Combining literature review, acoustic mapping and in situ observations: an overview of coralligenous assemblages in Liguria (NW Mediterranean Sea)

A review and update of the existing knowledge on the coralligenous assemblages of Liguria (NW Italy) was conducted as an essential step towards management measures for their conservation according to the EU Marine Strategy Framework Directive. By combining a literature review, acoustic mapping and in situ observations on a geographic information systems platform, we were able to assess the distribution and heterogeneity of coralligenous assemblages and the main pressures affecting them. The reliability of the literature was previously estimated using a dependability index. The coralligenous assemblages cover an area of 130.9 ha and range from 10 to 113 m depth. Twelve different biological facies (five of them not included in the EUNIS list) were identified and four main geomorphotypes (plungingcliffs, paleocliffs, rockfalls and shoals) were recognized. Incident light values influenced the distribution of four facies in Portofino promontory. Pressures were found on 33% of the coralligenous assemblages investigated, mainly due to fishing activities, mass mortality events, invasive species and occasional mucilaginous events. Our results showed a high spatial, geomorphological and biological heterogeneity of coralligenous assemblages in Liguria.


INTRODUCTION
Since Marion (1883) first coined the word "coralligène", coralligenous assemblages have been studied, classified and defined as hard substrata of biogenic origin that are mainly produced by the accumulation of calcareous encrusting algae, light being the main environmental factor affecting their development (Ballesteros 2006).Considered one of the most important and characteristic habitats of the Mediterranean Sea (Laborel 1961, 1987, Laubier 1966), coralligenous assemblages represent the climax biocenosis of the circalittoral zone (Pérès and Picard 1964), where endangered or commercially important fish species are known to live, feed or reproduce (Salomidi et al. 2012).Coralligenous communities are considered of high conservation interest because of their biogeographic uniqueness, their highly varied physical structure, their high species biodiversity, their highly diversified occurrence stratified throughout the different benthic marine zones and their slow growth (Ballesteros 2006, Agnesi et al. 2008, Salomidi et al. 2012).
Trawling has effects on seafloor that resemble those of forest clear-cutting (Watling and Norse 1998).It is the most destructive impact affecting the coralligenous habitat, but it is not the only one.Artisanal fishing, anchoring, diving, waste water discharges, aquaculture facilities, changes in land use, coastal infrastructure construction, urbanization, anomalous high water temperatures, proliferation of mucilage and invasive species are currently affecting the biodiversity and structure of coralligenous assemblages (Baldacconi and Corriero 2009, Ballesteros 2009, Linares et al. 2010).
Although the importance of coralligenous assemblages and the threats affecting them are commonly accepted within the scientific community and among managers, they were not listed among the priority habitats defined by the EU Habitat Directive (92/43/EEC), thus remaining without formal protection.Only one European law (Council Regulation (EC) Nº 1967/2006) prohibits destructive fishing over Mediterranean coralligenous and mäerl communities but it remains ineffective in the current scenario of a lack of relevant geospatial data (UNEP-MAP-RAC/SPA 2008, Salomidi et al. 2012).Although some efforts have been made to increase information on the distribution of this sensitive habitat (Georgiadis et al. 2009, Giakoumi et al. 2013, Martin et al. 2014), knowledge of the geographical and depth distribution of coralligenous assemblages as well as their biodiversity and its relation to their functioning is still needed for their conservation and sustainable use (UNEP-MAP-RAC/SPA 2009, European Commission 2011, Pergent 2011).The current situation is not helped by the fragmented geopolitical scenario characterizing the Mediterranean basin, with weak, uncoordinated, conflicting approaches or non-existent regulatory frameworks, policy mechanisms and enforcement (Fraschetti et al. 2011).Improving the current and future monitoring of marine biodiversity is essential in order to implement the Marine Strategy Framework Directive (MSFD) and to achieve Good Environmental Status (GES).In addition to enhancing regional cooperation, there is a need to further develop innovative and cost-effective monitoring methods and to incorporate state-of-the-art technological developments into current monitoring practices (Patrício et al. 2014).
Habitat mapping is crucial in order to obtain reliable estimates of the total area occupied by each habitat, so as to reach the required conservation targets and to implement appropriate management measures (UNEP-MAP-RAC/SPA 2008, Costello 2009).A bibliographic database plays a major role in the study of habitat distribution and its changes over time.In some Mediterranean regions, information on coralligenous distribution is available in the literature but its reliability must be estimated (Leriche et al. 2001).Acoustic technologies are increasingly being used for monitoring benthic habitats (Zapata-Ramirez et al. 2013), and those used for mapping coralligenous communities normally include side-scan sonars and multibeam sonars (Bonacorsi et al. 2012, Gordini et al. 2012).Acoustic mapping using multibeam echo sounders allows us to infer seafloor topography, and therefore the geomorphologies on which coralligenous assemblages thrive, and the best locations for ground truthing.The European Commission decision on descriptors for the MSFD states that data of geographic information systems (GIS) are a prerequisite for ecosystem-based management of human activities and for developing related spatial tools.Modelling and mapping using GIS provide the basis for informed decision making (Sardá et al. 2012).
In this work we combined a literature review, acoustic mapping and in situ observations to (i) build a map of the distribution of coralligenous assemblages in Liguria (NW Italy) using GIS; (ii) characterize their biology and geomorphology and study the influence of geomorphology and light conditions on biological features; and (iii) consider the main pressures affecting coralligenous assemblages and provide some management recommendations.

Study area
The Ligurian coast falls into the Ligurian Sea, the northernmost part of the western Mediterranean Sea, with 350 km of arch-shaped coastline from Ventimiglia at the west border with Provence-Alpes-Côte d'Azur (France) to La Spezia at the east border with Tuscany, Italy.Almost two thirds of the coast is composed of cliffs and promontories alternating with sandy to gravelly pocket beaches, especially along the Eastern Riviera.Soft coasts are less developed and are typically found adjacent to small coastal plains along the Western Riviera (Cattaneo-Vietti et al. 2010, Rovere 2011).

Literature review and database creation
We combined a literature review, acoustic mapping and in situ observations to build a data layer of coralligenous presence using the ArcGIS ® software by Esri (Environmental Systems Resource Institute), ArcMap 10.2 (www.esri.com).The existing literature was collected from scientific publications, technical reports ("grey literature"), diving and fishing books; we also considered unpublished in situ observations by experts and recreational divers (Cánovas- Molina et al. 2014).In order to estimate the reliability of the literature reviewed, a dependability index (DI) was applied, considering two parameters: geospatial positioning (five points) and method adopted for data acquisition (five points) (Table 1).The DI was not calculated for documents taken from the literature reviews because of the variety of methods used for data acquisition.A score of 1 for at least one of the parameters measured in the DI, was considered the critical value for low dependability and therefore rejection.Where digital spatial information was not available, polygons of presence of coralligenous assemblages were created by extracting spatial information from textual descriptions, helped by acoustic mapping (multibeam echo sounder) provided by Regione Liguria, when available.We carried out ground truthing through SCUBA diving at selected sites (Vado Ligure, Cogoleto, Arenzano and Punta Manara), following expert divers' observations and/ or seafloor topography revealed by multibeam echo sounder.

Database analysis
A list of characteristic sessile megabenthos species of coralligenous assemblages, identified from the literature review and in situ observations, was compiled.Facies were identified following the EUNIS Habitat Classification (Davies et al. 2004).The habitat classification developed and managed by the European Topic Centre for Nature Protection and Biodiversity (ETC/ NPB in Paris) provides a comprehensive typology for the habitats of Europe and its adjoining seas.EUNIS habitats are arranged in a hierarchy, from level 1 to 4 for marine habitats.Information on the most conspicuous species was used to identify the facies for each site with coralligenous.
Geomorphotypes where coralligenous assemblages in Liguria thrive were identified using multibeam echo sounder, when available, and textual descriptions from the literature gathered.Four geomorphotypes were considered: i) cliffs, vertical or near-vertical walls from a steep rock face, either active (plunging cliffs) or inactive (paleocliffs), ending occasionally in a flat bottom; ii) rockfalls, sets of masses of rocks of different sizes and shapes that lay on a bedding plane that originated from the face collapse of the cliff; and iii) shoals, isolated outcrops surrounded by sand or biodetrital sediments, which originated through past erosional events of the cliff, sea level rise and deposition of sediments.
To study the influence of incident light on coralligenous assemblages, we calculated incident light values on coralligenous assemblages from satellite imagery and long term-Secchi Disc records in Liguria available for the Portofino coastal area (Rapallo, Genoa).Percent light values of surface irradiance at each station in Portofino were calculated using standard methods with the formula E z =E 0 e -kz (Kirk 1994), where E z is the irradiance at depth z, E 0 is the surface irradiance, K is the vertical attenuation coefficient and z is the depth.K was approximate to K dPAR and obtained using Secchi Disc transparency records (January 2002-December 2012) using the formula K dPAR =1.7/Secchi Depth (m).E 0 was approximated to photosynthetically active radiation and obtained from Goddard Earth Sciences Data and Information Services Centre (GES DISC) Interactive Online Visualization and Analysis Infrastructure, "Giovanni" (January 2002-December 2012).Percent light values of surface irradiance on facies of coralligenous assemblages for each station were calculated as the mean value among incident light values at the shallowest and the deepest distribution.To test the influence of incident light on the distribution of coralligenous facies in Portofino, a one-way ANOVA was applied.
Information on presence of mechanical and biological pressures, as well as signs of mass mortality events due to water temperature increasing, was reported for each site with coralligenous assemblages.

Literature reliability
From 1937 to date, 58 documents described some aspects of the coralligenous habitat in Liguria (Supplementary material Table S1), mainly scientific papers or technical reports written in Italian.The Eastern Riviera and Marine Protected Areas (MPAs), i.e.Portofino (Rapallo, Genoa), Punta Mesco (Cinque Terre, La Spezia), and Isole di Portovenere (La Spezia), were among the most studied localities, accounting for 70% and 58% of the documents respectively.Data were acquired mainly by means of scientific SCUBA diving during the last two decades.Acoustic mapping was available for 10 localities.Eleven documents showed score 1 for at least one of the parameters of the DI and therefore have low dependability.

Distribution of coralligenous assemblages
A total of 90 sites within 18 localities were found to have coralligenous assemblages with an extent of 130.9 ha (Fig. 1 and Supplementary material Table S2).Seven localities (59.6 ha) were found along the Eastern Ligurian Riviera and 11 localities (71.3 ha) along the Western Riviera, the furthest location being 4 km from the coast at Capo Berta (Imperia) and the closest near the shoreline at Portofino.Regional and National protection laws applied for 48.3 ha (37%) situated in Portofino, Cinque Terre and Portovenere.
The shallowest coralligenous assemblages in Liguria were those of Portovenere and Portofino at 10 m depth, the deepest that of Bergeggi at 113 m depth.

Biology, geomorphology and incident light values
A preliminary attempt to estimate the total number of sessile megabenthos taxa in the Ligurian coralligenous habitat is presented in Supplementary material Table S3.Sponges were the dominant group in terms of species richness (Fig. 2).The literature review permitted us to identify 12 facies of coralligenous communities (Table 2).In Liguria four different coralligenous geomorphotypes were identified (Fig. 3).Shoals were the dominant morphology in Liguria in terms of extent, with 50.6 ha, and were mainly represented in the Western Riviera.Plunging cliffs, paleocliffs and rockfalls covered 32.6, 29.8 and 17.9 ha, respectively (Table 3).Paleocliffs and rockfalls were mainly colonized by facies with Paramuricea clavata.In fact, this facies  In Portofino, incident light values on coralligenous communities varied from 0.003% to 28.37% of surface irradiance (Fig. 4).Eight facies of coralligenous assemblages thrived in Portofino promontory at different % values of surface irradiance (Fig. 5).A one-way ANOVA on incident light values showed significant differences between the facies with Paramuricea clavata, Eunicella cavolini, Corallium rubrum and Leptogorgia sarmentosa (p<0.01,F=6.13, df=3).Facies with Eunicella cavolini showed the highest distribution range of incident light values.On the other hand, Corallium rubrum, Paramuricea clavata and Leptogorgia sarmentosa thrived on lower and narrower ranges of light conditions.

Main pressures
Pressures reported in the literature were mainly due to fishing activities, such as the presence of abandoned gears such as nets, traps and "palamiti" (a type of longliner).Other pressures were anchors, plastic bags, invasive species such as the macroalgae Caulerpa cy-lindracea and Womersleyella setacea, occasional mucilaginous events, and mass mortality events linked to climate warming.In Liguria, evidence of one or more such pressures was found in 30 of 90 sites.

Database of coralligenous assemblages
A total of 47 documents describing aspects of coralligenous habitats in Liguria were considered reliable after applying the DI for literature, and were therefore used to assess the distribution, biology, geomorphology and pressures.As digital spatial information and GPS positioning were scarce in the literature review, combining it with acoustic mapping and in situ observations helped improve accurate positioning on a GIS platform.Literature dealing with coralligenous assemblages in Liguria is more detailed for the Eastern Riviera, accounting for 70% of the documents, where their distribution is mainly associated with cliffs.The Western Riviera has been poorly studied and rarely characterized due to its particular distribution on isolated rocky outcrops of little extent found patchily along the coast.

Distributional, biological and geological aspects
Coralligenous habitats can be defined as highly heterogeneous systems, where the environmental variables together with distribution and abundances of taxa can differ greatly on both a geographical and a local (tens of metres) scale (Ferdeghini et al. 2000, Balata et al. 2005, Virgilio et al. 2006).The coralligenous assemblages in Liguria therefore show high spatial, morphological and biological variability (Fig. 6).
The shallowest distribution of coralligenous communities in Liguria was found in Portovenere, where they developed from 10 to 30 m depth associated with low light irradiance.Sediment load coming from the Magra River, located about 10 km away, explains the high turbidity in the water, with Secchi Disc values of 6.2±1.6 m (Bassano et al. 2000).Twelve facies were identified in the Ligurian coralligenous assemblages.In the marine environment, light is a key environmental variable coupling physics to biogeochemistry and ecology.According to Ballesteros (1992), coralligenous communities develop between 0.05% and 3% of the surface irradiance."Dorsale Portofino" reaches light values below 0.05% (0.003-0.14%)where coralligenous assemblages grow on shoals between 52 and 84 m depth and are dominated by suspension feeders: gorgonians, antipatharians and sponges with few calcareous concretions.Light values greater than 3% correspond to coralligenous communities above 27 m depth with a high frequency of so-called "precoralligenous" species (Bianchi et al. 2004), such as Cladocora caespitosa, Codium bursa, Flabellia petiolata, Halimeda tuna and Sarcotragus fasciculatus as components of coralligenous communities with high light tolerance.
In coralligenous assemblages two main geomorphologies have been described (Pérès andPicard 1964, Ballesteros 2006): i) banks, which can fall into the coralligène de plateau definition and consist of flat frameworks with a variable thickness (0.5 to 4 m) built on horizontal substrate; and ii) rims, which can fall into the coralligène de l'horizon inférieur de la roche littorale definition and consist of structures developed along vertical cliffs with a thickness range of 0.2 to 2 m.Some problems arise in trying to classify coralligenous formations into these two categories due to the high variability of the build-ups in terms of geometry, dimension, areal distribution and settled substrate (Bracchi et al. 2014).In Liguria, the high variability of origin, extension, slope and shape of geomorphologies led us to identify, as did Laborel (1961), four different geomorphotypes: cliffs (paleocliffs and plunging cliffs), shoals, and rockfalls.

Main pressures
Pressures were found at 33% of the sites, mainly due to professional and recreational fishing activities, mass mortality events, presence of invasive species and mucilaginous events.In Portofino, a major cause of mortality in Paramuricea clavata is damage by fishing lines, followed by the attachment of numerous epibionts (Bavestrello et al. 1997).Coralligenous habitats have been invaded by alien species such as Caulerpa cylindracea (Piazzi et al. 2007) and/or Womersleyella setacea (Gatti et al. 2015), with consequent loss of structure and biodiversity of the original assemblages.In the late summer of 1999, when a sudden increase in sea water temperature in the Ligurian Sea occurred down to 50 m depth, an extensive mortality of gorgonians and other epibenthic organisms was observed in Liguria: the proportion of affected gorgonians ranged from 60% to 100%, suggesting that millions of sea fans died (Cerrano et al. 2000).

Management recommendations
We highly recommend including coralligenous habitats in the EU Habitats Directive (92/43/EEC) as a priority natural habitat type.In addition, MPAs must be established in order to protect representative coralligenous assemblages by applying the protection and management measures recommended by Articles 6 and 7 of the SPA protocol (Ballesteros 2009).In Liguria, 48.3 ha (37%) of coralligenous habitats are inside national or regional MPAs.Among the 82.6 ha of coralligenous assemblages without a specific conservation plan, the assemblages of Bergeggi and the deep (>70 m) assemblages of Portofino lie outside the limits of the MPAs of "Isola di Bergeggi" and "Portofino", respectively.Therefore, we recommend the extension of the two MPAs in order to include them.
To enforce the MSFD and achieve GES, we suggest the implementation of a common monitoring programme for coralligenous habitats of Liguria.This monitoring programme should fulfil the following requirements: -It should take into account their high geographic and biological heterogeneity and therefore be applied also in the poorly studied shoals of the Western Riviera.
-It should be able to assess ecological status and pressures and their linkage.Ecological status assessment should consider coralligenous structure (e.g.species richness and percentage cover), as well as functioning (e.g.connectivity and population dynamics).
-It should include new methods and technologies.The use of photo-video mosaicing can provide information on both landscape level (metre-scale) maps and high-resolution (sub-millimetre) images of individual colonial organisms (Reid et al. 2010).If further validated, the photo mosaic techniques and 3D reconstructions would provide a cost-effective tool for assessing ecological status and monitoring changes due to natural or anthropogenic disturbance in these environments (Zapata-Ramirez et al. 2013).
Combining a literature review (scientific publications, unpublished documents and books), acoustic mapping and in situ observations (ground truthing and personal communications) on a GIS platform resulted in an efficient, powerful tool for the assessment of the distribution and heterogeneity of coralligenous communities and the pressures affecting them.This methodology could give information on the real extent and diversity of these assemblages to help governments and stakeholders implement appropriate management measures to guarantee sustainable development of coastal areas.

Table 1 .
-Scores of the dependability index (up to 5+5 points) from the literature.

Table 3
. -Distribution of coralligenous facies on the different geomorphotypes in Liguria (values in hectares).N/A, no data available.

Table S1 .
-Literature review on coralligenous assemblages in Liguria.Scarce, scarce biological information or just focused on one species.DI (GP+DA), dependability index of literature (Geospatial Positioning + Data Acquisition).*Score=1for at least one of the parameters considered in the DI.

Table S2 .
-Ligurian localities with coralligenous and main characteristics.N/A, No data available.* Facies not included in the EUNIS Habitat 2004 classification for Mediterranean coralligenous communities.