Sponge community variations within two semi-submerged caves of the Ligurian Sea (Mediterranean Sea) over a half-century time span

Abstract In the last few decades, macrobenthic community structures and their species abundances have shown significant changes in the Mediterranean Sea, whose causes were attributed to anthropogenic activities and to global warming effects. The Mediterranean sponges have shown a peculiar sensitivity to these changes: the populations of some species showed significant decreases, while others, more thermophilous, increased. Therefore, sponges may be a good proxy for evaluating the effects of environmental changes. Thanks to the observations conducted by Sarà about 55 years ago, a comparative analysis of the sponge populations present within two semi-submerged caves in the Ligurian Sea was possible. The two sponge assemblages re-studied in 2016 showed an increase in terms of specific richness and a significant change in their structural aspects, since the three-dimensional growth forms were mostly replaced by two-dimensional ones, a process observed also in other littoral communities. Consequently, the sponge communities inside the semi-submerged caves may be considered poorly resilient: the massive sponges were hit by the positive thermal anomalies occurring in the Ligurian Sea in the last decade and were replaced by encrusting forms, within a possible phase of cave recolonisation.

Semi-submerged caves are particularly suitable to verify the temporal stability of or the advent of possible changes to the benthic coastal rocky communities at a local scale, because they are easily recoverable due to their certain geographical position. Moreover, light and water movement gradients within caves cause different physical conditions along the horizontal cave axis which, in turn, promote a wide richness in small confined spaces (Dimarchopoulou et al. 2018). Furthermore, the lack of algal coverage facilitates the observation of zoobenthic organisms.
Recently, an ecosystem-based method for evaluating the ecological quality of Mediterranean caves (CavEBQI) was developed, opening new perspectives to the study of disturbance impacts at large geographic and temporal scales (Rastorgueff et al. 2015). However, these kinds of studies tend to consider and evaluate affinity, stability and changes at large systematic or functional group levels, whereas variations are often recordable at the specific level only. Bearing this in mind, the study of the sponge fauna permits us to check not only the ecological quality of littoral habitats, but also the effects of possible long-lasting environmental changes (Bertolino et al. 2014a(Bertolino et al. , 2016. Unfortunately, the lack of historical data sets and time series on local sponge biodiversity does not allow the verification of possible changes in the structure of their populations, even in the light of the global change. Fortunately, in 1961-1963, Michele Sarà (1926, a pioneer of Italian sponge taxonomy, analysed the sponge assemblages present in two small semi-submerged caves in the Ligurian Eastern Riviera (Ligurian Sea) (Sarà 1964).
The main goals of this paper were to evaluate possible qualitative changes in the sponge biodiversity of the two semi-submerged caves after a halfcentury time spana period characterised by important global and human-driven structural changes (e.g. Parravicini et al. 2013;Bertolino et al. 2016;Betti et al. 2017;Longobardi et al. 2017).

Site characteristics
For three summers (1961)(1962)(1963), Sarà (1964) surveyed and studied the sponge communities of two semi-submerged caves located in the Eastern Ligurian Riviera, called Bonassola Cave and Zoagli Cave after the names of the closest respective villages (Figure 1(a)). The distance between the two caves is about 25 km.
The Bonassola Cave is a semi-submerged, horizontal cave which opens on gabbro rocks about 1 km eastward of the homonymous village (Figure 1(b)). The bottom of the cave consists of irregular boulders and pebbles. The cavity is accessed through a 4-5 m long passage, 2-2.5 m wide and with a ceiling 2 m above sea level. This corridor opens into a large cavity (12 m × 7 m), with a height of about 6 m. Numerous crevices and small galleries are present on the side walls. The cave continues in a south-east direction with a passage, 12 m long and 2 m wide, connected to the exterior through a second, fully submerged opening. Inside the cave, sunlight penetrates mainly through a wide opening in the vault. Light conditions create a semi-obscure environment, while total darkness is reached only in parts of the walls and in the south-east corridor. Depending on light intensity and cave morphology, it is possible to distinguish three zones, according to Sarà (1964): the semi-dark walls, the dark walls and the semidark bottom of the cave.
The Zoagli Cave is a small blind-end (cul-desac) semi-submerged cave, which opens in a marly cliff about 1 km west of the Zoagli village ( Figure 1(c)). It is about 16 m long and 4 m wide at the entrance, with a height of the vault decreasing from about 3 m at the entrance to about 1 m at the end of the cave. The pebble floor steadily rises from a depth of 2 m to the end of the cave where it forms a small beach. The walls are mostly smooth. The opening is oriented to the east. According to Sarà's (1964) observations, four different zones with different light intensity can be recognised: A-zone (exposed towards the south), external and semi-dark; B-zone (towards the south), internal and dark; C-zone (towards the north), external and semi-dark; and D-zone (towards the north), internal and dark. During summer 2016, the sponge assemblages of these caves were studied again; samples were collected for taxonomic determinations and photographs were taken on standard surfaces of 400 cm 2 in the different sectors defined by Sarà (1964), according to the light intensity. In the Bonassola Cave, 10 replicates were conducted in each of eight sectors: four on the cave walls (A, B, C, F) characterised by semi-dark conditions; two on the bottom in semi-darkness (G1, G2) and two on the dark walls (D, E), for a total of 80 images (Figure 1(b)). In the Zoagli Cave, 10 replicates were conducted in each of the four sectors, two (A, C) characterised by semi-dark conditions and two (B, D) by dark conditions, for a total of 40 images (Figure 1(c)). For each cave data were elaborated by grouping the different sectors according to the light intensity conditions (Sarà 1964). For each light condition the abundance of each species was evaluated as percent presence in the studied standard surfaces. According to the growth habit, the recorded sponges were divided into encrusting (En) and massive species (Ms).
To compare the recent data with those of Sarà (1964), who used a semi-quantitative analysis, we have transformed our data (presented in Tables I  and II) as follows: +, present (< 5% of the studied images); ++ abundant (5-24% of the studied images); +++, very abundant (> 24% of the studied images).

Results
During the 2016 survey, a total of 30 species were found in both of the studied semi-submerged caves; seven of them were shared by the two caves.
The Bonassola Cave, characterised by a complex morphology, hosted 21 species (Table I; Figure 2). Sarcotragus spinulosus Schmidt, 1862 was the most abundant species in each examined sector. Other frequent species were Erylus discophorus (Schmidt, 1862), Aplysina sp., Ircinia variabilis (Schmidt, 1862), Spongia (Spongia) officinalis Linnaeus, 1759, and Rhabderemia topsenti van Soest and Hooper, 1993 recorded with different frequencies in the three sectors ( Figure 3). The richest sector was the dark wall, followed by the semi-dark wall and by the semidark bottom (Figure 4(a)). In the complex the number of encrusting species was double that of the massive ones (Figure 4(b)). The sponge assemblages present in the three zones of the cave were widely different: 24% of the recorded species were shared by all zones, while 57% were recorded in a single zone.
The Zoagli Cave assemblage included 13 demosponges and the calcarean Clathrina coriacea (Montagu, 1814) (Table II; Figure 2). The different sectors of the cave are quite different in terms of species composition: Crambe crambe (Schmidt, 1862) and Ircinia oros (Schmidt, 1864) were the most abundant species on semi-dark walls while in the dark area the most frequent sponges were C. crambe, Aplysina sp. and Timea fasciata Topsent, 1934 (Figure 3). The number of species was higher in the semi-dark zone while, in accordance with the growth patterns, the encrusting species represented the higher number (Figure 4(c,d)). In this cave, half of the observed species were recorded only in one zone.
Particularly interesting was the finding of a species of Aplysina ( Figure 5) characterised by only one kind of pithed fibres devoid of foreign bodies, forming a regular network of large polygonal meshes as usual in the genus. Nevertheless, the growth habit of our specimens differs from that shown by the two Mediterranean species of the genus: A. aerophoba (Nardo, 1833) and A. cavernicola (Vacelet, 1959). In fact, the specimens recorded in the caves are small, yellow, cushion-shaped sponges (2-5 mm in diameter). They form clusters composed of 3-10 specimens joined together by thin, sometimes branching processes (Figure 5(a)). Sometimes several small specimens fuse together, forming large encrusting plates (Figure 5(b)). The colour and the conulose surface of the specimens are the same as in the other two Mediterranean Aplysina. The species is widely distributed in the semi-dark and dark portions of both caves. Several specimens recorded in the tidal zone of the Zoagli Cave remain completely emerged during low tides ( Figure 5(c)).
The comparative analysis of our data and those reported by Sarà (1964), who visited these semisubmerged caves several times between 1961 and 1963, was surprising. Firstly, the number of species recorded by Sarà in both the caves was about half that found in our survey (Tables I and II; Figure 2). Secondly, the species composition appeared widely changed (Tables I and II (Sarà, 1958), Mycale (Aegogropila) tunicata (Schmidt, 1862), Phorbas fictitius and Spongia (Spongia) virgultosa) were considered very abundant by Sarà (1964).
Despite changes in the specific composition, the sponge distribution in the different zones of the caves remained quite similar: in general, the species richness increased in each zone of the caves (Tables I  and II; Figure 4(a-c)).
As for growth patterns, the number of massive species decreased or remained constant in both caves, whereas the encrusting ones strongly increased (Tables I and II; Figure 4(b-d)).

Discussion
In the absence of long-term series of historical data that could provide a reliable picture of the environmental changes within marine communities (Boero et al. 2015), the comparative analysis of richness and abundance of a peculiar taxon such as Porifera in two different temporal periods remains a good tool for highlighting the changes that have occurred. Sponges have been recently suggested as a surrogate taxon for the structural and functional study of sessile benthic diversity in Mediterranean marine caves (Gerovasileiou et al. 2017). They appear therefore a good proxy, being the most specific taxon inside Mediterranean hard-bottom benthic communities and in particular within marine caves, where they were widely studied (Sarà 1958(Sarà , 1959(Sarà , 1961a(Sarà ,b, 1962(Sarà , 1968Labate 1965;Pansini et al. 1977;Bibiloni et al. 1989;Corriero et al. 1997Corriero et al. , 2000Corriero et al. , 2003Harmelin et al. 2003;Bussotti et al. 2006;Gerovasileiou & Voultsiadou 2012). Species composition, abundance and distribution of sponges within a cave are driven not only by its dimensions, but also by the differences of light and water-movement gradients related to exposition and topography of the cave. In particular, semi-submerged, tunnel-shaped caves generally show richer communities due to the more intense water movement usually recorded in these spots (Riedl 1966;Balduzzi et al. 1989;Martì et al. 2004;Gerovasileiou & Voultsiadou 2016). Moreover, caves are particularly selective environments and each seems to host its own sponge assemblage (Corriero et al. 2003). Only a small number of species are, in fact, widespread within Mediterranean marine caves: Petrosia (Petrosia) ficiformis, Ircinia variabilis, Agelas oroides and Spirastrella cunctatrix were found in one-third of the explored Mediterranean caves (Gerovasileiou & Voultsiadou 2012).
However, attention must be paid to the fact that 67% of the species recorded in total were found in less than five caves and 34.5% were found only in one (Gerovasileiou & Voultsiadou 2012). This is particularly true for the semi-submerged caves characterised by the presence of sciaphilous species, also able to live outside. The most frequent species, present in almost 70% of the 10 semi-submerged caves until now analysed along the Italian coast Pansini et al. 1977;Corriero 1989;Corriero et al. 1997Corriero et al. , 2000, were Clathrina coriacea, Crambe crambe, Spirastrella cunctatrix, Chondrosia reniformis and Ircinia variabilis. The abovecited species were recorded also during the study of the Sarà caves in the Ligurian Sea.
Moreover, in this type of caves massive species tend to show an encrusting growth habit, becoming flat, as a consequence of the intense water-movement (Corriero et al. 2003).
Long-term studies to evaluate possible changes in the sponge composition and structure within  Mediterranean caves are yet to come. Thanks to the observations conducted by Sarà (1964) about 55 years ago, a first comparative analysis of the sponge assemblages present within two semi-submerged caves in the Ligurian Sea was possible. During this time span, the sponge assemblages of the considered caves experienced an increase in terms of specific richness and a significant change in the structural aspects. The three-dimensional growth forms decreased or remained quite similar, while numerous new two-dimensional ones were recorded in both caves, regardless of differences in cave morphologies, sunlight gradients, and water-movement and  lithology variables which determine the spatial distribution of each species. Similar results were obtained by Montefalcone et al. (2018) studying the benthic community settled inside the Bergeggi Cave (Western Ligurian Riviera). A reduction in massive demosponges in the Mediterranean was also observed in other littoral communities (Bertolino et al. 2016;Betti et al. 2017). In particular, species of the genus Spongia have suffered intense mass mortalities since the last two decades of the past century (Cerrano et al. 2000;Garrabou et al. 2009). Our data are in agreement with this tendency. On the other hand, the Mediterranean rocky communities have experienced a widespread increase of encrusting species in terms of both diversity and abundance (Cerrano et al. 2000;Bianchi et al. 2014a,b;Bertolino et al. 2016;Betti et al. 2017). Also, palaeontological data, obtained for the coralligenous community and extended to the entire Holocene, strongly suggest that sponge diversity increased in warmer periods .
Over the past few decades, the coasts onto which the caves open were not particularly disturbed by anthropogenic activities. We may therefore hypothesise that the recorded changes could be exclusively attributed to climatic events, which favoured some species and disadvantaged others. In conclusion, the two sponge communities of the semi-submerged caves appear poorly resilient and the massive sponges might have suffered from the positive thermal anomalies which occurred in the Ligurian Sea in the last few decades, and were replaced by the encrusting ones. The process of cave recolonisation and recovery after major disturbances appears slow and studying it is necessary for the management and conservation of these habitats .

Funding
This work was supported by the SIR grant (Italian Ministry of Education) [RBSI14CH49].

Disclosure statement
No potential conflict of interest was reported by the authors.