the improving accuracy of technologies (e.g., side-scan sonar, multibeam, remotely operated vehicles, autonomous underwater vehicles, etc.) (Bianchi et al., 2004; Astruch et al., 2012; Valette-Sansevin et al., 2019) and prediction models (Martin et al., 2014; Vassallo et al., 2018), the mapping of CDB is still challenging. The assessment of coastal water body masses, marine habitats or ecosystems involves numerous indices with different purposes (Bianchi et al., 2022). None of these indices alone offers an adequate basis for solving all the issues: specific descriptors may assess the ecological status linked with water quality (WFD; e.g., Borja et al., 2000; Borja et al., 2003; Ballesteros et al., 2007; Gobert et al., 2009; Lopez y Royo et al., 2010; Blanfuné et al., 2017; Piazzi et al., 2021), ecosystem functioning (descriptor 1 of the MSFD; e.g., Personnic et al., 2014; Montefalcone et al., 2015a; Giakoumi et al., 2015), seafloor integrity (descriptor 6 of the MSFD; e.g., Montefalcone et al., 2007; Gatti et al., 2015; Enrichetti et al., 2019; Piazzi et al., 2019), the conservation status of natural habitats (HD; e.g., Maciejewski et al., 2016; Sartoretto et al., 2017), natural and anthropogenic stressors (Holon et al., 2015; La Rivière et al., 2017; Ruitton et al., 2020).

Obviously, the future of coastal zone management policies will not be based on a single integrative index (Bond, 2001), but an appropriate combination of the most relevant ones (Pikaver et al., 2004; Paoli et al., 2016; Borja et al., 2021; Oprandi et al., 2021). Indicators are currently used to underpin guidelines for managers and stakeholders. In this context, the ecosystem-based approach should be deployed (Laffoley et al., 2004; Boudouresque et al., 2020a), with the aim of achieving a better understanding of ecosystem functioning as a basis for more suitable management strategies. Ecosystem-Based Quality Indices (EBQIs) have been already defined for a number of coastal marine ecosystems, namely Posidonia oceanica seagrass meadows (Personnic et al., 2014), coralligenous reefs (Ruitton et al., 2014), undersea caves (Rastorgueff et al., 2015) and rocky reefs (Thibaut et al., 2017). The aim of the present work is to develop and apply a new EBQI expressly designed for CDB (Astruch et al., 2022a). Although CDB is traditionally identified mostly on the basis of the study of the infaunal component, which requires sampling with grabs or dredges operated from large vessels, the approach presented here will focus on the epibenthic component, which can be studied with a lighter and less costly approach, i.e., photographic, video and/or visual techniques - to be privileged in a conservation context (Bianchi et al., 2022). The effectiveness of this novel index was tested along the coasts of Provence and French Riviera (France, Mediterranean Sea) (Figure 1).

Figure 1 Study area along the French Mediterranean coast and location of the 29 sampling sites.

2 Material and methods

2.1 Conceptual model of coastal detrital bottom functioning

A workshop involving 15 experts in benthic ecology (mainly the authors of the present work) was held the 2^nd of April 2019 (i) to propose and validate a conceptual model of the CDB functioning, (ii) to define the weight of each functional compartment (=box) corresponding to its importance in the ecosystem, and (iii) to develop a sampling protocol with suitable operational descriptors to assess the different compartments. A DELPHI approach was implemented (Rowe and Wright, 1999). The conceptual model thus defined schematizes the functioning of CDB (Figure 2). The compartments of the conceptual model are described in the following pages.

Figure 2 Conceptual functional model of the Coastal Detrital Bottom Ecosystem. Arrows correspond to a flux of organic matter between boxes. Light grey boxes are not assessed directly. *See corresponding paragraph for more details about Carbon sink.

2.2 Sampling strategy

Twenty-nine sampling sites were investigated (Table 1, Figure 1). Sampling took place between June and July 2020 and 2021 during the maximum development of the macroalgal communities (Piazzi et al., 2002). Site selection took into account different levels of pressures linked with human activities, from pristine-like to urbanized sites. The study area presents a gradient in hydrological conditions, involving the Rhône River to the west, the oligotrophic waters of eastern Provence and the run-offs in the urbanized French Riviera.

Table 1 Sampling sites position (DD: Decimal degrees), depth at upper and deep zones (m) and management level (MPA, Marine Protected Area; NTZ, No-take Zone; MUM, Multi-use Management).

Two depth ranges were sampled at each site: the upper zone (between 28 and 43 m, according to the lower limit of the P. oceanica meadow), and the deep zone (between 48 and 79 m depth). Upper sites were selected far enough away from P. oceanica meadows to avoid the transitional zone and ensure a correct assessment of CDB compartments.

Sampling was carried out by SCUBA diving, photoquadrats and seascape video. In the upper zone, observations were carried out by scuba divers for inventory of in situ recognizable taxa and assessment of their abundance. Each dive corresponds to the observation of 2 divers during ~12 minutes. Assuming the movements of the divers at an average speed of ~10 m·min^-1 (including stops) observing the bottom on a 2 m wide strip, each dive corresponds to an investigated surface area of ~500 m². Three cores of sediment were sampled per site to assess organic matter and sediment grain size. In both upper and deep zone, an original device, the ‘Biocube’ (Figure 3), was deployed to sample: (i) photoquadrats with an 80 cm × 80 cm sized frame using a compact camera with a 15 s time lapse. On-board video monitoring using a GoPro^® Hero 3 was used to control the position of the quadrat during sampling. Each quadrat was randomly positioned according to the drift of the operating boat. (ii) Seascape videos were taken during the photoquadrat sampling. The aim of the diversity of sampling methods is to provide complementary observations at the scale of each sampling site for a standardized assessment.

Figure 3 Diagram of the Biocube device. The pictures on the right show different types of CDB sampled on photoquadrats.

2.3 Compartments assessed

2.3.1 Box 1: Rhodoliths: Free living corallinales, peyssonneliales, and sporolithales

This compartment is involved in the structuring of the ecosystem by increasing sediment grain size and providing a support for other compartments (see boxes 2, 3, 10 in Figure 2). Among Corallinales, several species can be found such as Lithothamnion corallioides. Peyssonneliales can be calcareous or not, several species occur in CDB, such as Peyssonnelia rosa-marina (Boudouresque and Denizot, 1973) or P. squamaria. The Sporolithale Sporolithon mediterraneum is a frequent rhodolith species of CDB. For some species, in situ identification is challenging or impossible and collection is needed. The cover naturally decreases according to depth and the corresponding reduction of light availability. In oligotrophic and clear waters, rhodolith beds can be found down to 80 m depth (Ballesteros, 1994). Aguilar et al. (2009) described rhodolith beds at 140-150 m depth around the Balearic Islands. Rhodoliths play an important role in the deep benthic fixing and sequestration of organic and inorganic carbon (Basso, 2012; Basso et al., 2016). The cover (mean percentage) of living rhodoliths appears to be the easiest descriptor to assess the structure of CDB, which conditions its functioning in both the upper and the deep zones (Bosellini and Ginsburg, 1971; Basso et al., 2016).

2.3.2 Box 2: Perennial non-calcified macroalgae

The abundance of perennial non-calcified macroalgae can be linked to good ecological conditions and the absence of significant pressures (mechanical erosion, terrigenous inputs, diminution of light). Fucales (Sargassum spp., Ericaria spp., Gongolaria spp., Cystoseira spp.), Laminariales (Laminaria rodriguezii Bornet), Osmundaria volubilis and Carpomitra costata (Stackhouse) Batters are among the most representative taxa (Gautier and Picard, 1957; Joher et al., 2012; Thibaut et al., 2016; Aouissi et al., 2018; Bermejo et al., 2018; Bruno de Sousa et al., 2019; Jódar-Pérez et al., 2020; Reynes et al., 2021). These long-lived species provide a stand for seasonal macroalgae (boxes 3 and 4) and sessile fauna (box 10), a habitat (boxes 9 and 12) and a trophic resource (box 5). They contribute significantly to the primary production of the ecosystem (Ballesteros, 1990). Codium bursa (Olivi) C.Agardh, C. vermilara (Olivi) Delle Chiaje and Sphaerococcus coronopifolius Stackhouse, perennial species observed during the present work, were not considered. The abundance of these species known as tolerant can be linked to a stressor (Vidondo and Duarte, 1995; Ruitton et al., 2017). In addition, C. fragile (Suringar) Hariot is a non-indigenous species (Provan et al., 2005). The mean percentage cover in the upper zone is a suitable descriptor to assess this compartment.

2.3.3 Box 3: Seasonal erect macroalgae

Macroalgal species from this compartment, characterized by fast growth rate and seasonal cycle, thrive mainly from spring to a maximum development during summer. Seasonal macroalgae also contribute significantly to primary production (Ballesteros, 1989). The most frequent and abundant taxa are the Phaeophyceae Arthrocladia villosa (Hudson) Duby and Sporochnus pedunculatus (Hudson) C.Agardh, widespread in the circalittoral both in Coralligenous reefs and CDB. Some species usually observed in the infralittoral can be found such as Acetabularia acetabulum (L.) P.C.Silva or Padina pavonica (Linnaeus) Thivy. Umbraulva dangeardii M.J.Wynne & G.Furnari is an Ulvophyceae that can be observed often associated with rhodolith beds. Chrysymenia ventricosa (J.V.Lamouroux) J.Agardh and Sebdenia spp. (J.Agardh) Berthold are among the main characteristic Florideophyceae. Identification at species level is impossible based on in situ observation or video and photograph analysis. The mean cover (percentage) of seasonal macroalgae appears to be a suitable descriptor. Non-indigenous and invasive species with a seasonal dynamic can thrive in the CDB: Caulerpa cylindracea Sonder (Klein and Verlaque, 2008), C. taxifolia (M.Vahl) C.Agardh (Meinesz et al., 2001; Montefalcone et al., 2015b), mucilage (several species including NIS: Acinetospora crinita (Carmichael) Sauvageau, Chrysonephos lewisii (W.R.Taylor) W.R.Taylor, Nematochrysopsis marina (J.Feldmann) C.Billard, Zosterocarpus oedogonium (Meneghini) Bornet) (Giuliani et al., 2005; Schiaparelli et al., 2007; Bianchi et al., 2019a). The latter taxa have been considered as stressors and are therefore not taken into account in the assessment of Compartment 3.

2.3.4 Box 4: Seasonal turf

Seasonal turf is defined as an algal community with a low height (a few centimetres at most). Worldwide, turf is more and more often observed replacing more structuring communities, influenced by stressors (Littler and Littler, 1980; Connell et al., 2014). Seasonal turfs include numerous species but very few can be identified in situ like the Ulvophyceae Pseudochlorodesmis furcellata (Zanardini) Børgesen and Valonia macrophysa Kützing. Womersleyella setacea (Hollenberg) R.E.Norris and Acrothamnion preissii (Sonder) E.M.Wollaston, two alien species (Boudouresque and Verlaque, 2002), can be present and are considered as stressors. These two species are often invasive in coralligenous reefs, Posidonia oceanica meadows, and infralittoral photophilic reefs (Piazzi and Cinelli, 2000; Piazzi et al., 2007; Bianchi et al., 2019b), and were found in CDB at Port-Cros National Park (Provence, France; Astruch et al., 2019). The distinction by visual assessment between seasonal turf and seasonal erect macroalgae is difficult. Considering that the two compartments progress with the same dynamics, it is relevant to consider their cumulative cover as the descriptor of their status in both upper and deep zones. The ‘Compartments 3 and 4’ are therefore named hereafter ‘Compartment 3-4’.

2.3.5 Box 5: Herbivores

The high diversity of the macroalgal communities on CDB (boxes 1, 2, 3, 4) influences the abundance and the diversity of epibenthic herbivores. In the present work, the following taxa were considered as herbivores: echinids (Sphaerechinus granularis (Lamarck, 1816), Stylocidaris affinis (Philippi, 1845); occasionnaly Centrostephanus longispinus (Philippi, 1845), Echinus melo Lamarck, 1816), and the gastropods Cerithium sp. Bruguière, 1789 and Aplysia spp. L., 1767. While S. granularis is strictly a herbivore (Weinberg, 1996; Wirtz and Debelius, 2003), S. affinis, C. longispinus and E. melo are omnivores. The latter two are not characteristic of CDB, even if they can be found in the vicinity of coralligenous beds. According to food availability and habitat (Wilkie et al., 1996), they are able to shift into herbivorous-dominant behaviour (De Ridder and Lawrence, 1982; Francour, 1989; Francour, 1991; Bergbauer and Humberg, 2000). Sphaerechinus granularis and S. affinis are known to be observed gathering on detrital bottoms (Harmelin and Duval, 1983). Vertebrate herbivores are mainly represented seasonally by the fish Sarpa salpa (L., 1758) during its breeding period (late summer) (Verlaque, 1990). Considering the short duration of the phenomenon, this species is not taken into account in this compartment. The alpha taxonomic richness (number of taxa per site) is a relevant descriptor to assess the herbivores compartment.

2.3.6 Box 6: Particulate organic matter (POM) of the sediment

Particulate organic matter (POM) of the sediment observed in abundance can be linked to natural eutrophic conditions or to a stressor (Sciberras et al., 2009). Based on our sampling strategy, the assessment of this compartment is feasible by considering the cover of muddy sediment on the bottom observed on the photoquadrats. This descriptor is already applied to assess the POM sedimented on coralligenous reefs (Ruitton et al., 2017). However, its assessment on CDB proved difficult. We therefore propose the percentage of pelitic fraction (<63 µm) as a proxy of the POM of the sediment. Organic matter accumulation is determined by environmental dynamics (currents, fine particles transfer to deep areas) (Orekhova and Ovsyany, 2020). A considerable percentage of pelitic fraction can also correspond to Muddy Detrital Bottoms (MDB).

2.3.7 Box 7: Infauna

Infauna constitutes a key compartment, influenced by sediment grain size and organic matter content (Pérès and Picard, 1964; Word, 1978; Cocito et al., 1990). Part of the Infauna can also be part of the Epifauna. We distinguished here three trophic groups within the infauna: (Box 7a) Filter and suspension-feeders: Represented by Phoronida, bivalves (ie. Laevicardium oblongum (Gmelin, 1791) or Acanthocardia deshayesii (Payraudeau, 1826)), Porifera (e.g., Suberites domuncula (Olivi, 1792)), Ophiuroidae and annelids such as Serpulidae; (Box 7b) Detritivores: Regroup surface and sub-surface deposit feeders such as annelids such as Petta pusilla Malmgren, 1866 (or Capitellidae, particularly in case of organic enrichment) and other different groups of worms (Echiura, Sipuncula, Nematoda), and molluscs such as Turritellinella tricarinata (Brocchi, 1814), Moerella donacina (L., 1758), Abra nitida (O. F. Müller, 1776) or A. prismatica (Montagu, 1808) (Zenetos, 1996); (Box 7c) Carnivores> (including scavengers): annelids such as Nephtyidae, Eunicidae or Polynoidae, molluscs gastropods such as Naticidae or Conidae, decapods and starfishes.

CDB, particularly in the upper circalittoral zone, are characterized by a relatively coarse grain size, decreasing with depth. It directly influences the trophic network in the sediment. Consequently, coarse sediment hosts a lower content in organic matter and a lower abundance of the infauna (Blanchard et al., 2009). According to Bellan et al. (1980), high organic matter content can be a stressor for infauna, by preventing the recycling of organic matter due to polluted water inputs. The organic matter content is often used as a descriptor of the trophic status of infauna (Cocito et al., 1990; Dell’Anno et al., 2002). However, the natural abundance of organic matter can be linked to the vicinity of seagrass meadows, (i) providing dead leaves that partly remain in the sediment, or (ii) with dead matte a few cm under the sediment surface (Boudouresque et al., 2019). In the present work, we assess the whole infauna compartment on the basis of the organic matter content in the sediment (percentage).

2.3.8 Box 8: Organic microdetritus

Posidonia oceanica seagrass meadows produce and export a huge number of dead leaves to adjacent ecosystems, including CDB (Boudouresque et al., 2016). Infralittoral photophilous reefs and coralligenous reefs also export macroalgae, which accumulate or move over circalittoral soft bottoms. Organic macrodetritus composition and biomass vary according to source ecosystems: dead leaves of P. oceanica, macroalgae or terrestrial plant remains. Before reaching greater depth, off the continental shelf, organic macrodetritus are used as a trophic resource by the detritivores of the CDB.

2.3.9 Box 9: Detritivores (epibenthic)

Detritivores play an important role, degrading litter and organic debris on the sediment. This compartment is composed of Echinodermata (Holothuroidea, Echinoidea and Asteroidea), Crustacea and Gastropoda. The abundance of some taxa can be favoured by eutrophic (i.e., HOM: High Organic Matter) or oligotrophic (LOM: Low Organic Matter) conditions. Two categories are distinguished: (9a) HOM Detritivores: Including Holothuria spp. L., 1767, Bonellia viridis Rolando, 1822 and Paguridea; (9b) LOM Detritivores: Represented by Spatangus purpureus O.F. Müller, 1776, Chaetaster longipes (Bruzelius, 1805), Echinaster sepositus (Retzius, 1783), Gastropoda and Crustacea (Mysidacea, Malacostraca). While a high abundance of HOM detritivores can be linked with a CDB that is potentially degraded (eutrophication), the abundance of LOM detritivores, more sensitive to eutrophication, indicates a good status of this compartment. Mean density (individuals per m² observed in the photoquadrats at both depths) is a relevant descriptor to assess detritivores 9a and 9b.

2.3.10 Box 10: Filter and suspension feeders (epibenthic)

Benthic filter and suspension feeders on CDB belong to various taxonomic groups: Annelida, Ascidiacea, Bryozoa, Cnidaria, Porifera) (Harmelin, 1978; Souto et al., 2010). On coralligenous reefs, the abundance of Cliona viridis (Schmidt, 1862) is considered as a consequence of stressors (bioerosion in a eutrophic configuration) (Cerrano et al., 2001). However, the presence of C. viridis with a relatively low density on CDB, as observed in our case study, cannot be considered as an indicator of stress. Mean density of individuals or colonies per m² (observed in the photoquadrats at both depths) seems to be the most suitable descriptor for filter and suspension feeder assessment, unlike taxonomic richness, due to the difficulty of identification in situ or from videos or photographs.

2.3.11 Box 11: Planktivorous teleosts

This compartment includes planktivorous pelagic or demersal fish, with life traits significantly linked with the substrate. Strictly pelagic planktivorous teleosts with no interaction with the bottom (e.g., Clupeidae, Engraulidae) were not considered. The main species observed were Spicara maena (L., 1758), S. smaris (L., 1758), Boops boops (L., 1758), Chromis chromis (L., 1758) and Anthias anthias (L., 1758). The latter two indicate the vicinity of hard substrates. CDB play an important role for these species: a spawning ground for Spicara spp. (Harmelin and Harmelin-Vivien, 1976) and an essential habitat for Boops boops juveniles. Planktivorous species are also essential in the trophic network providing a food supply for high level predators (Bănaru et al., 2013; Cresson et al., 2020). The alpha specific richness (number of species per site, considering all methods) was measured to assess the planktivorous teleosts compartment.

2.3.12 Box 12: Carnivores

This box encompasses all carnivorous species, predators of crustaceans, annelids, fishes and other benthic invertebrates. This box is divided in two: Invertebrate carnivores (12a) and Vertebrate carnivores (12b). (12a) Invertebrate carnivores: This includes mainly motile species: crustaceans, echinids (Asteroidea, Ophiuroidea), Gastropoda, Nudibranchia, Cnidaria, etc. The alpha taxonomic richness of invertebrate carnivores (per site considering all methods) is a suitable descriptor for this compartment. (12b) Vertebrate carnivores: This includes teleost fish species (e.g., Pleuronectidae, Trachinidae, Sparidae, Gobiidae, Blenniidae, Mullidae, Labridae, Triglidae) and chondrichthyans (e.g., Scyliorhinus spp. Blainville, 1816, Rajidae). Carnivorous fish find an important trophic resource at the surface or inside the sediment of CDB. Some use CDB as a nursery habitat (e.g., Serranus hepatus (L., 1758), Blennius ocellaris L., 1758) or a breeding habitat (e.g., Symphodus cinereus (Bonnaterre, 1788), Solea spp. Quensel, 1806). The assessment of abundance or diversity of fish assemblages normally applied in other marine habitats, such as reefs or seagrass meadows (Harmelin-Vivien et al., 1985), proved of little applicability in CDB. The most suitable descriptor to assess the compartment 12b is the alpha taxonomic richness (number of species observed per site using all methods).

2.4 Compartments not assessed

In the frame of an ecosystem-based approach, it is necessary to considerer all the main functional compartments even if some are not assessed.

2.4.1 Phyto- and zooplankton

Together with POM, phytoplankton and zooplankton are the basis of planktivore diet (Khoury, 1987; Cresson et al., 2014; Chen et al., 2022). This compartment can vary very quickly, and its proper assessment would require too much time and resources relative to its weight in the ecosystem functioning (Table 1). In the framework of an index assessment, it is more suitable to assess it indirectly through its consumers, as is done for other ecosystems (see Personnic et al., 2014; Thibaut et al., 2017).

2.4.2 Pelagic dissolved and particulate organic matter (DOM and POM)

Dissolved organic matter (DOM) and particulate organic matter (POM), provided by phytoplankton exudation, grazing by zooplankton and terrestrial inputs, play a key-role in the marine microbial trophic network (Mostajir et al., 2012) and in some coastal food webs (Darnaude, 2005; Bănaru et al., 2007). As far as the CDB ecosystem is concerned, this compartment is considered as indirectly assessed by the filter feeders and suspension feeder compartment (Box 10) and therefore was not assessed directly.

2.4.3 Microbial loop

The microbial loop is driven by the DOM and the POM (Velimirov et al., 1984; Buffoni et al., 1990). DOM is an important substrate for bacterial production. The microbial loop is defined as a community including heterotrophic micro-organisms, namely Archaea, bacteria, flagellated eukaryotes and ciliates (Azam et al., 1983; Boudouresque, 2015). This compartment is not directly assessed in the present work but it is strictly linked with POM and DOM, plankton communities, filter and suspension feeders (Box 10), planktivores (Box 11) and carnivores (Box 12).

2.4.4 Piscivores

Piscivorous teleosts strictly linked with CDB are very few. High level predators that can be found are mostly casual visitors from adjacent ecosystems (see below). Lophius spp. L., 1758 and Scorpaena scrofa L., 1758, two nocturnal and benthic predators, more frequently occupy CDB. At the scale of a given sampling site, the occurrence of piscivores is very low.

2.4.5 Transiting live organisms

Many organisms transit through the CDB mainly from the infralittoral and circalittoral ecosystems to the bathyal and abyssal bottoms. Consequently, numerous species from coralligenous reefs, infralittoral reefs and Posidonia oceanica meadows can be found in CDB. The assessment of this transiting fauna was indirect through the main compartments of CDB assessed (Boxes 1, 2, 3-4, 5, 9, 10, 12).

Numerous high-level predators can visit CDB occasionally (see the compartment ‘Visiting predators’ in Figure 2), such as Dentex dentex (L., 1758), Pagrus pagrus (L., 1758), Zeus faber L., 1758, and feed on planktivorous or mesocarnivorous fish species. Because of a very low and occasional occurrence, this compartment was not taken into account.

2.4.6 Carbon sink

Rhodolith beds act as major carbon sinks through their high storage capacity of both organic (C_org) and inorganic carbon (C_inorg). The effectiveness of this role is strictly associated with the vitality of calcareous organisms living with these ecosystems (Martin et al., 2007; Basso, 2012; Burrows et al., 2014; Watanabe et al., 2020). While the C_inorg sequestration results mainly from the significant calcium carbonate contents of the rhodolith skeleton (Van der Heijden and Kamenos, 2015), the capture and the trapping of C_org within rhodolith beds come from autochthonous and allochthonous sources (Mao et al., 2020). The C_org production is mainly completed through the photosynthesis of rhodolith organisms (B1) and the occurrence of facies of calcareous invertebrates (e.g., Bryozoa, Echinidae, Gastropoda, Annelida, Bivalvia). Though the C_org fixation is well known for rhodolith beds (Watanabe et al., 2020), its long-term sequestration in the sediments is still to be specified. For P. oceanica meadows, the fixation and C_org sequestration capacity proved to be closely related with water depth and its state of conservation. Thus, healthy and shallow meadows are considered as carbon sinks whereas deeper and altered ones may act as carbon sources (Monnier, 2020; Monnier et al., 2021; Pergent-Martini et al., 2021; Monnier et al., 2022).

2.5 Descriptors selection

Fifteen descriptors were selected to assess the 12 boxes of CDB (Table 2). Boxes 3 and 4 were considered together as seasonal macroalgae, so that 11 boxes were actually considered. The weight of each box (W), first defined during the workshop, was successively adjusted according to the results of the test in the field. The compartments selected as providing the main contribution to the functioning of the ecosystem are: Box 1-rhodoliths (W=6), Box 5-infauna (W=5), box 10-filter and suspension feeders, Box 2-perennial and Box 3-4-seasonal macroalgae (W=4).

Table 2 Boxes, descriptors, weight (W) and status for the EBQI assessment (POM, Particulate Organic Matter; HOM, High Organic Matter; LOM, Low Organic Matter).

Descriptors for each compartment were selected among a wide range of available descriptors tested during the sampling. The criteria for descriptor selection were (i) its applicability, (ii) a wide enough range of values across the sampled sites, (iii) the absence of redundancy with the descriptors of other compartments in order to reduce the total number of descriptors.

The method to assess each descriptor is presented in Table 3. The monitoring of one sampling site can be carried out in less than one day with a small boat and 3 operators. It involves (i) one dive with two operators in the upper zone with 12 minutes of investigations (sampling of 3 sediment cores, photographs and videos), covering a mean surface of about 500 m², and (ii) the deployment of the Biocube at the two depths (upper and deep zones) to sample 30 photoquadrats per depth and seascape videos. (iii) Analysis of the cores involves measurement of grain size and percentage of organic matter.

Table 3 Method for the assessment of the descriptors for each functional compartment.

2.6 EBQI calculation

The EBQI is then calculated using the following formula:

W_i is the weight of the box, S_i the status of the box and S_max maximum status possible (4/4). A notation between 0 and 10 is obtained. An ecological status is defined according to the EBQI notation (Personnic et al., 2014; Table 4). The ecological status limits are similar to those defined for Posidonia oceanica seagrass meadow (Personnic et al., 2014), coralligenous reefs (Ruitton et al., 2014) and rocky reefs (Thibaut et al., 2017).

Table 4 Ecological status according to EBQI notation.

A confidence index (CI) was used to assess the quality of the data (suitable protocol, recent data, expert judgment) (see Personnic et al., 2014 for CI calculation) (Table 5). The CI allows assessment of the quality of the sampling according to the recommended methods. If a sampling does not combine all the suitable methods, it will still be possible to calculate an EBQI with a low CI (e.g., data available with other protocols, expert judgement).

Table 5 Confidence index (CI) related with data quality or expert judgement (Criteria from Personnic et al., 2014).

2.7 Statistical analysis

Correlation between descriptors notation has been tested by calculating Spearman rank-order correlation and assessing its significance with p-value. Principal Component Analysis (PCA) was used without standardization to show the relation between boxes and EBQI sites notations.

The robustness of the EBQI for CDB has been tested with regard to the ecological status per box and the weight per box. A robustness analysis involving 1000 simulations with 25, 50, 75 or 100% of probability of perturbation of ±1 or ±2 (of each ecological status per box and each weighting per box) has been tested. The analysis recalculates the EBQI notation considering the corresponding perturbation ( ± 1 or ±2) and the probability of perturbation (25, 50, 75 or 100%).

Box status has been plotted on radar charts and the uniformity of box status has been assessed through circularity of polygons representing those charts. Circularity is the ratio between the polygon area and the area of a circle having a circumference equal to the polygon perimeter (Oprandi et al., 2021). Perimeter and area have been calculated using ArcGIS Pro 3.0^®.

The influence of the management level was tested on EBQI notations and circularity value using PERMANOVA analysis, using Euclidian distance and 999 permutations (Anderson, 2001). Following Thibaut et al. (2017), four management levels were defined: (i) sites without protection (hereafter named ‘No protection’), (ii) Natura 2000 sites (EU Habitats Directives) with no effective protection (hereafter named ‘Natura 2000), (iii) protected sites, including MPA with Multi-Use Management (hereafter named MPA-MUM) and (iv) highly protected, including MPA with No-Take Zone (hereafter named ‘MPA-NTZ’).

P-value threshold was set at 5% for all tests. Analyses were performed with R software (R core team, 2022).

3 Results

3.1 EBQI notation

EBQI notation ranged from 2.05 (27TH) to 8.06 (21PAM) (Table 6; Supplementary Figure 1). Eight sites out of 29 exhibited a Good or Very Good notations while 10 showed a Poor or Bad status. The Confidence CI was lower in Box 7 (infauna) for all sites because of the method used (organic matter assessment instead of macrofauna analysis). The CI was Very Good for all sites, ranging from 88 to 93%.

Table 6 EBQI notation, status per box, Confidence Index (CI in %) and Circularity value (Cir.) for the 29 sampling sites (see Figure 1 for location).

EBQI notation and box status attributed to each site are mainly explained by local specificities and the vicinity or not of human activities (i.e., urbanized areas). PCA showed a gradient along Axis 2 (16.8% of the total variance) from Poor to Very Good sites. Good and Very Good EBQI notations were mainly explained by the high status of Boxes 1, 2 and 3-4 (primary producers). Boxes 6 (sedimented POM), 7 (infauna), 8 (organic macrodetritus) and 12 (carnivores) were less discriminating (Figure 4).

Figure 4 Principal Component Analysis. Sampling sites are coloured according to their EBQI notation. Boxes descriptors are projected in light grey. Red: Poor; Yellow: Bad; Green: Intermediate; Blue: Good; Dark Blue: Very Good.

Radar plot projection helps in assessing the heterogeneity of ecological status per box for each sampling site (Supplementary Figure 2). For example, 21PAM (Pampelonne), with the best notation (EBQI = 8.06), is defined by an ecological status ranging from 2 to 4 while 09EMB (Embiez), also with a good notation (EBQI = 6.98), is characterized by an ecological status between 0 and 4.

A Spearman rank correlation has been tested between the fifteen descriptors used for EBQI calculations (Figure 5). Positive correlations moderately to highly significant are observed between descriptors related to primary producers (Box 1 - rhodoliths, Box 2 - perennial macroalgae, and Box 3-4: seasonal macroalgae and turf). These same descriptors also show significant correlations with the mean density of LOM detritivores (Box 9b). The diversity of herbivores (Box 5) shows a significant negative correlation with the cover of organic macrodetritus (Box 8). The percentage of organic matter (Box 7) shows low but significant correlations (R²< 0.34) with sedimented POM (Box 6). The percentage of pelitic fraction of the sediment (Box 6) shows a significant negative correlation with the cover of perennial macroalgae (Box 2), rhodoliths (Box 1-UZ) and seasonal macroalgae and turf (Box 3-4). HOM detritivores density (Box 9a) shows no correlations with any other descriptor.

Figure 5 Spearman rank correlation between the fifteen descriptors related to ecosystem boxes of the coastal detrital bottoms. The number in bold is the correlation coefficient R. *: p< 0.05; **: p< 0.01; ***: p< 0.001. Blue, positive correlation; Red, negative correlation.

3.2 Circularity analysis

The circularity of radar charts reflects the uniformity of box status. Values range from 0.25 (28ANT) to 0.73 (21PAM) and are independent of the EBQI notation (Spearman correlation: R=0.242 and p=0.206). According to the EBQI notation and circularity value, 4 different situations are identified (Figure 6): (i) CDB sites with moderate-good notation (EBQI≥5) and high homogeneity of box status (circularity≥0.5) such as 24DRA, 04PLA, 09EMB, 16BAG and 21PAM; (ii) CDB sites with moderate-good notation but low homogeneity of box status such as 13FG, 18CAV, 15HB, 20CAM, 26EST, 24PQR and 17 VAI; (iii) CDB sites with poor notation (EBQI<5) and high homogeneity of box status status like 22ST, 10AMP, 07 CC and 29VIL; (iv) CDB sites with poor overall notation and low homogeneity of quality status such as 27TH, 01COU, 06SUG, 11DF, 05PR, 25AGA, 03PRA, 02CAR, 12TGR, 28ANT, 23SAR and 19CT.

Figure 6 Scores of EBQI and Circularity for the 29 sites. Red: Poor; Yellow: Bad; Green: Intermediate; Blue: Good; Dark Blue: Very Good.

3.3 Influence of the management level

The management level at the 29 sampling sites is heterogeneous. Four sites have no protection, 12 sites are included in Natura 2000 sites, 5 sites are within MPA-MUMs, and 8 are within MPA-NTZs. Analysis of EBQI mean notation and Circularity mean value per management level shows no significant differences between management levels (Permanova: respectively p=0.866 and 0.737; Figure 7).

Figure 7 Box plots showing EBQI notation (left) and circularity value (right) according management level. The black crossbar corresponds to the mean EBQI notation, the grey rectangle corresponds to the standard error, the vertical bars correspond to the 95% confidence interval.

3.4 Robustness analysis

3.4.1 Robustness regarding perturbation of the ecological status

The higher the perturbation and the probability of perturbation, the less similarity is observed in the iterations compared to the initial EBQI calculation (Figure 8; Supplementary Table 1). The percentage of similarity decreases with the probability of perturbation or the number of perturbations ( ± 1 or ±2). With perturbation of the ecological status per box, the mean similarity (all sites included) moves from 81.8% (25%) to 61.7% (100%) for a ±1 perturbation and from 68.8% (25%) to 38.8% (100%) for a ±2 perturbation. A high perturbation of the ecological status implies a lower stability of the EBQI.

Figure 8 Robustness of the EBQI with regard to the Ecological status per box. Sampling sites (x-axis) are sorted in ascending order according to the EBQI notation from left to right. In order to test the effect of the status per box on the EBQI (robustness), status values have been randomly perturbed (above, ± 1; below, ± 2) 25% (left) and 100% (right) of probability; 1000 iterations were performed. The change of the EBQI notation (Bad through Very Good) of a site, for a given iteration, is shown by the colour of the new class in which it falls). Red: Poor; Yellow: Bad; Green: Intermediate; Blue: Good; Dark Blue: Very Good.

3.4.2 Robustness regarding perturbation of the weighting

When perturbing the weighting per box, the mean similarity is higher, moving from 91.5% (25%) to 84.5% (100%) for a ±1 perturbation and from 86.5% (25%) to 73.3% (100%) for a ±2 perturbation (Figure 9; Supplementary Table 1). The perturbation of the weighting does not affect the stability of the index as much as the perturbation of the ecological status per box.

Figure 9 Robustness of the EBQI with regard to the weighting per box. Sampling sites (x-axis) are sorted in ascending order according to the EBQI notation from left to right. In order to test the effect of the weighting per box on the EBQI (robustness), status values have been randomly perturbed (above, ± 1; below, ± 2) 25% (left) and 100% (right) of probability; 1000 iterations were performed. The change of the EBQI notation (Bad through Very Good) of a site, for a given iteration, is shown by the colour of the new class within which it falls. Red: Poor; Yellow: Bad; Green: Intermediate; Blue: Good; Dark Blue: Very Good.

The lower values of similarity for high levels of perturbation ( ± 2) of both the ecological status per box and the weighting confirm the fact that weighting the boxes is consistent and useful, as already shown by Personnic et al. (2014) for Posidonia oceanica seagrass meadows. It is worth noting that the sites with the lowest similarity are often the sites where the initial EBQI notation is close to a threshold between two statuses (e.g., 7.53 is a Very Good notation but very close to the Good status<7.50).

4 Discussion

4.1 An original Ecosystem-Based Quality Index for Coastal Detrital Bottoms

This paper proposes an original Ecosystem-Based Quality Index to assess the health status of the Coastal Detrital Bottoms in the North-Western Mediterranean context. The goal is to provide (i) a useful tool, embracing the concept of the ecosystem-based approach, for coastal managers (MPAs, local authorities), and (ii) arguments for a better consideration of this key habitat according to the MSFD and Barcelona Convention. To determine a relevant biological indicator, the main issue is finding the balance between the most suitable descriptors versus the feasibility of methods of implementation according to available human and material resources. Unlike ecosystems of priority interest such as Posidonia oceanica meadows (Montefalcone, 2009; Lopez y Royo et al., 2010; Personnic et al., 2014; Oprandi et al., 2021), no indices or descriptors are available for the functional or ecosystem-based analysis of CDB.

The EBQI is a suitable tool to broadly assess the functioning of an ecosystem (Boudouresque et al., 2020a; Boudouresque et al., 2020b). It aims at providing a first integrated insight with a general notation and status per compartment, useful at the scale of a sampling site or in the frame of a monitoring network at larger scale. The CI for CDB, between 88 to 93%, is far higher than the mean CI assessed for sites used to defined other EBQI indices: 53% for Posidonia oceanica seagrass meadows (Personnic et al., 2014), 62% for undersea caves (Rastorgueff et al., 2015), and 74% for rocky reefs (Thibaut et al., 2017).

As expected, our analysis highlights the influence of both natural features and anthropogenic activities on the ecological status of CDB. Sites in areas without trawling or dredging activities, far from urbanized zones, show the best notation. The Provence and French Riviera coasts provide a wide range of anthropogenic pressure and local natural specificities. This heterogeneity was the key for defining an EBQI for CDB and testing the reliability of the descriptors of the functional compartments. A further analysis will have the benefit of new and detailed data to identify the causes of such a status: natural feature (hydrological context, geology, adjacent habitats) or pressures linked with human activities (trawling, terrigenous input, coastal development, etc.). As for all point measurement methods to assess the ecological quality, the EBQI requires a high number of replicates per site to give a better insight into the structure of the ecosystem assessed. The aim of this approach is to offer a broader view of the ecosystem functioning of a given ecosystem.

According to our results, the protocol proposed is accurate, using accessible methods involving SCUBA diving, photographs, photoquadrats and videos. The robustness analysis showed an effect of a perturbation of the ecological status per box or weighting. Nevertheless, the percentage of similarity remained high for a relatively moderate perturbation. The probability of 25% was tested to simulate a theoretical application with one in four chances of being wrong, which is consistent and closer to the reality of field application. Keeping in mind that the EBQI aims at a broad assessment of ecosystem quality and functioning, the results confirmed the robustness of the approach.

4.2 Criticism and improvement perspectives

It would have been a much-improved method if the sediment macrofauna had been properly assessed instead of just measuring the organic matter (Box 7). The quality of the assessment would have been improved, at ecosystem scale, but this would have been very time-consuming during the sampling process. The AMBI index (the AZTI’s Marine Biotic Index: Borja et al., 2000; Borja et al., 2003), and other indexes that have been developed, such as e.g., BENTIX (Simboura and Zenetos, 2002), the BQI (Benthic Quality Index; Rosenberg et al., 2004), the GPBI (General Purpose Biological Index; Labrune et al., 2021) and MEDOCC (specifically for Mediterranean assemblages; Pinedo et al., 2015), allow the assessment of the ecological quality of soft bottoms based on sediment macrofauna (see also Labrune et al., 2006). They consider the abundance in the sediment of species sensitive to stressors, tolerant and opportunistic. They can provide reliable information about infauna status to complete and improve the ecosystem-based assessment, but can be ineffective for detection of an impact of physical pressures (e.g., AMBI: see Labrune et al., 2006, but also Borja et al., 2015). The macrofauna of the sediment is of importance in terms of ecological functioning and trophic interactions with epibenthic and demersal communities (Austen et al., 2002), and obviously an assessment considering only the organic matter content is far from being entirely satisfactory (Cocito et al., 1990). It could also be argued that while the proper identification of CDB must be based on the infauna (Pérès and Picard, 1964), the index presented here focuses mostly on the epibenthos. However, characterization and assessment are two distinct steps of environmental diagnostics (Bianchi et al., 2012): assessment criteria can be applied only after the habitat in question has been characterized. We do believe that the health status of the CDB epibenthic facies is an acceptable proxy for the quality of the whole ecosystem; that is why we have decided to assign the highest weighting to Box 1 (Rhodoliths), which may require discussion. All the direct anthropogenic physical pressures mentioned in the previous pages obviously threaten the epibenthos first, rather than the infauna, which is known to be more resilient. Furthermore, as far as we are aware, the ongoing seawater warming in the Western Mediterranean region (Cerrano et al., 2000; Garrabou et al., 2009) is affecting epibenthic species more than the infauna. But it may be true that the degradation of the CDB epibenthic facies may not necessarily imply a bad status for the infauna, and further research is needed. An accurate assessment of the entire ecosystem should imply more than just a selection of functional compartments, as evoked by de Jonge et al. (2019).

Rhodolith beds in CDB can be considered both soft and hard substrata, according to the associated benthic communities (Jacquotte, 1962; Hall-Spencer, 1998; Basso et al., 2017). Rhodolith beds are also described as a step towards the edification of coralligenous banks, from a soft bottom towards a hard bottom, under specific environmental conditions (Basso et al., 2007). The pioneer zoologist André Fortuné Marion named as ‘coralligenous gravels’ the rhodolith beds of the Gulf of Marseille (Marion, 1883). The ‘Coralligène de plateau’ described by Pérès and Picard (1951, 1964); Laborel (1961) and Bellan-Santini et al. (1994) is a coralligenous bank developing from the growth and coalescence of rhodoliths and calcified invertebrates. This highlights the key-role of calcareous macroalgae in inducing autogenic ecological succession (Laborel, 1961; Laubier, 1966; Odum, 1971; Begon et al., 1986; Basso et al., 2007). In some circumstances, the epibenthic associations of the CDB, particularly rhodolith beds, can be considered as being at the threshold between a circalittoral sandy bottom and an actual upcoming coralligenous bank. A CDB facies with small concretionary boulders has also been described by Meinesz et al. (1983). This does not mean that CDB are necessarily hard substrates. The slow evolution towards a coralligenous bank cannot be generalized to all CDB, even in presence of rhodolith beds.

Bryozoan facies, as described by Bianchi (2009a) and Harmelin (2017), have not been found at the 29 sampling sites considered in the present work. The patchiness of epibenthic associations and facies and the rarity of such a feature might explain our result. Yet facies of Pentapora fascialis and other branched bryozoans on CDB are known for example around Port-Cros Island (Laborel et al., 1976; Holon and Harmelin, 2014; Harmelin, 2017) but were not found during our investigations. Such rarity highlights the conservation issue and the importance of protecting and monitoring CDB. Bryozoans are known to contribute significantly to the bioclastic part of soft bottoms worldwide (Halfar et al., 2006).

The assessment of widespread and patchy habitat such as CDB requires a seascape approach. It is known that the seascape can directly influence the trophic network and the ecosystem functioning (Boström et al., 2011; Abadie et al., 2018; Santos et al., 2022). In the light of that, a seascape approach should be a potential suitable descriptor to assess the quality of an ecosystem (e.g., marine forests, habitat structure, etc.; Cheminée et al., 2014; Thiriet et al., 2014). We have proposed a combination of methods to provide as broad and as representative an assessment as possible, involving direct observation via SCUBA diving, photoquadrats and seascape videos, taking into account functional ecosystem compartments structuring the seascape. The compartments assessed exhibited structuring communities (filter and suspension feeders, rhodoliths) which contribute to the seascape. For example, Gobert et al. (2014) proposed the LIMA index, which is based on a topographical and a biological description and could be applied on all types of habitat from 0 to 40 m depth. Such an approach could be of interest for CDB hosting structuring facies and associations; in the case of CDB without such features, a relatively unstructured seascape cannot be interpreted as a low-functioning ecosystem, considering the macrofauna of the sediment, a compartment of importance but with little or barely perceptible consequences for the seascape structure (bioturbation, calcified organisms providing bioclastic elements, etc.).

Our analysis showed strong and expected correlations between some of the descriptors. Defining an indicator implies the selection of suitable descriptors but that are not too numerous for an operational implementation. Our purpose was to assess the main compartments of the ecosystem, as defined in Figure 2, using as few descriptors as possible. The correlation observed between the final set of descriptors could lead us to reduce the number of descriptors and then simplify the assessment and avoid redundancy; at the same time, a combination of correlated descriptors assessing different ecological functions improves the robustness. As reported by Personnic et al. (2014), the removal of correlated boxes might result in a loss of accuracy of the method.

An ecosystem-based approach can be useful for Ecosystem Services assessment, providing quantitative data for the different ecosystem compartments (Seitz et al., 2014; Burgos et al., 2017). Such an approach has already been carried out for several ecosystems, including seagrass meadows and considering different bundles of ecosystem services according to different quality status (Kermagoret et al., 2019; Scémama et al., 2020).

4.3 A tool to support assessment of anthropogenic pressures

The next step should be the design of an anthropogenic pressure assessment tool to estimate the links between ecological status and cumulated anthropogenic pressures (Holon et al., 2015; Guarnieri et al., 2016; Quemmerais-Amice et al., 2020). Assessing anthropogenic pressures is challenging. Moreover, the assessment of human activities can be considered as a good proxy. In the present work, we did not identify visible evidence of trawling or dredging, the main physical pressures affecting CDB. Such evidence can be difficult to find, even in long-time trawled areas (Ordines et al., 2017). The main evidences of stress observed was the occurrence of macrowastes (e.g., glass, plastic, fishing gears) and invasive alien species (e.g., Caulerpa cylindracea) or the transition to Muddy detrital bottoms. A complementary approach is needed to better characterize human pressures. Our study sites are not located in areas with heavy impact of trawling activity, compared to e.g., the Gulf of Lion, the Catalan Coast or Balearic Islands (Farriols et al., 2017; Jac and Vaz, 2020). New data in different configurations with regard to anthropogenic pressures need to be sampled and assessed.

4.4 The need for an effective protection

The increase in the surface area of Marine Protected Areas driven by the European Union since the 2010s has led to the extension of MPAs over the continental shelf, including coastal detrital bottoms. The new borders of these MPAs do not automatically mean an effective protection and management, and it takes time for coastal managers to integrate these new areas in dashboards and management plans (Vassallo et al., 2020). As a result, the ecological status of CDB sites shows no correlation with the management level in contrast with the findings of Thibaut et al. (2017), who evidenced differences in sublittoral rocky reefs according to the level of management. In addition, we observed no statistical difference between Natura 2000 and unprotected sites, pointing some weaknesses already discussed by Gianni et al. (2022). These results highlight the need to enhance the conservation of widespread ecosystems of the continental shelf such as the CDB, requiring the implementation of effective marine protected areas (Sala et al., 2021).

The conservation or the restoration of CDB implies the reduction or at least the regulation of anthropogenic pressures to allow the natural resilience of communities identified as impacted by such activities. In that case, the role of effective MPAs is limited to the regulation or prohibition of certain human activities: e.g., trawling, dredging, artisanal fishing, mooring. The regulation of other pressures (e.g., terrigenous inputs, the spread of NIS, eutrophication, sedimentation) calls for larger scale policies.

4.5 EBQI implementation for CDB

As an indicator embracing the concept of the ecosystem-based approach, the aim of the EBQI defined for CDB is to provide relevant and useful data for both European Directives (MSFD, WFD, HD) and local management purposes. Compared to other indicators dedicated to soft bottoms (e.g., Borja et al., 2000; Labrune et al., 2006; Labrune et al., 2021), using the EBQI does not require a strong expertise in taxonomy even if more parameters are involved for its calculation. Fifteen descriptors are needed to assess the EBQI for CDB, 13 for Posidonia oceanica seagrass meadows (Personnic et al., 2014) and 10 for infralittoral rocky reefs (Thibaut et al., 2017). Data sampling can be carried out in less than one day per site, involving three operators, which is reasonable if planning a long-term monitoring network at regional (e.g., French Mediterranean coast) or local scale (e.g., MPA; Astruch et al., 2022b). The frame proposed by the EBQI can provide a standardized method for assessing the conservation status according to the Habitats Directive (e.g., assessment of the structure and the function of a given habitat), where diagnosis is usually based on expert judgement and broad data. Repeated applications over time would be welcome to assess the suitability of this index for describing change following management actions (Mancini et al., 2020; Oprandi et al., 2022).

The EBQI on CDB should be applied and tested in contexts with other biogeographical specificities and pressures: oligotrophic, warmer and impacted by non-indigenous species, where very little knowledge is available (e.g., Southern and Eastern Mediterranean Sea). The same approach could also be applied on subtidal rhodolith beds in other areas worldwide of high concern with regard to related human activities (e.g., Brittany). The strength of the EBQIs implemented from now in the Mediterranean Sea is the possibility to measure an ecological status even without all the descriptors assessed with the appropriate method, using the Confidence Index (CI) (see Bevilacqua et al., 2020; Güreşen et al., 2020). This point could allow a preliminary evaluation of CDB in areas with limited human and material resources.

5 Conclusion

An original Ecosystem-Based Quality Index (EBQI) for the Coastal Detrital Bottoms (CDB) has been defined, inspired by already existing EBQIs for Posidonia oceanica seagrass meadows, coralligenous reefs, underwater marine caves, and infralittoral rocky reefs, all based mainly on the assessment of their epibenthic assemblages. Relatively easy to implement and robust, the CDB EBQI provides a fit-for-purpose tool to assess the quality of this widespread ecosystem in accordance with public policies at national and European levels. In the context of the increase in the surface areas of Marine Protected Areas, management goals should include CDB in upcoming plans to secure the conservation of a habitat of importance at Mediterranean scale. Our approach is mainly focused on epibenthic assemblages and improvement is needed in order to better take into account the assessment of the macrofauna of the sediment. We suggest applying and testing this approach in other areas within the Mediterranean Sea and beyond.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

All authors (excepting AO, NM, BM, and BD) participated in a workshop in Marseille in April 2019. PA, BB, TS, and AO carried out the field investigations and the data sampling in 2020 and 2021. All authors contributed to the content for their associated sections. TS and AO produced the statistical analysis. PA compiled content from all authors and prepared the final version. All authors contributed to the article and approved the submitted version.

Funding

The present work was funded by the French Biodiversity Office in the frame of the ACDSea project (Assessment of Coastal Detrital ecosystem conservation Status: an ecosystem-based approach).

Acknowledgments

This work was carried out in the frame of the ACDSea project (Assessment of Coastal Detrital ecosystem conservation Status: an ecosystem-based approach). We kindly thank the French Biodiversity Office (OFB) for financial support, the experts involved in the working out of the index, Michèle Perret-Boudouresque for documentation assistance, and Michael Paul for proof-reading of the text. This work is dedicated to the memory of our late colleague Boris Daniel, involved in the ACDSea project since the first steps.

Conflict of interest

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

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmars.2023.1130540/full#supplementary-material

Glossary

Life Science Identifiers

Taxa are sorted in order of appearance.

Posidonia oceanica (L.) Delile urn:lsid:marinespecies.org:taxname:145794

Cystoseira C. Agardh urn:lsid:marinespecies.org:taxname:144126

Ericaria Stackhouse urn:lsid:marinespecies.org:taxname:1535414

Gongolaria Boehmer urn:lsid:marinespecies.org:taxname:369694

Osmundaria volubilis R.E.Norris urn:lsid:marinespecies.org:taxname:144841

Lithophyllum racemus (Lamarck) Foslie urn:lsid:marinespecies.org:taxname:145160

Lithothamnion corallioides (P.Crouan & H.Crouan) P.Crouan & H.Crouan urn:lsi:marinespecies.org:taxname:145165

Lithothamnion crispatum Hauck urn:lsid:marinespecies.org:taxname:145166

Lithothamnion minervae Basso urn:lsid:marinespecies.org:taxname:145174

Lithothamnion valens Foslie urn:lsid:marinespecies.org:taxname:145180

Neogoniolithon hauckii (Rothpletz) R.A.Townsend & Huisman urn:lsid:marinespecies.org:taxname:169942

Spongites fruticulosa Kützing urn:lsid:marinespecies.org:taxname:494816

Peyssonnelia crispata Boudouresque & Denizot urn:lsid:marinespecies.org:taxname:145277

Peyssonnelia heteromorpha (Zanardini) Athanasiadis urn:lsid:marinespecies.org:taxname:1311383

Peyssonnelia rosa-marina Boudouresque & Denizot urn:lsid:marinespecies.org:taxname:145285

Peyssonnelia rubra (Greville) J.Agardh urn:lsid:marinespecies.org:taxname:145287

Peyssonnelia squamaria (S.G.Gmelin) Decaisne ex J.Agardh urn:lsid:marinespecies.org:taxname:145288

Sporolithon mediterraneum Heydrich urn:lsid:marinespecies.org:taxname:496041

Pentapora fascialis (Pallas, 1766) urn:lsid:marinespecies.org:taxname:111082

Smittina cervicornis (Pallas, 1766) urn:lsid:marinespecies.org:taxname:111551

Turbicellepora avicularis (Hincks, 1860) urn:lsid:marinespecies.org:taxname:111285

Hornera Lamouroux, 1821 urn:lsid:marinespecies.org:taxname:111041

Cladocora caespitosa (Linnaeus, 1767) urn:lsid:marinespecies.org:taxname:135146

Eunicella singularis (Esper, 1791) urn:lsid:marinespecies.org:taxname:125365

Eunicella verrucosa (Pallas, 1766) urn:lsid:marinespecies.org:taxname:125366

Leptogorgia sarmentosa (Esper, 1791) urn:lsid:marinespcies.org:taxname:125369

Alcyonium palmatum Pallas, 1766 urn:lsid:mariespecies.org:taxname:125334

Pennatula rubra (Ellis, 1764) urn:lsid:marinespecies.org:taxname:128519

Pteroeides griseum (Bohadsch, 1761) urn:lsid:marinespecies.org:taxname:181504

Palinurus elephas (Fabricius, 1787) urn:lsid:marinespecies.org:taxname:107703

Pinna nobilis L. 1758 urn:lsid:marinespecies.org:taxname:140780

Laminaria rodriguezii Bornet urn:lsid:marinespecies.org:axname:145729

Carpomitra costata (Stackhouse) Batters urn:lsid:marinespecies.org:taxname:145910

Codium bursa (Olivi) C.Agardh urn:lsid:marinespecies.org:taxname:145079

Codium vermilara (Olivi) Delle Chiaje urn:lsid:marinespecies.org:taxname:145093

Sphaerococcus coronopifolius Stackhouse urn:lsid:marinespecies.org:taxname:145908

Codium fragile (Suringar) Hariot urn:lsid:marinespecies.org:taxname:145086

Arthrocladia villosa (Hudson) Duby urn:lsid:marinespecies.org:taxname:145306

Sporochnus pedunculatus (Hudson) C.Agardh urn:lsid:marinespecies.org:taxname:145915

Acetabularia acetabulum (L.) P.C.Silva urn:lsid:marinespecies.org:taxname:494795

Padina pavonica (Linnaeus) Thivy urn:lsid:marinespecies.org:taxname:145385

Umbraulva dangeardii M.J.Wynne & G.Furnari urn:lsid:marinespecies.org:taxname:840293

Chrysymenia ventricosa (J.V.Lamouroux) J.Agardh urn:lsid:marinespecies.org:taxname:145845

Sebdenia (J.Agardh) Berthold urn:lsid:marinespecies.org:taxname:144177

Caulerpa cylindracea Sonder urn:lsid:marinespecies.org:taxname:660621

Caulerpa taxifolia (M.Vahl) C.Agardh urn:lsid:marinespecies.org:taxname:144476

Acinetospora crinita (Carmichael) Sauvageau urn:lsid:marinespecies.org:taxname:145398

Chrysonephos lewisii (W.R.Taylor) W.R.Taylor urn:lsid:marinespecies.org:taxname:375956

Nematochrysopsis marina (J.Feldmann) C.Billard urn:lsid:marinespecies.org:taxname:375863

Zosterocarpus oedogonium (Meneghini) Bornet urn:lsid:marinespecies.org:taxname:145476

Pseudochlorodesmis furcellata (Zanardini) Børgesen urn:lsid:marinespecies.org:taxname:144485

Valonia macrophysa Kützing urn:lsid:marinespecies.org:taxname:145883

Womersleyella setacea (Hollenberg) R.E.Norris urn:lsid:marinespecies.orgtaxname:146371

Acrothamnion preissii (Sonder) E.M.Wollaston urn:lsid:marinespecies.org:taxname:144488

Sphaerechinus granularis (Lamarck, 1816) urn:lsid:marinespecies.org:taxname:124427

Stylocidaris affinis (Philippi, 1845) urn:lsid:marinespecies.org:taxname:124268

Centrostephanus longispinus (Philippi, 1845) urn:lsid:marinespecies.org:taxname:124331

Echinus melo Lamarck, 1816 urn:lsid:marinespecies.org:taxname:124294

Cerithium Bruguière, 1789 urn:lsid:marinespecies.org:taxname:137760

Aplysia L., 1767 urn:lsid:marinespecies.org:taxname:137654

Sarpa salpa (L., 1758) urn:lsid:marinespecies.org:taxname:127064

Laevicardium oblongum (Gmelin, 1791) urn:lsid:marinespecies.org:taxname:139006

Acanthocardia deshayesii (Payraudeau, 1826) urn:lsid:marinespecies.org:taxname:138991

Suberites domuncula (Olivi, 1792) urn:lsid:marinespecies.org:taxname:134282

Petta pusilla Malmgren, 1866 urn:lsid:marinespecies.org:taxname:130597

Turritellinella tricarinata (Brocchi, 1814) urn:lsid:marinespecies.org:taxname:1381415

Moerella donacina (L., 1758) urn:lsid:marinespecies.org:taxname:147021

Abra nitida (O. F. Müller, 1776) urn:lsid:marinespecies.org:taxname:141435

Abra prismatica (Montagu, 1808) urn:lsid:marinespecies.org:taxname:141436

Holothuria L., 1767 urn:lsid:marinespecies.org:taxname:123456

Bonellia viridis Rolando, 1822 urn:lsid:marinespecies.org:taxname:110363

Spatangus purpureus O.F. Müller, 1776 urn:lsid:marinespecies.org:taxname:124418

Chaetaster longipes (Bruzelius, 1805) urn:lsid:marinespecies.org:taxname:124004

Echinaster sepositus (Retzius, 1783) urn:lsid:marinespecies.org:taxname:125161

Cliona viridis (Schmidt, 1862) urn:lsid:marinespecies.org:taxname:134146

Spicara maena (L., 1758) urn:lsid:marinespecies.org:taxname:126828

Spicara smaris (L., 1758) urn:lsid:marinespecies.org:taxname:126830

Boops boops (L., 1758) urn:lsid:marinespecies.org:taxname:127047

Chromis chromis (L., 1758) urn:lsid:marinespecies.org:taxname:127000

Anthias anthias (L., 1758) urn:lsid:marinespecies.org:taxname:127031

Scyliorhinus Blainville, 1816 urn:lsid:marinespecies.org:taxname:105729

Serranus hepatus (L., 1758) urn:lsid:marinespecies.org:taxname:127042

Blennius ocellaris L., 1758 urn:lsid:marinespecies.org:taxname:126761

Symphodus cinereus (Bonnaterre, 1788) urn:lsid:marinespecies.org:taxname:273567

Solea Quensel, 1806 urn:lsid:marinespecies.org:taxname:126132

Lophius L., 1758 urn:lsid:marinespecies.org:taxname:125802

Scorpaena scrofa L., 1758 urn:lsid:marinespecies.org:taxname:127248

Dentex dentex (L., 1758) urn:lsid:marinespecies.org:taxname:273962

Pagrus pagrus (L., 1758) urn:lsid:marinespecies.org:taxname:127063

Zeus faber L., 1758 urn:lsid:marinespecies.org:taxname:127427

References