Bull Volcanol (2009) 71:95–112 DOI 10.1007/s00445-008-0211-4
RESEARCH ARTICLE
Geology of a complex kimberlite pipe (K2 pipe, Venetia Mine, South Africa): insights into conduit processes during explosive ultrabasic eruptions R. J. Brown & M. Tait & M. Field & R. S. J. Sparks
Received: 19 July 2006 / Accepted: 4 March 2008 / Published online: 8 April 2008 # Springer-Verlag 2008
Abstract K2 is a steep-sided kimberlite pipe with a complex internal geology. Geological mapping, logging of drillcore and petrographic studies indicate that it comprises layered breccias and pyroclastic rocks of various grain sizes, lithic contents and internal structures. The pipe comprises two geologically distinct parts: K2 West is a layered sequence of juvenile- and lithic-rich breccias, which dip 20–45° inwards, and K2 East consists of a steep-sided pipe-like body filled with massive volcaniclastic kimberlite nested within the K2 pipe. The layered sequence in K2 West is present to > 900 m below present surface and is interpreted as a sequence of pyroclastic rocks generated by explosive eruptions and mass-wasting breccias generated by rock fall and sector collapse of the pipe walls: both processes occurred in tandem during the infill of the pipe. Several breccia lobes extend across the pipe and are truncated by the steep contact with K2 East. Dense pyroclastic rocks within the layered sequence are interpreted as welded deposits. K2 East represents a conduit that was blasted through the layered breccia sequence at a late stage in the eruption. This phase may have involved fluidisation of trapped pyroclasts, with loss of fine particles
Editorial responsibility: J McPhie R. J. Brown (*) : M. Field : R. S. J. Sparks Department of Earth Sciences, Wills Memorial Building, University of Bristol, Queen’s Road, Bristol BS8 1RJ, UK e-mail:
[email protected] M. Tait De Beers Group Services, Cornerstone Office, Southdale 2135, Johannesburg, South Africa
and comminution of coarse clasts. We conclude that the K2 kimberlite pipe was emplaced in several distinct stages that consisted of an initial explosive enlargement, followed by alternating phases of accumulation and ejection. Keywords Kimberlite . Volcanic conduit . Pyroclastic . Explosive eruption
Introduction Conduit deposits derived from the explosive eruption of kimberlite magma occur throughout the geological record, but no such eruptions have been witnessed and we know comparatively little about the near-surface eruptive behaviour of such magmas. Kimberlite magmas are characterised by low viscosities (0.1–10 Pa s), by SiO2 contents of <<35 wt.% and by high volatile contents (CO2 and H2O; Mitchell 1986; Sparks et al. 2006). Kimberlite pipes share many characteristics (e.g., steep-sided funnel shapes, marginal breccias, pseudo-concentric or nested lithofacies distributions and steeply-dipping fabrics) with other volcanic conduits, such as ignimbrite vents and diatremes (e.g., Lorenz 1975, 1986; Ekren and Byers 1976; Reedman et al. 1987; White 1991; Stoppa 1996; Kano et al. 1997; Kurszlaukis and Lorenz 1997; White and McClintock 2001; Junquiera-Brod et al. 2005). These similarities imply a general overlap in conduit processes between different styles of explosive eruption. Little consensus has yet been reached on the emplacement mechanisms of kimberlite pipes. Some authors consider that volatile exsolution is the primary driving mechanism for kimberlite eruptions (e.g., Dawson 1967, 1971, 1980; Clement and Reid 1989; Field and Scott-Smith 1999; Skinner and Marsh 2004; Sparks et al. 2006; Wilson
96
and Head 2007). Others have proposed, on the grounds of striking similarities with maar–diatreme volcanoes, that interaction with ground water is essential for the formation of kimberlite pipes (e.g., Lorenz 1975, 1986; Kurszlaukis et al. 1998; Lorenz and Kurszlaukis 2007). Some models of kimberlite volcanism have involved the formation of a subsurface kimberlite pipe by intrusive brecciation followed by a short-lived volatile-driven explosion (e.g., Clement and Reid 1989; Field and Scott-Smith 1999; Skinner and Marsh 2004). This explosion is thought to be directly responsible for the formation of the pipe-filling deposits. Sparks et al. (2006) argued that kimberlite eruptions are volatile-driven and that they may comprise numerous phases of excavation and infill. The duration of kimberlite eruptions is considered to range from an hour (e.g., Wilson and Head 2007) through to several days-to-months (Lorenz 1975; Kurszlaukis and Barnett 2003; Sparks et al. 2006). Much of the controversy stems from an incomplete understanding of the properties of kimberlite magma. An emerging debate concerns the origin of coherent dense rocks within kimberlite pipes that have been termed either hypabyssal or magmatic kimberlite. Such rocks are characteristic of, but not confined to, the root zones of kimberlite pipes (Hawthorne 1975; Clement and Reid 1989; Field and Scott Smith 1999). Facies transitional between coherent dense kimberlite and volcaniclastic rocks have been observed in several kimberlite pipes (e.g., Hetman et al. 2004; Skinner and Marsh 2004). The dense rocks commonly exhibit ‘segregationary textures’, sub-millimetresized irregular-shaped domains of serpentine and calcite in between irregular and globular masses of groundmass. This texture has been attributed to liquid immiscibility (Mitchell, 1975) and the in-situ exsolution of volatiles during the fragmentation of kimberlite magma into pyroclasts (e.g., Skinner and Marsh 2004; Hetman et al. 2004). Sparks et al. (2006) argued that these coherent dense rocks have a pyroclastic origin and form by welding processes, in particular the sintering together of hot juvenile grains. Here we describe and interpret a well-exposed complex kimberlite pipe (K2) of the Venetia Kimberlite cluster, South Africa. The pipe uniquely preserves a layered sequence of pyroclastic and mass-flow breccias to a considerable depth in the pipe (∼1 km below original ground surface) that we interpret as the deposits of episodic explosive eruptions and gravitational collapses of the pipe walls. Some rock types in K2 have textures that we attribute to welding processes. The evidence suggests that the K2 eruption was a protracted explosive event that underwent several distinct phases of excavation and infill. The preserved pipe-filling deposits are interpreted as predominantly waning-stage products. The data support the theoretical model of kimberlite volcanism presented in Sparks et al. (2006).
Bull Volcanol (2009) 71:95–112
Geological setting The Venetia kimberlite cluster is located in the Limpopo mobile belt, an ancient collision zone between the Zimbabwe and Kaapvaal cratons. The cluster comprises 14 known pipes and dykes outcropping over ∼4 km2 (Fig. 1a). Only the two largest (K1 and K2) are currently mined. The kimberlite pipes were intruded ∼519 Ma (Phillips et al. 1999) into a complex, north-verging shallowly eastward-plunging synclinal fold in Proterozoic gneiss, schist, amphibolite and metasedimentary rocks (Barton et al. 2003). The limbs of the fold dip 40° NE around the southern side of K2. A 50m-thick Proterozoic dolerite sill that pre-dates the kimberlites occurs at ∼450 m below surface level. Shale, rhyolite and basalt clasts occur as inclusions in the pipe-filling deposits and indicate that Early Proterozoic Waterberg Group rocks covered the basement at the time of kimberlite emplacement but have since been eroded away. About 300– 600 m of post-emplacement erosion has been estimated for the Venetia area (Kurszlaukis and Barnett 2003). Preexisting structures in the basement had a major control on the position and shape of the Venetia kimberlite pipes: many appear to be elongated along faults and shear zones in the metamorphic basement (Kurszlaukis and Barnett 2003). Pipe growth during eruptions is inferred to have been influenced by joint-controlled tensional failure of the pipe walls along pre-existing structural weaknesses and by regional stress orientations, accounting for the elongate and straight rectilinear shapes of the pipes (Fig. 1a; e.g., Barnett 2008). Previous work Previous workers divided K2 into an eastern and western half (Seggie et al. 1998). It was considered that the western half comprised the products of up to three intrusive (hypabyssal) phases and that the western pipe margin was extensively brecciated and infiltrated by kimberlite. The eastern half comprised a deposit termed ‘tuffisitic kimberlite breccia’ (of Clement 1982; Clement and Skinner 1985), which is essentially a homogeneous pyroclastic deposit Fig. 1 a Map of the Venetia kimberlite cluster showing the seven largest kimberlite bodies. Modified from Seggie et al. (1998). b Surface geology of K2 pipe as reconstructed from mine bench maps. Inset stereographs show poles of lithic clast a–b planes at the margins of the pipes. See text for lithofacies descriptions. c Plans of the mine at 120 m depth increments, as reconstructed from drillcore intersects, to illustrate the inferred position and extent of the NW extension, K2 East and the inferred changing shape of the pipe with depth. d Crosssection of the K2 pipe interpreted from drillcore data and illustrating the inferred geology at depth. The depth of the viewing field for the illustrated drillcores is 25 m either side of the cross-section plane. Subsurface data was constructed using CAD software (Gemcom GEMS™). Inset map of South Africa. (VM=Venetia Mine). All observations taken from 60–108 m bps
Bull Volcanol (2009) 71:95–112
a
c
d
97
b
98
(massive volcaniclastic kimberlite in this work; after Sparks et al. 2006). Kurszlaukis and Barnett (2003) documented volcanological and structural aspects of the Venetia kimberlite cluster. They reported a sequence of layered breccias in K2 west with variable country rock contents and interpreted them as talus deposits derived from subsidence of the pipe walls. They concluded that the geology did not fit a simple intrusive origin (sensu Clement 1982; Clement and Reid 1989) and instead suggested that the pipe was emplaced over several phases. They proposed that phreatomagmatic explosivity could have been important during the Venetia eruptions.
Bull Volcanol (2009) 71:95–112
over five benches (60–108 m bps) were mapped on a 1:300 scale over 2 years. This accounts for >1,200 m of horizontal exposure. Information on the subsurface geology was obtained from logging of drillcore and from examining existing drillcore data: K2 has been pierced by >80 boreholes over the past 30 years, accounting for >22 km of core. Over 250 thin sections taken from surface outcrops and drillcore have been examined. The geology was modelled in 3D using computer software (Gemcom GEMS™). A lithofacies-based approach has been adopted with lithofacies distinguished on the basis of grain size, lithic clast abundance and composition.
Terminology K2 kimberlite pipe The term macrocryst is used for anhedral crystals, typically >0.6 mm in diameter inferred to be mantle-derived xenocrysts; phenocryst for euhedral–subhedral crystals, <0.6 mm in diameter, inferred to have crystallised from kimberlite melt during magma ascent; xenocryst for crystals liberated from the country rock, and free crystals for crystals inferred to have been explosively liberated from the magma. Alteration is pervasive in K2 and all olivine crystals have been replaced by serpentine, talc and carbonate. For brevity and unless stated otherwise, the term olivine refers to serpentinised olivine pseudomorphs; unaltered olivine crystals are rare. The term pelletal clast refers to subspherical ash or lapilli-sized grains (typically olivine crystals) that are coated in a thin (<50 µm) smooth-surfaced shell of altered fine-grained kimberlite groundmass (e.g., “pelletal lapilli” of Mitchell 1986). We use the term coated clast to refer to a lithic clast that is surrounded by a rim (<2 mm thick) of finegrained kimberlite groundmass. They are distinguished from pelletal clasts by generally having a lithic clast at the centre, by irregular surfaces and by abundant phenocrysts in the rims. The term coherent refers descriptively to a rock that does not have an apparent fragmental rock texture (following Cas et al. 2008). Coherent in this context does not have a genetic connotation: i.e., it does not imply that the rock was formed by crystallisation from a magmatic melt (e.g., a dyke, or lava; cf. McPhie et al. 1993). All depths given are relative to the present surface (not mining level) at 700 m elevation and are referred to as metres below present surface (m bps). The current mining level is at 108 m bps. Each mined bench is approximately 12 m high. For simplicity we use standard terminology for volcaniclastic rocks (based on that of Fisher 1961) and we follow kimberlite terminology introduced by Sparks et al. (2006).
K2 is a steep-sided volcanic pipe, ∼250×300 m in diameter (at present surface), with a known vertical extent of > 900 m bps (Fig. 1b). It is crudely pear-shaped in plan at the present surface and changes shape with depth (Fig. 1c). It has a minimum volume of approximately 1.4×107 m3. The pipe walls dip steeply inwards (82–88°) and remain steep at depth, although the presence of coarse lithic-rich breccias and megablocks in the west half means that it can be difficult to locate the margin precisely in drillcore. K2 sits on the junction between a NE-trending fault and NW striking joints (Kurszlaukis and Barnett 2003). A narrow NW-striking fissure-like extension protrudes from the main body of the pipe at depth (Fig. 1c) and follows NW-striking joints. It is inferred to represent a cavity which may have been formed by large-scale spalling of the pipe walls (Kurszlaukis and Barnett 2003). K2 has a complex internal geology comprising layered breccia units of varying grain sizes and compositions (Fig. 1b, d; Kurszlaukis and Barnett 2003). Thin kimberlite dykes occur around the margins. We divide K2 into two geologically distinct parts, K2 West and K2 East, which we describe separately below (Fig. 1). K2 West K2 West comprises a lithofacies association of layered breccias and pyroclastic rocks. It is >700 m thick and occupies western parts of the pipe in the upper 400 m and the whole of the pipe at depths greater than 400 m bps. This lithofacies association accounts for ∼85% of the known pipe volume. The constituent lithofacies are described in detail below.
Methods
Country rock megablocks (Bl)
A detailed field and petrographic investigation of the K2 kimberlite pipe was undertaken. All available mine faces
This lithofacies comprises largely coherent blocks of basement country rock 5 to >20 m in diameter (Fig. 2a,
Bull Volcanol (2009) 71:95–112
99
or two types of basement rock (predominantly gneiss, schist, amphibolite, dolerite or quartzite). Clasts are angular with joint-bounded surfaces, although a minor proportion of them are moderately to well-rounded. Some clasts are intensely fractured and show jigsaw-fit textures. The poorly sorted fine-coarse ash matrix is composed of abundant country rock and xenocrysts (particularly biotite, amphibole, quartz and feldspar), serpentine, clay and minor olivine crystals. Calcite occurs as a patchy pore-fill in some Br units.
b). The margins of some blocks are highly fragmented with jigsaw-fit textures (Fig. 2a). Most megablocks have been rotated relative to their orientation in the adjacent wall rock and some have undergone as much as ∼90° rotation about a sub-horizontal axis. Megablocks are found throughout K2 West (Fig. 1d) and the largest are associated with the country rock breccia lithofacies (Br). Country rock breccia (Br) Country rock breccias comprise >90 vol.% lithic lapilli and blocks with an interstitial matrix (Fig. 2c). They occur in units several metres thick to >30 m thick around the western margin of K2, interbedded with other breccia lithofacies (e.g., Fig. 2d); several Br units extend across the whole of K2 West at depth (Fig. 1d). The clast population in the country rock breccias is commonly dominated by one
a
Matrix-supported kimberlite breccia This volumetrically important lithofacies comprises 20– 50 vol.% angular to well rounded lithic lapilli and blocks in an altered kimberlite-derived matrix containing abundant free olivine crystals (Fig. 3a). The clast population is
b
d c
e
Fig. 2 Country rock breccias (Br) in K2 West. a Jigsaw-fit textured country rock megablock has been comprehensively fractured but fragments have not been rotated or moved. b Country rock breccia lithofacies showing abundant basement clasts in sparse matrix (bench 6). Mine benches are 12 m high. c >20 m diameter gneiss country rock megablock in country rock breccia (Br). Block has been rotated ∼90°
about a horizontal axis relative to in-situ country rock. Metre-rule for scale. All observations from 72–96 m bps. d Inward-dipping contact between country rock breccia and clast-supported volcaniclastic kimberlite breccia in K2 West. e Jigsaw textured gneiss clasts in core (DDH221, 243.78 m)
100
dominated by gneiss, schist, amphibolite and dolerite with subordinate marble and basalt. Dolerite accounts for >40% of the clast population in some units between 370–570 m bps, below the outcrop level of a dolerite sill in the country rock. Extremely well-rounded basement lithic lapilli and blocks are common and many exhibit concentric-foliated margins of altered rock around fresh cores. Angular, platy fragments apparently derived from these altered layers occur in the matrix (see Fig. 3a). Scattered country rock megablocks occur in association with this lithofacies. The long axes of platy lapilli and blocks dip inwards at ∼ 40– 80° locally within <10 m of the pipe margins (Fig. 1b). Crude layering within this lithofacies is defined by gradational-to-sharp variations in clast abundance on a metre-to-tens of metres scale (Figs. 4 and 5). This layering crudely dips into the pipe and is typically discontinuous
Bull Volcanol (2009) 71:95–112
over a similar scale and essentially comprises more or less clast-rich pods or lenses. This lithofacies occurs as metre to tens-of-metres-thick layers and lenses within other breccia lithofacies. Clasts are commonly partially to completely altered to serpentine, diopside and calcite (Fig. 3c). The matrix comprises poorly sorted whole and broken free olivine crystals (10–30 vol.%; Fig. 3c), basement xenocrysts and country rock fragments. A range of juvenile clasts are present in the matrix. Pelletal clasts are common and comprise a ∼50 µm -thick rim of altered kimberlite groundmass around an olivine crystal (Fig. 3c). In some matrix-supported kimberlite breccia units, the rims of pelletal clasts appear to have coalesced, forming irregularshaped patches that can enclose olivine crystals (Fig. 3c). The relative proportions of coalesced pelletal clasts versus discrete pelletal clasts is variable. A distinguishing charac-
a
b
c
d
Fig. 3 a Scattered lithic clasts in matrix-supported kimberlite breccia. Platy clast from a concentrically-foliated layered margin of rounded clast in the centre of photograph. b Well-rounded clast within clastsupported kimberlite breccia; 60 cm shown on rule. c Photomicrograph of matrix-supported kimberlite breccia illustrating agglutinated pelletal clasts, which each comprise a serpentinised euhedral olivine
crystal (O) surrounded by a concentric shell of finely crystalline kimberlite. Coated lithic clasts in left and upper right. Pore-space is filled with diopside-serpentine-calcite cement (Borehole DDH075, 110.17 m). d Photomicrograph of juvenile clasts (J) in matrixsupported volcaniclastic kimberlite breccia. Note irregular shape and eccentrically-sited olivine grain in juvenile clast on left
Bull Volcanol (2009) 71:95–112
101
Clast-supported kimberlite breccias This lithofacies is distinguished from the matrix-supported kimberlite breccia lithofacies by a higher lithic clast content (50–90 vol.%). It has a sparse matrix, similar to that in the matrix-supported kimberlite breccia lithofacies, but this differs in having abundant basement xenocrysts and ashgrade country rock fragments. Scattered country rock megablocks are commonly found in association with this lithofacies. Most breccia units are dominated by one or two particular clast types (i.e., comprising >40% clast population) with subordinate amounts of other basement rocks (see Kurszlaukis and Barnett 2003). Jigsaw-fit clast textures are common (Fig. 2e). Approximately 5–15% clasts are moderately to extremely well-rounded and display altered concentric foliated margins as found in the matrixsupported volcaniclastic kimberlite breccia lithofacies (Fig. 2b). This lithofacies occurs conformably interbedded with the other K2 West lithofacies as metre to tens-of-metres-thick layers, lenses and pods (Fig. 1d). Volcaniclastic kimberlite This lithofacies occurs in units up to several tens of metres thick. It is lithologically and petrographically similar to the matrix-supported kimberlite breccia lithofacies but contains <20% lithic lapilli. Scattered lithic blocks are present. Diffuse layering is locally defined by variations in lithic clast abundance. The orientation of this layering is not known as this lithofacies has only been encountered in drillcore. Fig. 4 Graphic logs of drillcores in K2. The contact between K2 East and K2 West is intersected by DDH361 (right). The layers are defined by lithic clast size and abundance and by variations in olivine abundance
teristic of this lithofacies is the presence (<1 vol.%) of subspherical to irregular-shaped juvenile clasts of crystallised kimberlite groundmass. They can be up to several millimetres in diameter. Many have an eccentrically-placed olivine crystal or crystals (Fig. 3d). Coated lithic clasts are common (2–6 vol.%) and vary from several millimetres up to several centimetres in diameter. They comprise a lithic lapilli or lithic ash core surrounded by a 0.1 to >2 mm coating of kimberlite groundmass. They commonly have irregular margins (Fig. 3c). The coatings comprise olivine, phlogopite and spinel crystals (<100 µm) within a finergrained irresolvable serpentine–diopside alteration assemblage (Fig. 3c). The lithofacies has a remnant porosity; pores are fringed with diopside microlites and filled with serpentine and calcite.
Olivine-rich volcaniclastic kimberlite This volumetrically minor lithofacies consists of centimetre-thick units of olivine-rich volcaniclastic kimberlite. It comprises free crystals of olivine (20 mm in diameter) in an altered matrix similar to that of the matrix-supported kimberlite breccia lithofacies. Olivine crystals account for >40 vol.% of the lithofacies. It has been recognised in drillcore in upper levels of the pipe (Figs. 4 and 5), but has not been observed in pit exposures. It occurs interstratified with the matrix-supported volcaniclastic kimberlite breccia lithofacies but the lateral extent of single layers is not known. Coherent kimberlite This lithofacies comprises dense, crystalline kimberlite (Fig. 6). It typically contains <10 vol.% of highly altered lithic clasts (Fig. 6), although units with elevated volumes of lithic inclusions (up to 30 vol.%) are present. Contacts
102
Bull Volcanol (2009) 71:95–112
Fig. 5 Fence diagram constructed from graphic logs of 5-m spaced drillcore across the K2 East-K2 West contact. Two packages can be defined within K2 West i.e., an upper matrix supported kimberlite breccia dominated package and a lower package comprising, matrix-
supported kimberlite breccia, olivine-rich volcaniclastic kimberlite, clast-supported volcaniclastic kimberlite breccia and country rock breccia. Note impersistence of lithic rich units. Inset map shows location of borehole pattern. Top of drillcore at 96 m bps
between units of varying clast abundances are gradational (Fig. 6). It contains 10–15 vol.% of partially fresh to serpentinised olivine macrocrysts and phenocrysts in a variably altered fine-grained (<100 µm) crystalline groundmass dominated by phlogopite, serpentine, calcite, and spinel with minor apatite and diopside. Spinel can be poikilitically enclosed by phlogopite crystals. Some units of coherent kimberlite exhibit small (<50– 100 µm) irregular-shaped, serpentine and calcite domains devoid of opaque groundmass crystals (Fig. 6). These domains can partially surround rounded and irregular blebs of kimberlite groundmass that in some cases look similar to pelletal clasts (Fig. 6). In some coherent kimberlite units, euhedral calcite crystals radiate into the domains from the edges of the groundmass domains. Other coherent kimber-
Fig. 6 Graphic log through part of DDH273 illustrating the layering of units of coherent kimberlite (coherent kimberlite) and massive volcaniclastic kimberlite breccia (matrix-supported kimberlite breccia). Contacts between the two lithofacies are gradational and are recorded by changes in colour, clast content and petrography. The top photograph set illustrates a matrix-supported kimberlite breccia unit with abundant lithic lapilli and blocks and a matrix texture comprising discrete pelletal olivine clasts. Pore space between clasts is filled with serpentine (s) and calcite (cc) cement. Middle set of photographs illustrates dark blue-grey coherent kimberlite (coherent kimberlite) with low volumes of lithic inclusions (<15 vol.%). In thin-section, a fragmental texture is observed and defined by the presence of partially agglutinated pelletal clasts. Pore space (“segregations”) is filled with calcite (white). O=olivine macrocrysts. Lower set of photographs illustrates coherent kimberlite unit with abundant lithic inclusions (L) and olivine macrocrysts (O). In centre of photomicrograph partially agglutinated pelletal clasts (P) can be seen partially surrounded by calcite-filled pore space (white irregular-shaped patches). Note the similarities between the petrographic textures in all three samples
Bull Volcanol (2009) 71:95–112
103
104
Bull Volcanol (2009) 71:95–112
Fig. 7 Fence diagram of K2 pipe illustrating the complex internal structure determined from geological mapping of mine faces from 48– 96 m bps. Contacts between breccia units in K2 West generally dip
into the pipe. Maps of several successive mined faces on benches 7 and 8 are shown. The layered breccias are truncated by the contact with K2 East
lite units are locally texturally heterogeneous, with small patches (<500 µm) devoid of olivine phenocrysts but that have elevated percentages of groundmass spinel. In several units, elevated concentrations of olivine macrocrysts form diffuse clusters up to several centimetres in diameter. Coated clasts, similar to those in the matrix-supported kimberlite breccia lithofacies, are present in some coherent kimberlite units. Lithic clasts have been completely replaced by calcite, phlogopite and/or serpentine (Fig. 6). This lithofacies occurs in one 11-m-thick drillcore intersection at 160 m bps and in many units up to several tens-of-metres thick interbedded with breccia lithofacies (matrix and clast-supported kimberlite breccias) below 400 mbsl. In drillcore, it exhibits gradational contacts into the matrix-supported kimberlite breccia lithofacies (Fig. 6).
The geometries of these units are as yet poorly constrained by drillcore data. Structure of K2 West K2 West exhibits an inward-dipping layered sequence of clastic lithofacies in the upper 500 m bps, first recognised by Kurszlaukis and Barnett (2003; Figs. 7 and 8). Single breccia units dip 20–45° into the pipe from the west, north and north–east margins (Fig. 1d). Sub-vertical contacts between units have been recognised in outcrop (Fig. 7). Five major clast-dominated lobes (country rock breccia and clast-supported kimberlite breccia lithofacies), each 10– 40 m thick, extend 50–200 m into the pipe from the western margin (Fig. 2) and are truncated by the East K2 body. The
a
b
Fig. 8 Selected bench maps showing the complicated layering seen in K2 West. Note long axes of large megablocks (dark shaded) also dip into pipe centre
Bull Volcanol (2009) 71:95–112
lobes are separated by crudely layered juvenile-rich packages up to 100 m thick that are dominated by matrixsupported kimberlite breccia and volcaniclastic kimberlite with minor country rock breccia and clast-supported volcaniclastic kimberlite breccia lithofacies (Fig. 1d). The juvenile clast-rich packages are confined to interior parts of K2 West. Each juvenile clast-rich package is up to ∼40 m thick and most are laterally impersistent over a similar scale (Fig. 5). The NW extension is filled with this layered breccia sequence. At depths greater >500 m bps, the entire breadth of the pipe is dominated by crudely layered juvenile clast-rich lithofacies (clast and matrix-supported kimberlite breccia lithofacies, coherent kimberlite and volcaniclastic kimberlite). Lithic-rich lithofacies (country rock breccia) are present at depth, but the thick extensive lithic-rich breccia lobes seen in higher levels are not apparent. However, the geometries of lithological units at these depths remain poorly constrained by the available data. Dolerite-rich breccias occur between 370–570 m bps, below the level of a ∼40 m-thick dolerite sill intruded in the basement. Close to the pipe margins, dolerite megablocks and boulders are common within country rock breccia and clast-supported kimberlite breccia lithofacies units (Fig. 9). Away from the margins, dolerite clasts comprise up to 70 vol.% of the lithic population in matrix-supported kimberlite breccia and volcaniclastic kimberlite lithofacies. Within some matrix-supported kimberlite breccia units, the coarsest lithic clasts (>10–20 cm in diameter) comprise almost entirely (>95 vol.%) dolerite, whereas the finer clasts (<5 cm in diameter) comprise a wide range of basement lithologies (i.e., dolerite, schist, gneiss and amphibolite). Dolerite-rich facies are not found below ∼570 m bps in the pipe. K2 East K2 East comprises a tapering pipe-like body persistent to at least 500 m bps. It narrows from 120 m diameter at 96 m bps to <50–70 m diameter at >250 m bps but its deep subsurface geometry (>500 m bps) remains poorly
105
constrained (Fig. 1d). It has a minimum volume of ∼1.2× 106 m3. The western margin of K2 East is marked by a N– S striking sheared contact (Fig. 10a) that dips >80° E at the surface and steepens to ∼88° E below 140 m bps (Fig. 1d). K2 East consists entirely of massive volcaniclastic kimberlite. Lithofacies in K2 East Massive volcaniclastic kimberlite This lithofacies is a massive quasi-homogeneous volcaniclastic deposit that typically comprises ∼8–20 vol.% angular lithic clasts supported in a matrix of poorly sorted crystal-lithic tuff (Fig. 10b). The clast population is dominated by basement clasts (dolerite, biotite schist, gneiss, amphibolite gneiss) with subordinate Waterberg Group mudstone and basalt clasts. Diffuse decimetre-to metre-scale pods and zones, containing up to 30 vol.% lithic lapilli, are present in outcrop and in drillcore, but are volumetrically minor (Kurszlaukis and Barnett 2003). Lithic clasts are commonly partially replaced by carbonate and serpentine, particularly in their margins. Rare rounded juvenile clasts (<5 cm) of both volcaniclastic and coherent varieties of kimberlite are present. The matrix comprises a poorly sorted framework of free crystals of olivine (∼30–45 vol.%), lithic clasts and country rock xenocrysts in a pervasive serpentine–diopside cement (Fig. 10c). Olivine crystals are dominantly fine to coarsegrained (<1–6 mm) rounded anhedral crystals or highly angular crystal fragments. Pelletal clasts (<0.5 mm) are common and comprise an olivine crystal coated in a thin (<50 µm) concentric shell of altered fine-grained kimberlite comprising serpentinised olivine and phlogopite, randomly oriented diopside microlites and rare spinel (Fig. 10c). The inter-grain pore-space is filled with isopachous radiating acicular diopside microlites, massive serpentine and clay minerals (Fig. 10d; e.g., Stripp et al. 2006). Country rock xenocrysts include feldspar, amphibole, quartz and biotite. The massive volcaniclastic kimberlite lithofacies of K2 East has a grain size distribution comparable to that seen in other kimberlite pipes and appears depleted in very fine particles (<62.5 µm; Fig. 10d) and very coarse clasts (>10 cm; Fig. 10b) compared to what would be expected of a deposit derived by explosive fragmentation (Walters et al. 2006). K2 East–K2 West contact
Fig. 9 Dolerite-rich clast-supported kimberlite breccia from borehole DDH221 (410–420 m) in K2 West. Clasts >10 cm in diameter are dominantly dolerite (dol), whereas those <10 cm diameter comprise a range of basement lithologies (e.g. gneiss, schist and amphibolite)
The change from K2 west to K2 East is marked by a shear zone a few decimetres wide that comprises millimetre-tocentimetre thick lenses up to several decimetres long of massive volcaniclastic kimberlite (Fig. 10a). The shear zone is apparent at depth in drillcore as a decimetre-thick zone of
106
a
Bull Volcanol (2009) 71:95–112
b
c d
Fig. 10 Lithofacies in K2 East. a The contact of K2 East with K2 West is marked by an eastward-dipping shear zone visible on two benches (7 and 8). To the west are country rock breccia and matrixsupported kimberlite breccia lithofacies and to the east is massive volcaniclastic kimberlite lithofacies. Benches are approximately 12 m high. b Close-up of massive volcaniclastic kimberlite lithofacies illustrating its relatively fine grain size, abundant olivine crystals (O)
and lithic lapilli (L). Note that it typically lacks clasts larger than ∼15 cm. c Photomicrograph of massive volcaniclastic kimberlite illustrating granular nature of deposit, abundant pelletal clasts (O) and filled pore space between grains. d Photomicrograph of pelletal clasts comprising an olivine phenocryst (O) surrounded by a thin shell of altered kimberlite groundmass. Biotite xenocryst is visible along the lower border of the image. Observations taken from 84–96 m bps
broken core. No sense of movement could be established within the shear zone and it does not extend into a fault in the country rock.
proportions of lithic inclusions (<2 vol.%) and some locally exhibit a vertical fabric defined by long axes of olivine crystals. In some intrusions millimetre-thick margin-parallel layering is defined by variations in olivine macrocryst and phenocryst abundance and in groundmass composition. Within these dykes, olivine macrocrysts comprise <5 vol. % whereas euhedral to subhedral olivine phenocrysts comprise 10–30 vol.% (Fig. 11a). Euhedral to subhedral spinels are ubiquitous (5–20 vol.%) and can form necklace textures around olivine phenocrysts. The crystalline groundmass contains calcite, phlogopite and pseudomorphs after monticellite. Groundmass calcite is commonly partially replaced by serpentine (Fig. 11b).
Kimberlite intrusions Dark grey to black, aphanitic to finely porphyritic kimberlite dykes, typically <20–50 cm thick, have intruded along the margins of the pipe and at depth along the contact between K2 East and K2 West. They have characteristically sharp, sinuous contacts with their host and can thicken and thin over a few metres by up to 50%. At some locations decimetre-thick sub-horizontal sill-like bodies branch off from dykes. The dykes contain low
Bull Volcanol (2009) 71:95–112
a
107
(Brown et al. 2007). The upper parts of the pipe and surface crater would have cut through Waterberg Group rocks, but Waterberg Group clasts only form a small component of the pipe fill (<2 vol.% of K2 East), implying that almost all Waterberg Group strata were explosively ejected into pyroclastic deposits beyond the crater rim during early stages. Emplacement of layered breccias in K2 West
b
Fig. 11 a Photomicrograph of kimberlite dyke in K2 East. Olivine crystals (O) are serpentinised, and have started to lose their outlines. b Cross-polar photomicrograph of dyke encountered in drillcore in K2 West (Borehole DDH034, 160.44 m). Olivine crystals (O) have been completely replaced by serpentine (dark grey) and are surrounded by a zone of amorphous serpentine (serp) which partially replaces adjacent primary calcite crystals (cc)
Discussion Here we discuss the evolution of the K2 pipe in terms of the hypothetical framework of kimberlite eruptions proposed by Sparks et al. (2006). Pipe excavation We infer that the opening stages of the eruption involved dyke propagation, explosive cratering and early growth of the pipe from the top down (Fig. 12a; Stage I of Sparks et al. 2006). This stage was followed by pipe widening and deepening as under-pressured conditions developed in the conduit (Stage II of Sparks et al. 2006). These stages are considered to be essentially erosive and evidence of them in the preserved geology in K2 is limited. The rectilinear pipe shape and the elongated NW extension at depth are consistent with structural trends in the country rock (Fig. 1c; Kurszlaukis and Barnett 2003; Barnett and Lorig 2007). Degradation of the pipe is attributed to tensile failure of the rock mass (e.g., Barnett 2008). The presence of rounded altered lithic clasts with concentric-foliated layered margins and fresh cores in the lithofacies suggests pipe localisation in zones of chemically corroded country rock
Sparks et al. (2006) argued that under conditions of explosive flows adjusted to 1 atm surface pressure at the vent, the vertical velocity of pyroclasts exiting a kimberlite pipe decreases as a function of increasing vent radius for any given magma supply rate. Thus as the conduit widens, a critical threshold can be reached at which most fragments cannot be ejected and thus start to accumulate within the pipe (Stage III of Sparks et al. 2006); such a condition can also develop due to waning magma supply. It is at this stage that we consider that the breccias of K2 West started to accumulate within the pipe (Fig. 12b). Development of layering in K2 West The layers in K2 are defined primarily by changes in the lithic clast to juvenile clast ratio and in the composition of the lithic clasts. A continuum is observed from lithofacies that contain >90 vol.% country rock clasts (e.g., country rock breccias) through to breccias in which country rock clasts account for <20 vol.% (e.g., volcaniclastic kimberlite). We infer that this layering records the layer-by-layer filling of the pipe by two main processes: gravitational collapse of the pipe margins and deposition from explosive eruptions (Fig. 12b). The high lithic clast content, presence of megablocks, typically monolithic or weakly heterolithic compositions and sparse matrix of the country rock breccias are consistent with formation by gravitational collapses resulting from tensile failure of the wall rock (e.g., Barnett 2008). Jigsaw-fit textures may have resulted from stresses induced on impact in the pipe and are similar to those seen in clasts in debris avalanche deposits (e.g., Glicken 1990; Reubi and Hernandez 2000; Clavero et al. 2004). The inward-dipping slopes defined by bedding surfaces in West K2 (Fig. 1d) are consistent with the construction of talus slopes by the accumulation of collapsed lithic clasts (Kurszlaukis and Barnett 2003). Major gravitational collapses produced large-volume breccia lobes (∼4×105 m3) that covered the entire West K2 (Fig. 1d). The dips of some breccia lobes are significantly higher than the angles of repose expected for talus (∼35–38°; Selby 1982). Several factors may have contributed to these elevated angles of repose including high inter-clast friction between irregular shaped and angular clasts (e.g., Carrigy 1970) and the
108 Fig. 12 Cartoon illustrating the inferred eruption history of K2. a Opening eruptive phases were highly explosive due to large magma overpressures and they probably involved large-scale surface cratering. We infer that rapid widening and deepening of the conduit was achieved when the pressure difference allowed tensile failure of the wall-rock. b At some point, exit conditions were such that the erupting mixture did not have the ability to clear material from the conduit and pipe started to fill. Deposition occurred from explosive volcanic eruptions and from gravitational collapse of the pipe walls. This phase deposited the layered breccia sequence in K2 West. c Late-stage explosivity excavated a large portion of K2 West and deposited massive volcaniclastic kimberlite to form K2 East. The eruption may have developed into a partially fluidised system as pyroclasts accumulated in the pipe. d After the eruption the pipe comprised a remnant layered sequence (K2 West) and a cross-cutting internal conduit filled with massive volcaniclastic kimberlite (K2 East)
Bull Volcanol (2009) 71:95–112
a
b
c
d
presence of fine cohesive particles (i.e., a matrix). Riedel et al. (2003) showed that angle of repose of cohesionless particles increases to over 40° in the interiors of cones (analogous to the interior of a pipe), compared to ∼30° on the outer slopes of particle piles due to circular chaining and slope chaining processes that provide 3D-support for inward-facing slopes. Steep grain fabrics in marginal breccias in West K2 could result from frictional drag achieved by compaction of the pipe filling deposits, where
compaction is greater in central parts of the pipe (Lorenz 2007). We infer that juvenile-rich lithofacies (matrix and clastsupported kimberlite breccia, volcaniclastic kimberlite, olivine-rich volcaniclastic kimberlite, coherent kimberlite) were deposited by fall-back of pyroclasts into the pipe from gas-thrust tephra columns (e.g., Downes et al. 2007). Proximal fallout is complicated and can occur simultaneously from the upper parts of a column, from the margins
Bull Volcanol (2009) 71:95–112
of the lower column and from ballistic trajectories (e.g., Sable et al. 2006). Fallout may be further complicated by occurring within a narrow pipe. We infer that the resultant deposits are chaotic, massive to very poorly bedded, exhibit poor sorting and may contain outsized clasts (cf. White 1991). Recycling of clasts within the pipe could lead to the polymict clast populations seen in the matrix-supported kimberlite breccia lithofacies. The variable lithic clast abundances in matrix-supported kimberlite breccia in K2 West indicate that the ejecta varied in composition with time. Large collapses of the pipe wall would have fed freshly fragmented country rock into the pipe that could be subsequently entrained by the explosive eruptions. Whether these eruptions were sustained or comprised a series of discrete explosions remains uncertain. The inter-layering of lithic- and juvenile-rich lithofacies indicates that both mass-movement and pyroclastic processes operated contemporaneously to fill the pipe. The occurrence of these deposits to depths of >700 m bps suggests that the pipe must have been deeply excavated prior to infill (Kurszlaukis and Barnett 2003). Large-volume subsidence breccias in other kimberlite pipes have been used to infer significant volume deficits at depth during kimberlite eruptions (e.g., Venetia K1, Barnett 2004). Welding processes We interpret the coherent kimberlite lithofacies (coherent kimberlite) as a welded pyroclastic rock on the basis of (A) its porphyritic texture and crystalline igneous groundmass that indicate that it crystallised from kimberlite melt, and (B) on its gradational contacts between coherent kimberlite and matrix-supported volcaniclastic kimberlite breccia, locally elevated volumes (up to 30 vol.%) of lithic clasts and textural heterogeneity indicative of a clastic genesis (Fig. 6; e.g., Sparks and Wright 1979). We interpret the irregular calcite and serpentine patches (Fig. 6) as cement filling remnant pore space between sintered juvenile particles. This texture is strongly reminiscent of that in the clastic lithofacies (massive volcaniclastic kimberlite; matrix-supported volcaniclastic kimberlite breccia; Figs. 3c, 6 and 10). Welding in pyroclastic rocks comprises two different and independent processes, namely sintering together of hot clasts at point contacts (Guest and Rogers 1967) and deformation of the hot clasts accompanied by reduction of pore space. The former process can occur without the latter to form incipiently welded rocks (Smith 1960). Although compactional foliation (e.g. fiamme and eutaxitic textures) commonly forms as pore space is reduced, there are examples of densely welded rocks with no foliation (Sparks et al. 1999). We thus emphasise that the occurrence of welding does not require compactional foliations to develop.
109
Serpentine–calcite domains imparting a “globular segregationary” or “segregationary” texture on the rocks (Clement and Skinner 1985; Mitchell 1986; Seggie et al. 1998) have been interpreted as the products of frozen degassing fronts formed at the point of volatile exsolution of erupting kimberlite magma (e.g., ‘transitional kimberlite’ at Premier and Kamfersdam pipes, South Africa, Field and Scott Smith 1999; also Hetman et al. 2004; Skinner and Marsh 2004). The rocks in which they are found are typically interpreted as intrusions. There are difficulties with this interpretation. Firstly, violent dynamic processes are impossible to preserve directly in a deposit (Sparks et al. 2006). Secondly, during vesiculation, the textural relationship between the melt phase and the gas phase is characteristically governed by surface tension effects, which produce thinned asymptotic melt walls (Sparks et al. 2006). The textures of the coherent kimberlite lithofacies in K2, however, are the exact opposite with the “segregation” domains forming the thinned asymptotic spaces between the melt regions (i.e., the pelletal clasts). We therefore propose that this texture can instead be interpreted as the product of sintering of juvenile particles. We propose that periodic increases in mean pyroclast temperature or accumulation rate (e.g., Head and Wilson 1989) during the K2 eruption resulted in the pyroclasts welding on deposition. We infer that the original outlines of many juvenile clasts in coherent kimberlite have been lost due to agglutination, alteration and diagenesis. Welded rocks are common proximal deposits from many styles of explosive eruptions (e.g., Hawaiian, Strombolian and Plinian activity; e.g., Sparks and Wright 1979; Head and Wilson 1989) and are particularly common during the explosive eruption of low viscosity magmas (e.g., Sumner 1998). They should be expected during the eruption of extremely low viscosity kimberlite magmas. Partial excavation of K2 West and deposition of massive volcaniclastic kimberlite in K2 East The steep-sided pipe-like geometry of K2 East and the discordant contact with the layered K2 West breccias indicate that K2 East must post-date K2 West, but the time gap is not known (Fig. 2; Kurszlaukis and Barnett 2003). The sub-vertical walls of K2 East suggest that the K2 West breccias were lithified by the time K2 East was emplaced. The K2 East eruption must have had an initial highly explosive phase in order to partially excavate K2 West. We interpret the fragmental texture and the abundant juvenile and mantle-derived clasts (pelletal lapilli, olivine macrocrysts and phenocrysts) of massive volcaniclastic kimberlite as indicative of a primary pyroclastic deposit. We infer that massive volcaniclastic kimberlite accumulated in the newly-formed conduit within the K2 pipe
110
during the waning stages of the K2 East eruption (Fig. 12c). The well-mixed nature of massive volcaniclastic kimberlite together with the absence of very fine particles and very coarse clasts (see Fig. 10b, c) is consistent with an emplacement mechanism that involved both elutriation of fine particles and sinking and break-up of large lithic clasts, e.g., a vigorously fluidised system (Sparks et al. 2006; Walters et al. 2006; Gernon et al. 2008). We propose that as the eruption waned, pyroclasts started to accumulate in the pipe. Exsolved gas streamed through this bed and elutriated fines from the top of the bed. We infer that gas was derived from eruptions. Experimental work on gas fluidisation of particle beds in diverging conduits (Gernon et al. 2008) illustrates that at relatively high gas fluxes, central parts can become vigorously fluidised and well mixed. In the experiments of Gernon et al. (2008), clasts are transported upwards within central parts by gas bubbles, and are carried back down by subsidence at the margins. Under steady conditions, a steady-state conveyor system develops. Shearing is concentrated at the margins of the fluidised body: this process may explain the sheared contact between K2 East and West (Fig. 10a). Re-entrainment of subsided pyroclasts is an efficient means of mixing together lithic clasts of varying lithology and provenances from the country wall stratigraphy (Sparks et al. 2006). Heterogeneities seen in massive volcaniclastic kimberlite, in terms of clast size and abundance may reflect incomplete mixing or segregation processes during the shut-down of fluidisation. Stage IV—late stage intrusion and diagenesis of the pipe-fill Following the filling of the pipe, late-stage intrusions of degassed kimberlite magma occurred along the margins of the pipe (Fig. 12d). Margin-parallel layering in some dykes indicates multiple phases of intrusion. The diopside– serpentine–calcite assemblages in the pore-spaces of the massive volcaniclastic kimberlite and matrix-supported volcaniclastic kimberlite lithofacies are interpreted as posteruption cements precipitated by hydrothermal fluids that circulated through the hot and porous deposit after the eruption. Recent work by Stripp et al. (2006) indicated that the diopside–serpentine assemblage forms at temperatures between ∼250–380°C from meteoric-derived fluids. This assemblage provides a minimum estimate on the emplacement temperature of massive volcaniclastic kimberlite. We similarly interpret the calcite and serpentine domains in the coherent kimberlite facies (coherent kimberlite) as remnant pore-space in densely-welded rock that has been filled with hydrothermally precipitated secondary minerals. The projection of dog-tooth calcite crystals into some of these domains is reminiscent of isopachous textures in the
Bull Volcanol (2009) 71:95–112
cements of the massive volcaniclastic kimberlite and matrix-supported kimberlite lithofacies.
Conclusions The K2 kimberlite pipe is filled with a complex sequence of layered rocks (K2 West) that resulted from a protracted explosive eruption. The presence of layered breccias and pyroclastic rocks to a considerable depth (>900 m) requires that the pipe was deeply excavated prior to filling (Kurszlaukis and Barnett 2003). The breccias are interpreted as mass-movement (subsidence) deposits that were emplaced concurrently with pyroclastic deposits derived from explosive eruptions. During this phase, eruptive conditions were periodically suitable for the deposition of welded rocks. A later phase of volcanic activity partly excavated the layered breccia sequence and deposited massive volcaniclastic kimberlite in a steep-sided conduit (K2 East). The accumulating pyroclasts became fluidised during the infill of this later conduit. Following the eruption, the deposits were extensively altered (serpentinised) and lithified by meteoric-derived hydrothermal fluids circulating through the hot, permeable deposits (e.g. Stripp et al. 2006). The geological evidence is consistent with the model of kimberlite volcanism outlined by Sparks et al. (2006) and appears contradictory to intrusive models of kimberlite pipe formation (sensu Clement 1982; Clement and Reid 1989; Field and Scott Smith 1999) and models invoking very short eruption durations (e.g., Wilson and Head 2007). The presence of a layered sequence to significant depths in K2 West questions whether a simple tripartite kimberlite pipe model (crater–diatreme–root zone divisions; Hawthorne 1975; Clement and Skinner 1985; Field and Scott Smith 1999) is valid for all southern African kimberlite pipes. Acknowledgements We acknowledge funding by De Beers Group Services, UK Ltd. and De Beers Consolidated Mines. Geological staff at Venetia Mine, South Africa are thanked for on-site support. The paper benefited greatly from many useful discussions with W. Barnett, T. Gernon, S. Kurszlaukis, J. Robey, J. Schumacher, S. Trickett and A. Walters. RSJS acknowledges a Royal Society-Wolfson Merit award. Thorough reviews by J.D.L. White, T. Thordarson and J. McPhie (editor) greatly improved the science and style of the manuscript.
References Barnett WP (2004) Subsidence breccias in kimberlite pipes—an application of fractal analysis. Lithos 76:299–316 Barnett WP (2008) The rock mechanics of kimberlite volcanic pipe excavation. J Volcanol Geotherm Res DOI 10.1016/j. jvolgeores.2007.12.021
Bull Volcanol (2009) 71:95–112 Barnett WP, Lorig L (2007) A model for stress-controlled pipe growth. J Volcanol Geotherm Res 159:108–125 Barton JM Jr, Barnett WP, Barton ES, Barnett M, Doorgapershad A, Twiggs A, Klemd R, Martin J, Mellonig L, Zenglein R (2003) The geology of the area surrounding the Venetia kimberlite pipes, Limpopo Belt, South Africa: a complex assembly of terranes and granitoid magmatism. South African J Geol 106:109–128 Brown RJ, Kavanagh J, Sparks RSJ, Tait M, Field M (2007) Mechanically disrupted and chemically weakened zones in segmented dike systems cause vent localization: evidence from kimberlite volcanic systems. Geology 35:184–188 Carrigy MA (1970) Experiments on the angles of repose of granular materials. Sedimentology 14:147–158 Cas RAF, Porritt L, Pittari A, Hayman P (2008) A new approach to kimberlite facies terminology using a revised general approach to the nomenclature of all volcanic rocks and deposits: descriptive to genetic. J Volcanol Geotherm Res DOI 10.1016/ j.jvolgeores.2007.12.018 Clavero JE, Sparks RSJ, Pringle MS, Polanco E, Gardeweg MC (2004) Evolution and volcanic hazards of Taapaca Volcanic Complex, Central Andes of Northern Chile. J Geol Soc London 161:603–618 Clement CR (1982) A comparative geological study of some major kimberlite pipes in northern Cape and Orange Free State. Ph.D. Thesis (unpublished) University of Cape Town pp 432 Clement CR, Reid AM (1989) The origin of kimberlite pipes: an interpretation based on the synthesis of geological features displayed by southern African occurrences. In: Ross J, Jaques AL, Fergusan J, Green DH, O’Reilly SY, Danchin RV, Janse AJA (eds) Sydney Geol Soc Australia Kimberlites and related rocks, 14:632–646 Clement CR, Skinner EMW (1985) A textural-genetic classification scheme of kimberlites. Trans Geol Soc South Africa 88:403–409 Dawson JB (1967) A review of the geology of kimberlite. In: Wyllie PJ (ed) Ultramafic and Related Rocks. Wiley, New York, pp 241–251 Dawson JB (1971) Advances in kimberlite geology. Earth Sci Rev 7:187–214 Dawson JB (1980) Kimberlites and their xenoliths. Springer, New York, p 252 Downes PJ, Ferguson D, Griffin BJ (2007) Volcanology of the Aries micaceous kimberlite, Central Kimberley Basin, Western Australia. J Volcanol Geotherm Res 159:85–107 Ekren EB, Byers FM Jr (1976) Ash-flow fissure vent in west–central Nevada. Geology 4:247–251 Field M, Scott Smith BH (1999) Contrasting geology and near-surface emplacement of kimberlite pipes in Southern Africa and Canada. VIIth Int Kimberlite Conf 1:217–237 Fisher RV (1961) Proposed classification of volcaniclastic sediments and rocks. Geol Soc Am Bull 71:127–132 Gernon TM, Gilbertson MA, Sparks RSJ, Field M (2008) Gasfluidisation in an experimental tapered bed: insights into processes in diverging volcanic conduits. J Volcanol Geotherm Res DOI 10.1016/j.jvolgeores.2007.12.034 Glicken H (1990) The rockslide-debris avalanche of the May 18 1980, eruption of Mount St Helens—10th anniversary perspectives. Geosci Canada 17:150–153 Guest JE, Rogers PS (1967) The sintering of glass and its relationship to welding in ignimbrites. Proc Geol Soc Lond 1641:174–177 Hawthorne JB (1975) Model of a kimberlite pipe. Phys Chem Earth 9:1–15 Head JW, Wilson L (1989) Basaltic pyroclast eruption: influence of gas-release patterns and volume fluxes on fountain structure, and the formation of cinder cones, spatter cones, rootless flows, lava ponds and lava flows. J Volcanol Geotherm Res 37:261–271
111 Hetman CM, Scott Smith BH, Paul JL, Winter FW (2004) Geology of the Gahcho Que kimberlite pipes, NWT, Canada: root to diatreme transition zones. Lithos 76:51–74 Junqueira-Brod TC, Gaspar JC, Brod JA, Kafino CV (2005) Kamafugitic diatremes: their textures and field relationships with examples from the Goiás alkaline province, Brazil. J South Am Earth Sci 18:337–353 Kano K, Matsuura H, Yamauchi S (1997) Miocene welded tuff infilling a funnel-shaped eruption conduit, Shiotani, southeast of Matsue, SW Japan. Bull Volcanol 59:125–135 Kurszlaukis S, Barnett WP (2003) Volcanological and structural aspects of the Venetia Kimberlite cluster—a case study of South African kimberlite maar–diatreme volcanoes. S Afr J Earth Sci 106:165–192 Kurszlaukis S, Lorenz V (1997) Volcanological features of a low viscosity melt: the carbonatitic Gross Brukkaros volcanic field, Namibia. Bull Volcanol 58:421–431 Kurszlaukis S, Buttner R, Zimanovski B, Lorenz V (1998) On the first experimental phreatomagmatic explosion of a kimberlite melt. J Volcanol Geotherm Res 80:323–326 Lorenz V (1975) Formation of phreatomagmatic maar–diatreme volcanoes and its relevance to kimberlite diatremes. Phys Chem Earth 9:17–29 Lorenz V (1986) On the growth of maars and diatremes and its relevance to the formation of tuff rings. Bull Volcanol 48:265–274 Lorenz V (2007) Syn- and post-eruptive hazards of maar–diatreme volcanoes. J Volcanol Geotherm Res 159:285–312 Lorenz V, Kurszlaukis S (2007) Root zone processes in the phreatomagmatic emplacement model and consequences for the evolution of maar–diatreme volcanoes. J Volcanol Geotherm Res 159:4–32 McPhie J (1993) The Tennant Creek porphyry revisited: a synsedimentary sill with peperite margins, Early Proterozoic, Northern Territory. Australian J Earth Sci 40:545–558 Mitchell RH (1975) Geology, magnetic expression, and structural control of the Central Somerset Island kimberlites. Can J Earth Sci 12:757–764 Mitchell RH (1986) Kimberlites: Mineralogy, geochemistry and petrology. Plenum, New York, p 442 Phillips D, Kiviets GB, Barton ES, Smith CB, Viljoen KS, Fourie LF (1999) 40Ar/39Ar dating of kimberlites and related rocks: problems and solution. Proc 7th Int Kimberlite Conf Cape Town South Africa 677–687 Reedman AJ, Park KH, Merriman RJ, Kim SE (1987) Welded Tuff infilling a volcanic vent at Weolsong, Republic of Korea. Bull Volcanol 49:541–546 Reubi O, Hernandez J (2000) Volcanic debris avalanche deposits of the upper Maronne Valley (Cantal Volcano, France): evidence for contrasted formation and transport mechanisms. J Volcanol Geotherm Res 102:271–286 Riedel C, Ernst GGJ, Riley M (2003) Controls on the growth and geometry of pyroclastic constructs. J Volcanol Geotherm Res 17:121–152 Sable JE, Houghton BF, Wilson CJN, Carey RJ (2006) Complex proximal sedimentation from Plinian plumes: the example of Tarawera 1886. Bull Volcanol 69:89–103 Seggie AG, Hannweg GW, Colgan EA, Smith CB (1998) The geology and geochemistry of the Venetia kimberlite cluster, Northern Province, South Africa. Proc 7th Int Kimberlite Conf 1:750–756 Selby MJ (1982) Hillslope materials and processes. Oxford University Press, Oxford, p 451 Skinner EMW, Marsh JS (2004) Distinct kimberlite classes with contrasting eruption processes. Lithos 76:183–200 Smith RL (1960) Ash-flows. Geol Soc Am Bull 71:795–842
112 Sparks RSJ, Wright JV (1979) Welded air-fall tuffs. In: Chapin CE, Elston WE (eds) Geol Soc Am Special Paper Ash-flow tuffs, 180:155–166 Sparks RSJ, Tait SR, Yanev Y (1999) Dense welding caused by volatile resorption. J Geol Soc 156:217–225 Sparks RSJ, Baker L, Brown RJ, Field M, Schumacher J, Stripp G, Walters A (2006) Dynamical constraints on kimberlite volcanism. J Volcanol Geotherm Res 155:18–48 Stoppa F (1996) The San Venanzo maar and tuff ring, Umbria Italy: eruptive behaviour of a carbonatite–melilite volcano. Bull Volcanol 57:563–577 Stripp GR, Field M, Schumacher JC, Sparks RSJ (2006) Postemplacement serpentinization and related hydrothermal metamorphism in a kimberlite from Venetia, South Africa. J Met Geology 24:515–534
Bull Volcanol (2009) 71:95–112 Sumner JM (1998) Formation of clastogenic lava flows during fissure eruption and scoria cone collapse: the 1986 eruption of Izu-Oshima Volcano, eastern Japan. Bull Volcanol 60:195– 212 Walters A, Phillips J, Brown RJ, Field M, Gernon T, Stripp G, Sparks RSJ (2006) The role of fluidisation processes in the formation of volcaniclastic kimberlite: grain size observations and experimental investigation. J Volcanol Geotherm Res 155:119–137 White JDL (1991) Maar-diatreme phreatomagmatism at Hopi Buttes, Navajo Nation (Arizona), USA. Bull Volcanol 53:239–258 White JDL, McClintock M (2001) Immense vent complex marks flood-basalt eruption in a wet, failed rift: Coombs Hills, Antarctica. Geology 29:935–938 Wilson L, Head JW III (2007) An integrated model of kimberlite ascent and eruption. Nature 447:53–57
