Marine invasions on a subtropical island : fouling studies and new records in a recent marina on Madeira Island ( Eastern Atlantic Ocean )

In recent years, several marine non-indigenous species (NIS) lists have been produced for many European countries but little is known about the diversity and distribution of fouling NIS in Portugal (mainland and islands). We conducted a six-year survey of a marina located on the south coast of Madeira island, Portugal to assess NIS diversity on the island, constituting the first NIS inventory for the archipelago. We found 16 NIS, of which 9 are new records. Both species richness and abundance changed during the course of colonization whether total, NIS, or native diversity were considered. The number of native species decreased with colonization while the number of NIS significantly increased. More importantly, we demonstrated that the number of NIS detections in the marina was correlated with increasing ship traffic over the years.


Introduction
Biological invasions of marine non-indigenous species (NIS) are frequent in coastal communities, and the rate of documented invasions has significantly increased in the last two decades (Ruiz et al. 2009).The greatest proportion of recent marine invasions has been facilitated through the transport of species in ballast water and as hull fouling (Ruiz et al. 2000); the latter vector conveying numerous "fouling" organisms around the world, including sponges, hydroids, tube worms, barnacles, mussels, bryozoans, sea squirts, and algae.In fact, every hard artificial substrate in the marine realm (e.g.ship hulls, piers, pontoons, pilings, seawalls and buoys) is subject to biofouling, i.e., the accumulation of sessile microorganisms and macroorganisms with time (Wahl 1997;Railkin 2004;Dürr and Thomason 2010).The species composition of any fouling community can change considerably over space and time (Railkin 2004;Canning-Clode et al. 2010;Dürr and Thomason 2010).Biofouling assemblages represent an ideal study system for ecologists, due to their amenability for experimentation, their ease of access (without requiring low tides or vessel use), and due to the organisms' proclivity to settle on artificial substrates and their subsequent fast growth (Canning-Clode and Wahl 2010).
In the last decade, lists of marine non-indigenous species have been produced for several European countries such as the UK, Belgium, Denmark, Germany, France, Greece, and Italy (Goulletquer et al. 2002;Occhipinti Ambrogi 2002;Jensen and Knudsen 2005;Pancucci-Papadopoulou et al. 2005;Kerckhof et al. 2007;Minchin et al. 2013).The only NIS inventory for Portuguese waters is for the Azorean archipelago (Cardigos et al. 2006).In addition, although a few new records have been reported for the island of Madeira (e.g.Wirtz 1998;Wirtz and Canning-Clode 2009), little is known about the diversity and distribution of fouling NIS on the island archipelago.Consequently, we conducted a sixyear survey of a recently constructed marina located on the southeast coast of Madeira Island to contribute to a first inventory of marine NIS in Madeira.We deployed settling plates and examined diversity and abundance at different stages of the colonization process of the recruiting fouling assemblages.We also examined vessel traffic in the marina in recent years and hypothesized that i) species richness would change with time, regardless of what category was considered (i.e.total, NIS, native or cryptogenic richness); ii) species composition would change in time; and iii) and NIS numbers were positively correlated to vessels traffic in the marina over time.

Sampling
This study was conducted at Quinta do Lorde Marina (QLM), located at the southeast part of Madeira Island (32º44.5'N,016º42.8'W; Figure 1 At each sampling event, settling plates were detached, retrieved from the field, and photographed.Plates were then brought back into the field within 3 hours of each sampling event.Species richness and abundance at plate level were determined by recording the number of species identified from the photographs using the image analysis software CPCe (Kohler and Gill 2006).Voucher specimens of each species were further submitted, as necessary, to expert taxonomists for verification or identification.Algae and sessile macroinvertebrates were identified to the lowest possible taxonomic group based on existing literature or taxonomic experts and assigned to one of four categories: native, NIS, cryptogenic (i.e., unspecified origin, that is, unknown whether native or introduced) or unresolved (based on an inability to identify to species level).
Each photographic image was sub-divided into a 3 × 3 grid of 9 cells, with 11 random points per cell resulting in 99 points analyzed per picture.This stratified random sampling method attributes points to each image region (Kohler and Gill 2006).In addition, each plate was examined using a dissecting microscope for better resolution.During this procedure, fouling assemblages were always submerged in seawater and protected from direct sun exposure.These analyses permitted us to place species into one of three bins in terms of spatial cover: ≤ 1% mean cover; < 10% mean cover, and > 10% mean cover (Appendix 1). Finally

Data analysis
For each category of diversity (i.e., total, NIS, native and cryptogenic), linear and quadratic models were used to test for relationships between diversity of fouling assemblages and colonization age.If both linear and quadratic models were significant, the best fit was accepted based on its R 2 .For the purpose of this analysis, species classified as unresolved were pooled into the cryptogenic category.
In addition, percent cover data were used for multivariate analyses of similarity (ANOSIM) to evaluate effects of colonization age on species composition.Non-metric multidimensional scaling (MDS) was applied to visualize similarity between compared groups.We then used the SIMPER routine to evaluate the contribution of each taxon to average dissimilarities between groups.The more significant taxa causing these dissimilarities were identified (Clarke and Warwick 1994).Multivariate analysis was performed in the R platform (R Core Team 2013) with the 'vegan' package (Oksanen et al. 2012).
Finally, the relationship between ship traffic at QLM over the years and number of accumulated NIS was tested with linear regression.All statistical analysis was performed in the R platform (R Core Team 2013).

Results
Over the course of six years, 49 taxa were recorded in Quinta do Lorde Marina, of which 16 were categorized as native (33%), 16 as NIS (33%) and 8 as cryptogenic (16%) (NIS shown in Table 1.Full species list and abundances in Appendix 1).Nine of the species classified as NIS were new records for the Madeira Archipelago.
The most frequent taxonomic groups amongst NIS were bryozoans (44%) and ascidians (31%), while other taxonomic groups contributed a smaller number of NIS, such as sponges and barnacles (13% and 6%, respectively).NIS diversity was higher after 44 and 50 months with 13 and 10 species, respectively.In addition, the sponge Mycale senegalensis and the ascidian Distaplia corolla were both detected after month 44 and have shown the highest abundances thereafter (Table 1 and Appendix 1).As well, certain species here categorized as cryptogenic will probably turn out to be NIS in future surveys once their identification is resolved (e.g.

Haliclona cf. indistincta, Celleporaria cf inaudita, Didemnum cf. perlucidum, Distaplia cf. bermudensis, see Appendix 1).
Species richness and abundance of fouling communities colonizing settling plates at QLM changed significantly over the course of the study (Table 2, Figure 2).Quadric models of the relationship between diversity and colonization fit better than linear models (Table 2).Total species richness changed significantly over the course of colonization reaching maximum species richness at 12 months, followed by an apparent climax community until month 50, and dropping at month 74 (Figure 2A).In contrast, NIS diversity increased with time (Figure 2B) while native and cryptogenic richness displayed a near linear decrease after 24 months (Figure 2C-D).Species abundances showed very similar patterns to the ones of species richness (Figure 2E-H).All species produced maximum cover after one year of colonization, when communities reached an apparent equilibrium (Figure 2E).As with species richness, NIS cover significantly increased over the course of colonization (Figure 2F).Native species cover showed a similar pattern to the one   displayed by total species cover with maximum cover after 24 months of colonization (Figure 2G).Finally, percent cover of species included in the cryptogenic category significantly decreased with time (Figure 2H).
Finally, according to SIMPER routines, the invasive ascidian Distaplia corolla was essential in differentiating early colonization communities from late assemblages (Table 3).Accordingly, average abundances of Distaplia corolla increased over the colonization process and had a 20% positive contribution to dissimilarities between T3 and T50 communities and over 45% positive contribution to dissimilarities between T3 and T74 assemblages (Table 3).
The number of vessels arriving at Quinta do Lorde Marina increased after we commenced our work in 2006, from pre-2006 levels of 300 vessels or less, to vessel arrivals of more than 400 (and up to 600) per year (black line, Figure 4).
Similarly, we detected a significant positive relationship between the accumulative number of vessel arrivals and cumulative number of NIS during the survey period (P > 0.05; R 2 = 0.76).
Approximately 50% of vessels arriving every year at Quinta do Lorde Marina come from the neighboring island of Porto Santo.Other significant ports of origin were: Portugal mainland (6%); Canary Islands (6%); Mediterranean (5%); The relationship between the colonization and total richness (and total abundance to some extent) displayed a unimodal pattern with maximum diversity at middle stages of colonization.Moreover, while the number of native species seemed to decrease with colonization time, the number (and abundance) of NIS increased.These findings seem to be consistent with other studies focusing invasions in fouling assemblages.For example, in a study to assess the effects of diversity on the invasion of sea squirts in a fouling community, Stachowicz et al. (2002) observed a decline in native diversity as the abundance of invaders increased over time.
The number of NIS detections in the marina was not independent from vessel traffic; there is a positive relationship between accumulative number of NIS and accumulative vessel traffic.The majority of vessels arriving at QLM came from the neighboring island of Porto Santo; a significant portion of arrivals in both marinas (Quinta do Lorde and Porto Santo) came from Portugal mainland, Mediterranean and Northern Europe.This fact may suggest that most NIS detected in this study were likely secondary or tertiary introductions.That is, Madeira is receiving species introduced earlier to other European ports, and is not (at this time) a site for novel invasions of the European theatre.The resulting introduction of NIS into marinas via secondary introduction has been recently detected in different biogeographic regions (e.g.Johnson et al. 2001;Floerl and Inglis 2005;Ashton et al. 2006;Clarke Murray et al. 2012;Minchin et al. 2013).Therefore, recreational boating may play a key role in the secondary spread of marine NIS (Ashton et al. 2006;Clarke Murray et al. 2011).
Finally, due to budget restrictions we conducted this survey in only one marina.One marina is unlikely to be representative of the full NIS diversity in the archipelago.For this reason, future studies should incorporate other marinas, as well as other habitats, on these islands.We believe additional surveys in other marinas and habitats (e.g.pontoon, dock) would result in other marine NIS detections.

Figure 1 .
Figure 1.Location of the Quinta do Lorde Marina (QLM) on the southeast coast of Madeira, Portugal.

Figure 2 .
Figure 2. Relationship between species richness and colonization (panels A-D) and relationship between species abundances and colonization (panels E-H) at Quinta do Lorde Marina after 74 months of colonization.Means and 95% confidence intervals are indicated (n = 10).

Figure 4 .
Figure 4. Number of vessels arriving at Quinta do Lorde Marina since its construction and accumulated number of NIS detected during the present study.Black line and black circles for number of vessel arrivals.Gray line and gray squares for accumulative number of NIS.Dashed line indicates the beginning of our survey.

Figure 5 .
Figure 5. Percentage of yearly vessel arrivals from different biogeographic regions at Quinta do Lorde Marina (A) during 2008 and 2011 and at Porto Santo Marina (B) during the same period.Means and standard deviations are indicated.

Table 1 .
List of the 16 non-indigenous species (NIS) found in Quinta do Lorde Marina (QLM) from 2006 to 2012.For a detailed full species list and abundances over time see Appendix 1.
, since QLM only opened in 2002, we were able to compare the number of NIS relative

Table 2 .
Effects of colonization age on diversity of fouling communities after a 6 year fouling survey conducted at QLM. Results of the linear and polynomial regression analyses are shown for species richness and percent cover.

Table 3 .
Stachowicz et al. 2002;Valdivia et al. 2005;Sugden et al. 2008;Canning- Clode et al. 2011;Crooks et al. 2011ch taxa contributed more (>10%) to observed changes in community composition between colonization ages.Contribution (%) and direction of change (+ positive; -negative) are indicated.Clode 2009), little was known about the diversity and distribution of fouling NIS in this region.Although the data we present here were restricted to one single marina in the island of Madeira, the present work constitutes the first NIS inventory for the Madeira archipelago.During this study we found 16 NIS, of which 9 are new records to Madeira archipelago.Most of the NIS present in the marina were bryozoans, ascidians and sponges and they were detected after month 44.Indeed, several recent studies conducted in marinas and bays across the globe have found ascidians and bryozoans among the most abundant organisms colonizing settling plates (e.g.Stachowicz et al. 2002;Valdivia et al. 2005;Sugden et al. 2008;Canning- Clode et al. 2011;Crooks et al. 2011).Both species richness and abundance changed during the course of colonization of settling plates in QLM.