2.1.1. Location and General Context of the Study Area

Core C61 was collected in the Gulf of Guinea, on the central part of the Cameroonian continental shelf, at a depth of approximately 200 meters (Figure 1). This area lies along the upper margin of the continental slope and is strongly influenced by riverine inputs from the Sanaga River, as well as other coastal rivers such as the Wouri, Nyong, and Ntem. These rivers transport fine-grained sediments rich in organic matter and pollen into the marine environment. The Sanaga River watershed, which spans roughly 133,000 km², covers a broad ecological gradient from humid equatorial forests in the south to drier savanna regions in the center and north of Cameroon ([5]). This wide range of source environments contributes to the mixed continental signature of the sedimentary inputs observed offshore. Since most detrital particles reaching the shelf originate from continental erosion, the marine sedimentation in this area primarily reflects terrigenous input, aside from biogenic components produced in situ.

Figure 1. Location of the sampling sites on the Cameroonian continental shelf.

The regional oceanic circulation, dominated by the Equatorial Countercurrent and the Guinea Current ([8]; [9]), facilitates the dispersion of particles and the development of a continuous sedimentary record that accurately reflects paleoenvironmental variations from the Late Pleistocene to the recent Holocene ([28]; [6]).

2.1.2. Geological and Geomorphological Framework

The Cameroonian continental shelf is a geological structure inherited from the opening of the South Atlantic during the Early Cretaceous, characterized by N60-oriented faults that produced an alternation of horst and graben-type tectonic blocks ([1]). The sedimentary cover mainly consists of clayey and silty deposits of fluvial origin, largely supplied by the Sanaga River basin. The continental relief, organized in successive steps from the coastal plain to the Adamawa Plateau, promotes intense erosion and a rapid transfer of terrigenous material to the sea. In marine settings, sedimentary dynamics are expressed by recent deltaic deposits in the north, whereas in the south, relict sediments such as glauconitic and bioclastic sands—remnants of the Last Glacial Maximum and subsequent Holocene transgressions—predominate ([27]; [54]). Overall, this region represents a complex sedimentary system shaped by the combined influence of tectonic, climatic, and hydrological processes, which together account for the high heterogeneity of the deposits.

2.1.3. Pedological Framework of the Sanaga River Basin

The Sanaga River basin, the main source of sediments supplying the Cameroonian continental shelf, exhibits a wide diversity of soils that directly influence the composition of particles exported to the marine environment. In the humid equatorial zone of southern Cameroon, kaolinite-rich ferralitic soils, developed on ancient Precambrian basement rocks, are predominant. Swampy areas host hydromorphic soils enriched in organic matter, while the volcanic regions of the west and southwest are covered by Andosols derived from basic volcanic rocks ([66]; [49]). Under the combined effects of heavy rainfall and stepped topography, these soil formations undergo intense erosion, resulting in the transport of clays, silts, and pollen grains to the ocean.

2.1.4. Atmospheric and Oceanographic settings

The climate of the Cameroonian coast is largely controlled by the Intertropical Convergence Zone (ITCZ), whose seasonal movements determine an alternation between a wet season dominated by oceanic air masses and a dry season marked by the continental Harmattan winds ([36]; [57]; [68]). These atmospheric variations directly influence regional ocean circulation in the Gulf of Guinea ([8]; [9]; [33]), characterized by a northward coastal drift current and the southward-flowing Guinea Current (Figure 1). At the surface, these currents facilitate the transport and dispersion of fine particles and pollen material along the coast, while in deeper waters, the Guinea Current promotes sediment redistribution offshore.

2.1.5. Current Climate and Vegetation Formations

The coastal region of Cameroon experiences a humid equatorial climate, characterized by mean annual temperatures ranging from 27 to 29˚C and exceptionally high rainfall, locally exceeding 10,000 mm per year at the foot of Mount Cameroon ([68]). This high level of humidity promotes the development of dense and diverse vegetation, dominated by humid evergreen forests rich in Caesalpiniaceae, Euphorbiaceae, and Ulmaceae ([37]; [72]). Along the margins of flooded and lagoonal areas, swamp forests and mangroves dominated by Rhizophora and Avicennia thrive, whereas inland from the coast, semi-deciduous forests give way to wooded savannas in drier regions. This vegetation mosaic, governed by climatic and topographic gradients, illustrates the gradual transition between forest and savanna biomes. Its present-day floristic composition is reflected in the pollen signal preserved in marine sediments, thereby providing an essential reference framework for the reconstruction of Holocene paleoenvironments ([44]; [5]; [6]).

2.2.1. Sampling Protocol

A total of 26 samples were collected: 16 at 10 cm intervals in the upper section (0 - 150 cm), where sedimentation was relatively slow, and 10 at 30 cm intervals in the lower section, characterized by more rapid sedimentation. This sampling strategy provides representative coverage of paleoenvironmental variations spanning from the end of the Pleistocene to the Late Holocene.

2.2.2. Palynological Processing

The palynological processing of sediments from core C61 was conducted following the standard procedures described by [22]. Each sample underwent a series of chemical treatments to remove carbonates, silicates, and non-pollen organic matter, followed by repeated washing and centrifugation to obtain clean residues. The resulting fractions were then mounted on microscope slides using glycerine jelly as the mounting medium. This procedure ensured good preservation and optimal visibility of pollen grains and spores for both quantitative and qualitative analyses.

2.2.3. Pollen Observation, Identification, and Classification

Observations were performed using an optical microscope at magnifications ranging from ×400 to ×1000. For each sample, a minimum of 300 pollen and spore grains were counted to ensure adequate statistical representativeness. Taxonomic identification was based on published atlases ([43]; [7]; [62]) and on the reference collection of the Palynology Laboratory at the University of Montpellier 2. The identified taxa were subsequently grouped according to their ecological affinities: dense evergreen forests, semi-deciduous forests, savannas, swamp formations, mangroves, and aquatic plants. This classification made it possible to establish ecological indicator groups used for the reconstruction of paleoenvironmental and paleoclimatic conditions.

2.2.4. Ecological Significance and Chronological Framework

The interpretation of pollen assemblages is based on the identification of bioindicator taxa that reflect climatic and ecological conditions through time. Poaceae and Bridelia indicate dry phases dominated by savanna formations, whereas Caesalpiniaceae, Sapotaceae, and Ulmaceae point to humid periods favorable for the development of dense forest vegetation. The presence of Podocarpus is associated with cooler conditions at higher elevations or during episodes of climatic cooling. When considered within the context of the Late Glacial and Holocene, these indicators allow the establishment of a relative chronology of paleoenvironmental changes and enable the correlation of regional vegetation dynamics with major global climatic fluctuations.

Geochemical analyses of core C61 show a progressive decrease in CaCO₃ content toward the top of the sequence, while total organic carbon increases. This trend is interpreted as a shift toward enhanced marine productivity and environmental stabilization in the uppermost levels. The clay mineral assemblage is dominated by kaolinite, with subordinate amounts of smectite and illite, indicating a predominantly continental origin of detrital inputs, primarily from the Sanaga River basin.

The lithological profile of the core, based on previous studies by [54] and [26] (Figure 2), reveals three main sedimentary units. The basal unit (540 - 355 cm) consists of grey-green clayey mud enriched in plant debris, reflecting strong fluvial input and rapid deposition. The middle section (355 - 145 cm) is composed of sandy mud containing micro-shell fragments, suggesting a transitional depositional environment with increasing marine influence. The uppermost unit (145 - 0 cm) is made up of soft, homogeneous grey mud rich in mollusk remains, characteristic of a low-energy, stabilized marine setting.

To establish the chronological framework of the sequence, six levels of core C61 were dated by radiocarbon (¹⁴C) analysis performed on both bulk organic matter and mollusk shells ([54]; [27]). The resulting calibrated ages (cal yr BP) range from approximately 11,330 cal yr BP at the base to 1450 cal yr BP at the top of the core.

The sedimentation rate was not uniform throughout the sequence. Between 11,300 and 9100 cal yr BP, the average sedimentation rate was approximately 157 cm per 1000 years. After 9100 cal yr BP, the rate dropped sharply to about 24 cm per 1000 years, and during the last millennium, it decreased further to approximately 10 cm per 1000 years (Figure 3).

A—Mollusk shell debris and Nests of soft fecal pellets; B—% Water; C—% Sand and % Pelites; D—% Kaolinite + % Smectite + % Illite; E—% CaCO₃; F—% Organic carbon. C¹⁴ dating (0 - 11,300). Estimated sedimentation rate (cm/10³ years).

Figure 2. Lithological profile of core C61 ([54]).

Figure 3.Age-depth model and sedimentation rates derived from calibrated ¹⁴C dates ([54]).

Table 1.Count of main pollen types in samples from the Cameroonian continental shelf.

Table 2.Pollen count data (absolute values).

Examining three to four transects per slide was generally sufficient to reach the required 300 pollen grains, except at the 70 - 80 cm level (seven transects) and the 430 - 440 cm level (two transects). In total, 26,444 pollen and spore grains were identified. Among other palynomorphs, approximately 939 were indeterminate, 132 were damaged, and 785 corresponded to dinoflagellate cysts (Table 1). The 224 recorded taxa, distributed among 93 families, show a marked dominance of Rhizophora (34.3%), spores (18.3%), Poaceae (10.1%), Cyperaceae (8.2%), and Alchornea (4.7%). The highest taxonomic diversity was observed in Rubiaceae (23 genera), followed by Euphorbiaceae (16) and Phyllanthaceae (11). These assemblages reflect the regional mosaic of forested and open vegetation types.

The pollen spectra forming the overall pollen diagram were constructed using two calculation matrices. The bioindicator taxa, which represent the two main pollen groups, were derived from matrix (1), which includes all identified pollen types. Matrix (2), used for the analysis of the other major and characteristic taxa, was generated from a dataset in which Rhizophora and spore taxa were excluded (Table 2).

Based on the temporal comparison of pollen spectra, significant changes in floristic composition were identified, allowing the distinction of three major successive palynological zones (Table 3), which correspond to the main paleoecological phases recorded in core C61.

Table 3.Description of the palynological zones.

Table 4. Plant types (herbaceous and arboreal forms) and major taxa (Rhizophora and spores).

The pollen groupings (Table 4), organized by plant type within the three palynological zones of the overall pollen diagram (Figure 4), are characterized by the following main plant assemblages: 1) biological forms (herbaceous and arboreal types); 2) taxa dominant in pollen proportions (Rhizophora, spores); 3) taxa serving as environmental and climatic bioindicators (Rhizophora, Pandanus, Podocarpus); and 4) taxa characteristic of the major ecosystems—savanna (Poaceae, Bridelia, Combretum), dense humid forest (Caesalpiniaceae, Sapotaceae, Ulmaceae, Sacoglottis), and ubiquitous or pioneer understorey taxa (Alchornea, Tetrorchidium, Drypetes).

Zone 1: Regression of ubiquitous pioneer and undergrowth taxa (9400-11300 years). Zone 2: Increase in spores during the Podocarpus phase, both alternating in contrast with Rhizophora (5400-9400 years). Zone 3: Forest resurgence in favor of Podocarpus (0-5400 years).

Figure 4.Pollen diagram of ecological bioindicators.

Table 5.Vegetation types and ecological bioindicators.

The bioindicator diagram, constructed from the data presented inTable 5 and illustrated in Figure 5, reveals marked variations through time. Alternating fluctuations in intensity are observed between, on the one hand, plant formations (forested, savanna, and mixed types) and, on the other hand, specific taxa such as Rhizophora, spores, and Podocarpus. Rhizophora pollen is particularly abundant in Zone 1, whereas spores and Podocarpus are only sparsely represented. In Zone 2, the proportion of spores increases significantly, indicating a more open and less humid phase. Finally, in Zone 3, Podocarpus exhibits a steady increase toward its maximum abundance, reflecting the expansion of montane or temperate vegetation under cooler and more humid climatic conditions.

Zone 1: Regression of ubiquitous pioneer and undergrowth taxa (9400-11,300 years). Zone 2: Increase in spores during the Podocarpus phase, both alternating in contrast with Rhizophora (5400-9400 years). Zone 3: Forest resurgence in favor of Podocarpus (0-5400 years)

Figure 5.Evolution of ecological bioindicateurs (Rhizophora, Podocarpus, Spores).

Pollen analysis reveals a substantial representation of numerous plant families, indicating high diversity at both the genus and species levels. The most abundant taxa, in terms of both absolute number and relative proportion, correspond to those identified in recent dredging samples from the Cameroonian continental shelf, notably Rhizophora, spores, Poaceae, Cyperaceae, Alchornea, Drypetes, Uapaca, and Podocarpus ([5]; [6]). These results corroborate previous observations on marine pollen sedimentation, emphasizing the key role of fluvial inputs, particularly from the Sanaga River—the main drainage system collecting runoff from watersheds that extend from the dense forests of the south to the savannas of central and northern Cameroon ([56]; [27]). Thus, the floristic diversity recorded in core C61 faithfully reflects the botanical richness of the adjacent continent.

However, overall taxonomic diversity remains slightly lower than that observed on land, likely due to taphonomic processes and the constraints imposed by marine dynamics ([24]; [14]; [71]). Indeed, incomplete fossilization and the long-distance transport of pollen grains from their continental source can result in the underrepresentation or even absence of certain taxa originating from distant ecosystems. In particular, arboreal taxa characteristic of savanna or semi-deciduous forest, montane or semi-deciduous forests—such Hymenocardia,Brideliaor Annona—are often poorly represented or absent in marine records, despite being common in the inland vegetation. This limitation, well documented in marine palynological studies from Central Africa and the Gulf of Guinea ([11]; [70]; [29]), may affect the accuracy of paleoenvironmental reconstructions based solely on marine records.

The palynological data from core C61 allow for a detailed reconstruction of paleoenvironmental dynamics on the Cameroonian continental shelf from the onset of the Holocene to approximately 1450 cal yr BP. Rather than beginning with general considerations, we first present the main vegetational and sedimentary stages identified in the core and interpret them in light of regional climatic trends and sea-level changes. Three distinct paleoecological phases are identified, reflecting the evolving balance between coastal dynamics, continental inputs, and climatic forcing.

First Phase (11,400-9400calyr BP): Coastal Expansion and Fluvio-lagoonal Environments Dominated by Rhizophora and Cyperaceae

This initial phase, corresponding to Zone 1, is characterized by the dominance of Rhizophora and Cyperaceae, with a co-occurrence of taxa such as Combretaceae, Bridelia, and various herbaceous pollen types (Figure 4). These assemblages indicate the development of mangrove systems and wetland savannas at the land–sea interface, under the influence of a warm and humid climate. The high abundance of Rhizophora suggests a well-established mangrove belt, while Cyperaceae and NAP taxa reflect seasonally waterlogged open habitats. Podocarpus is poorly represented, likely restricted to mid-altitude zones due to still-elevated regional temperatures ([44]; [34]).

This period follows the Late Dryas and corresponds to a time of increasing monsoon strength and hydrological intensification in Central Africa. Enhanced runoff from the Sanaga watershed, possibly linked to glacial meltwater inputs, likely contributed to accelerated sedimentation on the continental shelf ([27]; [54]). However, the role of glacial melting as a climatic driver must be interpreted with caution in this tropical basin, where local hydrodynamics and ITCZ displacement are more significant.

Second Phase (9400-5400calyr BP): Stabilization of Coastal Systems and Expansion of Savanna-Grassland Mosaics

During this interval, Cyperaceae decline while Poaceae and spores increase, accompanied by a retreat of shrubs such as Bridelia and Combretaceae. This floral transition suggests a shift toward more open environments, such as wooded savannas or anthropogenic clearings, but does not unequivocally signal aridification. Contrary to earlier interpretations, the available data do not support a markedly dry climate during this phase.

Instead, this period likely corresponds to the Holocene Climatic Optimum, characterized regionally by high rainfall and increased vegetation productivity ([53]; [25]). The expansion of grasses may reflect disturbance regimes, seasonal contrasts, or hydro-edaphic variability, rather than long-term drought. Simultaneously, a marine transgression reached its maximum around 6000 cal yr BP, gradually distancing the core site from the shoreline. This change in sedimentary context explains the dominance of marine muds and the reduced terrigenous input in this unit ([27]; [54]).

Third Phase (5400-1450calyr BP): Forest Expansion and Climatic Oscillations in a Stabilized Marine Context

This phase, corresponding to the upper part of the core, is marked by a sharp decrease in sedimentation rate (Figure 5), indicating post-transgressive stabilization of the marine depositional environment. The pollen assemblage reveals a diversification of arboreal taxa, including Caesalpiniaceae, Sapotaceae, and Ulmaceae, along with forest species such as Sacoglottis, Lophira, and Elaeis. Pioneer taxa (Alchornea, Tetrorchidium, Pycnanthus, Uapaca, Drypetes) are also well represented, pointing to mature forest development with dynamic understorey regeneration.

The occurrence of Podocarpus, coupled with peak values of spores, indicates episodes of cooler and wetter conditions, possibly linked to cloud-based occult precipitation in upland areas ([44]). Rather than a simple trend, this period reflects alternating wet and dry subphases, consistent with regional climatic variability during the Late Holocene ([34]; [68]; [53]). The observed vegetational transitions support a scenario of increased forest cover in the last two millennia, aligning with other Central African paleoecological records ([59]; [46]; [73]).

4.3.1. Variations in the Proportion of Rhizophora Spores and Pollen

Analyses performed on core C61 reveal that the high proportions of spores observed at various levels mainly result from a decrease in Rhizophora pollen. Spores are produced in large quantities by terrestrial or epiphytic ferns in a wide range of environments, from savannas to forests, including transitional zones. Such spore abundance is observed near the mouths of all coastal rivers. This interpretation is supported by modern observations from the Cameroonian continental shelf, which show that present-day concentrations display similar trends—that is, peaks in Rhizophora and spore abundance often coincide. However, the relative magnitudes of these peaks vary among dredged samples, and the proportions of each taxon depend on their proximity to mangrove development areas ([6]).

However, a clear spatial differentiation is observed: in the southern sector, where the coastline is devoid of mangroves, spore values reach high levels, whereas in the north, the presence of tidal influence, extensive mangroves, and the high production of Rhizophora pollen maintain these values at lower levels.

4.3.2. Dynamics of Rhizophora, Spores, and the Paleoecological Significance of the Coastal Signal

The variations in Rhizophora pollen frequency observed in core C61 reflect changes in the extent or proximity of mangrove ecosystems during the Holocene. Rhizophora is an excellent indicator of estuarine and lagoonal zones under saline influence, and its abundant presence generally suggests that the recording site was located relatively close to the shoreline under a stabilized sea level ([24]; [6]; [27]). Sharp declines in Rhizophora within certain sections of the pollen spectrum may reflect either a gradual offshore shift of the coastline due to Holocene transgression ([54]) or local ecological factors (e.g., sedimentation, salinity, turbidity) unfavorable to mangrove development ([14]).

In the same intervals, relative increases in spores probably reflect dilution effects in the pollen signal rather than an actual rise in spore production. The inverse alternation between Rhizophora and spores, particularly in Zones 2 and 3, therefore indicates variations in fluvial input and coastal vegetation cover, without necessarily implying a regressive or markedly cooler climatic episode ([6]; [35]).

Moreover, the intermittent presence of Podocarpus cannot be interpreted as direct evidence of altitudinal migration due to cooling. It is more likely that these occurrences reflect sporadic inputs from mid-altitude areas during wetter phases favorable to its regional development ([44]; [34]). In the absence of independent data on sea level or temperature, caution is warranted in interpreting these signals. Thus, the opposite variations among Rhizophora, spores, and Podocarpus should be viewed as the combined result of sedimentary dynamics, distance from the coastline, local ecosystem fluctuations, and taphonomic effects inherent to marine environments ([24]; [14]; [70]).

The palynological sequence from core C61 shows several transitions in vegetation composition that are partly comparable to continental records from Central Africa. However, some key intervals well documented in lake records—such as the forest retreat and climatic instability around 2500 cal yr BP—are not clearly expressed in this marine core.

This retreat has already been highlighted in recent syntheses of Holocene environmental change in Central Africa ([26]; [55]), which emphasize the widespread nature of the 2.5 ka dry phase, often associated with savanna expansion, lake-level fluctuations, and cultural changes. Its absence or attenuation in core C61 could reflect several factors: the limited number of palynological levels in the upper part of the core, the large intervals between radiocarbon dates, or taphonomic constraints linked to low sedimentation rates (Figure 3).

Despite this limitation, the overall trends observed in core C61—namely, mid-Holocene forest expansion and Late Holocene fluctuations—remain broadly consistent with other regional archives, such as Lake Barombi Mbo ([10]), Lake Ossa ([58]), and Kitina ([20]). These comparisons highlight the potential of marine pollen records to capture large-scale climatic phases, but also their limits in recording short-term or abrupt events, especially when sampling resolution is low.

Future work should therefore focus on improving the age-depth model of core C61 through additional radiocarbon dates, particularly in the uppermost sections, and increasing palynological resolution between 3000 and 1000 cal yr BP to verify whether regional signals such as the 2.5 ka dry event are preserved in the marine domain.