Brown algae as heavy metals and nitrogen biomonitors of coastal areas: a review of protocols

TESE DE DOUTORAMENTO BROWN ALGAE AS HEAVY METALS AND NITROGEN BIOMONITORS OF COASTAL AREAS: A REVIEW OF PROTOCOLS Rita García Seoane ESCOLA DE DOUTORAMENTO INTERNACIONAL PROGRAMA DE DOUTORAMENTO EN MEDIO AMBIENTE E RECURSOS NATURAIS SANTIAGO DE COMPOSTELA 2019 DECLARACIÓN DO AUTOR/A DA TESE Brown algae as heavy metals and nitrogen biomonitors of coastal areas: a review of protocols Dna. Rita García Seoane Presento a miña tese, seguindo o procedemento axeitado ao Regulamento, e declaro que: 1) A tese abarca os resultados da elaboración do meu traballo. 2) De selo caso, na tese faise referencia ás colaboracións que tivo este traballo. 3) A tese é a versión definitiva presentada para a súa defensa e coincide coa versión enviada en formato electrónico. 4) Confirmo que a tese non incorre en ningún tipo de plaxio doutros autores nin de traballos presentados por min para a obtención doutros títulos. En Santiago de Compostela, Novembro de 2019. Asdo. Rita García Seoane AUTORIZACIÓN DO DIRECTOR / TITOR DA TESE Brown algae as heavy metals and nitrogen biomonitors of coastal areas: a review of protocols D. Jesús Ramón Aboal Viñas Dna. María Teresa Boquete Seoane INFORMA/N: Que a presente tese, correspóndese co traballo realizado por Dna. Rita García Seoane , baixo a nosa dirección, e a utorizamos a súa presentación, considerando que reúne os r equisitos esixidos no R egulamento de Estudos de Doutoramento da USC, e que como directores desta non incorre nas causas de abstención establecidas na Lei 40/2015. En Santiago de Compostela, Novembro de 2019. Asdo. Asdo. D. Jesús Ramón Aboal Viñas Dna. María Teresa Boquete Seoane To carry out this Doctoral Thesis, Rita García Seoane was supported by a predoctoral fellowship from the Ministerio de Ciencia, Innovación y Universidades (FPU14/02015) (Spanish Government). During the development of this Doctoral Thesis, two research stays were done, at the Department of Integrative Biology (University of South Florida, USA) in 2017, and at the Marine Biophysics Unit (Okinawa Institute of Science and Technology Graduate University, Japón) in 2018. Both research stays were funded by the Ministerio de Ciencia, Innovación y Universidades (Spanish Government). The doctoral candidate declares no conflicts of interest related to her Thesis. The articles presented in this Thesis have the signed consent of all authors who have participated in them. A mis padres approach. Heavy metals as Cd, Cu, Hg, Pb and Zn, which are persistent, toxic, and liable to bioaccumulate in organisms and transfer to the trophic chains, are also included to provide a more realistic estimate of the impact of these bioavailable elements in the aquatic environment. Although the analysis of marine macroalgae is not required under the current legislation, these organisms have been widely used as biomonitors of marine pollution (including heavy metals, organic pollutants or radioactive elements among others) since the middle of the 20th century. Brown algae have demonstrated to strongly bind a variety of metal ions, making them one of the most widely used bioabsorbents (Volesky & Holan, 1995; Davis et al., 2003). Species such as Fucus vesiculosus, F. spiralis or Ascophyllum nodosum have been routinely used in metal pollution monitoring programmes around the world (García-Seoane et al., 2018a,b). The use of macroalgae offers important advantages for effective biomonitoring, when compared to other marine organisms included in some programmes, as e.g. fishes or invertebrates: i) they are sessile, easily identified, collected and transplanted, ii) they are widely distributed and available all year round, iii) they have a long life span, iv) they present wide ranges of environmental tolerance, surviving in highly adverse and polluted environmental conditions, v) the concentrations of pollutants in their tissues are strongly correlated with the concentration in seawater and sediments, etc. (e.g. Haug et al., 1974; Phillips, 1980, 1990). One of the main reasons for the non-inclusion of macroalgae in environmental monitoring programmes may be the lack of standardized protocols describing how this technique needs to be applied. As it will be pointed out in Chapters I and VI of this dissertation, researchers are currently using a wide variety of protocols differing in aspects such as the species used, or the way in which the algae are collected, processed and analyzed. These variations obey mostly to practical and economic reasons rather than scientific criteria. The absence of protocols based on such criteria limits the interpretation of the results obtained and hampers the comparison of results between studies. Additionally, the scarcity of studies addressing methodological issues also reveals a general lack of concern for GENERAL INTRODUCTION 17 the standardization of the protocol. For this reason, in this dissertation we also addressed some basic aspects of the methodology of the technique to contribute to establish the use of macroalgae in extensive monitoring studies regulated by legislation. The structure of this Thesis responds to the research gaps identified in the previous paragraph. First of all, we performed a literature review of all the studies using macroalgae as biomonitors of marine pollution, including those using both passive (collection of specimens growing naturally in the area of interest) and active biomonitoring (transplanting individuals from their natural areas to the study sites). Based on these reviews, we proposed a more standardized protocol that served as the basis for the other studies included in this Research Thesis. Some of the most important methodological aspects of the technique for which there was no consensus recommendation, or which have never been studied, were addressed by means of experimental studies. Because the use of native algae has been preferred over the use of transplanted algae, our research efforts focused on optimizing some methodological aspects of the passive biomonitoring techniques: i) number of subsamples and subsample collection strategy to adequately represent the intra-site variability in concentrations (Chapter III), ii) number of samples to significantly differentiate the levels of pollution between sampling sites (Chapter IV), iii) collection period of samples to better represent temporal variation in concentrations (Chapter V), and iv) selection of material for analysis in order to minimize differences in concentrations resulting from tissue variation (Chapter V). In addition, the suitability of macroalgae for assessing natural levels of pollution in the marine environment and detecting sources of pollution was studied in Chapter II. Finally, a study using transplantation techniques was carried out in Chapter VII to assess the possible variation in the uptake capacity of algae under chronic metal exposure. The species selected for carrying out the experimental studies that make up this Doctoral Thesis is Fucus vesiculosus (Linnaeus 1753) because: i) it is the species of brown algae most widely studied in biomonitoring of water quality (García-Seoane et al., 2018b), ii) it meets the characteristics to be a good biomonitor (Martin et al., 1997), and iii) it is widely spread throughout the Galician coast. The bladder-wrack, as it is commonly known, is a perennial brown macroalgae that belongs to the Phylum Ochrophyta, Class Phaeophyceae, Order Fucales, and Genus Fucus (Strasburger et al., 2008). The etymology ‘vesiculosus’ derives from the characteristic air bladders on both sides of midrib that confer buoyancy to the thalli (Fig. 1B). Although this is the main distinctive feature of the species, bladderless forms have also been described on more wave exposed shores, even in the Galician coast (Pazó & Romarís, 1979; Bárbara et al., 1995). The species is widely distributed in cold waters at both sides of the Atlantic Ocean, frequent on the Atlantic coasts of Europe (North Sea, Baltic Sea, Greenland, Azores, Canary Islands and Madeira), and North America (from Ellesmere Island and Hudson Bay to North Carolina), but absent in the southern hemisphere (Fig. 1A). Fucus vesiculosus is common on semi-exposed or sheltered shores, and dominant in the mid-intertidal on rocky shores, often with A. nodosum, and occupying a position in the intertidal zonation below the populations of F. spiralis, and in a zone further up from F. serratus (Carlson, 1991; Bárbara et al., 1995). As regards its morphology, this fucoid is characterized by its flattened, and pseudo-dichotomously branched thallus with apical growth crisscrossed by a midrib. The thallus can reach 100 cm in length and it is fixed to the rocky substrate by means of a conical disc or holdfast (van den Hoek et al., 1995) (Fig. 1B). As all members of the genus Fucus, the life cycle of F. vesiculosus is oogamous diplont with sexual zygotes (South & Whittick, 1987; van den Hoek et al., 1995; Serrão et al., 1999). Unlike other species, like for example F. spiralis which is monoecious, F. vesiculosus is dioecious, with separate male and female plants. During the fertile season the gametes are produced by meiosis and multiplied by mitosis in the gametangia from male plants (antheridia) and from female plants (oogonia). Antheridia and oogonia are formed in reproductive structures, called receptacles, that develop apically at the tips of the branches (Fig. 1C, 1D). Once mature, haploid gametes (spermatozoids and egg cells) are GENERAL INTRODUCTION 19 released, and diploid zygotes are produced by external fertilization. After fertilization, zygotes attach to the substrate and grow into a new diploid gametophyte (van den Hoek et al., 1995; Serrão et al., 1996; Ladah et al., 2003, 2008). Although sexual reproduction is the most common strategy in fucoids, asexual propagation by clonal reproduction (i.e. detachment of adventitious branches from the parental plant that act as vegetative propagules) has been described in species of the genus Fucus, including F. vesiculosus (Bergström et al., 2005; Tatarenkov et al., 2005). Fig. 1. A) Distribution of Fucus vesiculosus in the northern hemisphere. B) Fucus vesiculosus thalli attached to substrate. Detailed images of reproductive structures of F. vesiculosus, antheridia (C) and oogonia (D). Images obtained from www.aphotomarine.com (A), www.bioweb.uwlax.edu (B), and www.seaweed.ie (C, D). General Objectives GENERAL OBJECTIVES 23 The overall aim of this PhD research was to evaluate the potential of marine macroalgae to monitor heavy metals, metalloids and nitrogen in coastal environments, and to propose a scientifically based protocol for the application of this technique. For this purpose, the information presented in this dissertation is divided into two parts, each consisting of several chapters. Part I (Chapters I to V) encompasses the main body of this work and is dedicated to the use of native macroalgae as biomonitors of pollution (passive monitoring). Part II (Chapters VI and VII) focuses on the use of transplants of macroalgae in biomonitoring studies (active monitoring), as an alternative to the use of native algae. The specific objectives of each chapter are summarized as follows: 1. To carry out a critical review of the methodology used in studies concerning pollution biomonitoring in coastal environments worldwide, and involving the use of native marine macroalgae (passive biomonitoring). The specific aims of this review were: to establish the state of the art of the methodology, to identify its degree of harmonization, and to propose an updated protocol for monitoring pollutants with these organisms (Chapter I). 2. To study the potential use of the brown macroalgae Fucus vesiculosus to characterize the natural range of variability in the δ15N signal, and concentrations of N and Hg in coastal areas not affected by local sources of pollution, and to detect small scale sources of contamination. To achieve this objective, samples of the species were collected in more than 150 sampling sites distributed along the shoreline in 3 study zones affected by different degrees of pollution (Chapter II). 3. To propose a sampling strategy for biomonitoring with F. vesiculosus that encompasses the intra-site variability (local variability), and accounts for the spatial structure in the concentrations of Al, As, Cd, Co, Cr, Cu, Fe, Hg, Ni, Pb, Zn and N, and in the δ15N signal, while optimizing the number of subsamples used to make a composite sample, the sampling effort and costs. For this purpose, 50 subsamples of the algae were collected at random using a regular sampling grid design, in each of 3 sites affected by different levels of pollution (Chapter III). 4. To determine the optimal number of samples of F. vesiculosus needed to detect statistically significant differences in the mean concentrations of Al, As, Cd, Co Cr, Cu, Fe, Hg, Ni, Pb, Zn and N, and in the δ15N signal between 2 sampling sites affected by different levels of pollution (Chapter IV). The experimental set-up is the same as described in the previous chapter. 5. To assess the existence of temporal variation of bioconcentration of heavy metals, metalloids and nitrogen in F. vesiculosus, as well as to study the intra-thallus variability in these elements. For this purpose, the concentrations of Al, As, Cd, Co, Cr, Cu, Fe, Hg, Ni, Pb, Zn, N, and the δ15N values were determined in samples of the species collected monthly over a period of 3 years at 3 sampling sites. The implications that the temporal (intra-annual/seasonal) and intra-thallus variability in the concentrations may have on the results of biomonitoring studies with algae are discussed. In addition, an appropriate sampling design, representative of the existing variability in the tissue contents of elements, is proposed for application in biomonitoring studies. The growth of the species (in terms of the number of dichotomies formed in the thallus through the year) is also considered, and the implications for interpretation of the results of temporal studies of pollutants are also discussed (Chapter V). 6. To evaluate the essential aspects of the methodology and the degree of standardization of the protocols used in studies involving biomonitoring of inorganic pollutants and nutrients in coastal environments worldwide with transplants of marine macroalgae (active biomonitoring). On the basis of the conclusions reached, a scientifically based protocol is proposed for application of the biomonitoring technique with transplanted algae (Chapter VI). 7. To compare the accumulation capacity of metals and metalloids in specimens of F. vesiculosus growing naturally under different pollution scenarios. Thalli of the species were reciprocally transplanted between 2 polluted and 2 unpolluted sites. The differences in the dynamics of uptake/release rates of Al, As, Cd, Co, Cr, Cu, Fe, Hg, Ni, Pb and Zn were GENERAL OBJECTIVES 25 studied after exposure of the thalli in these environments for 90 days (Chapter VII). 33 CHAPTER II Application of macroalgae analysis to assess the natural variability in selected pollution concentrations (N and Hg), and to detect sources of it in coastal environments Science of the Total Environment, 650 (2019), 1403–1411 R. García-Seoane1, J.A. Fernández1, M.T. Boquete2,3, J.R. Aboal1. 1Ecology Unit, Dept. Functional Biology, Universidade de Santiago de Compostela, Fac. Biología, Lope Gómez de Marzoa s/n, Santiago de Compostela, 15782, A Coruña, Spain. 2Estación Biológica de Doñana, CSIC, Avenida Américo Vespucio 25, Isla de la Cartuja, Sevilla 41092, Spain. 3Department of Integrative Biology, University of South Florida, 4202 E Fowler Ave, Tampa, FL 33620, USA. https://doi.org/10.1016/j.scitotenv.2018.09.156 . Chapter III 37 CHAPTER III Sampling optimization for biomonitoring metal contamination with marine macroalgae Environmental Pollution, 255 (2019), 113349 R. García-Seoane1, J.A. Fernández1, Z. Varela1, C. Real2, M.T. Boquete3,4, J.R. Aboal1. 1Ecology Unit, Dept. Functional Biology, Universidade de Santiago de Compostela, Fac. Biología, Lope Gómez de Marzoa s/n, Santiago de Compostela, 15782 A Coruña, Spain. 2Ecology Unit, Dept. Functional Biology, Universidade de Santiago de Compostela, Escola Politécnica Superior de Enxeñaría, Rúa Benigno Ledo, 2, Campus Universitario, 27002 Lugo, Spain. 3Estación Biológica de Doñana, CSIC, Avenida Américo Vespucio 25, Isla de la Cartuja, Sevilla 41092, Spain. 4Department of Integrative Biology, University of South Florida, 4202 E Fowler Ave, Tampa, FL 33620, USA. https://doi.org/10.1016/j.envpol.2019.113349 Chapter IV 41 CHAPTER IV Optimal number of Fucus vesiculosus subsamples to differentiate between sites affected by distinct levels of heavy metal contamination Submitted for publication. R. García-Seoane, J.R. Aboal, J.A. Fernández. Ecology Unit, Dept. Functional Biology, Universidade de Santiago de Compostela, Fac. Biología, Lope Gómez de Marzoa s/n, Santiago de Compostela, 15782 A Coruña, Spain. Abstract The presence of trace elements in marine habitats is a serious environmental problem which increasingly affects ecosystem and human health. The use of macroalgae as contamination biomonitors represents a valuable alternative approach to traditional physicochemical methods. The present study was carried out to determine the optimal number of subsamples of Fucus vesiculosus needed to detect statistically significant differences in the mean concentrations of Al, As, Cd, Co Cr, Cu, Fe, Hg, Ni, Pb, Zn, N and δ15N between two sampling sites affected by different levels of contamination. For this purpose, we plotted the density distributions of the concentrations of the different elements and examined the local variability at three sites. For sites with mean concentrations differing by more than 30%, 20 subsamples were sufficient to detect significant differences for all of the elements, except Cr. The proposed methodology could be used in other studies in the absence of specific research on each species and region. Keywords: Contamination monitoring; Heavy metals; Coastal environments; Algae; Local variability; Standardization. 2 42 2 1 XXY   222 2 1 XXXY   As an explicit expression is not available for the inverted square root transformation, µY and 𝜎2Y were calculated as the mean and variance of the data transformed from the raw variable (then yi = xi-0.5, Eqs. 1 and 2), which although not strictly correct, is more appropriate than using the data from the original variable. All statistical tests and Box-Cox transformations were performed using IBM SPSS Statistics 24 (SPSS Inc., USA). The density functions for each element were estimated by kernel smoothing, with the Kern-Smooth package (Wand & Ripley, 2006) in R (R development Core Team, 2008). The technique was applied with a Gaussian kernel, the width of which was optimised for each data set by using direct introduction with two levels of functional estimation (Wand & Jones, 1995). 3. Results The descriptive statistics of the concentrations of the elements determined in the algae collected at the three SS studied are shown in Table 1. In general, contamination levels were higher at SS3 than at SS1 and SS2. As regards the variability of the data, the coefficients of variation and dispersion (median absolute deviation/median) were, in comparison with the other elements, higher for Al, Fe and Pb in all SS, with values of around 40–50%. Additionally, the density distributions for some of the elements and sites studied are shown as an example in Fig. 1. These graphs confirm that the concentrations of most of the elements (other than N) in algae were non-normally distributed at all SS (see Fig. 1 and Table 1). In general, polymodal distributions with positive asymmetry (skewed to the right) were observed (e.g. for δ15N in SS1, and Ni in SS2), and in some CHAPTER IV 49 cases the skew to the right was so large that there appeared to be some negative asymmetry (e.g. As in SS3). These types of distributions are indicative of the existence of contamination at the SS (Viana et al., 2010). For example, the concentrations of Cr, Fe and Zn include an extreme value 3 times higher than the modal value of the distributions in SS2. Some of these atypical values in the density distributions were identified in the same subsamples for some of the elements determined in the same SS (e.g. Al and Cr in SS1, Fig. 1). The comparisons between the density distributions of Cu in SS2–SS3 and Zn in SS1–SS2 are shown as examples (see Fig. 1) to illustrate the reasoning behind the calculation of the minimum number of subsamples required to significantly differentiate two sites in terms of metal concentrations (values presented in Table 1). As previously explained, the number of subsamples will be determined by the d value (i.e. the difference between population means µ1 and µ2) and by the deviation of each distribution compared (σ). For Cu, the distributions compared are clearly different, with means differing by ca. 50% (high d value), whereas for Zn, the distributions almost overlap and only differ by 5% (low d value). In addition, for Cu the σ values for both distributions (~0.8) are lower than those for Zn (4.9 and 8.8). With this information and using the previously defined equations, we found that by collecting a minimum of 3 subsamples from SS2 and another 3 from SS3, we can significantly differentiate (p≤0.05) between the mean concentrations of Cu at both SS, as the minimum detectable difference between population means, δ3=2.22, is lower than d=2.37. On the contrary, Zn distributions were very variable and the mean values were very similar, so that significant differences would not be detected with 3 subsamples (as δ3 > d), and a much larger n (119 in each SS) would be required to differentiate between the mean concentrations of Zn at SS1 and SS2 (as δ119 < d=1.71). Thus, the more dissimilar the means (µ2 - µ1) are (i.e. the higher the value of d) and the less variable the data from each distribution are (σ1 and σ2) (i.e. the smaller the value of σ), the easier it will be to detect significant differences between SS (as a small n is required). Likewise, for most of the remaining elements, when pairs of SS with very different means were compared (e.g. Al, As, Fe, Cu, Hg and Pb, for the SS1–SS3 and SS2–SS3 comparisons), the number of subsamples required was small (between 1 and 8), whereas to differentiate between the concentrations for those elements with similar means in the SS compared (e.g. Cr, Cu and Zn in the SS1–SS2 comparison), a larger number of subsamples would be required (more than 400 for Cr and Cu) (Table 1). Table 1. Descriptive statistics of concentrations (µg g-1) of each element determined in 50 subsamples of Fucus vesiculosus collected at the sampling sites (SS) under study. Mean values and coefficient of variation (%) are shown in regular typeface for those elements that are normally distributed, and median values and coefficients of dispersion (median absolute deviation/median, %) are shown in italics for those elements and SS that are not normally distributed. The minimum number of subsamples required to differentiate (p≤0.05) the mean tissue concentrations of the elements studied in F. vesiculosus growing at the three SS compared is also shown. Pairs of SS in which the mean concentrations differ by more than 30% are shown in bold type. a: ng g-1; b: %; c: ‰. Sampling sites Comparisons SS1 SS2 SS3 SS1-SS2 SS1-SS3 SS2-SS3 Al 200 (53) 352 (77) 1437 (50) 24 2 2 As 55.6 (10) 51.3 (11) 28.1 (14) 47 2 1 Cd 0.82 (17) 1.02 (20) 0.47 (9) 32 13 9 Co 0.52 (19) 1.18 (16) 0.99 (10) 3 35 48 Cr 0.31 (24) 0.40 (50) 0.98 (37) 581 29 84 Cu 2.31 (12) 2.53 (17) 4.80 (17) 415 7 3 Fe 191 (35) 299 (58) 1002 (44) 7 1 1 Hga 22.5 (30) 14.5 (20) 59.3 (19) 6 1 1 Ni 1.07 (25) 2.02 (17) 3.75 (6) 9 3 16 Pb 0.40 (41) 0.25 (39) 1.62 (39) 86 8 6 Zn 32.5 (15) 33.1 (10) 46.6 (8) 119 2 4 Nb 0.40 (11) 0.77 (7) 0.59 (8) 2 4 5 δ15Nc 1.94 (4) 2.59 (2) 2.81 (4) 2 12 5 CHAPTER IV 51 Fig. 1. Density functions for the concentrations of Al, Cr and δ15N in sampling site (SS) 1, Hg and Ni in SS2, and As, Fe and N in SS3, determined in the brown algae Fucus vesiculosus. Subsamples showing extreme concentrations are highlighted in Al and Cr at SS1 as examples. The upper graphs show the comparisons made between two SS that differ in terms of mean concentrations (µ1 and µ2) and deviations (𝜎1 and 𝜎2) for Cu and Zn. The size of each arrow indicates the detectable minimum difference in concentration (δ), and the difference between the mean values (d) of each pair of SS compared, for n=3 and n=119 for Cu and Zn respectively. Thin and thick lines represent the values of both populations compared. For each kernel, vertical lines along the X-axis correspond to the individual values of the subsamples collected at each SS (n=50). The concentrations are expressed in µg g-1, except for Hg (ng g-1), N (%) and δ15N (‰). 4. Discussion The variability in the concentrations of elements in tissues of algae largely determines the number of subsamples required to differentiate between two SS separated in space, or to differentiate temporal changes in the same SS (e.g. in time trend studies). This variability can be generated by factors that locally affect the accumulation of elements in algae (i.e. factors causing intra–SS variability). Some of these factors have been investigated in detail (e.g. Nickless et al., 1972; Martin et al. 1997; García-Seoane et al., 2019) and are primarily related to the effect of the vertical position of the algae on the shoreline and to variations in environmental conditions (Fuge & James, 1974; Villares et al., 2002). In view of this, and by using the method recommended by Zar (2010), we have verified that in order to detect significant differences between SS in which the mean concentrations of the element under consideration are very different, very few subsamples are needed, e.g. 3 for Cu (mean concentrations differing more than 50% between SS compared, Fig. 1), see Table 1. These results are consistent with those reported by Barreiro et al. (1993) for F. ceranoides and F. vesiculosus collected in the same region, in which 3–4 subsamples were considered sufficient to detect significant differences between the mean concentrations of Al, Fe, Co, Mn, Ni and Zn, which differed by 50% in the different sites. On the other hand, we observed that the number of subsamples required increases substantially when the differences between the means are small, leading to impractical sample sizes, e.g. 199 subsamples in the case of Zn (Fig. 1 and Table 1). In light of this, we estimated the minimum number of subsamples that would yield the greatest number of significant differences between SS. We found that collection of 20 subsamples at each SS was sufficient to detect significant differences (p≤0.05) for all the elements whose concentrations differed by more than 30% between the SS compared (except for Cr for SS1–SS3 and SS2–SS3 comparisons, Table 1). For most of these elements (e.g. Al, As, Cu, Fe and Hg), a smaller number (<10 subsamples) would also have been sufficient. CHAPTER IV 53 These results (in terms of the number of subsamples required to differentiate SS) are only strictly applicable to studies using F. vesiculosus and carried out in the same region, because as mentioned above, the number of subsamples will depend on the existing local variability in the SS and on the similarity or divergence in the level of contamination between the SS compared. Thus, as the vast majority of studies using this technique to compare heavy metal or nutrient levels between SS do not justify the sample size chosen, and most collect a small number of subsamples at each SS (García-Seoane et al., 2018), the results of such studies could be questioned. In particular, in those cases where no significant differences were found on comparing contamination levels between SS, this may be due to the fact that the number of subsamples collected was not sufficiently representative of the high variability existing. For example, Riget et al. (1997) examined differences in element concentrations between different locations by collecting 5 subsamples of each species (F. vesiculosus, F. distichus, and A. nodosum) per site. However, a two-way ANOVA test revealed that the concentrations of the elements displaying the highest levels of natural variability (i.e. Cr and Pb), did not differ significantly between any of the 4 locations compared. According to our estimations, the concentrations of Cr and Pb in F. vesiculosus are highly variable (Table 1), and at least for this species, 5 subsamples were not sufficient to differentiate the SS with certainty. We estimated that at least 30 subsamples would be needed to differentiate the sites on the basis of Cr concentrations, and between 6 and 86 subsamples would be required to the same for Pb (Table 1). Likewise, Rainbow et al. (2002) reported no significant differences (Tukey’s post-hoc test) between SS with similar mean concentrations of elements such as Pb and Zn, when 10 subsamples of F. vesiculosus were collected in each SS. According to our estimations (Table 1), >80 subsamples may have been necessary to detect differences in Pb and Zn concentrations between SS. 5. Conclusions Although the technique for monitoring contamination with macroalgae is still far from being standardized, one aspect of the sampling protocol has been defined for monitoring heavy metal contamination in coastal waters. Despite the high variability in the concentrations of the different elements (especially Al, Fe and Pb) determined in the subsamples of F. vesiculosus, we were able to establish the minimum number of subsamples that should be collected in each SS to allow robust differentiation between SS affected by different levels of metal contamination (ca. 20 subsamples). From a practical point of view, this number of subsamples can easily be collected during sampling surveys. Although the number of subsamples was estimated using F. vesiculosus and applying our definition of a subsample (see Material and methods), our recommendation is valid for studies carried out in other regions and for other algal species, in the absence of new studies under different settings. Finally, in order to improve the results of the studies and increase the reliability and comparability of the data, further research is needed to support and revise the current methodology for using algae to monitor contamination levels. Acknowledgements The authors belong to the Galician Competitive Research Group GRC/GPC2016-002 and to the CRETUS Strategic Partnership (AGRUP2015/02). Both of these programmes are co-funded by FEDER (UE). The authors are grateful to the RIAIDT-USC for use of analytical facilities. Rita García Seoane was in receipt of a grant awarded by the Spanish Ministerio de Ciencia, Innovación y Universidades within the Programa de Formación de Profesorado Universitario (FPU 2014). CHAPTER IV 55 References Barreiro, R., Picado, L., Real, C., 2002. Biomonitoring heavy metals in estuaries: A field comparison of two brown algae species inhabiting upper estuarine reaches. Environ. Monit. Assess. 75(2), 121–134. Barreiro, R., Real, C., Carballeira, A., 1993. Heavy–metal accumulation by Fucus ceranoides in a small estuary in north–west Spain. Mar. Environ. Res. 36(1), 39–61. Box, G.E., Cox, D.R., 1964. An analysis of transformations. J. R. Stat. Soc. Series B Stat. Methodol. 211–252. Bryan, G.W., Langston, W.J., Hummerstone, L.G., 1980. The use of biological indicators of heavy metal contamination in estuaries: with special reference to an assessment of the biological availability of metals in estuarine sediments from south–west Britain. Occas. Publ. Mar. Biol. Assoc. U.K. 1, 1–73. Carro, N., García I., Ignacio M., Mouteira A., 2010. Spatial and temporal trends of PCBs (polychlorinated biphenyls) in mussel from Galician coast (1998 – 2008). Environ. Int. 36, 873–879. Chernova, E.N., Sergeeva, O.S., 2008. Metal concentrations in Sargassum algae from coastal waters of Nha Trang Bay (South China Sea). Russ. J. Mar. Biol. 34(1), 57–63. Cobelo–García, A., Prego, R., 2004. Influence of point sources on trace metal contamination and distribution in a semi–enclosed industrial embayment: the Ferrol Ria (NW Spain). Estuar Coast Shelf Sci. 60(4), 695–703. Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, New Jersey: Erlbaum. Cohen, J., 1992a. A power primer. Psychol. Bull. 112(1), 155–159. Cohen, J., 1992b. Statistical Power Analysis. Curr. Dir. Psychol. Sci. 1(3), 98–101. Conti, M.E., Cecchetti, G., 2003. A biomonitoring study: Trace metals in algae and molluscs from Tyrrhenian coastal areas. Environ. Res. 93(1), 99–112. Desu, M., Raghavarao, D., Boston, M.A., 1990. Sample size methodology. Academic Press, Inc. Deutsch, B., Voss, M., 2006. Anthropogenic nitrogen input traced by means of δ15N values in macroalgae: results from in situ incubation experiments. Sci. Total Environ. 366(2 3), 799–808. Fong, P., Kamer, K., Boyer, K.E., Boyle, K.A., 2001. Nutrient content of macroalgae with differing morphologies may indicate sources of nutrients for tropical marine systems. Mar. Ecol. Prog. Ser. 220, 137– 152. Fuge, R., James, K.H., 1973. Trace metal concentrations in brown seaweeds, Cardigan Bay, Wales. Mar. Chem. 1(4), 281–293. Fuge, R., James, K.H., 1974. Trace metal concentrations in Fucus from the Bristol channel. Mar. Pollut. Bull. 5(1), 9–12. García-Seoane, R., Fernández, J.A., Varela, Z., Real, C., Boquete, M.T., Aboal, J.R., 2019. Sampling optimization for biomonitoring metal contamination with marine macroalgae. Environ. Pollut. 255, 113349. García–Seoane, R., Fernández, J.A., Villares, R., Aboal, J.R., 2018. Use of macroalgae to biomonitor pollutants in coastal waters: optimization of the methodology. Ecol. Indic. 84, 710–726. Hintze, J.L., 2008. Power analysis and sample size system (PASS) for windows User's Guide I. NCSS. USA: Utah, Kaysville. CHAPTER IV 57 IBM Corp. Released 2016. IBM SPSS Statistics for Windows, Version 24.0. Armonk, NY: IBM Corp. Marmolejo–Rodríguez, A.J., Cobelo–García, A., Prego, R., 2007. Background values, distribution and contamination of metals in the sediments of the Pontevedra Ria (NW Spain). Soil Sediment. Contam. 16(6), 557–568. Maroli, L., Pavoni, B., Sfriso, A., Raccanelli, S., 1993. Concentrations of polychlorinated biphenyls and pesticides in different species of macroalgae from the Venice Lagoon. Mar. Pollut. Bull. 26(10), 553– 558. Martin, M.H., Nickless, G., Stenner, R.D., 1997. Concentrations of Cadmium, Copper, Lead, Nickel and Zinc in the alga Fucus serratus in the Severn estuary from 1971 to 1995. Chemosphere 43(2), 325–334. Miramand, P., Bentley, D., 1992. Heavy metal concentrations in two biological indicators (Patella vulgata and Fucus serratus) collected near the French nuclear fuel reprocessing plant of La Hague. Sci. Total Environ. 111(2–3), 135–149. Moher, D., Dulberg, C.S., Wells, G.A., 1994. Statistical power, sample size, and their reporting in randomized controlled trials. JAMA 272(2), 122– 124. Nawakowski, C., Nicholson, M.D., Kershaw, P.J., Leonard, K.S., 2004. Modelling 99Tc concentrations in Fucus vesiculosus from the north–east Irish Sea. J. Environ. Radioact. 77(2), 159–173. Net, S., Henry, F., Rabodonirina, S., Diop, M., Merhaby, D., Mahfouz, C., Amara, R., Ouddane, B., 2015. Accumulation of PAHs, Me–PAHs, PCBs and total mercury in sediments and marine species in coastal areas of Dakar, Senegal: Contamination level and impact. Int. J. Environ. Res. 9(2), 419–432. exhibited a decrease in concentrations from the youngest to the oldest dichotomies. To minimize de effect of intra-annual variability, a sampling strategy consisting on collect subsamples 2 times per year separated by 6 months and combined in a composite sample was recommended for biomonitoring purposes with this species. To minimize the effect of intra-thallus variability, it is proposed to make composite samples of the three apical dichotomies of thallus. Keywords: Aquatic contamination; Biomonitoring; Seasonality; Intra-thallus variability; Algae; Metals. CHAPTER V 65 1. Introduction From its inception as useful environmental tools in biomonitoring marine pollution in the early 1950s (Wort, 1955; Black & Mitchell, 1952), macroalgae have become one of the most commonly used bioindicators worldwide (Malea & Kevrekidis, 2014; Bonanno & Orlando-Bonaca, 2018; García-Seoane et al., 2018). However, the methods implemented in this type of study have not been properly addressed, and many basic aspects of the technique have not been taken into consideration despite being issues that affect the interpretation of the results obtained. One such aspect is the temporal representativeness of the concentrations of pollutants in the algae collected in a given area. Temporal variability (intra-annual) in element concentrations in algae must be characterized, otherwise samples collected within the same sites in different months/seasons will not be comparable, and inter-study comparisons will be also restricted. Although sample collection is usually limited to a particular time of the year, typically in summer to promote good sampling conditions (García-Seoane et al., 2018), if the concentrations of elements in algae vary over time, the values obtained making a single sampling survey may not well represent the annual mean value within the site, but an over or underestimation of it. Previous studies in which samplings were carried out over a minimum period of one year covering all seasons, revealed intra-annual variability in more than 75% of the elements considered (García-Seoane et al., 2018). For example, Ferreira & Oliveira (1988) found coefficients of variation (CV) for Hg of 45% in samples of the brown algae Fucus vesiculosus collected throughout a year, and similarly, Villares et al. (2013) found annual CV of 17%, 25%, 32%, 38% and 133% for Mn, N, Fe, Zn and Cu respectively for the same species. These variations in the concentrations of elements in algae have been attributed to changes in environmental factors (Stoeppler et al., 1986; Haroon et al., 1995), although most authors think that biological factors, such as metabolism, reproduction and/or growth, constitute the most determinant factors of the temporal variability (Rao & Indusekhar, 1989; Malea, 1995; Wright & Mason, 1999). Several authors have recommended that to minimize the effect of this variability, various samples should be collected during the year in each site, commonly with monthly samplings, and integrated in a single composite sample, a strategy also known as “Time-bulking” of samples (Phillips & Segar, 1986; Stoeppler et al., 1986). The main problem of this approach is that the greater the magnitude of the intra-annual variability (due or not to seasonality), the largest the number of samplings would need to be performed, involving a huge sampling and economic effort. Thus, it is necessary to characterize the temporal variability of the concentrations of elements in algae in order to establish the sampling frequency required to yield representative values of annual concentrations (Malea et al., 2015; García-Seoane et al., 2018). Along with the intra-annual variability, other aspect of the biomonitoring technique with macroalgae that needs to be closely investigated in order to yield representative and comparable data among studies is the intra-thallus variability (Riget et al., 1997; Burger et al., 2007; Sáez et al., 2012). Several studies have been reported to show differences in metal concentrations between different parts of the algal thalli, especially in fucoid species (García-Seoane et al., 2018), differences mainly related to the age and physiological condition of the tissues (Favero & Frigo, 2002; Savage & Elmgren, 2004). In consequence, selection of the part of the algae used as biomonitor should be previously standardized for the results to be consistent across studies and not draw erroneous conclusions (Sáez et al., 2012). In addition, it is still not known whether there is any relationship between intra-thallus and intra-annual variability. The present study aims to investigate the existence of temporal variation patterns of intra-annual bioconcentration of nitrogen, δ15N and trace elements in F. vesiculosus along with the intra-thallus variability in the concentrations of these elements over a three-year period in three sampling sites. The possible implications that the existence of temporal and intra-thallus variability in the tissue contents of these elements would have on the interpretation of results in biomonitoring studies will be discussed, and an appropriate sampling design for routine use in CHAPTER V 67 biomonitoring programmes with macroalgae, representative of the intraannual and intra-thallus variability, will be proposed. In addition, the periodicity in the occurrence of new dichotomies in the thallus of this species and their implications in the interpretation of the results of temporality studies of pollutants will also be discussed. 2. Material and methods 2.1 Sampling Samples of the brown seaweed Fucus vesiculosus L. (Class Phaeophyceae) were collected monthly from three sampling stations (SS) located in the coast of Galicia (NW Spain) from November 2015 to November 2018. The SS were located far from point sources of pollution, such as port infrastructures or submarine outfalls. The first site (SS1) was located in the Ría of Ferrol (centered in X = 557811, Y = 4812476; UTM 29N ETRS89), considered one of the most contaminated rias by heavy metals in the coast of Galicia (Cobelo-García & Prego, 2004). The second site (SS2) was selected within the Ría de Muros e Noia (X = 506599, Y = 4737191), an area with scarce industrial activity. The last site (SS3) was located in the Ría de Pontevedra (X = 523837, Y = 4697147), characterized by the presence of a chlor-alkali industry and several small cities in its margins. In each SS, 30 subsamples consisted of individual thalli with similar size attached to rocks, were collected within three 50 m bands parallel to the coastline (10 subsamples per band), and combined into a single composite sample to achieve greater representativeness of the intra-SS variability in the concentrations. A more detailed description of the sampling protocol, washing and sample processing can be found in García-Seoane et al. (2019). 2.2 Sample processing Samples were manually cleaned by removing any adhering material and discarding receptacles, damaged or old tissues and sections heavily affected by epiphytes (García-Seoane et al., 2018). Consecutive dichotomous sections of the seaweed thallus were separated at the base of every pair of air bladders underneath the angle formed by each fork with a glass spatula, as illustrated in Fig. 1, following the method used in previous studies (e.g. Savage & Elmgren, 2004, Stengel et al., 2005; Carballeira et al., 2014). The three most apical dichotomies were separated during the first year under study. One more consecutive dichotomy (4th dichotomy) was separated the next year. The same procedure was followed over the third year, with the 5th dichotomy being included, making a total of 5 dichotomies at the end of the 3 years of survey. Prior to analysis, each dichotomy was individually dried in a forced air oven at 40ºC (72 h), then homogenized in a tangential mixer mill with zirconium oxide grinding vessels (Retsch MM400), and weighed on a precision balance (Mettler ToledoXP26). Dried dichotomies (between ca. 2 and 8 g dry weight, d.w., each) were then stored at room temperature in hermetically sealed vials until chemical analysis. Fig. 1. Diagram showing the different sections (from 1st–apical to 5th–basal) selected over the three-year survey in the dichotomously divided thallus of Fucus vesiculosus. CHAPTER V 69 2.3. Growth measurements Fucus vesiculosus is characterized by apical growth and pseudo-dichotomous branching of the thallus. The thallus grows forming dichotomous ramifications, leading to a parallel although unequal development of the branches (Hoek et al., 1995). Several authors have suggested that the branches in this species are dichotomously divided once a year, which has led to the interpretation that each dichotomy corresponds to an annual growth segment (e.g. Carlson, 1991; Savage & Elmgren, 2004; Stengel et al., 2005; Carballeira et al., 2014). However, during past field work campaigns, the authors have observed that the frequency of the thallus forking is presumably higher than the previously described in these studies. Because we are unaware of any previous studies which addresses the growth periodicity of F. vesiculosus (in terms of number of dichotomies formed each year), we believe that it is convenient to assess whether each dichotomy truly corresponds to an age cohort or if several dichotomous divisions occur consecutively during the same year. To study growth of F. vesiculosus, a group of 30 thalli was randomly selected within each SS and individually labeled with coded plastic tags in the basal section of the thallus. The number of new dichotomies formed and the size growth were periodically monitored by taking photographs of each thallus against a background of graph paper. Growth monitoring was carried out every 3 months for a period of 9 months (January 2019 – October 2019). All individuals were measured to the nearest mm from the base of the holdfast to the tip of the longest frond. 2.4 Chemical analysis Before analysis, samples were dried again at 40ºC in a forced air oven. The mineralization of the samples (1 g d.w.) was performed in Teflon vessels in a microwave oven (CEM MDS2100) in three successive steps (10 min at 100ºC, 7 min at 150ºC, 25 min at 190ºC), by adding 10 mL of HNO3 (65%), 2 mL of H2O2 (30%) and 2 mL of MilliQ water. The concentrations of Al, As, Cd, Co, Cr, Cu, Fe, Ni, Pb and Zn were determined by ICP– MS (VARIAN 820-MS ICP quadrupole mass spectrometer) at the Research Support Services Unit from Universidade de Santiago de Compostela. The concentrations of Hg were determined in an elemental analyzer (DMA 80 Milestone). For determination of %N and δ15N signal, 3±0.1 mg (d.w.) of sample were packaged in tin capsules (EuroVector) and processed in an elemental analyser (FlashEA1108 Carlo Erba Instruments) coupled to a mass spectrophotometer (MAT253 ThermoFinnigan). Determinations were made at the Unit of Instrumental Techniques of Analysis from Universidade da Coruña. To ensure the analytical quality of the process, one analytical blank, one replicate sample and two certified reference materials, Sea lettuce–Ulva lactuca [BCR–279] and Bladderwrack–Fucus vesiculosus [ERM–CD200] (Institute for Reference Materials and Measurements, IRMM, Belgium) were analyzed once every ten samples. In the case of %N, only the reference material BCR–279 was analyzed. The global error associated with the analytical process was usually lower than 6%, except for Hg (around 10%) and Al (around 17%). The percentage of recovery from the reference materials ranged for BCR–279, between 65% (Pb) and 110% (Hg), usually around 85%, and for ERM–CD200, between 65% (Pb) and 122% (Hg), usually around 90%. Determinations were above the corresponding limits of quantification (LOQ), with exception of Ni (in 5% of the cases), Cu (8%), Co (11%), Cd (15%), Al (33%), Pb (44%) and Cr (69%). Data from those SS where element concentrations were below the LOQ in more than 30% of the samples, i.e. from SS1 for Cr, from SS2 for Cr and Pb, and from SS3 for Al, Cr and Pb, were not included in the data treatment. CHAPTER V 71 2.5 Data analysis 2.5.1 Intra-thallus variability Lilliefors modifications of the Kolmogorov–Smirnov test was used to check the normality of the data. Except for N and δ15N, which concentrations were normally distributed, those elements with nonnormal distributions were successfully normalized using Box–Cox transformations: log(x) transformation for As, Cd, Co, Cu, Fe, Ni and Pb, sqrt(x) transformation for Al, and 1/sqrt(x) transformation for Hg and Zn. A Three–way ANOVA test was used to test differences between the five dichotomous sections of the thallus and study the interaction between the factors (categorical variables): “Sampling Month * SS * Dichotomy”. The concentrations of the elements were considered as the dependent variables. When there was no interaction between factors and when significant differences were detected (p≤0.05), a Tukey’s Post–Hoc test was used to identify differences. Spearman’s rank correlation coefficients were also calculated to determine the relationship between the concentrations of the elements and the weight (d.w.) of the dichotomies. Statistical analysis was performed using R-3.4.0 (R development Core Team, 2008). 2.5.2 Analysis of the structure of the temporal series The classical analysis of time series is based on the assumption that the values taken by the variable are the consequence of three components (trend, seasonal and random components) whose combined actuation results in the measured values (Brockwell & Davis, 2002; Chatfield, 2003; Anderson, 2011). To distinguish between trend and seasonal components, trends in raw data were isolated by lineal regression analysis of the time series. A detrending procedure was then applied to remove the trend component from those non-stationary time series (with trend and/or variability changing systematically over time) with significant trends (p<0.01) (Chatfield, 2003; Box et al., 2014). To assess the degree of dependence in observations of the time series, autocorrelograms were calculated from the regression residuals in those detrended series, or directly from raw data in those stationary series (with constant mean and variance over time), for each dichotomy and for each element separately at all SS. The “acf( )” function available in R-3.4.0 (Package gstat) was used to plot correlograms (Pebesma, 2004; Gräler et al., 2016). Correlations outside the 95% confidence interval (bounds ±1.96/√𝑛, where n is the number of lags and 1.96 is the 0.975 quantile of the standard normal distribution) were deemed significant at the 5% significance level (Brockwell & Davis, 2002). In the absence of data for the three years in the 4th and 5th dichotomies, and to get a more consistent temporal analysis, correlograms and time plots were only studied in the three more apical dichotomies. In order to study the seasonal variation and to identify the dominant periods (or frequencies) of the series, periodic (cyclic) regression models with a sine and a cosine component were fitted (Chatfield, 2003), using the “spectrum” (to calculate spectral density and determine the periods of the series) and “lm” (to illustrate the model estimated) functions available in R-3.4.0 (Package TSA, R development Core Team, 2008) (Cryer & Chan, 2008; Shumway & Stoffer, 2017). A description of the periodic regression model and its parameters can be found in Table S1 (Supplementary Material). Although the terms inside the sine and cosine functions are known, the regression coefficients are automatically estimated by the regression model fitted of each series. The fitting of the models was done by calculating determination coefficients (r2), considered significant at p<0.01. 3. Results Annual range of variation in concentrations of the different elements/δ15N signal in the five dichotomies of the thallus at each SS are shown in Table 1. Although δ15N signal is not an element per se, but is an CHAPTER V 73 isotopic relation of elements, we will refer from now on to all the elements determined (including δ15N signal) as “elements”, to simplify the readability throughout the text. Intra-annual variability in the concentrations at each dichotomy was studied by calculating the coefficients of dispersion (COD), as the ratio between the MAD (median absolute deviation) and the data median. Concentrations varied between 3% for δ15N at SS1 (third year, 1st dichotomy) and 747% for Cd at SS3 (third year, 3rd dichotomy). The COD usually ranged between 30% and 60% for most of the elements in all the SS, with the exception of N and δ15N (usually <20%), and were especially high (>100%) for Al in SS2, As and Cd in all SS, and Co, Cu and Fe in SS3. A general comparison between dichotomies showed that the apical dichotomy was clearly the least variable in SS1 in the three years studied. No differences were apparently observed between dichotomies in the other SS. Table 2. Results from significant lineal regression analysis of time series (p<0.01) in Fucus vesiculosus dichotomies from three sampling sites (SS). Determination coefficients (r2) and the slope of each regression are shown. Significant regression models parameters adjusted to the seasonal component of time series (p<0.01). Sig.: signification level; n.s.: nonsignificant; β0: mean of the series; β1 and β2: regression coefficients; ϵt: residuals; per: period (further details in Table S1, Suppl. Mat.). Trend component Seasonal component Dichotomy r2 Slope r2 Sig. β0 β1 β2 ϵt per 1st 0.23 3.44 - n.s. - - - - - SS1 2nd - - 0.22 <0.01 130 8.86 45.8 2.26 10 Al 3rd - - 0.42 <0.001 146 29.7 64.5 4.07 10 1st - n.s. - - - - - SS2 2nd 0.26 -6.56 - n.s. - - - - - 3rd 0.28 -8.99 - n.s. - - - - - 1st 0.48 2.63 - n.s. - - - - - SS1 2nd 0.50 1.99 0.24 <0.01 -1.30E-04 12.2 -9.66 -1.99 12 3rd 0.28 0.76 0.33 <0.001 4.56E-05 8.72 -6.37 -1.94 12 1st 0.33 2.26 - n.s. - - - - - As SS2 2nd 0.24 1.63 0.59 <0.001 -2.09E-04 -8.95 -32.54 1.33 36 3rd - - 0.44 <0.001 44.6 -12.9 -17.7 0.504 36 1st 0.23 1.75 0.36 <0.001 2.99E-04 17.0 -24.7 -1.328 12 SS3 2nd 0.34 1.39 - n.s. - - - - - 3rd 0.19 0.45 0.20 <0.01 -0.424 1.82 -6.61 -1.56 12 1st 0.41 0.05 0.53 <0.001 -1.46E-04 0.664 -0.091 0.040 12 SS1 2nd 0.21 0.01 0.58 <0.001 7.25E-05 0.293 -0.081 -0.021 12 3rd - - 0.40 <0.001 0.606 0.223 -0.101 -0.036 12 1st - - 0.35 <0.001 1.80 0.209 -0.530 -0.103 12 Cd SS2 2nd - - 0.60 <0.001 0.896 0.193 -0.371 -0.020 12 3rd - - 0.44 <0.001 0.766 0.156 -0.330 -0.041 12 1st - - - n.s. - - - - - SS3 2nd - - 0.23 <0.01 0.425 -0.053 -0.126 -0.022 11 3rd - - - n.s. - - - - - 1st 0.28 -0.02 - n.s. - - - - - SS1 2nd - - 0.25 <0.01 1.01 0.406 -0.467 -0.089 12 3rd - - 0.31 <0.001 1.35 0.415 -0.705 -0.076 12 1st 0.23 -0.10 - n.s. - - - - - Co SS2 2nd 0.24 -0.14 - n.s. - - - - - 3rd 0.21 -0.16 0.16 <0.01 -6.36E-04 0.177 -2.48 -0.518 12 1st 0.26 -0.03 - n.s. - - - - - SS3 2nd 0.21 -0.03 - n.s. - - - - - 3rd - - 0.26 <0.01 0.752 0.100 -0.498 -0.088 12 1st - - 0.32 <0.001 3.99 1.37 -0.486 0.087 12 SS1 2nd - - 0.20 <0.01 3.72 0.840 -1.01 -0.391 12 3rd - - 0.35 <0.001 3.78 0.338 -1.44 -0.340 12 1st - - 0.39 <0.001 3.62 0.373 -0.899 -0.227 12 Cu SS2 2nd 0.31 -0.04 0.58 <0.001 -3.10E-04 0.042 -0.748 0.044 12 3rd 0.40 -0.06 - n.s. - - - - - 1st - - 0.35 <0.001 2.38 -0.752 0.310 -0.040 10 SS3 2nd - - 0.24 <0.01 1.89 -0.339 -0.353 -0.140 12 3rd - - 0.47 <0.001 1.84 -0.596 -0.192 -0.115 12 CHAPTER V 81 TABLE 2 (Continued) Trend component Seasonal component Dichotomy r2 Slope r2 Sig. β0 β1 β2 ϵt per 1st - - - n.s. - - - - - SS1 2nd - - 0.30 <0.001 141 23.7 38.3 -1.21 10 3rd - - 0.36 <0.001 169 38.0 49.9 -3.88 10 1st - - - n.s. - - - - - Fe SS2 2nd 0.41 -5.54 - n.s. - - - - - 3rd 0.34 -7.02 - n.s. - - - - - 1st - - 0.31 <0.001 68.9 -12.1 -26.1 -6.36 16 SS3 2nd 0.31 -2.35 - n.s. - - - - - 3rd 0.32 -2.46 0.34 <0.001 0.662 0.203 31.2 -1.97 12 1st 0.23 -0.29 - n.s. - - - - - SS1 2nd 0.48 -0.48 - n.s. - - - - - 3rd 0.28 -0.43 - n.s. - - - - - 1st - - - n.s. - - - - - Hg SS2 2nd 0.57 -0.48 - n.s. - - - - - 3rd 0.43 -0.36 - n.s. - - - - - 1st 0.23 -0.43 0.27 <0.01 -1.03E-03 -1.21 -6.38 -0.045 10 SS3 2nd 0.27 -0.47 - n.s. - - - - - 3rd 0.38 -0.46 - n.s. - - - - - 1st 0.36 -0.03 - n.s. - - - - - SS1 2nd - - 0.20 <0.01 1.78 0.858 -0.740 -0.342 12 3rd - - 0.25 <0.01 2.67 0.993 -1.27 -0.283 12 1st 0.34 -0.10 - n.s. - - - - - Ni SS2 2nd 0.37 -0.18 - n.s. - - - - - 3rd 0.35 -0.20 0.27 <0.01 -8.04E-04 0.683 -2.20 -0.383 12 1st 0.27 -0.05 - n.s. - - - - - SS3 2nd 0.27 -0.07 - n.s. - - - - - 3rd - - 0.21 <0.01 2.35 -0.480 -1.02 -0.300 12 1st - - 0.32 <0.001 0.350 0.006 0.090 -0.009 10 Pb SS1 2nd - - 0.32 <0.001 0.457 0.053 0.145 -0.016 10 3rd - - 0.23 <0.01 0.604 0.130 0.175 -0.058 10 1st 0.18 -1.26 0.37 <0.001 -0.987 24.3 -8.51 -5.82 12 SS1 2nd - - 0.39 <0.001 74.8 44.3 -33.0 -5.09 12 3rd - - 0.51 <0.001 95.0 49.3 -48.8 -3.32 12 1st - - 0.38 <0.001 43.2 17.7 -19.0 -4.74 12 Zn SS2 2nd 0.19 -1.77 0.32 <0.001 -2.10E-04 14.7 -28.4 -6.13 12 3rd - - 0.48 <0.001 56.2 18.9 -39.8 -4.80 12 1st 0.22 -1.20 0.30 <0.01 3.46E-04 12.3 -15.1 -1.62 12 SS3 2nd 0.24 -1.52 0.32 <0.001 -4.33E-04 7.90 -22.5 -5.57 12 3rd - - 0.38 <0.001 51.6 8.17 -26.2 -4.75 12 3.3 Growth measurements Regarding growth in F. vesiculosus, a very short time of experiment revealed the high mortality among the individuals monitored. Several individuals had disappeared after an interval of a few weeks and gradually at some point during the checkup, so these could not be followed for the entire period. Due to a strong herbivory in the apical tissues of algae and the significant loss of individuals in SS3, growth measurements in this site were only possible at 3 months after the initial follow-up. In addition, many individuals experienced significant size growth that it was impossible to follow the number of dichotomies formed from 6 months onwards. Although the loss of individuals not guaranteed data of statistical significance, it allowed a preliminary inquiry into size and dichotomous branching variation in this species. Thus, healthy thalli formed 2–3 new dichotomies every three months (up to 5 new dichotomies in some cases), what is ca. 1 dichotomy/month (Fig. S2, Suppl. Mat.). The growth in thallus length was more variable between measurements. Thalli in SS1 and TABLE 2 (Continued) Trend component Seasonal component Dichotomy r2 Slope r2 Sig. β0 β1 β2 ϵt per 1st - - 0.65 <0.001 0.719 0.172 -0.083 0.020 12 SS1 2nd - - 0.65 <0.001 0.626 0.180 -0.089 0.029 12 3rd - - 0.71 <0.001 0.555 0.178 -0.087 0.021 12 1st - - 0.67 <0.001 0.932 0.029 -0.133 -0.004 12 N SS2 2nd - - 0.50 <0.001 0.830 0.021 -0.150 0.019 12 3rd - - 0.67 <0.001 0.778 0.042 -0.116 0.006 12 1st - - 0.71 <0.001 0.757 0.147 -0.149 0.013 12 SS3 2nd - - 0.58 <0.001 0.690 0.124 -0.140 -0.003 12 3rd - - 0.65 <0.001 0.624 0.097 -0.167 0.012 12 1st - - 0.51 <0.001 2.49 0.294 -0.019 0.012 12 SS1 2nd - - 0.50 <0.001 2.29 0.269 -0.132 0.011 12 3rd - - 0.66 <0.001 2.15 0.244 -0.156 0.016 12 1st 0.36 0.01 - n.s. - - - - - δ15N SS2 2nd 0.34 0.01 - n.s. - - - - - 3rd 0.44 0.01 - n.s. - - - - - 1st - - - n.s. - - - - - SS3 2nd 0.31 0.02 0.21 <0.01 4.72E-05 -0.242 -0.076 -0.028 24 3rd 0.34 0.02 0.23 <0.01 -1.04E-02 -0.230 -0.044 0.009 24 CHAPTER V 83 SS2 had grown 7±3.5 cm (mean±SD) three months after the first measurement and 5±3 cm after the second measurement. The growth after the third measurement was not considered given the significant loss of individuals and the strong herbibory in those who survived. 4. Discussion 4.1 Intra-thallus variation The concentrations of all elements varied significantly depending on the dichotomy analyzed, but this variation was not dependent on the “Sampling month” neither on “SS” (Table S2, Suppl. Mat.). Regardless the type of trend in concentrations observed along the thallus, the lack of significant differences in concentrations of most elements between the oldest dichotomies suggests that bioconcentration rates in F. vesiculosus stabilize with the age of the tissue. Intra-thallus variation in concentrations of elements was widely reported in algae (García-Seoane et al., 2018), and previously observed in F. vesiculosus between dichotomies (e.g. Savage & Elmgren, 2004; Carballeira et al., 2014), and between different structural parts of the thallus (e.g. Bryan & Hummerstone, 1973; Carvalho et al., 1997). However, none of these studies has applied robust statistics to determine whether concentrations in the different dichotomies of the thallus significantly interact with factors as the month of collection or the SS. Furthermore, most of them did not even use statistical tests to determine whether the differences observed between the different parts of the thallus were significant or not, and neither cross-checked the representativeness of their results by comparing among various SS (García-Seoane et al., 2018). In view of this, the gradual increase in concentrations of Al, Co, Fe, Ni, Pb and Zn observed from the growing tips towards the older dichotomies of the thallus (Fig. S1, Suppl. Mat.) is in accordance with the results reported from previous studies for the same elements in F. vesiculosus, e.g. for Al, Fe, Pb and Zn (Bryan & Hummerstone, 1973), Fe, Ni and Zn (Barnett & Ashcroft, 1985), and Al, Co, Fe, Ni and Zn (Försberg et al., 1988; Söderlund et al., 1988). Furthermore, the decreasing trend in concentrations of As, Cd, Hg, N and δ15N with the age of dichotomy (Fig. S1, Suppl. Mat.), although is opposed to the δ15N values found in the same species by Savage & Elmgren (2004), is also consistent with the results from Carballeira et al. (2014) and Viana et al. (2015), who found highest δ15N values in the apical dichotomies compared to the rest of the frond. However, although these authors noted the existence of concentration gradients within the thallus of F. vesiculosus and, in most cases, their conclusions agree with those reached in this study, the absence of statistical tests confirming the existence of significant differences between dichotomies, especially in older studies, leads these results to be considered inconclusive. The presence of a concentration gradient related with the thallus age could be the result of several factors, including: i) the intrinsic uptake characteristics of algae tissues (e.g. variations in cation exchange capacity); ii) the different growth rates and metabolic activity of tissues (e.g. Rice & Lapointe, 1981; Stengel et al., 2005); iii) but also the contamination by epiphytes (e.g. Kangas et al., 1982) and fine particles (e.g. Bryan & Hummerstone, 1973; Försberg et al., 1988; Malinovskaya, 1998); and iv) the presence of pre-existing contaminant load in tissues. However, according to the results of the present study, the preferential concentration of the elements in young or old tissues cannot be explained by a single factor. The physicochemical characteristics of the elements may influence the capacity to compete for binding sites (e.g. Stengel et al., 2005; Ryan et al., 2012), but the variation in the number and/or type of binding sites along the thallus may also explain the differences observed in the distribution patterns of elements. Meanwhile, the different metabolic functions of the elements across different thallus regions can also contribute to the observed differences (e.g. Rice & Lapointe, 1981; Stengel et al., 2005). As some authors (e.g. Bryan & Hummerstone, 1973; Higgins & Mackey, 1987) suggested, the metal uptake capacity in algae may increase CHAPTER V 85 with growth rate, as algae growth rates determine the quantity of tissue per unit of time that is exposed to the environmental levels of pollutants. Although the relationship between bioconcentration and growth rate is likely to be metal-specific (Stengel et al., 2005). Thus, higher growth rates in young tissues than in old parts would lead to a rapid increase of biomass and, consequently, to a high proportion of free binding sites for capture elements (higher cation exchange capacity in tissues). Furthermore, it is assumed that the active uptake and release processes are faster in new and metabolically more active tissues because the ability to regulate both processes in the macroalgae decrease in the course of aging (Malinovskaya, 1998). All of the above would explain the large differences observed in the concentrations of As, Cd, Hg, N and δ15N between young and old dichotomies (Fig. S1, Suppl. Mat.). Nevertheless, high concentrations of elements such as Al, Co, Fe, Ni, Pb and Zn in the oldest areas of the thallus (Fig. S1, Suppl. Mat.) is probably more related to factors as the appearance of epiphytes (Malinovskaya, 1998) or particulate matter (Bryan et al., 1985; Barreiro et al., 2002). But the possible pre-existing contaminant load compared to youngest tissues or, as suggested by Higgins & Mackey (1987), the increase in metal binding sites in older tissues may also cause those high concentrations. Finally, contrary to the mentioned in the previous paragraph, some authors (e.g. Fuge & James, 1973; Markham et al., 1980; Villares et al., 2002) have pointed out that metal uptake capacity in algae decrease with growth rate, as higher growth rates in apical parts would result in an increase in biomass of tissue, leading to a dilution of concentrations by new growth, and therefore to lower concentrations in young than in old tissues. However, this theory can be discarded based on that, although young tissues grow proportionally more than olds, all tissues will uptake/release metals until they are balanced with the concentrations present in the environment, even while the new tissues is growing, so there could never be a dilution by growth. To conclude, the strong negative correlations between elements such as As and Cd, and the dry weight of the dichotomies, could indicate that the greater the amount of tissue, the lower the bioconcentrated concentrations. This could be related to the amount of specific surface available to adsorb these elements which, in proportion, may be less than the rest of the tissue as the algae grow. 4.2 Temporal variation The concentrations of all elements in all the dichotomies of F. vesiculosus varied considerably over years. A significant decreasing pattern in concentrations of elements such as Co, Hg, Ni and Zn was observed in the tissues from 2015 to 2018 in all SS (Table 2), with concentrations decreasing annually between 10% and 20%. Although Viana et al. (2010) studied temporal changes only in apical dichotomies in this species in the same study area, they observed a significant decrease in the concentrations of Al, Cd, Co, Fe, Hg and Zn of ca. between 5% and 10% from 2001 to 2007. The results of the present study also demonstrate a large degree of intra-annual variation in the concentrations of the elements determined in the F. vesiculosus tissue (Table 1), being comparable to the intra-annual variability found for the same elements in a multitude of studies of temporal trends using macroalgae as biomonitors (i.e. coefficients of variation, CV between 35% and 58%, except N, CV <30%) (see García- Seoane et al., 2018). According to the correlograms calculated and to the models fitted to the time series (Fig. 2, Table 2, Table S3 in Suppl. Mat.), this intra-annual variation in concentrations is not random, and can be reasonably attributed to the existence of seasonality in the bioconcentration process of these elements. Except for Al, Hg and Pb, the other elements showed cyclical fluctuations every 12 months in at least one of the SS, consisting on maximum values in winter and minimum values in summer. Concentrations of Cu, Fe, N and Zn exhibited more than two-fold changes between both seasons. Seasonal variation was previously documented in over 80% of publications assessing the effect of temporal variability in macroalgae, and most of these concluded, as in CHAPTER V 87 the present study, that the highest concentrations of metals typically occur during winter and are lower in summer (García-Seoane et al., 2018). Seasonality was observed in F. vesiculosus for example for As (Stoeppler et al., 1986), Cd (Riget et al., 1995), Zn (Riget et al., 1995; Villares et al., 2013) and δ15N (Lemesle et al., 2015, 2016), but also in other fucoid species, e.g. for Cd, Zn (Miramand & Bentley, 1992; Martin et al., 1997) and Ni (Martin et al., 1997). However, although there has been much discussion about the existence of seasonal fluctuations in bioconcentration of elements in algae, until now it has never been proven using autocorrelation analysis. Furthermore, most of those studies using the term seasonality to describe variations between winter and summer, have done so erroneously from our point of view, since in order to establish the existence of seasonality it is necessary to collect samples at a high frequency (e.g. monthly) repeatedly over several years. Those studies comparing concentrations in different seasons collecting samples for a one-year period or less, cannot clearly state that the variations observed correspond to a seasonal pattern, since it cannot be discarded that these differences are due to a different cause, for example, to an isolated pollution event at the time when the highest concentrations were detected. Finally, since there was no significant interaction between “Sampling month” and “Dichotomy” for any of the elements, it can be concluded that the intra-thallus variation occurs independently of the month/season of the year in which the algae are collected. Seasonal variation in algae concentrations may be attributed to changes in environmental factors such as temperature, precipitation, salinity, pH or light conditions (Stoeppler et al., 1986; Haroon et al., 1995; Wright & Mason, 1999), as the bioavailability of metals in seawater and sediments changes depending on the physico-chemical conditions of the environment (Burdon-Jones et al., 1982; Wright & Mason, 1999). Although seasonal variation could also reflect seasonal changes in the levels of the elements in solution due to occasional fluvial and terrestrial inputs (Lacerda et al., 1985; Fink & Manley, 2011). However, although most authors attributed this pattern to a diluting effect due to growth of algae in the warmest periods (e.g. Riget et al., 1995; Martin et al., 1997; Vasconcelos & Leal, 2001; Villares et al., 2002, 2013), we do not believe that seasonal fluctuations can be explained by the differences in algae growth rates between summer and winter, since as it was mentioned above, algal tissues uptake/release metals until they are balanced with the ambient concentrations, irrespective of their growth rate. Thus, the possibility that the seasonality of concentrations be due to other factors intrinsic to the algae, such as metabolism (photosynthesis and respiration) and reproduction cannot be discarded (e.g. Rao & Indusekhar, 1989; Wright & Mason, 1999). In the case of N, as it is the main limiting nutrient for primary production in marine coastal waters, the occurrence of minimum values in summer may suggests nutrient limitation by this element, as it corresponds to the period of maximum productivity (Villares et al., 2013). Regardless, any one of these possibilities fully explain the seasonal variation differences between elements. Bioconcentration of elements in F. vesiculosus is most likely due to complex inter-relations among environmental and biological factors (Vasconcelos & Leal, 2001; Villares et al., 2002), and the interaction between these factors can result in seasonal cycles of different period. Finally, because the thalli of F. vesiculosus have been found to form several new dichotomies each year, dichotomies can no longer be considered equivalent to annual growth cohorts, as done in biomonitoring studies studying temporality in this species (e.g. Savage & Elmgren, 2004; Carballeira et al., 2014). 5. Conclusions The concentrations of Al, As, Cd, Co, Fe, Hg, Ni, Pb, Zn, N, as well as δ15N signal depend on the tissue age in F. vesiculosus. This species preferentially bioconcentrates As, Cd, Hg, N and δ15N in young dichotomies and Al, Co, Fe, Ni, Pb and Zn in old dichotomies. As the concentrations of nitrogen and trace elements will largely depend on the CHAPTER V 89 dichotomy selected for analysis, we assume that the results from biomonitoring studies that use different sections of the thallus of F. vesiculosus are not strictly comparable. In addition, accordingly to our findings, the concentrations of nitrogen and trace elements in F. vesiculosus are subject to high seasonal variation. Therefore, comparison of samples collected at different times of the year is not acceptable. Thus, the present study proposes the following sampling protocol, that should serve as a guideline for future biomonitoring studies using marine macroalgae on the SS scale: 1. To reduce the effect of variability along F. vesiculosus thallus and to ensure inter-comparability between studies, we propose to make composite samples of the three apical dichotomies of thallus (usually available even in the youngest thalli). 2. To minimize the effect of the seasonal variation and provide annual representative data while optimize the sampling effort, subsamples should be collected 2 times per year separated by 6 months (the time lag between the annual maximum and minimum concentrations) and combined in a single composite sample. Acknowledgements The authors are members of the Galician Competitive Research Group GRC/GPC2016-002 and to the CRETUS Strategic Partnership (AGRUP2015/02), which are co-funded by FEDER (EU). Authors would like to thank the use of RIAIDT-USC analytical facilities. Rita García Seoane is grateful to the Spanish Ministerio de Ciencia, Innovación y Universidades for a grant awarded within the Programa de Formación de Profesorado Universitario (FPU 2014). Villares, R., Puente, X., Carballeira, A., 2002. Seasonal variation and background levels of heavy metals in two green seaweeds. Environ. Pollut. 119(1), 79–90. Wort, D.J., 1955. The seasonal variation in chemical composition of Macrocystis integrifolia and Neroecystis luetkeana in British Colombia coastal waters. Can. J. Botany 33(4), 323–340. Wright, P., Mason, C.F., 1999. Spatial and seasonal variation in heavy metals in the sediments and biota of two adjacent estuaries, the Orwell and the Stour, in eastern England. The Sci. Total Environ. 226, 139– 156. Chapter VI 99 CHAPTER VI Biomonitoring coastal environments with transplanted macroalgae: a methodological review Marine Pollution Bulletin, 135 (2018), 988–999 R. García-Seoane1, J.R. Aboal1, M.T. Boquete2,3, J.A. Fernández1. 1Ecology Unit, Dept. Functional Biology, Universidade de Santiago de Compostela, Fac. Biología, Lope Gómez de Marzoa s/n, Santiago de Compostela, 15782, A Coruña, Spain. 2Estación Biológica de Doñana, CSIC, Avenida Américo Vespucio 25, Isla de la Cartuja, Sevilla 41092, Spain. 3Department of Integrative Biology, University of South Florida, 4202 E Fowler Ave, Tampa, FL 33620, USA. https://doi.org/10.1016/j.marpolbul.2018.08.027 Chapter VII 103 CHAPTER VII Phenotypic differences in heavy metal accumulation in populations of the brown macroalgae Fucus vesiculosus : a transplantation experiment Accepted in Ecological Indicators. R. García-Seoane1, J.R. Aboal1, M.T. Boquete2,3, J.A. Fernández1. 1Ecology Unit, Dept. Functional Biology, Universidade de Santiago de Compostela, Fac. Biología, Lope Gómez de Marzoa s/n, Santiago de Compostela, 15782, A Coruña, Spain. 2Estación Biológica de Doñana, CSIC, Avenida Américo Vespucio 25, Isla de la Cartuja, Sevilla 41092, Spain. 3Department of Integrative Biology, University of South Florida, 4202 E Fowler Ave, Tampa, FL 33620, USA. Abstract The concentrations of Al, As, Cd, Co, Cr, Cu, Fe, Hg, Ni, Pb and Zn in the thalli of Fucus vesiculosus transplanted reciprocally among four sites affected by different degrees of metal pollution (two unpolluted and two industrial environments) were measured with the aim of comparing the capacity of the algae to accumulate these elements under such conditions. At the beginning of the experiment, the concentrations of all elements differed significantly between the individuals from at least one of the unpolluted and one of the polluted sites. After exposure of the algae for 90 days, the concentrations of all of the elements except As, Cd and Cu in individuals transplanted from the unpolluted sites to the most polluted site increased to higher levels than in the algae transplanted within the site of origin. The same was observed for all elements except As and Cd in the thalli transplanted from one of the unpolluted sites to the second most polluted site. By contrast, the concentrations of the elements in transplants from the polluted sites exposed in the unpolluted sites usually decreased to approximately the same levels as in the autotransplants. These results suggest that the heavy metal uptake capacity may be limited in F. vesiculosus populations exposed to long-term pollution as an adaptive response to toxicity by metals. We therefore recommend the use of algal transplants to study water quality in highly polluted sites, rather than of native algae already growing in the sites, to avoid the possible effects of such adaptation. Keywords: Biomonitoring; Marine pollution; Heavy metals; Fucus; Transplants; Adaptation. CHAPTER VII 105 1. Introduction Biomonitor are organisms that provide quantitative information (measurable responses) about environmental changes, whether natural or human-induced, through chemical analysis of their tissues (Hertz, 1991; Carballeira et al., 2000). In the context of heavy metal pollution in marine environments, an ideal biomonitor should be a net accumulator of metals, i.e. an organism in which a simple correlation between the concentrations of metals in its tissues and the bioavailable concentrations in its surrounding environment is observed (Phillips, 1980; Rainbow, 1995; Markert et al., 1999). The linear relationship between exposure concentrations and body burden should be maintained to provide unequivocal information about the levels of metals available in the environment (Stengel & Dring, 2000). Marine macroalgae are often used in biomonitoring programmes to estimate metal pollution in a wide range of environments (Phillips, 1980). Some species of brown algae (Class Phaeophyceae) are able to survive in highly polluted environments, such as areas receiving waste from fish farms (Rönnberg et al., 1992), different types of industries (Marsden et al., 2003), and sewage treatment plants (Oczkowski et al., 2008), and can accumulate high levels of heavy metals in their tissues. Although some metals play important roles in metabolic processes (viz. Cu, Fe and Zn), they may have toxic effects when present at high concentrations and cause physiological stress in algae (Gaur & Rai, 2001; Pearson et al., 2010), affecting growth rates (e.g. Collén et al., 2003), survival rates (e.g. Marsden et al., 2003; Sales et al., 2011) and cover (e.g. Marsden et al., 2003). According to several authors, in algae excess metal can lead to the substitution of essential co-factors and thus to the inactivation of enzymes and proteins and the subsequent inhibition of photosynthesis (Plötz, 1991; Küpper et al., 2002). Several studies have demonstrated the existence of resistance mechanisms in algal populations continuously exposed to high levels of concentrations of metals in the samples transplanted from P2 (for all elements except As, Co and Pb), from U1 (for Co, Fe, Hg and Ni), and from U2 (for Co, Hg and Ni) to P1 were, significantly higher than the levels reached in the autotransplants. A similar pattern of increase in the concentrations was observed in thalli transferred to P2, for fewer elements (i.e. Al, Cr, Cu, Fe, Hg, Pb, and Zn), and with the crosstransplants from U1 (for Al, Cu and Hg) and from P1 (for Pb) reaching significantly higher concentrations than in the autotransplants. On the other hand, the concentrations of all elements in crosstransplants from P1 exposed in U1 and U2 tended to decrease to approximately the same levels as in the autotransplants. The decrease was of more than half the initial values for Cr, Ni, Pb and Zn, and up to ca. sixfold times for Co and Cu. The concentrations of As, Cu, Hg and Zn (for Zn only in U2) decreased in the algae transplanted from P2 to U1 and U2. For the other elements (i.e. Al, Cd, Co, Cr, Fe, Ni, Pb), the concentrations in the crosstransplants from P2 tended to increase, rather than decrease. By contrast, the final concentrations of Cd were significantly lower in crosstransplants from P1 and U2 exposed in U1, and from P1 exposed in U2 than in the corresponding autotransplants. As an exception, the concentrations of Cd in the crosstransplants from U1 and U2 exposed in P1 and P2 were lower than in the autotransplants, whereas the concentrations in the algae transplanted from P2 to U1 and U2 tended to increase above the concentrations in the autotransplants. Finally, the two–way ANOVA test applied to the final concentrations (see Table 1) also revealed a significant exposure site*origin site interaction for Al, Cd, Cu, Fe and Hg. The exposure site had a different effect on the final concentration of metals depending on the site of origin of the algae. Although no clear pattern was observed regarding the significant interactions, the Simple Effects Test showed that the effect in the concentrations was associated with exposure sites P1 and P2 (exceptionally also U1 and U2 in the case of Cd). The uptake/release rates for each element and transplant, along with the variability expressed as the coefficient of variation (CV), were also CHAPTER VII 113 calculated (Table 2). Transplants exposed in unpolluted sites (U1 and U2) showed negative rates (release) for all elements, which were especially high in transplants from P1. Release rates were also observed for As and Cd in transplants transferred to P2. On the contrary, the highest positive rates (uptake) were observed in transplants exposed in the polluted sites (mainly for Al, Cu, Fe and Zn). The CV revealed greater variability in the concentrations of some elements in the autotransplants from P1 that in the crosstransplants (for Al, Co, Cr, Cu, Fe, Ni, Pb and Zn). The same was found in P2 for other elements (i.e. As, Cd and Hg). In U1 and U2, auto and crosstransplants were not clearly distinguished in terms of variability. Fig. 1. Changes in the mean concentrations (µg g-1 d.w. ± SE, n=5) of Al, As, Cd, Co, Cr and Cu in transplants of Fucus vesiculosus exposed in two polluted sites (P1 and P2) and in two unpolluted sites (U1 and U2). Squares and continuous lines: specimens transplanted from the polluted sites (black lines: P1; grey lines: P2). Triangles and discontinuous lines: specimens transplanted from the unpolluted sites (black lines: U1; grey lines: U2). Asterisks indicate the comparisons in which the mean concentrations at initial time were significantly different (p≤0.05) among sites (results of Kruskal–Wallis test). Note that the scale on the ordinate axis is different for polluted and unpolluted sites. t0: initial time; tf: final time (90 days). CHAPTER VII 115 Fig. 2. Changes in the mean concentrations (µg g-1 d.w. ± SE, n=5) of Fe, Hg, Ni, Pb and Zn in transplants of Fucus vesiculosus exposed in two polluted sites (P1 and P2) and in two unpolluted sites (U1 and U2). Squares and continuous lines: specimens transplanted from the polluted sites (black lines: P1; grey lines: P2). Triangles and discontinuous lines: specimens transplanted from the unpolluted sites (black lines: U1; grey lines: U2). Asterisks indicate the comparisons in which the mean concentrations at initial time were significantly different (p≤0.05) among sites (results of Kruskal–Wallis test). Note that the scale on the ordinate axis is different for polluted and unpolluted sites. t0: initial time; tf: final time (90 days). Table 1. Results of the two–way ANOVA comparing final concentrations (tf) in autotransplants and crosstransplants of Fucus vesiculosus, according to the exposure and origen site. F = F statistic; df = degrees of freedom; and p = p value. Significant p values are shown in bold italics (p<0.001), in bold type (p<0.01) and in italics (p<0.05). Final concentrations Source of variation Dependent variable df Mean squares F Sig. (p) Exposure site Al 3 1.09E+07 53.43 <0.001 As 3 7613 30.44 <0.001 Cd 3 1.708 64.20 <0.001 Co 3 96.81 87.46 <0.001 Cr 3 30.31 32.03 <0.001 Cu 3 9412 216.8 <0.001 Fe 3 1.42E+07 89.35 <0.001 Hg 3 1992 192.5 <0.001 Ni 3 66.02 33.76 <0.001 Pb 3 99.34 5.38 0.003 Zn 3 3.31E+05 67.66 <0.001 Origin site Al 3 7.26E+05 3.569 0.020 As 3 2218 8.869 <0.001 Cd 3 0.378 14.22 <0.001 Co 3 0.809 0.731 0.538 Cr 3 1.513 1.599 0.200 Cu 3 218.3 5.028 0.004 Fe 3 8.72E+05 5.474 0.002 Hg 3 29.09 2.810 0.048 Ni 3 4.646 2.376 0.080 Pb 3 4.612 0.250 0.861 Zn 3 2.10E+04 4.291 0.009 Exposure site * Origin site Al 9 5.50E+05 2.703 0.011 As 9 327.0 1.308 0.254 Cd 9 0.094 3.546 0.002 Co 9 1.025 0.926 0.510 Cr 9 1.170 1.237 0.293 Cu 9 244.6 5.635 <0.001 Fe 9 3.45E+05 2.163 0.039 Hg 9 45.69 4.414 <0.001 Ni 9 1.853 0.948 0.493 Pb 9 13.07 0.708 0.699 Zn 9 4287 0.876 0.552 Error Al 55 2.04E+05 As 55 250.1 Cd 55 0.027 Co 55 1.107 Cr 55 0.947 Cu 55 43.42 Fe 55 1.59E+05 Hg 55 10.35 Ni 55 1.955 Pb 55 18.47 Zn 55 4894 CHAPTER VII 117 Table 2. Mean uptake/release rates (ng g-1 day-1) of the elements bioconcentrated in transplants of Fucus vesiculosus exposed for 90 days at each study site (n=5 replicates). For Al, Fe and Zn, the rates are expressed in µg g-1 day-1. Value of the coefficient of variation: CV (%); polluted sites: P1, P2; unpolluted sites: U1, U2. 4. Discussion In the present study we used an experimental design based on a transplantation technique with the objective of studying the differences in heavy metal accumulation capacities among transplants of F. vesiculosus exposed in different environments. To accomplish this, we selected a number of sites with high and low levels of contamination and exposed at each of those sites, algal samples from other sites (crosstransplants) and native algae from the exposure site (autotransplants). We found that our approach (in relation to the choice of sites) was adequate for the established objectives, because the differences in the concentrations among sites at the beginning of the experiment confirmed that P1 was the most polluted site, followed by P2, and that U1 and U2 were not polluted, or were affected by much lower levels of pollution than P1 and P2 (Figs. 2 and 3). Our methodology is an improvement on that used in other studies assessing the response of algae to environmental heavy metal exposure using transplantation techniques (e.g. Eide et al., 1980; Ho, 1984; Amado Filho et al., 1999; Hédouin et al., 2008; Sáez et al., 2015), because these researchers did not use autotransplants, but compared the concentrations in crosstransplants with those in the native algae growing in the site. However, as previously mentioned, this comparison is not strictly valid because the algae collected directly from natural populations and analysed are not subjected to the effect of the transplantation, that may add an extra layer of variation in the final concentrations obtained (Sáez et al., 2015). After comparing the final concentrations of heavy metals in autotransplants and crosstransplants reached after 3 months of exposure in the different sites, we observed the following: i) the concentrations in the transplanted algae from the unpolluted sites (U1 and U2) and from the polluted site P2 (generally less contaminated than P1) increased and exceeded those in the autotransplants from P1 for all elements studied, in at least one of the autotransplants/crosstransplants comparisons, and ii) algae from P1 and P2 transplanted to the unpolluted sites responded CHAPTER VII 119 rapidly to the low metal concentrations, with a large reduction in the metallic content of their tissues to values close to or even lower than the autotransplants for all elements (except Cu and Hg in transplants from P2) (Figs. 2 and 3). These observations confirm, first of all, that metal levels in algae vary dynamically, i.e. algae are able to modify their metal burden in response to the levels of these elements in their surrounding environment. Secondly, the fact that transplants originally growing in less polluted environments are capable of taking up higher levels of metals than native algae when growing under the same conditions, suggests that populations of F. vesiculosus differ in their capacity to take up heavy metals. Although it was initially thought that the accumulation process entailed such a strong bond that the release of metals to seawater was not possible (Munda, 1978), reduction (release) of metal concentrations in algal specimens transferred from polluted to unpolluted sites was previously observed in brown algae (e.g. Eide et al., 1980; Andrade et al., 2006) and described as a metal detoxification mechanism, probably associated with exudation of chelated metal compounds (e.g. Karez & Pereira, 1995; Gledhill et al., 1999; Vasconcelos & Leal, 2001; Pinto et al., 2003). Other authors such as Eide et al. (1980) and Amado Filho et al. (1999) attributed the reduction in metal levels in two species of brown algae to a dilution effect associated with tissue growth during exposure in a metal–free environment. However, our findings are not consistent with the dilution effect hypothesis as we did not observe a reduction in the concentrations of metals in the crosstransplants or autotransplants exposed in the unpolluted sites. On the contrary, concentrations in autotransplants from U1 and U2 remained basically unchanged or increased slightly between t0 and tf, the final concentrations were not significant affected by growth of the algae. The total metal content in the algae is thus the result of the inputs and outputs from different cellular compartments (i.e. extracellular, intercellular and intracellular) plus the particulate matter adhered to the surface of the thallus and intracellular particles. Metals can also be adsorbed extracellularly to the cell wall or taken up and sequestered in cells (Morris et al., 1999; Costas & López, 2001; Fink & Manley, 2011). Polyanionic polysaccharides (negatively charged and composed mainly of alginic acids and sulphated fucans) are the main constituents of the cell wall and intercellular matrix and act as ion-exchanger and ionic barrier (Percival, 1979; Veroy et al., 1980; Andrade et al., 2004). These compounds are able to retain metal cations from the surrounding environment by exchanging them with counter ions (e.g. Na+, Ca+2 and Mg+2) bounded to polyanionic sites, such as amino, carboxyl, phosphate and sulphate functional groups (Percival & McDowell 1967; Myklestad et al., 1978; Żbikowski et al., 2007). The content of a particular metal in the cellular wall will be therefore determined by the ion selectivity of the polyanions and the competition among metallic ions (Haug & Smidsrød, 1967; Sinnott, 2007; Ryan et al., 2012). When absorbed in the cytoplasm, metal ions are chelated by intracellular macromolecules such as polyphenols, metallothioneins and phytochelatins (Myklestad et al., 1978). It has been suggested that polyphenols, which have a high capacity to chelate heavy metals and are present in high proportions in brown algae (Ragan, 1976; Targett et al., 1992), may accumulate some metals in brown algal tissues (e.g. Ragan et al., 1979; Pedersén, 1984; Karez & Pereira, 1995; Hédouin et al., 2008). Based on our knowledge of the mechanisms involved in metal uptake in algae, and on our previous experience, we believe that high proportions of metals in algae are bound to the cell wall polysaccharides and are not localized in the intracellular compartment of the cells, as concluded by Ryan et al. (2012). Our findings in a recent field study with F. vesiculosus transplants (unpublished data), showed that the species is able to take up and release Hg in only 5 days; this would probably not be possible if most of the Hg were localized intracellularly, as it would possibly entail a high cost in terms of energy. We therefore hypothesize that the differences we observed in the capacity of the different populations to take up/release heavy metals (intraspecific variability) may be due to the modification of one or more structural components of the cellular compartments involved in metal uptake, and probably to differences in the cell wall polysaccharide CHAPTER VII 121 composition. Native algae from the most polluted sites (P1 in the present study), subjected to chronic elevated metal exposure, would have suffered such changes in response to polluted-induced stress, and would have a lower metal uptake capacity than those living in slightly polluted or unpolluted environments (e.g. Morris et al., 1999; Ma et al., 2000; Hédouin et al., 2008). Some authors have suggested that adaptation to this type of environments may imply the genetic differentiation of organisms subjected to the selective pressure generated by contamination (Hédouin et al., 2008; Ritter et al., 2010; Sáez et al., 2015), giving rise to ecotypes of the same species with differential tolerance to heavy metal excess (Moenne et al., 2016). For this reason, algal transplants may therefore be more sensitive biomonitors than native algae, reflecting more accurately the level of metal pollution in the environment, as they will not be affected by the resistance mechanisms that may occur in native populations (e.g. Hédouin et al., 2008; Pereira et al., 2014; García-Seoane et al., 2018), in addition to other advantages over the use of native macroalgae, as e.g. the ability to be exposed in places where resident species do not occur (e.g. Søndergaard et al., 2014; Jona-Lasinio et al., 2015). 5. Conclusions The linear relationship between metal concentrations in algae and in the surrounding environment is used to justify the use of these organisms as pollution biomonitors. However, our findings suggest that this relationship may not occur in algae continuously exposed to high concentrations of metals. Hence, we recommend the use of transplants (active biomonitoring) rather than native algae (passive biomonitoring) to assess the level of metal contamination, in order to prevent the effect of changes in the metal uptake capacity (tolerance/resistence mechanisms) as a result of chronic exposure to contamination. proteomics highlights local adaptation of two strains of the model brown alga Ectocarpus siliculosus. Proteomics 10, 2074–2088. Rönnberg, O., Ådjers, K., Roukolathi, C., Bondestam, M., 1992. Effects of fish farming on growth epiphytes and nutrient content of Fucus vesiculosus L. in the Åland archipelago, northern Baltic Sea. Aquat. Bot. 42(2), 109–120. Ryan, S., McLoughlin, P., O'Donovan, O., 2012. A comprehensive study of metal distribution in three main classes of seaweed. Environ. Pollut. 167, 171–177. Sáez, C.A., González, A., Contreras, R., Moody, J., Moenne, A., Brown, M.T., 2015. A novel field transplantation technique reveals intraspecific metal-induced oxidative responses in strains of Ectocarpus siliculosus with different pollution histories. Environ. Pollut. 199, 130– 138. Sales, M., Cebrian, E., Tomas, F., Ballesteros, E., 2011. Pollution impacts and recovery potential in three species of the genus Cystoseira (Fucales, Heterokontophyta). Estuar. Coast. Shelf S 92(3), 347–357. Sinnott, M., 2007. Carbohydrate Chemistry and Biochemistry: Structure and Mechanism. RSC Publishing, Cambridge. Smith, K.L., Hann, A.C., Hrwood, J.L., 1986. The subcellular localisation of absorbed copper in Fucus. Physiol. Plant. 66(4), 692–698. Søndergaard, J., Bach, L., Gustavson, K., 2014. Measuring bioavailable metals using diffusive gradients in thin films (DGT) and transplanted seaweed (Fucus vesiculosus), blue mussels (Mytilus edulis) and sea snails (Littorina saxatilis) suspended from monitoring buoys near a former lead–zinc mine in West Greenland. Mar. Pollut. Bull. 78(1–2), 102–109. Stengel, B.D., Dring, M.J., 2000. Copper and iron concentrations in Ascophyllum nodosum (Fucales, Phaeophyta) from different sites in CHAPTER VII 129 Ireland and after culture experiments in relation to thallus age and epiphytism. J. Exp. Mar. Biol. Ecol. 246(2), 145–161. Targett, N.M., Coen, L.D., Boettcher, A.A., Tanner, C.E., 1992. Biogeographic comparisons of marine algal polyphenolics: evidence against a latitudinal trend. Oecologia 89(4), 464–470. Vasconcelos, M.T.S., Leal, M.F.C., 2001. Antagonistic interactions of Pb and Cd on Cu uptake, growth inhibition and chelator release in the marine algae Emiliania huxleyi. Mar. Chem. 75(1–2), 123–139. Veroy, R.L., Monta, N., de Guzman, M.L.B., Laserna, E.C., Cajipe, G.J., 1980. Studies on the binding of heavy metals to algal polysaccharides from Philippine seaweeds. I. Carrageenan and the binding of lead and cadmium. Bot. Mar. 23(1), 59–62. Viana, I.G., Aboal, J.R., Fernández, J.A., Real, C., Villares, R., Carballeira, A., 2010. Use of macroalgae stored in an Environmental Specimen Bank for application of some European Framework Directives. Water Res. 44(6), 1713–1724. Żbikowski, R., Szefer, P., Latała, A., 2007. Comparison of green algae Cladophora sp. and Enteromorpha sp. as potential biomonitors of chemical elements in the southern Baltic. Sci. Total Environ. 387(1–3), 320–332. 131 APPENDIX. SUPPLEMENTARY MATERIAL APPRENDIX. SUPPLEMENTARY MATERIAL (CHAPTER V) Fig. S1. Boxplots of the element concentrations in the five dichotomies of the thallus of Fucus vesiculosus (1st–apical, 5th–basal) from all sites under study. The whiskers extend from 1.5 IQR (interquartile range) of the lower quartile to the 1.5 IQR of the upper quartile. Those data points that exceed these limits are drawn outside the boxplot. Different letter(s) above the error bars denote significant differences of mean concentrations among dichotomies based on results from Tukey’s Post–Hoc test (p≤0.05). The concentrations are expressed in µg g-1 (except *: % and **: ‰). In each boxplot: n=108 (1st, 2nd and 3rd dichotomies), n=72 (4th dichotomy), and n=36 (5th dichotomy). For Al and Pb, n value is respectively one-third and two-thirds lower in each dichotomy, as some of the sampling sites were eliminated from data analysis. 133 Fig. S2. Growth progression in one Fucus vesiculosus individual from site 2 (SS2). A) January; B) April; C) July. Growth monitoring was not possible in October. Table S1. Description of the general periodic regression model and its parameters fitted to the time series. Table S2. Results of the Three-way ANOVA test comparing the concentrations of elements in the five dichotomous sections of the thallus of Fucus vesiculosus. df=degrees of freedom; SS=sum of squares; MS=mean square; F=F statistical value; Sig.=statistical significance level (p–value). Significant p–values are shown in bold type. Source Variable df SS MS F Sig. Month Al 11 267 24.3 1.41 .166 As 11 1.16 0.105 1.93 .036 Cd 11 4.44 0.404 10.0 .000 Co 11 18.8 1.72 14.2 .000 Cu 11 2.62 0.238 13.3 .000 Fe 11 0.992 0.090 2.21 .014 Hg 11 0.308 0.028 7.35 .000 Ni 11 9.61 0.874 9.77 .000 Pb 11 0.676 0.061 1.09 .373 Zn 11 0.380 0.034 54.2 .000 N 11 7.05 0.641 79.1 .000 δ15N 11 3.57 0.324 3.75 .000 Site Al 2 3007 1504 87.7 .000 As 2 2.56 1.28 23.5 .000 Cd 2 7.32 3.66 90.8 .000 Co 2 19.4 9.69 80.2 .000 Cu 2 5.92 2.96 165 .000 Fe 2 7.41 3.71 90.9 .000 Hg 2 0.149 0.075 19.6 .000 Ni 2 0.794 0.397 4.44 .012 Pb 2 19.9 9.96 176 .000 Zn 2 0.092 0.046 72.2 .000 N 2 3.53 1.76 218 .000 δ15N 2 24.1 12.0 139 .000 Dichotomy Al 4 187 46.9 2.73 .029 As 4 10.9 2.73 50.2 .000 Cd 4 13.1 3.28 81.3 .000 Co 4 10.7 2.68 22.2 .000 Cu 4 0.542 0.135 7.56 .000 Fe 4 3.11 0.777 19.1 .000 Hg 4 0.072 0.018 4.70 .001 Ni 4 20.6 5.16 57.7 .000 Pb 4 5.55 1.39 24.5 .000 Zn 4 0.087 0.022 34.3 .000 N 4 2.17 0.542 66.9 .000 δ15N 4 9.26 2.32 26.8 .000 135 TABLE S2 (Continued) Source Variable df SS MS F Sig. Month * Site Al 22 711 32.3 1.88 .011 As 22 0.282 0.013 0.235 1.000 Cd 22 1.96 0.089 2.21 .002 Co 22 1.69 0.077 0.635 .897 Cu 22 1.49 0.068 3.78 .000 Fe 22 1.76 0.080 1.97 .007 Hg 22 0.099 0.005 1.19 .255 Ni 22 2.87 0.130 1.46 .089 Pb 22 1.66 0.075 1.33 .151 Zn 22 0.034 0.001 2.41 .000 N 22 1.23 0.056 6.93 .000 δ15N 22 5.81 0.264 3.05 .012 Month * Dichotomy Al 44 200 4.54 0.265 1.000 As 44 0.335 0.008 0.140 1.000 Cd 44 0.663 0.015 0.373 1.000 Co 44 1.63 0.037 0.307 1.000 Cu 44 0.693 0.016 0.879 .689 Fe 44 0.349 0.008 0.195 1.000 Hg 44 0.075 0.002 0.451 1.000 Ni 44 1.61 0.037 0.410 1.000 Pb 44 0.618 0.014 0.248 1.000 Zn 44 0.030 0.001 1.08 .341 N 44 0.130 0.003 0.364 1.000 δ15N 44 1.51 0.034 0.398 1.000 Site * Dichotomy Al 8 191 23.9 1.39 .200 As 8 0.122 0.015 0.280 .972 Cd 8 0.248 0.031 0.769 .630 Co 8 0.143 0.018 0.148 .997 Cu 8 0.186 0.023 1.30 .244 Fe 8 0.095 0.012 0.293 .968 Hg 8 0.018 0.002 0.608 .771 Ni 8 0.305 0.038 0.426 .905 Pb 8 0.208 0.026 0.460 .884 Zn 8 0.008 0.001 1.65 .110 N 8 0.021 0.003 0.328 .954 δ15N 8 0.113 0.014 0.163 .995 TABLE S2 (Continued) Source Variable df SS MS F Sig. Month * Site * Dichotomy Al 88 714 8.11 0.473 1.000 As 88 0.475 0.005 0.099 1.000 Cd 88 1.13 0.013 0.318 1.000 Co 88 1.29 0.015 0.121 1.000 Cu 88 0.703 0.008 0.446 1.000 Fe 88 1.12 0.013 0.313 1.000 Hg 88 0.112 0.001 0.334 1.000 Ni 88 1.18 0.013 0.150 1.000 Pb 88 1.77 0.020 0.355 1.000 Zn 88 0.010 0.000 0.177 1.000 N 88 0.236 0.003 0.331 1.000 δ15N 88 1.39 0.016 0.184 1.000 Residuals Al 233 3993 17.1 As 252 13.7 0.054 Cd 251 10.1 0.040 Co 250 30.2 0.121 Cu 250 4.48 0.018 Fe 252 10.3 0.041 Hg 250 0.953 0.004 Ni 251 22.4 0.089 Pb 252 14.3 0.057 Zn 251 0.162 0.001 N 252 2.04 0.008 δ15N 252 21.8 0.086 137 Table S3. Description of the correlation patterns in time series of the different elements, for each dichotomy of the Fucus vesiculosus thallus at the sampling sites (SS). Correlations are showed from lag 1 to lag 35. Upward and downward arrows represent positive and negative correlation coefficients, respectively. Significant correlations (whether positive or negative) are showed in bold type. Background shading has been included to facilitate interpretation of correlation patterns. those in the surrounding environment (Philips, 1990). The results reported in Chapter VII indicate that the linearity of the bioconcentration process in algae growing nearby pollution sources has been altered, breaking one of the basic principles of biomonitoring. For this reason, we proposed the use of algae transplants over the use of native algae for the quantitative evaluation of water quality in coastal systems at least when sources of pollution are known in the area. Finally, despite the advances reported in this dissertation, and as stated in the two literature reviews (Chapters I and VI), some other key methodological aspects of the biomonitoring technique using macroalgae still remain to be standardized. Some aspects related to the use of native algae comprise e.g. issues related with the sample collection (as the size of the SS), and other aspects associated with the sample processing (as the sample cleaning method before chemical analysis to remove particulate material adhered to the thallus surface). Although the use of transplantation techniques has been less frequently implemented in biomonitoring studies (see Chapter VI), it offers certain advantages over the use of native plants as e.g.: i) enables studies to be extended to areas where native populations of the algae species of interest are absent (Fernandes et al., 2012; Jona-Lasinio et al., 2015) and, ii) avoids the influence of some factors that may affect the interpretation of the results, since reduces the effect of the possible spatial variability of concentrations in native algae at the SS (Alquezar et al., 2013), improves the temporal interpretation of the results because the time of exposure is known (Sáez et al., 2015) and, as discussed in Chapter VII, eliminates the possible phenotypic variation of native algae in highly polluted sites (Hédouin et al., 2008). Therefore, future efforts to standardize procedures should be directed towards the optimization of the protocols implemented in monitoring technique using algae transplants rather than native algae. General Conclusions GENERAL CONCLUSIONS 147 It has been demonstrated that Fucus vesiculosus can be used to establish the natural range of variability in concentrations of δ15N, N and Hg in coastal areas. The results also showed that a large part of this natural variability is explained by the presence of a spatial pattern (deterministic component of the variability) that is characteristic of each zone and each contaminant under study. The evaluation of the natural range of variability in concentrations of δ15N using this species was useful to detect the presence of sources of small scale contamination by δ15N, but also enabled to determine the spatial representativeness of the sampling site (SS) (the SS can me moved a distance of up to 900 m without yielding large variations in the δ15N signal, at least for the SS studied). High inter-individual variability was observed between subsamples of F. vesiculosus at 3 different SS, with concentrations varying longitudinally and transversally along the SS. This revealed that the position of the subsamples within the SS influences the bioconcentrated concentrations in the tissues. The findings evidenced that, to minimize the spatial intra- SS variability in the concentrations and adequately represent the concentrations within the SS, a minimum of 30 subsamples evenly distributed should be collected within three bands parallel to the coastline along a 30–50 m wide stretch (10 subsamples in each band), and also at different heights on the rocks if possible, and bulked in a single composite sample. The ability to differentiate significantly the mean concentrations of any element between two SS using F. vesiculosus will depend on the number of samples of this species collected and individually analyzed at each SS. The evidences provided in this Thesis allowed to conclude that the number of samples needed to detect such differences will be largely determined by the mean and variance in the concentrations of elements in algae tissues at each SS. Taking into account the results, a minimum of 20 samples should be collected at each SS to detect significant differences for elements whose concentrations differed by more than 30% between the SS compared. Factors as the time of collection and the part of the algal thallus selected for chemical analysis should be taken into account when standardize the methodology for monitoring pollution levels using algae. Our findings demonstrated the existence of high intra-annual/seasonal, as well as high intra-thallus variation in the concentrations of elements in F. vesiculosus tissues. In order to minimize the influence of these sources of variability on the data obtained in biomonitoring studies and to ensure a good representativeness of the results, we suggested that subsamples should be collected 2 times per year separated by 6 months and combined in a composite sample. To minimize the effect of variability along F. vesiculosus thallus, it is proposed to make composite samples of the three apical dichotomies of thallus. The ability of F. vesiculosus to take up and/or release elements can be altered when native populations of this algae grow subjected to long-term and high heavy metal pollution levels. We hypothesized that the stress induced by metallic toxicity has altered some of the mechanisms involved in the process of bioconcentration of these elements, restricting their ability to concentrate these compounds. This phenotypic alteration is interpreted as a response of the algae to overcome such extreme conditions. Therefore, we recommend to use algae transplants rather than native algae for biomonitoring studies of water quality in order to prevent the possible effect of such adaptations. An exhaustive study of the methods used in passive biomonitoring studies with macroalgae revealed that, up to date, there is no standard protocol that enables the widespread implementation of this technique. This lack of standard methodology hampers routine use of these organisms in biomonitoring programmes. Thus, with the objective of normalizing the use of native algae as environmental tools for water quality assessment, we proposed a standard protocol. Although most of the methodological aspects included in this protocol have not yet been thoroughly investigated, our recommendations are based on the conclusions of previous studies involving this technique, especially those who considered methodological issues and, of course, in the results from the experimental studies carried out in the course of this Doctoral Thesis (the latter GENERAL CONCLUSIONS 149 are shown in bold type in the Recommendations section). The protocol is described below: Methodological issues Recommendations 1. Pre-sampling issues 1.1 Sampling design Regular sampling (grids) or transects 1.2 Number of sampling sites (SS) As many as possible (based on the scope, degree of spatial resolution, economical and operational constraints of the survey) 1.3 Sampling density 1 SS/16 km or 1 SS/32 km2 (1 SS per 16/2n km or 1 SS per 32/2n km2, when more resolution is needed) 1.4 Survey frequency As frequent as possible (economic limitations), but firstly check for seasonality and/or intra-annual variation 1.5 Sampling period 2 samplings per year separated by 6 months (annual composite sample) 1.6 Algae species 1 species (Ulva sp. or Fucus sp.). If more than 1 species is used, qualitative interpretation 2. Sample collection 2.1 Nº of samples/subsamples To significantly differentiate the levels of pollution between SS: minimum 20 samples (individually analyzed) per SS To represent the intra-site variability in concentrations: composite sample (minimum 30 subsamples) 2.2 Amount of sample collected Minimum of 5 g dry weight (d.w.) 2.3 Subsamples collection strategy Evenly distributed in the SS within 3 bands parallel to the coastline (10 subsamples/band), at different heights on the rocks if possible 2.4 Area occupied by the SS Between 200 and 500 m2 2.5 Size of the SS 30–50 m wide stretch 2.6 Moment of collection Low tide (intertidal zone) 2.7 Collection depth Less than 1 m deep in intertidal zone at low tide 2.8 Spatial representativeness of the SS SS at 900 m away from the original SS 2.9 Distance between SS and pollution focus Further research 2.10 Sampling precautions Use of plastic gloves (pre-cleaned and trace metal free)/sample only material attached to substrate/no smoking 2.11 In situ cleaning Washed thoroughly in seawater in the field 2.12 Transportation In sealed plastic bags in refrigerated conditions 2.13 Storage pre-processing At 4°C no more than 6 days. Never freeze samples before analysis 3. Sample processing 3.1 Cleaning Manual cleaning and scraping of epiphytes, without washing 3.2 Selection of material for analysis Composite sample of 3 apical dichotomies in Fucus sp. 3.3 Drying Lyophilizing or drying at 40°C + an aliquot at 105ºC (d.w.) 3.4 Homogenization Ground in metal free mills (not sieve) 3.5 Certified reference material for analysis Certified reference material for macroalgae 4. Additional issues 4.1. Correction for particulate contamination Apply at least once per site to rule out the possible contribution of sediment Since the development of methodology for application of the active biomonitoring technique with algae is practically inexistent, the protocol proposed below is supported by the findings reported in the available literature. In this respect, methodological issues which have never been studied in detail, like the preparation of the algae for transplant (including acclimation, cloning and culture in the laboratory), the vital status of the transplants (whether using living or devitalized material), and exposure of the transplants (e.g. duration of exposure and exposure conditions) should be addressed in the near future. Methodological issues Recommendations for species with laminar growth Recommendations for species with apical growth 1. Selection and preparation of the macroalgae 1.1 Monitoring species 1 single species (Ulva sp.) 1 single species (Fucus sp.) 1.2 Selection of material for transplant Discs Whole thalli 2. Preparation of the transplants 2.1 Pre-exposure treatments Acclimation to low nutrient conditions in studies on nutrients Acclimation to low nutrient conditions in studies on nutrients Further research Further research 2.2 Sample cleaning Washed in seawater in the field Washed in seawater in the field 2.3 Cloning and culture in the laboratory Further research Further research 2.4 Type of transplant Cages Bags 2.5 Amount of algae/transplant 1 g fresh weight (f.w.)/30 cm3 1 g f.w./100 cm2 2.6 Vital status Further research Further research 3. Exposure of the transplants 3.1 Support Buoys Buoys 3.2 Depth of exposure ca. 1 m ca. 1 m 3.3 Duration of exposure Further research Further research 3.4 Number of replicates/transplant Minimum of 3 Minimum of 3 3.5 Initial concentrations and controls 3–5 initial times and 3 controls 3–5 initial times and 3 controls 4. Post-exposure treatments 4.1 Transportation Under refrigeration Under refrigeration 4.2 Storage pre-processing At 4°C no more than 6 days. Samples never frozen before analysis At 4°C no more than 6 days. Samples never frozen before analysis 4.3 Cleaning Washed in seawater in the field and manual cleaning Washed in seawater in the field and manual cleaning 4.4 Selection of material for analysis Algal discs Green parts of the shoot or apical segments (2–3 cm) 4.5 Drying Lyophilizing or drying at 40°C for 24 h + an aliquot at 105ºC (d.w.) Lyophilizing or drying at 40°C for 24 h + an aliquot at 105ºC (d.w.) 4.6 Homogenization Milling in metal free mills Milling in metal free mills 4.7 Certified reference material for analysis Certified reference material for macroalgae Certified reference material for macroalgae General References GENERAL REFERENCES 153 Alquezar, R., Glendenning, L., Costanzo, S., 2013. The use of the brown macroalgae, Sargassum flavicans, as a potential bioindicator of industrial nutrient enrichment. Mar. Pollut. Bull. 77, 140–146. Bárbara, I., Cremades, J., Pérez-Cirera, J.L., 1995. Zonación de la vegetación bentónica marina en la Ría de A Coruña (N.O. de España). NACC Nova Acta Cient. Compostel. Biol. 5, 5–23. Beiras, R., 2018. Marine Pollution. Sources, Fate and Effects of Pollutants in Coastal Ecosystems. Elsevier, pp. 3–20. Bergström, L., Tatarenkov, A., Johannesson, K., Jonsson, R.B., Kautsky, L., 2005. Genetic and morphological identification of Fucus radicans sp. nov. (Fucales, Phaeophyceae) in the brackish Baltic Sea. J. Phycol. 41, 1025–1038. Bradl, H., 2005. (Ed). Heavy Metals in the Environment: Origin, Interaction and Remediation Volume 6. Academic Press, London. Carlson, L., 1991. Seasonal variation in growth, reproduction and nitrogen content of Fucus vesiculosus L. in the Öresund, Southern Sweden. Bot. Mar. 34(5), 447–454. Carpenter, P.D., Butler, E.C.V., Higgins, H.W., Mackey, D.J., Nichols, P.D., 1991. Chemistry of trace elements, humic substances and sedimentary organic matter in Macquarie Harbour, Tasmania. Mar. Freshw. Res. 42(6), 625–654. Carral, E., Villares, R., Puente, X., Carballeira, A., 1995. Influence of watershed lithology on heavy metal levels in estuarine sediments and organisms in Galicia (North-West Spain). Mar. Pollut. Bull. 30(9), 604– 608. Conti, M.E., Finoia, M.G., 2010. Metals in molluscs and algae: A northsouth Tyrrhenian Sea baseline. J. Hazard. Mater. 181(1–3), 388–392.