Deep-water macroalgal-dominated coastal detritic assemblages on the continental shelf off Mallorca and Menorca (Balearic Islands, Western Mediterranean)

Abstract We present a quantitative physiognomic characterization of major macroalgal-dominated assemblages on coastal detritic bottoms of the continental shelf off Mallorca and Menorca (Balearic Islands, Western Mediterranean). In late spring of 2007 and 2008, 29 samples were collected by bottom trawling at depths between -52 and -93 m. These samples were then sorted and identified to their lowest taxonomic level. Statistical analyses distinguished six different assemblage types: shallower water environments (-52 to -65 m in depth) were characterized by Osmundaria volubilis and Phyllophora crispa meadows and two types of Peyssonnelia beds; two assemblage types, Laminaria rodriguezii beds and maërl beds, were only present in deep-water environments (-77 to -81 m); and an assemblage dominated by P. crispa and Halopteris filicina was found in both shallow and deep waters (-57 to -93 m). We assess the distribution of these six assemblage types through the studied area.


Introduction
Coastal detritic bottoms are characterized by a large amount of particles of organic origin, a low percentage of silt, and, typically, by the absence of muddy particles ( P é r è s 1985 ). They constitute a main habitat within the sedimentary bottoms of continental shelves occurring close to shore, usually at more than -25 m in depth ( P é r è s 1985 ). Assemblages developing in these coastal detritic bottoms encompass a large range of species and functional diversities and also harbor a vast number of rare and interesting species that are often restricted to these kinds of assemblages (Cabioch 1969, Ballantine et al. 1994, Bellan -Santini et al. 1994, Grall and Gl é marec 1997, Foster 2001, Steller et al. 2003 ). In addition, coastal detritic assemblages act as nursery grounds for various invertebrates and fishes, including commercial species (Massut í et al. 1996, Colloca et al. 2003, Kamenos et al. 2004 , Massut í and Re ñ ones 2005 ), and they shelter many calcareous algae and calcareous invertebrates, which indicates that these assemblages are major carbonate producers (Ballesteros 1994 , Canals andBallesteros 1997 ).
The Mediterranean Sea has a long tradition in the study of coastal detritic assemblages. Research on the flora began at the end of the 19th century and the beginning of the 20th century (Rodr í guez-Femen í as 1889 , Mazza 1903, de Buen 1905, Bell ó n-Uriarte 1921. Later, diverse authors highlighted the high biodiversity and reported different assemblage types such as Laminaria rodriguezii beds, ma ë rl beds, and free-living Peyssonnelia beds ( P é r è s and Picard 1963, 1964, Picard 1965, Giaccone 1973, Augier and Boudouresque 1978, Bourcier 1982, Ballesteros et al. 1993. Besides this major literature, other significant studies have focused on ma ë rl ( Dieuzede 1940, Feldmann 1943, Gautier and Picard 1957, Jacquotte 1961, 1962, Ballesteros 1988, Basso 1995a,b , Bordehore et al. 2003, Piazzi et al. 2003, Agnesi et al. 2011, L. rodriguezii beds (Feldmann 1934, Molinier 1956, 1960, and Peyssonnelia beds (Huv é 1954, Carpine 1958, Laborel et al. 1961, Ballesteros 1994 ). However, all these different assemblage types have been described separately, and although they have several species in common, no attempts have been made to find out whether they are different and can be distinguished by statistical methods.
Direct sampling (e.g., scuba diving) in the deep waters of the continental shelf, where coastal detritic bottoms develop, has severe limitations due to restricted sampling time, long decompression stops, and diver performance decrease (UNEP -MAP-RAC/SPA 2008 ). Although scuba diving has been occasionally used to describe species composition on coastal detritic bottoms (Giaccone 1972, Ballesteros 1988, 1994, Piazzi et al. 2003, indirect sampling methods (e.g., dredging, trawling, or video surveys with remotely operated vehicles) are those usually used (Cailliet et al. 1999 , Bax andWilliams 2001 ). Thus, dredges are the most common practice used in the description of algal assemblages (Basso 1995a,b , Bordehore et al. 2003, whereas trawls are probably the most frequently used method in the description of fish and invertebrate assemblages (Bertrand et al. 2002, Massut í and Re ñ ones 2005, Fanelli et al. 2007, Garc í a-Mu ñ oz et al. 2008, Ordines and Massut í 2009. The Balearic Islands have a central position in the western Mediterranean Sea and are the emergent parts of a promontory including four major islands. The Balearic shelf can be divided into two geographic sectors: the Mallorca-Menorca shelf, covering 6418 km 2 , and the smaller Eivissa-Formentera shelf, with a total surface of 2709 km 2 (Acosta et al. 2002 ). Coastal detritic bottoms have been largely identified on the Mallorca-Menorca shelf (de Buen 1934, Alonso et al. 1988, Forn ó s et al. 1988, Canals and Ballesteros 1997, Forn ó s and Ahr 1997, Massut í and Re ñ ones 2005 and have been characterized in terms of Foraminifera (Milker et al. 2009 ) and mega benthos (Massut í et al. 1996 , Massut í and Re ñ ones 2005 , Ordines and Massut í 2009 ). Seaweeds are a major component of these coastal detritic bottoms, but they have been taken into account in a somewhat cursory manner in descriptive studies that try to cover all benthic groups (e.g., de Buen 1934 , Canals andBallesteros 1997 ); the only literature providing a full list of seaweeds thriving on these bottoms is the description of Peyssonnelia and ma ë rl beds by Ballesteros (1994) .
Hence, in this study we attempted to identify the main assemblage types that can be distinguished on coastal detritic bottoms of the continental shelf off Mallorca and Menorca, according to the abundance of the different algal species. We used bottom trawling as a sampling technique. In addition, we provide geographical and bathymetric distributions of these assemblages.

Materials and methods
The present study was located on the continental shelf off Mallorca and Menorca (Balearic Islands, Western Mediterranean; Figure 1 ). This shelf is generally narrow, except at the south of Mallorca and in the channels between Mallorca and Menorca and Mallorca and Cabrera, where it becomes larger and has a gentle slope (Acosta et al. 2002 ). The absence of rivers reduces the presence of terrestrial sediments, and therefore, most of the sediments are usually of biogenic origin Ballesteros 1997 , Forn ó s andAhr 1997 ) and contain a high percentage of carbonates (Acosta et al. 2002 ). In addition, light transmittance through the water is very high Zabala 1993 , Canals andBallesteros 1997 ), allowing algaldominated benthic communities to develop deeper than on other Mediterranean continental coasts (Ballesteros and Zabala 1993 ).
Samples of coastal detritic assemblages were collected at depths ranging from -52 to -93 m ( Figure 1) in the late spring (May/June) of 2007 and 2008 during the MEDITS_ES05 bottom trawl surveys. Samples were collected with experimental GOC73 equipment and followed the general specifications for the MEDITS surveys (Dremiè re et al. 1999, Bertrand et al. 2002. A total of 29 samples were collected; each sample consisted on a haul obtained during 30 min at a vessel speed of 3 knots. When the hauls arrived on board, algae were sorted and a 6 l sample was obtained from every haul and preserved in 4 % formalin in seawater. Once in the laboratory, samples were washed, sorted, and identified to the lowest taxonomic level. For each sample (6 l), the abundance of each taxon was measured as algal surface area (Sa i ) in square centimeters (Ballesteros 1992a ). Subsequently, these data were standardized as algal surface area per haul (Sa iH ), expressed in square centimeters per square meters , as follows: where Sa i is the algal surface area of each species (cm 2 ), WW H is the wet weight of algae for each haul (g), S H is the total haul surface (m 2 ), and WW s is the wet weight of the algae sampled (g).
To visualize the affinities between samples, we used a nonmetric multidimensional scaling (nMDS) ordination (Kruskal and Wish 1978 ) based on a Bray-Curtis similarity matrix calculated from Sa iH data. A hierarchical group average agglomerative clustering method accompanied by the SIMPROF test (Clarke et al. 2008 ) adjusted to 9999 permutations and a 0.1 % significance level according to Potter et al. (2001) was used to explore the potential grouping structures among samples. Finally, the percentage similarity routine (SIMPER) was run to quantify the contribution of each species to the similarity/dissimilarity between the varying SIMPROF groups. All the analyses were performed with PRIMER version 6 software ( PRIMER-E Ltd., Plymouth, UK; Clarke and Warwick 2001 ).
The nMDS plot based on species abundances shows patterns of resemblance among the species composition of different trawls (nMDS stress 0.1; Figure 2 ). According to the SIMPROF test, the samples may be classified into six different groups assigned to six different assemblages. They are O. volubilis and P. crispa meadows (Ov/Pc_m), two different kinds of Peyssonnelia beds, one dominated by Peyssonnelia inamoena (Pi_b) and another by Peyssonnelia rubra (Pr_b), L. rodriguezii beds (Lr_b), ma ë rl beds (M_b), and P. crispa and H. filicina meadows (Pc/Hf_m) (Table 1). Looking at the species composition in the six assemblages, some of them had exclusive species, whereas some other species were found throughout the continental shelf off Mallorca and Menorca: Flabellia petiolata , Halymeniaceae unidentified 1, Lithothamnion valens , Peyssonnelia rubra , Peyssonnelia squamaria , P. crispa , and Spongites fruticulosa . This list of common species increased up to 20 when we considered those present in all the assemblages other than ma ë rl beds where erect species were scarce. The assemblage with the maximum number of exclusive species was Pc/Hf_m, with 24 exclusive species, followed by Ov/Pc_m with 7 exclusive species, Pr_b with 5 exclusive species, and Pi_b with 3 exclusive species. The remaining assemblages did not have any exclusive species (Table 2).
The assemblages Ov/Pc_m, Pi_b, and Pr_b were located in shallow waters (-52 to -65 m) and characterized by a great abundance of O. volubilis , P. crispa , and Peyssonnelia spp., whereas they differed from each other not only in the relative abundance of these species but also in the abundances Figure 2 nMDS ordination with standardized Sa iH (cm 2 m -2 ). Samples are displayed indicating their code, depths range, and corresponding assemblage type. Code shows the year of sampling (7, 2007; 8, 2008) followed by the number of the sample. Assemblage types: Lr_b, Laminaria rodriguezii beds; M_b, ma ë rl beds; Ov/Pc_m, Osmundaria volubilis and Phyllophora crispa meadows; Pc/Hf_m, Phyllophora crispa and Halopteris filicina meadows; Pi_b, Peyssonnelia inamoena beds; Pr_b, Peyssonnelia rubra beds.

Brought to you by | CSIC -Unidad Coordinacion Bibliotecas
Authenticated | 161.111.254.116 Download Date | 10/17/12 10:52 AM of accompanying species. Hence, Ov/Pc_m was formed by 5 samples located at depths between -52 and -60 m, in which both O. volubilis and P. crispa were identified as the main species according to the SIMPER test. These meadows were characterized by a high number of species (47 ± 6 per haul) and an Sa TH of 30,945 ± 15,358 cm 2 m -2 per haul. Pi_b and Pr_b were both Peyssonnelia beds, but they had very different species composition and Sa TH values. Thus, Pi_b, formed by only 2 samples located at -62 m depth, was characterized by the species Peyssonnelia inamoena , A. preissii , and Peyssonnelia rosa-marina , and had a low number of species (34 ± 11) but a high Sa TH (620,000 ± 495,000 cm 2 m -2 ). In contrast, the 3 samples of Pr_b, located between -60 and -65 m, were characterized by P. crispa , O. volubilis , and Peyssonnelia rubra and had a greater number of species (mean = 55 ± 7) and lower Sa TH (138,451 ± 31,210 cm 2 m -2 ) than Pi_b.
In deeper waters (from -77 to -81 m), two groups were identified, M_b and Lr_b. The M_b included only 1 sample dominated by the corallines Spongites fruticulosa and Lit. valens and had a very low number of species (10) and Sa TH (306 cm 2 m -2 ) in comparison with the rest of the samples. In contrast, Lr_b, comprising 3 samples, was dominated by L. rodriguezii and had a higher number of species (29 ± 10) and Sa TH (17,926 ± 5005 cm 2 m -2 ).
Finally, Pc/Hf_m included 15 samples collected in a large range of depths (from -57 to -93 m). Phyllophora crispa , H. filicina , O. volubilis , and Spongites fruticulosa were the most abundant species in this group, with 38 ± 9 species and an Sa TH of 3511 ± 1963 cm 2 m -2 .

Discussion and conclusions
Six different coastal detritic bottom assemblages were distinguished on the continental shelf off Mallorca and Menorca based on haul sampling. In general, depth was an important correlate of the distribution of the different assemblages, with the exception of Pc/Hf_m, which occurred in a wide range of depths. Moreover, the limit that separated the relatively shallow from the deep assemblages was situated at around -70 m in depth. Although the decrease in light availability may certainly play an important role in the segregation of these assemblages (Ballesteros and Zabala 1993 ), we argue here that hydrographic characteristics at the different depths might also contribute to this segregation. The deep waters (from -70 to -200 m) on the Balearic shelf have characteristics of the Intermediate Western Mediterranean Waters whose temperatures are always between 12.5 ° C and 13 ° C and have salinities at around 38.15. This is in contrast with shallower waters, which are influenced by both the North Atlantic Surface Waters and the Gulf of Lion Cold Waters and have salinities usually < 38 and seasonal temperatures ranging from 14 ° C to 27 ° C (Salat andFont 1985 , Vives andL ó pez-Jurado 1988 ). The minimal oscillation of temperature in the deeper waters may allow the development of stenothermal species whose growth would be limited in the shallow waters.
The Osmundaria volubilis and Phyllophora crispa meadows (Ov/Pc_m) were found off southeastern Mallorca and western Menorca and were closely related to the assemblage of O. volubilis from the coastal detritic bottoms described previously by P é r è s and Picard (1963) at Port-Cros (France). However, these authors reported that O. volubilis is usually associated with Rytiphlaea tinctoria rather than P. crispa . Even though O. volubilis is very common in these and other coastal detritic assemblages, it is not a species that can be considered as exclusively associated to these kinds of assemblages, as it is also very abundant on some types of coralligenous outcrops (Ballesteros 1992b ) and on infralittoral rocky bottoms (Boudouresque 1973, Augier and Boudouresque 1975, Serio and Pizzuto 1990, 1992, Ballesteros et al. 1993. Detritic coastal bottoms dominated by different species of Peyssonnelia have been previously reported on several Mediterranean coasts (see Ballesteros 1994 for a review). P é r è s and Picard (1963Picard ( , 1964 reported Peyssonnelia beds in Port-Cros (France) that were dominated by Peyssonnelia rosa-marina f. saxicola together with a smaller quantity of Peyssonnelia harveyana . Later, Augier and Boudouresque (1978) reported coastal detritic assemblages dominated by Peyssonnelia rosa-marina f. rosa-marina and Peyssonnelia rubra . Off the Balearic Islands, Peyssonnelia beds have been reported by de Buen (1905) off the northeastern coast of Menorca, Cape Formentor (Mallorca), and Cabrera at depths between -60 and -160 m and by Ballesteros (1994)

Kelp beds dominated by the endemic brown alga
Laminaria rodriguezii had a low number of species per sample probably because the dense canopy of this kelp prevents the growth of other algae. In fact, Picard (1965) already highlighted that these kelp beds are an impoverished facies on coastal detritic bottoms. Other species that can occasionally be found in these kelp beds are Cystoseira spinosa var . compressa and Cystoseira zosteroides. These two species and Phyllariopsis brevipes are usually considered to be common in L. rodriguezii beds (Giaccone 1967 , UNEP/UICN/GIS Posidonie 1990 ). The low abundance of these accompanying species in our samples may be related to the frequent trawling in the area sampled. In fact, recent collections (2011) from better preserved detritic bottoms in Mallorca and Menorca showed that C. spinosa var . compressa and C. zosteroides can be also extremely abundant in the L. rodriguezii beds off the Balearic Islands (C. Rodr í guez-Prieto and S. Joher personal observations). The development of Lr_b seems to be restricted to particular places, such as the Menorca Channel. Hence, its growth is probably determined by the presence of a detritic bottom composed of rhodoliths, dim light conditions, low and constant water temperature (about 14 ° C), and unidirectional and constant currents (Molinier 1960 , P é r è s and Picard 1964, Giaccone 1967, 1971. In fact, the lower limit of L. rodriguezii distribution seems to depend on light availability, whereas the upper limit probably depends on temperature (UNEP/UICN/GIS Posidonie 1990 ).
The ma ë rl bed (M_b) sampled in this study on the eastern coast of Menorca was mainly characterized by a great abundance of Spongites fruticulosa and Lithothamnion valens . However, this assemblage did not have the usual diversity found on these kinds of detritic bottoms off the Balearic Islands, which may be also dominated by other species such as Lit. corallioides and Phymatolithon calcareum or even Lithophyllum racemus (de Buen 1905, Ballesteros et al. 1993, Ballesteros 1994 , E. Ballesteros personal observations). In contrast to other ma ë rl beds from the southwestern Mediterranean (Bordehore et al. 2003, Piazzi et al. 2003, the M_b found in our study was very poor in erect algae. The development of ma ë rl beds in the Mediterranean seems to depend on the existence of high to moderate unidirectional water currents (Picard 1965 ), and most of them correspond to the facies of Spongites fruticulosa described by P é r è s and Picard (1964) .
The eurybathic meadow (Pc/Hf_m) mainly differed from the other assemblages by a lower abundance of O. volubilis . Due to the fact that this meadow does not seem to correspond to any other known assemblage, we cannot rule out the possibility that it corresponds to a methodological artifact due to a mixed sampling in relatively small and patchy heterogeneous areas. Moreover, this bottom may correspond to an algal drift accumulation zone.
Among the five introduced species found in the area, Botryocladia madagascariensis and Acrothamnion preissii were the most widely distributed, whereas Caulerpa racemosa var . cylindracea , Asparagopsis taxifomis , and Lophocladia lallemandii were found only occasionally and they were never abundant. Only Caulerpa racemosa var. cylindracea has been reported to behave as an invasive alien in coastal detritic assemblages of the northwestern Mediterranean (Klein and Verlaque 2009 ).
In summary, using algal composition and abundances in the samples collected, bottom trawling proved to be a useful method to characterize and describe assemblages of the continental shelf off Mallorca and Menorca. However, doubts arose for some trawls (e.g., those identified as the eurybathic Pc/Hf_m) because we did not know whether sampling had been performed on highly heterogenous bottoms or if the unique trawl content may represent a mixture of two or more different assemblages. These doubts could also arise when identifying the assemblages according to faunal composition. Dredging, a more frequently used method in the characterization of these algal assemblages, is probably not as useful in describing species groupings, as it does not take into account high spatial heterogeneity. In fact, most of the assemblages found in this study had already been described by other authors using different methodologies (e.g., P é r è s and Picard 1964, Ballesteros et al. 1993, Ballesteros 1994, thus confirming that characterization based on algal composition and abundance determined by bottom trawls is a feasible approach for studying deep-water assemblages.