\makeatletter
\@ifundefined{HCode}
{\documentclass[screen,CRGEOS,Unicode,biblatex,published]{cedram}
\addbibresource{crgeos20260285.bib}
\newenvironment{Table}{\begin{table}}{\end{table}}
\newenvironment{noXML}{}{}
\let\citep\parencite
\let\citet\textcite
\def\xcitealp#1#2{\citeauthor{#1}, \citelink{#1}{#2}}
\newcommand*{\citelink}[2]{\hyperlink{cite.\therefsection @#1}{#2}}
\def\defcitealias#1#2{}
\let\citetalias\textcite
\let\citepalias\parencite
\def\thead{\noalign{\relax}\hline}
\def\endthead{\noalign{\relax}\hline}
\def\tabnote#1{\vskip4pt\parbox{.95\linewidth}{#1}}
\def\botline{\\\hline}
\def\dollar{\$}
\def\tbody{\noalign{\relax}}
\def\tsup#1{\textsuperscript{#1}}
\def\tsub#1{\textsubscript{#1}}
\def\og{\guillemotleft}
\def\fg{\guillemotright}
\def\0{\phantom{0}}
\let\permil\textperthousand
\usepackage[figuresright]{rotating}
\RequirePackage{etoolbox}
\usepackage{upgreek}
\usepackage{wasysym}
\usepackage{pifont}
\def\jobid{crgeos20260285}
%\graphicspath{{/tmp/\jobid_figs/web/}}
\graphicspath{{./figures/}}
\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}}
\let\MakeYrStrItalic\relax
\def\refinput#1{}
\def\back#1{}
\skip\footins10pt
\csdef{Seqnsplit}{\\}
\newcommand*{\hyperlinkcite}[1]{\hyper@link{cite}{cite.#1}}
\newenvironment{inftab}{}{}
\DOI{10.5802/crgeos.341}
\datereceived{2026-04-28}
\daterevised{2026-06-09}
\dateaccepted{2026-06-09}
\ItHasTeXPublished
}
{
\PassOptionsToPackage{authoryear}{natbib}
\documentclass[crgeos]{article}
\usepackage[T1]{fontenc} 
\usepackage{pifont}
\def\CDRdoi{10.5802/crgeos.341}
\newenvironment{sidewaystable*}{\begin{table*}}{\end{table*}}
\def\href#1#2{\url[#1]{#2}}
 \def\citelink#1#2{\citeyear{#1}}
 \def\xcitealp#1#2{\citealp{#1}}
\makeatletter
\def\CDRsupplementaryTwotypes#1#2{}
 \def\selectlanguage#1{} 
 \def\hyperlinkcite#1#2{\citeyear{#1}}
\let\newline\break
\def\CDRsupplementaryTwotypes[#1]#2#3{}
\newcommand\@coi{}
\newcommand\COI[1]{\gdef\@coi{#1}}
\newcommand\printCOI{\ifx\@coi\@empty\else%
\section*{Declaration of interests}
\@coi\fi
}
}
\makeatother

\COI{The authors do not work for, advise, own shares in, or receive
funds from any organization that could benefit from this article, and
have declared no affiliations other than their research organizations.}
\newcommand{\noopsort}[1]{}
\dateposted{2026-07-22}
\begin{document}

\begin{noXML}

\CDRsetmeta{articletype}{research-article}

\TopicFR{Oc\'eanographie, biog\'eosciences oc\'eaniques}
\TopicEN{Oceanography, oceanic biogeosciences}

\title{Deep-sea benthic foraminifera in the Cassidaigne Canyon (NW
Mediterranean): assessing the ecological recovery six years after the
cessation of red mud dumping at a bauxite industrial waste site}

\alttitle{Les foraminif\`{e}res benthiques profonds du canyon de
Cassidaigne (nord-ouest de la M\'{e}diterran\'{e}e) : \'{e}valuation de
la r\'{e}silience \'{e}cologique six ans apr\`{e}s l'arr\^{e}t des
rejets de boues rouges sur un site de d\'{e}chets industriels de
bauxite}

\author{\firstname{Christophe} \lastname{Fontanier}\CDRorcid{0000-0003-4849-6634}\IsCorresp}
\address{FORAM, Study Group, 9 rue des Fauvettes, F-49125 Tierc\'{e}, France}
\address{Universit\'{e} de Bordeaux, UMR CNRS 5805 EPOC -- OASU,
All\'{e}e Geoffroy Saint-Hilaire, CS 50023, F-33615 Pessac, France}
\address{Universit\'{e} d'Angers, 4 boulevard Lavoisier, F-49000 Angers, France}
\email[C. Fontanier]{c.fontanier@foram.eu.com} 

\author{\firstname{Briony} \lastname{Mamo}}
\address{Department of Biology, Macquarie University, North Ryde, NSW, 2109, Australia}

\author{\firstname{D\'eborah} \lastname{Mille}}
\address{CREOCEAN, Agence PACA Corse, 230 Avenue de Rome,
Valparc-B\^{a}t B, F-83500 La Seyne sur Mer, France}
\address{SINAY, rue Alfred Kastler F-14000 Caen, France}

\author{\firstname{S\'ebastien} \lastname{Thorin}}
\addressSameAs{5}{CREOCEAN, Agence PACA Corse, 230 Avenue de Rome,
Valparc-B\^{a}t B, F-83500 La Seyne sur Mer, France}

\keywords{\kwd{Benthic foraminifera}\kwd{Bauxite
residues}\kwd{Cassidaigne Canyon}\kwd{Opportunistic
species}\kwd{Ongoing recovery}}

\altkeywords{\kwd{Foraminif\`{e}res benthiques}\kwd{R\'{e}sidus de
bauxite}\kwd{Canyon de Cassidaigne}\kwd{Esp\`{e}ces
opportunistes}\kwd{R\'{e}silience en cours}}

\shortrunauthors

\begin{abstract}      
During an environmental survey performed in winter and spring 2022,
living (Rose Bengal stained) benthic foraminiferal faunas were
investigated at 13 stations sampled within the \mbox{Cassidaigne} Canyon (NW
Mediterranean Sea) and surrounding area. These stations are located
between \mbox{265--2300 m} water depth. For many decades, industrial bauxite
residues of red mud have been dumped into the canyon via a submarine
pipe, causing physical disturbance and chemical contamination. In
January 2016, solid waste underwater dispersal ceased and was replaced
with the dumping of a low-density liquid effluent. Six years after the
cessation of red mud dispersal, our observations at the 725 m-depth
station closest to the Cassidaigne Canyon submarine outlet show a
better ecological quality compared to the 2012 (during the red mud
dumping) and 2016 (ten months after the cessation of dumping) sampling,
suggesting a putative biotic recovery at the seafloor. That being said,
this station still presents the highest abundance of opportunistic
species, and a noticeably altered benthic diversity. At the other
twelve stations, foraminiferal standing stocks and simple diversity
decrease with decreasing food input to the seafloor and increasing
water depth. There foraminiferal composition, with a minor contribution
of opportunistic and stress-tolerant species, echoes (1) the overall
meso-oligotrophic patterns of a relatively stable ecosystem, and (2)
the putative trophic effect of phytodetritus exportation in spring
2022.
\end{abstract}

\begin{altabstract} 
Au cours d'un suivi environnemental men\'{e} durant l'hiver et le
printemps 2022, les faunes de foraminif\`{e}res benthiques vivants
(color\'{e}s au rose Bengale) ont \'{e}t\'{e} \'{e}tudi\'{e}es sur 13
stations \'{e}chantillonn\'{e}es dans le canyon de Cassidaigne
(nord-ouest de la mer M\'{e}diterran\'{e}e) et ses environs. Ces
stations sont situ\'{e}es \`{a} des profondeurs comprises entre 265 et
2 300 m\`{e}tres. Depuis plusieurs d\'{e}cennies, des r\'{e}sidus
industriels de bauxite sous forme de boue rouge sont d\'{e}vers\'{e}s
dans le canyon via un tuyau sous-marin, provoquant des perturbations
physiques et une contamination chimique. En janvier 2016, le rejet
sous-marin de d\'{e}chets solides a cess\'{e} et a \'{e}t\'{e}
remplac\'{e} par le d\'{e}versement d'un effluent liquide de faible
densit\'{e}. Six ans apr\`{e}s l'arr\^{e}t du rejet de boues rouges,
nos observations \`{a} la station situ\'{e}e \`{a} 725 m de profondeur,
la plus proche de la sortie sous-marine du canyon de Cassidaigne,
montrent une meilleure qualit\'{e} \'{e}cologique par rapport aux
\'{e}chantillonnages de 2012 (pendant le rejet de boues rouges) et de
2016 (dix mois apr\`{e}s l'arr\^{e}t du rejet), sugg\'{e}rant une
possible r\'{e}silience biotique du fond marin. Cela dit, cette station
pr\'{e}sente toujours la plus forte abondance d'esp\`{e}ces
opportunistes et une diversit\'{e} benthique sensiblement
alt\'{e}r\'{e}e. Aux douze autres stations, les stocks permanents de
foraminif\`{e}res et la diversit\'{e} simple diminuent avec la
diminution de l'apport organique vers les fonds marins et
l'augmentation de la profondeur de l'eau. La composition
foraminif\`{e}re, avec une contribution mineure d'esp\`{e}ces
opportunistes et tol\'{e}rantes au stress, refl\`{e}te (1) les
caract\'{e}ristiques m\'{e}so-oligotrophes g\'{e}n\'{e}rales
d'\'{e}cosyst\`{e}mes relativement stables, et (2) l'effet trophique
pr\'{e}sum\'{e} de l'exportation de phytod\'{e}tritus au printemps
2022.
\end{altabstract} 

%\input{CR-pagedemetas}

\maketitle

{\pagebreak}

\twocolumngrid

\end{noXML}

\defcitealias{Fontanieretal2015}{ibid.}
\defcitealias{Fontanieretal2020}{ibid.}
\defcitealias{Fontanieretal2008a}{ibid.}

\section{Introduction}\label{sec1}

In natural marine settings, both temporal and spatial dynamics of
deep-sea foraminifera (Eukaryota, Rhizaria) are controlled by three
major physico-chemical parameters \citep[see the review
by][]{Gooday2003, Zeppillietal2015}. Organic-matter flux toward the
sea floor is the most important ecological constraint as far as it
defines the benthic ecosystem's trophic level, and the related food
availability for foraminiferal communities \citep{Gooday2003}. When
high, this food supply supports high-density and high-diversity
foraminiferal fauna. When too excessive, the organic-matter flux is
also considered an ecological limiting factor by inducing either
temporary or long-term hypoxia either in the sediment or bottom water
\citep{Fontanieretal2014, Goodayetal2000, Kurbjeweitetal2000,
Schumacheretal2007}. Below a certain oxygen concentration threshold
(i.e., 45~$\upmu$M/l), the oxygenation level of the bottom water then
becomes the second ecological constraint, limiting the diversity of
benthic foraminiferal fauna \citep{Fontanieretal2014, Goodayetal2000,
Kurbjeweitetal2000, Schumacheretal2007}. Often neglected or
underestimated in ecological studies, hydro-sedimentary processes
constitute the third set of environmental constraints in deep-sea
environments. Sediment gravity flows running down mature submarine
canyons can supply organic detritus and terrigenous particles to the
deep ocean. Foraminiferal faunas living in these naturally disturbed
habitats are characterized either by various stages of
low-diversity-fauna colonization occurring after physical disturbance
(e.g.\ turbidity flows), or by equilibrium phases (with a net increase
in faunal diversity) related to the gradual burial of organic matter
(e.g.\ eutrophication) \citep{Durosetal2013, Durosetal2011,
Fontanieretal2008a, Fontanieretal2008b, HessJorissen2009, Hessetal2005,
Kohoetal2008, Kohoetal2007}. Because of the abovementioned,
foraminifera are a remarkable group of organisms for studying rapid
disturbances in benthic environments, whether natural or anthropogenic.
These unicellular organisms, characterized by their relatively short
life cycle and their metabolic plasticity, make ideal candidates for
the monitoring of environmental stress related to human activities
\citep{Schonfeldetal2012, Zeppillietal2015}.

Between 1967 and 2015, bauxite residues (namely red mud) were dispersed
into the Cassidaigne Canyon (NW Mediterranean) by the Gardanne alumina
refinery \citep[see review by][]{Dauvin2010}. Bauxite red mud (a
combination of liquid effluent and residual solid) was drained away by
a submarine pipe and discharged at a water depth of 320~m, about 8~km
offshore. This sedimentary material passed down along the axis of the
Cassidaigne Canyon and its lateral flanks to great depths (${>}$2000~m)
with a total coverage estimated to be more than 900~km\tsup{2}
\citep{Dauvin2010, Fabrietal2013, Fontanieretal2020, Fontanieretal2015,
Fontanieretal2012}. In January 2016, bauxite residue dispersal ceased
(see
\href{https://alteo-environnement-gardanne.fr/-Fabrication-et-stockage}{https://alteo-environnement-gardanne.fr/} 
for further information). Since then, only residual liquid effluent has
been released from the pipeline outlet into the Cassidaigne Canyon.
This liquid, characterized by a density lower than the ambient sea
water, is gradually diluted as it rises up through the water column.

Ecological studies have been carried out for the last five decades in
order to assess the impact of red mud on deep-sea metazoan 
\citep{Bourcier1969, Bourcieretal1993, BourcierZibrowius1973,
Fabrietal2013, VitielloVivier1974, Vivier1978a, Vivier1978b}.  Close to
the pipe outlet, the hydro-sedimentary pollution related to the
flooding of red mud (i.e., high sedimentation rate and potential
submarine erosion) precluded benthic meiofauna and macrofauna
settlement along the canyon axis. In the surrounding areas, normal
hemipelagic deposit- and suspension-feeding benthic macrofauna could
thrive, despite the presence of a thin layer of red mud.
\citet{Fontanieretal2012} investigated foraminiferal faunas from two
stations located at 725~m and 1528~m along the axis of the Cassidaigne
Canyon (ESSROV cruise, October 2011). Both studied sites were highly
contaminated by iron, titanium, vanadium and chromium compared to
normal hemipelagic sedimentary environments. At the shallower station
located close to the pipe outlet, the living (Rose Bengal stained)
benthic foraminiferal community was characterized by very low diversity
(i.e.\ only three species; \textit{Gyroidina umbonata} (Silvestri,
1898),  \textit{Bulimina marginata} d'Orbigny, 1826 and
\textit{Bulimina costata} d'Orbigny, 1852). The physical disturbance
related to red mud deposition/remobilisation was likely the major
hydro-sedimentary parameter precluding the settlement of diverse fauna.
Conversely, bauxite residues had no environmental impact on
foraminiferal faunas living at the deeper site. In September 2012 (one
year later), fourteen stations located between 288--2432 m water depth
at varying proximity to the pipe outlet were sampled
\citep{Fontanieretal2015}. Due to more extensive coring,
\citetalias{Fontanieretal2015} evaluated the impact of red mud dispersal in
the Cassidaigne Canyon, not along its axis but on its flanks and its
surrounding area (adjacent canyons and the deep basin). Deposits of red
mud were observed in the Cassidaigne and Planier Canyons down to
${\sim}$2000 m (coverage area ${\sim}$900~km\tsup{2}). The diversity,
composition and standing stocks of foraminiferal faunas were
predominantly constrained by overall meso-oligotrophic conditions. The
reduction of sedimentary organic detritus with varying water depth and
the ecological constraint determined by bottom currents generated
gradual changes in foraminiferal communities, regardless of red mud
presence. Compared to the canyon axis studied by
\citet{Fontanieretal2012}, there was no obvious environmental impact of
dispersed bauxite residues on benthic biodiversity in the sampling
period (September 2012). In September and October 2016, four years
after the previous study, and ten months after the cessation of red mud
dumping (January 2016), more extensive core collections than those
performed by \citet{Fontanieretal2012, Fontanieretal2015} were
gathered in the frame of an environmental survey
\citep{Fontanieretal2020}. Foraminiferal communities were sampled at 16
stations located between 265--2500 m with varying proximity to the pipe
outlet. Most of these sites were within the geographical zone where
historical bauxite residues have been previously detected 
\citep{Dauvin2010, Fontanieretal2012, Fontanieretal2015}.  At the 725
m-depth station closest to the Cassidaigne Canyon submarine pipe,
\citet{Fontanieretal2020} recorded the highest abundance of species
considered opportunistic (\textit{B.~marginata} and \textit{Gyroidina
altiformis}, Stewart \& Stewart, 1930) and a strongly altered benthic
diversity. At the other fifteen stations, foraminiferal standing stocks
and simple diversity decreased by decreasing food input to the seafloor
and increasing water depth. There, foraminiferal composition with a
minor contribution of stress-tolerant species echoed the overall
meso-oligotrophic patterns of a relatively stable and unpolluted
ecosystem.

During an environmental survey performed in winter and spring 2022
(this study), living (stained) benthic foraminiferal faunas were
gathered at 13 stations sampled within the Cassidaigne Canyon (NW
Mediterranean Sea) and surrounding area. These stations, which are
located between 265--2300 m water depth, are the same as those sampled
in 2016 and studied by \citetalias{Fontanieretal2020}. The only difference
with the 2016 campaign is that the three stations furthest from the
outlet (${>}$60~km) were not sampled for this study. The main objective
of our study is to determine the ecological patterns of benthic
environments by investigating foraminiferal fauna (diversity indices
and faunal composition) almost six years after the last sampling cruise
and the cessation of solid waste dispersal in the Cassidaigne Canyon
\citepalias{Fontanieretal2020}. It is hypothesized that historically
impacted benthic ecosystems may be resilient, unless the particular
hydro-sedimentary conditions prevailing in the Cassidaigne canyon (i.e.\ 
instability of historical red mud deposits in the canyon head) limit
the ecological recovery of the most vulnerable benthic environments.

\begin{figure*}
\vspace*{-3pt}
\includegraphics{fig01}
\vspace*{-3pt}
\caption{Bathymetry of the study area and location of the
13 stations sampled during the 2022 oceanographic cruises
(winter-spring 2022). ``Outlet'' refers to pipeline outlet that was
used to disperse bauxite residues.} \label{fig1}
\vspace*{-4pt}
\end{figure*}

\section{Study area}\label{sec2}

The Cassidaigne Canyon lies in the eastern Gulf of Lions (NW
Mediterranean) (Figure~\ref{fig1} insert). The 200~m-deep canyon  head
borders the Cassis Bay at only 7~km from the coast and is characterized
by a narrow canyon axis (2~km in width at 1700~m depth downstream)
(Figure~\ref{fig1}) \citep{Fabrietal2017}. The Northern Current (NC),
which forms the northern branch of the cyclonic Liguro-Proven\c{c}al
Current (LPC), follows the continental margin from the Provence coast
(France) to the coast of Catalonia (Spain) \citep{BethouxPrieur1983,
Millot1990}. The NC determines the general surface water circulation
patterns. Below the surface waters (${>}$200~m), spreads the modified
Levantine Intermediate Water (LIW), which is characterized by a
salinity maximum (${\sim}$38.5) and a relative temperature maximum
(${>}$13~\textdegree C). The Western Mediterranean Deep Water (WMDW)
runs below the LIW with a diffusive boundary at 500--800~m
\citep{Bethouxetal2002, BethouxPrieur1983}.  It is characterized by a
homogeneous temperature (${\sim}$13~\textdegree C) and salinity
(38.40--38.45) \citep{Bethouxetal2002, BethouxPrieur1983}. Primary
production in surface waters shows a classic seasonal variation in
temperate latitudes, with a remarkable bloom in the boreal spring
(March to May) \citep{Fraysseetal2013}. Moreover, in late winter,
spring and summer, coastal upwellings are triggered by north-westerly
winds (Mistral) \citep{Fabrietal2017, Brunetal2023}. They generate
enhanced phytoplankton production in the surface water
\citep{Millot1990, Fraysseetal2013}. The transitional period between
autumn and winter is the least productive \citep{Fraysseetal2013}. The
changes in the circulation of the NC over the course of a year, the
variability in the nature of the winds blowing over the Bay of Cassis
and the general morphology of the continental shelf and the Cassidaigne
canyon generate very specific hydro-sedimentary processes of sediment
remobilisation (i.e.\ turbidity currents) in the axis of the canyon down
to its greatest depths \citep{Brunetal2023}.

Our present study is based on sediment cores collected during a
monitoring oceanographic cruise, which took place in winter and spring
2022. Thirteen stations were sampled within and around the Cassidaigne
Canyon (Table~\ref{tab1}; Figure~\ref{fig1}). Eleven of these stations,
starting with ``U'', have already been studied by
\citet{Fontanieretal2015, Fontanieretal2020} and the remaining two
stations, SR1 and SR2, correspond approximately to sampling sites
investigated in \citet{Fontanieretal2012}. Both were studied by
\citet{Fontanieretal2020}. Stations U03 (292~m) and U05 (751~m) are
located at the head and on the eastern flank of the Cassidaigne Canyon.
Stations SR2 (747~m) and SR1 (1553~m) are situated along the
Cassidaigne Canyon axis. Stations U06--U09 are along the Planier Canyon
between ${\sim}$600--2000~m water depth. Both stations U02 and U10 are
located along the deep valley where both the Marseille and Planier
tributary canyons converge (${>}$1800~m). Stations U11 and U12
(${>}$2200~m) are under the influence of the
Marseille/Planier/Cassidaigne Canyon system. Station U13 is located in
the western branch of the Cap-Sici\'{e} Canyon (France), less than 7~km
from the coast and around 25~km south-east of the pipe outlet. 

%tab1
\begin{sidewaystable*}
\caption{\label{tab1}Water depth, coordinates and physiographic
settings of all stations sampled during the 2022 oceanographic cruises
(winter and spring 2022)}
%\tabcolsep=3pt
\fontsize{8.5}{10.5}\selectfont
\begin{tabular}{cccccccc}
\thead
Station & \parbox[t]{3pc}{\centering Sampling date} & 
Latitude & Longitude & \parbox[t]{2.5pc}{\centering Depth (m)} & Settings &
\parbox[t]{7.5pc}{\centering Horizontal distance from the pipeline outlet (km)} & 
\parbox[t]{10.5pc}{\centering Visual detection of red mud deposits}\vspace*{2pt} \\
\endthead
U03 & 11/01/2022 & 43\textdegree 07.05$^{\prime}$N & 05\textdegree
26.11$^{\prime}$E & \0292 & \parbox[t]{10.5pc}{\centering Head of the
Cassidaigne Canyon} & 5.9    & \parbox[t]{10.5pc}{\centering Reddish brown
surface layer (several cm)}\vspace*{2pt} \\

SR2 & 11/01/2022 & 43\textdegree 07.38$^{\prime}$N & 05\textdegree
28.88$^{\prime}$E & \0747 & \parbox[t]{10.5pc}{\centering Axis of the
Cassidaigne Canyon} & 2.4    & \parbox[t]{10.5pc}{\centering Reddish brown
sediment}\vspace*{2pt} \\

U05 & 12/01/2022 & 42\textdegree 59.40$^{\prime}$N & 05\textdegree
31.85$^{\prime}$E & \0751 & \parbox[t]{10.5pc}{\centering Eastern flank of
the Cassidaigne Canyon}\vspace*{2pt} & 17.3 & No \\

SR1 & 12/01/2022 & 43\textdegree 00.13$^{\prime}$N & 05\textdegree
25.49$^{\prime}$E & 1553  & \parbox[t]{10.5pc}{\centering Axis of the
Cassidaigne Canyon} & 16.3   & \parbox[t]{10.5pc}{\centering Reddish brown
surface layer (several cm)}\vspace*{2pt} \\

U13 & 27/04/2022 & 43\textdegree 00.78$^{\prime}$N & 05\textdegree
45.54$^{\prime}$E & \0952 & \parbox[t]{10.5pc}{\centering Western Branch
of the Cap-Sici\'e Canyon}\vspace*{2pt}  & 25 & No \\

U06 & 11/01/2022 & 43\textdegree 02.34$^{\prime}$N & 05\textdegree
21.00$^{\prime}$E & \0605 & \parbox[t]{10.5pc}{\centering Head of the
eastern branch of the Planier Canyon} & 16.6 &
\parbox[t]{10.5pc}{\centering Reddish brown surface layer (several cm)}
\vspace*{2pt}\\

U07 & 11/01/2022 & 43\textdegree 00.09$^{\prime}$N & 05\textdegree 19.21$^{\prime}$E & 1056  & \parbox[t]{10.5pc}{\centering Eastern branch of the Planier Canyon} & 21.2 & \parbox[t]{10.5pc}{\centering Reddish brown surface layer (several cm)} \\

U08 & 12/01/2022 & 42\textdegree 57.43$^{\prime}$N & 05\textdegree
14.04$^{\prime}$E & 1530  & \parbox[t]{10.5pc}{\centering Axis of the
Planier Canyon} & 29.7 & \parbox[t]{10.5pc}{\centering Reddish brown
surface layer (several cm)}\vspace*{2pt} \\

U09 & 13/01/2022 & 42\textdegree 51.53$^{\prime}$N & 05\textdegree
14.58$^{\prime}$E & 1968  & \parbox[t]{10.5pc}{\centering Axis of the
Planier Canyon} & 37.5 & \parbox[t]{10.5pc}{\centering Reddish brown
surface layer (several cm)}\vspace*{2pt} \\

U10 & 13/01/2022 & 42\textdegree 46.22$^{\prime}$N & 05\textdegree
21.95$^{\prime}$E & 1800  & \parbox[t]{10.5pc}{\centering Connection
between both Marseille and Planier Canyons} & 42.3 &
\parbox[t]{10.5pc}{\centering Reddish brown patches at the sediment
surface}\vspace*{2pt} \\

U02 & 27/04/2022 & 42\textdegree 48.83$^{\prime}$N & 05\textdegree
29.58$^{\prime}$E & 2100  & \parbox[t]{10.5pc}{\centering Connection
between both Marseille and Planier Canyons} & 36 &
\parbox[t]{10.5pc}{\centering Reddish brown patches at the sediment
surface}\vspace*{2pt} \\

U11 & 26/04/2022 & 42\textdegree 46.22$^{\prime}$N & 05\textdegree
40.80$^{\prime}$E & 2222  & \parbox[t]{10.5pc}{\centering Connection
between Marseille/Planier and Cassidaigne Canyons} & 43.3 &
\parbox[t]{10.5pc}{\centering Reddish brown surface layer (cm)}
\vspace*{2pt}\\

U12 & 26/04/2024 & 42\textdegree 49.01$^{\prime}$N & 05\textdegree
46.97$^{\prime}$E & 2290  & \parbox[t]{10.5pc}{\centering Connection
between Marseille/Planier and Cassidaigne Canyons} & 42.3 &
\parbox[t]{10.5pc}{\centering Reddish brown patches at the sediment
surface}\vspace*{2pt}

\botline
\end{tabular}
\vspace*{-19pc}
\end{sidewaystable*}

In accordance with previous studies by \citet{Fontanieretal2012,
Fontanieretal2015, Fontanieretal2020}, reddish brown surface sediment
was observed at most stations providing (with other physicochemical
indicators such as the geochemical composition of sediment) qualitative
evidence regarding the geographical and \mbox{historical} dispersal of bauxite
residues \citep[Table~\ref{tab1};][]{CREOCEAN2018}. Surface
sediment \mbox{(0--4~cm interval)}  Titanium (Ti) content, considered a
geochemical proxy for red mud dispersal  \citep{Dauvin2010}, matches
relatively well with visual observations of the sediment--water
interface \citep{Fontanieretal2020}. Extraordinarily high Ti values
were recorded at station SR2  (${\sim}$32\,000 $\upmu$g${\cdot}$g$^{-1}$ DW)
and to a lesser degree, station SR1  (${\sim}$20\,500
$\upmu$g${\cdot}$g$^{-1}$ DW) confirming that bauxite residues
accumulated \mbox{preferentially} along the Cassidaigne Canyon axis
\citep{CREOCEAN2018}. For comparison, the Ti content of the pipeline
dispersed red mud before January 2016 was ${\sim}$70\,000
$\upmu$g${\cdot}$g$^{-1}$ DW \citep{SAFEGE2011}. Nepheloid layers and
sediment gravity flows are considered the main hydro-sedimentary
processes responsible for transferring the bauxite-derived material
from the pipeline outlet along the Cassidaigne Canyon axis
\citep{Dauvin2010, Fabrietal2013, Fontanieretal2015, Fontanieretal2012,
Brunetal2023}. Moderate to high Ti values were recorded at most of the
other stations in adjacent canyons (between 3300 and 4400
$\upmu$g${\cdot}$g$^{-1}$ DW) even at great depths (${\sim}$5100
$\upmu$g${\cdot}$g$^{-1}$ DW at station U10, 1800~m). As suggested by
\citet{Fontanieretal2015}, the region's episodically strong up- and
downwelling currents coupled with efficient sediment transfer by both
gravity and suspension flows could trigger the large spatial coverage
of the natural and Ti-laden seafloor sediments. In contrast, samples
from station U05 (725~m) and U13 (958~m) yielded relatively low Ti
content (respectively ${\sim}$3400 and ${\sim}$3100 $\upmu$g${\cdot}$g$^{-1}$
DW)  \citep{Fontanieretal2015, Fontanieretal2012}. As already discussed
in \citet{Fontanieretal2015}, both stations U05 and U13 are located in
canyon areas not accessible by bauxite residue. Stations U11 (2222~m),
U12 (2290~m) and U02 (2100~m) located at the deeper connections between
the Marseille, Planier and Cassidaigne Canyons, also exhibit low Ti
content (between 2700 and 3100 $\upmu$g${\cdot}$g$^{-1}$ DW).

\section{Material and methods}\label{sec3}

Although our intention was to retrieve samples in boreal autumn (as had
been the case in 2010, 2012 and 2016) \citep{Fontanieretal2012, 
Fontanieretal2015, Fontanieretal2020}, particularly unfavourable 
weather conditions in \mbox{September} and October 2021 led to the
expedition's cancellation and sampling was delayed until the \mbox{boreal}
winter of 2022. Again, due to bad weather, only nine of the thirteen
planned stations were sampled in January 2022. The four remaining
stations were finally investigated in April 2022 (Table~\ref{tab1}).
For this study, all the samples from winter and spring 2022 are grouped
together for comparison with those from previous campaigns (e.g.,
2016). However, we are aware that the 2022 samples incorporate the
spatio-temporal variability inherent in sampling over two seasons.
Therefore, this paper constitutes a snapshot of ecological conditions
prevailing during January and April 2022 in the Cassidaigne Canyon and
surrounding area, almost six years after the last sampling cruise
\citep{Fontanieretal2020}.

\subsection{Sampling}\label{ssec31}

Sediment samples were collected with an USNEL-type box corer (surface
area of 2500~cm\tsup{2}). Two deployments were conducted at all sites
(Figure~\ref{fig1}). In the first box core, two sectors with equal
surfaces (1250~cm\tsup{2}) were defined with a plastic plate. One
sector was subsampled with a Plexiglas tube (internal diameter of 
9.3~cm, surface area of 68~cm\tsup{2}). The uppermost 2~cm of this
sediment core were sliced into half-centimeter intervals and dedicated
to foraminiferal analyses. As explained above, because of
meteorological constraints (strong swell), the box corer could not be
deployed at stations U13 (952~m), U02 (2100~m), U11 (2222~m) and U12
(2290~m) in January 2022. There, sediment cores were collected in April
2022, three months later. To understand overall ecosystem variability,
triplicates are recommended at each sampling site
\citep{Schonfeldetal2012}. Despite this, most ecological papers
studying deep-sea living (stained) foraminiferal communities use only
one core per site. To facilitate effective comparisons between previous
work of this kind, we have also only used one core per site.
Nevertheless, readers should consider our observations and
interpretations with care as they may be biased by potential spatial
(cm- to m- scale) variability that we cannot fully account for with our
datasets.

\subsection{Benthic foraminiferal analysis}\label{ssec32}

Whilst on board, sediment samples dedicated to foraminiferal study were
transferred to 250~cm\tsup{3} bottles filled with 95\% ethanol
containing 2~g${\cdot}$L$^{-1}$ Rose Bengal stain, commonly used to
identify live foraminifera \citep{MurrayBowser2000, Walton1952}. All
samples were gently shaken for several minutes to obtain a homogeneous
mixture. Some weeks after the spring-2022 cruise they were sieved
through a 125~$\upmu$m screen and the sieve residues were stored in
95\% ethanol. Well-stained foraminifera (all chambers excluding the
final stained bright pink) were sorted into wet samples and stored in
Plummer slides. Strict staining criteria were applied and doubtful
individuals without perfectly stained tests were not included. Miliolid
and non-transparent agglutinated taxa were broken for inspection of the
interior of the test. Most live foraminifera were identified at the
species level. All data generated or analysed during this study are
included in this published article (see Supplementary Material).
At each station, the total number of stained individuals found in each
core (surface area of 68~cm\tsup{2}) was normalized to an area of
100~cm\tsup{2} so that the resulting density could be compared with
previous surveys and other studies conducted in other geographic areas.
Diversity indices including simple diversity S (representing the number
of species), Shannon index H$'$ (log~base~e), Rarefied Species Richness
E(S\tsub{35}) and Dominance index D were calculated
\citep{HayekBuzas1997, Murray2006} \citep[using PAST software
by][]{Hammeretal2001}. These faunal descriptors were based on counts of
stained specimens from the four depth horizons analysed in each core.

Although we understand the value of studying either juvenile or
preadult individuals (belonging to the {${<}$}125~$\upmu$m size fraction)
and any opportunistic foraminifera belonging to this size class, this
work was carried out as part of a contractual service for a consulting
firm, and the samples were processed according to a protocol
established in accordance with previous foraminiferal studies conducted
in the study area \citep{Fontanieretal2015, Fontanieretal2020}. As in
the temporal monitoring, only the ${>}$125~$\upmu$m size fraction was
studied; the {${<}$}125~$\upmu$m size fraction residues were not
preserved and cannot be studied.

\section{Results}\label{sec4}
\subsection{Foraminiferal standing stocks and diversity}\label{ssec41}

Foraminiferal standing stocks ranged between ${\sim}$100 (U12, 2290~m)
and ${\sim}$1500 (U03, 292~m) individuals per 100~cm\tsup{2}
(Figure~\ref{fig2}). Values were
lower (${<}$260 individuals${\cdot}$100~cm$^{-2}$) at depths greater than
1500~m compared to shallower stations. Simple diversity (S) varied
between 16 (SR1, 1553~m) and 82 (U03, 265~m) taxa (Figure~\ref{fig3}).
Diversity generally decreased with increasing water depth, with S
values lower than 34 species below 1600~m. The only exception is the
station SR2 (747~m) located along the Cassidaigne Canyon axis, where
only 26 taxa were identified. Shannon index H$'$ and Rarefied Species
Richness E(S\tsub{30}) followed the same trend (Figure~\ref{fig3}) with
higher values recorded at both stations U13 (952~m) and U03 (292~m).
Station SR1 presents low H$'$ (1.9) and E(S\tsub{30}) (9) values
corresponding to the very low simple diversity and relatively high
dominance. Dominance index D and Shannon index values were inversely
related. When the above data are compared with the data from the autumn
2016 campaign, faunas gathered in winter and spring 2022 are generally
denser and more diverse, particularly at the shallower stations.

\begin{figure}
\includegraphics{fig02}
\caption{Standing stocks (No. Ind. 100~cm$^{-2}$) of living
(stained) foraminiferal faunas at the 13 investigated stations.
Stations are arranged by both physiographic setting and increasing
depth. A graphical comparison is proposed between the data from this
study (winter and spring 2022) and the autumn 2016 data 
\citep{Fontanieretal2020}. Asterisks indicate stations sampled in
spring 2022 (compared to others sampled in winter 2022).} \label{fig2}
\end{figure}

\begin{figure}
\includegraphics{fig03}
\caption{Simple diversity S, Shannon diversity index H$'$,
Rarefied Species Richness E(S\tsub{30}) and Dominance index D of living
(stained) foraminiferal faunas at the 13 investigated stations.
Stations are arranged by both physiographic setting and increasing
depth. The data from this study (winter and spring 2022) and the autumn
2016 study \citep{Fontanieretal2020} is illustrated for
effective comparison. The shaded column corresponds to station SR2,
which showed a historical alteration of foraminiferal diversity during
previous sampling periods 
\citep[i.e., autumns 2010 and 2016,][]{Fontanieretal2012,
Fontanieretal2020}. Asterisks indicate
stations sampled in spring 2022 (compared to others sampled in winter
2022).} \label{fig3}
\end{figure}

\subsection{Faunal composition}\label{ssec42}

At the head of the Cassidaigne Canyon (station U03, 292~m),
\textit{Hoeglundina elegans} (d'Orbigny, 1826) (13\%),
\textit{Uvigerina elongatastriata} (Colom, 1952) (11\%) and
\textit{Melonis barleeanus} (Williamson, 1858) (9.5\%) dominated a
relatively well-diversified living fauna (Figure~\ref{fig4}). Along the
axis of the Cassidaigne Canyon (station SR2, 747~m), \textit{Gyroidina
orbicularis} d'Orbigny, 1826 (29\%), \textit{Bulimina marginata}
d'Orbigny, 1826 (16\%) and \textit{Gyroidina altiformis} Stewart \&
Stewart, 1930 (11\%) were dominant. At the same depth on the eastern
flank of the Cassidaigne Canyon (station U05, 751~m), \textit{Uvigerina
mediterranea} Hofker, 1932 (31\%) dominated the living fauna.
\textit{Melonis barleeanus} (Williamson, 1858) (13\%) and
\textit{Uvigerina peregrina} Cushman, 1923 (7\%) were secondary taxa.
At station SR1 (1553~m) located along the Cassidaigne Canyon axis,
foraminiferal fauna was dominated by \textit{M.\ barleeanus}
(Williamson, 1858) (45\%), \textit{U.\ mediterranea} Hofker, 1932 (17\%)
and \textit{U.\ peregrina} Cushman, 1923 (7\%). In the Sici\'{e} Canyon,
at station U13 (952~m) sampled in spring 2022, \textit{M.\ barleeanus}
and \textit{U.\ mediterranean} (${\sim}$11\%) were the most abundant
species. Along the upper part of the Planier Canyon axis (stations U06,
U07 and U08, ${<}$1530~m water depth), \textit{U.\ mediterranea} and
\textit{M.\ barleeanus} were dominant. The relative contribution of
\textit{U.\ mediterranea} ranged between 14\% and 24\%, whereas the
relative abundance of \textit{M.\ barleeanus} was ${\sim}$20\%. At station
U06 (605~m), \textit{Bigenerina nodosaria} d'Orbigny, 1826 was a
substantial faunal component (${\sim}$10\%). \textit{Uzbekistania
charoides} (Jones \& Parker, 1860) (13\%) and \textit{Nodellum
membranaceum} (Brady, 1879) (12\%) were secondary taxa at station U08
(1530~m). In the deepest part of the Planier Canyon (U09, 1968~m),
\textit{M.\ barleeanus} was still an important species (14\%). But
\textit{N.\ membranaceum} (22\%) was the most abundant taxon. At station
U09 (1800~m), \textit{Thurammina albicans} Brady, 1879 (25\%) dominated
the living fauna. \textit{Melonis barleeanus} was abundant (20\%)
whereas \textit{N.~membranaceum} (13\%) was a substantial faunal
component. Deeper than ${\sim}$2000~m, \textit{N.~membranaceum} was the
dominant species with percentages between 20\% (U11, 2222~m) and 36\%
(U12, 2290~m) (Figure~\ref{fig4}). \textit{Melonis barleeanus} was also
abundant in living fauna with a relative contribution of 13\% at
station U12 and 30\% at station U11. \textit{Lagenammina calcarea}
(Cushman, 1947) was a substantial taxon at both station U02 (2100~m)
(19\%) and station U11 (2222~m)~(8\%).\looseness=1

\begin{figure*}[p!]
\includegraphics{fig04}
\caption{Composition of benthic live (stained)
foraminiferal faunas at the 13 investigated stations. Only major
species (at least ${>}$5\% at one site) are illustrated. Stations are
arranged by both physiographic setting and increasing depth. Asterisks
indicate stations sampled in spring 2022 (compared to others sampled in
winter 2022).} \label{fig4}
\end{figure*}

\section{Discussion}\label{sec5}

There are multiple studies on living foraminifera from the Gulf of
Lions and the Ligurian Sea prior to our present study
\citep{BizonBizon1984, ContrerasRosalesetal2012, DeRijketal2000, 
Fontanieretal2015, Fontanieretal2008a, Fontanieretal2008b, 
Goineauetal2012, Goineauetal2011, Schmiedletal2000}. They provide
reliable information concerning what we might expect in terms of
natural foraminiferal abundance and distribution in the region.
Furthermore, a foraminiferal response to red mud pollution in the axis
of the Cassidaigne Canyon has already been documented by
\citet{Fontanieretal2012, Fontanieretal2020}. Both works and other
recent papers regarding foraminiferal recolonization in canyon settings
\citep{Durosetal2013, Durosetal2011, HessJorissen2009, Hessetal2005}
provide a reliable basis on which to assess the potential impact of red
mud dispersal on foraminiferal biodiversity and its potential
resilience since January 2016, when red mud dispersal\break ceased.

\subsection{Trophic control on foraminiferal faunas in the Cassidaigne
Canyon and surrounding area}\label{ssec51}

Low-diversity (S ${<}$ 34 taxon; $\mathrm{H}'< 2.9$) and low-density (${<}$260 
individuals${\cdot}$100~cm$^{-2}$) foraminiferal faunas are observed on
the distal lower slope (${>}$1500~m) compared to more diverse and
densely populated communities documented from almost all shallower
stations (except station SR2) (Figures~\ref{fig2} and~\ref{fig3}). This
faunal distribution pattern was already documented by
\citet{Fontanieretal2015,  Fontanieretal2020} based on samples
collected in September 2012 and in  September--October 2016. It is
likely related to the natural scarcity of food (i.e.\ sedimentary
organic matter) with water depth. This deep basin food impoverishment
echoes (1) the natural decrease in exported primary productivity (i.e.\ 
fresh phytodetritus) with increasing water depth and (2) the naturally
diminishing lateral advection of degraded organic compounds from
neritic areas to deeper stations \citep{Fontanieretal2015,
Fontanieretal2020}.  Compared to autumn 2016, Station U13 (952~m) is
particularly diverse in spring 2022 (78 compared to 42 taxa in 2016)
and presents relatively high foraminiferal standing stocks (850
compared to 330 individuals${\cdot}$100~cm$^{-2}$ in 2016)
(Figures~\ref{fig2} and~\ref{fig3}). This interannual difference is
likely a result of organic matter enhanced flux related to
phytoplankton spring bloom, in April 2022. It is notable that the
export of phytodetritus to the deepest stations (U02, U11 and U12;
${>}$2100~m) that were also sampled in spring 2022 had no remarkable
effect on the diversity and density of foraminiferal faunas compared
with the autumn 2016 samples.

A detailed analysis of the faunal composition of all the stations (with
the exception of station SR2) shows that three major species dominate.
\textit{Melonis barleeanus} is a major taxon (${>}$10\%) at all
stations (between 265--2280~m), particularly at station SR1 (45\%,
1530~m). This species is abundant in mesotrophic and well-oxygenated
environments  \citep{Caralp1989a, Caralp1989b, Durosetal2013,
Durosetal2011, Fontanieretal2002, Fontanieretal2005, Fontanieretal2003,
Fontanieretal2008a, Fontanieretal2008b, Fontanieretal2015, 
Kohoetal2007, Kurbjeweitetal2000, Licarietal2003, Schmiedletal2000}. 
In both open slope and canyon settings, \textit{M.\ barleeanus}  thrives
generally in intermediate infaunal microhabitats, some centimetres
below the sediment--water interface, where it feeds on degraded organic
matter. This species is generally absent in mature canyons where
gravity flows trigger destruction/recolonisation of benthic
foraminiferal habitats \citep{HessJorissen2009, Hessetal2005}.  Our
observations support the assumption that most of the bathyal stations
are characterized by the deposition of low-quality organic compounds,
either transported laterally by along-slope currents (i.e.\ nepheloid
layer) or related to decaying phytodetritus, previously exported to the
seafloor during a bloom. The dominance of \textit{U.~mediterranea} at
most stations located at a depth of less than 1500~m (with the
exception of station SR2) is in agreement with upper slope faunas
described in the western Mediterranean Sea. This species is generally
documented as a shallow infaunal taxon able to feed on relatively fresh
organic phytodetritus in mesotrophic ecosystems
\citep{ContrerasRosalesetal2012, DeRijketal2000, Durosetal2013,
Durosetal2011, EberweinMackensen2006, Fontanieretal2006,
Fontanieretal2003, Fontanieretal2008b, Fontanieretal2002, Kohoetal2008,
Kohoetal2007, Schmiedletal2000}. The notable presence of
\textit{Rosalina bradyi} (Cushman, 1915) at stations U05 (751~m), U13
(952~m) and U06 (U06) is in perfect agreement with previous
observations in autumn 2016 \citep{Fontanieretal2020}. This taxon is
abundant in shelf ecosystems with a preference for an epiphytic and/or
epilithic life habit \citep{Fontanieretal2008a}. Whilst attached to
vegetation, individuals of this species can be transported by bottom
currents into canyons \citepalias{Fontanieretal2008a}. Therefore, the
occurrence of \textit{R.\ bradyi} at our sample sites further underlines
a natural source-to-sink connection in terms of organic supply and
sediment transfer between upper-slope environments and deeper adjacent
shelves. Below 2000~m, \textit{Nodellum membranaceum}, 
\textit{Thurammina albicans} Brady, 1879, and \textit{Lagenammina
calcarea} constitute substantial components of living faunas,
coinciding with autumnal 2012 and autumnal 2016 faunal patterns
\citep{Fontanieretal2015, Fontanieretal2020}. All above-mentioned
species are typical of oligotrophic basins from the western
Mediterranean Sea \citep{BizonBizon1984, DeRijketal2000,
Fontanieretal2015, Fontanieretal2012, Fontanieretal2008b,
Fontanieretal2020}. This suggests that our deeper sample sites
(${>}$2000~m) are not affected by the high input of organic\break compounds.

To summarize, the ecological observations made in winter and spring
2022 show that the benthic fauna at almost all the stations (except
station SR2, see below) are constrained by meso-oligotrophic
conditions. The least diverse fauna, living on little degraded organic
matter, occupy the deep basin, while the densest and most diverse fauna
develop on the upper part of the slope, where the accumulation of fresh
and degraded organic matter is higher.

\subsection{Questionable ecological recovery along the axis of the
Cassidaigne Canyon}\label{ssec52}

In autumn 2012, a 725 m-deep sample site located very close to station
SR2 (747~m) was characterised by a very low-diversity community (S ${=}$\
 3 and $\mathrm{H}'= 0.76$) which was dominated by \textit{Gyroidina umbonata}
(70\%) and \textit{Bulimina marginata} (25\%)
\citep{Fontanieretal2012}. \textit{Bulimina marginata} has been
documented as an opportunistic species living in outer-shelf and
upper-slope environments, at both early and advanced stages of
recolonization in mature canyons \citep[e.g.][]{Fontanieretal2003,
Hessetal2005, Langezaaletal2006, HessJorissen2009, Goineauetal2011}.
Considered an opportunistic and pioneer taxa, \textit{G.\ umbonata} and
\textit{B.\ marginata} were then indicative of intense hydro-sedimentary
pollution due to red mud deposition and remobilisation along the axis
of the Cassidaigne Canyon. In autumn 2016 (ten months after the
cessation of bauxite residues dumping), station SR2 (747~m) was
characterized by a slightly higher diversity (but still low) (S ${=}$\
 13; $\mathrm{H}'= 1.67$) suggesting an ongoing recolonization of contaminated
substrate. \textit{Bulimina marginata} (50\%), \textit{Gyroidina
altiformis} (27\%) and \textit{G.\ umbonata} (3\%) dominated,
contributing 80\% of the living community. Although \textit{G.\ 
altiformis} had been documented as a very low contributor (${<}$2\%) of
bathyal foraminiferal faunas in the Western Mediterranean Sea and the
North-east Atlantic Ocean \citep[e.g.,][]{ContrerasRosalesetal2012,
Durosetal2011, Durosetal2013, Fontanieretal2015, Fontanieretal2002,
Fontanieretal2008a, Fontanieretal2008b},  its strong contribution in
the axis of the Cassidaigne canyon demonstrated its ability to
proliferate in a stressed community recovering from ecosystem upheaval.
In winter 2022 (our present study), station SR2 (747~m) is
characterized by a diversity higher (S ${=}$\ 26; $\mathrm{H}' = 2.43$) than
previous samplings. However, diversity indices remain still lower
compared to station U05 (S ${=}$\ 53; $\mathrm{H}' = 2.82$) located at the same
depth on a flank of the Cassidaigne canyon. The two opportunistic
species \textit{B.\ marginata} and \textit{G.\ altiformis} account for
16\% and 11\% respectively of living fauna, which is dominated by 
\textit{Gyroidina orbicularis} (29\%).  \textit{Gyroidina orbicularis}
has already been described between 500 and 2000 m depth on the open
slopes of the Gulf of Lions and the Bay of Biscay
\citep{Fontanieretal2002, Fontanieretal2008b}. With lower contributions
than in our study area (${\sim}$10\% between 1000 and 1500~m in the Gulf
of Lions, and 10\% at 2000~m in the Bay of Biscay), it is considered an
indicator species for meso-oligotrophic conditions prevailing in the
middle and lower well-oxygenated slopes \citep{Fontanieretal2002,
Fontanieretal2008b}.  \textit{Gyroidina orbicularis} is a further
species of note in that it generally occupies an intermediate infaunal
microhabitat, several centimetres below the water-sediment interface
\citep[e.g.,][]{Fontanieretal2002, Fontanieretal2008b}. Such a living
position suggests an ability to tolerate the stress (i.e.\ hypoxia) of
living in subsurface sediments, but certainly does not suggest
opportunistic behaviour such as that observed in species normally
proliferating disturbed sediments. Furthermore, the faunal association
observed at station SR2 in winter 2022 (association between \textit{G.\ 
orbicularis} and the opportunistic species \textit{B.\ marginata} and
\textit{G.\ altiformis}) as well as the relatively high diversity
indices of the sampled fauna supports the hypothesis of an ecosystem in
biotic recovery, marked episodically by benthic habitat disturbance. In
Figure~\ref{fig5}, we illustrate the proportion of opportunistic
foraminiferal taxa which were documented as potential recolonizers of
freshly disturbed areas (\textit{Psammosphaera} spp.,
\textit{Saccammina} spp.,  \textit{Technitella} spp., \textit{R.\ 
scorpiurus},  \textit{Quinqueloculina seminula} (Linneaus, 1758), 
\textit{G.\ altiformis},  \textit{G.\ umbonata}, \textit{B.\ marginata}) 
\citep{Fontanieretal2012, Fontanieretal2013, HessJorissen2009,
Hessetal2005, HessKuhnt1996, Kaminski1985, Fontanieretal2020}. At
station SR2, opportunistic and pioneer taxa which constituted
${\sim}$80\% of the fauna in autumn 2016, represent ${\sim}$30\% of the
community sampled in 2022 (six years after the cessation of red mud
dumping). Yet at all stations except SR2, opportunistic recolonizers
account for less than 10\% of the living faunas (Figure~\ref{fig5})
where benthic foraminifera thrive in relatively stable ecosystems and
natural trophic conditions control diversity, density and
composition.\looseness=1

\begin{figure*}[t!]
\includegraphics{fig05}
\caption{Relative abundance (\%) of opportunistic and
stress-tolerant foraminiferal taxa that are considered potential
recolonizers of freshly disturbed areas (\textit{Psammosphaera} spp.,
\textit{Saccammina} spp., \textit{Technitella} spp.,
\textit{Quinqueloculina seminula}, \textit{Gyroidina umbonata},
\textit{Gyroidina altiformis}, \textit{Bulimina marginata}). To
facilitate effective comparison, the data from this study (winter and
spring 2022) and the autumn 2016 study
\citep{Fontanieretal2020} are both illustrated. Asterisks
indicate stations sampled in spring 2022 (compared to others sampled in
winter 2022).} \label{fig5}
\end{figure*}

Before drawing any hasty conclusions from our observations concerning
station SR2, whose benthic fauna has historically been impacted by red
mud \citep{Fontanieretal2012, Fontanieretal2020}, it is important to
remember that the samples taken in winter 2022 (our study) do not
correspond to the previous sampling periods  
\citep[autumns 2012 and 2016;][]{Fontanieretal2012,
Fontanieretal2020}. Autumn is generally a period
of very low primary production in surface waters in our study area
\citep{Fraysseetal2013}, whereas upwellings in winter (spring and
summer) can occasionally increase the productivity of surface waters.
With this in mind, it should therefore be considered that the increase
in diversity in the Cassidaigne Canyon axis in winter 2022 could simply
be linked to a higher density of fauna in relation to greater inputs of
organic matter compared with autumnal periods. To better assess the
possible recovery of the Cassidaigne Canyon ecosystems, it would be
important to collect new samples in the autumn to compare with
historical faunas \citep{Fontanieretal2012, Fontanieretal2020}.

\section{Conclusions}\label{sec6}

During an environmental survey performed in winter and spring 2022,
living (stained) benthic foraminiferal faunas were investigated at 13
stations sampled within the Cassidaigne Canyon (NW Mediterranean Sea)
and surrounding area. These stations are located between 265--2300 m
water depth. For many decades, industrial bauxite residues of red mud
have been dumped into the canyon via a submarine pipe, causing physical
disturbance and chemical contamination. In January 2016, underwater
solid waste dispersal ceased and was replaced with the dumping of a
low-density liquid effluent. Six years after the cessation of red mud
dispersal, our observations at the 725-m-depth station closest to the
Cassidaigne Canyon submarine outlet show a better ecological quality
compared to the 2012 (during the red mud dumping) and 2016 (ten months
after the cessation of dumping) samplings, suggesting a putative biotic
recovery at the seafloor. However, this 725-m-depth station still
presents the highest abundance of opportunistic species (e.g.\ 
\textit{Bulimina marginata}), and a noticeably altered benthic
diversity (compared to other stations at a similar depth but not the
canyon axis). At the other twelve stations, foraminiferal standing
stocks and simple diversity decrease by decreasing food input to the
seafloor and increasing water depth. There, foraminiferal composition
with a minor contribution of opportunistic and stress-tolerant species
echoes (1) the overall meso-oligotrophic patterns of a relatively
stable ecosystem and (2) the putative trophic effect of phytodetritus
exportation for samples gathered in spring 2022. 

\section*{Acknowledgements}

We thank the crew members of R/V ``JANUS II'' and all scientific
participants on the 2022 oceanographic cruises. CF (first author of
this paper) performed foraminiferal analyses in the framework of an
industrial contract linking financially the FORAM Research Group and
CREOCEAN (\url{www.creocean.fr}) to ALTEO. ALTEO allowed authors to use
foraminiferal data for this publication. The Titanium dataset generated
during the current study is not publicly available due contractual
constraints linking CREOCEAN to ALTEO ALUMINA but is available from the
corresponding author and the co-authors working in CREOCEAN on a
reasonable request. Finally, we want to thank an anonymous reviewer for
his/her comments on the first manuscript.

\printCOI

\section*{Supplementary materials}

Supporting information for this article is available on the journal's
website under \printDOI\ or from the author.

\CDRsupplementaryTwotypes[application/zip]{supplementary-material}{\cdrattach{crgeos-341-suppl.xlsx}}

\back{}

\printbibliography
\refinput{crgeos20260285-reference.tex}

\end{document}