South Africa - Inkaba.org

New insights on the role of a mantle plume in South Atlantic
breakup
R.B. Trumbull, J.K. Keiding, I.V. Veksler
GFZ German Research Centre for Geosciences, Telegrafenberg, 14473 Potsdam, Germany, bobby@gfzpotsdam.de,
jakob@gfz-potsdam.de, ilya.veksler@gfz-potsdam.de
ABSTRACT
There is controversy surrounding the existence and possible role of a deep-seated mantle plume associated with the
Paraná-Etendeka large igneous province. Observational evidence is limited to the erupted and intruded magmatic
rocks of mantle derivation: flood basalts, central complexes and mafic dyke swarms. Key information needed is
pressure-temperature conditions of mantle melting (how hot? how deep?) and composition of primary magmas.
Estimating primary magma compositions from geochemical models based on mineral and rock data is subject to
large uncertainties because rocks are often poor proxies for melt compositions. Preferable is direct analysis of melt
inclusions trapped within early-formed crystals and isolated from the rest of the magma during ascent and eruption.
Here we present data from melt inclusions and host olivines from the Henties Bay-Outjo (HOD) dyke swarm and
Tafelkop lavas in the Etendeka province of NW Namibia. The prime target for this study are dykes of high- MgO
(picritic) compositions, which are thought to be plume-derived, and are therefore of considerable significance for
understanding mantle dynamics and magmatism in this province.
Preliminary microprobe results show that olivines with 85 mol.% forsterite or higher are common in the HOD
dykes, while olivine in the lavas analyzed are less magnesian (Fo78 or lower). We reconfirmed extremely Mg-rich
olivines (up to Fo93.4) that were reported by Thompson and Gibson (2000) from SW of the Brandberg Complex,
and we have discovered more localities with exceptionally Mg-rich olivines (up to Fo93.6), so the distribution of
such high-Mg olivines appears to be larger than previously thought. Most olivines with Fo >85 are too Mg-rich to
be in equilibrium with the whole-rock composition, indicating that the grains are entrained xenocrysts which
crystallized from an earlier, much more magnesian melts.
Most important is the finding that melt inclusions are abundant in the xenocrysts with Fo > 90, so we have access
to melt compositions from which these enigmatic olivines formed. The inclusions range in size from 10-70
microns, are typically round in shape and are partly or completely crystallized, with Al-rich clinopyroxene , Al-rich
orthopyroxene, and spinel (chromite). Trapped melts are tholeiitic picritic basalts with maximum 17.5 wt. % MgO.
Olivine-melt themometry using these compositions yields estimates of mantle potential temperature of 1520°, in
contrast to the 1700°C proposed by Thompson et al. (2000) based on whole-rock models and more in line with
petrologic estimates for plume temperatures in Hawaii and Iceland. Further analysis of trace element compositions
from these inclusions will give potentially important information on magma type and mantle source composition.
Reference: Thompson, R. N. & Gibson, S. A. (2000). Transient high temperatures in mantle plume heads inferred
from magnesian olivines in Phanerozoic picrites. Nature 407, 502-506
KEYWORDS: Etendeka Province, mantle temperature, olivine, melt inclusions
80