Deep chemical structure of the Southern African mantle from kimberlite megacrysts

Xenoliths of the Cr-poor megacryst suite were formed by the crystallization of magmas in the deeper regions of subcontinental lithospheric mantle of southern Africa, and may be used to probe its compositional structure. Some 2300 pyrope garnet megacrysts from 61 kimberlites in southern Africa define four broad arrays on the basis of Cr (sub 2) O (sub 3) -Mg# systematics, referred to as H-, M-, L- and diffuse trends. M-trend megacrysts, representing typical Cr-poor megacryst compositions, occur over much of the southwestern Kaapvaal Craton, continuing into the Proterozoic Namaqua Province to the west. In contrast, all suites examined from the Zimbabwe Craton, as well as the Premier kimberlite cluster on the Kaapvaal craton, are dominated by H-trend megacrysts. The boundary between the two domains corresponds broadly to the locus of a region of slower seismic velocity in the uppermost mantle. L-trend megacrysts with very low Cr contents and diffuse, steeply inclined trends occur around the periphery of the continent and in zones of previous tectonic activity. Variations in Cr (sub 2) O (sub 3) -Mg# trends are attributed to different megacryst magma sources and variations in the chemical character and depth of lithospheric mantle into which megacryst magmas are emplaced. In addition to the classic fractional crystallization processes, partial melting and magma-wall rock interaction are recognized as potentially important influences on megacryst composition. However, uncertainties in the quantitative contributions of each process render speculative the broader-scale geodynamic interpretations of the regional distribution patterns. Megacrysts in orangeites appear to reflect an origin and evolution in depleted lithospheric mantle. High-Cr megacrysts in kimberlites of the northern cratonic areas bear the influence of lithospheric mantle and their distribution its accessibility to infiltration. Commonly coherent M-trends of the southern Kaapvaal could be interpreted to indicate that the lithosphere in this region was locally less easily infiltrated, that complete conversion of depleted lithosphere to fertile mantle resembling the megacryst magma source by high magmatic flux occurred, or that shallow emplacement of megacryst-magma source mantle into the lower regions of the lithosphere preceded megacryst emplacement, effectively shielding the magmas from overlying lithosphere. The steeply trending, diffuse patterns are thought to result from some combination of changes in phase proportions in response to shallower intrusion of megacryst magmas, mixed source regions that include involvement of low-Cr, possibly oceanic mantle, and the enhanced degree of lithospheric interaction afforded by tectonized and more fertile lithosphere of Proterozoic orogenic belts and intra-continental rifts. Present megacryst sampling in space and time allows for the lateral differences to be long-lived features of the subcontinental mantle, and permit a possible origin for megacryst magmas with OIB affinities within the continental mantle root or boundary layer. These speculations may be resolved with an increased sampling density and understanding of megacryst petrogenesis.