Introduction

Iron-manganese (Fe–Mn) oxide crusts and nodules (‘manganese nodules’) cover substantial areas of the global seafloor, attract considerable quantities of trace and rare earth elements due to their properties and slow accumulation rates (Koschinsky and Hein 2017), and are considered as a potential ore source for rare metals (Hein et al. 2013). Multidisciplinary approaches have been employed, especially in recent years, to understand oceanographic factors that influence spatial distribution and growth rates of Fe–Mn oxide crusts and nodules (e.g. Shkolnik et al. 2012; Josso et al. 2019, 2020; Benites et al. 2020; Charles et al. 2020). Fe–Mn oxide encrustations can be of different geochemical origin (Koschinsky and Hein 2017), although bacteria and even coccolithophores also play an important role in initiation and development of Fe–Mn oxide deposition (e.g. Wang and Müller 2009; Wang et al. 2011; Shkolnik et al. 2012; Jiang et al. 2019). Most work on dating Fe–Mn oxide crusts and nodules has been done based on trace and rare earth element isotopes and radiometric techniques (Josso et al. 2019, 2020, and references cited therein), but calcareous nannofossils have also been used to estimate the age of polymetallic concretions (Janin 1985, 1987) and the growth rates of ferromanganese nodules and crusts (Harada and Nishida 1976; Cowen et al. 1993) in the North-West and Central Pacific.

Very little work has been done on understanding Fe–Mn oxide deposition in SW Indian Ocean and, in particular, off South Africa and in the Mozambique Channel (Rogers 1995; Albarède et al. 1997; Frank et al. 2002; Charles et al. 2020).

During the SO-183 cruise of the RV Sonne (Jokat 2006), dredging of the Mozambique Ridge and in the Mozambique Basin resulted in the collection of a suite of manganese nodules, rocks encased by manganese encrustations, and individual manganese encrustations (Watkeys et al. 2006). Geochemical analyses of these samples, combined with previous data from the region, revealed that the manganese mineralisation was formed primarily by hydrogenetic processes, with secondary diagenetic enrichment having taken place on the Mozambique Ridge crest and possibly some hydrothermal enrichment on the Ridge flanks (Perritt and Watkeys 2007).

This paper reports on the calcareous nannofossils from the manganese nodules and encrustations of the Mozambique Ridge and in the Mozambique Basin. This provides the first model for establishing both the age and rate of Fe–Mn oxide precipitation in the SW Indian Ocean using micropalaeontological data and allows the genesis of these nodules and encrustations to be linked to major changes in oceanographic conditions and tectonic activity in the region during the late Cenozoic.

Materials and methods

The calcareous nannofossil assemblages come from samples collected from three localities (DL1, DL2, DL3) on the Mozambique Ridge and one (DL4) on the Jaguar seamount in the Mozambique Basin (Fig. 1, Table 1). The sites are up to 1500 km apart and range in depths from 4077 to 756 m. At all but one site, two dredges were undertaken. The dredging methodology was described by Watkeys et al. (2006).

Fig. 1
figure 1

Map of the Mozambique Ridge and environs with sample sites. The lines show the course of the ship, along which dredging was carried. The map is modified after König and Jokat (2010)

Full size image
Table 1 Summary of dredge localities
Full size table

Dredge samples of manganese nodules and encrustations were cut on-board the RV Sonne and examined under a binocular microscope. This revealed the presence of calcareous centres to manganese nodules, recorded as chalk cores, and thin calcareous layers at the base of and within the manganese encrustations (Watkeys et al. 2006). It was suspected that these layers contained calcareous nannofossils, which are a very useful tool for establishing the age of sediments, especially when only very small quantities of sediments are available for analysis (Bosman et al. 2007). In the laboratory on-shore, a standard technique of nannofossil identification was used for the preparation of the smear slides (Green et al. 2008). Totally 30 samples were analysed under a light microscope Olympus BX53 at 1250 × magnification with phase contrast and cross-polarised light. Photographs are taken with an Olympus digital camera (Fig. 2). Scanning electron microscopy (SEM) was employed to confirm identifications and for photographing biostratigraphically important species (Fig. 3).

Fig. 2
figure 2

Calcareous nannoplankton from the Mozambique Ridge and the Jaguar seamount in the Mozambique Channel. All illustrations are light micrographs. The abbreviations ‘pol’ and ‘tr’ denote polarised and transmitted light, respectively. a Discoaster triradiatus, sample DL3:05–36, tr; b Discoaster tamalis, sample DL3:05–36, tr; c Discoaster blackstockae, sample DL3:05–43, tr; d Discoaster asymmetricus, sample DL3:05–43, tr; e Discoaster brouweri, sample DL3:05–36, tr; f Discoaster brouweri, sample DL3:05–34, tr; g Discoaster pansus, sample DL3:05–36, tr; h Discoaster pentaradiatus, sample DL3:05–36, tr; i Hayaster perplexus, sample DL3:05–34, tr; j Helicosphaera carteri, sample DL3:05–48, pol; k Helicosphaera sellii, sample DL3:05–43, pol; l Reticulofenestra pseudoumbilicus, sample DL3:05–37, pol; m Gephyrocapsa oceanica, sample DL3:05–55, pol; n Gephyrocapsa parallela, sample DL3:05–35, pol; o Gephyrocapsa carribbeanica, sample DL3:05–48, pol; p Pseudoemiliania lacunosa, sample DL1:05–02, pol; q Calcidiscus macintyrei, sample DL3:05–43, pol; r C. macintyrei, sample DL3:05–43, tr; s C. leptoporus, sample DL3:05–38, pol; t Umbilicosphaera sibogae, sample DL3:05–38, pol

Full size image
Fig. 3
figure 3

SEM pictures of calcareous nannoplankton from the samples of the manganese deposits of the Mozambique Ridge and Mozambique Basin. a Discoaster brouweri, DL3:05–36; b Discoaster pentaradiatus, DL3:05–43; c Discoaster tamalis, DL3:05–36; d Syracosphaera pulchra, proximal view, DL3:05–43; e Discoaster cf. asymmetricus, DL3:05–36; f Discoaster tamalis, DL3:05–36; g Umbilicosphaera sibogae, distal view, DL3:05–37; h Gephyrocapsa carribbeanica, distal view, DL3:05–48; i Gephyrocapsa oceanica, distal view, DL1:05–01; j Calcidiscus leptoporus, distal view, DL3:05–48; k Calcidiscus macintyrei, distal view, DL3:05–36; l ‘small’ Gephyrocapsa spp., distal view, DL3:05–48; m Helicosphaera carteri, proximal view, DL3:05–38; n Calciosolenia fossilis, general appearance, DL1:05–02; o ‘large’ Gephyrocapsa spp., distal view, DL3:05–38; p Rhabdosphaera claviger, general appearance, DL3:05–43; q Scyphosphaera sp., lateral view, DL3:05–36; r Emiliania huxleyi, distal view, DL1:05–01

Full size image

The abbreviations FO and LO are used for the first known occurrence and the last known occurrence of the nannofossil species in the geological record.

Abundance of individual calcareous nannofossil taxa is recorded as follows:

  • A—abundant (> 10 specimens per field of view (FOV));

  • C—common (> 1–10 specimens per FOV);

  • F—few (1 specimen per 1–10 FOV);

  • R—rare (< 1 specimen per 10 FOV);

  • VR—very rare (1 specimen per slide).

Nannofossil preservation has been determined based on following criteria:

  • G—good, most specimens exhibit little or no secondary alteration, little or no evidence of dissolution and/or recrystallization;

  • M—moderate, all specimens are easily identifiable to the species level, but exhibit some etching and/or recrystallization from calcite dissolution and/or overgrowth;

  • P—poor, identification of species is hampered at the species and/or generic level but possible in some cases; specimens are severely etched or overgrown, primary morphological characteristics are largely destroyed and fragmentation occurs.

The overall preservation of the nannofossils is good, except for dredge site DL2. Here, the samples were collected at about 3.5–4.0 km depth, which is close to the carbonate compensation depth (CCD) (Perritt and Watkeys 2007). Consequently, the poor preservation of the calcareous nannofossils is not unexpected.

Taxonomy of the calcareous nannofossil species follows Perch-Nielsen (1985), Rio et al. (1990), Sato and Takayama (1992), Young and Bown (1997), Bown (2005) and Young et al. (2017). Identification of the genus Gephyrocapsa has been based on the measurements provided by Backman et al. (2012) for gephyrocapsid placoliths: ‘small sized’, < 4.0 µm; ‘medium sized’, 4.0–5.5 µm (Gephyrocapsa spp. ≥ 4 µm in the zonation of Backman et al. (2012)) and ‘large sized’, > 5.5 µm.

For biostratigraphic interpretations of the samples by calcareous nannofossils, standard zonations by Martini (1971) and Okada and Bukry (1980) and a new scheme of Backman et al. (2012) have been applied. The combination of schemes by Martini (1971) and Backman et al. (2012) provides a better age constrain for the studied samples.

The minimum growth rate of Mn encrustations has been calculated as a range, using the earliest and latest possible datings provided by calcareous nannofossils. The age of calcareous nannofossil zonal boundaries is given after Gradstein et al. (2020).

Results

Preliminary biostratigraphic results were published elsewhere (Ovechkina et al. 2009). Out of 30 examined samples, 21 samples contained representative assemblages of nannoplankton, which included 61 species, with nine species being reworked (Appendix); the other samples were barren. The preservation of nannofossils is generally good and moderate.

Dredge locality DL1

This locality, which is towards the southern end of the Mozambique Ridge (Fig. 1; Table 1), yielded some mafic igneous rocks encrusted by manganese but the samples largely comprised encrustations only. The surfaces of some samples were covered in small black nodules approximately 1 mm in diameter, typical of manganese growth on a surface exposed to ocean water. The encrustations exhibit very fine laminations, which are sometimes cross-cut by straight burrows that are about 5 mm in diameter and up to 7 cm deep. In two cases, the burrows terminate in a presumed Barnea sp. cast (Mollusca: Bivalvia: Pholadidae). There is the suggestion of 1–2 cm thick ‘bedding’ in some samples. Close examination reveals possible low angle cross-laminations within one ‘bed’ while others contain very subtle fining-upwards grading. The laminations in some samples display some gentle undulations that may be due to the draping effect, when manganese precipitate mimics the underlying surface. In other cases, however, it appears to involve soft-sediment deformation. In one instance, folding seems to have commenced as box-folds, with limbs of 2.5 cm long and straight crests connecting the limbs 4 cm long, after which the layering becomes overfolded and the upper limb thrusts out.

Calcareous nannofossils were found in layers in four samples (Fig. 4). In three of them (05–01, 05–02 and 05–03), the nannofossils were found in the basal layer of the manganese encrustation in contact with the underlying rock, while in 05–13 the nannofossils were recovered from the basal layer of the encrustation sample. The presence of Emiliania huxleyi in 05–01 indicates the uppermost Pleistocene–Holocene Zone NN21 or the upper part of CNPL11 Zone. The FO of E. huxleyi delineates the lower boundary of Zone NN21; the base of the Zone CNPL11 is marked by the LO of Pseudoemiliania lacunosa. The absence of Ps. lacunosa and the presence of scarce E. huxleyi indicate that this sample most probably refers to the upper part of CNPL11 Zone. The dominant species are Gephyrocapsa oceanica, Gephyrocapsa sp. (medium), Calcidiscus leptoporus, Helicosphaera carteri, Reticulofenestra sp. (small) and Umbilicosphaera sibogae.

Fig. 4
figure 4

Calcareous nannofossil distribution from the samples of the manganese deposits in the dredge localities DL1, DL2 and DL4

Full size image

The other three samples (05–02, 05–03 and 05–13) have moderately preserved nannofossil assemblages (Fig. 4). The dominant species are Calcidiscus leptoporus, Gephyrocapsa carribbeanica (medium), Gephyrocapsa sp. (small) and Reticulofenestra sp. (small). Gephyrocapsa oceanica (medium) and Ps. lacunosa are found in fewer numbers, and Calcidiscus macintyrei is rare. The presence of C. macintyrei and the absence of Discoaster brouweri confirm the assignment of these samples to the lower part of Zone NN19 (lower Pleistocene). According to the zonation of Backman et al. (2012), these samples may be referred to the Zone CNPL8, the base of which is fixed by the base of the Gephyrocapsa (≥ 4 µm) and the top of Gephyrocapsa (≥ 5.5 µm). In the lower part of Zone CNPL8 the LO of C. macintyrei is recorded (Backman et al. 2012). The ‘medium-sized’ Gephyrocapsa which including medium sized G. caribbeanica and G. oceanica are found in all three samples. The presence of group of medium gephyrocapsid and C. macintyrei allows referring these samples to the lower part of the Zone CNPL8 (basal Calabrian).

Dredge locality DL2

This dredge locality was at the southernmost extremity of the Mozambique Ridge and only yielded a total of seven samples, in each case involving a rock sample encrusted by manganese. In one specimen (05–27), calcareous nannofossils were recovered from the basal layer of the encrustation. The nannofossil preservation is poor, and the uncommon occurrence of Gephyrocapsa oceanica allows to place this sample only in a very broad interval of Zones NN19–NN21 or CNPL8–CNPL11, i.e. early Pleistocene–Holocene (Fig. 4). Due to the poor preservation of nannofossils, and therefore unreliable dating, this sample has been excluded from further analysis.

Dredge locality DL3

At this locality, dredging took place on the south-west flank and top of a small hill towards the centre of the southern section of the Mozambique Ridge (Fig. 1; Table 1). All of the rocks recovered were either encrusted in manganese or were fragments of encrustations and nodules. Some were burrowed, bioturbated and contained bivalve remains or their imprints, while two others contained bone fragments. Some samples had contorted laminations. Manganese nodules ranged from 5 to 10 cm diameter and when sectioned, commonly had a calcareous core which, in the larger nodules, was porous and decomposed. This locality provided 15 samples with calcareous nannofossils.

The richest nannofossil assemblages are found in the cores of three samples, all with good preservation (05–34, 05–36 and 05–37) (Fig. 5). The assemblages of these samples are characterised by dominant species Calcidiscus leptoporus, C. macintyrei and Coccolithus pelagicus. Dictyococcites sp., Discoaster asymmetricus, D. deflandrei, D. pentaradiatus, D. surculus, Helicosphaera carteri, Pseudoemiliania lacunosa, Reticulofenestra pseudoumbilicus, Syracosphaera pulchra and Umbilicosphaera sibogae are less common to rare. A rich assemblage of Miocene–Pliocene discoasters has been identified in these samples: Discoaster brouweri, D. pentaradiatus, D. surculus, D. variabilis, D. tamalis and D. asymmetricus. The species D. tamalis, D. asymmetricus, Reticulofenestra pseudoumbilicus and Pseudoemiliania lacunosa are stratigraphically most important and present in all three samples, except for D. tamalis which is absent from 05–37.

Fig. 5
figure 5

Calcareous nannofossil distribution from the samples of the manganese deposits in the dredge locality DL3

Full size image

In the sample 05–36, Discoaster triradiatus and Discoaster pansus have been found. Discoaster triradiatus ranges from the upper Miocene (NN9) through terminal Pliocene (top NN18) (Browning et al. 2017). The FO of Discoaster pansus is recorded within NN5 (lower/middle Miocene) and its LO is within NN15 (lower/upper Pliocene).

The FO of Discoaster asymmetricus delineates the base of Zone NN15 (early/late Pliocene) and disappears within Zone NN18 (upper Pliocene). Discoaster tamalis is distributed in a narrow interval. Its FO is recorded within Zone NN14 (lower Pliocene), and its LO is established near the top of Zone NN16 (upper Pliocene).

Reticulofenestra pseudoumbilicus is stratigraphically widely distributed. Its FO is recorded within the Zone NN4 (lower Miocene), and its LO is established at the top of Zone NN15 (lower/upper Pliocene). The FO of Pseudoemiliania lacunosa is recorded within Zone NN15 (lower/upper Pliocene), and its LO is recorded at the top of Zone NN19 (upper Pleistocene). These samples correspond to Martini’s (1971) Zone NN15 (lower/upper Pliocene) due to the presence of Reticulofenestra pseudoumbilicus and the absence of Amaurolithus tricorniculatus. The presence of D. asymmetricus and R. pseudoumbilicus allows referring all three samples to Zone CNPL3 (Backman et al. 2012). Pseudoemiliania lacunosa has been found in all three samples; the presence of this species is fixed in the upper part of CNPL3 Zone. Co-occurrence of R. pseudoumbilicus and Ps. lacunosa indicates that these samples fall into the upper part of Zone CNPL3 (upper Zanclean, lower Pliocene) by Backman et al. (2012).

Sample 05–43, an encrustation with a similarly rich nannofossil assemblage with good preservation, yielded the stratigraphically important yet rare species: D. asymmetricus, D. blackstockae, D. pentaradiatus, D. surculus, D. tamalis and Ps. lacunosa. Calcidiscus macintyrei, Helicosphaera sp., Coccolithus pelagicus and Reticulofenestra sp. (small) are abundant. Discoaster surculus appears at the base of Zone NN11 (upper Miocene) and disappears at the top of Zone NN16 (upper Pliocene/basalmost Pleistocene). Discoaster blackstockae has a wider stratigraphic range of distribution, its FO is fixed within Zone NN10 (upper Miocene) and the LO is within Zone NN18 (lower Pleistocene) (Browning et al. 2017).

The stratigraphic distribution of three other species has been discussed earlier. Using Martini’s (1971) zonation, sample 05–43 can be referred to the lower part of Zone NN16 (uppermost Zanclean/Piacenzian) due to the absence of R. pseudoumbilicus and the presence of D. surculus and D. tamalis, the latter disappearing in the upper part of the zone. The absence of R. pseudoumbilicus and the presence of D. tamalis place this sample to Zone CNPL4 (Backman et al. 2012), which is defined as the interval from the top of Reticulofenestra pseudoumbilicus and the top of D. tamalis.

Sample 05–52 is an encrustation with a well preserved but very impoverished nannofossil assemblage, which is represented only by Pseudoemiliania lacunosa, Calcidiscus leptoporus and Helicosphaera carteri. The last two species are distributed in a broad stratigraphic range. The FO of Ps. lacunosa is recorded within the lower Pliocene Zone NN15 (middle part of CNPL3), while its disappearance demarcates the upper boundary of the middle Pleistocene Zone NN19 (CNPL10). Reticulofenestra pseudoumbilicus, which LO marks the upper boundary of NN15 (CNPL3), has not been found in this sample. Sample 05–52 can conditionally be referred to a broad stratigraphic interval of zones NN16 (CNPL4)–NN19 (CNPL10), but is excluded from further analysis due to unreliability of dating.

Five samples (05–53, 05–49, 05–54, 05–48 and 05–42) yielded nannofossils from the lower Pleistocene. Sample 05–49 is an encrustation with moderate preservation, samples 05–42 and 05–54 are cores with moderate preservation and samples 05–53 and 05–48 are cores with good preservation of nannofossils. The dominant species are Calcidiscus leptoporus and Gephyrocapsa carribbeanica (medium). Calcidiscus macintyrei, Helicosphaera carteri, H. sp., Dictyococcites sp. and Reticulofenestra sp. (small) are recorded in low numbers.

The presence of Calcidiscus macintyrei and the absence of D. brouweri place these samples to the lower part of Zone NN19 (lower Pleistocene). The LO of D. brouweri defines the lower boundary of this zone and the LO of C. macintyrei is recorded within the lower part of NN19.

According to the zonation of Backman et al. (2012), these samples can be referred to Zone CNPL8, which lower boundary is fixed by the base of Gephyrocapsa (≥ 4 µm) and upper boundary by the top of Gephyrocapsa (≥ 5.5 µm). In the lower part of Zone CNPL8, the LO of C. macintyrei is recorded (Backman et al. 2012). The group of ‘medium sized’ Gephyrocapsa, which includes G. caribbeanica and G. oceanica, was recorded from almost all five samples. Gephyrocapsa caribbeanica was recorded in all samples, and G. oceanica was found in samples 05–54, 05–48, 05–49 and 05–42. The presence of the medium gephyrocapsid group and C. macintyrei allows referring these samples to the lower part of the Zone CNPL8.

The assemblage of samples 05–38 (an interlayer in an encrustation with good preservation) and 05–55 (a core with poor preservation) is characterised by dominant Calcidiscus leptoporus, Gephyrocapsa carribbeanica (medium), Pseudoemiliania lacunosa and Reticulofenestra sp. (small). Gephyrocapsa oceanica (medium), Gephyrocapsa sp. (large), Helicosphaera carteri and H. sellii are recorded in low quantities. Both samples are placed within the lower part of the lower Pleistocene Zone NN19 by Martini (1971), based on the presence of G. carribbeanica and H. sellii and the absence of Calcidiscus macintyrei. According to the zonation by Backman et al. (2012), these samples belong to Zone CNPL8, which lower boundary is fixed by the base of Gephyrocapsa (≥ 4 µm) and upper boundary by the top of Gephyrocapsa (≥ 5.5 µm). In the latter zonation, the FO of ‘large’ Gephyrocapsa is recorded in the middle part of the Zone CNPL8. Thus, the co-occurrence of medium forms of Gephyrocapsa (G. caribbeanica and G. oceanica) and large forms of Gephyrocapsa and the presence Helicosphaera sellii allow placing these samples in the upper part of CNPL8.

Nannofossil assemblages from two cores (05–40 with good preservation and 05–44 with moderate preservation) and an interlayer (05–35 with good preservation) are all similar. Calcidiscus leptoporus, Gephyrocapsa carribbeanica (medium), G. oceanica (medium), Gephyrocapsa sp. (small) and Reticulofenestra sp. (small) are the dominant species. All three samples may be assigned to the interval of the upper part of the lower-middle Pleistocene Zone NN19 according to Martini’s (1971) scheme. This is based on the presence of the dominant Gephyrocapsa carribbeanica (except in 05–35), the absence of C. macintyrei and Helicosphaera sellii and the presence of G. oceanica and Ps. lacunosa, which disappearance demarcates the upper boundary of this zone. According to the zonation of Backman et al. (2012), these samples can be referred to Zone CNPL10, which is defined as the interval between the top absence of Gephyrocapsa (≥ 4 µm) and the top of Ps. lacunosa. The presence abundant ‘medium’ and ‘small’ Gephyrocapsa spp. and common Gephyrocapsa parallela was recorded in all three samples, which is typical for this zone.

Dredge locality DL4

This site on the Jaguar seamount (Fig. 1; Table 1) was the shallowest dredge locality. It yielded samples of volcanic rocks encased in manganese encrustations. One sample (05-78B) contained a moderately preserved nannofossil assemblage that indicated Zone NN21 or the upper part of CNPL11 Zone (upper Pleistocene–Holocene) due to the presence of E. huxleyi (Fig. 4). The FO of E. huxleyi delineates the lower boundary of Zone NN21; the base of Zone CNPL11 is marked by the LO of Ps. lacunosa. The absence of Ps. lacunosa and the presence of rare E. huxleyi most probably place this sample into the upper part of CNPL11 Zone. The dominant species are Gephyrocapsa oceanica, Calcidiscus leptoporus, Gephyrocapsa sp. (medium), Helicosphaera carteri and Umbilicosphaera sibogae.

Discussion

It is important to emphasise that, with only a few exceptions, the identified nannofossils come from the basal layer of the manganese encrustation sample or from the core of nodules. Therefore, they are yielding critical information about ages of encrustations and Fe–Mn oxide precipitation rates. Firstly, the age of nannofossils is established in a particular sample. Secondly, this age marks the maximum age of the onset of conditions favourable to Fe–Mn oxide precipitation in the sample, so the thickness of the layer yields a minimum growth rate.

Overall, the samples fall into three age groups and six subgroups (Fig. 6):

  • Pliocene: (1) DL3: 05–34, 05–36, 05–37; (2) DL3: 05–43.

  • Pleistocene: (3) DL1: 05–02, 05–03, 05–13; DL3: 05–53, 05–49, 05–54; 05–48, 05–42; (4) DL3: 05–38, 05–55; (5) DL3: 05–35, 05–40, 05–44.

  • Pleistocene–Holocene: (6) DL1: 05–01; DL4: 05-78B.

Fig. 6
figure 6

Calcareous nannofossil biostratigraphic summary of the Pliocene–Holocene on the Mozambique Ridge and Mozambique Basin, showing the bi-modal (Pliocene and middle Pleistocene–Holocene) age distribution for the nannofossils of subgroups 1–2 vs 3–6 (see ‘Discussion’ section). Sample 05–52 can only be referred to a broad stratigraphic interval due to a very impoverished nannofossil assemblage. Pliocene and Pleistocene biozones by Backman et al. (2012)

Full size image

In the absence of isotope chronometry, only broad ages in millions of years can be estimated for each subgroup (Backman et al. 2012; Gradstein et al. 2020).

Pliocene

  1. (1)

    DL3: 05–34, 05–36, 05–37 (4.04–3.82 Ma)

    Although the presence of Pseudoemiliania lacunosa places all three samples in the upper part of Zone CNPL3, Backman et al. (2012) do not provide the exact age for the FO of Ps. lacunosa. Thus, we use the age of the lower boundary of Zone CNPL3 (4.04 Ma) in calculation of the minimum Fe–Mn oxide precipitation rate, bearing in mind that the actual minimum rate can be slightly higher.

  2. (2)

    DL3: 05–43 (3.82–2.76 Ma)

    Backman et al. (2012) discuss the age of the disappearance of Discoaster tamalis in the interval of 3.01–2.70 Ma, but accept 2.76 Ma in their scheme.

    Pleistocene

  3. (3)

    DL1: 05–02, 05–03, 05–13 (1.71–1.60 Ma)

    The range of the entire Zone CNPL8 is 1.71–1.25 Ma, but Backman et al. (2012) note that Calcidiscus macintyrei disappears in the lower part of the zone (1.60 Ma), just before the appearance of ‘large’ Gephyrocapsa spp. (> 5.5 μm) at 1.59 Ma. Since we have observed both C. macintyrei and ‘medium’ Gephyrocapsa spp. (≥ 4 μm) in all three samples, we use 1.60 Ma as the upper limit for calculating the minimum Fe–Mn oxide precipitation rate.

  4. (4)

    DL3: 05–53, 05–49, 05–54; 05–48, 05–42 (1.71–1.60 Ma)

    These samples have also yielded both C. macintyrei and ‘medium’ Gephyrocapsa spp. (≥ 4 μm), so we use 1.60 Ma as the upper limit for calculating the minimum Fe–Mn oxide precipitation rate, as discussed above.

  5. (5)

    DL3: 05–38, 05–55 (1.59–1.25 Ma)

    Contrary to subgroup 3, samples 05–38 and 05–55 are characterised by the presence of ‘large’ Gephyrocapsa spp. (> 5.5 μm) and the absence of C. macintyrei, so 1.59 Ma is used for calculating the minimum Fe–Mn oxide precipitation rate.

  6. (6)

    DL3: 05–35, 05–40, 05–44 (1.06–0.43 Ma)

    This age interval corresponds to the entire range of Zone CNPL10.

    Pleistocene–Holocene

  7. (7)

    DL1: 05–01; DL4: 05-78B (0.29–0.00 Ma)

Both samples contain Emiliania huxleyi, which FO is commonly accepted at 0.29 Ma (Backman et al. 2012; Gradstein et al. 2020).

There are several major routes of the Fe–Mn oxide precipitation in the ocean, i.e. hydrogenetic, hydrothermal, mixed hydrogenetic/hydrothermal and diagenetic (Koschinsky and Hein 2017). Each results in considerably differing rates of Fe–Mn oxide accumulation. The hydrogenetic process is the slowest, with only few millimetres per 1 Ma. Thus, for example, for hydrogenous Fe–Mn encrustations, Ingram et al. (1990) reported growth rates of 0.9–5.4 mm/Ma from the Pacific; Eisenhauer et al. (1992) estimated rates of 6.2 mm/Ma in the Quaternary of the Pacific. Usui and Terashima (1997) observed rates of 1.1–2.8 mm/Ma in the NW Pacific, and Frank et al. (2002) identified rates of 0.7–6.14 mm/Ma in different parts of the Atlantic, Southern and Indian oceans. In the SW Atlantic, Benites et al. (2020) calculated growth rates of 0.57–1.39, 2.07–4.31 and 1.48–4.38 mm/Ma using different Co-chronometer equations for non-phosphatized and phosphatized layers. Josso et al. (2020) mentioned rates of between 0.8 and 14 mm/Ma for the NE Atlantic. Most importantly for this study, Charles et al. (2020) noted rates of 1.3–3.6 mm/Ma for the Mozambique Channel, without, however, explaining how these values were obtained.

The hydrothermal route manifests in much higher, sometimes by several orders of magnitude, rates of Fe–Mn oxide precipitation. Ingram et al. (1990) provided rates of ≥ 33 mm/Ma for hydrothermally originated encrustations in the Pacific, and Usui and Terashima (1997) suggested rates of up to 1000 mm/Ma for young hydrothermal crusts in the NW Pacific. The mixed hydrogenetic/hydrothermal and diagenetic accumulation is characterised by intermediate growth rates, which, however, never reach extremely high values. Thus, Ingram et al. (1990) suggest Fe–Mn oxide precipitation rates of about 20 mm/Ma in the Pacific, whereas Glasby and Andrews (1978) mentioned 1.5, 3.5–36 and 10–20 mm/Ma while discussing diagenetic processes of accretion on the Hawaii Ridge.

Noteworthy are estimates of Fe–Mn accretion rates based on the calcareous nannofossil biostratigraphy. Cowen et al. (1993) calculated 0.8–2 mm/Ma for different intervals between the Palaeocene/Eocene boundary and lower Pliocene, applying calcareous nannofossil biostratigraphy to a Fe–Mn crust from Hawaii. The authors did not explicitly indicate the geochemical origin of the crust, which, however, can be identified as hydrogenous based on the inferred rate of its accretion.

Harada and Nishida (1976) analysed a Mn nodule dredged at the summit of the Makarov Guyot (depth 1378 m), NW Pacific, and identified calcareous nannofossils from outer and inner surfaces of its layers. The nodule had a 2-mm thick outer, 8-mm thick middle and 50-mm thick innermost layer. The calcareous nannofossil assemblage identified on the inner surface of the outmost layer allowed the authors to attribute the complex to NN19 Zone, for which they accepted the interval of 2–0.5 Ma, and to calculate the growth rate of 1.0–6.7 mm/Ma for this layer. Harada and Nishida identified Pseudoemiliania lacunosa and Calcidiscus macintyrei among other stratigraphically broadly distributed species in that assemblage. While the presence of the former (and the evident absence of discoasters) justifies the placement of the assemblage to NN19, the latter species disappears in the lower part of CNPL8 Zone at 1.60 Ma according to Backman et al. (2012). This does not affect much the calculation of the minimum growth rate of the outer layer, but has a far-reaching consequence for estimation of the accretion rate of the two inner layers. Harada and Nishida (1976) identified Gephyrocapsa oceanica in the innermost layer adjacent to the basaltic nucleus; this, coupled with the absence of Discoaster spp., allowed the authors to suggest 2 Ma as the maximum age for the onset of the Mn precipitation and to calculate its minimum rate for the two inner layers as 39 mm/Ma (there is apparently a typo, since the rate should be 29 mm/Ma). However, the new age constraint by C. macintyrei from above puts the formation of the two inner layers in the interval of 1.93–1.60 Ma and allows us to calculate a reliable Mn precipitation rate of 176 mm/Ma for the two inner layers. Harada and Nishida (1976) did note that the massive texture of the inner layer was indicative of a rapid (i.e. hydrothermally driven) growth. The West Pacific Seamount Province is generally characterised by slow (i.e. hydrogenetic) rates of the Fe–Mn oxide precipitation (Okamoto and Usui 2014), but there is evidence of hydrothermal processes of manganese accretion in the region (e.g. Usui and Nishimura 1992; Usui and Terashima 1997).

In terms of the rate of the Fe–Mn oxide precipitation, our material from the Mozambique Ridge falls into three categories. Two samples (05–34 and 05–43, from a core and an encrustation dated back to the Pliocene) show the slowest growth of 6.9–7.3 and 4.7–6.5 mm/Ma (Table 2), respectively; these rates are in agreement with the hydrogenetic processes, although they lie at the upper boundary of ever recorded values for this mode of accretion.

Table 2 Calculated growth rates (mm/Ma) based on the average age of nannofossil-dated basal layers in samples
Full size table

All other samples, save for 05–36, 05–37 and 05–53, demonstrate much higher rates of precipitation between 20.8 and 248.3 mm/Ma (Table 2). This is suggestive of a hydrothermal origin. The remaining three samples (05–36, 05–37 and 05–53), with their calculated growth rates of 9.9–14.4 mm/Ma, may well originate from a mixed hydrogenetic/hydrothermal process.

Overall, the calculated Fe–Mn oxide precipitation rates of 4.7 to 248.3 mm/Ma (Table 2) are indicative of the hydrothermal rather than hydrogenetic origin, contrary to the earlier obtained geochemical data (Perritt and Watkeys 2007). Further to this, manganese precipitation was apparently not continuous. Samples from dredge sites DL1 and DL3 contain older manganese laminations with burrows that do not continue upwards into the overlying younger laminations. Consequently, there must have been a period of time when the contact between these two series of laminations formed an exposed surface, thus allowing burrowing and, in places, bioturbation to take place prior to manganese precipitation being resumed.

The bi-modal (Pliocene and middle Pleistocene–Holocene) ages for the nannofossils (subgroups 1 and 2 vs 3–6, Fig. 6) probably reflect oceanographic changes in the SW Indian Ocean in response to major global events taking place elsewhere. The mid-Pliocene was the time when the Panama Isthmus finally closed (Lunt et al. 2008). This event caused a reorganisation of the currents in the Atlantic Ocean, resulting in an initial pulse of North Atlantic Deep Water (NADW) into this region of the SW Indian Ocean as attested by an increased isotopic signal around 3–4 Ma (Burton et al. 1997). Precipitation of manganese from the middle Pleistocene to Holocene was probably brought about by onset of the Northern Hemisphere glaciations that influenced the vector and dynamics of the NADW during the Quaternary (Rutberg et al. 2000; Frank et al. 2002; Bell et al. 2015; Bergh et al. 2021). However, more data on nannofossils from Mn nodules from a broader stratigraphic interval preceding the Pliocene and from other localities in SW Indian and SE Atlantic oceans are needed to test this hypothesis.

Reworked calcareous nannofossils are found at all dredge sites except DL2, which is close to the CCD and therefore not conducive to preservation of calcareous microfossils. Despite the small number of species (see Appendix), when taken overall they have been reworked from the Eocene to early Pliocene rocks. These ages correspond to the primarily Palaeocene/Eocene to early Pliocene sequences that have been identified in the Transkei Basin and Natal Valley (Niemi et al. 2000; Schlüter and Uenzelmann-Neben 2007). In addition, dredge sites DL1 and DL3 contain reworked Cretaceous nannofossils. This implies that in the Pliocene and Pleistocene, bottom currents must have eroded the Upper Cretaceous sedimentary sequences (Schlüter and Uenzelmann-Neben 2007).

With regards to the possibility of tectonic activity, the samples from dredge site DL1 contain features typical of soft-sediment deformation. As the Mozambique Ridge is an oceanic structure known to be associated with earthquakes (Craig et al. 2011), it seems likely that these features were formed due to instabilities caused by earthquakes during the Pleistocene. It is known that the axis of earthquake activity started to extend deep into the Natal Valley, following the oceanward propagation of the East African Rift System during the Neogene (Wiles et al. 2014).

Conclusions

Manganese precipitation on the Mozambique Ridge as nodules during the mid-Pliocene possibly occurred in response to changes in ocean currents that took place when the Panama Isthmus closed and the NADW penetrated the region. A return of the NADW to the region due to the onset of the Quaternary glaciations in the Northern Hemisphere might have caused the precipitation of further manganese deposits as encrustations on the Mozambique Ridge and the Jaguar seamount from the terminal Pleistocene to Holocene. Fluctuations in oceanographic conditions due to glacial and interglacial periods were possibly responsible for the ages of precipitation recorded in this study. On the basis of the calculated growth rates, which are 4.7–248.3 mm/Ma, the manganese deposits are considered mainly hydrothermal in origin.

There is an unfortunate void with respect to detailed geological, palaeontological and palaeoceanographic information in this part of the SW Indian Ocean. The results of this study reveal the potential of the Mozambique Ridge and environs with respect to filling this gap. The samples obtained by dredging indicate that in all probability the Mozambique Ridge has a complete Pleistocene and most likely Tertiary sedimentary record as well. The ridge is aligned N–S, extending from 35°S to 22°S, and ranges in depth from about 700 m to over 4 km. Consequently, it provides a large latitudinal range and depth variations that have preserved a wide range of oceanographic conditions during the Tertiary–Quaternary. Further detailed investigation will result in a better understanding of the palaeoenvironmental history of the region.