Species Interactions and Regime Shifts in Intertidal and Subtidal Rocky Reefs of the Mediterranean Sea

The Mediterranean Sea is an enclosed basin connected to the Atlantic through the Strait of Gibraltar in the west and to the Sea of Marmara and the Black Sea through the Dardanelles in the east. Its surface waters cover 2,969,000 km2, which makes the Mediterranean the largest enclosed sea in the world (Bianchi and Morri, 2000). A recent analysis has estimated the occurrence of some 17,000 marine species in the Mediterranean, with taxa such as the Phaeophyta, Rhodophyta and Porifera including more than 10 per cent of the total number of species known globally (Coll et al., 2010). These figures are remarkable, considering that the Mediterranean Sea covers less than 1 per cent of the surface and volume of the world’s oceans. The origin of the highly diversified Mediterranean biota is relatively recent and is mainly derived from the Atlantic Ocean. The Mediterranean Sea has undergone various geological, climatic and hydrological transformations that have contributed to generate the hotspot of marine biodiversity that we see today (Bianchi and Morri, 2000; Coll et al., 2010; Lejeusne et al., 2010). In particular, isolation from theAtlantic during the Messinian crisis (6 Mya) resulted in strong evaporation with consequent dramatic changes in climate, sea level and salinity (Bianchi and Morri, 2000). The Messinian crisis decimated the biota of the ancient Mediterranean, which was largely dominated by Indo- Pacific species of warm-water affinities. When the connection with the Atlantic was reestablished (5 Mya), the newly colonising species mixed with those surviving the Messinian crisis. Alternating glacial and interglacial periods during the Quaternary, favouring the colonisation of boreal and subtropical species, further contributed to the diversification of the biota in the Mediterranean Sea. The opening of the Suez Canal in 1869 has also impacted on the native biodiversity of the Mediterranean through the introduction of new species from the Red Sea (Galil et al., 2014), which will likely increase in the near future following the expansion of the Suez Canal in 2015 (Galil et al., 2015b). There is a long tradition of descriptive studies in the Mediterranean Sea, with detailed accounts on taxonomic composition and regional patterns of distribution of marine organisms that exploded in the 1960s owing to intensifying oceanographic cruises and the advent of scuba diving (reviewed in Coll et al., 2010). Ourunderstanding of region-wide patterns of marine biodiversity in the Mediterranean Sea has increased considerably in the last fifteen years as a result of large-scale field surveys (Sala et al., 2011), extensive reviews of the literature (Bouillon et al., 2004; Danovaro et al., 2010; Martin and Giannoulaki, 2014; Telesca et al., 2015), modelling (Sarà et al., 2013; Marras et al., 2015) and synthesis of expert opinions (Micheli et al., 2013). This integration has generated new insights into the present distribution of the Mediterranean biota, as well as focussing attention on the main threats to marine biodiversity and the need to implement better conservation practices at the regional scale (Airoldi and Beck, 2007; Claudet and Fraschetti, 2010; Coll et al., 2010, 2012; Mouillot et al., 2011). In contrast to descriptive studies, experimental ecology has only been introduced recently in the Mediterranean, with an initial focus on biological interactions. The first experiment that incorporated the logical requirements of replication, randomisation and independence examined the effect of removing a canopy-forming alga in littoral rock pools (Benedetti-Cecchi and Cinelli, 1992a). A parallel study in the same system expanded the range of species interactions examined to include the effects of herbivory and competition between algal turfs (see Connell et al., 2014, for a clarification of the term turf ) and canopy recruitment (Benedetti-Cecchi and Cinelli, 1992b). Experimental work proliferated in the following years, examining species interactions in the context of ecological succession (Benedetti-Cecchi, 2000a, 2000b) and extending this approach to subtidal environments (Airoldi et al., 1995; Airoldi, 2000a). As we will discuss, some of these studies established the foundation for the theory that canopy-forming algae and turfs represent alternative states in shallow temperate rocky coasts under different disturbance and stress regimes (Airoldi, 1998, 2000b; Benedetti-Cecchi et al., 2001b), and that changes in sediment loads are one of the main triggers of these shifts in subtidal habitats (Airoldi et al., 1996; Airoldi and Cinelli, 1997; Airoldi, 2003; Irving et al., 2009). More recently, manipulations on subtidal rocky reefs have shown how low levels of herbivory in combination with the localincrease of nutrients may foster the recovery of macroalgal canopies and associated biota (Guarnieri et al., 2014). Experiments in the Mediterranean have also contributed to focus attention on the variance of ecological interactions (Benedetti-Cecchi, 2000c, 2003). Novel experimental designs have been developed that facilitate the separation of the effects of changing the mean intensity from the variance of biological interactions or any other spatially or temporally variable ecological process (Bertocci et al., 2005; Benedetti-Cecchi et al., 2006). These experiments revealed, for example, how spatial variance and mean intensity of grazing may interactively maintain spatially heterogeneous patterns of algal cover, illustrating the great potential for grazing to generate alternative states in marine benthic habitats (Benedetti-Cecchi et al., 2005). Motivated by the need to understand the effects of intensifying anthropogenic impacts and climate change on marine biodiversity, an increasing number of studies are now examining species interactions in relation to regional stressors and global threats such as ocean warming, acidification, extreme climate events and biological invasions (Bulleri et al., 2016). A better understanding of biotic interactions in the Anthropocene is also essential to guide habitat rehabilitation and restoration efforts. Here, we provide an overview of this type of research that will likely characterise future experimental research in the Mediterranean and elsewhere, using regime shifts as a conceptual framework to guide this approach. The concept of regime shifts, the abrupt transition between alternative states, unifies key aspects of species interactions and their responses to degrading environmental conditions, including resilience, early warning signals of collapse, extinction and hysteresis, all of which have direct bearing on environmental management. To achieve this, we start with a brief introduction to regime shifts and the underlying theory, followed by a discussion of ongoing regime shifts in the Mediterranean; such as the transition from macroalgal forests to turfdominated assemblages and the widespread collapse of sessile organisms in response to heatwaves, species invasions, infectious diseases andpest metabolites. We then examine the implications of threshold-like biological responses and hysteresis that are typically associated with regime shifts for habitat restoration and rehabilitation. Finally, we conclude with an overview of the research that is needed to understand the interplay between species interactions and rapid environmental change, for which the Mediterranean is providing several dramatic examples.