Decapod crustacean assemblages off the West coast of central Italy ( western Mediterranean )

Community structure and faunal composition of decapod crustaceans off the west coast of central Italy (western Mediterranean) were investigated. Samples were collected during five trawl surveys carried out from June 1996 to June 2000 from 16 to 750 m depth. Multivariate analysis revealed the occurrence of five faunistic assemblages: 1) a strictly coastal community over sandy bottoms at depths <35 m; 2) a middle shelf community over sandy-muddy bottoms at depths between 50 and 100 m; 3) a slope edge community up to 200 m depth as a transition assemblage; 4) an upper slope community at depths between 200 and 450 m, and 5) a middle slope community at depths greater than 450 m. The existence of a shelfslope edge transition is a characteristic of the western and central Mediterranean where a Leptometra phalangium facies is found in many areas at depths between 120 and 180 m. The brachyuran crab Liocarcinus depurator dominates the shallow muddy-sandy bottoms of the shelf, while Parapenaeus longirostris is the most abundant species from the shelf to the upper slope assemblage. The dominant species on the middle slope are mainly nektobenthic (Aristeus antennatus, Aristaeomorpha foliacea, Plesionika martia), bentho-pelagic (Pasiphaea sivado) and benthic species (Nephrops norvegicus). The average diversity of the five assemblages defined increased with depth and the highest diversity values were found in the slope assemblages. Depth is the main factor influencing decapod assemblage composition, but sediment type and dissolved organic matter seem to be secondary structuring factors particularly in the shelf zone.


INTRODUCTION
Decapod crustaceans are a dominant or subdominant faunal component in temperate to subtropical assemblages, such as the Mediterranean Sea, and constitute a key taxon linking lower and higher trophic levels (Wenner and Boesch, 1979;Fariña et al., 1997;Cartes, 1998).In the Mediterranean, they are one of the dominant groups of megabenthic invertebrates on the continental shelf and slope and a very important fraction of demersal communities (Abelló et al., 1988;Maynou and Cartes, 2000;Colloca et al., 2003).The distribution and abundance of decapod crustacean species have been investigated in detail in the north-west basin (Abelló et al., 1988;Cartes and Sardà, 1993;Maynou and Cartes, 2000), particularly on the continental slope (Cartes, 1993;Cartes et al., 1994;Sardà et al., 1994).No data on the crustacean assemblages have been published from central and eastern Mediterranean areas.In Italian waters (western Mediterranean), Arena and Li Greci (1973), Relini (1981) and Tunesi (1986) investigated the quantitative composition of the decapod crustacean fauna occurring in trawlable fishing grounds.
Demersal fauna off the central-western coasts of Italy is exploited by an intensive trawl fishery targeting both commercial fish and invertebrates.Studies on demersal assemblage distribution in this area demonstrated the importance of several crustacean species (Biagi et al., 2002;Colloca et al., 2003) due to their abundance.The main purpose of this study is to examine the faunistic composition and the main decapod crustacean assemblages occurring off the west coast of Central Italy.The factors that may affect species distribution are also discussed.

Study area and data collection
This study was carried out within the framework of the Mediterranean International Trawl Survey (MEDITS) (Bertrand et al., 2002).A total of five bottom trawl surveys were carried out during the summers (June-July) of 1996 to 2000.The area investigated covered 15,289 km 2 off the west coast of central Italy (western Mediterranean) between 16 and 750 m depth (Fig. 1).The project used a randomized stratified sampling design based on depth (five bathymetric strata: 10-50, 51-100, 101-200, 201-500 and 501-700 m) and area.The depth limits were adopted to best cover the distribution areas of the most exploited, or potentially exploitable, species.Locations of stations were selected randomly within each bathymetric stratum at the beginning of the project (1994).Over the following years, a total of 46 fixed stations were sampled in day-time hours (6.00 a.m. to 6.00 p.m.) (Bertrand et al., 2002).Haul duration varied according to depth: 30 min at depths of less than 200 m and 60 min deeper than 200 m.All the surveys were carried out with the same otter bottom trawler using a net with a cod end stretched mesh size of 10 mm.The towing speed of the vessels was about 3.0 knots.The distance covered by the net on the bottom was calculated using GPS positions when the net reached and left the bottom.The mean opening of the net mouth was directly calculated for most of the hauls by a commercial SCANMAR system.The swept area was then obtained by multiplying the mean mouth opening of the net by the towing distance.
In the study area, the continental shelf extends up to 120-150 m depth covering a distance of 15-30 km from the shoreline.The limit between the upper and middle slope ranges between 400 and 500 m depth (Carpine, 1970;Relini et al., 1986).The temperature of 13°C is generally constant down to 120-200 m depth (Ifremer, 1997).A layer of Levantine intermediate water, characterized by maximal salinity, flows anticlockwise along the slope between 250 and 700 m depth (Serravall and Cristofalo, 1999).Temperature and salinity are even more stable (12.8°C; 38.4 psu S) in the deep-sea water masses of the Mediterranean (Reyss, 1971).Data on the main macro-epibenthic species found for each assemblage are available in Colloca et al. (2003).Based on grain size composition, according to Emylianov and Schimkus (1986), the continental shelf shallower than 50 m is characterized by sand (1-0.1 mm fraction dominates), aleuritic (0.1-0.001 mm) and pelitic (<0.001 mm fraction dominates) sediments are found between 50 and 100 m, and bottom sediments are mainly composed of pelitic muds below 100 m.
The study area presents a heterogeneous sedimentary situation which is strongly influenced by terrigenous contributions from several rivers.The northern part is supplied by sediments from the Fiora, Marta, Mignone and Arrone rivers.The sediment type is mainly sandy-clayey (A.A.V.V., 1985).Wave action distributes sediments based on textural characteristics: sandy and sandy-clayey bottoms typify the coastal zone with an increase in the muddy-clayey fraction off-shore (Tortora, 1989).The central part is dominated by the deltic cusp of the Tiber River.The distribution of sediments is the same as on the right side of the Tiber, but in this area the limit between sandy-muddy and pelitic sediments is located at 70-90 m depth (Bellotti and Tortora, 1985).In the central part, grain size strongly decreases, while in the southern part of the central area it increases due to the strong hydrodynamics and the steep bottoms.In the southern part of the study area fine sands are located shallower than 20-30 m depth.Finally, the Gulf of Gaeta is influenced by the Garigliano River contribution.In this area, sedimentation is more homogeneous and reaches the highest rates and the highest thickness of the Latium coasts (Chiocci and La Monica, 1999).

Analysis of abundance data
Abundance index data (Number km -2 ) of decapod crustaceans obtained during the five trawl surveys were pooled in a matrix of species abundance according to station and survey.A Bray-Curtis similarity matrix was then constructed to perform a cluster analysis and a non-parametric multidimensional scaling ordination (nMDS) using PRIMER v5 software (Clarke and Warwick, 1994).
An analysis of similarities (ANOSIM) was used to test the significance of species assemblage differences between trawl surveys (Clarke, 1993).This test assesses significant differences between groups of replicates against a series of random simulations, so that a test statistic (R) is calculated.The R-statistic value, which describes the extent of similarity between each pair in the ANOSIM analysis, can range between 1, when two groups, surveys in this case, are entirely separated and 0 if there is no difference between the two groups.
Since no significant differences between surveys were found in the structure of the assemblages, the data were pooled and average species abundances in relation to trawl station were calculated.An nMDS was then performed on a Bray-Curtis dissimilarity matrix based on average species abundance in relation to trawl station.Uncommon species that were seldom present (which appeared once or twice during the whole sampling period) or exhibited low abundance levels were not included in the data matrix.
The SIMPER procedure (Similarity Percentage analysis) was used to identify those species that characterize each demersal assemblage and those responsible for discriminating between assemblages (Clarke, 1993).
The structure of decapod assemblages was analyzed using community descriptors to obtain a general view.The diversity of each identified assemblage was calculated using Hill's index (N1) as a univariate measure of the species diversity (Hill, 1973).This index, which is the exponential of the Shannon-Wiener function (exp H'), has already been used in studies on fish assemblages (Greenstreet and Hall, 1996;Blanchard, 2001).Margalef's index (D) was used to assess the species richness of each assemblage.One-way analysis of variance (ANOVA) was used to test if differences between the diversity indexes of the assemblages were significant, based on biomass data.We also calculated Spearman's correlation coefficients between assemblages and the projections of species per haul along dimensions 1 and 2 of MDS.The variables were: depth, latitude and longitude.

RESULTS
During the trawl survey a total of 53 species of decapod crustaceans were sampled, most of them belonged to the Caridea and Penaeoidea.Only 5 species of Anomura and 12 of Brachyura were collected during the sampling period (Table 1).
The pair-wise test comparisons (ANOSIM) performed on catch data obtained in each survey showed a low level of overlap between surveys: the pair-wise  R values ranged between 0.05 and 0.40, which indicates a very low segregation level between surveys.The dendrogram of similarities among stations showed that five main groups can be clearly defined along the bathymetric gradient (Fig. 2).The first branching of the similarity tree clearly discriminates between shelf and slope stations.The shelf group can be further sub-divided respectively into coastal shelf stations on sandy bottoms (7 trawl stations, mean depth 33.4 m, SD=11.3),middle-shelf stations on muddy bottoms (10 stations, mean depth 96.7 m, SD=12.7) and shelf-slope edge stations on muddydetritic bottoms (9 stations, mean depth 160.7 m, SD=26.4).There are two main groups on the slope, which join respectively upper slope stations (7 stations, mean depth 333.7 m; SD=51.6) and middle slope stations (13 stations, mean depth 546.5 m; SD=97.1).The bathymetrical division between the coastal and the shelf group is located at about 35 m, while the division between shelf and slope is at about 180 m.The limits between the upper and the middle slope are located at about 400 m depth.Identified assemblages were also separated in nMDS analysis (Fig. 3).A counter-clockwise increase in depth related to a horseshoe effect can be seen in the distribution of trawl stations in the nMDS analysis.The nMDS shows depth as the first key factor, but the horseshoe effect may denote the existence of a structuring factor that is dependent on depth.This is probably a sedimentary factor that represents a gradient of increasingly fine particles (fine silt and mud) in sediments.The stations were distributed gradually from the coastal sandy-detritic bottoms to muddy detritic bottoms on the slope edge.The pair-wise comparisons between groups (Table 2) defined by multivariate analysis showed high separation between assemblages (R= 0.83).
With regards to spatial and bathymetric distribution of decapod assemblages, out of the explored variables, decapods were highly correlated to depth (Spearman coefficient r= 0.951, p<0.001), and a quasi-significant correlation was found with latitude (r= -0.281, p = 0.0581).
The most frequent species differed between depth strata (Table 3): Liocarcinus depurator and Medorippe lanata were the most abundant species between 0 and 100 m depth.The pink shrimp Parapenaeus longirostris dominated from 100 to 300 m depth, below these depths Norway lobster Nephrops norvegicus, pasipheids, aristeids and pandalids showed the highest values of occurrence.
Table 4 shows the percentage contribution of the main contributor species to within-group similarity calculated for the assemblages identified in the survey.Thus, Liocarcinus depurator and Medorippe lanata were the main contributors in the shelf-sand assemblage.Parapenaeus longirostris was the most typifying species in both the shelf-mud (65.7%) and the upper slope assemblages (31.1%).Nephrops norvegicus (17.6%),Macropipus tuberculatus (15.3%) and three species of pandalid shrimps contributed to defining the upper-slope assemblage.The most typical species of the middle-slope assemblage were Aristaeomorpha foliacea (30%), Nephrops norvegicus (22.8%) and Plesionika martia (13.2%).
The average diversity, according to Hill's index (N1) (Table 5), decreased from the coastal shelf to the shelf-slope edge assemblage, but increased strongly from the shelf-slope edge to the upper and middle slope communities (ANOVA: F=19.77; p<0.0001, see Table 6).

DISCUSSION
The decapod crustacean fauna of the central Tyrrhenian presented a clear zonation effect with a series of well-defined bathymetric boundaries.The boundaries seemed to be connected with a variety of factors all related to depth (Macpherson, 1991).
The distribution of species depends on a number of environmental variables, such as temperature and food availability, which can affect both the maximum density attained and the extreme limits of their distribution.The species distribution is wider or narrower in function of physical changes and biotic interactions (Therborg, 1970).Although in this study depth appeared to be one of the main structuring factors for the spatial distribution of species, other environmental variables directly related to depth, such as sediment type and organic matter, could play an important role in the distribution of some taxa (Paguridae, Crangonidae, Astacidea, etc.) mainly linked to edafic factors, especially in the coastal area.In this study the second structuring factor seems to be a sedimentary aspect, which repre-  sents a gradient of increasingly fine particles (fine silt and mud) in sediments (Emylianov and Schimkus, 1986).Sediment grain size, which is determined by hydrodynamic sorting mechanisms, and organic content are considered the second most influential characteristics that determine occupancy by marine benthos (Gray, 1981); this factor could be more important for benthic decapods.Changes in the degradation of organic matter with depth are also important, both in relation to the vertical downward drift through the water column (Wishner, 1980) and the organic matter on the slope (Carpine, 1970;Rowe, 1983).In the coastal zone, this factor could play an important role in the distribution pattern: a significant input of organic matter is brought by different rivers, and this contribution influenced the whole study area.
Cluster analysis showed five faunistic assemblages: the main ecological change in structure and composition of species assemblages occurred moving from the continental shelf to the slope.This zonation pattern of decapod crustacean fauna is similar to that observed in the north-west Mediterranean (Abelló et al., 1988) and other areas of the north Atlantic (Lagardère, 1973(Lagardère, , 1977;;Haedrich, 1975;Haedrich et al., 1980;Markle et al., 1988;Olaso, 1990;Fariña et al., 1997), but in our study we found a clear shelf-slope transition group, between 120 and 180 m depth, not observed in other areas.
Three main assemblages were detected along the continental shelf, which grouped coastal shelf, middle shelf and shelf-slope edge stations respectively.In the coastal zone, the analysis did not allow us to identify differences in the assemblage composition, despite the heterogeneity of environmental conditions due to the presence of eutrophic areas like those affected by the water from the Tiber River (central Latium) and other rivers, lagoons connected with the sea, seagrass meadows, etc.Most of the coastal stations were located on sandy-muddy bottoms characterized by a high sedimentation rate and usually interpreted as ecotonal habitats between the coastal sandy bottoms and the terrigenous muddy bottoms located offshore (Colloca et al., 2003).
The crustacean community on middle shelf soft bottoms is dominated by Brachyurans and hermit crabs with very abundant and frequent species like Liocarcinus depurator, Goneplax rhomboides, Medorippe lanata, Pagurus prideaux and P. excavatus.These species are linked more to sediment type and to the amount of suspended organic matter (Vernberg and Vernberg 1983;Rufino et al., 2004).Sediment characteristics have been found to strongly influence occurrence and distribution of many benthic decapod crustaceans, especially those with burrowing and burying habits (Atkinson and Taylor, 1988;Pinheiro et al., 1996;Comeau et al., 1998).The dominance of benthic crustaceans mainly preying on benthos resources (Cartes and Abelló, 1992;Cartes, 1998) is probably due to benthos resources being more important as a food-source in shelf communities, which agrees with a generally recognized decrease in benthos biomass with depth (Rowe, 1983).
The shelf-slope transition, at depths between 120 and 170 m, is characterized by detritic organogenic sediments colonized by the crinoid Leptometra phalangium, a suspension-feeding macro-epibenthic species confined in the Mediterranean to the shelfbreak area (Colloca et al., 2003).Its abundance in the studied area can reach 12-15 ind.m -2 .Two characteristic species in this transition zone, Parapenaeus longirostris and Macropipus tuberculatus, showed a wide bathymetric distribution that covered both the upper shelf and the slope.
Within the shelf and slope assemblages, the occurrence of species reaching higher concentrations over a restricted depth range determined the occurrence of distinguishable assemblages at a lower dissimilarity level.
On the upper slope, Brachyura and Paguridae are replaced by several groups of Decapoda Natantia, Macrura (N.norvegicus, Polycheles typhlops) and Galatheidae (e.g.Munida intermedia).The only brachyuran species reaching significant abundance on the upper slope is M. tuberculatus.
The transition between upper-slope fauna and a strictly bathyal fauna appeared to be located at about 450 m.The upper slope assemblage can be considered as transitional towards the bathyal environment where a well defined demersal community exists.The upper-slope was dominated by the pink shrimp P. longirostris and the Norway lobster N. norvegicus together with three species of pandalid shrimp: Plesionika heterocarpus, P. antigai and P. gigliolii.The middle slope is characterized by Aristaeidae and by the substitution within pandalid shrimps by the two deepest, P. martia and P. acanthonotus (Carbonell and Abelló 1998;Fanelli et al., 2004).The observed differences in the distribution of decapod crustaceans on the slope seem to be strongly linked to the depth gradient which may affect changes in other environmental factors such as sed-iments, benthic communities, light intensity, etc. (Gaertner et al., 1999).Maynou et al. (1996) showed that depth-related zonation of decapod crustaceans is complemented by a significant horizontal component that reflects a strong two-dimensional spatial heterogeneity.Abundant species like Pandalidae and Aristeidae shrimps and N. norvegicus (Abelló et al., 1988norvegicus (Abelló et al., , 2002;;Cartes and Sardà 1993;Maynou and Sardà 1997;Maynou et al., 1998;Fanelli et al., 2004) have a patchy distribution related to geomorphology (e.g.canyons), sediment re-suspensions and other factors that may increase local production and food availability (Cartes and Sardà, 1993;Cartes, 1994;Maynou et al., 1996).
Available data are not sufficient to show the effect of other oceanographic conditions on the distribution pattern of decapod crustacean species on the slope.Nevertheless, Fanelli and Cartes (2004) in the Alboran Sea observed a strong influence of some environmental variables (primary production 4 and 6 weeks before the sampling, hour of sampling in addition to depth) on some biological aspects and the pattern distribution of abundance of two dominant pandalid shrimps, P. martia and P. heterocarpus.
In the Mediterranean Sea the temperature does not change significantly below 150 m depth and therefore it may affect species distribution only on the shelf.In this region Gaertner (2000) observed a high persistence of fish assemblages during the year despite the seasonal changes in temperature, thus excluding any effect of this factor on the species distribution.Salinity in the western Mediterranean Sea shows changes due to the occurrence of a layer of Levantine Intermediate Water along the slope between 250 and 700 m depth (Serravall and Cristofalo, 1999).The effect of these depth-related differences in salinity on species distribution is currently unknown even though Abelló et al. (1988) hypothesized that salinity can play a structuring role for crustacean decapod communities.The differences in hydrological conditions between the shelf (high variability) and the slope (stability) can be interpreted as being responsible for the strong differences in the composition and structure of the demersal fauna in these two regions.Deep-water species are competitive in oligotrophic waters and adapted to restricted ranges of environmental parameters (temperature, salinity, light) (Ghidalia and Burgois, 1961;Maynou and Cartes, 2000), and may not be able to survive in a more heterogeneous environment such as the shelf.In the same way, shelf species are confined to this region as they are not able to compete and adapt to the stable environmental conditions of the slope.
Crustacean diversity appeared to be affected by the depth increasing significantly from the continental shelf to the middle slope.The lowest diversity and species richness values were observed on the shelfbreak where only two species of euribathic decapods, M. tuberculatus and P. longirostris, dominated the assemblage.This decrease in diversity on the shelfbreak can be attributed to the disappearance of typical shelf species, firstly brachyuran and hermit crabs, which are not replaced by slope species.
The greater environmental stability of the slope zone allows a more mature and thus more diverse community to develop (Sanders, 1968;Sanders and Hessler, 1969).The amplitude of seasonal temperature fluctuations decreases with depth.In addition, the occurrence of steep bottoms in this area favours the co-existence of pelagic (e.g.Pasiphaea spp.) and nektobenthic species (Aristaeidae and Pandalidae), which are relatively independent of benthic resources, with the strictly benthic species, resulting in increased diversity.
This study provides useful information about the composition and distribution of decapod fauna in an oligotrophic and over-exploited area.Further investigations should be devoted to analyzing species relationships and resource partitioning between species to better understand the dynamic of crustacean assemblages exposed to fishing exploitation.

FIG. 1
FIG. 1. -Map of the study area and positions of trawl stations in the MedITS project off the central-western coasts of Italy (western Mediterranean ) during 1996-2000.
FIG. 2. -Dendrograms of trawl stations using group-average clustering from Bray-Curtis similarity on average abundance of decapod crustaceans obtained during 1996-2000 surveys off the centralwestern coasts of Italy.Mean depth of each sample is given.The upper line indicates groups at the 40% level of similarity.

TABLE 2 .
-Analysis of similarities (ANOSIM) performed on the groups identified by cluster and nMDS analysis.Sample statistic (Global R): 0.826; Significance level of sample statistic: 0.1%; Number of permutations: 999 (Random sample from a large number); Number of permuted statistics greater than or equal to Global R: 0.

TABLE 5 .
-Average number of species (S), Species richness (D), Species diversity by Hill's index (N1) for each of the five assemblages of decapod crustaceans distinguished by cluster analysis off the central-western coasts of Italy.S.D.: standard deviation