Author's personal copy - SRK. Jelsma et al. / Lithos 112S (2009) 155 165 157 Author's personal copy...

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Tectonic setting of kimberlites

Hielke Jelsma a,⁎, Wayne Barnett b, Simon Richards c, Gordon Lister d

a De Beers Exploration, Johannesburg, South Africab SRK Consulting, Vancouver, Canadac School of Earth and Environmental Sciences, James Cook University, Townsville, Australiad Research School of Earth Sciences, The Australian National University, Canberra, Australia

a b s t r a c ta r t i c l e i n f o

Article history:Received 21 October 2008Accepted 19 June 2009Available online 3 July 2009

Keywords:KimberliteTectonic settingSupercontinentGondwanaLithospheric discontinuityMelt focusing

Kimberlites can be viewed as time capsules in a global plate tectonic framework. Their distribution illustratesclustering in time and space. Kimberlite ages span the assembly and break-up of a number ofsupercontinents, such as Rodinia and Gondwana. These supercontinents show time lines with (i) broadperiods devoid of kimberlite magmatism corresponding to times of continent stability, and (ii) narrowkimberlite emplacement windows corresponding to times of fundamental plate reorganizations. Thisepisodicity implies that kimberlite emplacement events are intrinsically related to particular stages in the lifecycle of supercontinents. The onset of kimberlite magmatism is closely associated with thermal perturbations(thermal insulation, mantle upwelling?) beneath a stagnant or sluggish supercontinent. These perturbationsmay have caused uplift and the onset of continental break-up through fracture zones propagating into thesupercontinent. Subsequent spreading and ocean floor development is marked by apparent cusps and jogs inplate motion paths. Resultant strain is accommodated along trans-lithospheric corridors with episodic upliftand erosion and focused kimberlite melt migration. The corridors are manifest as discontinuities in thelithosphere mantle, measured as geophysical gradients and as changes in mantle lithosphere composition.Within the crust, these corridors are expressed as (a) terrane boundaries, (b) incipient continental rifts, (c)fracture zones, or (d) major dyke swarms. Some kimberlite populations are clustered along parallel sets ofcorridors widely distributed across a large part of a subcontinent and repeated magmatism is seen withinmany of the clusters. The association of kimberlite occurrences with discontinuities may be ascribed tofavorable conditions for melt production and to resultant melt focusing along high strain zones that containfractures and faults. Such conditions may be attained during different stages in the evolution of continents:(a) supercontinent formation; (b) incipient rifting (driven by far-field stresses?) and onset of continentalbreak-up; and (c) strain accommodation along the continental continuation of oceanic fracture zones duringspreading. Type (c) may show concomitant kimberlite magmatism in separate continents after break-up.

© 2009 Elsevier B.V. All rights reserved.

1. Introduction

Kimberlites are part of a suite of volatile-rich, high-MgO, ultrabasicrocks (kimberlites, sensu lato). These ultrabasic rocks includepotentially diamondiferous members such as kimberlites and olivinelamproites and superficially similar but uneconomic rocks such asmelnoites and carbonatites (e.g., Mitchell, 1995; Chakhmouradianet al., this issue; Nielsen et al., this issue).

Kimberlite ages span the assembly and break-up of a number ofsupercontinents from the Neoarchean to the Cenozoic Eras, theyounger of which are Rodinia and Gondwana/Pangea. Kimberliteoccurrences are found on all continents, and in a variety of geologicalsettings emplaced as small volcanic pipes, sills or dykes (Fig. 1).Amongst these, the diamondiferous members are generally restricted

to cratons, shields or mobile belts and underlain by thick subconti-nental lithosphere mantle (SCLM; Clifford, 1966; Janse, 1991).

The lowermost part of the SCLM is “a complex, chemically activezonewhere depletedmantle lithosphere is injected andmetasomatisedby small-volume melts sourced from the adjacent fertile, adiabatic,convecting mantle asthenosphere” (O'Reilly and Griffin, 2006). Generalagreement exists about the source region for the Group 2 kimberlites inSouthAfrica (metasomatisedmantle lithosphere peridotite, e.g., Le Roexet al., 2008), but opposing views persist in the case of the Group 1kimberlites, with models ranging from sub-lithospheric sources (e.g.,Nowell et al., 2004; Paton et al., this issue; Pearsonet al., 2008; Breyet al.,this issue) to deep lithospheric sources (e.g., Le Roex et al., 2003; Beckerand Le Roex, 2006).

Differing models have also been proposed for the trigger ofkimberlite melt formation, including (a) enhanced mantle plumeactivity producing hotspot tracks with kimberlite age migration (e.g.,Morgan, 1983; Le Roex, 1986; Skinner, 1989; Heaman and Kjarsgaard,2000), (b) subduction of oceanic lithosphere and partial melting of

Lithos 112S (2009) 155–165

⁎ Corresponding author.E-mail address: [email protected] (H. Jelsma).

0024-4937/$ – see front matter © 2009 Elsevier B.V. All rights reserved.doi:10.1016/j.lithos.2009.06.030

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r.com/ locate / l i thos


overlying mantle producing traces of deep-seated subduction (e.g.,Helmstaedt and Gurney 1984; McCandless 1999), (c) thermalperturbations associated with tectonic events involving lithosphericfaults formed or reactivated during break-up of continents with strainlocalization andmelt focusing (e.g., Dawson,1970;Marsh,1973; Sykes,1978; Bailey 1992; White et al., 1995; Jelsma et al., 2004; Moore et al.,2008), or (d) multiple origins (e.g., Heaman et al., 2004).

In this study spatial and temporal data for kimberlites fromSouthern Africa are compared with the lithosphere anisotropy andsurface geological features. We will discuss the setting of kimberlitemagmatism within a global tectonic context and outline differenttectonic environments inwhich kimberlite melts are formed, that mayhelp elucidate their origins.

2. Temporal and spatial constraints

Over time, continents show alternating periods set apart either byabundance or scarcity of kimberlite magmatism. Heaman et al. (2003)noted a 110 Ma period of absence of kimberlite activity between360 Ma and 250 Ma worldwide. In Southern Africa a quiescent periodis evident between 500 Ma and 250 Ma, following the assembly ofpalaeocontinent Gondwana.

Appendix A and Fig. 2 provide an overview of available age datausing Southern Africa as an example. The database comprises 195occurrences representing a wide spectrum of rock types (including122 kimberlites, 22 melnoites, and 33 carbonatites and syenites),analytical methods (mostly U–Pb zircon and perovskite and Ar–Ar,Rb–Sr and K–Ar mica and whole rock techniques), and precision andaccuracy, and is the collective effort of a multitude of researchersduring a period of over 30 years. Clustering of ages can be seen duringthe Mesoproterozoic (1635 Ma, 1330 Ma,1200 Ma, 1100 Ma), the earlyPaleozoic (510 Ma, Pan-African), the Mesozoic (240 Ma, 145 Ma,120 Ma, 85 Ma, 73 Ma), and the Cenozoic (54 Ma and 31 Ma). Someage populations appear to represent a local, short-lived magmaticepisode (e.g., 1635 Ma Kuruman event), whereas others aredistributed across an entire region (e.g., the Cretaceous 120 Ma and85 Ma events) and show a wider variation of ages.

Kimberlite preservation depends on the geological environmentand on age (younger kimberlites tend to be better preservedcompared to older kimberlites, but this is certainly not always thecase). This is basically an exhumation and burial account. Over 1 km oferosion off-the-top of a typical kimberlite pipe removes most of thecrater- and diatreme-facies components, and therefore lowers thelikelihood of discovery. Burial of kimberlites beneath a syn- or post-emplacement sedimentary cover sequence temporarily preserves thekimberlites. Following burial, the kimberlites may (or may not) be re-exhumed during subsequent erosion.

In space, the continents are transected by sets of systematiclineaments or mega-lineaments that are considered to be corridors of

Fig. 1. World digital elevation model with superimposed cratonic and shield domains (light grey) and the global distribution of kimberlites (●), lamproites (○), melnoites (x), andcarbonatites (+).

Fig. 2. Kimberlite age data for Southern Africa.

156 H. Jelsma et al. / Lithos 112S (2009) 155–165


Fig. 3. Distribution of kimberlites in Southern Africa. (a) Digital elevation model (90 m SRTM) with ocean floor bathymetry. Note the blocky rectilinear coastline of the subcontinent.(b) Geological of Southern Africa (Jelsma et al., 2004), with superimposed kimberlite age data (Table 1). (c) Structural elements mapped from geological, geophysical, remotelysensed and elevation data and extent of Permo-Triassic Karoo basins (rift and sag) and recent earthquake locations (source: USGS, 1973-2008, MN1.5).

157H. Jelsma et al. / Lithos 112S (2009) 155–165


concentrated, aligned tectonic activity. Fig. 3 shows the distribution ofkimberlites in Southern Africa with the major crustal structuralelements. These structures occur as lineaments visible on digitalelevation models, high resolution satellite (or aerial photo) imagery,and on geophysical datasets or geological maps. They are marked bychanges in drainage and vegetation patterns, fault-line valleys orscarps, density of linear photo-anomalies and breaks and displace-ments of geophysical anomalies (gravity, magnetic, seismic andresistivity features). Many of these lineaments are associated withzones of neo-tectonic faulting (e.g., faulting of escarpments, faulting ofdune patterns in Namibia and Botswana, drainage system re-orientations, etc.) and current (low magnitude) seismic activity.

Some of these lineaments are lithosphere-scale structural corridorsthat formed focal areas for magmas and fluids ascending from themantle. These corridors are manifest as anisotropies associated withgradients in the SCLM, measured by seismic velocity, seismicanisotropy, resistivity, potential field and/or compositional hetero-geneities (e.g., Fouch et al., 2000; Jaques and Milligan 2004; Jones andCraven2004; Snyder et al., 2004; O'Reilly andGriffin 2006).Within thecrust, these corridors are expressed as (a) terrane boundaries,(b) incipient continental rifts, (c) fracture zones, either as the con-tinental continuation of oceanic fracture zones or as associatedaccommodation zones, or (d) major dyke swarms. On a regionalscale, the distribution of kimberlites is typically clustered along thesetrans-lithospheric corridors. An association is apparent betweenkimberlites and tips or shoulders of rifts, major pre-existing dykeswarms, and bends, step-overs and fault intersections within struc-tural corridors. Many corridors have been repeatedly reactivated overtime, suggesting they are through-going fundamental “flaws” withinthe continental architecture, and “pathways” for kimberlite magmas.

In Southern Africa, a strong association was noted (a) betweenJurassic-age kimberlites and ENE–WSW and NNW–SSE trending

corridors and (b) between Cretaceous-age kimberlites and NE orNW trending corridors (e.g., Jelsma et al., 2004). Similar lithospheretrends are observed as anisotropies associated with the cratonic keel(James et al., 2001), craton edges and marginal orogenic belts(Fig. 3b), the Permo-Triassic Karoo Rift System (Fig. 3c; Catuneanuet al., 2005) and the Cenozoic East Africa Rift System (Fig. 3c), as wellas recent earthquake loci (Fig. 3c; cf. Dirks et al., 2003). Both corridorsystems are directly related to the break-up of Gondwana and thecontinuation of oceanic fracture zones into the African continent. Forinstance, the Jurassic corridor trends were imprinted on the continentduring the separation of West and East Gondwana and the opening ofthe Somali Basin and the Weddell Sea along NNW–SSE trendingfracture zones (such as the Davie Fracture Zone); the Cretaceouscorridor trends during the separation of South America and Africawith the opening of the South Atlantic. Similar examples can be foundelsewhere. The corridors are expressed by, respectively, the blockyrectilinear outline of segments of the coastline of east Africa (Somalia–Kenya–Tanzania–Mozambique–South Africa) and southwest Africa(Namibia and Angola), visible on digital elevation models. Some ofthese corridors have older origins, such as the well-described ENE–WSW trending Thabazimbi–Murchison Lineament and ZoetfonteinFault in South Africa, the NE–SW trending Omaruru–OmatakoLineament in Namibia and the NE–SW trending Quilenges–AnduloFault Zone in Angola.

The ENE–WSW trending Zoetfontein Fault is a high angle faultzone that marks the northern edge of the Kaapvaal Craton and has ahistory of repeated reactivation since the late Archean and continuingto the present. It controlled deposition of both the MesoproterozoicWaterberg and Carboniferous–Jurassic Karoo sediments and basalts.The Thabazimbi–Murchison Lineament to the southmarks the zone ofaccretion between theWitwatersrand Block and the Northern Domainof the Kaapvaal Craton and has shown 2700 Ma of episodic

Fig. 3 (continued).



deformation (Good and de Wit, 1997). The Omaruru–OmatakoLineament is a prominent and long-lived NE–SW trending fault zonethat has displaced Karoo basalts and members of the Karoo-ageOkavango dyke swarm. It controlled deposition of Cretaceoussediments and is still seismically active. The basement-cover basalCretaceous surface shows a marked step across this structure.Neotectonic activity is indicated by linear depressions and displace-ment and erosion of dunes. Towards the Atlantic coast the structure isassociated with complexes of carbonatites and syenites. The NE–SWtrending Quilenges–Andulo Fault Zone in Angola has a strike length ofover 400 km between the towns of Quilenges and Andulo and has awidth of 90 km (De Boorder 1982). It is a composite set ofdiscontinuous faults that are deep-seated and associated withcarbonatites and kimberlites. The fault zone is along strike with theLucapa “Graben” in northeastern Angola. The latter was active duringthe Palaeoproterozoic, the Permo-Triassic, the Cretaceous and theCenozoic (with neotectonic faulting and associated earthquakeactivity). It can be traced to the southwest to one of the transferzones of the South Atlantic Ocean (Sykes, 1978).

Fig. 4 shows the orientation data for major faults as well as the spatialdistribution of kimberlites at a local scale (b40 km) and a regional scale(40–250 km) using autocorrelation. Note the marked similarity betweenfault orientations and the inter- and intra-cluster kimberlite spatialdistribution. On a local scale, many of the kimberlites are composites ofmultiple emplacement stages with dykes commonly acting as precursorsto subsequent eruptive events, such as observed at Finsch and Kimberleymines (Barnett and Preece, 2002; Dawson, 1970). Circular pipes arefrequently found at the intersection of kimberlite dyke trends and otherstructures such as dolerite dykes or faults. A key feature however is apreferred common dyke-fracture orientation, ellipticity of pipes, as wellas alignment of lobes and pipes amongst coeval kimberlites. For instance,the c. 120 Ma kimberlites within the western part of the Kaapvaal Cratonin South Africa are manifest as kimberlite dyke swarms with segmenteden-echelon dyke-fracture arrays and long axes of blows along dykesoriented between 065° and 015° (N=73). In contrast, the c. 85 Makimberlites within the Letseng cluster in Lesotho show dyke orientationstrending 115° (N=233, Fig. 5). Within the Orapa cluster in Botswanapipes are well preserved and pipe long axes and dykes trend 135°(N=46). Such a commonality of orientations is observedwithin a clusterbut also across several clusters (for example, Finsch and Kimberleyclusters) within a continental domain and implies a regional tectoniccontrol on kimberlite emplacement. If we assume that the orientation ofdykes and elongated pipes best reflects the orientation of maximumcompressive stress (e.g. Nakamura, 1977), this orientation may havechanged from NE–SW during emplacement of the c. 120 Ma kimberlitesto NW–SE toWNW–ESE during emplacement of the c. 85Ma kimberlites.

The NW–SE to NNW–SSE orientations may relate to the WegenerStress Anomaly (WSA) described by Andreoli et al. (1996). Thisanomaly extends over the western parts of Southern Africa and wasprobably attained in theMid Cretaceous (ca.100Ma). A compilation ofstress tensors measured on mines suggests that the NW–SE stressorientation is still currently active (Stacy and Wesseloo, 1998). It ispossible that the WSA existed prior to the opening of the SouthAtlantic, based on coast-parallel mafic dyke orientations (132Ma FalseBay Swarm, Reid et al., 1991), and weakened at various times, such asduring the emplacement of the c. 120 Ma kimberlites (Andreoli et al.,1996).

3. Bottom-up or top-down?

The Jurassic–Cretaceous kimberlites in Southern Africa are clus-tered along sets of parallel lineaments widely distributed across alarge part of the subcontinent, with the younger groups found closertowards the continent margins (Fig. 6). The linear distribution inSouth Africa has been used as an argument for a mantle plume hotspot track (e.g., Skinner 1989). However, similar-age magmatism canbe seen over vast areas, such as shown by the 120 Ma and 85 Ma agepopulations, each covering areas of in excess of 2000 km by 1000 km.Repeated magmatism is observed in many clusters (Appendix A;Fig. 3b), such as Kimberley (119 Ma, 85 Ma), Sutherland (120 Ma,74Ma), Klipspringer (148Ma, 74Ma), and Luxinga/Alto Cuilo (135Ma,115 Ma). Parallelism of same-age kimberlite “tracks” and repeatedmagmatism at the same sites suggests emplacement along structuraldiscontinuities rather than an association with hotspot tracks.Systematic younging of kimberlites between 70 Ma (E) and 46 Ma(W) in the western part of South Africa and in Namibia probablyreflects strain propagation and melt migration along fracture zones,associated with the opening of the South Atlantic Ocean.

Table 1 compares the timing of kimberlite magmatism with majorevents affecting Gondwana. Kimberlite emplacement age windowscoincide with: (i) apparent cusps (variation in direction) and jogs(variation in velocity) in the relative plate motion paths of continents(Africa–Eurasia, Africa–Antarctica), based on a fit of seafloor magneticanomaly isochrons with oceanic FZ traces and on the construction ofapparent polar wander paths from palaeomagnetic data (but note thatdata points are few and far apart); (ii) stages of uplift and erosion, andthe formation of unconformities in the offshore and onshorestratigraphic record documenting episodic tectonic instability (cf.Jelsma and Smith, 2004; Moore et al., 2008); (iii) Large IgneousProvince (LIP) manifestation. Marsh (1973) andHargraves and Onstott(1980) already noted the synchronicity of kimberlite ages and changesin Apparent Polar Wander curves. In Hargraves words, “these

Fig. 4. Fault orientations and kimberlite spatial distribution in Southern Africa. Major faults (N=818, mapped from Landsat, SRTM) show dominant orientations at 055° and 150°(a). Autocorrelation on kimberlite distributions (irrespective of age, N=2173) within clusters (0–40 km) shows preferential NW–SE alignment (b) and between clusters (40–250 km) 053°, 078° and 145° trends (c).



relationships suggest that the emplacement of kimberlites maycoincide with episodes of changes in the direction of plate motions”.Jelsma et al. (2004) and Moore et al. (2008) linked kimberlitemagmatism to changes in the direction and velocity of plate motion.

Kimberlite age data for “Laurasia” and West Gondwana are shownin Fig. 7. Members within each of these supercontinents (e.g. SouthernAfrica and South America inWest Gondwana) show similar age peaks,reflecting a common history, but different peaks are visible whenmembers across the two supercontinents are compared. For instance,the 370 Ma reconstruction shows continent break-up and associatedkimberlite magmatism affecting Siberia, and continent stability andlack of kimberlite magmatism in Southern Africa and South America.In North America Jurassic kimberlite magmatism follows the break-upof Pangea with the opening of the Atlantic and the eruption of theCentral Atlantic Magmatic Province at 200 Ma. In Southern Africa,kimberlite magmatism follows (1) the break-up of West and EastGondwana with the opening of the Weddell Sea and Somali Basin andthe eruption of the Karoo–Ferrar basalts at 180Ma, (2) the break-up ofSouth America and Africa with the opening of the South Atlantic andthe eruption of the Parana–Etendeka basalts at 135 Ma. Kimberlitemagmatism is marked by a number of distinct peaks. At 135–115 Ma,seafloor spreading in the South Atlantic causes rifting of Africa andSouth America and is accompanied by a major peak in kimberlitemagmatism. The subsequent peak at 95–80 Ma is marked byconcomitant magmatism in Southern Africa and South America,after continental break-up, and heralds major plate reorganization.

The temporal association between Large Igneous Province mag-matism (the Karoo–Ferrar Province at 180 Ma, the Parana–EtendekaProvince at 135 Ma and perhaps the Reunion Province at 85 Ma) andthe coinciding and subsequent main pulses of kimberlite magmatism(180–145 Ma, 135–115 Ma, 95–80 Ma) is significant. Sears (2001)proposed that the stalled Gondwana supercontinent “drove its ownbreak-up by insulating the underlying mantle, causing thermal

expansion, and uplift, fracturing and associated LIP magmatism”.Using compilations of data from xenoliths and xenocrysts, Bell et al.(2003) and Griffin et al. (2000) presented evidence for a protractedthinning and heating event during the Jurassic that affected SouthernAfrica. Geochemical imaging of the SCLM by Kobussen et al. (2008)has demonstratedmarked changes between time slice sections for theGroup 2 kimberlites (143–117 Ma, listing their age grouping) and theGroup 1 kimberlites (108–74 Ma, ibid.) across the SW Kaapvaal cratonboundary. The work shows that the SCLM was heated and chemicallyrefertilised by infiltrating asthenosphere–derived melts, thinning theSCLM layer in the order of 40 km. The authors attribute this event tochanges in the stress field associated with the opening of the SouthAtlantic Ocean, or to mantle upwelling. The regional extent of thisthermal event is poorly constrained. Seismic tomography (James et al.,2001 — Kaapvaal Seismic Program) showed no pervasive thermalerosion of the mantle root beneath the interior parts of the KaapvaalCraton.

4. Discussion

Experimental modeling by O'Neill et al. (2005) helps to explainsome of the observations and interpretations, in particular theimportance of discontinuities in the lithosphere mantle and the roleof plate motions in creating or reactivating discontinuities. Theauthors modeled sub-lithospheric melt production and showed thatthe geometry of the continental keel plays a critical part indetermining both location and extent of partial melting as well astime-dependent thermal perturbations in particular in proximity tocraton margins and gradients of the lithosphere–asthenosphereboundary. Beneath thick cratonic lithosphere a thermal upwelling ormantle plume will only produce a small volume of melt, unlessvolatiles are introduced to the system. Significant volumes are onlygenerated once hot upwelling material has advected around the thick

Fig. 5. Lesotho kimberlite pipes and dykes, 90 Ma.



root zone into regions of thinner lithosphere. The modeling showedthe importance of large stress gradients on craton boundaries andother lithospheric discontinuities in localizing strain and focusing(kimberlite) melt migration in these areas.

The association of kimberlite occurrences with trans-lithospherediscontinuities may be ascribed to favorable conditions for meltproduction as a result of thermal perturbations (thermal insulation,mantle upwelling?) and to resultant melt focusing along networks offaults. Such conditions may be attained during different stages in theevolution of continents (Fig. 8).

Type A — Supercontinent formation. Emplacement of kimberlitesfollows the assembly of continental fragments into a supercontinent(such as Rodinia or Gondwana). Examples include 1100 Ma kimber-lites in Zimbabwe (Mwenezi cluster) and India (Narayanpet andWajrakarur clusters) and 520 Ma kimberlites in Southern Africa(Venetia, The Oaks, Sese, Colossus, Uitkomst clusters). In SouthernAfrica, the 1100 Ma kimberlite event may be linked to the emplace-ment of the Umkondo LIP (Hanson et al., 1998), following accretionaryevents along the southern and eastern margins of the Kaapvaal Cratonwithin the Namaqua–Natal orogenic belt. In India, a similar associa-tion may be made between concomitant kimberlite magmatism, athermal anomaly beneath the eastern Dharwar Craton and tectonicevents that took place along the easternmargin of the Dharwar Craton

within the Eastern Ghats Mobile Belt. The 520 Ma kimberlites inSouthern Africa define a broad ENE–WSW trending lineament thatparallels the Pan-African age Damaran Orogenic Belt to the north. Inall examples given, these events may be linked to the formation oflarge-scale mantle thermal anomalies following orogenic events,perhaps as a result of late-stage thermal relaxation following collision,with thermal equilibration and resultant partial melting of (sub?)lithospheric mantle and with kimberlite melt migration within thelithosphere focused along reactivated pre-existing structures.

Type B — Incipient rifting. Following a period of supercontinentstability, thermal insulation of the lithosphere beneath a stagnant orsluggish supercontinent or the impact of a hot spot or plumeimpinging on a supercontinent leads to uplift and the onset ofcontinental break-up. During the initial stages of break-up, a networkof rifts and fault zones is formed as a result of extensional ortranstensional stresses that propagated into the supercontinent. TheKaroo and East Africa Rift Systems in Africa in many places reactivatepre-existing discontinuities. Examples of associated magmatisminclude kimberlites within the Jwaneng cluster (244 Ma; Smith,1983) in Botswana, emplaced along a major NE–SW trending Karoofault zone that forms an accommodation structure between theZoetfontein Lineament and the Thabazimbi–Murchison Lineament.The Kapamba cluster (240Ma; Phillips,1986) in Zambiawas emplaced

Fig. 6. Kimberlite distribution in Southern Africa divided in eight time windows.



Fig. 6 (continued).

Table 1Time-event chart for Southern Africa.

Time Kimberlites“sensu lato”

Events Plate vector cusps Sediments unconformities Hotspot manifestation thermalperturbation

180 Ma ? Karoo–Ferrar LIP (183 Ma), onset spreadingSomali sea (180 Ma?)

183 Ma: Bouvet(Karoo–Ferrar)

145 Ma 148–141 Ma West and East Gondwana separate (145 Ma) 145 Ma Onshore (142 Ma)132 Ma Etendeke–Paraňa LIP (137–133 Ma), opening of

South Atlantic (133–120 Ma)135 Ma Onshore (135 Ma) 132 Ma: Tristan, Shona,

Kerguelen120 Ma 135–115 Ma Africa and South America, India and Antarctica

separate (120 Ma)128 Ma 5 Atl (127–124 Ma)

Development of Central Africa Cretaceous RiftSystem (118 Ma)

121 Ma 6 Atl (120–118 Ma) and onshore(120 Ma)

102 Ma 103–101 Ma 13 Atl (112 Ma) onshore(101 Ma)

92 Ma 95–83 Ma India and Madagascar separate 84 Ma 15 Atl (94–91 Ma) 85 Ma: Marion, St Helena,90°East, Comores, Reunion84 Ma 16 Atl (86–84 Ma) and onshore

(84 Ma)72 Ma 76–70 Ma India and Seychelles separate 21–22Atl (74–67 Ma)65 Ma 65–59 Ma 65 Ma Onshore (66 Ma)52 Ma 56–50 Ma 55 Ma Offshore (54–43 Ma)37 Ma ? 37 Ma Offshore (40–37 Ma) and

onshore (37 Ma)Appendix A Jelsma et al., 2004 Fairhead and Wilson, 2004;

Eagles and König, 2008McMillan and Dale, 2002;Fairhead and Wilson, 2004

http://www.mantleplumes.com



within the NE–SW trending Luangwa Rift, at the intersection with aNW–SE trending fault zone. The Kundelungu kimberlites in the DRC(32 Ma; Batumike et al., 2008) are located at the tip of the NE–SWtrending Lake Mweru–Luapula graben, which is the southwestern

extension of one of the branches of the East Africa Rift System.Individual kimberlites were emplaced at the intersection of NE–SWandNNW–SSE faults. Along the proto-Indian Ocean and proto-AtlanticOcean margins, continental break-up is associated with thermal

Fig. 7. Kimberlite age data for "Laurasia" (top) and West Gondwana (bottom) with continent reconstructions at 370 Ma and 170 Ma (Torsvik and Cocks, 2003).



erosion of mantle lithosphere, and the emplacement of Large IgneousProvinces such as the Karoo–Ferrar Province at 180 Ma (Jourdan et al.,2007), the Central Atlantic Magmatic Province (Hames et al., 2003) at200 Ma and the Parana–Etendeka Province at 135 Ma (Erlank et al.,1984). During uplift, the core of the supercontinent, including thecratons, undergoes extensional strain. Pre-existing lithospheric dis-continuities are reactivated extending into cratonic domains andshields as fracture zones. Related to this stage are the 135–115 Makimberlite populations that include Finsch in South Africa and Catocain Angola.

Type C — Tectonic triggers. Spreading is in progress and platemotion paths are marked by cusps (variation in direction) and jogs(variation in velocity). Changes in the plate motion may have causedshearing of sub-lithospheric ridges with strain accommodation alonglithospheric discontinuities, in this case in particular the continentalcontinuation of oceanic fracture zones. Kimberlites that may berelated to this stage are the 95–80 Ma kimberlite populations thatinclude Orapa in Botswana (93 Ma), Kimberley in South Africa(85 Ma) and Kao in Lesotho (89 Ma). Of note is the connectedmagmatism in Southern Africa and South America (Fig. 7).

5. Conclusions

Kimberlites can be viewed as time capsules in a global platetectonic framework. They record tectonic events in a changing world,marking stages in the evolution of continents and are collectively keypieces in palaeocontinent reconstructions. The trigger for kimberlitemagmatism appears to be related to thermal perturbations (thermalinsulation, mantle upwelling?) beneath a stagnant or sluggishsupercontinent and associated tectonic change — stress changeinduced within the continental plate because of supercontinentfragmentation or supercontinent assembly.

Many kimberlites are associated with trans-lithospheric structuralcorridors. Large stress gradients develop at lithospheric discontinu-ities during tectonic events leading to strain localization. This wouldhave caused localized decompression melting near the base of theSCLM (cf. Vaughan and Scarrow, 2003) or merely focused meltmigration paths following the impingement of a coinciding thermalanomaly at the lithosphere–asthenosphere boundary (O'Neill et al.,2005). In contrast, periods devoid of kimberlite magmatism corre-spond to times of supercontinent stability (a sluggish Gondwana at500–250 Ma), or apparently smooth plate motion paths.

The three tectonic stages, (a) supercontinent formation, (b) incipientrifting and break-up and (c) tectonic triggers represent specific stages in

the evolution of any supercontinent. We have given examples ofkimberlites that can be related to Gondwana break-up, but otherexamples can be provided for other continents.

Variable endowment across continental regions may be related to(a) the chemical composition and physical state of potentialkimberlite source areas, (b) a required minimum size of continentsto be able to cause effective insulation of the underlying mantle anddeflect root zone destructive melt volumes, (c) the size and rootgeometry of continental (and cratonic) lithosphere in determiningmelt production (O'Neill et al., 2005), (d) the presence andorientation of anisotropies within continents in localizing strain andkimberlite melt migration and, interlinked, (e) tectonic change.

Acknowledgements

The Exploration Manager of De Beers Group Services, CharlesSkinner, is thanked for permission to present this work. Dave Apter,Bruce Wyatt, Craig Smith, Chris Hatton, Erika Barton, Paul Dirks,Maarten de Wit and Mike Watkeys are thanked for valuablediscussions over the years. This manuscript has benefited from theconstructive comments by Lynton Jaques and David Snyder. GordonLister acknowledges research support provided by ARC DiscoveryGrant DP0343646 "Tectonic Reconstruction of the Evolution of theAlpine-Himalayan Orogenic Chain".

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.lithos.2009.06.030.

References

Andreoli, M.A.G., Doucoure, M., van Bever Donker, J., Brandt, D., Anderson, N.J.B., 1996.Neotectonics of southern Africa — a review. Africa Geoscience Review 3, 1–16.

Bailey, D.K., 1992. Episodic alkaline igneous activity across Africa. In: Storey, B.C.,Alabaster, T., Pankhurst, R.J. (Eds.), Magmatism and the causes of continental break-up: Geological Society, Special Publications, vol. 68, pp. 241–255.

Barnett, W., Preece, C., 2002. Expanding the geological model for Finsch Mine. SouthAfrican Journal of Geology 105, 381–400.

Batumike, J.M., Griffin, W.L., Belousova, E.A., Pearson, N.J., O'Reilly, S.Y., Shee, S.R., 2008.Africa's youngest kimberlites? LAM-ICPMS U–Pb dating of kimberlitic perovskite:Eocene–Oligocene kimberlites from the Kundelungu Plateau, D.R. Congo. Earth andPlanetary Science Letters 267, 609–619.

Becker, M., Le Roex, A., 2006. Geochemistry of South African on- and off-craton Group Iand group II kimberlites: Petrogenesis and source region evolution. Journal ofPetrology 47, 673–703.

Bell, D.R., Schmitz, M.D., Janney, P.E., 2003. Mesozoic thermal evolution of the southernAfrican mantle lithosphere. Lithos 71, 273–287.

Fig. 8. Schematic diagrams illustrating the three described types of tectonic settings related to kimberlite magmatism. (a) Assembly of supercontinent Gondwana following Pan-African orogenesis (blue). (b) Incipient rifting (grey Karoo basins). The dashed circle represents the postulated thermal anomaly. (c) Tectonic triggers and concomitant magmatism inSouthern Africa and South America, after continental break-up. SF: Sao-Francisco Craton, RP: Rio Plato Craton. The thin lines give examples of trans-continental lithosphericdiscontinuities that have been reactivated as the continental extension of oceanic fracture zones. The white diamonds represents graphic groupings of kimberlites and related rocks.



Brey, G.P., et al. 2009, this issue. Influence of water and fluorine on melting ofcarbonated peridotite at 6 and 10 GPa. Proceedings of the 9th InternationalKimberlite Conference. Lithos 112S, 249–259.

Catuneanu, O., Wopfner, H., Eriksson, P.G., Cairncross, B., Rubidge, B.S., Smith, R.M.H.,Hancox, P.J., 2005. The Karoo basins of south-central Africa. Journal of African EarthSciences 43, 211–253.

Chakhmouradian, A.R., et al. 2009, this issue. “Kimberlite” from Wekusko Lake,Manitoba: Actually a diamond-indicator-bearing dolomite carbonatite. Proceedingsof the 9th International Kimberlite Conference. Lithos 112S, 347–357.

Clifford, T.N., 1966. Tectono-metallogenetic units and metallogenic provinces of Africa.Earth and Planetary Science Letters 1, 421–434.

Dawson, J.B., 1970. The structural setting of African kimberlite magmatism. In: Clifford,T.N., Gass, I.G. (Eds.), Africanmagmatism and tectonics. Oliver and Boyd, Edinburgh,pp. 321–335.

De Boorder, H., 1982. Deep-reaching fracture zones in the crystalline basementsurrounding the West Congo System and their control of mineralization in Angolaand Gabon. Geoexploration 20, 259–273.

Dirks, P., Jelsma, H., Blenkinsop, T., 2003. The Geology of Africa. In: Stuwe, K., Grasemannde Vivo (Eds.), EOLSS. UNESCO Publishing, Oxford, UK.

Eagles, G., König, M., 2008. A model of plate kinematics in Gondwana breakup.Geophysical Journal International 173, 703–717.

Erlank, A.J., Marsh, J.S., Duncan, A.R., Miller, R.M., Hawkesworth, C.J., Betton, P.J., Rex, D.C.,1984. Geochemistry and petrogenesis of the Etendeka volcanic rocks from SouthWestAfrica. Namibia. Geological Society of South Africa Special Publication 13, 195–246.

Fairhead, J., Wilson, M., 2004. Sea-floor spreading and deformation processes in theSouth Atlantic Ocean: are hotspots needed? www.mantleplume.org/satlantic.

Fouch, M.J., Fischer, K.M., Parmentier, E.M., Wysession, M.E., Clarke, T.J., 2000. Shear wavesplitting, continental keels, and patterns of mantle flow. Journal of GeophysicalResearch 105, 6255–6275.

Good, N., de Wit, M.J., 1997. The Thabazimbi–Murchison Lineament of the KaapvaalCraton, South Africa: 2700 Ma of episodic deformation. Journal of the GeologicalSociety of London 154, 93–97.

Griffin, W.L., O'Reilly, S.Y., Davies, R.M., 2000. Subduction-related diamond deposits:constraints and possibilities? Reviews in Economic Geology 11, 291–310.

Hanson, R.E., Martin, M.W., Bowring, S.A., Munyanyiwa, H.,1998. U–Pb zircon age for theUmkondo dolerites, eastern Zimbabe: 1.1 Ga large igneous province in southernAfrica–East Antarctica and possible Rodinia correlations. Geology 26, 1143–1146.

Hargraves, R.B., Onstott, T.C., 1980. Palaomagnetic results from some southern Africankimberlites, and their tectonic significance. Journal of Geophysical Research 85,3587–3596.

Heaman, L.M., Kjarsgaard, B.A., 2000. Timing of eastern North American kimberlitemagmatism: continental extension of the Great Meteor hotspot track? Earth andPlanetary Science Letters 178, 253–268.

Heaman, L.M., Kjarsgaard, B.A., Creaser, R.A., 2003. The timing of kimberlite magmatismin North America: implications for global kimberlite genesis and diamondexploration. Lithos 71, 153–184.

Heaman, L.M., Kjarsgaard, B.A., Creaser, R.A., 2004. The temporal evolution of NorthAmerican kimberlites. Lithos 76, 377–397.

Helmstaedt, H., Gurney, J.J., 1984. Kimberlites of Southern Africa — are they related tosubduction processes? Proceedings of the 3rd IKC, vol. 1, pp. 425–434.

Hames, W.E., McHone, J.G., Ruppel, C., Renne, P. (Eds.), 2003. The Central AtlanticMagmatic Province: American Geophysical Union Monograph, vol. 136, p. 267.

James, D.E., Fouch, M.J., VanDecar, J.C., van der Lee, S., the Kaapvaal Seismic Group, 2001.Tectospheric structure beneath southern Africa. Geophysical Research Letters 28,2485–2488.

Janse, A.J.A., 1991. Is Clifford's rule still valid? Affirmative examples from around theworld. Proceedings of the 5th IKC, Araxa, Brazil, pp. 196–198.

Jaques, L.R, Milligan, P.R., 2004. Patterms and controls on the distribution ofdiamondiferous intrusions in Australia. Lithos 77, 783–802.

Jelsma, H., Smith, C., 2004. Geodynamic setting of kimberlites. AGU, Montreal, May.Abstract.

Jelsma, H.A., deWit, M.J., Thiart, C., Dirks, P.H.G.M., Viola, G., Basson, I.J., Anckar, E., 2004.Preferential distribution along transcontinental corridors of kimberlites and relatedrocks of southern Africa. South African Journal of Geology 107, 301–324.

Jones, A.G., Craven, J.A., 2004. Area selection for diamond exploration using deep-probing electromagnetic surveying. Lithos 77 (1–4), 765–782.

Jourdan, F., Féraud, G., Bertrand, H., Watkeys, M.K., 2007. From flood basalts to theinception of oceanization: example from the 40Ar/39Ar high-resolution picture of

the Karoo large igneous province. Geochemistry Geophysics Geosystems 8, Q02002.doi:10.1029/2006GC001392.

Kobussen, A.F., Griffin, W.L., O'Reilly, S.Y., Shee, S.R., 2008. Ghosts of lithospheres past:imaging an evolving lithospheric mantle in southern Africa. Geology 36, 515–518.

Le Roex, A.P., 1986. Geological correlation between southern African kimberlites andsouth Atlantic hot-spots. Nature 324, 243–245.

Le Roex, A., Bell, D.R., Davis, P., 2003. Petrogenesis of Group I kimberlites fromKimberley, South Africa: evidence from bulk rock chemistry. Journal of Petrology44, 2261–2286.

Le Roex, A., Coe, N., Gurney, J., Pearson, D.G., Nowell, G., 2008. Petrogenesis of Group IIkimberlites: a case study from southern Africa. 9th IKC. Extended Abstract, A-00041.

Marsh, J., 1973. Relationships between transform directions and alkaline igneous rocklineaments in Africa and South America. Earth and Planetary Science Letters 18,317–323.

McCandless, T.E., 1999. Kimberlites: mantle expressions of deep-seated subduction.Proceedings of the 7th IKC, vol. 2, pp. 545–549.

McMillan, I.K., Dale, D.C., 2002.West Coast Model. Germany-RSAWorkshop, Cape Town,18–20 Nov.

Mitchell, R.H., 1995. Kimberlites, Orangeites, and Related Rocks. Plenum Press, NewYork. 410 pp.

Moore, A., Blenkinsop, T., Cotterill, F., 2008. Controls on post-Gondwana alkalinevolcanism in Southern Africa. Earth and Planetary Science Letters 268, 151–164.

Morgan, J.W., 1983. Hot spot tracks and the early rifting of the Atlantic. Tectonophysics94, 123–139.

Nakamura, K., 1977. Volcanoes as possible indicators of tectonic stress orientation —

principal and proposal. Journal of Volcanology and Geothermal Research 2, 1–16.Nielsen, T.F.D., et al. 2009, this issue. Distribution of kimberlite and aillikite in the

Diamond Province of southern West Greenland: A regional perspective based ongroundmass mineral chemistry and bulk compositions. Proceedings of the 9thInternational Kimberlite Conference. Lithos 112S, 358–371.

Nowell, G.M., Pearson, D.G., Bell, D.R, Carlson, R.W., Smith, C.B., Kempton, P.D., Noble, S.R.,2004. Hf isotope systematics of kimberlites and their megacrysts: new constraints ontheir source regions. Journal of Petrology 45, 1583–1612.

O'Reilly, S.Y., Griffin, W.L., 2006. Imaging global chemical and thermal heterogeneity inthe subcontinental lithospheric mantle with garnets and xenoliths: Geophysicalimplications. Tectonophysics 416, 289–309.

O'Neill, C.J., Moresi, L., Jaques, A.L., 2005. Geodynamic controls on diamond deposits:implications for Australian occurrences. Tectonophysics 404, 217–236.

Paton, C., et al. 2009, this issue. Identifying the asthenospheric component of kimberlitemagmas from the Dharwar Craton, India. Proceedings of the 9th InternationalKimberlite Conference. Lithos 112S, 296–310.

Pearson, D.G., Nowell, G.M., Kjarsgaard, B.A., Dowall, D.P., 2008. The genesis ofkimberlite: geochemical constraints. 9th IKC. Extended Abstract, A-00149.

Phillips, D., 1986. 40Ar/39Ar age analyses – progress. Princeton University Report.Reid, D.L., Erlank, A.J., Rex, D.C., 1991. Age and correlation of the False Bay dolerite dyke

swarm, southwestern Cape, Cape Province. South African Journal of Geology 94,155–158.

Sears, J., 2001. Icosahedral fracture tessellation of early Mesoproterozoic Laurentia.Geology 29, 327–330.

Skinner, E.M.W., 1989. Contrasting Group I and Group II kimberlite petrology: towards agenetic model for kimberlites. Geological Society of Australia, Special Publication14, 528–544.

Smith, C.B., 1983. PhD thesis, University of the Witwatersrand, Johannesburg.Snyder, D.B., Rondenay, S., Bostock, M.G., Lockhart, G.D., 2004. Mapping the mantle

lithosphere for diamond potential using teleseismic methods. Lithos 77, 859–872.Stacy, T.R., Wesseloo, J., 1998. Final project report: evaluation and upgrading of records

of stress measurement data in the mining industry. Safety in Mines ResearchAdvisory Committee (SIMRAC). Mine Health and Safety Council, Johannesburg,South Africa, Report GAP 511b, 31p.

Sykes, L., 1978. Intraplate seismicity, reactivation of pre-existing zones of weakness,alkaline magmatism, and other tectonism postdating continental fragmentation.Reviews Geophysics Space Physics 16, 621–688.

Torsvik, T.H., Cocks, L.R.M., 2003. Reconstruction methods and palaeogeography. Schoolof Geosciences, the University of the Witwatersrand.

Vaughan, A., Scarrow, J., 2003. K-rich mantle metasomatism in control of localizationand initiation of lithospheric strike-slip faulting. Terra Nova 15, 163–169.

White, S.H., de Boorder, H., Smith, C.B., 1995. Structural controls of kimberlite andlamproite emplacement. Journal of Geochemical Exploration 53, 245–264.


Author's personal copy - SRK. Jelsma et al. / Lithos 112S (2009) 155 165 157 Author's personal copy...

Documents

Transcript of Author's personal copy - SRK. Jelsma et al. / Lithos 112S (2009) 155 165 157 Author's personal copy...