Geology of the Mwadui kimberlite, Shinyanga district, Tanzania

The Mwadui pipe represents the largest diamondiferous kimberlite ever mined and is an almost perfectly preserved example of a kimberlitic crater in-fill, albeit without the tuff ring.

The geology of Mwadui can be subdivided into five geological units, viz. the primary pyroclastic kimberlite (PK), re-sedimented volcaniclastic kimberlite deposits (RVK), granite breccias (subdivided into two units), the turbidite deposits, and the yellow shales listed in approximate order of formation. The PK can be further subdivided into two units—lithic-rich ash and lapilli tuffs which dominate the succession, and lithic-poor juvenile-rich ash and lapilli tuffs. The lower crater is well bedded down to at least 684 m from present surface (extent of current drill data). The bedding is defined by the presence of juvenile-rich lapilli tuffs vs. lithic-rich lapilli tuffs, and the systematic variation in granite content and clast size within much of the lithic-rich lapilli tuffs. Four distinct types of bedding have been identified in the pyroclastic deposits. Diffuse zones characterised by increased granite abundance and size, and upward-fining units, represent the dominant types throughout the deposit.

Lateral heterogeneity was observed, in addition to the vertical changes, suggesting that the eruption was quite heterogeneous, or that more than one vent may have been present. The continuous nature of the bedding in the pyroclastic material and the lack of ash-partings suggest deposition from a high concentration (ejecta), sustained eruption column at times, e.g. the massive, very diffusely stratified deposits. The paucity of tractional bed forms suggest near vertical particle trajectories, i.e. a clear air-fall component, but the poorly sorted, matrix-supported nature of the deposits suggest that pyroclastic flow and/or surge processes may also have been active during the eruption.

Available diamond sampling data were examined and correlated with the geology. Data derive from the old 120 (37 m), 200 (61 m), 300 (92 m) and 1200 ft (366 m) levels, pits sunk during historical mining operations, drill logs, as well as more recent bench mapping. Correlating macro-diamond sample data and geology shows a clear relationship between diamond grade and lithology. Localised enrichment and dilution of the primary diamond grade has taken place in the upper reworked volcaniclastic deposits due to post-eruptive sedimentary in-fill processes. Clear distinction can be drawn between upper (re-sedimented) and lower (pyroclastic) crater deposits at Mwadui, both from a geological and diamond grade perspective.

Finally, an emplacement model for the Mwadui kimberlite is proposed. Geological evidence suggests that little or no sedimentary cover existed at the time of emplacement. The nature of the bedding within the pyroclastic deposits and the continuity of the bedding in the vertical dimension suggest that the eruption was continuous, but that the eruption column may have been heterogeneous, both petrologically as well as geometrically. Volcanic activity appears to have ceased thereafter and the crater was gradually filled with granite debris from the unstable crater walls and re-sedimented volcaniclastic material derived from the tuff ring.

The Mwadui kimberlite exhibits marked similarities compared to the Orapa kimberlite in Botswana.