Micropollutant content of Sargassum drifted ashore: arsenic and chlordecone threat assessment and management recommendations for the Caribbean

Massive Sargassum beachings occurred since 2011 on Caribbean shores. Sargassum inundation events currently involve two species, namely S. fluitans and S. natans circulating and blooming along the North Atlantic subtropical gyre and in the entire Caribbean region up to the Gulf of Mexico. Like other brown seaweeds, Sargassum have been shown to bioaccumulate a large number of heavy metals, alongside with some organic compounds including the contamination by historical chlordecone pollution in French West Indies (FWI), an insecticide used against the banana’s weevil Cosmopolites sordidus. The present study reports, during two successive years, the concentration levels of heavy metals including arsenic in Martinique and Guadeloupe (FWI). We found that Sargassum can also accumulate a high concentration of chlordecone. Sargassum contamination by chlordecone is observed in areas close to contaminated river mouth but can be partly due to chlordecone desorption when secondary drifted on chlordecone-free shore. Our results further demonstrate that algae bleaching raises a number of questions about inorganic and organic pollutant (i) bioaccumulation, at sea for arsenic and close to river plumes for chlordecone, (ii) transport, and (iii) dissemination, depending the shoreline and the speciation for arsenic and/or metabolization for both.


Introduction
Sargassum (Fucales, Phaeophyceae) are present worldwide in tropical and subtropical environments (Niermann, 1986). All the described Sargassum species are benthic except two pelagic ones: Sargassum natans and Sargassum fluitans, initially reported by C. Columbus (Lapointe and Hanisak, 1985), and mainly confined to the Gulf of Mexico and Sargasso Sea, where they are found accreted together into lines and mats (Marmorino et al., 2011). Even if climatic events (e.g., hurricanes) could spread them onshore, pelagic Sargassum landings in massive amounts were never reported elsewhere than the Caribbean area (Loffler and Hoey, 2017). However, since 2011, large beaching events have been occurring on West Africa and Greater Caribbean shores (van Tussenbroek et al., 2017). Firstly episodic, such events tend to be more common in part due to an unexpected growth location in front of the Amazonian mouth, and volumes involved are yearly expressed in millions of tons of algae drifted to shores (Wang et al., 2019). Sargassum piled up on shores and perishes, producing foul and corrosive hydrogen sulfate, oxygen depletion in water, and colloidal bleed is observed (Brylinsky, 1977), and can also have impacts on human health (Resiere et al., 2018). Such events overwhelmed the public authorities, endangering economic activities, mainly tourism and port ones (Langin, 2018).
Seaweeds are well known to concentrate metals from seawater in the different parts of their thallus, and a wide range of essential and non-essential metals have been measured (Malea and Kevrekidis, 2014;Bonanno and Orlando-Bonaca, 2018;García Seoane et al., 2018). Among heavy metals, arsenic (As) is a notorious and toxic metalloid, ubiquitous in the environments which is accumulated in algae including Sargassum spp. and can potentially also be found in algal food or derived products (Devault et al., 2020a;Magura et al., 2019;Milledge et al., 2020 ;Rodriguez-Martinez et al., 2020).
Arsenic chemical behavior in the environment is close to phosphorus (Neff, 2002). As such, phosphate and As(V) can compete for adsorption sites (Neff, 1997). Arsenate (As(V)) is predominant in inorganic aqueous and aerobic environments and is strongly adsorbed onto the surface of several aquatic organisms and oxidized minerals of Fe, Mn, and Al (Al Mamun et al., 2018), including inner iron plaques in Sargassum. Arsenite (As(III)) is highly abundant under anoxic environments as oxyanion (Smedley and Kinniburgh 2002) and more toxic than As(V) (Ferguson and Gavis, 1972). As can change in the environment quickly and repeatedly from one species to the other; however, methylation preserves the As speciation in organic As compounds.
Along with heavy metals, and especially As, Sargassum spp. are known to accumulate organic micropollutants, even if poorly documented, as summarized by Devault et al. (2020a): polycyclic aromatic hydrocarbons accumulated into algae driven by aromaticity level (Bertilsson and Widenfalk, 2002;Tobiszewski and Namiesnik, 2012;Seepersaud et al., 2018;Stout et al. 2017) confirmed for S. natans and S. fluitans during the offshore Deepwater Horizon MC 252 blowout in the Gulf of Mexico and the subsequent oil spill. For pesticides, sorption behavior is rather complex. Even if hydrophobicity broadly induces similar behavior, especially for older formulations, more recent pesticides are partly designed to delude plant metabolism.
Chlordecone (CLD) is also a well-known pesticide impacting French West Indies (Cavelier 1980;Bocquené and Franco, 2005;Coat et al., 2006;Dubuisson et al., 2007, Gourcy et al., 2009Le Déaut and Procaccia, 2009;Jondreville et al. 2014;Devault et al., 2016;among many). Notwithstanding the worldwide spread of this pesticide, the climactic area where CLD was spread is the West Indies. French West Indies were selected in order to determine the extent of such pollution: French West Indies were selected to study the As and chlordecone content of Sargassum drifted to shore. Moreover, French public agency for energy and environmental management (ADEME) supports the attempts for collecting and valorizing the large volumes of Sargassum to turn this disadvantage into an asset: contamination assessment has to be performed in order to substantiate such perspectives. In this article, authors focused on Sargassum beached in Martinique and Guadeloupe shores. Our study aims at quantifying both (i) the organic micropollutant chlordecone and its derived metabolites and (ii) the different species of arsenic present in Sargassum samples found along the shore of two islands of the French West Indies during two successive years.

Sampling strategy
Sampling was performed during two campaigns in Martinique and in Guadeloupe. The 2018 campaign was between the 26th and 28th in July for the Guadeloupe Archipelago, between July 30 and August 2 in Martinique, and from March 12 to April 4 in Guadeloupe. Most of the samples were collected either on the beach or nearshore (with a maximum of 500 m from the coast). For the 2019 campaign, most of the time we collected at the same site two onshore samples, one sample of fresh Sargassum beached on the sand and one sample corresponding to dry Sargassum again lying on the sand.
Samples were collected, during the first survey, in order to be representative to the low-scale heterogeneity, i.e., sampled threefold at few distance on the shore of the same site. Samples collected during the second survey for each site were collected following the four following conditions: (1) grounded and at least partly dried and browned, being drifted to shore but not yet grounded; (2) recently grounded, i.e., still wet; (3) close to being grounded, in the beating of the waves; and (4) still in seawater. Each condition required was filled as possible: sampling was submitted to presence or absence of algae at sampling date and site. Practically, sampling was not significantly impacted and lacking samples were rare.

Investigated sites
Sampled sites were selected because they are on the windward coast of Guadeloupe and Martinique. The Basse-Terre windward coast in Guadeloupe, i.e., the mountain part, and the north of Martinique windward coast are known to have surface water and groundwater discharge contaminated by chlordecone. Moreover, Sargassum mats were being drifted to the shore by wind and currents, and windward coast is the most impacted by Sargassum groundings, whatever the island-so authors sampled only windward coast, i.e., eastward coast. The distance between sampling sites in Guadeloupe was 10 km and in Martinique was 5 km. Over the 14 sampled sites in Guadeloupe, 7 were in Basse-Terre shores and 7 in Grande-Terre shores. Over the 20 sampled sites in Martinique, 9 were in an area expected to be contaminated by chlordecone (restricted area for fishing and gathering), 8 were apart from the restricted area, and 3 were piled-up area, i.e., sampled in Sargassum piles, formed by drifted algae brought together in order to de-clog the ports and beaches. The three piles were only sampled during the first campaign. The sampling locations are presented in Fig. 1.

General sample preparation
After material sampling in Guadeloupe and Martinique, samples were frozen in dry ice and dispatched to La Drôme Laboratoire in metropolitan France in the 48 h following sampling. The overall sample quantity was grinded with a Grindomix GM 200 Retsch apparatus till < 63 μm. After this first grinding, dry matter and mineral matter were determined thanks to NF EN 12880 (Chiffre et al., 2015) and NF EN 15169 standard (Dia et al., 2019), respectively.

TOC and N-Kjeldahl analysis
Total organic content (TOC) and N-Kjeldahl were carried out after freeze-drying of previous grinded samples on a CHRIST Alpha 1-2 LD Plus. TOC was determined following the NF ISO 14235 (Remon et al., 2005) operating procedure including hot sulfochromic oxidation and spectrocolorimetry quantification (LQ = 1000 mg (C) per kg of dried sample (DS)). N-Kjeldahl analysis followed the NF EN 13342 standard involving hot sulfuric acid mineralization with selenium catalyst (Guilayn et al., 2020). Titrimetry analysis allowed to determine N-Kjeldahl content (LQ = 200 mg (N) per kg of dried sample (DS)).

Metal and semi-metal speciation analysis
The list of analyzed metals included As, Cd, Cr, Co, Cu, Ni, Pb, Zn, Hg, P, and S. For the overall metal content and metal speciation, previous grinded samples were mineralized with nitric acid in micro-waves following the NF EN 13805 procedure (Chevallier et al., 2015). Quantitative metal analysis was achieved thanks to ICP-MS X-serie apparatus from Thermo Fisher. Metal determination followed the NF EN 17294-2 standard (Glorennec et al., 2010). For total metal content, the limits of quantification in mg (metal) per kg of dried sample (DS) are listed in Table 1. Fig. 1 Maps of related Martinique and Guadeloupe (sources: Direction de la Mer). Because of putative contamination by chlordecone, some of the marine areas within which fishing is totally forbidden (red) or restricted (purple) are presented 1.00 mg (Zn)/kg dw Metal speciation allowed to determine As and Hg species. As compounds were As (total), AsIII (arsenite), AsV (arsenate), MMA (monomethylarsonic acid), DMA (dimethylarsinic acid), AsB (arsenobetain), and AsC (arsenocholin). Hg compounds were Hg (total), iHg (inorganic Hg), and MMeHg (monomethylHg). Arsenic speciation analysis was carried out on a ICP-MS Nexion 300 apparatus from PerkinElmer coupled with a liquid chromatography supplied by Flexar. Hg speciation was carried out on the same ICP-MS Nexion 300 apparatus from PerkinElmer but coupled with a gas chromatography 680 supplied by Clarus. For all metal species, the limit of quantification is summarized in Table 2.

Chlordecone and chlordecone metabolite analysis
Following the protocol detailed in Devault et al. (2016), chlordecone and chlordecone metabolite (5b-hydro and chlordecol) analysis was carried out on the previous grinded samples. One hundred fifty microliters of 13 C chlordecone tracer was added to 10 g of grinded sample. Twenty milliliters of acetone was mixed and stirred with the previous mixing for 24 h. After filtration, samples were washed two times with 5 mL of acetone. Chlordecone compounds were extracted after stirring with a dichloromethane solution (20 mL of CH 2 Cl 2 in 175 mL of H 2 O saturated with NaCl salt). After two extractions, the CH 2 Cl 2 recovered was concentrated to 1 mL by dried evaporation (N 2 ). One milliliter of hexane was added to the previous CH 2 Cl 2 concentrated solution. For liquid chromatography analysis, the sample had to be changed. Chlordecone compounds were extracted from 200 μL of CH 2 Cl 2 and hexane solution thanks to 1 mL of H 2 O/acetonitrile solution. Chlordecone compounds were analyzed by Acquity UPLC System with XevoTQ-S Mass spectrometer (LC-MS-MS). LC condition and MS-MS result exploitation are presented in Table 3. The LQ in μg (chlordecone compound) per kg ww is listed in Table 4.

Data analysis
Statistical analyses were performed with R version 4.0.3 (R Core Team (2020), R: A language and environment for statistical computing, R Foundation for Statistical Computing, Vienna, Austria, URL https://www.R-project.org/). Data were visualized using the ggplot2 package (Wickham, 2016). Differences in chemical concentrations or ratios between treatments were evaluated a Kruskal-Wallis ranksum test followed by a Dunn test with Benjamini-Hochberg corrections.

Results and discussion
Contamination pattern in Guadeloupe and in Martinique involves both (1) Sargassum offshore contaminated by oceanic As, mainly resulting from the human activities at global or local scales, and (2) Sargassum contamination by chlordecone in the coastal environment due to water resource and soil pollution by banana weevil control during the 1970s   (Bocquené and Franco, 2005). All the results are detailed in Appendix 1.

Arsenic
As content offshore has been ensured by Atlantic Ocean samplings in front of Santa Lucia Channel (identified as "Transect" in Table 5 and 6 of Appendix), As(V) was the dominant form of arsenic (mean ± standard deviation, 58.2 ± 18.0% of total As), corroborating international literature (Michel, 1985). Climactic As content (125.7 mg/kg dw) is observed since pelagic sampling sites and higher As concentrations in Sargassum ashore, whatever island or sampling sites, are related to drifting algae in opposition to dried ones ( Fig. 2): ashore, Sargassum transudate, and then leak As (Devault et al., 2020a, b). We found significantly less total As, As(V), MMA, and AsBet in the dried samples from the beach than in the floating algae (P < 0.05). For the As(III) and DMA species, no significant differences were observed (Fig. 2). Sargassum are floating algae and beaching often involve a first step of compression of massive Sargassum biomass along the shoreline, during which Sargassum density will exceed the density in the mat. At this step, which have to be more studied to properly understand the inducting parameter(s), Sargassum will be stressed and will transudate. Transudation content is poorly known but authors identified that chlordecone and As are significantly transudate in few hours for floating and sank algae (Devault et al., 2021). Due to tide, gravity waves, and wind, a part of Sargassum could beach and be dried in the sun; due to rain and tissue degradation, Sargassum will leach; this leach can reach 22,000 μg/L (Chevalier, pers. com.). Another part can sink and, in case of cup-shaped dent coastline from which Sargassum could not drift away, Sargassum biomass densely accreted could form a pseudo-vase by sedimentation in shallows. In the present study, authors did not focus on such environments, but limited them to floating Sargassum, beached and fresh ones, beached and dried ones, and piled ones, but two sampling sites correspond to an extreme accretion situation.
Comparing the results of our study with the current literature is delicate because the analysis proposed elsewhere about As was mainly performed on total As content, even including inorganic As species (Maher et al., 2009), focuses on the protocol enhancement (Han et al., 2009) or the metabolic pathways (Edmonds et al., 2009;Pichler et al., 2006;Leal-Acosta et al., 2013) not about detailed environmental monitoring (Devault et al., 2020a) or did not include As speciation detail (Rodríguez-Martínez et al., 2019). As such, the detailed protocol proposed in the present article provided new insights into As dynamics in algae. Total As concentrations in pelagic Sargassum exceed those known for other algal species but are in agreement with results found in previous studies for those species (Michel, 1985;van Tussenbroek et al., 2017). This upper (at least double) As concentration could be due to pelagic conditions: the other Sargassum species are benthic and phosphate concentration in shallow waters is higher than in the open oceans, implying a higher need of this element. Benson (1984) and fellows hypothesized that orthophosphate and As are competitive for phosphate transporter and that S. natans and fluitans overuse of phosphate transporter to compensate the phosphate lack in the open ocean, increasing As accumulation. Total As content heterogeneity is limited: if piled Sargassum presents a significantly lower As content than other (26.6 ± 20.9 μg/g and 80.9 ± 29.8 μg/g dw, respectively), piled Sargassum sampling sites, located in Martinique, are scarce. Total concentration, at 80.9 ± 29.8 μg/g dw (minimum 28.8 and maximum 127 μg/g dw), is very in agreement with Rodríguez-Martínez et al. (2019) who reported a median of 80, a minimum of 24, and a maximum of 172 μg/g dw. This statement has to be limited to total As but, based on the 20 million ton estimation per month of floating Sargassum drifted to Caribbean shores (Wang et al., 2019), even in considering a minimized 50 μg/g dw concentration, the total flux of As is yearly about 1000 ton. Regarding the sites where algae were not piled, a significant distinction could be made between the two places where "pseudo-vase" (previously defined as densely accreted, Sargassum amounts decay in shallow waters) was sampled: at surface of the amounts, algae will be parched by sun, forming a few centimeter-thick compact crust (able to support a crawling adult but not a walking one) covering decimeter-thick wet rotten algae. They are stuck in place due to the limited tidal range (about 40 cm) and permanent trade winds and currents. Such sites have total As content not significantly different to piled ones but significantly different with other non-piled Sargassum (respectively 22.6 ± 13.5 μg/g and 83.4 ± 27.7 μg/g); in such clogged places, interstitial water into the pseudo-vase reaches more than 1000 μg/L due to As transudate from Sargassum but also due to limited water circulation. For such concentrations, the sanitary impact of As is unknown: the highest contamination studied for dermic way as 200 μg/L (Mink et al., 2008;Tsuji et al., 2014) for non-scratched skin but Sargassum spp. lugs harsh the skin: impacts at 200 μg/L are controversial for intact skin and have to be comforted for scratched one. Progressive accumulation of Sargassum, sank or accumulated as pseudo-vase, is a threat for the coastal environment because of this As input at each Sargassum event.
As speciation has been detailed for arsenate, arsenite, methylarsonic acid (MMA), dimethylarsinic acid (DMA), arsenobetaine, and arsenocholine. All the organic forms are based on pentavalent As. As(V) prevails in oxic conditions (Michel, 1985;Muse et al., 1989;Nekk, 2002;Pell et al., 2013): 41.8 ± 25.2 mg/kg dw is observed for whole samples and, in details, 55.7 ± 18 μg/kg dw in Guadeloupe and 33.0 ± 26.04 μg/kg dw in Martinique. However, average concentration observed in Martinique is 39.7 ± 25 μg/ kg dw without piled Sargassum. Minimum concentration is 0.025 μg/kg dw (0.8 μg/kg dw without piled Sargassum) and maximum 106.5 μg/kg dw. Considering the other species of As, they are formed under anoxic conditions (Brown 2010) by reduction of As(V) into As(III), inorganic both but in a reversible process, and microbial pathways to form arseno-organic molecules, which lead to a non-reversible speciation (Muse et al., 1989). Comparing Martinique and Guadeloupe results, the presence and abundance of such non-As(V) species are the more segregating: non-As(V) are, in Martinique, more present than in Guadeloupe: As(III) concentration was about 3.8 ± 5.2 μg/kg dw for whole samples: 2.4 ± 2.2 μg/kg dw in Guadeloupe and 4.9 ± 6.4 μg/ kg dw in Martinique (but 4.3 ± 6.3 μg/kg dw without piled Sargassum), minimum 0.03 μg/kg dw and maximum 29.9 μg/kg dw, compared to concentration in MMA (8.5 ± 10.8 μg/kg dw for whole samples: 3.7 ± 4.6 μg/kg dw in Guadeloupe and 12.2 ± 12.4 μg/kg dw in Martinique, but 14.8 ± 12.5 μg/kg dw without piled Sargassum), minimum 0.1 μg/ kg dw and maximum 42.9 μg/kg dw (40.3 μg/kg dw without off-sampled Sargassum), and concentration in DMA (8.5 ± 10.8 μg/kg dw for whole samples: 3.7 ± 4.6 μg/kg dw in Guadeloupe and 12.2 ± 12.4 μg/kg dw in Martinique, but 14.8 ± 12.5 μg/kg dw without piled Sargassum), minimum 0.01 μg/kg dw and maximum 28.1 μg/kg dw. In other words, non-As(V) species concentration in Martinique is at least twofold of the corresponding concentration in Guadeloupe, and this ratio often reaches fourfold for the organoarsenic species, even if they are based on As(V). The concentrations of arsenite, MMA, and DMA are in the same order, and arsenobetaine and arsenocholine in an order less, but with the same higher concentrations in Martinique samples (respectively, 1 ± 1 μg/kg dw in Guadeloupe and 2.9 ± 3.5 μg/kg dw in Martinique and 0.5 ± 0.4 μg/kg dw in Guadeloupe and 1.3 ± 2.2 μg/kg dw in Martinique). Those differences are not significant due to wide standard deviation associated with the spatial heterogeneity and suggest that the fate of As speciation is related to sites. The wide standard deviation is related to spatial heterogeneity, about which the first campaign focused on, but this result is not satisfying and led the authors to sample depending on location on the beach: in water, grounded but still wet, and grounded and desiccated (Fig. 2). However, if the ratios between As(V) and the other As forms were island-dependent, the considered ratios from the same island are comparable, whatever the location on the shore, i.e., all the As forms are transuding with the same intensity whatever the island. Such interisland difference might be explained by the presence of the genes driving the organoarsenial metabolization (Héry et al., 2008) but a dedicated study to state is needed. This interisland difference might also be related to geomorphologic patterns; Guadeloupe shoreline being straight but Martinique shoreline being rugged in front of Sargassum inputs, authors could hypothesize that Sargassum tends to be more often under anaerobic condition in Martinique than in Guadeloupe. Further studies about current and water column oxygenation have to test this hypothesis, including the frequency of the genes driving the organoarsenial metabolization.
Considering As fate for drifted Sargassum, authors observed that As content decreases dramatically between floating Sargassum to dried ones (Fig. 2). This phenomenon cannot be explained by volatilization (As and organoarsenial compounds studied having a volatilization temperature upper than 350 °C) nor degradation-not pertinent for elements. If photolysis of organic forms of arsenic is reported (authors only found Howard and Comber, 1992 as reference), this aspect is too weakly studied to decide, and it cannot justify the total As decrease. Figure 2 informs about the temporal process because dried Sargassum is the oldest, but is this arsenic leaching due to rain, due to daily dipping by tide, or both? The dataset does not allow to determine but further studies have to be performed to understand. Nevertheless, Devault et al. (2021) highlighted that Sargassum transudate the main part of their As content in few hours in the water column when stressed, addressing the environmental manager reactivity. Considering the second campaign, during which authors distinguished dried and wet algae, the ratios between As species are not modified whatever the concentration decrease level but As species sums still represented 105.1 ± 10.1% of total As measurement, discarding the hypothesis of an apparent decrease due to formation of no-analyzed As species. The homogeneous decrease for most of the measured As species (except DMA and As(III)) is in favor of leaching in opposition to a species-dependant metabolization. As consequence, As leaching, as leachate on beach sand as transudate in the water column, is an As input that can affect the coastal environment in Caribbean environments where As presence has not been reported as critical in geochemical background.

C, N, and P
We also performed an elemental analysis of the C, N, and P contents (Fig. 3). No significant difference was observed for the C/N ratio of the sample from different drifted statuses (Fig. 3A). It is noteworthy that this C/N ratio (mean 24.85) would make Sargassum potentially compatible with its use as organic amendment (Doleasha et al., 2021). Regarding the N/P ratio, we found a significant increase (P < 0.001) in the dried samples from the beach compared to the wet samples and those collected in the sea (Fig. 3B). Such increase in the N/P ratio indicates a decrease of P, which could be explained by a phosphorous lixiviation from Sargassum. In oligotrophic environments, such input would have to be quantified to determine how this flux can contribute to eutrophication, but bearing in mind the massive scale of Sargassum beaching.

Chlordecone
Sargassum contamination by organic micropollutants has already been reported by Yasmeen et al. (2018) for 48 different molecules in/on S. wightii, Stout et al. (2017) and Torralba et al. (2017) for petroleum hydrocarbons due the Fig. 3 A, B C/N and N/P ratio of the samples considering their drifted status: still floating, fresh algae (In_seawater); recent deposit still humid, algae supple and not purplish (Shore_wet); and dated deposit dry even brittle, algae color turned purplish even grayed (Shore_dry) Deepwater Horizon oil spill, but the contamination by organic micropollutants have not been reported with the same intensity that inorganic ones, particularly As, reported by Devault et al. (2020a). For the present study, results for the chlordecone concentration are presented in Fig. 4A, B and highlight that organic micropollutants can contaminate Sargassum to concerning concentrations.
Chlordecone content is due to terrestrial pollution, leading to coastal contamination. In this way, concentration level of chlordecone could be directly related to chlordeconecontaminated watersheds. In Guadeloupe, chlordecone has been used in the southern part of Basse-Terre, i.e., the place where Sargassum is the most highly contaminated (Figs 4A, B and 5). Pérou river, already studied intensively (Coat et al., 2011;Crabit et al., 2016;Méndez-Fernandez et al., 2018), is known to be a major source of chlordecone to the surrounding shell.
Chlordecone has only been recorded in sites of Basse-Terre of Guadeloupe, where average concentration was 127.0 ± 169.7 μg/kg dw (minimum: 0.8 μg/kg dw, maximum 616.4 μg/kg dw) for the first sampling campaign and 495 ± 828.5 μg/kg dw for the five sites concerned for the second campaign (minimum 15.9 μg/kg dw, maximum 2697 μg/kg dw). In Guadeloupe, chlordecone was not recorded in any of the seven sampling sites in Grande-Terre, while in Basse-Terre, it was found in places where the contaminated watersheds discharge into the sea (Dromard et al., 2019). In the concerned area, fishing and collecting seafood are restricted due to the chlordecone threat. However, chlordecone was observed in the Saintes archipelagos (first campaign), at 58.8 ± 13.2 μg/kg dw, and in Marie-Galante Island, at 7.7 μg/kg dw. These concentrations highlight the risk of chlordecone marginal spread due to mats first landing on contaminated shores and then drifting to chlordecone-free coastal areas. In a  Devault et al. (2021) suggested that chlordecone and other micropollutants can transudate because Marie-Galante Island and the Saintes archipelagos are chlordecone-free.
As for Guadeloupe, campaigns performed in Martinique included sites of restricted area due to chlordecone contamination. Sargassum samples contaminated by chlordecone have an average chlordecone concentration of 137.9 ± 90.9 μg/kg dw for the first campaign and 246.6 ± 271.1 μg/kg dw (minimum 4.4 μg/kg dw, maximum 798.9 μg/kg dw) for the second campaign. In opposition to Guadeloupe results, where all the samples from the same sites where contaminated, samples from Martinique show heterogeneity in the level of contamination among replicates within sites that can range from undetectable to notable. 5b-Hydrochlordecone was only observed in high-CLD concentration samples, i.e., a ratio about 0.06 ± 0.01 with chlordecone concentration in agreement with Devault et al. (2016).
In the Sargassum piled sampled in Martinique, no chlordecone was observed but algae collected were sampled on chlordecone-free shores despite few chlordecone vestigial concentrations. Notwithstanding, other pollutions are reported in the Caribbean, due to crop as house contamination (Devault et al., 2020b), and have to be considered as well.

Consequences for Sargassum issue management
Arsenic contamination of Sargassum mat is inevitable because As is stored by the algae in the open ocean. Thus, the strategy should be to limit Sargassum beaching and accumulation in shallow waters. This could be achieved with the installation of barriers close to shore. Boom cover could be a preliminary action: (1) they are efficient to avoid that sand be picked up with beached algae, sand limiting industrial valorization; (2) Sargassum decay induces leaks and anaerobiosis, worsening the Sargassum effect on biota by oxygen deprivation; (3) As leaks will be more easily diluted into deeper water column than ashore; (4) chlordecone concentration will be limited due to seawater concentration dilution for estuaries to booms; (5) ease the transport from gathering places to stock even valorization plants.
The As content involves that Sargassum have to be picked up as soon as possible, ideally before a day, in order to limit the leak-and before it sank, which occurs in about 3 days, as authors observed. Sargassum collected have to be gathered in dedicated landfill limiting the groundwater pollution and leaching of Sargassum piles have to be treated. As content limits the use of picked up algae: contamination levels exceed animal even nutrition, neither crude Sargassum deposition on ground, as it occurs illegally. Shore piles on sand ground, the actual way of stock, have to be avoided. Large amounts of Sargassum, as observed in Pointe rouge (Martinique) and Porte d'Enfer (Guadeloupe), have to be monitored because population could have to walk the "pseudo-vase" in order to reach fishing device, as observed in the field. Abrasive for the human skin, this "pseudo-vase" can increase the risk of As contamination due to the interstitial water concentration. Chlordecone is adsorbed on Sargassum in polluted bays: polluted and chlordecone-free Sargassum have to be piled separately in order to allow distinct valorization. Drift along the shoreline have to be obstructed in order to limit contaminated mat transfer to chlordecone-free areas, particularly in Guadeloupe where the coastline is less cup-shaped. Even more than for chlordecone-free Sargassum, chlordecone content induces that non-authorized Sargassum valorization has to be tracked.
Leaching from Sargassum occurs in dissolved phase, i.e., under labile and/or colloidal phase, even if the present study does not allow to conclude the prevalence of each. But in the dissolved phase is when micropollutants are bioavailable, as for inorganic, and the authoritative free-ion activity model (Morel, 1983) as for organic ones (Bosma et al., 1997;Leppänen, 2000Leppänen, , 2003Leppänen, , 2006Kraaij et al., 2003;Kukkonen et al., 2004;Landrum et al., 2007;De Weert et al., 2008;Cui et al., 2013). Progressive biota contamination by As is a concern for the whole food web, and a special attention might firstly focus on filtering species, and especially bivalves, which are eaten by local populations (Modestin, 2020).

Recommendations for sampling strategy
During the present study, two sampling strategies were performed: along the shoreline or perpendicularly to the shoreline in order to propose to environmental managers a way to collect beached Sargassum properly. Based on Table 7 and focusing on As (total) value (i.e., not the sum of the As species) because of the better quantification frequencies, metrological repeatability, and lower standard deviation, blind sampling (i.e., whatever the position) leads to 81.6 ± 30.8 μg/kg dw (standard deviation/mean: 37.8%). In details, these are as follows: (1) grounded and at least partly dried and browned, being drifted to shore but not yet grounded (53 ± 30.3 μg/ kg dw; standard deviation/mean 57.2%); (2) recently grounded, i.e., still wet (91.7 ± 29.3 μg/kg dw; standard deviation/mean/mean 32%); (3) close to be grounded, in the beating of the waves (92.2 ± 21.9 μg/kg dw standard deviation/mean 23.7%); and (4) still in seawater (92.1 ± 17.7 μg/kg dw; standard deviation/mean 19.2%).
Considering such results, and regarding that sampling in condition (4) is the less ambiguous, (floating, fresh, and old gold-colored algae tend to reach the highest As concentrations and values are more homogeneous), authors propose to environmental managers to sample fresh algae not yet drifted ashore during their surveys. Considering chlordecone, condition 4 sampled Sargassum reach also the higher values.
Those recommendations are in agreement with the valorization proposed by Devault et al. (2020a, b), Oxenford et al. (2021), Milledge et al. (2020, and Doleasha et al. (2021) who, all, urge that algae cannot be valorized if decayed and if they are polluted by sand.

Conclusions
The present article highlights that Sargassum reaches West Indian shores already contaminated by the As natural content of the open ocean. In coastal waters, it can also become polluted with chlordecone residues due to the 1970s-1990s anthropic uses in French West Indies. Accumulation of chlordecone is rapid: algae reaching their in situ accumulation level in a day. However, the concentrations reached in the range predicted by international literature. Notwithstanding, as for As and for chlordecone, pollutants are able to also leak and pollute the water column, drained out by intense transudation of dissolved matters-an unexpected behavior for chlordecone. Thus, chlordecone pollution into Sargassum is more segregating polluted bays because chlordecone content in Sargassum decreases elsewhere but contaminated Sargassum mats can drift from polluted to chlordeconefree shores and could pollute them secondarily. It is also a severe concern because micropollutants in dissolved phase are bioavailable and could enter the trophic web easily. Sampling protocols to monitor micropollutant concentrations in pelagic Sargassum need to be developed at international scale to improve the management of massive landings and help decide valorization pathways. Appendix 1. Field contamination: Sargassum content (average, n = 3) Table 5  Table 6 Table 7