Marine Invasion in the Mediterranean Sea: The Role of Abiotic Factors When There Is No Biological Resistance

The tropical red alga Womersleyella setacea (Rhodomelaceae, Rhodophyta) is causing increasing concern in the Mediterranean Sea because of its invasive behavior. After its introduction it has colonized most Mediterranean areas, but the mechanism underlying its acclimatization and invasion process remains unknown. To understand this process, we decided i) to assess in situ the seasonal biomass and phenological patterns of populations inhabiting the Mediterranean Sea in relation to the main environmental factors, and ii) to experimentally determine if the tolerance of W. setacea to different light and temperature conditions can explain its colonization success, as well as its bathymetric distribution range. The bathymetric distribution, biomass, and phenology of W. setacea were studied at two localities, and related to irradiance and temperature values recorded in situ. Laboratory experiments were set up to study survival, growth and reproduction under contrasting light and temperature conditions in the short, mid, and long term.Results showed that, in the studied area, the bathymetric distribution of W. setacea is restricted to a depth belt between 25 and 40 m deep, reaching maximum biomass values (126 g dw m−2) at 30 m depth. In concordance, although in the short term W. setacea survived and grew in a large range of environmental conditions, its life requirements for the mid and long term were dim light levels and low temperatures. Biomass of Womersleyella setacea did not show any clear seasonal pattern, though minimum values were reported in spring. Reproductive structures were always absent. Bearing in mind that no herbivores feed on Womersleyella setacea and that its thermal preferences are more characteristic of temperate than of tropical seaweeds, low light (50 µmol photon m−2 s−1) and low temperature (12°C) levels are critical for W. setacea survival and growth, thus probably determining its spread and bathymetric distribution across the Mediterranean Sea.


Introduction
The spread of non-indigenous species is claimed to cause dramatic ecological impacts and is considered a major threat for biodiversity conservation [1]. The Mediterranean Sea is one of the areas of the world most severely hit by those impacts with about 955 introduced species [2], among which macroalgae are considered to be especially worrying [3] because they may alter both ecosystem structure and function by monopolizing space and developing into ecosystem engineers [4]. The red alga Womersleyella setacea (Hollenberg) R.E. Norris is one of at least eight species that can be assigned to the category of invasive macroalgae in the Mediterranean [3]. In many Mediterranean localities it is exceedingly abundant, forming thick, persistent carpets that completely cover deep sublittoral rocky substrata [5][6][7][8][9], have substantial negative effects on native communities [7,[10][11], modify benthic assemblages [5,8,[12][13][14], and outcompete key species [15]. Womersleyella setacea was described originally from the Hawaiian Islands [16] and later reported for other tropical localities both of the Pacific and Atlantic oceans [17]. It was first observed in Mediterranean coastal waters in the eighties in the Var region, France [18] and in Italy [19], and rapidly increased its distribution throughout Mediterranean waters: Corsica, Mediterranean coasts of Spain, the Balearic Islands, the Adriatic Sea, Malta and Greece [17]. The origin and way of introduction of this species remain unknown, but a suggested vector is ship hull fouling [20].
Only a small fraction of the many marine species introduced outside of their native range are able to invade and thrive in new habitats [21]. Studies of traits that make non-indigenous marine species invasive are essential to understanding the invasion procedure and to identify the key processes and filters that determine their success [22]. One of the first suggested determinants of the invasion process is the climate, since it sets broad limits to invader distribution and may cause introduced species to fail immediately during colonization [23]. However, environmental conditions that suit W. setacea are unknown. In fact, there is little published information about the phenology of the introduced Mediterranean populations (but see [5,9,13]), and the only previous study of its physiology is restricted to short response observations concerning a few weeks and a narrow range of light and temperature conditions [24], preventing further generalizations in the long term. Therefore, the main goal of the present paper is to describe the seasonal biomass and phenological patterns of natural W. setacea populations, relating them with the main environmental factors. In addition, we set up two batteries of laboratory experiments to 1) study seasonal patterns of W. setacea in relation to daylength, light and temperature, and 2) assess short-(1 month), mid-(3 months), and long-term (1 year) light and temperature requirements and tolerance for survival, growth and reproduction of specimens of a Mediterranean population of W. setacea. Further, we try to relate these results to its natural bathymetric distribution and its colonization success. And the adaptive capacity of W. setacea to winter environmental conditions, the most critical season in the Mediterranean Sea for tropical algae, was also investigated.

Study site
The present study was carried out in the Scandola Natural Reserve (Parc Naturel Régional de la Corse), Corsica, France (Fig. 1). This marine protected area (hereafter Scandola MPA) was established in 1975 and covers 1000 ha. It has been recognized by the United Nations as a natural World Heritage Site and was inscribed in the world heritage list in 1983. The invasive species W. setacea was first recorded in the Scandola MPA in 1989 although it was probably introduced somewhat earlier [18], and at the beginning of this study it was already widespread in all the MPA (authors' pers. obs.).

Bathymetric distribution of Womersleyella setacea in the Scandola MPA
Within the Scandola MPA two rocky bottom sites (Imbutu and Cala Solana, Fig. 1) were chosen and a perpendicular transect to shore was established at each site. Depth of both transects ranged between 0 and 40 m. Coverage of W. setacea was estimated by means of quadrats of 25625 cm divided into 25 subquadrats of 565 cm [25][26], and the number of subquadrats in which W. setacea appeared was recorded and used as a unit of measure. Twenty quadrats (total area of 1.25 m 2 ) were randomly positioned within each 5 m depth range.

Biomass sampling procedure
The study of the biomass annual cycle lasted from November 2007 to October 2008 and was performed at both sites, Imbutu and Cala Solana, at 30 m depth, in order to cope with the observed maximum W. setacea development (see bathymetric distribution results, Fig. 2).
At each study site three randomly positioned plots of 20620 cm were periodically sampled throughout one year by SCUBA divers. The plots were scraped and specimens were stored in individual plastic bags. Sorting of W. setacea was performed in the laboratory, and its biomass was measured as dry weight (dw) by drying the sample in an oven at 60uC to constant weight.

Photon flux density and temperature recording
To characterize light intensity in W. setacea distribution habitats, photon flux density was recorded at both localities by HOBOH LI data loggers (Onset Computer, EME Systems, Berkeley, CA, USA), at 25, 30, 35 and 40 m depth, hourly during one week both in summer (June) and in autumn (October).

Laboratory experiments
Assays were performed in laboratory culture with specimens submitted to conditions simulating the seasonal changes of daylength, light and temperature in the field ( = seasonal experiments), in order to confirm the seasonal pattern of W. setacea populations in relation to these factors. Furthermore, some specimens were submitted to constant conditions of light and temperature to study the tolerance of W. setacea to those factors ( = ecophysiological experiments).
Specimens were collected by SCUBA divers at Imbutu, on 16 October 2009, from coralligenous assemblages situated at 30 m depth. Material was placed in situ in opaque cotton bags to avoid high light intensities, transported in insulated boxes to the laboratory, and kept at 16uC overnight. The day after collecting, specimens were rinsed in sterilized seawater. Vegetative tufts of similar size (around 1.5 cm 2 ) were exited and cultured in 250 ml vessels in incubators (Radiber, Barcelona, Spain) equipped with 30 W cool white fluorescent bulbs. Photosynthetic photon flux densities (PPFD) were measured by means of a Li-1400-501 quantometer (Li-Cor, Lincoln, NE, USA) and adjusted to the desired levels by neutral density filters. The culture medium was a quarter-strength modified Von Stosch enriched seawater medium [27] in which diatom growth was controlled by 5 mg l 21 GeO 2 [28] and (cyano-) bacteria growth by 3 mg l 21 Penicillin-G [29]. Cultures were shaken daily by hand. The medium was changed weekly to avoid nutrient depletion, using a medium previously preheated to the experimental temperature. Specimens were cleaned of epiphytes weekly.
The survival of cultured thalli was determined as the time between the collecting date and the death of the specimen. At the beginning of the cultures (initial time), all the thalli looked rosy and healthy, but over the course of the culture the specimens went through two different survival phases. The first phase, Phase SI, began at the initial time and finished when the thalli were more than 95% damaged; and a second phase, Phase SII, began at the end of Phase SI and continued until the death of the specimens.
In seasonal experiments, cultures (with three replicates per culture) were maintained successively at daylength and temperature conditions of winter, spring, summer and autumn, combined with PPFD of 5, 10, 20, 50 and 75 mmol photon m 22 s 21 ( Table 1). Changes of season in culture conditions coincided with changes of season throughout the year. In ecophysiological experiments, cultures (with four replicates per culture) were grown in a variety of temperature and PPFD regimes and 8:16 h light:dark (Table 2).
In both seasonal and ecophysiological experiments the length of the SI phase, the survival, and reproduction of the thalli were followed for up to one year. Cultures were monitored every week during the first three months of culture and monthly thereafter, by means of a photograph taken with a Canon EOS 350D (Canon, Tokyo, Japan). If quick changes were observed in the specimens, photographs were once again taken weekly.
The thallus surface (in mm 2 ) was determined at the initial time of the culture and every two months using ImageJ (National Institutes of Health, USA). Representative specimens of material used in cultures have been deposited in the Herbarium of the University of Girona (HGI).

Statistical analysis
In situ biomass was analyzed for differences between sites (Imbutu and Cala Solana) and within each site for time (five levels) by means of a Kruskal-Wallis non-parametric analysis. A nonparametric approach was used because of heteroscedasticity and departures from normality. In seasonal experiments and ecophysiological experiments a one-way ANOVA on thallus surface values obtained during the first day of the experiment was carried out to verify that all thalli undergoing different treatments started at similar conditions. In seasonal experiments, after one year of culture, final thallus surface, survival and length of Phase SI were analyzed with a one-way ANOVA with PPFD (four levels) as a fixed factor, whereas in ecophysiological experiments, thallus surface, survival and length of Phase SI were analyzed by a twoway ANOVA, where temperature (six levels) and PPFD (four levels) were fixed orthogonal factors. Tukey tests were used for a posteriori multiple comparisons of means data. Assumptions of normality and homogeneity of variances were examined using the Kolmogorov-Smirnov and Barlett tests, respectively. Variables were rank-transformed prior to the analysis when assumptions were not fulfilled. The analyses were performed using the STATISTICA 6.0 package.

Bathymetric distribution
Womersleyella setacea displayed the highest quantitative dominance between 30 and 35 m depth, with a maximum coverage at 30 m depth, both at Imbutu and Cala Solana, while it was absent at depths shallower than 20 m (Fig. 2). W. setacea maximum coverage was observed at depths where incident light levels ranged between 10 and 150 mmol photon m 22 s 21 , and especially between 10 and 50 mmol photon m 22 s 21 (Fig. 2).

Biomass annual cycle
Womersleyella setacea biomass was similar at both localities (K-W test; p = 0.247), ranging between 0.6260.34 (mean 6SE) g?dw?sample 21 on April 08 and 5.0360.47 (mean 6SE) g?dw?sample 21 on November 07 at Imbutu (Fig. 3). During most of the year W. setacea biomass remained high and stable, without significant differences through time in Cala Solana (K-W test; p = 0.807), and only in spring W. setacea biomass showed a decrease in Imbutu (Tukey test p.0.05) (Fig. 3). The minimal biomass in spring is more evident at Imbutu probably because the maximal biomass is higher by comparison to Cala Solana.

Seasonal experiments
In seasonal experiments the surface of cultured specimens were initially similar ( Fig. 4A; ANOVA-test, p.0.05), and during the experiments all specimens of all the treatments grew progressively, increasing their surface at different rates depending on the PPFD (Fig. 4A, ANOVA-test, p,0.05). Survival, and length of Phase SI, were PPFD dependent too (

Ecophysiological experiments
At the beginning of the experiment, all cultured specimens had a similar initial thallus surface ( Fig. 5A; ANOVA-test, p.0.05). In the short term (one month), W. setacea was able to survive in all the combinations of temperature and PPFD assays, whereas in the mid term (3 months) most specimens cultured at high temperatures died, and in the long term (1 year) only those cultured below 18uC survived (Figs. 5A, 5B).
Womersleyella setacea survival was both temperature and PPFD dependent ( Table 3, (Fig. 5B). The optimal conditions were very different in the short, mid and long term. At short and mid term, the optimal conditions were 16uC combined with 50-75 mmol photon m 22 s 21 , but at the end of cultures (1 year), the specimens that still were in Phase SI, looked healthier and continued to maintain an increase in thallus surface were those cultured at 10uC and 20-75 mmol photon m 22 s 21 and at 12uC and 50 mmol photon m 22 s 21 (Fig. 5A). The maximum growth was observed in these last conditions (Figs. 5A, 6). Most specimens cultured at 24 and 26uC only persisted in Phase SI for less than 1 month (Fig. 5B).

Phenology
We did not observe any reproductive structure (sporangia or gametangia), either in the field or in the cultured specimens.

Discussion
The present study provides data on the abundance and temporal variability of W. setacea in the field. We observed high values of W. setacea biomass (up to 126 g dw m 22 in November at Imbutu), which are of the same order of magnitude as those recorded in other Mediterranean areas [5]. Similarly, the values obtained are comparable to those of other well-known invasive species thriving in benthic Mediterranean assemblages. For instance, values reported for the Chlorophyta Caulerpa racemosa, although they can be much higher, rarely exceed 100 g dw m 22 [30][31][32][33], and those reported for the Rhodophyta Lophocladia lallemandii range between 50 and 200 g dw m 22 [34]. Similarly, although maximum values of Caulerpa taxifolia in Mediterranean rocky bottoms range between 200 and 500 g dw m 22 [35][36][37], most C. taxifolia populations exhibit lower biomass values (from 50 to 200 g dw m 22 ) [38][39][40][41].
In general, no seasonal variation pattern has been found in the Scandola MPA populations of W. setacea. A dense thick red filamentous turf was widespread and persistent throughout the year at both localities studied, except in April, when biomass showed a slight decrease, in agreement with previous studies on W. setacea populations in other Mediterranean areas [5,9,13]. Populations of W. setacea from the Scandola MPA seem to propagate only by vegetative ways, what also agrees with other observations available for other Mediterranean regions in the field [5] or in cultures [24]. However, the presence of tetrasporangia was reported in the original collections from Hawaiian populations [16].
Mediterranean W. setacea is considered to be a sciaphilic species constituting thick carpets in deep waters [5]. However, its bathymetric distribution may vary in different areas, probably in function of the environmental features of each region (e.g., in Greece it was observed between 10 and 20 m depth [6], in the northern Aegean Sea between 15 and 30 m depth [13], in the Adriatic Sea, France and Italy between 10 and 30 m depth [5,9,18], and in the Spanish coast between 15 and 30 m depth (Cebrian, pers. obs.)). The bathymetric distribution of W. setacea shows that in the Scandola MPA it is restricted to depths between 25 and 35 m.  The present light requirements confirm the sciaphilic behavior of W. setacea, which presents a maximal survival and growth at PPFD ranging between 10 and 75 mmol photon m 22 s 21 . Depth limits of algal distributions at increasing depths are related to the decrease of survival at lower PPFD [42][43][44]. Lower survival and growth of W. setacea were observed at lower PPFD (5 mmol photon m 22 s 21 ), explaining that in the Scandola MPA W. setacea almost disappeared or, at least, did not form dense carpets below 40 m depth, where PPFD were usually below 10 mmol photon m 22 s 21 .
Results show that Mediterranean populations of W. setacea are adapted to relatively cold waters, specifically below 16uC. In the western Mediterranean these temperature conditions are common  in the entire water column in winter and spring but are restricted to below the thermocline in summer and autumn (Fig. 7), at depths also characterized by low light conditions. In fact, all the specimens submitted to a constant temperature of 10uC experienced a small but persistent growth during the entire culture period, explaining why the thick carpets of W. setacea are present all year long in the field, and why winter temperatures are not a limiting factor for W. setacea invasiveness. These clear temperate water affinities compromise the until now stated tropical origin of W. setacea, highlighting the need for further comparative molecular investigation on the phylogeny and biogeography of Mediterranean and extra-Mediterranean populations of W. setacea to clarify the relationship of the invasive strain to the original tropical strain. Tolerance to low temperatures also occurs in other invasive species dwelling in the Mediterranean. For instance, the invasive Mediterranean strain of Caulerpa taxifolia is able to survive three months at temperatures ranging from 10 to15uC [45][46]. Nevertheless, the Mediterranean strain of W. setacea differs from C. taxifolia in its inability to support warm temperatures. W. setacea would be able to survive summer periods (of less than two months) in the upper infralittoral only at 20 mmol photon m 22 s 21 combined with temperatures #24uC. This explains in part the absence of W. setacea at the shallow infralittoral algal assemblages in the Scandola MPA, where temperatures commonly exceed 24uC (Fig. 7). Similarly, temperature requirements support the fact that in all Mediterranean sites where W. setacea has been reported it never develops thick carpets above 10 m depth, where temperatures can rise up to 25-30uC ([5-6,9,18,], this work).
The lack of seasonality of W. setacea biomass in the field is in agreement with the seasonal experiments at the laboratory, where no differences were found in the surface increase among the different periods of the year, probably because seasonality of underwater PPFD, temperature, hydrodynamics, and nutrient availability are minimized with depth [47][48][49][50]. Besides, W. setacea is capable of continuous asexual vegetative spread throughout the year [6,24,51], creating a very stable and homogenous habitat which may contribute to the loss of seasonality in the macrobenthic community structure [13].
Womersleyella setacea can thrive as an epiphyte and is able to overgrow other sessile benthic organisms. Therefore it is not limited by the availability of free substrate, and avoids competition with natives for substrate. On the other hand, invulnerability to native herbivores may be an important determinant of invasion success of marine macroalgae. While some studies reject the biological resistance hypothesis for marine macroalgae (e.g. [52][53]), others provide evidence that invasive algae are actively grazed by native herbivores, which can control their populations [54][55]. However, recent research addressed whether native generalist herbivores may provide resistance to macroalgae invasion in the Mediterranean Sea reveal that W. setacea is not consumed either by sea urchins nor Sarpa salpa [56][57][58]. Thus, we can hypothesize that in absence of a biological resistance in the new habitat, the main factors driving W. setacea development are light, temperature, and probably physical disturbance. The importance of light and temperature as major factors determining seaweed distribution has been stressed by many authors (see review in [59]); likewise, our field observations and experimental work revealed these factors as being critical for W. setacea survival and growth, and they are probably determining the spread and bathymetric distribution of this species across the Mediterranean Sea.
The invasion success of W. setacea in Mediterranean deep water assemblages probably relies on two different abilities. The first one is the ability to maintain permanent carpets all year long, outcompeting both perennial and ephemeral native species. The second ability would be the enhanced and sustained growth of W. setacea, higher than those recorded for other Mediterranean perennial or pseudoperennial native macroalgae growing in deep waters and cultured under the same conditions [60][61][62]. Thus, high growth and persistence should be the basis of Womersleyella setacea's capacity to outcompete native and engineering macroalgae and invertebrates from deep-water Mediterranean bottoms [8,15]. Additionally, Womersleyella setacea traps sediment and, in this way, it may pre-empt space and prevent the attachment of possible spatial competitors [9] as described for other filamentous species [63]. This makes the settlement of native species and the survival of their juvenile stages impossible [64], thus reducing the species diversity and equitability of phytobenthic communities [5,7,[65][66].  In brief, populations of W. setacea thriving in Mediterranean coastal waters exhibit thermal features of a temperate seaweed rather than a tropical one. Thus, winter temperatures are not a limiting factor in the survival and spread of W. setacea all over the Mediterranean Sea, and hot summer temperatures prevent its distribution in shallow waters. Womersleyella setacea also performs much better at low light intensities around 50 mmol photon m 22 s 21 , which is in accordance with the depth distribution found in the field. We suggest that the invasion success of W. setacea in Mediterranean environments relies on its high growth, persistence and resistance to herbivores, which outcompetes both perennial and ephemeral native species growing at the same depth intervals where W. setacea develops.