\makeatletter
\@ifundefined{HCode}
{\documentclass[screen,CRGEOS,Unicode,biblatex,published]{cedram}
\addbibresource{crgeos20250249.bib}
\newenvironment{noXML}{}{}
\let\citep\parencite
\let\citet\textcite
\def\xcitealp#1#2{\citeauthor{#1}, \citelink{#1}{#2}}
\def\citealt#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\tbody{\noalign{\relax}}
\def\tabnote#1{\vskip4pt\parbox{.95\linewidth}{#1}}
\def\tsup#1{$^{{#1}}$}
\def\tsub#1{$_{{#1}}$}
\def\ndash{\text{--}}
\def\hyphen{\text{-}}
\def\og{\guillemotleft}
\def\fg{\guillemotright}
\def\0{\phantom{0}}
\def\bcaption#1#2#3{\caption{#1\unskip\break\\[-5pt]\hspace{\fill}}
\end{figure}\setcounter{figure}{#2}\begin{figure}\vspace*{-.7pc}\caption{(\textbf{cont}.) #3}}
\makeatletter
\g@addto@macro{\UrlBreaks}{\UrlOrds}
\gappto{\UrlBreaks}{\UrlOrds}
\RequirePackage{etoolbox}
\usepackage{upgreek}
\def\inlinefig#1{\includegraphics{#1}}
\def\jobid{crgeos20250249}
%\graphicspath{{/tmp/\jobid_figs/web/}}
\graphicspath{{./figures/}}
\def\xmorerows#1#2{#2}
\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{}
\let\ubreak\break
\csdef{Seqnsplit}{\\}
\def\botline{\\\hline}
\def\back#1{}
\newcommand*{\hyperlinkcite}[1]{\hyper@link{cite}{cite.#1}}
\DOI{10.5802/crgeos.316}
\datereceived{2025-03-19}
\daterevised{2025-10-14}
\dateaccepted{2025-10-15}
\ItHasTeXPublished
}
{
\PassOptionsToPackage{authoryear}{natbib}
\documentclass[crgeos]{article}
\usepackage[T1]{fontenc}
\def\CDRdoi{10.5802/crgeos.316}
\def\citelink#1#2{\citeyear{#1}}
\def\xcitealp#1#2{\citealp{#1}}
\let\ubreak\relax
\def\bcaption#1#2#3{\caption{#1~#3}}
\makeatletter
\def\CDRsupplementaryTwotypes[#1]#2#3{}
\def\hyperlinkcite#1#2{\citeyear{#1}}
}
\makeatother

\usepackage{upgreek}

\dateposted{2025-12-09}
\begin{document}

\begin{noXML}

\CDRsetmeta{articletype}{review}

\TopicFR{Pal\'eoenvironnements, pal\'eoclimats}
\TopicEN{Paleoenvironments, paleoclimates}

\title{Timing and amplitude of the dessication of Sahel at the end of
African Humid Period: Senegal case study}

\alttitle{Chronologie et ampleur de la dessiccation du Sahel \`{a} la
fin de la p\'{e}riode humide africaine : \'{e}tude de cas du
S\'{e}n\'{e}gal}

\author{\firstname{Magloire} \lastname{Mandeng-Yogo}\CDRorcid{0009-0009-2948-3601}\IsCorresp}
\address{Laboratoire d'Oc\'{e}anographie et du Climat -
Exp\'{e}rimentations et Approches Num\'{e}riques (LOCEAN/IPSL), IRD,
CNRS UMR 7159, Sorbonne Universit\'{e}, Paris, France}
\email[M. Mandeng-Yogo]{Magloire.Mandeng-Yogo@ird.fr}

\author{\firstname{Anne-Marie} \lastname{L\'{e}zine}\CDRorcid{0000-0002-3555-5124}}
\addressSameAs{1}{Laboratoire d'Oc\'{e}anographie et du Climat -
Exp\'{e}rimentations et Approches Num\'{e}riques (LOCEAN/IPSL), IRD,
CNRS UMR 7159, Sorbonne Universit\'{e}, Paris, France}
%\email{anne-marie.lezine@locean.ipsl.fr}

\keywords{\kwd{AHP}\kwd{Sahel}\kwd{Senegal}\kwd{Niayes}\kwd{Dessication}\kwd{Pollen analysis}\kwd{Geochemical analysis}}

\altkeywords{\kwd{AHP}\kwd{Sahel}\kwd{S\'{e}n\'{e}gal}\kwd{Niayes}\kwd{Dessiccation}\kwd{Analyse pollinique}\kwd{Analyse g\'{e}ochimique}}

\thanks{ACCEDE ANR Belmont Forum (18 BELM 0001 05), IRD, CNRS}

\begin{abstract}
Available archives of past environments and climates in Senegal are
relatively limited in number and concentrated in ``Niayes'' coastal
area. The review of available data for the 6 ka--2 ka period presented
here provides a major contribution, revealing that the end of AHP
occurred in two successive stages. A first dry event was dated at 4.5
ka. However, its impact was considerably attenuated in this region due
to specific conditions on the Atlantic coast, compared to other areas
of the Sahel. Depending on the site, gallery forests were either
minimally affected or, conversely, replaced by wooded grasslands. After
a brief return to humid conditions, the Niayes region experienced
dessication at 2.5 ka, with the drying up of water bodies and dramatic
disruption of the gallery forests.
\end{abstract}

\begin{altabstract}
Les archives disponibles sur les environnements et les climats
pass\'{e}s au S\'{e}n\'{e}gal sont relativement peu nombreuses et
concentr\'{e}es dans la zone c\^{o}ti\`{e}re des \og Niayes \fg.
L'examen des donn\'{e}es disponibles pour la p\'{e}riode comprise entre
6000 et 2000 ans avant notre \`{e}re pr\'{e}sent\'{e} ici apporte une
contribution majeure, r\'{e}v\'{e}lant que la fin de l'AHP s'est
produite en deux \'{e}tapes successives. Un premier \'{e}pisode de
s\'{e}cheresse a \'{e}t\'{e} dat\'{e} \`{a} 4500 ans avant notre
\`{e}re. Cependant, son impact a \'{e}t\'{e} consid\'{e}rablement
att\'{e}nu\'{e} dans cette r\'{e}gion en raison des conditions
sp\'{e}cifiques de la c\^{o}te atlantique, par rapport \`{a} d'autres
zones du Sahel. Selon les sites, les for\^{e}ts-galeries ont
\'{e}t\'{e} soit tr\`{e}s peu affect\'{e}es, soit remplac\'{e}es par
des prairies bois\'{e}es. Apr\`{e}s un bref retour \`{a} des conditions
humides, la r\'{e}gion des Niayes a connu une dessiccation \`{a} 2,5
ka, avec l'ass\`{e}chement des plans d'eau et une perturbation
dramatique des for\^{e}ts-galeries.
\end{altabstract}

\maketitle

\vspace*{4pt}

\twocolumngrid

\end{noXML}

\defcitealias{Lezineetal2011b}{ibid.}
\defcitealias{Bouimetarhanetal2009}{ibid.}

\section{Introduction}

Over the past two decades, the timing and amplitude of the end of the
African Humid Period (AHP) in tropical North Africa have been the
subject of extensive debates. Was the transition from wet conditions
that favored the expansion of tropical forest trees in the Sahel and
the Sahara during the Holocene \citep{Watrinetal2009} to the modern
semi-arid and arid landscapes abrupt or gradual? While
\citet{deMenocaletal2000} then \citet{Shanahanetal2015} identified the
end of AHP in Sahara as an abrupt event around 5~ka,
\citet{Kropelinetal2008} and \citet{Lezineetal2011a} showed that the
humid-arid transition was more gradual, spanning between 4.3~ka and
2.7~ka.\ High-resolution records of past \mbox{hydrological} conditions
and their impact on natural vegetation are scarce in the Sahel.
However, the Atlantic coastline north of the Cape Verde peninsula in
Senegal, known as the ``Niayes region'', is one of the most
well-documented areas. Several factors, including rainfall, sea-level
variations, and groundwater behavior, have influenced the hydrology of
this region during the Holocene, contributing to the complexity of
local hydrology and distribution of natural vegetation
\citep[e.g.][]{Chateauneufetal1986, LezineChateauneuf1991,
Maugisetal2009, Putallaz1962}. In 1989, L\'{e}zine demonstrated that
this region underwent large-scale environmental changes, with the
widespread expansion of tropical-humid gallery forests during AHP.
Recent high-resolution geochemical and palynological studies\break
\citep{Falletal2010, LemonnierLezine2021, Lezineetal2019,
Ndiayeetal2022} have shed new light on the chronology and amplitude of
the collapse of these gallery forests at the end of AHP and the
establishment of \mbox{present-day} 
\mbox{landscape.} This review presents Holocene
data from the entire ``Niayes region'', from the Cape Verde peninsula
in the south to the mouth of the Senegal River in the north, with the
aim of clarifying the different stages of the shift from humid to dry
conditions at the end of AHP and the related collapse of forests. In
the absence of archaeological data in this specific region---unlike the
Senegal river valleys \citep[e.g.,][]{Bocoum1986, McIntoshBocoum2000}
and Fal\'{e}m\'{e} \citep{Chevrieretal2016}, or the Sine Saloum and
Gambia river valleys \citep[e.g.,][]{Camaraetal2017, Laporteetal2017,
Holl2022}---Paleoenvironmental changes discussed here are thought to be
linked to rainfall change. Data from lakes and rivers in other regions
of Senegal are used to provide a regional picture of environmental
change at the\break end of AHP.

\section{The Niayes region of Senegal} \label{sec2}

``Niayes'' are interdunal depressions located behind the Atlantic
coastal strand of Senegal, spanning roughly 15{\textdegree}N to
16{\textdegree}N. These small, isolated basins or elongated depressions
perpendicular to the coast correspond to former channels whose outlets
to sea have been obstructed by sand dune formations over the last
millennia.\ Interdunal depressions support plant communities that
include an extension of humid tropical forests to the north, favored by
local moisture conditions and low evaporation along the coast. They
form the ``sub-Guinean domain'' described by \citet{Trochain1940},
including oil palms (\textit{Elaeis guineensis}) and other Guinean and
Sudano-Guinean trees (\textit{Phoenix reclinata, Syzygium guineense,
Alchornea cordifolia, Lophira alata}\,\ldots) that form gallery forests
along water bodies (\mbox{Figure}~\ref{fig1}B).\looseness=-1

\begin{figure*}
\includegraphics{fig01}
\vspace*{1pt}
\caption{\label{fig1}(A)~Relations between fresh groundwater and
salt-water front in Niayes \citep[from][]{Chateauneufetal1986}; 
(B)~Guinean and
Sudano-Guinean trees at dune's foot in a Niaye. Dune tops are covered
with~steppic (Sahelian) shrubs and herbs.}
\vspace*{2pt}
\end{figure*}

\looseness=-1
Today, the sub-Guinean domain of Niayes lies in an azonal position
compared to the steppe vegetation of the inland Sahel. Tropical plants
it contains benefit from a freshwater table close to the surface at the
bottom of the interdunes. The level of this freshwater table mainly
depends on the rainfall amount and sea-level position. Freshwater
recharge originates from Thi\`{e}s plateau located approximately 25~km
to southeast and gradually \mbox{decreases} \mbox{northward} and westward
(Figure~\ref{fig2}).\ Contact between the freshwater table and seawater
front reaches the \mbox{surface} along the beach, descending through
Quaternary sands towards the internal littoral domain
\citep{Chateauneufetal1986}.

\begin{figure*}
\includegraphics{fig02}
\caption{\label{fig2}Piezometric map of the Niayes region in 1962
\citep{Putallaz1962}. Blue and grey dots show location of pollen and
geochemical records cited in the text, respectively. Arrows indicate
lowering of the water table from Thi\`{e}s plateau.}
\end{figure*}

Tropical humid forests were widely developed in the Niayes region
during the African Humid Period \citep{Lezine1988}, due to heavy
monsoon rainfall over West Africa \citep{Lezineetal2011b}. However,
these forests are now severely degraded by agricultural and industrial
practices, accompanied by heavy water table pumping, as well as the
severe drought that occurred in the Sahel between around 1970 and 1980
\citep{Aguiaretal2010}.

\section{Paleoenvironmental context} \label{sec3}

Specific environmental conditions in the Niayes region are influenced
by two major factors: rainfall and sea-level variations.


\subsection{Sea-level variations along the Senegalese coast}
\label{sec3.1}

Numerous radiocarbon dates on marine mollusk shells and mangrove
sediments have been used to reconstruct sea-level changes along the
Senegalese coast \citep{Faureetal1980}. These dates indicate a gradual
rise in the sea after the lowest level recorded during the Last Glacial
Maximum \citep[ca.\ ${-}$100~m,][]{Vacchietal2025}, with modern
level reached around 7000 years ago. The sea-level then fluctuated
within a small amplitude of plus or minus 2~m \citep{Faureetal1980}.
However, due to the low elevation in the coastal zone, these variations
led to significant environmental changes, such as the invasion of
\textit{Rhizophora} mangroves into the Senegal River delta and along
the river bed for several hundred kilometers upstream, including Lakes
Rkiz \citep{GrouardLezine2023} and Guiers \citep{Lezine1988}. While
there is consensus on the identification of a transgressive period
peaking around 5~ka, the ``Nouakchottian'' \citep{Elouard1968},
subsequent period is subject to more debate \citep{Barusseauetal1995}
due to fewer available dates. Nevertheless, \citet{GrouardLezine2023}
showed the retreat of the \textit{Rhizophora} mangrove at Lake Rkiz
between 4.2 and 2.1~ka, coinciding with marine regression identified
further north in Mauritania, the \mbox{``Tafolian''} \citep{Hebrard1978}.
Then, during the last millennium, a final, modest marine
\mbox{transgression} led to the intrusion of saltwater into the bed of
the Ferlo River, a~tributary of the Senegal River, and the
\mbox{redevelopment} of \textit{Rhizophora} near Lake Rkiz
\citep{Faureetal1980, Fofanaetal2020, GrouardLezine2023,
Monteilletetal1981}. 

\begin{figure*}
\vspace*{-2pt}
\includegraphics{fig03}
\vspace*{-2pt}
\caption{\label{fig3}(A)~African Humid Period derived from radiometric
dating of four lake sediment categories from across Sahara and Sahel
perennial (lacustrine) and seasonal (playas) lakes, rivers
(fluviatile) and wetlands (palustrine). Insolation curve from
\citet{BergerLoutre1991}. (B)~Northward migration of tropical trees
across North Africa during AHP. Dots show the first appearance of
selected tree pollen taxa (\textit{Celtis, Alchornea}, \textit{Syzygium,
Piliostigma}) at the latitude of the studied sites \citep{Lezine2017}.}
\vspace*{-2pt}
\end{figure*}

\subsection{The African Humid Period (AHP)} \label{sec3.2}

The increase in summer insolation induced by orbital changes during the
last glacial--interglacial transition enhanced a thermal contrast
between land and sea, producing heavy summer monsoon rains that, in
turn, formed numerous lakes in the current arid regions of Sahara and
Sahel \citep{COHMAP1988}. Dated hydrological records from these {regions}
show that the number of deep freshwater lakes \mbox{gradually} increased from
as early as 15.5~ka, then accelerated from 12~ka to reach a maximum
between 9.5 and 8.5~ka, in a phase with a summer insolation maximum
\citep{BergerLoutre1991} (Figure~\ref{fig3}). Then the number of lake
records regularly decreased over time until present. From the
mid-Holocene, the increase in palustrine (marsh/swamp) records, coupled
with a gradual decrease in lake records, clearly confirms the
progressive dryness in the region. In more detail, the drying up of the
Saharan lakes occurred as early as the mid-Holocene, between 7.6~ka and
5.2~ka in Sudan \citep{Ritchieetal1985, RitchieHaynes1987} or
around 5.9~ka in Algeria {\citep{Lecuyeretal2016}.} In northern Chad, at
Lake Yoa, the {transition} from a monsoon-dominated climate regime
to one dominated by northern Mediterranean influences occurred around
4~ka, and present-day desert conditions were definitively established
by 2.7~ka \citep{Kropelinetal2008, Lezineetal2011a}.\looseness=-1

Vegetation responded to increased precipitation during the African
Humid Period by migrating tropical plants northwards, up to the tropics
\citep{Helyetal2014, Watrinetal2009} (Figure~\ref{fig3}B). These plants
used lake and riverside wetlands as migration routes to enter the
desert. \citet{Ritchieetal1985} and \citet{RitchieHaynes1987} suggested
that the southern limit of desert steppes was located 400~km north of
its current position, around 16{\textdegree}N, while the northern limit
of the Sudanian woodlands and wooded grasslands, currently at
14{\textdegree}N, shifted northwards to around 19{\textdegree}N during
the period between 10~ka and 5~ka. \citet{Watrinetal2009} and
\citet{Helyetal2014} have also shown that the distribution of
vegetation in Sahara and Sahel during AHP was more complex than a
simple displacement of vegetation zones. Plants now in distinct
phytogeographical zones were capable of co-occurrence, with tropical
trees (Sudanian to Guineo-Congolian) forming gallery forests near
rivers and lakes, and Saharan steppic taxa on the surrounding dunes.

By the end of the AHP, the tropical trees retreated southwards. In the
central Sahel, the forest cover, expressed as the total percentage of
arboreal pollen in Figure~\ref{fig4}, decreased dramatically from 4~ka
onwards to its modern aspect. In the western Sahel (i.e., in the Niayes
region), dense forest cover persisted for over about 2000 years due to
the specific conditions of the Atlantic littoral
\citep{Lezineetal2021}.

\begin{figure}
\includegraphics{fig04}
\caption{\label{fig4}Tree cover at the end of the AHP in the central
Sahel. Kajemarum \citep{SalzmannWaller1998} and Jikaryiya
\citep{Walleretal2007} sites (northern Nigeria) show that trees
(Arboreal Pollen~\%) fall from 4~ka onwards, more or less spectacularly
depending on the site.}
{\vspace*{-.3pc}}
\end{figure}

\begin{figure}
\includegraphics{fig05}
\bcaption{\label{fig5}The end of the AHP in Senegal: (A)~bleu dots:
dated record of sea-level fluctuations as a function of altitude
(Table~01); (B)~dated hydrological records as a function of latitude
(Table~02) (dark blue: lacustrine/high water level; green:
palustrine/low water level; light blue: fluviatile); (C)~isotopic
geochemistry ($\updelta^{13}$C) of bulk organic matter from two distinct
Niaye depressions at Mboro \citep{Falletal2010, Ndiayeetal2022};
}{4}{(D)~pollen records (in percentages) of tropical humid (sub-Guinean)
taxa in western Senegal (Niayes region and Senegal River delta), from
the bottom to the top: Diogo, Lompoul, Mboro-Baobab, St Louis,
ToubaN'Diaye and Potou
\citep[][{\url{https://africanpollendatabase.ipsl.fr}}]{Fofanaetal2020,
LemonnierLezine2021, Lezine1988}. Grey bands indicate dry periods.}
\end{figure}

\section{The data set} \label{sec4}

Sea-level variations are deduced from fossil shells and marine terraces
\citep{Faureetal1980}. Of the 88 dated records, only 61 are associated
with altitude (Figure~\ref{fig5}A; Table~01). Hydrological records
follow the methodology developed in \citet{Lezineetal2011b}. We use 71
dated records of fluviatile, lacustrine (high~water~level), and
palustrine (low water level) hydrological conditions from 12 localities
in Senegal (Figure~\ref{fig5}B; Table~02). These data were extracted
from literature and field campaigns. Each sample was assigned to its
specific category by \mbox{considering} the original interpretation of
the authors and our own knowledge of tropical paleohydrology based on
sedimentological (grain-size distribution), geochemical
(${}^{13}$C~data) and micropaleontological data: e.g., diatoms, aquatic
pollen types \citepalias{Lezineetal2011b}. Dating is based on accelerator
mass spectrometry (AMS) and conventional radiocarbon dates. Raw
${}^{14}$C dates were converted to calendar ages using CALIB~8.2
\citep{Reimeretal2013}. The reservoir age of 511 ${\pm}$ 176 years
calculated for the sector of the West African coast closest to the site
was applied to marine shells \citep{Ndeye2008}. Organic isotope records
from \citet{Falletal2010} and \citet{Ndiayeetal2022} are shown in
Figure~\ref{fig5}C. These records are expressed as ${\updelta}^{13}$C,
with less negative values (${-}$14 to ${-}$16{\textperthousand}) indicating
dry environmental conditions and grassland expansion, and the most
negative values (${-}$25, ${-}$27{\textperthousand}) indicating moist and
forest expansion \citep{Desjardinsetal2020}. Pollen percentages of
tropical humid taxa (Guineo-Congolian to Sudanian) are shown in
Figure~\ref{fig5}D. As is usual in paleopalynology, they are calculated
against a sum excluding aquatics and ferns. Age models used for both
isotope and pollen records are those from the authors, except for
\citet{Falletal2010}, for which it has been calculated using a linear
interpolation between two dated levels, with the top of the sequence
corresponding to zero.

\section{Results} \label{sec5}
\subsection{Paleohydrological changes} \label{sec5.1}

Hydrological records show that end of the AHP occurred in two
successive drying phases (Figure~\ref{fig5}B,D): 
\begin{itemize}
\item An early drying event is recorded between 4.8 and 3.4~ka. At
Mboro, \citet{Ndiayeetal2022}
indicate that arid conditions set in around 4.8~ka, based on ${}^{13}$C
data (Figure~\ref{fig5}B,C). This arid phase continued until 3.8~ka in
the Niayes region, as shown by the record of low water levels at Potou
(Figure~\ref{fig5}B). In addition, the absence of fluvial sediments in
the Bao Bolon between 4.2 and 3.4~ka suggests that runoff ceased,
confirming the setting of dry conditions (Figure~\ref{fig5}B).
\item After a return to wet conditions, a deep and widespread drought
set in at 2.5~ka.\ All dated hydrological records from the Niayes
region reveal a lowering of the water table
(\mbox{Figure}~\ref{fig5}B). This drought followed the gradual, though
fluctuating, drying of hydrological conditions at Mboro from 4.3~ka
onward (Figure~\ref{fig5}C). These observations agree with previous
data from \citet{Bouimetarhanetal2009}, who showed the dramatic
decrease of the Senegal River input starting from 2.5~ka then peaking
at 2.2~ka.
\end{itemize}

Characterization of hydrological conditions during the last two
millennia is more complex: \citetalias{Bouimetarhanetal2009} report periodic
flash flood events in Senegal River after 2.2~ka, while
\citet{Nizouetal2010} suggest a return to humid conditions between 1.0
and 0.7~ka. River flows in the Fal\'{e}m\'{e} and Bao Bolon ceased
between 1.2 and 0.2~ka \citep{Chevrieretal2016, Davidouxetal2018,
Sternetal2019} (Figure~\ref{fig5}B). A similar trend was recorded
northward in the Senegal River Valley in St.~Louis
\citep{Fofanaetal2020} where dry conditions occurred at 0.2~ka. Dry
conditions are also observed at Mboro in the Niayes region between 0.7
and 0.4~ka \citep{Falletal2010} (Figure~\ref{fig5}C). Aridity was
definitively established during the 19th century \citep{Fofanaetal2020,
Lezineetal2019}.

\subsection{Forest evolution} \label{sec5.2}

Pollen data from the Niayes region show two different situations likely
influenced by sites' proximity to the water table (Figure~\ref{fig2}).
Diogo and Lompoul record a spectacular and continuous development of
sub-Guinean gallery forests throughout the Holocene, with pollen taxa
percentages equal to or exceeding 70\% even after early Holocene forest
\mbox{optimum} \citep{Lezine1988}. At both sites, gallery forests collapsed
abruptly around 2.5~ka, with a dramatic loss of 35--50\% of tropical
forest taxa. At Touba N'Diaye and Potou, forest development was more
variable. Sub-Guinean gallery forests were severely disturbed between
5.4 and 4.5~ka. Then they experienced a significant resurgence, peaking
at 3.5~ka at Potou and 3.9~ka at Touba N'Diaye. This final period of
forest development ended ca.~2.6~ka. A discrepancy between the two
sites is likely due to the low resolution of data and the limited
number of dates used to construct age models.

\section{Discussion} \label{sec6}
\subsection{The end of AHP in the Niayes region:{\hfill\break} a
2-stage process} \label{sec6.1}

The evolution of West Tropical African vegetation during the Holocene
shows that, after the 9~ka forest optimum \citep{Helyetal2014}, Climate
became more seasonal, allowing the development of Sudanian or
Sudano-Guinean plants tolerant of a 5--7 month dry season at the
expense of earlier, more humid forests. The first widespread forest
disruption, however, was recorded later at 4.5~ka, corresponding to the
strong disturbance or disappearance of gallery forest in Nigeria
(Figure~\ref{fig4}) and Chad \citep{Kropelinetal2008, Lezineetal2011a,
SalzmannWaller1998, Walleretal2007}. In Senegal, this event coincided
with the opening up of Niayes landscape at Mboro, Potou, and Touba
N'Diaye, as well as the cessation of Bao Bolon river activity. At
Mboro, the landscape deteriorated to the point that gallery forests
were replaced by grasslands \citep{Ndiayeetal2022}
(Figure~\ref{fig5}C). At Potou and Touba N'Diaye however the forest
strongly declined but never completely disappeared
(Figure~\ref{fig5}D). Unlike the central Sahel, this initial warning
signal in the Niayes region was followed by a phase of forest
redevelopment between 4~ka and 2.5~ka. However, local hydrology
fluctuated greatly, with alternating wet and dry periods as the overall
climate gradually dried, as shown by Mboro's isotopic data
\citep{Falletal2010, Ndiayeetal2022}. Senegal River runoff was
significant during the 2.9--2.5~ka \citep{Bouimetarhanetal2009} or
2.75--1.9~ka interval \citep{Nizouetal2010}. The tipping point
\mbox{toward} the end of the AHP was reached at 2.5~ka, with the
collapse of many gallery forests and establishment of a grass-dominated
landscape. After 2.5~ka, all dated hydrological records indicate the
lowering of Niayes water bodies, with maximum drying between 2.1~ka and
1.4~ka. Yet, the Senegal River Basin experienced periodic flash floods
\citep{Bouimetarhanetal2009}, and Bao Bolon and Fal\'{e}m\'{e} rivers
continued flowing until 1.3~ka \citep{Chevrieretal2016,
Davidouxetal2018, Sternetal2019}. This chronology of events agrees well
with the sequence in northern Chad \citep{Lezineetal2011a}, where
monsoon influence declined around 4~ka, followed by the establishment
of modern climatic conditions around 2.7~ka.

\looseness=-1
Over the last two millennia, gallery forests have not disappeared
entirely from the Niayes region and Senegal River banks. They increased
slightly between 1100--1680~CE \cite{LemonnierLezine2021,
Ndiayeetal2022}, due to favorable climate conditions as indicated by
Senegal River freshwater discharge \citep{Fofanaetal2020} and reduced
dust transport to the ocean \citep{Mulitzaetal2010}. The present-day
Niayes and Senegal River delta environment dates back 1850~CE
\citep{Lezineetal2023}.

\subsection{Have changes in seawater levels influenced groundwater
levels in the Niayes and the{\hfill\break} development of gallery
forests at the end of the AHP?} \label{sec6.2}

Configuration of piezometric level in the Niayes region
(Figure~\ref{fig2}) and proximity of the salt front may have
significantly influenced forest cover heterogeneity and hydrological
conditions revealed by palynological and geochemical analyses. Here, we
attempt to assess their impact during the two dry events that marked
the end of the African Humid \mbox{Period}.

The 4.5~ka dry event occurred during the ``Nouakchottian'' marine
transgression, when sea-level was ${+}$3~m above present
(Figure~\ref{fig5}A). Contact between salt- and freshwater in the
Niayes region had therefore shifted inland compared to today
(Figure~\ref{fig1}). This situation may have maintained humidity at the
bottom of the interdunes, as the groundwater table lay above the
saltwater table. This can be seen at Diogo and Lompoul, where dense
gallery forests persisted despite arid climate. In contrast, the impact
of the 4.5~ka dry event was more pronounced at Mboro-Baobab, Touba
N'Diaye, and Potou, leading to a reduction in forest cover. This
suggests these sites were located closer to the contact zone between
salt and freshwater, and thus more sensitive to hydrological changes
than previous sites. However, the situation in the Niayes was never as
severe as in the central Sahel and Sahara, where tropical trees were
definitively eliminated.

Between 4~ka and 2~ka, as sea-level generally declined, the Niayes
environment underwent profound transformations directly linked to
climate change: the phase of forest development and rising water table
between 4~ka and 2.5~ka clearly indicates the establishment of humid
conditions in the Niayes region. The filling of ponds at Sil and flu
disruption of forests at 2.5~ka reflects activity of the Bao Bolon
(Figure~\ref{fig5}B) show this humid phase extended beyond the Niayes,
into the Sudanian savannah zone \citep{LezineCasanova1989}. Conversely,
widespread disruption of forests at 2.5~ka reflects the abrupt onset of
arid conditions in the Niayes region. The retreat of the coastline and
salt front likely amplified the impact of this drought episode on
forest disturbance. By 2.5~ka, the Sudanian zone had returned to wetter
conditions, as demonstrated by continuous fluvial records in the Bao
Bolon.

\section{Conclusion} \label{sec7}

Combined records from various sources for the period between 6~ka and
2~ka in Senegal reveal a nuanced picture of the end of the African
Humid Period in the region. Rather than a single, abrupt shift from
humid to arid conditions, the Niayes region experienced a two-phase
transition. An initial drying event around 4.5~ka, marked by forest
disruption, was followed by a period of fluctuating hydrology and
forest regeneration. However, a tipping point occurred at 2.5~ka,
leading to widespread forest collapse and the establishment of a
grassland-dominated landscape. These two major phases demonstrate the
non-linear response of continental systems to the gradual weakening of
monsoonal fluxes over North Africa. This aligns with broader trends in
the Sahel and the Sahara \citep{Collinsetal2017, Gasseetal1990}, but it
highlights the significant influence on local \mbox{factors,} particularly
sea-level changes and their effect on groundwater, which modulated the
regional drying trend. In this highly diverse area, due to its
proximity to the shoreline and the subsequent influence of sea-level
variations on \mbox{groundwater} \mbox{levels,} \mbox{significant} differences can be
observed in the response of each interdunes depression to climate
change. For instance, forest cover at Diogo and Lompoul sites exhibited
an abrupt response at 2.5~ka, in contrast to the more gradual and/or
highly variable response observed at the other Niayes sites. The
heterogeneity of the vegetation response within this relatively small
region underscores the complex interplay between climate, hydrology,
and vegetation dynamics in this area during the end of the AHP. This
emphasizes the importance of considering local hydrological conditions
when reconstructing past climates and predicting future environmental
changes.

\section*{Author contribution}

MMY and AML designed the study and contributed to writing of the
manuscript. 

\section*{Acknowledgements}

This work contributes to ACCEDE ANR Belmont Forum project (18 BELM 0001
05) (France). MMY was funded by IRD and AML by CNRS in France. AML
thanks the African Pollen Database for making pollen data stored and
freely available.

\CDRGrant[ANR]{18 BELM 0001 05}

\section*{Declaration of interests} 

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.

\section*{Data availability}

Pollen data are stored in the African Pollen database
(\url{https://africanpollendatabse.ipsl.fr}).\ Other data used in this
paper are available in the supplementary section or directly from the
literature.

\section*{Underlying data}

Supporting information for this article is available on the journal's
website under \printDOI\ or from the author.
The underlying data for this article is available at
\url{https://doi.org/10.23708/RPBPQH}.

\CDRsupplementaryTwotypes[application/zip]{supplementary-material}{\cdrattach{Table-01.xlsx}}
\CDRsupplementaryTwotypes[application/zip]{supplementary-material}{\cdrattach{Table-02.xlsx}}

\back{}

\printbibliography
\refinput{crgeos20250249-reference.tex}

\end{document}