RESULTS AND DISCUSSION

Two profiles for Be isotope analyses with a resolution of 2 mm/sample were sampled across the Fe-rich and the Mn-rich laminae of the seafloor-surface crust in Core 194-1196B-1R. In addition, Be isotopes were measured on two other recent hardground samples and on the deep subsurface hardground crust (Table T2). Unlike many other studies on hydrogenetic ferromanganese crusts (cf. Frank, 2002, and references therein), no downward decrease of the ¹⁰Be concentrations is observed (Fig. F6). This is also not the case after normalizing to stable ⁹Be to account for possible dilution effects. The ¹⁰Be/⁹Be ratios vary between 1 x 10^–7 and 1.5 x 10^–7 with an average of 1.16 x 10^–7, which is within errors identical to the range of recent growth surfaces of hydrogenetic Pacific deepwater ferromanganese crusts (von Blanckenburg et al., 1996). Similar ratios have also been obtained for the two other surface hardground crusts. This clearly documents that the hardground crusts are hydrogenetic precipitates that have incorporated the isotopic composition along ambient deep waters during growth. The range and the similarity of the ¹⁰Be/⁹Be ratios to other present-day surface values of deep-sea ferromanganese crusts also limits the growth period to within the past few 100 k.y. Growth rates must therefore have been significantly higher than those of deep-sea hydrogenetic ferromanganese crusts (0.5–15 mm/m.y.). We would like to point out that there is the possibility that the crusts had an initial ¹⁰Be/⁹Be ratio of up to 2.5 x 10^–7, similar to water column measurements from ~300 m water depth at other locations in the Pacific Ocean (Kusakabe et al., 1987). The crusts would accordingly have stopped growing for the past ~2 m.y.; however, we consider this very unlikely in view of the favorable conditions for the growth of ferromanganese crusts on the SMP.

In view of the narrowly constrained present-day initial ¹⁰Be/⁹Be ratio, the age of the subsurface hardground crusts from 117 m depth in the core of Site 1194 on the NMP can be determined using beryllium isotopes. Under the assumption of a constant flux of ¹⁰Be into the ocean, a constant initial ¹⁰Be/⁹Be ratio at the surface of the crusts, and the identical ¹⁰Be/⁹Be ratios (0.023 ± 0.005 x 10^–7) of the two subsurface crust samples (from Section 194-1194A-14X-1), we calculate an age of 8.65 ± 0.50 Ma. This is in very good accordance with the biostratigraphic and seismic stratigraphic age of the hiatus (Fig. F7) during which the crust grew (7.7–11.8 Ma) (Isern, Anselmetti, Blum, et al., 2002). This hardground crust and the surface crust samples measured so far clearly do not represent the entire periods of the inferred hiatuses but only relatively brief periods within the hiatuses. The reason for these interruptions of growth are not clear, but these crusts can obviously not contribute to determine the durations of the hiatuses.

The Nd isotope composition of all samples measured ranges between –2.7 and 0.15 _Nd. The two profiles (Fig. F8) show internally consistent trends from identical surface _Nd values around –2.5 toward more radiogenic _Nd values of up to 0 at the base of the hardground. No other Nd isotope data were available from the water column or from shallow ferromanganese crusts from the Tasman Sea or Coral Sea with which our data can be directly compared. They are, however, at the higher end but within the range of upper water column Nd isotope compositions measured at other locations in the central and South Pacific Ocean (Piepgras and Jacobsen, 1988; Lacan and Jeandel, 2001). They therefore confirm the hydrogenetic, seawater-derived origin of the trace metals in these hardground crusts. The surface values of the two profiles are also within error identical to the surface values of a hydrogenetic crust from 1500 m water depth on the Lord Howe Rise farther south in the Tasman Basin (van de Flierdt et al., 2004). The pronounced increase of 2.5 _Nd units over the short growth period represented by the hardground indicates a significant change in the composition of the ambient deep water at the location of the SMP. This variation may be related to a glacial–interglacial change in the prevailing surface current on the SMP or, less likely, to changes in the isotopic composition of the current itself, potentially caused by changes in weathering intensity in its source region. With the currently available data in the area, this cannot be further constrained. As discussed below, it may also be possible that the pattern is a consequence of diagenetic remobilization.

The major element composition of the hardground crusts was measured on the same aliquots used for Nd and Be isotope analyses in order to evaluate their hydrogenetic origin. Concentrations of the minor and trace metals (Co, Cu, Nd, and Pb) are about a factor of 50 lower than normally observed in deepwater ferromanganese crusts, whereas Be concentrations are only about a factor of 10 lower. This is consistent with the inferred high growth rates, but it may also be that metal supplies did not only originate from the water column. Mn/Fe ratios in all samples except the Mn-rich profile are between 0.001 and 0.2, which is far below typical hydrogenetic ratios, whereas ratios for the Mn-rich layer are between 1.3 and 5.5. This is comparable to hydrogenetic ratios (Frank et al., 1999), although a value of 5 already points to diagenetic addition of Mn. The origin of ferromanganese crusts can be further evaluated by plotting the abundances of Fe, Mn, Co, Ni, and Cu (Table T3) in a ternary diagram (Bonatti et al., 1972) (Fig. F9). Although this tool to distinguish between hydrogenetic, hydrothermal, and diagenetic origin can strictly only be applied to mixed-type ferromanganese nodules, it can provide first-order information. Only data for the Mn-rich profile plot near the field of hydrogenetic growth, although some of them already plot to the right of it, which is consistent with a diagenetic addition of Mn. All other samples plot in the lower left corner of the diagram, reflecting very high relative Fe concentrations. Given that hydrothermal additions at this location can be excluded, we argue that trace metals and Mn were mobilized out of the hardgrounds to different extents. This is most clearly visible in the subsurface hardground, which has essentially lost all Mn, Co, Ni, and Cu, whereas more particle-reactive metals, such as Be, Nd, and Pb, have almost completely remained in place. The Mn-rich hardground has the highest trace metal concentrations, suggesting that it is least altered by diagenesis. Its largely very low Be concentrations, below the detection limit, and the evidence from Figure F10, however, suggest that some kind of diagenesis occurred here as well.

Supporting evidence for diagenetic alteration comes from a comparison of elemental maps (Fig. F10) with thin section images (Fig. F4) that show that the reddish brown laminae are iron rich and the black opaque sections are rich in manganese. A very thin, sharp fracture separates the compositional change from Fe rich to Mn rich (Figs. F4, F10). Most of the primary growth structures can be traced across this change. This confirms the postdepositional diagenetic alteration of the Fe-rich layers by Mn loss but does not exclude that the Mn-rich layers were also diagenetically altered.

Despite the fact that Be and Nd isotope data are consistent with seawater origin, it cannot be excluded that Be and Nd have also been redistributed during the process of diagenesis, most likely on very small spatial scales, which may explain the missing decrease of the ¹⁰Be/⁹Be ratios with depth in the two profiles and may also be the cause for the trend in the two Nd isotope profiles.

Until now the elevated topography of the SMP, combined with the strong bottom currents, prevented particle sedimentation on top of the SMP, resulting in absent sedimentation since the late Miocene–early Pliocene. This environment allowed ferromanganese crusts to grow, as shown on seafloor photographs (Fig. F11) taken during the site survey cruise onboard the Franklin (AGSO 209; CSIRO FR 03/99) (Heck et al., 1999).

Unlike the SMP, carbonate production on the NMP never reinitiated after subaerial exposure of the late middle Miocene sea level fall. During this lowstand, a small carbonate platform grew on the upper slope of the exposed NMP. This lowstand platform eventually drowned, as did the entire NMP, when sea level rose again and flooded these platforms. Because strong currents swept the seafloor, sedimentation started much later and the slope platform became finally buried under drift deposits in the late Pliocene. Thus, similar to the SMP, the NMP has been an ideal location for formation of ferromanganese crusts because sedimentation was prevented for an extended period of time.