La Isla de Gorgona, Colombia: A petrological enigma?

A wide range of intrusive (wehrlite, dunite, gabbro and olivine gabbro) and extrusive (komatiites picrites and basalts) igneous rocks are found on the small pacific island of Gorgona. The island is best known for its ~90 Ma spinifex-textured komatiites: the only true Phanerozoic komatiites yet discovered. Early work led to suggestions that the rocks of the island formed at a mid-ocean ridge, however more recent research supports an origin as part of a hot mantle plume-derived oceanic plateau. One of the main lines of evidence for this origin stems from the inferred high mantle source temperatures required to form the high-MgO (N15 wt.%) komatiites and picrites. Another remarkable feature of the island, considering its small size (8�2.5 km), is the degree of chemical and radiogenic isotopic heterogeneity shown by the rocks. This heterogeneity requires a mantle source region with at least three isotopically distinctive source regions (two depleted and one enriched). Although these mantle source regions appear to be derived in significant part from recycled oceanic crust and lithosphere, enrichments in 187Os, 186Os and 3He in Gorgona lavas and intrusive rocks, suggest some degree of transfer of material from the outer core to the plume source region at DW. Modelling reveals that the komatiites probably formed by dynamic melting at an average pressure of 3–4 GPa leaving residual harzburgite. Trace element depletion in Gorgona ultramafic rocks appears to be the result of earlier, deeper melting which produced high-MgO trace element-enriched magmas. The discovery of a trace-element enriched picrite on the island has confirmed this model. Gorgona accreted onto the palaeocontinental margin of northwestern South America in the Eocene and palaeomagnetic work reveals that it was formed at ~26 8S. It has been proposed that Gorgona is a part of the Caribbean–Colombian Oceanic Plateau (CCOP), however, the CCOP accreted in the Late Cretaceous and was derived from a more equatorial palaeolatitude. This evidence, and differing geochemical signatures, strongly suggests that Gorgona and probably other coastal oceanic plateau sequences in Colombia and Ecuador, belong to a completely different oceanic plateau to the CCOP