Response of top shell assemblages to cyclogenesis disturbances. A case study in the Bay of Biscay

Marine Environmental Research xxx (2015) 1e9

Contents lists available at ScienceDirect

Marine Environmental Research journal homepage: www.elsevier.com/locate/marenvrev

Response of top shell assemblages to cyclogenesis disturbances. A case study in the Bay of Biscay ~ oz-Colmenero a, G.-J. Jeunen a, Y.J. Borrell a, J.L. Martinez b, P. Turrero c, M. Mun E. Garcia-Vazquez a, * a b c

Departamento de Biología Funcional, Universidad de Oviedo, Spain Servicios Científico-T ecnicos, Universidad de Oviedo, Spain n a Distancia, Campus de Gijo n, Spain Universidad Nacional de Educacio

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 November 2014 Received in revised form 15 June 2015 Accepted 18 June 2015 Available online xxx

Cyclones and other climate disturbances profoundly affect coastal ecosystems, promoting changes in the benthic communities that require time, sometimes even years, for a complete recovery. In this study we have analysed the morphological and genetic changes occurred in top shell (Gibbula umbilicalis and Phorcus lineatus) assemblages from the Bay of Biscay following explosive cyclogenesis events in 2014. Comparison with previous samples at short (three years before the cyclogenesis) and long (Upper Pleistocene) temporal scales served to better evaluate the extent of change induced by these disturbances in a more global dimension. A significant increase in mean size after the cyclogenesis was found for the two species, suggesting selective sweeping of small individuals weakly adhered to substrata. Loss of haplotype variants at the cytochrome oxidase subunit I gene suggests a population bottleneck, although it was not intense enough to produce significant changes in haplotype frequencies. The high population connectivity and metapopulation structuring of the two species in the area likely help the populations to recover from disturbances. At a wider temporal scale, cyclogenesis effects seemed to compensate the apparent decreasing trends in size for P. lineatus occurred after the Pleistocene eHolocene transition. Considering disturbance regimes for population baselines is recommended when the long-term effects of climate and anthropogenic pressures are evaluated. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Cyclone events Genetic diversity Gibbula umbilicalis Metapopulation Phorcus lineatus Size selection Temporal variability

1. Introduction Marine communities are heavily impacted by different environmental factors, both anthropogenic and natural. Among the latter, “natural disasters” such as tropical storms or hurricanes, despite their relatively infrequent occurrence, can affect the benthic marine biota, from continental shelves to deep and abyssal bottoms (Harris, 2014). Natural disturbances are an important ecological process for benthic ecosystems and might be a key factor controlling the spatial distribution of many species in the marine environment, with mud and sand bottom communities recovering from disturbances faster than gravel and reef benthos (Harris, 2012, 2014). Although big storms are known to cause profound modifications

n Clavería s/ * Corresponding author. Departamento de Biología Funcional, C/ Julia n, 33006-Oviedo, Spain. E-mail address: [email protected] (E. Garcia-Vazquez).

in nearshore communities (e.g. Morton, 1988), their effect on local benthic communities is not well known yet, at least in the Bay of Biscay, mainly due to a lack of spatio-temporal baselines for model taxa and species assemblages; this hampers impact assessments in the zone (e.g. Juanes et al., 2007; Puente et al., 2009). In the last years cyclogenesis and storm events are increasing in the region e some are derived from subtropical processes (Liberato et al., 2013) and may be due to global climate change that promotes instability in temperate areas (e.g. Lozano et al., 2004). Storminess is expected to increase the vulnerability of coastal zones in these regions, with more intense erosion and subsequent changes in biodiversity (Lozano et al., 2004). In this study we focused on the effect of the cyclogenesis that affected the Bay of Biscay in the 2013e2014 winter (cyclone Dirk; see for example http://alert.air-worldwide.com/EventSummary. aspx?e¼727&tp¼31&c¼1, last accessed June 2015), taking the top shells Gibbula umbilicalis (da Costa 1778) and Phorcus lineatus (da Costa 1778; previously Osilinus lineatus) as model species. We have

http://dx.doi.org/10.1016/j.marenvres.2015.06.012 0141-1136/© 2015 Elsevier Ltd. All rights reserved.

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

2

chosen top shells because we have a baseline of previous genetic and morphological population data (e.g. Turrero et al., 2014), and because top shells are model species useful for assessing the impact of climate (e.g. Mieszkowska et al., 2006; Hawkins et al., 2009). Rapid alterations in the distributional limits of P. lineatus and G. umbilicalis have been described following climate change in the mid-1980s (Mieszkowska, 2009), and top shells are vulnerable to wave exposure since their sizes change along with this and other factors (e.g. Crothers, 2001; Preston and Roberts, 2007). Following this, we expected them to be sensitive to the effects of the cyclogenesis, which modify different parameters of the environment such as turbidity, nutrients dissolved in the water, intensity of wave action on the coast, etc. (Harris, 2014). Selection against vulnerable size classes, too big or too small for proper anchoring to the substrate under the influence of strong windstorms, would be expected in more exposed beaches, affecting population size distributions and even population genetic diversity if the storms caused very intense mortality. Although the removal of individuals is not necessarily the only effect of these storms, due to the strength of these events in recent years we considered it as one of the factors with most modifying power over these populations. The sequence of events would be: A storm comes to shore. More vulnerable (small and exposed) individuals are swept out. If many individuals are lost from an exposed population, a bottleneck and subsequent genetic drift is expected. These effects are expected to be less intense in less exposed beaches, and therefore differences between exposed (open) and sheltered beaches are expected for both size and population genetic variation (populations from more exposed beaches are expected to be less variable).

visually identified based on morphological traits (see for example Crothers, 2003) and stored in 96% ethanol until genetic analyses were performed. The ethanol was changed twice on consecutive days to improve tissue preservation. 2.2. Prehistoric and contemporary (pre-cyclogenesis) baselines Details of the baselines can be found in Turrero et al. (2014), and can be summarised as follows: remains of the species P. lineatus (formerly named Osilinus lineatus) were found in 6 archaeological sites (Balmori, Coberizas, Cuetu la Mina, La Lloseta, La Riera and Tito Bustillo) in the archives of the Archaeological Museum of Asturias, and the maximum widths of 1106 archaeological top shells were measured. These were found in strata from three different cultural/ technological phases: Solutrean, from ~20000 years ago (20 ka) to ~17 ka; Magdalenian, from ~17 ka to ~11.5 ka; and Epipalaeolithic, from ~11.5 ka to 6 ka (roughly coincident with the beginning of the Holocene epoch). Contemporary shellfish were randomly sampled during winter (December to January) from 2009 to 2011, with average temperatures between 9 and 10  C (comparable to the 10.2  C average for March 2014 in the region; see A.E.MET, 2015 for weather data), from the accessible intertidal level of rocky coast close to the archaeological sites. This area corresponds to the eastern part of the . A sampling area of the present study, between La Griega and Toro total of 135 top shells were collected from six areas, visually identified and stored in 96% ethanol for further genetic identification employing the Barcoding COI gene (Donald et al., 2012). 2.3. Morphometric analysis

2. Materials and methods 2.1. Studied area and post-cyclogenesis sampling The taxonomic nomenclature we use follows that currently accepted in the World Register of Marine Species (WoRMS, Boxshall et al., 2014). The top shell species G. umbilicalis (da Costa 1778) and P. lineatus (da Costa 1778) were chosen as models due to the existence of previous observations and genetic data from the region. The study area was the central coastal part of the Spanish Bay of Biscay, in the region of Asturias (43 400 0000 N/42 530 5000 N to 7120 0000 W/4 320 2000 W; see Fig. 1). Rocky beaches were visited in late March just after the end of the cyclogenesis that occurred during the previous winter (cyclone Dirk in December 2013 and later episodes; see for example http://poleshift.ning.com/profiles/ blogs/north-atlantic-wave-bombs, last accessed June 2015). The type of geological substrate, mainly calcareous, is very similar in all the considered beaches. None of tem is significantly affected by contamination. At least partially sheltered rocky shores were preferred for sampling to minimize wave exposure, although even sheltered beaches were impacted during the cyclogenesis event. This was apparent when sampling the beaches of Andrín and Vidiago (Fig. 1), where no gastropods could be found except for a few G. umbilicalis specimens. Sampling of the two considered species was conducted from three affected (more open) beaches  ), where sand had been removed and more (Otur, Verdicio and Toro rocks than usual were exposed, and three less affected (more sheltered) beaches, where sand was still in place and rock exposure o). had not changed noticeably (Perlora, La Griega and Po At each location, samples were obtained at random (i.e. without selection for size) from the intertidal transect, covering an area of approximately 2000 m2. For reasons of comparable habitat and sampling effort, naked rocks with <10% algae coverage were targeted for sampling. Sampling effort was roughly 10% (one individual collected per ten individuals observed). The samples were

The maximum widths of the shells were measured using a Vernier Caliper (±0.1 mm). For graphical representation, shell measurements were grouped into mean-centred shell width categories every 2.5 mm. 2.4. Genetic analysis DNA extraction was carried out following Estoup et al.’s (1996) resin-based Chelex protocol. The region of the mitochondrial gene cytochrome oxidase I (COI) was PCR-amplified with Applied Biosystem‘s Veriti Thermal Cycler from specimen DNA according to the procedure revised by Geller et al. (2013). The primers jgLCO1490 (TITCIACIAAYCAYAARGAYATTGG) and jgHCO2198 (TAIACYTCIGGRTGICCRAARAAYCA) were used. PCR was prepared with 4 mL 5 PCR buffer, 0.2 mL Taq polymerase (Promega), 2 mM MgCl2, 0.5 mM of each primer, 2 mL of 2.5 mM dNTP and 2 mL of genomic DNA in 20 mL reactions. PCR conditions were an initial denaturing step at 95  C for 5 min; then 35 cycles of 1 min at 95  C, 1 min at 46  C, 1.30 min at 72  C; and a final 7 min at 72  C. PCR products were examined on a 2% agarose gel stained with ethidium bromide. The illustra™ ExoStar™ 1-Step GE Healthcare Life Sciences protocol was applied to the PCR products, that were sequenced employing the BigDye Terminator Cycle Sequencing Kit v3.1 and analysed on a 3130 xl Genetic Analyzer (Applied Biosystems) Automated Sequencer at the Unit of Genetic Analysis of the University of Oviedo. Sequence chromatograms were edited using Seqman Lasergene v7.0.2. (DNASTAR). DNA sequences were treated using different programs. FASTA files were compiled per species after morphological classification was confirmed by comparison of the DNA dataset with reference sequences using the nBLAST program within NCBI (http://blast.stva.ncbi.nlm.nih.gov/). Alignment of COI sequences per species

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

3

Fig. 1. Location of the studied region and the sites mentioned in the text. Beaches in clear grey: 1, Otur (affected); 2, Verdicio (affected); 3, Perlora (less affected); 4, La Griega (less o de Llanes (less affected); 6, Toro  (affected); 7, Andrín (not sampled); 8, Vidiago (not sampled). Archaeological sites in white: A, La Lloseta, Tito Bustillo; B, Coberizas; affected); 5, Po C, Cuetu la Mina, La Riera; D, Balmori.

was performed with the ClustalW program within Mega 5.10 (Tamura et al., 2013). COI sequences were loaded into DnaSP v 5.10 (Librado and Rozas, 2009) to generate a haplotype data file. The haplotypes found for each species were submitted to GenBank (http://www.ncbi.nlm.nih.gov/) for availability in a public repository. The level of sequence polymorphism (genetic diversity) was measured by different parameters: Nh (number of sequence variants or haplotypes), Hd (haplotype diversity, that is, the probability of two randomly chosen haplotypes in the sample being different) and п (nucleotide diversity, or the mean number of differences between all pairs of haplotypes in the sample). They were calculated with ARLEQUIN v3.5 software (Excoffier et al., 2005). Haplotype networks were constructed with NETWORK v4.6.1.2 (www.fluxus-engineering.com). Different haplotypes are represented by circles connected by lines, where the mutations between them are represented. The diameter of each circle is proportional to the frequency of the corresponding haplotype.

order to obtain the same information, graphic and non-parametric comparisons were done between different statuses of exposure (open versus sheltered) and/or periods (Solutrean, Magdalenian, Epipalaeolithic, contemporary baseline in 2009e2011 and postdisturbance in 2014), not taking into account the status of the beaches from the paleontological samples because we did not know the beach of origin for each sample. Statistical analyses of morphological measures were performed with SPSS Statistics software version 17.0.2 and PAST v3.0 (Hammer et al., 2001). The significance of the differences between locations or chronological periods for species distribution was tested using non-parametric contingency Chi-Square. Finally, we compared, by parametric ANOVA two-ways analyses, the effect in P. lineatus (a species for which we had a baseline) with that in G. umbilicalis, to determine if both species were similarly affected or not after the disturbances. 3. Results

2.5. Statistical analyses 3.1. Population genetic diversity For the statistical analyses, samples from each chronological period, and status (affected/less-affected beaches) for postdisturbance samples, were pooled spatially. Dispersal distances for Phorcus populations, that is, their advance into new locations, can reach 1 km in less than one decade (Little et al., 2012), and the potential for population connectivity at this spatial scale has been confirmed by little or null regional population differentiation for top shells (e.g. Keith et al., 2011). Between-population differences at a smaller geographic scale, as is the present case, are therefore not expected. After a normalization test, non-parametric tests were performed to compare the effect of the cyclogenesis on shell size at a short scale (comparing post-disturbance data with the contemporary baseline, affected eopen beaches- and less affected esheltered beaches-, pairwise to check all possible comparisons). We used ManneWhitney tests to compare means and KolmorogoveSmirnov tests to compare the distributions of frequencies. In

PCR amplification of COI gene provided a total of 146 sequences 495-nucleotide long for G. umbilicalis samples collected after cyclone Dirk, 72 from affected beaches and 74 from less-affected locations (Table 1). They were submitted to GenBank, where they are available with accession numbers KP064694eKP064756. For P. lineatus 63 sequences from less-affected sites and 25 from affected locations were obtained, all 524 nucleotide long. Their GenBank accession numbers are KP064757eKP064788. For G. umbilicalis only a few sequences (GenBank accession numbers JN241973eJN241977) were available before the cyclogenesis (Table 1). These individuals were highly variable at this gene with Hd higher than 0.9, and five different haplotypes for seven individuals sampled (Nh/N ¼ 0.71). In contrast, after the disturbance this ratio was Nh/N ¼ 0.437, with 63 different haplotypes for 144 individuals sampled. The individuals sampled from less-affected locations exhibited only slightly higher diversity (both

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

4

Table 1 Genetic diversity at the cytochrome oxidase I gene for top shells Gibbula umbilicalis and Phorcus lineatus at six locations sampled after 2014's cyclogenesis, by location, by exposure status (affected or less-affected) and in the whole region, as well as regional data from before the disturbance. N, sample size; Nh, Hd and p: number of haplotypes, haplotype diversity and nucleotide diversity respectively (standard error given in parenthesis). Location

Otur Verdicio Perlora La Griega o Po  Toro Pre-disturbance Post-disturbance (whole) Affected (more open) Less-affected (sheltered)

Gibbula umbilicalis

Phorcus lineatus

N

Nh

Hd

28 8 42 22 10 36 7 146 72 74

14 5 27 14 8 27 5 63 36 35

0.913 0.857 0.963 0.952 0.956 0.976 0.905 0.957 0.948 0.959

p (0.030) (0.108) (0.017) (0.027) (0.059) (0.015) (0.103) (0.008) (0.015) (0.010)

Hd and p) than those sampled from affected locations (Table 1). The global picture given by the haplotype network (Fig. 2) for Gibbula, with two main lineages and many singletons, does not indicate a preferential loss of haplotypes from exposed versus lessexposed sites. FST values for pairwise comparisons between preand post-disturbance samples were not significant (FST values of 0.06622 with P ¼ 0.06306 and 0.03102 with P ¼ 0.13514, for comparison of pre-cyclone with affected and less-affected postcyclone samples respectively, both not significant). Differences between affected and less-affected post-cyclone samples were not significant (FST ¼ 0.00145 with P ¼ 0.54955).

0.0055 0.0058 0.0085 0.0064 0.0075 0.0079 0.0077 0.0072 0.0068 0.0076

(0.0036) (0.0039) (0.0048) (0.0038) (0.0047) (0.0045) (0.0050) (0.0041) (0.0039) (0.0043)

N

Nh

Hd

21 2 20 22 21 2 24 88 25 63

16 2 7 14 10 2 14 32 18 23

0.962 1.000 0.842 0.944 0.881 1.000 0.938 0.924 0.957 0.913

p (0.030) (0.500) (0.049) (0.029) (0.047) (0.500) (0.030) (0.016) (0.029) (0.019)

0.0098 0.0134 0.0075 0.0097 0.0096 0.0134 0.0095 0.0093 0.0097 0.0093

(0.0055) (0.0143) (0.0044) (0.0055) (0.0054) (0.0143) (0.0053) (0.0051) (0.0054) (0.0051)

Baseline P. lineatus samples obtained before the cyclogenesis (N ¼ 24), with GenBank accession numbers JN241979eJN241991 (formerly named O. lineatus following older taxonomic nomenclature), were moderately variable with Hd and p values of 0.938 and 0.009 respectively (Table 1), and 14 haplotypes for 24 individuals (Nh/N ¼ 0.58). After the disturbance, a total of 32 haplotypes were found for a total of 88 samples (Nh/N ¼ 0.36), which is a considerable reduction in general diversity. Differences between affected and less-affected locations for haplotype and nucleotide diversity  and were not consistent across sites; two affected localities, Toro Verdicio, have only two sequences. The haplotype network (Fig. 3)

Fig. 2. Haplotype network constructed from Gibbula umbilicalis cytochrome oxidase I gene sequences. Each circle represents a haplotype, and their diameters are proportional to their frequency. Black, pre-disturbance samples; white and grey, post-cyclone samples from more exposed and less exposed locations respectively.

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

5

Fig. 3. Haplotype network constructed from Phorcus lineatus cytochrome oxidase I gene sequences. Each circle represents a haplotype, and their diameters are proportional to their frequency. Black, pre-disturbance samples; white and grey, post-cyclone samples from affected and less-affected locations respectively.

exhibited a complex shape with different lineages, somewhat simpler than the network obtained for G. umbilicalis (Fig. 2) but with the similar feature of showing less shared haplotypes than singletons. For changes in haplotype frequencies between pre- and post-cyclone samples, FST value was not significant (FST ¼ 0.0196 with P ¼ 0.099). As expected, FST between affected and less-affected post-cyclone samples was also not significant (FST ¼ 0.0203 with P ¼ 0.946).

3.2. Top shell size A total of 455 G. umbilicalis and 384 P. lineatus specimens were sampled and measured after cyclone Dirk and subsequent episodes affected the region (Table 2A). A size baseline for P. lineatus from the region, sampled before 2014, is available (Turrero et al., 2014). The comparison of mean size between pre- (contemporary baseline) and post-cyclogenesis samples shows that differences are not significant (U ManneWhitney ¼ 24,785, P ¼ 0.803), but a comparison between size class frequencies shows their differences are highly significant (Z KolmorogoveSmirnov ¼ 2.074, P ¼ 0.000). As a whole, post-disturbance samples exhibited a higher proportion of large individuals than pre-disturbance samples (Fig. 4). All the comparisons between more exposed (open) versus less exposed (sheltered) beaches within periods and between periods were significant too, as can be seen on Table 2B. In the pre-cyclogenesis samples a wider variety was found, with extensive differences between open and sheltered beaches. These differences were reduced after the storms due to a clear decrease of the frequency of smaller top shells in the more sheltered beaches, and a decrease of the intermediate size classes in the more open beaches (Fig. 4), resulting in a greater similarity between the different beaches.

Table 2 Comparison of top shell sizes from different sampling locations in the Bay of Biscay after 2014's cyclogenesis. A: Mean size (in mm) ± standard error (N) of the considered species in the six sampling locations (Otur, Verdicio, Perlora, La Griega, o and Toro ), listed from west to east. B: Non-parametric analyses performed on Po Phorcus lineatus data to compare the post-disturbance samples with the predisturbance baseline, taking into account the exposure status of the sampling sites (exposed or less exposed, affected and less affected by the cyclogenesis respectively). Location

Status

Gibbula umbilicalis

A) Otur Verdicio Perlora La Griega o Po  Toro All exposed All less-exposed

Exposed Exposed Less-exposed Less-exposed Less-exposed Exposed Exposed Less-exposed

14.81 12.59 10.88 12.12 11.86 12.95 13.10 11.11

Comparisons B) Pre-disturbance vs. postdisturbance Affected beaches (pre vs. post) Less-affected beaches (pre vs. post)

± ± ± ± ± ± ± ±

0.92 1.54 1.33 2.55 1.73 1.26 1.59 1.61

(((40) (115) (188) (27) (18) (67) (222) (233)

ManneWhitney test

U ¼ 24,785, P ¼ 0.803 U ¼ 5544, P ¼ 0.005 U ¼ 1004.500, P ¼ 0.000 Post-disturbance (affected vs. less- U ¼ 11,279, affected) P ¼ 0.000 Pre-disturbance (affected vs. less- U ¼ 165.500, affected) P ¼ 0.000

Phorcus lineatus 14.94 15.54 13.92 12.66 13.13 15.00 15.31 13.04

± ± ± ± ± ± ± ±

4.23 2.40 3.15 3.67 3.36 1.87 3.12 3.49

(((54) (98) (39) (110) (76) (7) (159) (225)

Kolmorogov eSmirnov test Z ¼ 2.074, P ¼ 0.000 Z ¼ 2.031, P ¼ 0.001 Z ¼ 4.172, P ¼ 0.000 Z ¼ 2.592, P ¼ 0.000 Z ¼ 4.840, P ¼ 0.000

Compared with the older (prehistoric) samples, postcyclogenesis top shells were more comparable to our

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

6

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

considered as a single population the difference in size with Epipalaeolithic samples is not statistically significant (t ¼ 1.722, P ¼ 0.085 for samples with unequal variance). Taking into account the two species, top shell samples from locations more affected by the cyclogenesis were larger than those sampled from less-affected beaches, although for the same type of beach P. lineatus were always bigger than G. umbilicalis (Table 2A). A two-way ANOVA, performed to discriminate if the two species underwent the same effect after the cyclogenesis, was highly significant for differences in size between species as well as between affected and less-affected beaches (Table 3). The interaction between both factors was not significant (F ¼ 2.039, P ¼ 0.154, Table 3), indicating that the size difference between samples from affected and less-affected beaches within species was similar for the two species. The distribution in size classes (Fig. 6) shows the differences revealed by the ANOVA. The distribution obtained from samples of sheltered beaches contains a much higher proportion of top shells from small size classes than that from samples of more open beaches, for the two species. It is also evident that G. umbilicalis was smaller than P. lineatus, and that the shape of the size class distribution for open and sheltered beaches was similar within species, with a higher variance in P. lineatus. Fig. 4. Size classes of Phorcus lineatus before and after the cyclogenesis Dirk in open (above) and sheltered (below) beaches, presented as proportion of each size class in cm. Post-disturbance, N ¼ 225; pre-disturbance, N ¼ 135.

Epipalaeolithic samples than to the samples collected before the cyclogenesis (Fig. 5). The respective size means for Epipalaeolithic, pre- and post-disturbance samples are 14.39, 13.52 and 13.98 mm. The difference in mean size between Epipalaeolithic and postdisturbance 2014 samples is significant for both exposed (mean size 15.31; t ¼ 3.214, P ¼ 0.0014 for samples with equal variance) and less-exposed (mean size 13.04; t ¼ 4.815, P ¼ 2.06  106 for samples with unequal variance) sites. These differences are opposite in sign because top shells from exposed beaches were bigger than those found in Epipalaeolithic sites, whereas those from less exposed locations were smaller. When all post-cyclone samples are

3.3. Changes in top shell assemblages It is important to take into account the spatial component of species assemblages, because the relative abundance of Gastropod species varies along the studied coast due to a shift in environ~ as, in the centre of the region mental conditions around Cape Pen ~ oz-Colmenero et al., 2012; Turrero et al., 2012). Considering (Mun only the locations where data on top shell species composition were available before 2014 e that is, the eastern part of the studied , La Griega and Po  o) e, it can be noted that the assemarea (Toro blage changed (Fig. 7), with a significant increase in G. umbilicalis presence, which was scarce before the cyclogenesis (Chi-square of 10.98 for 1 degree of freedom, P ¼ 0.001). This species was clearly dominant in the more exposed beaches in 2014 (Table 1A). Moreover, in the two beaches visited in 2014 after the cyclogenesis where sampling was discarded (Vidiago and Andrín), only a few G. umbilicalis were found, and no P. lineatus. 4. Discussion The results obtained in this study reveal alterations in coastal top shell populations from the Bay of Biscay following exposure to cyclone Dirk. Small individuals were less abundant after the disturbance for the two model species studied, G. umbilicalis and P. lineatus, and regional genetic diversity (at COI DNA sequences) seems to have been reduced although haplotype frequencies did not change significantly. These effects are consistent with a sweeping effect of the cyclone at both morphological and genetic levels. Although there is no simple explanation linking the reduction of

Table 3 Two-way ANOVA comparing the mean sizes of the studied top shell populations. Status, open and sheltered beaches; Species, Gibbula umbilicalis and Phorcus lineatus.

Fig. 5. Size distributions (proportion of each size class in cm) of Phorcus lineatus populations inhabiting the studied areas in different chronological moments. Solutrean, N ¼ 74; Magdalenian, N ¼ 557; Epipalaeolithic, N ¼ 358; pre-disturbance, N ¼ 135; post-disturbance, N ¼ 225.

Exposure status: Species: Interaction: Within: Total:

Sum of squares

df

Mean square

F

p (same)

804.406 755.616 13.2923 5444.26 7132.8

1 1 1 835 838

804.406 755.616 13.2923 6.52007

123.4 115.9 2.039

7.87E-27 2.13E-25 0.1537

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

Fig. 6. Size distribution (proportion of each size class in cm) for the populations of Gibbula umbilicalis (Gu, green) and Phorcus lineatus (Pl, blue) inhabiting affected (dashed line) and less-affected (solid line) locations. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

7

locations, which on the other hand are also more exposed to waves in normal conditions. The lack of small size classes has been often considered an indication of sporadic settlement events (e.g. Lewis, 1986; Zacherl et al., 2003; Lima et al., 2006). However, the results found in this study suggest an alternative or complementary explanation in areas vulnerable to sediment disturbances, since a lack of small top shells could reflect a recent sweeping due to big storms or cyclones. From our results we can see that these climatic events acted in the direction of homogenizing the P. lineatus population, removing the small individuals from both more exposed and sheltered beaches refugees (Fig. 4A). The great strength of the waves brought on by these cyclones cancels the effect of refuge for juveniles of these species that the sheltered beaches provide in normal weather conditions. Related to this, some discussion on recruitment in these two species is necessary. Reproduction and development of the gonads for both species usually happens in the summer months (Mieszkowska et al., 2006, 2007), with the peak depending on the temperature at the area. The increase in temperature resulting from global warming has led to successful recruitment for these species in the North Atlantic (Mieszkowska et al., 2007). Because of this, successful recruitment is expected unless some disturbance modifies the usual conditions. One such disturbance would be the cyclogenesis processes in our study, which would remove juveniles already settled and would drag away those that were trying to settle, preventing them from adhering to the substrate. The autumn

o, Toro ) before and after 2014's cyclogenesis, Fig. 7. Gibbula umbilicalis and Phorcus lineatus assemblages found in the intertidal area of the eastern sampling points (La Griega, Po presented as the proportion of each species. Data for 2014 affected and less-affected beaches are also presented separately.

small size classes and cyclone Dirk, they do seem to be connected since this reduction occurred in both species regardless of their different species-specific mean size. Top shell juveniles attach relatively weakly to the substrate (e.g. Hayakawa et al., 2008); it could be hypothesized that small individuals exhibit less adhesiveness to the substrate than bigger ones. They would therefore be more vulnerable to sudden intense windstorms, which would involve massive movements of sediments, carrying sand and other particles that would especially affect smaller, more fragile individuals. However, very small individuals refuged inside rock crevices and small holes would be more protected against wave sweeping, being more difficult to remove. This would explain the apparent removal of intermediate size classes in the more open

months are usually the months of higher population density, corresponding to the later stages of settlement in the intertidal (Ahmedou Salem et al., 2014). The heavy storms the coast suffered would have impeded the new individuals to thrive. On the other hand, the post-storm recovery (or the resistance to sweeping) of G. umbilicalis and P. lineatus found in this study was better for the former, especially in open locations. This is consistent with the interpretation given above, based on the removal of small but intermediate size classes, i.e. the small individuals that were not smaller enough for hiding in crevices and intricate holes. This would be reflected in higher standard deviation for post-than for pre-disturbance samples (very small and big individuals would survive). Gibbula, which is generally smaller than Phorcus,

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

8

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun

recovered much better after the storm (Fig. 6), indicating that they resisted better the cyclogenesis impact than the other bigger top shell. Regarding genetic diversity, the results of this study are a good example of bottleneck effects. Top shells reproduce once a year in mid-summer and live for several years (e.g. Crothers, 2003). Since population sizes are large, they have long dispersal capacity and spatial connectivity is generally ensured (e.g. Keith et al., 2011; Little et al., 2012), pronounced gene drift and thus genetic differences between consecutive generations are not expected. Removing a part of the population, likely the younger generation/s, would reduce a part of the general diversity because some singletons would be lost, but the population's genetic pool would not change substantially. Moreover, replenishment of the lost variation by immigration from neighbouring, less exposed, beaches is expected in a short time thanks to the top shells' dispersal capacity. Top shells would function as metapopulations and genetic diversity losses in a part of their distribution would be rapidly compensated by high gene flow. These dynamics would enable top shells to colonize neighbouring environments; actually, these species and other intertidal gastropods are in expansion under the current climate change (e.g. Hellberg et al., 2001; Hawkins et al., 2008, ~ oz-Colmenero et al., 2012; Rubal et al., 2014). 2009; Mun We have found a significant change in size for both species under study that was dependent on the exposure to the cyclogenesis, which promoted spatial heterogeneity in the small region studied. Moreover, in some locations (Vidiago, Andrín) almost all top shells were swept away by the cyclogenesis, with only a few individuals of one species left. Patchy distributions of a species might therefore be explained from different exposure to disturbance events. Our results support the inclusion of disturbance regime concepts with other biophysical variables that define the fundamental niches of marine species for predictive habitat modelling, as proposed by Harris (2014). Under normal conditions, exposure of the beaches would be an important factor for intertidal populations, varying the intensity at which the sea affects them, but under conditions of hard climatic disturbances, the effect of homogenization of the affected coasts should be taken into account e here represented by a smaller range of sizes and the loss of genetic variation. Instability due to intense disturbances such as cyclogeneses would produce rapid changes in size in top shell populations, contributing to increase the average size in the long term by sporadically sweeping smaller size classes (when disturbances are produced, which may vary depending on the particular climatic regime of each region). This has enormous importance for comparative studies at large temporal scales: if samples from only one or a few years are considered as representative of modern populations, they may reflect just a transient status that may change significantly after a single disturbance event. Comparing samples from one or two undisturbed years with chronological periods that include at least various cyclonic events likely under represents the real evolution and status of the studied biota. For instance, known periods of global climatic change such as the Mediaeval Warm Period or the Little Ice Age (e.g. see Desprat et al., 2003; Eiríksson et al., 2006) are likely to have produced one or several such events. The inclusion of samples from years and zones affected by disturbances in regional baselines is therefore highly recommended. The results of this study also help to explain, at least partially, the scarcity of G. umbilicalis in archaeological sites of this region where Patella, Phorcus and Littorina were by far the most intensely rrez-Zugasti, 2011; exploited Gastropods (e.g. Ortea, 1986; Gutie Turrero et al., 2014). Ortea (1986) explained the almost total absence of this species from Upper Palaeolithic remains due to the

difficult extraction of the whole body from the shell with a pointed object like a pin (the Upper Palaeolithic is represented in our study by the Solutrean and Magdalenian periods). However, the shells can be broken as easily as those of the widely consumed Phorcus. Preference for bigger shellfish (selection for size, as seen for rrez-Zugasti, 2011 and Turrero et al., 2012) would example in Gutie also explain the differential Palaeolithic exploitation of these two species. Since under the same conditions Gibbula grows smaller than Phorcus, the latter species would be preferred for harvesting. 5. Conclusions The significantly bigger size of G. umbilicalis and P. lineatus specimens from beaches affected by 2014's cyclogenesis in the Bay of Biscay when compared to less-affected sampling points suggests the cyclogenesis had a sweeping effect over small individuals. A significant increase in the average regional size after the cyclogenesis compensated the apparent size decline of P. lineatus after the Upper Palaeolithic, suggesting that reference baselines should take into account disturbances when evaluating the long-term evolution of intertidal communities. On the other hand, significant changes in COI haplotype frequencies were not found despite the reduction of haplotype variation. High connectivity between populations and a metapopulation structuring of top shells would explain their high capacity for the recovery of population genetic variability. Acknowledgements The authors are grateful to the Museo de Arqueología de Asturias for granting permission to study part of their collections, and to an anonymous reviewer for his/her helpful suggestions for ~ oz-Colmenero holds a National improving data analysis. M. Mun Spanish Grant (reference AP-2010-5211). This study has been supported by the Regional Government of Asturias (grant number SV-PA-13-ECOEMP-41; GRUPIN14-093) and the Spanish National Project MINECO CGL2013-42415-R. This is a contribution from the Marine Observatory of Asturias (Spain). References A.E.MET, 2015. Weather reports from the Spanish Meteorological Agency (Agencia ~ ola de Meteorología). Data sets available at: http://www.aemet.es/es/ Espan serviciosclimaticos/vigilancia_clima/resumenes?w¼1&k¼ast (Last accessed June 2015). Ahmedou Salem, M.V., van der Geest, M., Piersma, T., Saoud, Y., van Gils, J.A., 2014. Seasonal changes in mollusc abundance in a tropical intertidal ecosystem, Banc d’Arguin (Mauritania): testing the ‘depletion by shorebirds’ hypothesis. Estuar. Coast. Shelf Sci. 136, 26e34. Boxshall, G.A., Mees, J., Costello, M.J., Hernandez, F., Gofas, S., Hoeksema, B.W., et al., 2014. World Register of Marine Species. Available from: http://www. marinespecies.org. at VLIZ. accessed 31.10.14. Crothers, J.H., 2001. Common top shells: an introduction to the biology of Osilinus lineatus with notes on other species in the genus. Field Stud. 10, 115e160. Crothers, J.H., 2003. Rocky shore snails as material for projects (with a key for their identification). Field Stud. 10, 601e634. nchez Gon ~ i, M.F., Loutre, M.-F., 2003. Revealing climatic variability of Desprat, S., Sa the last three millennia in northwestern Iberia using pollen influx data. Earth Planet. Sci. Lett. 213, 63e78.  Donald, K.M., Preston, J., Williams, S.T., Reid, D.G., Winter, D., Alvarez, R., Buge, B., Hawkins, S.J., Templado, J., Spencer, H.G., 2012. Phylogenetic relationships elucidate colonization patterns in the intertidal grazers Osilinus Philippi, 1847 and Phorcus Risso, 1826 (Gastropoda: Trochidae) in the northeastern Atlantic Ocean and Mediterranean Sea. Mol. Phylogenetics Evol. 62, 35e45.  nsdo  ttir, H.B., Cage, A.G., Gudmunsdo  ttir, E.R., KlitgaardEiríksson, J., Bartels-Jo Kristensen, D., Marret, F., Rodrigues, T., Abrantes, F., Austin, W.E.N., Jiang, H., Knudsen, K.-L., Sejrup, H.-P., 2006. Variability of the North Atlantic current during the last 2000 years based on shelf bottom water and sea surface temperatures along an open ocean ⁄ shallow marine transect in Western Europe. Holocene 16, 1017e1029. r, C.R., Perrot, E., Chourrout, D., 1996. Rapid one tube DNA Estoup, A., Largiade extraction for reliable PCR detection of fish polymorphic markers and

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012

~ oz-Colmenero et al. / Marine Environmental Research xxx (2015) 1e9 M. Mun transgenes. Mol. Mar. Biol. Biotechnol. 5, 295e298. Excoffier, L., Laval, G., Schneider, S., 2005. Arlequin ver. 3.0: an integrated software package for population genetics data analysis. Evol. Bioinforma. Online 1, 47e50. Geller, J., Meyer, C., Parker, M., Hawk, H., 2013. Redesign of PCR primers for mitochondrial cytochrome c oxidase subunit I for marine invertebrates and application in all-taxa biotic surveys. Mol. Ecol. Resour. 13 (5), 851e861. rrez-Zugasti, I., 2011. Coastal resource intensification across the PleistoceneGutie Holocene transition in Northern Spain: evidence from shell size and age distributions of marine gastropods. Quat. Int. 244, 54e66. Hammer, Ø., Harper, D.A.T., Ryan, P.D., 2001. PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4 (1), 9. http://palaeo-electronica.org/2001_1/past/issue1_01.htm. Harris, P.T., 2012. On seabed disturbance, marine ecological succession and applications for environmental management: a physical sedimentological perspective. In: Li, M., Sherwood, C., Hill, P. (Eds.), Sediments, Morphology and Sedimentary Processes on Continental Shelves, International Association of Sedimentologists Special Publication, 44, pp. 387e404 (Oxford). Harris, P.T., 2014. Shelf and deep-sea sedimentary environments and physical benthic disturbance regimes: a review and synthesis. Mar. Geol. 353, 169e184. Hawkins, S.J., Moore, P.J., Burrows, M.T., Poloczanska, E., Miezskowska, N., Herbert, R.J.H., Jenkins, S.R., Thompson, R.C., Genner, M.J., Southward, A.J., 2008. Complex interactions in a rapidly changing world: responses of rocky shore communities to recent climate change. Clim. Res. 37, 123e133. Hawkins, S.J., Sugden, H., Mieszkowska, N., Moore, P., Poloczanska, E., Leaper, R., Herbert, R.J.H., Genner, M.J., Moschella, P., Thompson, R., Jenkins, S., Southward, A.J., Burrows, M.T., 2009. Consequences of climate-driven biodiversity changes for ecosystem functioning of North European rocky shores. Mar. Ecol. Prog. Ser. 396, 245e259. Hayakawa, J., Kawamura, T., Ohashi, S., Horii, T., Watanabe, Y., 2008. Habitat selection of Japanese top shell (Turbo cornutus) on articulated coralline algae; combination of preferences in settlement and post-settlement stage. J. Exp. Mar. Biol. Ecol. 363, 118e123. Hellberg, M.E., Balch, D.P., Roy, K., 2001. Climate-Driven range expansion and morphological evolution in a Marine gastropod. Science 292, 1707e1710.  Juanes, J.A., Puente, A., Revilla, J.A., Alvarez, C., García, A., Medina, R., Castanedo, S., Morante, L., Gonz alez, S., García-Castrillo, G., 2007. The Prestige oil spill in Cantabria (Bay of biscay). Part II. Environmental assessment and monitoring of coastal ecosystems. J. Coast. Res. 23, 978e992. Keith, S.A., Herbert, R.J.H., Norton, P.A., Hawkins, S.J., Newton, A.C., 2011. Individualistic species limitations of climate-induced range expansions generated by meso-scale dispersal barriers. Divers. Distrib. 17, 275e286. Lewis, R., 1986. Latitudinal trends in reproduction, recruitment and populations characteristics of some rocky littoral molluscs and cirripedes. Hydrobiologia 142, 1e13. n ~ ez, P., Yuen, D., Trigo, I.F., Liberato, M.L.R., Pinto, J.G., Trigo, R.M., Ludwig, P., Ordo 2013. Explosive development of winter storm Xynthia over the subtropical North Atlantic Ocean. Nat. Hazards Earth Syst. Sci. 13, 2239e2251. Librado, P., Rozas, J., 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics 25, 1451e1452.

9

Lima, F.P., Queiroz, N., Ribeiro, P.A., Hawkins, S.J., Santos, A.M., 2006. Recent changes in the distribution of a marine gastropod, Patella rustica Linnaeus, 1758, and their relationship to unusual climatic events. J. Biogeogr. 33, 812e822. Little, C., Graham, M., Pilling, G.M., Trowbridge, C.D., Stirling, P., 2012. Osilinus lineatus: incursion and proliferation in Lough Hyne Marine Reserve, SW Ireland. Am. Malacol. Bull. 30 (1), 135e145. Lozano, I., Devoy, R.J.N., May, W., Andersen, U., 2004. Storminess and vulnerability along the Atlantic coastlines of Europe: analysis of storm records and of a greenhouse gases induced climate scenario. Mar. Geol. 210, 205e225. Mieszkowska, N., 2009. Climate Change: Observed Impacts on Planet Earth. In: Intertidal Indicators of Climate and Global Change. Elsevier B.V., pp. 281e296 (Chapter 15). Mieszkowska, N., Hawkins, S.J., Burrows, M.T., Kendall, M.A., 2007. Long-term changes in the geographic distribution and population structures of Osilinus lineatus (Gastropoda: Trochidae) in Britain and Ireland. J. Mar. Biol. Assoc. U. K. 87, 537e545. Mieszkowska, N., Kendall, M., Hawkins, S., Leaper, R., Williamson, P., HardmanMountford, N., Southward, A., 2006. Changes in the range of some common rocky shore species in Britain e a response to climate change? Hydrobiologia 555, 241e251. Morton, R.A., 1988. Nearshore responses to great storms. In: Clifton, H.E. (Ed.), Sedimentologic Consequences of Convulsive Geologic Events. Geological Society of America, pp. 1e22. ~ oz-Colmenero, A.M., Turrero, P., Horreo, J.L., Garcia-Vazquez, E., 2012. Evolution Mun of limpet assemblages driven by environmental changes and harvesting in North Iberia. Mar. Ecol. Prog. Ser. 466, 121e131. Ortea, J., 1986. The malacology of La riera cave. In: Straus, L.G., Clark, G.A. (Eds.), La Riera Cave. Stone Age Hunter-gatherers in Northern Spain, Anthropol. Res. Pap, 36. Arizona State University, Tempe, pp. 289e298. Preston, S.J., Roberts, D., 2007. Variation in shell morphology of Calliostoma zizyphinum (Gastropoda: Trochidae). J. Molluscan Stud. 73 (1), 101e104. Puente, A., Juanes, J.A., Calderon, G., Echavarri-Erasun, B., Garcia, A., GarciaCastrillo, G., 2009. Medium-term assessment of the effects of the Prestige oil spill on estuarine benthic communities in Cantabria (Northern Spain, Bay of Biscay). Mar. Pollut. Bull. 58, 487e495. Rubal, M., Veiga, P., Moreira, J., Sousa-Pinto, I., 2014. The gastropod Phorcus sauciatus (Koch, 1845) along the north-west Iberian Peninsula: filling historical gaps. Helgol. Mar. Res. 68, 169e177. Tamura, K., Stecher, G., Peterson, D., Filipski, A., Kumar, S., 2013. MEGA 6: molecular evolutionary genetics analysis version 6.0. Mol. Biol. Evol. 30, 2725e2729. ~ oz-Colmenero, A.M., Pola, I.G., Arbizu, M., Garcia-Vazquez, E., 2012. Turrero, P., Mun Morphological, demographic and genetic traces of upper Palaeolithic human impact on limpet assemblages in North Iberia. J. Quat. Sci. 27 (3), 244e253. ~ oz-Colmenero, A.M., Prado, A., Garcia-Vazquez, E., 2014. Long-term Turrero, P., Mun impacts of human harvesting on shellfish: North Iberian top shells and limpets from the upper Palaeolithic to the present. J. Mar. Syst. 139, 51e57. Zacherl, D., Gaines, S.D., Lonhart, S.I., 2003. The limits to biogeographical distributions: insights from the northward range extension of the marine snail, Kelletia kelletii (Forbes, 1852). J. Biogeogr. 30, 913e924.

~ oz-Colmenero, M., et al., Response of top shell assemblages to cyclogenesis disturbances. A case study in Please cite this article in press as: Mun the Bay of Biscay, Marine Environmental Research (2015), http://dx.doi.org/10.1016/j.marenvres.2015.06.012