Goldschmidt Conference Abstracts 2006 The interplay of tectonics, erosion and topography across the Drakensberg Escarpment, South Africa; A fission track study M.J. RAAB 1 , R.W. BROWN 2 , A.J.W. GLEADOW 1,3 1 The University of Melbourne, School of Earth Sciences, Australia (mraab@unimelb.edu.au) 2 University of Glasgow, Department of Geographical and Earth Sciences, Scotland, UK (r.brown@ges.gla.ac.uk) 3 The University of Melbourne, School of Earth Sciences, Australia (gleadow@unimelb.edu.au) South Africa’s eastern margin hosts one of the most spectacular geo- morphological features in southern Africa: The Drakensberg Escarp- ment. It has the greatest relief of any passive margin and thus became the historic type locality for many studies of passive margin evolution or landscape evolution in general (e.g., King, 1967). Applying thermo- chronology and numerical surface process models (e.g., van der Beek et al., 2002) has greatly contributed to a far more detailed understanding of the evolution of this area. These studies have estimated a minimum of 4.5 km of denudation for the coastal zone, with amounts decreasing inland to approx. 2 km, since formation of the margin some 130 Myr ago. Here, we discuss three deep boreholes from the coastal platform (Springfontein) and the elevated continental interior (Vrede and Weltevrede). While Vrede and Weltevrede provide evidence of Early Jurassic mag- matic overprinting (Karoo), Springfontein shows the denudational histo- ry of a structurally separated inlier during rifting in the Early Cretaceous. The most striking difference in the thermal histories of these three boreholes is, beyond the varying denudational pattern for Vrede and Weltevrede, that the borehole closest to the continental edge (Springfon- tein) has very little memory of the most recent Mid Cretaceous event. This is in contrast to the interior boreholes Vrede and Weltevrede and the Swartberg borehole, which is just seaward of the escarpment, and outcrop samples published by Brown et al. (2002). We interpret the Springfontein data as evidence for differential movement of a crustal inlier along a com- plex sequence of faults during the Mid Cretaceous, where this crustal block was down thrown and thus little affected by denudation during the Mid Cretaceous event. This structural interpretation is supported by the occurrence of several major E–W oriented rift basins within the vicinity of Springfontein which are all bounded to the north by normal faults and down thrown blocks to the south. References King, L.C., 1967. The Morphology of the Earth. van der Beek, P., Summerfield, M.A., Braun, J., Brown, R.W., 2002. J. Geophys. Res. 107 (11), 1–18. doi:10.1016/j.gca.2006.06.946 Effect of silicate melt composition and fO 2 on metal-silicate partitioning of Si, Nb, Ta and V N. RAI, M.J. WALTER, C.J. HAWKESWORTH Department of Earth Sciences, University of Bristol, BS81RJ, UK (glxnr@bris.ac.uk) High pressure experiments indicate that Nb and V have very similar metal-silicate partition coefficients and are slightly siderophile at reducing conditions (Wade et al., 2001). It is postulated that the upper mantle deple- tion in V, and the depletion in Nb relative to Ta in mantle reservoirs, could be due to sequestering of V and Nb into the core (Wade et al., 2001). Fur- ther, the mantle abundance of V together with high P–T partitioning data help to constrain the conditions of core segregation (Wade and Wood, 2005). Nb, V and Ta can exist in variable and high oxidation states. As such, their partitioning behaviour is dependent on oxygen fugacity, and by way of analogy with other high-valence cations (Mo 6+ ,W 4+,6+ ), partitioning may also be sensitive to silicate melt composition (Walter et al., 1995). Here, we report new experimental results on the effects of silicate melt composi- tion and oxygen fugacity on the partitioning behaviour of these elements along with Si. Experiments were made at 2 GPa and 2000 K and over a range of relative oxygen fugacities from IW-3 to IW-6. Silicate melts range from basaltic to peridotitic in composition. The melt compositional parameter, NBO/T, is used as a proxy for silicate melt composition, and has values from 0.7 to 4. Isobaric, isothermal data for each element were regressed using the equation: ln D met/sil = a + bfO 2 + c(nbo/T). Regres- sions show that at the experimental conditions both Nb and V exist in 3+ valence states, Si exists in 4+ state whereas Ta likely exists in a 5+ valence state. In contrast, higher pressure (25 GPa) partitioning data (Wade et al., 2001) show a 5+ valence state for Nb, which may imply a possible change in valence as the absolute fO 2 of the IW buffer becomes more oxidizing with pressure. Documenting such changes in valence is critical for parameteriza- tions that rely on exchange partition coefficients (Wade and Wood, 2005). Regressions indicate that melt composition has little effect on the partition- ing behaviour of V, Nb and Ta although Si shows strong compositional dependency. For trivalent Nb and V this is generally consistent with find- ings that melt composition has a small effect on partitioning of lower valence elements such as Ni 2+ and Co 2+ (Jaeger and Drake, 2000). A higher valence state of 5+ for Nb at higher pressures implies a greater effect of melt composition on partitioning. However, we find that Ta 5+ shows virtually no compositional dependence, in marked contrast to Si 4+ , and also that Ta 5+ becomes siderophile at less reduced conditions than Si 4+ . References Jaeger, Drake, 2000. Geochim. Cosmochim. Acta 64, 3887–3895. Wade et al., 2001. Nature 409, 75–78. Wade, Wood, 2005. Earth Planet. Sci. Lett. 236, 78–95. Walter et al., 1995. Science 270, 1186–1189. doi:10.1016/j.gca.2006.06.947 www.elsevier.com/locate/gca Geochimica et Cosmochimica Acta 70 (2006) A514–A551