\makeatletter
\@ifundefined{HCode}
{\documentclass[screen,CRGEOS,Unicode,Thematic]{cedram}
%\usepackage[sort&compress,square,comma,authoryear]{natbib}
\usepackage[square,comma,authoryear]{natbib}
\usepackage{wasysym}
\usepackage[T1]{fontenc}
\newenvironment{Table}{\begin{table}}{\end{table}}
\renewcommand\bibsection{\section*{\refname}}
\newenvironment{noXML}{}{}
\def\tabnote#1{\vskip4pt\parbox{.98\linewidth}{#1}}
\def\sfrac#1#2{{#1}/{#2}}
\def\sofrac#1#2{(#1/#2)}
\def\botline{\\\hline}
\def\thead{\noalign{\relax}\hline}
\def\endthead{\noalign{\relax}\hline}
\def\tbody{\noalign{\relax}\hline}
\def\tsup#1{$^{{#1}}$}
\def\tsub#1{$_{{#1}}$}
\def\ndash{\text{--}}
\def\Lbreak{\newline}
\def\hbreak{{-\break}}
\RequirePackage{etoolbox}
\usepackage{upgreek}
\def\jobid{crgeos20230113}
%\graphicspath{{/tmp/\jobid_figs/web/}}
\graphicspath{{./figures/}}
\def\0{\phantom{0}}
\usepackage{soul}
\newcounter{runlevel}
\usepackage{multirow} 
\def\nrow#1{\@tempcnta #1\relax%
\advance\@tempcnta by 1\relax%
\xdef\lenrow{\the\@tempcnta}}
\def\morerows#1#2{\nrow{#1}\multirow{\lenrow}{*}{#2}}
\def\xmorerows#1#2{#2}
\let\MakeYrStrItalic\relax
\def\refinput#1{}
\def\back#1{}
\csdef{Seqnsplit}{\\}
\DOI{10.5802/crgeos.236}
\datereceived{2023-01-30}
\daterevised{2023-07-12}
\datererevised{2023-09-08}
\dateaccepted{2023-09-11}
\ItHasTeXPublished
\makeatletter
\g@addto@macro{\UrlBreaks}{\UrlOrds}
\gappto{\UrlBreaks}{\UrlOrds}
}
{
\PassOptionsToPackage{authoryear}{natbib}
\documentclass[crgeos]{article}
\usepackage[T1]{fontenc} 
\def\CDRdoi{10.5802/crgeos.236}
\makeatletter
\def\sfrac#1#2{{#1}/{#2}}
\let\hbreak\relax
\def\sofrac#1#2{(#1/#2)}
\let\newline\relax
\def\href#1#2{\url[#1]{#2}}
\def\xmorerows#1#2{\morerows{#1}{#2}}
}
\makeatother

\usepackage{moreverb}

\begin{document}

\begin{noXML}

\makeatletter
\def\TITREspecial{\relax}
\def\cdr@specialtitle@english{Human Environment Observatory}
%\def\cdr@specialtitle@french{Observatoire Homme Milieu (OHM)}
\makeatother

\title{Pesticide upsurge, cross-contamination and biodiversity: case
studies from the Caribbean Coast Human-Environment Observatory}

\author{\firstname{Vincent} \lastname{Herv\'{e}}\CDRorcid{0000-0002-3495-561X}}
\address{Universit\'{e} Paris-Saclay, INRAE, AgroParisTech, UMR
SayFood, 91120 Palaiseau, France}
\email[V. Herv\'{e}]{vincent.herve@inrae.fr}

\author{\firstname{Pierre} \lastname{Sabatier}\CDRorcid{0000-0002-9620-1514}}
\address{EDYTEM, Universit\'{e} Savoie Mont Blanc, CNRS, UMR-5204,
73370 Le Bourget-Du-Lac, France}
\email[P. Sabatier]{pierre.sabatier@univ-smb.fr}

\author{\firstname{Josie} \lastname{Lambourdi\`{e}re}}
\address{UMR BOREA, Mus\'{e}um national d'histoire naturelle,
CNRS-8067, Sorbonne Universit\'{e}, IRD, Universit\'{e} de Caen
Normandie, Universit\'{e} des Antilles, Paris, France}
\email[J. Lambourdi\`{e}re]{josie.lambourdiere@mnhn.fr}

\author{\firstname{Richard} \lastname{Raymond}\CDRorcid{0000-0002-9544-3394}}
\address{UMR \'{E}co-anthropologie, Mus\'{e}um national d'Histoire
naturelle, CNRS-7206, Universit\'e Paris-Cit\'e, Paris, France}
\email[R. Raymond]{richard.raymond@mnhn.fr}

\author{\firstname{Eric} \lastname{Foulquier}\CDRorcid{0000-0002-9453-8885}}
\address{LETG, Universit\'e de Brest, CNRS, UMR 6554, F-29280 Plouzan\'{e}, France}
\email[E. Foulquier]{eric.foulquier@univ-brest.fr}

\author{\firstname{Iwan} \lastname{Le Berre}\CDRorcid{0000-0002-8933-743X}}
\addressSameAs{5}{LETG, Universit\'e de Brest, CNRS, UMR 6554, F-29280 Plouzan\'{e}, France}
\email[I. Berre]{iwan.leberre@univ-brest.fr}

\author{\firstname{Pascal Jean} \lastname{Lopez}\CDRorcid{0000-0002-9914-4252}\IsCorresp}
\addressSameAs{3}{UMR BOREA, Mus\'{e}um national d'histoire naturelle,
CNRS-8067, Sorbonne Universit\'{e}, IRD, Universit\'{e} de Caen
Normandie, Universit\'{e} des Antilles, Paris, France}
\email[P. J. Lopez]{pascal-jean.lopez@mnhn.fr}

\shortrunauthors

\thanks{The authors would like to thank the Labex DRIIHM; the French program
``Investissements d'Avenir'' (ANR-11-LABX-0010), which is managed by
the ANR}

\begin{abstract}
Similar to many small islands within the Wider Caribbean Region, the
French West Indies are rich but fragile island ecosystems threatened by
global warming, pollutants and other anthropic pressures. To understand
integrated and complex human impacts on the environment, the Caribbean
Coast Human-Environment Observatory (OHM) is developing disciplinary
and interdisciplinary studies related to the local-scale driving
factors that affect the coastal evolution in Guadeloupe. We present
here results that show the multiple causalities, causal asymmetries,
and equifinality of environmental pollution related to chlordecone,
glyphosate, arsenic  and \textit{Sargassum} and highlight the need to bring
further knowledge and awareness to the local populations on these
concerns.
\end{abstract}

\keywords{\kwd{Biodiversity}
\kwd{Human-environment observatory}
\kwd{Interdisciplinarity}
\kwd{Coastal transition zone}
\kwd{\textit{Sargassum}}
\kwd{Glyphosate}
\kwd{Chlordecone}}

\maketitle

\clearpage

\twocolumngrid

\end{noXML}

\section{Introduction}\label{sec1}


The small island developing states (SIDS) and member island countries
in the Caribbean Sea have been recognized to share similar sustainable
development challenges, as they often have limited resources and
fragile economies \citep{Robinson2018,Thomasetal2020}. SIDS
also correspond to hotspots of terrestrial and marine biodiversity with
a number of endemic species that need protection
\citep{Maunderetal2008,Miloslavichetal2010,Bellardetal2014}.
Unfortunately, the resilience of many fragile island ecosystems is
threatened by global warming and its\unskip\break collateral effects, including sea
level rise, coral bleaching, tropical cyclones, storm surges, and
droughts, as well as invasive species and other forms of ecological
interactions
\citep{Cohenetal2015,BenjaminThomas2019,Fernandez-Palaciosetal2021,Mycooetal2022}. 
The SIDS, like other places in the world, are also facing numerous
anthropic pressures from overfishing, pollution from the shore,
development operations, mass tourism, urban development, etc.
 \citep{Wolfetal2022}.

Among the nine French outermost regions, Guadeloupe and Martinique
(French West Indies) are exposed to the abovementioned threats,
especially marine submersion (i.e., cyclones and tsunamis), coastal
erosion and sea-level rise \citep{Parisetal2021}. They are also exposed
to several sources of pollution, such as polycyclic aromatic
hydrocarbons (PAHs) in mangroves \citep{Munozetal1997,Ramdineetal2012},
heavy metals in seagrass beds \citep{Bouchonetal2016} and pesticides in
soils and rivers \citep{Crabitetal2016,Comteetal2022}. Dumping of solid
and consumptive wastes by inhabitants and users has also been
documented, particularly in mangrove ecosystems
\citep{DahomeDiRuggiero2017}. Such improper solid waste deposition,
associated with a lack of suitable waste treatment plants and
inadequate pollutant storage areas, is one of the most severe threats
to the sustainability of the natural resources of sensitive habitats
and wildlife and of course to people. Despite these environmental
concerns, these islands and archipelagos of the Caribbean are at the
heart of important tourist flows, notably linked to cruises, because
they offer all the amenities and facilities expected of a tropical
destination. This tourist attraction is not devoid of environmental and
social consequences
\citep{MillerAuyong1991,Wilkinson2006,Scottetal2012,Wong2015,Adams2017,MackaySpencer2017,Wilkinsonetal2017,Wolfetal2022}.
Consequently, the development of ecotourism, as a form of more
sustainable tourism with lesser local impacts, corresponds to great
challenges that will require the development of management and
educational schemes and a better integration of historical contexts
\citep{Wood2000,DehoorneAugier2011,Nicolas-BraganceSaffache2016}.
Probably, such development should not conceal a better regulation of
the tourism flow and a reconsideration of the economic paradigm of
tourism. In fact, even if the French islands present unique potential
and assets, they have to adjust to their environmental context and to
sociohistorical considerations if they want to respond adequately to
both global and local challenges. 

During the year 2016, the French National Center for Scientific
Research (CNRS) had created one Human-Environment Observatories (OHM)
which was named the OHM ``Caribbean Coast'',\footnote{See the OHM
``Littoral Cara\"ibe'' at
\url{https://ohm-littoral-caraibe.in2p3.fr/}.} to investigate the
consequences of coastal landscape and resource transformations
(i.e., artificialization, accelerating urban-port dynamics,
use conflicts\,\ldots) on the Guadeloupean socio-ecosystem. Even if
the OHMs are in line with frameworks that have been developed to
address complex human--environmental processes
\citep{Binderetal2013,Bourgeronetal2018}, they are unique in the sense
that they have defined their own socio-environmental context along with
a disruptive event that is a human-based fact \citep{Chenorkian2020}.
This disruptive event has its own ability to self-disperse into a
plurality of positions, functions and implications. This concept of
disruptive event is also considered as a benchmark or an interstice
into which any OHM can dig through different approaches to study, at
the local scale, socio-ecosystem trajectories (functions, communities,
discourses, actors\,\ldots). In the case of the OHM Caribbean Coast, the
disruptive event corresponds to the modifications of urban-port
dynamics in a perspective of sustainable development of the Guadeloupe
archipelago. Through interdisciplinarity, integrative and multi-scalar
approaches, the OHMs also aim to help stakeholders in environmental and
\mbox{social} action/management planning, and to increase awareness of the
population. As an illustration of the wide range of approaches that are
needed to properly address the complexity and the plurality of
environmental concerns, we present here several studies performed by
the OHM Caribbean Coast on cross-contamination issues of the coastal
environment in Guadeloupe. We show here that the environmental impact
of pesticides that have emerged over time, and which can be linked to
changes in agricultural practices and policies, are also having impacts
on the \textit{Sargassum} crisis.

\section{Phytopharmaceutical products: use, diffusion and
resurgence}\label{2}

\subsection{A brief review of historical contamination by
chlordecone}\label{sec21}

In the French outermost regions, there are important ongoing concerns
related to the ban on the use of the organochlorine molecule
chlordecone. This molecule, which belongs to the same chemical family
as dichlorodiphenyltrichloroethane (DDT), was discovered in 1951 and
produced in the United States in the early 1950s under the trade name
Kepone\tsup{\text{\textregistered}} to combat leaf-eating insects, ants, roaches,
and fly larvae. From 1966 to 1975, the company Allied Chemical, with
contractor Life Sciences Products Company (LSP), produced Kepone at a
small plant in Hopewell, Virginia, along the James River
\citep{Goldfarb1978}. After a case report of employees suffering from
shivering, studies demonstrated negative effects on human health
\citep{Cannonetal1978,Epstein1978,Taylor1982,ReichSpong1983}.
Meanwhile, analyses showed that chlordecone was found throughout the
James River, having spread downstream on suspended sediment and in fishes
\citep{ReichSpong1983}. Chlordecone was shown to have negative effects
on marine unicellular algae \citep{Walshetal1977}, estuarine animals
\citep{SchimmelWilson1977} and fishes \citep{Couchetal1977}. Even
if there were debates on resistance forms, chlordecone and
pentachlorophenol were also shown to inhibit the growth of estuarine
bacteria \citep{OrndorffColwell1980,Mahaffeyetal1982}. By the end of
1975, approximately half a year after the first employee case report, the
production of Kepone was stopped in Hopewell, fishing was forbidden,
and in 1977, all Kepone \mbox{usage} was banned. Allied, LSP, the city of
Hopewell and six individuals were charged by the Federal court and
fined \citep{Goldfarb1978}. This case serves as a model of prosecution
for environmental pollution \citep{BacigalBacigal1987}. Starting from
these dates, several monitoring programs have shown that chlordecone
can still be detected in the majority of white perch and striped bass
samples even thirty years after the source of contamination was removed
\citep{Luellenetal2006}. Even partially, this information clearly
reveals the large spectrum and long-term impacts of chlordecone, which
was recognized as a persistent organic pollutant (POP) by the Stockholm
Convention in 2009 \citep{Madajetal2018}.

However, in the French West Indies, chlordecone (CLD) was used in
intensive cropping of bananas between 1972 and 1993, even though a
number of studies performed in the late seventies warned against the
putative consequences, including the contamination and impacts on human
health \citep{Snegaroff1977,Kermarrec1980}. Currently, we know that a
large portion of the population in Guadeloupe and Martinique is
contaminated with chlordecone. A reference study, named
KANNARI,\footnote{This study is currently being extended to better
evaluate the impregnation levels of the populations
(\url{https://www.chlordecone-infos.fr/sites/default/files/documents/impregnation-chlordecone-kannari.pdf}).}
which was performed from 2013--2014, revealed that a large proportion of
the population, generally over 18 years old, has been contaminated by
CLD, with differences depending on diet and place of residence. Persons
living in contaminated zones consuming seafood or freshwater products
from self-production of roots, tubers, poultry and eggs from
contaminated zones presented a higher percentage above the internal
chronic toxicological reference value (TRV)
\citep{Dereumeauxetal2020}. In fact, CLD has been described to alter
body-weight homeostasis \citep{Costetetal2022} and to be neurotoxic,
an endocrine disruptor, reprotoxic and likely carcinogenic
\citep{Multigneretal2010,Multigneretal2016}.\looseness=-1

Contamination by CLD not only affects a large proportion of the human
population but is also generalized to terrestrial, aquatic and likely
atmospheric ecosystems. Similar to what has been shown in the USA, CLD
has contaminated surface and ground \mbox{water,} and ineluctably through
watershed hydrology, rain runoff and soil erosion
\citep{BonanPrime2001,Mottesetal2016}, the molecule has reached marine
environments \citep{Bertrandetal2010}. However, a limited number of
studies have specifically focused on the transport of CLD, especially
between terrestrial and marine ecosystems
\citep{Woignieretal2018,Dromardetal2022}. Currently, various coastal
marine ecosystems, including mangroves, seagrass meadows and coral
reefs, are known to be contaminated. Even if details on bioaccumulation
processes are scarce, various organisms from biofilms to top predators,
including farmed animals, present moderate to high concentrations of
CLD
\citep{Coatetal2011,Dromardetal2016,Dromardetal2018a,Dromardetal2018b,Mendez-Fernandezetal2018,Montietal2020,Dromardetal2022,Hubasetal2022}.
Chlordecone can be detected in cetaceans living within the AGOA
sanctuary \citep{Mendez-Fernandezetal2018}. Nevertheless, important
debates still exist and further investigations have to address the actual
stability of chlordecone in various terrestrial and aquatic
environments \citep{Saaidietal2023}.

\subsection{Change in the use of pesticides}\label{sec22}

Since the end of the 1990s and the ban of CLD, several alternative
herbicides have been used for weed management. One of the cross-cutting
phytosanitary practices involves the use of glyphosate, the active
ingredient of broad spectrum post emergent, and nonselective systemic
herbicides \citep{Lietal2016}. Glyphosate-based formulations have
become the main herbicides used on a global scale
\citep{Cuhraetal2016}. Studies have revealed that glyphosate from both
urban and agricultural uses occurs in soil, surface water, and
groundwater, and residues such as aminomethylphosphonic acid (AMPA) are
found at all levels of the food chain, such as in drinking water,
plants, animals, and even humans
\citep{Aparicioetal2013,Maggietal2020,Rivas-Garciaetal2022}. However,
the recognition of pharmaceutical substances as being toxic, especially
to the environment and humans, and the associated regulations on usage
and accessibility in markets usually involve long processes that can
vary depending on the state.

To control the use of glyphosate and glyphosate-based herbicides
(GBHs), under the European initiative, the French government launched a
program in 2008 that has since been revised twice. The \mbox{program,} named
the Ecophyto plan, aims to reduce the uses, risks, and impacts of plant
protection products. According to accessible data from the General
Commission for Sustainable Development (GCSD), the glyphosate was over
the 2009--2017 period, the country's top-selling herbicide among the 119
herbicide active substances. For the French overseas territories, the
NGO G\'{e}n\'{e}rations Futures also reported the important use of
glyphosate.\footnote{\url{https://www.generations-futures.fr/cartes/carte-ventes-de-glyphosate-ha-de-sau-departement-2017/}.}
Nevertheless, like other governments in the European Union, the French
government has been committed to eliminate glyphosate use in cases for
which alternatives already exist. \looseness=-1

To our knowledge, only a few investigations on the presence of
glyphosate and derivative products have been performed in the French
West Indies. One report revealed that glyphosate is used year-round in
all cropping systems in Martinique
\citep{DeffontainesMottes2017}. The same authors also showed
intensified detection during the wet season in the upstream part of the
watershed, suggesting a seasonal intensification of treatments to
compensate for increased grass pressure. In this particular study, it
was shown that the metolachlor herbicide or probably its metabolites,
was present in 50\% of the samples, glyphosate was present in more than
20\%, and AMPA---the main degradation product of
glyphosate---was present in more than 90\% of the samples
\citep{DeffontainesMottes2017}. In addition, warnings on the use of the
present pesticide applications and potential pollution of watershed
outlets have been raised, especially in the case of their use in
horticulture \citep{Mottesetal2017}.

\subsection{Impact of agricultural uses within the coastal transition
zone}\label{sec23}

Recently, the OHM Caribbean Coast co-funded two paleoecology projects
on the consequences of agricultural uses of herbicides and pesticides.
The projects examined core sediments in the transition zone from
coastal areas with strict fishing prohibitions from Guadeloupe and
Martinique. Although the two corresponding watersheds studied were
different in terms of total cultivated area, other important
characteristics were similar: soil cover, types of cultures, and
climate. Using core \mbox{sediment}
\mbox{analyses,} a drastic increase in soil
erosion was demonstrated from the turn of the 21$^{\mathrm{st}}$
century (Figure~\ref{fig2}A), which corresponded to the beginning of the
usage of glyphosate in banana fields \citep{Sabatieretal2021}. Further
analyses of the core sample fluxes helped define two periods: one
period from 1950 to early 2000 characterized by a slow sedimentation
rate (SSR) and a period from 2001 to 2017 characterized by a fast
sedimentation rate (FSR). The two periods were also characterized by
differences in fluxes of glyphosate and chlordecone, with virtually no
glyphosate and low to moderate CLD flux found in marine sediments over
the SSR period and increasing amounts of these pesticides found over
the FSR period  (Figure~\ref{fig2}B--C).

These data allowed us to conclude that within the OHM studied areas and
also in Martinique, the application of glyphosate, which is known to
disrupted grass development, had a strong effect on soil erosion
\citep{Mickelsonetal2001}. We showed that the use of glyphosate has
induced an increase in CLD and its degradation products in coastal and
marine areas \citep{Sabatieretal2021}. Such an upsurge of pesticides
trapped in soils some time ago by an actual application of glyphosate
was also demonstrated for DDT in French vineyards
\citep{Sabatieretal2014}. These important results clearly demonstrate
that it is crucial to further study the consequences of herbicide use
regarding the mechanisms of remnant pesticide (i.e., CLD) mobility in
various terrestrial and aquatic environments of the French West Indies
\citep{Mottesetal2021}. Further erosion and ecotoxicological risk
assessments should also be conducted, and management practices should
be adjusted to the local context. We believe that the presented data
reinforce the need to develop nature-based solutions coupled with
sustainable agriculture using fewer pesticides, particularly in insular
territories that present limited agricultural lands that are often in
close vicinity to residential areas \citep{Deffontainesetal2020}.

\subsection{Impact of agricultural uses on biodiversity}\label{sec24}

The memory effect of agricultural uses can be investigated using
sedimentary ancient DNA (sedaDNA), even if the processes that govern
the production, transfer, incorporation, and preservation of \mbox{sedaDNAs}
remain only partially understood. Indeed, DNA preservation depends on
multiple abiotic and biotic processes 
\citep{Dellannoetal2002,Boereetal2011,Tortietal2015,Capoetal2016,Giguet-Covexetal2019,Capoetal2021}.
As a complement to the abovementioned results on the impacts of
glyphosate on the coastal transition zone in Guadeloupe, we explored
sedaDNAs to evaluate the temporal impacts on biodiversity.

The sedaDNAs of the core samples were extracted, and then a DNA
library was prepared by amplifying the V7 region of the 18S rRNA gene,
a universal eukaryotic marker. After sequencing the barcodes and
filtering processes, we obtained 41,081 eukaryotic operational
taxonomic units (OTUs; cutoff 95\%) that were taxonomically assigned
\citep{Herveetal2023}. Regarding the \textit{alpha} diversity (i.e.,
the species diversity in a site at a local scale), we found that both
OTU richness and diversity (the Shannon index and inverse of the
Simpson index) were significantly higher in the FSR samples than in the
SSR samples. Looking at the \textit{beta} diversity (i.e., the extent
of change in community composition or degree of community dissimilarity
between samples), nonmetric multidimensional scaling (NMDS) analyses
revealed that the FSR and SSR eukaryotic communities were different
(Figure~\ref{fig2}D) and that this difference was also significant
\citep{Herveetal2023}. Additionally, we found significant associations
between the ordination axis scores and both the terrigenous and
pesticide fluxes (Figure~\ref{fig2}D).

\begin{figure*}[t!]
\vspace*{-3pt}
\includegraphics{fig01}
\vspace*{-2pt}
\caption{\label{fig1}Maps of Guadeloupe. ({Left})
Location of Guadeloupe in the Lesser Antilles. ({Top right}) 
The Guadeloupe Archipelago. ({Bottom right})
Enlargement of Petit Cul-de-Sac Marin revealing the highly densified
area and location of Guadeloupe Port Cara\"{i}bes. (BD Carto IGN, 2019;
BD Topo IGN, 2019; GEBCO, 2020; TM World Borders, 2021.)}
\vspace*{-3pt}
\end{figure*}

We investigated in more detail variations in specific eukaryotic
clades, with a focus on three eukaryotic groups among the coastal
biomes that presented significant temporal patterns of variation and
played different ecological roles. As representative members of the
organic matter decomposer community, we chose Ascomycota (Opisthokonta,
Fungi). Bacillariophyta (commonly known as diatoms) and
Chloropicophyceae (Archaeplastida) were selected as members of the
carbon-fixing organisms. For these three groups of organisms, we found
no significant variation in their temporal relative abundance or
diversity indices during the SSR period   \citep{Herveetal2023}.
However, during the FSR period, we found a significant decrease in
Ascomycota abundance and a significant increase in relative abundance
for both Bacillariophyta and Chloropicophyceae relative abundance
(Figure~\ref{fig3}). Altogether, our data demonstrate that changes in
land uses alter the rates and patterns of sediment transport associated
with increases in the flux of pesticides, which in turn modify
biodiversity in the coastal transition zone, at least in one location
of the Petit Cul-de-Sac Marin, Guadeloupe, French West Indies
(Figure~\ref{fig1}) is also experiencing important anthropogenic forcing
(i.e., increases in marine traffic, coastal urbanization, {etc}.).

\begin{figure*}[t!]
\vspace*{-1pt}
\includegraphics{fig02}
\vspace*{-2pt}
\caption{\label{fig2}Changes in sedimentological data, pesticide flux
and eukaryotic community composition in a core archive from a marine
transition zone of Guadeloupe. ({A}) Terrigenous flux, ({B}) glyphosate
(GLY) $+$ aminomethyl-phosphonic acid (AMPA, a glyphosate primary
degradation product) flux (in 
$\mathrm{mol}{\cdot}\mathrm{cm}^{-1}{\cdot} \mathrm{yr}^{-1}$), and
({C}) chlordecone (CLD) + chlordecol (CLO, a chlordecone transformation
product) flux (in $\mathrm{mol}{\cdot}\mathrm{cm}^{-1}{\cdot}
\mathrm{yr}^{-1}$). ({D}) Nonmetric multidimensional scaling (NMDS)
ordination of the eukaryotic community composition based on
Bray--Curtis distances. Arrows indicate significant correlations
($p<0.05$) with environmental variables.}
\vspace*{-1pt}
\end{figure*}

\begin{figure*}[t!]
\includegraphics{fig03}
\vspace*{-2pt}
\caption{\label{fig3}Temporal evolution of the relative abundance of
some important eukaryotic clades over the SSR and FSR periods.}
\vspace*{-1pt}
\end{figure*}


For many kinds of aquatic organisms \citep{Tresnakovaetal2021}, it is
known that exposure to glyphosate can alter the diversity and abundance
of fungal \citep{Druilleetal2013,VanBruggenetal2021,Vazquezetal2021} 
or diatom communities
\citep{Debenestetal2010,Veraetal2010,Larrasetal2013,Corralesetal2021}.
Interestingly, it was also shown that glyphosate can differentially
affect the growth of marine diatom species, with some species being
capable of using glyphosate as the sole phosphorous source 
\citep{Wangetal2016}. Moreover, the impacts of glyphosate and its
uptake and degradation processes by both prokaryotic and eukaryotic
photosynthetic organisms are known to depend on physicochemical
conditions and nutrient concentrations
\citep{DrzyzgaLipok2018,Lozanoetal2018,Mesquitaetal2021,Wangetal2022,Zabaloyetal2022}.
Combined with our results, the latest information suggests that
glyphosate and chlordecone can affect the composition and structure of
coastal marine microorganism communities. Since photosynthetic
organisms are directly involved in primary productivity and microbial
decomposers in carbon and nutrient cycling, any alteration in their
composition could potentially have long-term consequences on \mbox{ecosystem}
functioning and stability as a result of changes in both green and
brown food webs \citep{Zouetal2016,Mougi2020}.\looseness=-1

\begin{figure}[t!]
\includegraphics{fig04}
\caption{\label{fig4}Abundance and diversity of
\textit{Sargassum}-related OTUs in sediment samples
from the coastal transition zone in Guadeloupe. 
(A)~Significant difference in the relative abundance of \textit{Sargassum}
between the SSR and FSR periods. (B) Maximum-likelihood tree
showing the relationships between the \textit{Sargassum}-related OTUs
recovered from the Guadeloupe sediment (FSR period in orange and SSR
period in blue) and other publicly available reference sequences of
\textit{Sargassum} species. This phylogenetic tree was inferred with
IQ-TREE version 2.2.0 \citep{Minhetal2020}. Evaluation of branch
supports was computed using SH-aLRT (1000 replicates). \textit{Fucus}
species were used as outgroups.}
\end{figure}

Interestingly, analyses of the diversity present in the core sample
also revealed the presence of OTUs assigned to the brown algae
\textit{Sargassum}. Even if \mbox{precise} taxonomic assignation was rather
difficult for these seaweeds because of the short length of the 18S
rRNA gene amplified here and probably also because this genetic marker
cannot resolve differences at the species level, we found a tendency
toward an increase in their relative abundance over the most recent
years. Over the period from 2012 to 2017, which corresponds to the
appearance of stranding, the relative abundance of \textit{Sargassum}
OTUs tended to increase (Figure~\ref{fig4}A). In fact, this relative
abundance was significantly higher during the FSR than during the SSR
period (Figure~\ref{fig4}A). Additionally, among these OTUs, we detected
7 abundant OTUs that were specific either to the SSR period or the FSR
period (Figure~\ref{fig4}B). Further studies will be necessary to show
whether it is possible to analyze the temporal evolution of
\textit{Sargassum} stranding based on sedaDNAs. However, such analyses
will have to consider the potential contribution of the increase in the
colonization of numerous coral reefs by algae, among which
\textit{Sargassum} can be an important \unskip\break actor
\citep{Thabardetal2011,LofflerHoey2018}. Finally, there is also great
interest in defining an appropriate molecular marker to resolve the
\textit{Sargassum} phylogeny, especially for the three pelagic
\textit{Sargassum} morphospecies, \textit{Sargassum fluitans} III,
\textit{S. natans} I and \textit{S. natans} VIII
\citep{Parr1939,Schelletal2015,Amaral-Zettleretal2017,Martinetal2021},
that are spreading in the Caribbean region.

\section{The golden-brown crisis and emerging concerns}\label{sec3}

Since 2011, the Caribbean and West Africa coastlines have been subject
to massive and episodic influxes of floating \textit{Sargassum} seaweed
\citep{Wangetal2019}. The large amounts of seaweed biomass washed up
along the coastlines have direct and indirect consequences for beaches
\citep{WilliamsFeagin2010,Innocentietal2018,Ruttenetal2021} and for the
functioning and resilience of nearshore ecosystems 
\citep{VanTussenbroeketal2017,Cabanillas-Teranetal2019,Rodriguez-Martinezetal2019}.
The mass stranding of these brown algae (Figure~\ref{fig5}) also has
significant negative impacts on livelihoods, public health, tourism,
fisheries and the coastal economy
\citep{Resiereetal2018,Schuhmannetal2022}.

\begin{figure*}[t!]
{\vspace*{-3pt}}
\includegraphics{fig05}
{\vspace*{-2pt}}
\caption{\label{fig5}Photographs of \textit{Sargassum} along the coasts
or at storage sites in Guadeloupe and Martinique. (A) Baie du Marigot
(Les Saintes, Guadeloupe), (B) \textit{Sargassum} spp., (C) Le Diamant
(Martinique), (D)~Le Marigot (Martinique), (E) Saint F\'{e}lix
(Guadeloupe), (F) Capesterre (Marie-Galante, Guadeloupe), (G)~storage
site of Pointe Faula (Martinique), and (H) storage site Capesterre,
(Marie-Galante, Guadeloupe).}
{\vspace*{-3pt}}
\end{figure*}

\begin{table*}[t!]
\caption{\label{tab1}Metal's composition in \textit{Sargassum} spp.
collected in 2018 from Guadeloupe and Martinique}
\fontsize{8}{10}\selectfont\tabcolsep=2.7pt
\begin{tabular}{cccccccccccc}
\thead
\multicolumn{2}{c}{Localization} & $n^{\mathrm{a}}$ & {Cd} & {Cr} & {Co} & {Cu} & {Hg} & {Ni} & Zn$^{\mathrm{b}}$ & {Pb} & {As} \\
\endthead
\morerows{1}{Coastal}  & Shore zone & \textit{97} & 0.47 $\pm$ 0.19 & 11.6 $\pm$ 23.7 & 1.13 $\pm$ 1.44 & 4.20 $\pm$ 7.31 & $<$LOD & 3.91 $\pm$ 1.11 & 7.66 $\pm$ 7.45 & 0.51 $\pm$ 0.71 & $71.9 \pm 28.4$ \\
 & Offshore & \0\textit{3} & 0.43 $\pm$ 0.05 & 4.04 $\pm$ 0.84 & 0.44 $\pm$ 0.07 & 3.73 $\pm$ 0.31 & $<$LOD & 4.91 $\pm$ 0.16 & 2.20 $\pm$ 0.40 & 0.13 $\pm$ 0.02 & $100.5 \pm 4.9$\0\vspace*{5pt} \\
Inland & Storage sites & \0\textit{9} & 0.28 $\pm$ 0.07 & 11.5 $\pm$ 1.2 & 0.98 $\pm$ 0.12 & 5.24 $\pm$ 1.58 & $<$LOD & 2.54 $\pm$ 1.58 & 6.31 $\pm$ 1.77 & 1.11 $\pm$ 0.43 & $25.6 \pm 15.9$ 
\botline
\end{tabular}
\tabnote{\fontsize{9.3}{11}\selectfont The concentrations correspond to mg of metal per kg of dried
\textit{Sargassum}.  Values are mean ${\pm}$ standard deviations.}
\tabnote{\fontsize{9.3}{11}\selectfont $^{\mathrm{a}}$ Number of samples analyzed.}
\tabnote{\fontsize{9.3}{11}\selectfont $^{\mathrm{b}}$ The results of the coastal shore zone for
the Zinc correspond to 96 and not 97 samples since one of them was
below the limit of detection (LOD).\vspace*{4pt}}
\end{table*}

\begin{table*}[t!]
\caption{\label{tab2}Arsenic concentration in \textit{Sargassum} spp.
collected in 2019 from Guadeloupe and Martinique}
\fontsize{9.3}{11}\selectfont\tabcolsep=4.2pt
\begin{tabular}{cccccccccc}
\thead
\multicolumn{2}{c}{Localization} & $n^{\mathrm{a}}$ & {As total} & {As(III)} & {As(V)} & {MMA} & {DMA} & {AsBet} & {AsC} \\
\endthead
Marine water &  & \textit{59} & 96.1 $\pm$ 24.8 & 1.00 $\pm$ 2.87 & 65.2 $\pm$ 23.6 & 16.0 $\pm$ 12.8 & 5.56 $\pm$ 7.16 & 4.59 $\pm$ 3.63 & 0.47 $\pm$ 0.73\vspace*{5pt} \\
\morerows{1}{Shore}  & Wet & \textit{31} & 96.8 $\pm$ 23.5 & 1.99 $\pm$ 6.29 & 64.7 $\pm$ 23.3 & 13.9 $\pm$ 12.3 & 6.75 $\pm$ 8.39 & 4.13 $\pm$ 3.17 & 0.61 $\pm$ 0.97 \\
 & Dry & \textit{28} & 54.4 $\pm$ 26.8 & 0.39 $\pm$ 0.46 & 34.3 $\pm$ 22.5 & \07.6 $\pm$ 10.2 & 5.72 $\pm$ 5.62 & 2.54 $\pm$ 2.20 & 0.70 $\pm$ 1.10
\botline
\end{tabular}
\tabnote{The concentrations correspond to mg of metal per kg of dried
\textit{Sargassum}. Values are mean ${\pm}$ standard deviations.}
\tabnote{$^{\mathrm{a}}$ Number of samples analyzed.}
\tabnote{The forms of arsenic are: arsenite (As(III)), arsenate
(As(V)), monomethylarsonate (MMA), dimethylarsinate (DMA),
arsenobetaine (AsBet), and arsenochlorine (AsC).}
\end{table*}


Members of the OHM Caribbean Coast associated with other colleagues
from the University of Antilles were involved in an interdisciplinary
project co-funded by the French agency for ecological transition (ADEME
in French) (see the complete report),\footnote{Eco3Sar: Ecologie,
Ecotoxicologie et Economie des Sargasses.
\url{https://librairie.ademe.fr/recherche-et-innovation/4039-eco3sar-ecologie-ecotoxicologie-et-economie-des-sargasses.html}.}
and dedicated to chemical and biological \mbox{contamination} issues related
to \textit{Sargassum} and to degrees of acceptability of their use in
organic amendments. The aims were (i) to determine levels of
contamination by heavy metals and CLD in \textit{Sargassum} seaweeds,
(ii)~describe the eukaryotic and prokaryotic biodiversity associated
with \textit{Sargassum} spp., (iii) test the potential use of these
algae in organic amendments, and (iv) reveal collective
representations of social actors and of the population through
diachronic and prospective approaches. Hereafter, we focus on
contamination and biodiversity-related issues.

\subsection{Heavy metals in \textit{Sargassum}}\label{sec31}

The accumulation of heavy metals by \textit{Sargassum} spp. has created
concerns that the inland storage and disposal of these seaweeds might
also impair high value subsidiaries such as agribusiness \mbox{opportunities}
(i.e.,~fertilizer in crop production or food supplements for livestock
production).\footnote{\url{https://guadeloupe.ademe.fr/sites/default/files/guide-valorisation-sargasses.pdf}.}
Our project was carried out in both Guadeloupe and Martinique with
sampling campaigns during 2018 and 2019, which correspond to periods
with high levels of \textit{Sargassum} stranding \citep{Wangetal2019}.
In 2018, we collected approximately 100 samples corresponding to
\textit{Sargassum} spp. from shore zones in Guadeloupe and Martinique,
offshore sites and storage sites in Martinique (Figure~\ref{fig6}A).
Among the heavy metals and metalloids tested, we found that even if
arsenic is only present at very low concentration in marine waters, it
is bioaccumulated by \textit{Sargassum} with average value of 78 mg/kg
dry wet (Table~\ref{tab1}). In fact, bioaccumulation of arsenic by
pelagic \textit{Sargassum} has also been demonstrated in other
Caribbean regions and territories
\citep{Florezetal2017,Rodriguez-Martinezetal2020,Cipollonietal2022,Devaultetal2022,Ortega-Floresetal2022}
and for various \textit{Sargassum} species \citep{Devaultetal2021}.

\begin{figure*}[t!]
{\vspace*{2pt}}
\includegraphics{fig06}
{\vspace*{2pt}}
\caption{\label{fig6}Maps showing the \textit{Sargassum}
sampling sites and chlordecone concentrations. (A)
Some sampling sites from campaigns 2018 and 2019 overlapped. The
Guadeloupe Archipelago ({left}) and Martinique Island
({right}). The color code corresponds to (yellow) cultivated
lands, marine zones with no restrictions on fishing (light blue),
marine zones with restrictions on fishing (medium blue), and marine
zones where fishing is forbidden (dark blue) (BD Carto IGN, 2019).
(B) Distribution of the chlordecone concentration
($\upmu$g/kg) in \textit{Sargassum} samples according to
fishing area regulations, with the Guadeloupe Archipelago
({left}) and Martinique Island ({right}).}
{\vspace*{2pt}}
\end{figure*}


To obtain further information on the putative effects of landing and
drying processes, we performed analyses of arsenic species (i.e.,
inorganic and methylated forms) in \textit{Sargassum} spp. but after
the separation of seaweeds by location, with seaweeds from the shore,
recently landed and dried (i.e., with a brownish color). We found a
significant decrease in total arsenic concentration upon drying
(Table~\ref{tab2}). The decrease in total arsenic concentration was
also \unskip\break significant for As(V) and MMA (Table~\ref{tab2}). Even if we do
not know the precise mechanisms underlying the decrease in
\textit{Sargassum}-associated arsenic concentration (i.e., direct or
indirect leaching by rainfall or due to seaweed drying, {etc}.),
our results warn us against the potential release of high concentrations of
harmful arsenic species. Leaching and transformation into various
compounds that might depend on complex physical, chemical, and
biological processes can probably occur along the usage chain. These
data reveal the need to further investigate the consequences of arsenic
accumulated by \textit{Sargassum} on coastal and terrestrial
ecosystems, but also on the putative health concerns for populations.
In addition, the presence of various arsenic forms in
\mbox{\textit{Sargassum}}~spp. might also hinder several downstream uses
\citep{Devaultetal2021}.

\subsection{Pesticide contamination}\label{sec32}

Because of the presence of chlordecone in various compartments of the
French West Indies, questions about the accumulation of this pesticide
in \textit{Sargassum} were raised in 2015. The first study performed at
that time in Martinique demonstrated its presence in \textit{Sargassum}
samples.\footnote{\url{https://www.guadeloupe.gouv.fr/content/download/8183/60117/file/1\%20-\%20valorisation\%20-\%20Institut\%20Technique\%20Tropical.pdf}.}

We further investigated contamination levels by CLD in
\textit{Sargassum} collected from coastal areas without or with
restricted uses (Figure~\ref{fig6}A). Indeed, since 2005, several
prefectural decrees have led to the delimitation of coastal areas that
are strictly prohibited from fishing a list of targeted fish,
crustacean, mollusk and echinoderm species and others with partial
protection. In both Guadeloupe and Martinique, CLD was found in samples
from the watersheds of land used for banana cultivation
(Figure~\ref{fig6}B). In Guadeloupe, the highest concentrations were
detected in sites that belong to the municipality Capesterre--Belle--Eau,
where soils and the P\'{e}rou River are highly contaminated with
chlordecone. In Martinique, the highest concentrations were found at
``\textit{Quartier bac}'', which belongs to the municipality La
Trinit\'{e}. Because of observed interannual
variabilities, we propose that the authorities should implement more
systematic analyses of CLD concentrations in \textit{Sargassum}, at
least for a number of local sites.

Contamination of \textit{Sargassum} by other organic pollutants has
been shown in the Caribbean with the presence of hydrocarbon compounds
due to liquid oil spilling during the Deepwater Horizon disaster
\citep{Torralbaetal2017}. Based on our findings, we suggest a more
systematic investigation of the presence of labile, persistent organic
pollutants (POPs) and the so-called contaminants of emerging concern
(i.e., pharmaceutical and personal care products, etc.) in
these seaweeds. Chemical contaminants of edible and nonedible seaweeds
are becoming a general concern
\citep{Maetal2018,Cherryetal2019,Anbazhaganetal2021,Hossainetal2022}
that will also require the adjustment of legislation and development of
relevant guidelines for their handling, storage, use and disposal.

\subsection{\textit{Sargassum}-associated biodiversity}\label{sec33}

Pelagic \textit{Sargassum} constitutes a floating ecosystem hosting
various marine species, such as sea turtles, seabirds, fishes and even
invertebrates
\citep{Parr1939,Weis1968,Fine1970,Dooley1972,Ryland1974,Haney1986,JobeBrooks2009,BallardRakocinski2012,Witheringtonetal2012,JacobucciLeite2014,Susilowatietal2015,Monroy-Velazquezetal2019},
with a few being endemic species. Concerning the less described
microbial (i.e., archaea, bacteria and small eukaryotes) composition,
it is likely that they also contribute to \textit{Sargassum} growth,
nutrient uptake, macroalgae spore release and germination, defense
against pathogens or reproduction
\citep{Eganetal2013,Florezetal2017,VanDerLoosetal2019}.
Exchanges between the \textit{Sargassum} holobiont and the surrounding
environment might have reciprocal impacts and might occur depending on
the environmental conditions; for example, microorganisms initially
involved in algae growth, raft formation and maintenance or algae
sinking might later contribute to algae biodegradation upon beaching
\citep{DeFouwetal2016}. In addition, microbial communities associated
with the seaweeds can proliferate in new habitats, as postulated for
the export of surface-dwelling fauna associated with \textit{Sargassum}
from the surface to the seafloor.

At the molecular level, the investigation of the richness of microbial
communities was mainly performed from surface drifting algae species 
\citep{Torralbaetal2017,Michoteyetal2020}. We investigated for the
first time the communities associated with \textit{Sargassum} at
beaches and storage sites  \citep{Herveetal2021}. For this purpose, we
amplified regions corresponding to the 16S rRNA gene for prokaryotes
and to 18S rRNA for eukaryotes from 100 samples collected in 2018 and
corresponding to seawater surrounding the seaweeds, \textit{Sargassum}
from beaches (e.g., along the shores and landed) or storage sites.
Concerning microbial richness, we identified 22,214 prokaryotic
operational taxonomic units (OTUs) and 17,679 eukaryotic OTUs. Among
them, we found that a large number of both prokaryotic and eukaryotic
OTUs were shared between these three compartments
(Figure~\ref{fig7}A--B). Nevertheless, statistical analyses
demonstrated that these three communities were different 
\citep{Herveetal2021}.

\begin{figure*}[t!]
\includegraphics{fig07}
\caption{\label{fig7}Venn diagrams of
\textit{Sargassum}-associated prokaryotic or eukaryotic OTUs.
(A) Prokaryotes and (B) Eukaryotes. The color code
corresponds to seawater surrounding \textit{Sargassum} (blue),
stranding \textit{Sargassum} on the shore (orange), and
\textit{Sargassum} from storage sites (green).}
\end{figure*}

Interestingly, among the prokaryotes, we found a large number of
bacterial OTUs that are probably involved in sulfur metabolism. From
inland storage sites, we detected methanogenic archaea, which suggests
that a portion of the carbohydrates from the seaweed alga is probably
degraded and used as a potential energy source by these prokaryotes 
\citep{Herveetal2021}. Further specific investigations of this
microbial diversity could probably facilitate the development of biogas
production and energy conversion procedures. We also found that the
\textit{Sargassum}-associated microbiota might contain potentially
harmful and nonnative species that are probably co-transported by
\textit{Sargassum}. If confirmed, such results may have direct
implications for risk assessment and reinforce the need to develop
further \mbox{studies} on \textit{Sargassum} holobionts, including those
focused on biosafety and biosecurity.

\section{Observation for knowledge and awareness}\label{sec4}

An important role of the OHM Caribbean Coast is also to promote and
disseminate research findings to local stakeholders and the population,
including for the teaching sphere and the younger generations. Indeed,
we believe that even if there is probably in Guadeloupe, following the
suggestions of Trevethan \citep{Trevethan2017}, a continuum on
environmental concerns from general awareness knowledge (GAK) to
detailed and specific knowledge (DSK), there might be considerable
personal heterogeneity in awareness within the population that must be
analyzed and documented for both the consequences of pesticide use and
\textit{Sargassum} influxes. In addition to scientists, social networks
and local media play important roles in the construction of GAK. In a
first step to analyze the putative differences between scientific
information and public media, we compared here the number of scientific
publications mentioning the words ``biodiversity'', ``arsenic'',
``glyphosate'', ``\textit{Sargassum}'' or ``chlordecone'' to the French
equivalent from the local media of France Antilles (we pooled both the
Martinique and Guadeloupe editions). Using the SCOPUS database and
searches of article titles, abstracts and keywords, we found an
increasing number of publications (Figures~\ref{fig8}A--C) \mbox{except} for
chlordecone, which presented a more erratic pattern of publication with
peaks in the mid-seventies corresponding to contamination issues at
Hopewell and since 2010 due to issues in the French West Indies
(Figure~\ref{fig8}D). In contrast, in the local media, the interest was
rather different, with a high number of publications focusing on
\textit{Sargassum} (Figure~\ref{fig8}G) and chlordecone
(Figure~\ref{fig8}H) and barely any reports on arsenic or glyphosate
(Figure~\ref{fig8}F). For \textit{Sargassum}, we found two peaks of
publication in 2015 and 2018 corresponding to years with high levels of
stranding, and potentially to the presence of an international
conference on \textit{Sargassum} that occurred in Guadeloupe during the
year 2018. However, it will be interesting to study in more detail the
link between press release on environmental concerns and other social
issues such as political events.

\begin{figure*}[t!]
{\vspace*{-2pt}}
\includegraphics{fig08}
{\vspace*{-2pt}}
\caption{\label{fig8}Environmental concern reports in the scientific
literature and media of France Antilles (Guadeloupe and Martinique).
Graphs correspond to searches of the SCOPUS database for 1970--2021
(A--D) or the France Antilles database for 2006 to 2021 (E--H). For
arsenic, the exceptionally high number of publications from 2008 is
likely related to the Second International Congress titled ``Arsenic
from Nature to Humans'' of 2008, which accounted for 1492 conference
papers published during this year. This survey was performed on October
2022.}
{\vspace*{-1pt}}
\end{figure*}


Even if these analyses are still very partial, they illustrate the
differences between scientific publications, local media and most
likely general public awareness. Such an observation reinforces our
motivation as an observatory to further develop actions to diffuse
knowledge and scientific information on contaminants along the
land-ocean continuum. In the future, we hope to further analyze public
awareness, concerns and priorities by comparing various publication
supports completed through surveys and interviews. We therefore hope to
be able to gain insight into the sensitivity of the local populations
and to draw comparisons to people from other European nations in regard
to, for example, marine environmental concerns \citep{Gelcichetal2014}.

\section{Conclusions}\label{sec5}

The most recent ``\textit{Sargassum} crisis'' that started with the
first stranding of \textit{Sargassum} in 2011 has\unskip\break created new
cross-contamination issues between seaweed, the coastal zone and inland
storage sites by heavy metals (i.e., arsenic) and organic pollutants
(i.e., chlordecone). Analyses of other pollution drivers within the
coastal transition zone have revealed dependencies not only between
past and current agricultural uses but also with soil erosion and the
composition of coastal eukaryotic assemblages. We should also develop
more theoretical approaches to evaluate social concerns, as proposed by
Fitzgerald \citep{Fitzgerald2019}, to properly address the multiple
causalities, causal asymmetries, and most likely equifinality of
environmental pollution concerns. We hope that the OHM Caribbean Coast,
which uses a long-term interdisciplinary framework, will properly
investigate social environmental trajectories, envision integrated
solutions and bring awareness to local populations.

\section*{Declaration of interests} 
The authors do not work for, advise, own shares in, or receive funds
from any organization that could \mbox{benefit} from this article, and have
declared no affiliations other than their research organizations.

\section*{Acknowledgments}
We would like to thank France-Antilles for allowing access to their
publication database. We would like to thank the Labex DRIIHM; the
French program ``Investissements d'Avenir'' (ANR-11-LABX-0010), which
is managed by the ANR. This work was supported  in part by the Agence
de l'environnement et de la ma\^{\i}trise de l'\'energie (ADEME). The
authors thank the anonymous referees for
their helpful comments that improved the quality of the manuscript. We
are also grateful to our colleagues working within the OHM Caribbean
Coast, and to Oscar for his support.

\CDRGrant[ANR]{ANR-11-LABX-0010}

\back{}

\bibliographystyle{crgeos}
\bibliography{crgeos20230113}
\refinput{crgeos20230113-reference.tex}

\end{document}