Biodiversity of deep-sea demersal megafauna in western and central Mediterranean basins

Abundance, biomass and diversity patterns of bathyal and abyssal Mediterranean megafauna (fishes and invertebrates) were analyzed in the western Balearic Sea, the western Ionian Sea and the eastern Ionian Sea. Sampling was conducted with a Otter-trawl Maireta System (OTMS) at depths ranging from 600 to 4000 m. A series of ecological indicators were computed: total abundance and biomass, Margalef species richness, Shannon-Wiener diversity and Pielou’s index of evenness. A multidimensional scaling was applied, indicating that the megafauna communities were grouped by depth, while geographic area had a less defined influence. Margalef richness declined with depth in all three areas, but more steeply in the western Ionian Sea. Pielou’s evenness behaved differently in the three zones, showing a V-shaped curve in the eastern Ionian while showing a decreasing pattern in the other two areas. At lower slope depths, massive presence of the fishes Alepocephalus rostratus in the western basin and Bathypterois mediterraneus in the central basin caused a sharp reduction in evenness.


INTRODUCTION
Deep-sea ecosystems are known to be the largest repository of biodiversity of the biosphere (Sanders, 1968;Gage and Tyler, 1991), but remain mostly unknown because of extreme conditions in sampling techniques.New insights into deep-sea biodiversity are thus necessary in order to better understand the grade of stability and vulnerability of these environments (Merrett and Haedrich, 1997), but should focus on largescale analyses (Bianchi and Morri, 2000) dealing, for example, with a whole enclosed deep-sea environment.
The Mediterranean Sea is an optimal environment for studying the patterns of biodiversity for several reasons.First of all, it is an enclosed sea with limited connection with the outer Atlantic ocean.Its constant deep water temperature of between 13 and 14°C and its high coastal input related to its closeness to land (Tyler, 2003;Sardà et al., 2004a) make it a natural laboratory for studying the factors involved in determining community diversity.Fishing pressure typically reaches depths of 800-900 m in the western basin, while in the rest of the Mediterranean it barely reaches 600 m (Mytilineou et al., 2005;Sardà et al., 2009).In addition, a trawling ban is in effect in the whole basin below 1000 m depth-a unique situation at present for an enclosed sea (EC Regulation 1967/2006), allowing us to study quasi-pristine environments.Below these depths direct human impact is reduced to a minimum and is usually caused by litter discharge (Galil et al., 1995), whose impact has not yet been clarified.
Most bathyal and abyssal soft bottoms-with the exception of chemosynthetic habitats such as hydrothermal vents and cold seeps-are driven by organic matter input from the upper strata (Smith et al., 2008;Corliss et al., 2009;Lampadariou et al., 2009).The case of the Mediterranean is unique, because it is a semi-enclosed sea that has a gradient of decreasing availability of organic matter from west to east (Bethoux et al., 1999;Danovaro et al., 1999).The Levantine deep basin, known to be generally highly oligotrophic (Azov, 1991;Psarra et al., 2000;Danovaro et al., 2001), also shows a clear scarcity of all size classes of benthic fauna (Danovaro et al., 1999;Basso et al., 2004;Lampadariou et al., 2009;Danovaro et al., 2010).
In the western Mediterranean basin, a series of studies have been conducted regarding the deep-sea megafaunal assemblages down to 2800 m depth (Abelló et al., 1988;Cartes and Sardà, 1993;Stefanescu et al., 1993;Sardà et al., 1994;Moranta et al., 1998;Carrassón and Cartes, 2002;D'Onghia et al., 2004b), but addressing only fishes and crustaceans.Few studies have dealt with non-crustacean invertebrates (Ramírez-Llodra et al., 2008;Ramírez-Llodra et al., 2010).Regarding the central basin, information about lower slope megafauna is scarce as studies conducted until now have focused on the upper and middle slope communities, and addressed only particular assemblages such as fishes (D'Onghia et al., 1998;Mytilineou et al., 2005) and crustaceans (Company et al., 2004;Politou et al., 2005).Concerning the whole Mediterranean Sea, studies conducted to date have shown the existence of a depth zonation of megafauna on the continental slope (Pérès, 1985;D'Onghia et al., 1998;Moranta et al., 1998;D'Onghia et al., 2004b), with peaks of biomass and abundance around 1200 m depth and a subsequent fall in ecological indicators below 1500 m depth (for a review, see Sardà et al., 2009).Another branch of studies, particularly carried out in the western Mediterranean, have dealt with the diet (Carrassón et al., 1997;Carrassón and Matallanas, 1998, 2001, 2002) and biology (Stefanescu et al., 1992;Massutí et al., 1996;Morales-Nin et al., 1996a, b) of the main fish species of the continental slope, as well as the biology of de-capod crustaceans (Company and Sardà, 2000;Puig et al., 2001;Company et al., 2003).
During the DESEAS trans-Mediterranean cruise conducted in 2001, benthic megafauna, including fishes and invertebrates of the western Balearic, western Ionian and eastern Ionian basins, was sampled (Sardà et al., 2004b).A series of community structure and distribution studies were conducted and published on fishes and crustaceans (Company et al., 2004;D'Onghia et al., 2004b;Sion et al., 2004), but the less abundant non-crustacean invertebrate species were not included in any of the subsequent analyses and detailed biodiversity studies of the whole megafauna community were not carried out.The objective of the present paper is to describe biodiversity patterns of whole benthic megafauna on Mediterranean continental slopes and deep basins, discussing the results in the light of recent advances in deep Mediterranean trophic characteristics.

MATERIALS AND METHODS
Megafaunal samples were collected during a trans-Mediterranean sampling cruise carried out in June 2001, in the framework of the DESEAS project (Sardà et al., 2004b), aboard the R/V García del Cid (Fig. 1).A Otter-trawl Maireta system (OTMS) was used as sampling gear (Sardà et al., 1998).A total of 26 trawls were conducted at depths ranging from 600 to 4000 m (Table 1): seven in the western Balearic basin, eight in the western Ionian Sea, nine in the eastern Ionian Sea, and two in the central abyssal plain.The OTMS is a benthic trawling net fitted with two divergent doors and a single warp cable.The total net length is 25 m, with a cod-end mesh size of 40 mm and an outer cover of 12 mm.Bridles of 67 m length were used.A SCAN-MAR system was used to estimate the mean net opening geometry during the trawl, showing a horizontal opening of 12.7±1.4m.As the SCANMAR system can only operate down to 1200 m depth, the same value was used for deeper trawls, considering that no technical changes were made to the sampler during the cruise.Vertical opening was estimated at 1.4 m, with substantially less variability due to the net building characteristics (Sardà et al., 1998).
Every megafaunal individual was identified to species level, counted and wet-weighted (±1 g).A faunal list was compiled, along with the total catch abundance of each species by sampling zone and the depth range of occurrence.The SCANMAR data were used to standardize abundance (number/km 2 ) and biomass (kg/ km 2 ) data between trawls.Species richness, Shannon-Wiener H', Margalef's richness index and Pielou's evenness index were calculated for each trawl (Shannon, 1948;Margalef, 1958;Pielou, 1966;Magurran, 2004).When replicate samples per depth were available, the mean Shannon index values were calculated.Similarity of community structure among samples from all geographic areas and depths was visualized with a non-metric multidimensional scaling (MDS), after abundance data had been log(X+1) transformed.This analysis is an iterative procedure that minimizes the difference between ranked Bray-Curtis similarity values and 2-dimensional distance between each pair of samples.MDS analysis was reiterated 1000 times with a minimum stress value of 0.01, and was applied to all samples from all depths and geographic areas.
To assess what species mostly caused dissimilarities between trawls, the two-way similarity percentages (SIMPER) non-parametric statistical routine was conducted, using "depth" and "area" as crossed factors and the Bray-Curtis measure as similarity (Clarke and Warwick, 2001).All diversity analyses were conducted using the PRIMER-E 6 software package.
Multidimensional scaling analysis calculated a stress value of 0.1, which indicates an acceptable grade of ordination with little risk of misinterpretation (Clarke and Warwick, 2001).The plot suggests a   bathymetric pattern of species distribution from 600 to 4000 m depth, as shown in Figure 2. Across the whole dataset, Margalef's richness index (Fig. 3) shows a significant decrease with depth (R 2 = 0.746, p<0.001, n=26).Non-linear regression found a negative exponential relation: where SR is Margalef's richness index and D is depth (m).
Negative linear correlations between species richness and depth were found for each of the three zones considered separately (Fig. 4).The slopes of the regression lines were higher for the two Ionian sea zones than for the Balearic basin, indicating a steeper decrease in species diversity with depth in the central basin than in the western zone.Margalef index ranged from a maximum of 3.66 at 600 m depth in the eastern Ionian to a minimum of 0.92 at 2600 m depth in the same zone.
Values of Pielou's evenness index are plotted against depth for the three sampling zones in Figure 4.In the western Balearic basin, values of Pielou's index for the deepest depths were higher than the shallowest ones, with an outlier at 1230 m, although no significant trend was observed.A similar pattern was present in the western Ionian, with a decrease in the index at 1500 and 1700 m depth.In the eastern Ionian Sea, the even-  ness index showed an inverted V-shaped tendency.Shannon's diversity index for the three zones is plotted against depth in Figure 4.In the western Balearic basin, the Shannon index is highest (2.64) at 800 m depth and then shows a significant fall, with the lowest value (1.67) at 1230 m.From 1500 m the diversity increases to relatively constant values of between 2.23 and 2.30 down to the deepest depths.In the western Ionian basin, the Shannon index shows its highest values between 600 and 1200 m (2.62-2.72),falling sharply afterwards.Its lowest value (1.36) stands at 1700 m depth, and below this point it does not increase above 1.8.The eastern Ionian basin shows a pattern with two peaks, one at 800 m (2.31) and one at 1300 m (2.60).Below 1300 m the index steadily decreases down to 1.07 at 2800 m depth, which is the lowest value found in our dataset.Total catch biomass in the three zones was higher at the shallower depths (600-1230 m) than at the deeper depths (Fig. 5).In the Balearic basin the biomass at 800-1000 m depth was dominated by the fishes Alepocephalus rostratus, Mora moro and the shark Galeus melastomus; at 1200 m a peak was present, mainly caused by A. rostratus, and at depths of 2500-2800 m the shark Centroscymnus coelolepis dominated the biomass.The western Ionian sea showed a single peak of biomass at 1200 m due to a high retrieval of the fish M. moro and the sharks G. melastomus and Dalatias licha.Two peaks were present at 600 and 1300 m in the eastern Ionian, the shallower one caused mainly by the fish Helicolenus dactylopterus and the crustacean Aristaeomorpha foliacea, while the deeper one was produced by a concentration of M. moro.At the two abyssal stations, biomass was comparable with that at shallower depths (1700 and 2000 m) in the same western Ionian zone.
According to SIMPER analysis, dissimilarity in community composition between the western Balearic basin, the western Ionian and the eastern Ionian across all depth ranges is mainly caused by the fishes A. rostratus and Coryphaenoides mediterraneus, the shark C. coelolepis and the decapods Aristeus antennatus, Polycheles typhlops and Geryon longipes (Table 3).

DISCUSSION
The paucity of robust biodiversity analyses in deep sea ecosystems is mainly caused by sampling difficulties that often result in low replicate numbers.The DESEAS sampling project provided a wide spectrum of novel data about deep-sea megafauna, making it possible to conduct new biodiversity analyses of the bathyal and abyssal megafauna communities as a whole along a longitudinal and bathymetric gradient (Sardà et al., 2004b).
Few non-crustacean invertebrates were retrieved, with molluscan cephalopods and gastropods making up the most part.The number of species in this group was considerably lower than that of deep-sea faunistic studies conducted in the northeastern Atlantic (Billett et al., 2001), where the dominance of non-crustacean invertebrates is well documented (Gage and Tyler, 1991).Integrating data with other types of gear that stay more in contact with the substrate (e.g. an Agassiz dredge) may help to clarify this point.Results from the multidimensional scaling routine showed a clear segregation by depth of benthic megafauna in the Mediterranean.The trawls at 600-800 m grouped together, confirming the middle slope subdivision (Pérès, 1985).Differences in this depth range between the western Balearic basin and the two study sites of the central basin were more marked than at the other depths.The depth range 1000-1700 m showed the highest dispersion between trawls, as was previously reported for fishes (Stefanescu et al., 1993;D'Onghia et al., 2004b) and crustacean assemblages (Maynou and Cartes, 2000;Company et al., 2004) in the Mediterranean.The present study included non-crustacean invertebrates in the analyses and this did not change the overall community structure pattern, suggesting that the megafaunal invertebrates follow the same pattern of depth zonation as the other taxa.The deepest assemblages at 1700-3000 m, along with the two trawls conducted at 3300 and 4000 m in the central Ionian abyssal plain, grouped together, indicating a reduction in depth influence below 1500 m.There was no clear pattern in the MDS plot that could be interpreted as an influence of longitude, except at middle slope depths where samples showed a slight grouping by geographic area.This suggests that depth is a stronger driver than geographic area for community composition, as was previously pointed out for the eastern Mediterranean basin megafauna (Kallianiotis et al., 2000) and meiofauna (Lampadariou et al., 2009).Influence of depth alone as a physical factor in determining assemblage composition is currently a matter of discussion, and some authors propose that megafaunal zonation is mainly controlled by food availability (Jones et al., 2003;Tyler, 2003;Soltwedel et al., 2009).Coupling biodiversity studies with trophic analyses of key species in each basin may shed light on this result.As reported by Politou et al. (2004), temperature showed a slight increase in a westeast axis (12.8-13.9 in the western basin, 13.3-14.2 in the central Mediterranean), while salinity showed no differences between zones.It is generally known that the hydrology of the deep Mediterranean is relatively stable and dissolved oxygen is also believed not to be limiting, with values constantly above 4 mlO 2 /l (Miller et al., 1970).Therefore, relating faunistic assemblages with available environmental data is currently difficult and further detailed abiotic studies are necessary.
A difference in equation fits for Margalef species richness was found between the whole dataset and the three zones considered separately.These differences could be caused by a relatively low number of samples.Further analyses with a larger dataset may bring to light patterns different from linear relations.Although Margalef index is more dependent than other diversity measures on sampling effort (Magurran, 2004), which in our case was higher at shallower stations, a negative linear correlation was found in all three basins, confirming the expected general reduction in megafaunal richness with depth (Gage and Tyler, 1991).The three zones exhibited a comparable species richness up to 1500 m depth, but after this limit only the western Mediterranean maintained a Margalef index of above 2.0, while in the Ionian basins (both western and eastern) values decreased rapidly.The Mediterranean sea is known to show a west-east gradient of decreasing productivity (Danovaro et al., 1999;Tselepides et al., 2004).It is therefore possible that higher surface productivity and coastal input of the western Balearic basin supply the necessary energy to sustain a diverse community down to deeper depths.Furthermore, the western Mediterranean is exposed to cyclic downslope dense shelf-water cascading events, which provide massive inputs of organic matter to bathyal and abyssal depths (Canals et al., 2006;Company et al., 2008).In the more oligotrophic Ionian Sea, organic matter reaching the lower slope and the deep basin is reduced, so there is a decrease in diversity early in the depth range.
Pielou's index of evenness expresses the degree of dissimilarity of the observed sample with a hypothetical situation of all equally abundant species.In the western Balearic Pielou's index showed a dramatic reduction at 1230 m depth, caused by the peak of abundance of the fish Alepocephalus rostratus.This dominance has been already pointed out in several studies (Stefanescu et al., 1993;Morales-Nin et al., 1996b;Sardà et al., 2009), and may be caused by specific trophic conditions (Carrassón and Matallanas, 1998) and by the occurrence of shelf-water cascading events that bring enriched water at these depths (Canals et al., 2006).A similar effect is present in the western Ionian at lower slope depths, and is caused by the spiderfish Bathypterois mediterraneus.The biology of this species is more adapted to the scarce trophic resources (Carrassón and Matallanas, 2001;D'Onghia et al., 2004a), although the exact reasons for this dominance of B. mediterraneus are, at present, not described.
In the two samples collected from abyssal depths, few species were collected: 11 at 3300 m and 8 at 4000 m.In both cases, the catch composition was dominated by the macrourid Coryphaenoides mediterraneus and by the deep water shrimps Acantephyra eximia and Nematocarcinus exilis.These occurrences have already been described as the deepest records for these species (Company et al., 2004;D'Onghia et al., 2004b).Species evenness at these two depths was low compared to the rest of the western Ionian basin.Food scarcity has been shown to limit the presence of large-sized species, favouring settlement of small-sized species that can crop more efficiently on a broader range of preys (sensu Dayton and Hessler, 1972).C. mediterraneus and N. exilis had lower mean sizes than other species of macrourid fishes and decapod crustaceans that inhabit the deep Mediterranean bottoms, suggesting that these species may be the only ones that can efficiently inhabit these low-energy environments.
Regarding biomass, in the western Balearic basin the pattern obtained is in accordance with that of previous studies conducted in this basin (Moranta et al., 1998;Massutí et al., 2004).In the western Ionian sea, biomass values remain low down to the abyssal plain at 4000 m depth.Biomass flow in the eastern Ionian sea essentially follows the same pattern, with a biomass fall starting from 1700 m, showing that, even with different diversity patterns, the three basins exhibit a consistent biomass decrease starting from lower slope depths.
According to SIMPER analysis, dissimilarity in community composition between the Balearic basin and the Ionian Sea (both western and eastern zones) is mainly caused by a few species of Actinopterygii and decapod crustaceans.A. rostratus has its distribution limited to the western Mediterranean basin (Morales-Nin et al., 1996b;D'Onghia et al., 2004b) but the reasons for this spatial segregation are yet to be identified.The brachyuran crab Geryon longipes were found in the western Ionian and in the eastern Ionian in relatively low abundances.The trophic strategy of this species is less adapted to the scarcity of energy resources in the eastern basin (Kitsos et al., 2005), where another crab of the family Geryonidae, Chaceon mediterraneus, is beginning to be present.C. mediterraneus was also found in the western Balearic basin at depths of 2500 and 2800 m.No occurrences of these two species at the same time were recorded, in agreement with another study which addressed the eastern basin megafauna (Jones et al., 2003).
The changes in megafaunal biodiversity across a geographical axis in the Mediterranean Sea have been partly addressed in this work, which has evaluated patterns of species substitution and biomass decline with depth and longitude.At depths of 1200-1500 m an interruption in the expected pattern of biodiversity across depth ranges takes place, with two distinct actinopterygian species-A.rostratus in the western basin and B. mediterraneus in the Ionian sea-dominating the abundance and biomass.This is an important matter for further exploration, with the hope of possibly extending the study to the whole longitudinal axis of the Mediterranean basin including the easternmost areas, where deep-sea megafauna sampling is at present very scarce.In addition, focused analyses to relate biodiversity to food availability and ecosystem functioning will provide essential information to better understand the observed community structure and diversity patterns in deep Mediterranean megafauna.

Fig. 3 .
Fig. 3. -Margalef species richness index for the whole studied area (western and central Mediterranean basins).

Fig. 4 .
Fig. 4. -Margalef species richness, Pielou evenness and Shannon-Wiener diversity indices for the western Balearic Sea, the western Ionian sea (including the two central abyssal plain samples), and the eastern Ionian Sea.Linear regression lines are shown only when significant (p<0.05).

Fig. 5 .
Fig. 5. -Standardized values of total catch biomass for the western Balearic Sea, the western Ionian sea (including the two central abyssal plain samples), and the eastern Ionian Sea.

Table 1 .
-List of analyzed trawls with time and spatial characteristics.

Table 2 .
-Species list of demersal megafauna sampled in the DESEAS cruise, along with their total collected abundance in each of the four samples areas, and their depth range of catch.

Table 2 (
cont.).-Species list of demersal megafauna sampled in the DESEAS cruise, along with their total collected abundance in each of the four samples areas, and their depth range of catch.

Table 3 .
-Dissimilarities between the three geographic zones as calculated by the SIMPER statistical routine.