Elsevier

Precambrian Research

Volume 249, August 2014, Pages 129-143
Precambrian Research

Dykes of the 1.11 Ga Umkondo LIP, Southern Africa: Clues to a complex plumbing system

https://doi.org/10.1016/j.precamres.2014.05.006Get rights and content

Highlights

  • We report six ∼1110 Ma U–Pb ages (Umkondo LIP) from dyke swarms of southern Africa.

  • The dykes define up to three radiating swarms centered on the craton margins.

  • Umkondo geochemistry is consistent with a significant SCLM contribution.

  • Dyke distribution has a strong lithospheric control.

  • A mantle plume or global-mantle warming are viable heat purveyors.

Abstract

The Umkondo Large Igneous Province (LIP) is represented by widespread (∼2.0 × 106 km2) mafic intrusions that were rapidly emplaced (1112–1108 Ma) into the Kalahari craton of southern Africa and the formerly adjacent Grunehogna Province of Antarctica during Rodinia assembly. Very few Umkondo-aged dykes have been identified before, resulting in a poor understanding of this LIP's plumbing system and origin. Here we report six new ∼1110 Ma U–Pb TIMS baddeleyite ages for various dolerite dykes, which, when coupled with geochemistry from some of the dykes, suggest association with the Umkondo LIP. The distribution of dykes defines distinct radiating swarms, which locate two separate magmatic centers on the northern margin of the Kalahari craton, and a third less robust center on the SE margin. The Umkondo intrusions’ geochemistry indicates significant partial melting of the sub-continental mantle lithosphere (SCLM) and requires a transient thermal anomaly in the mantle. A viable model sees a mantle plume ascend beneath the craton and split into different portions that moved and ascended to different lithospheric thin-spots along the margins of the craton. As an alternative, the rise in mantle temperature associated with continental aggregation at this time is considered sufficient to cause partial melting of the SCLM without any plume involvement. Specific features of the Umkondo LIP's plumbing system are supportive of either model, and an approach of multiple working hypotheses is recommended.

Introduction

The 1112–1108 Ma Umkondo LIP (Fig. 1) is a widespread intraplate magmatic event in southern Africa (Hanson et al., 2006). It is named for extensive dolerite sills that intrude the Proterozoic carbonate and siliciclastic strata of the Umkondo Group in eastern Zimbabwe and Mozambique (McElhinny, 1966, McElhinny and Opdyke, 1964). Based on paleomagnetism (McElhinny, 1966) and geochemistry (Munyanyiwa, 1999), the sills have been shown to be of approximately the same age as basaltic lava flows in the upper parts of the Umkondo Group. A distinct paleomagnetic signature has also been used to link numerous other intrusions across southern Africa to the Umkondo LIP (Gose et al., 2006, Hargraves et al., 1994, Jones and McElhinny, 1966, Wilson et al., 1987). Emplacement of the Umkondo sills in the type area was originally constrained by Rb–Sr isotopic methods to ∼1100 Ma (Allsop et al., 1989), but this age assignment has been refined to 1112–1108 Ma by U–Pb geochronological studies (Hanson et al., 2004a, Hanson et al., 1998, Wingate, 2001). Many additional intrusions and volcanic rocks from southern Africa have been dated to be synchronous with the Umkondo LIP (reviewed by Hanson et al., 2006, and shown in Fig. 1). Included amongst these are bimodal volcanic rocks that outcrop as isolated hills in northwestern Botswana. These hills form part of a geophysically defined narrow NE–SW trending belt in nortwestern Botswana known as the NW Botswna Rift (Key and Mapeo, 1999). Within the southwestern region of this belt is the ∼1106 Ma bimodal volcanic Kgwebe and Mabeleapodi hills and associated sedimentary rocks that are collectively refered to as the Kgwebe Formation (Schwartz et al., 1996, Singletary et al., 2003). These rocks have been shown to be deposited in an intercontinental rift (Mondie, 2000). Toward the northeast of the Kgwebe Hills are isolated outcrops of ∼1106 Ma felsic rocks known as the Goha and Gubatsha hills, which are correlated to the Kgwebe Formation (Hanson et al., 2006, Singletary et al., 2003). Several granites from the subsurface of northwestern Botswana can be reagrded as the plutonic equivalents of the Kgwebe Formation, while one borehole intersected an ∼1107 Ma mafic intusion (see Hanson et al., 2006, and refrences therein). The occurence of ∼1.1 Ga felsic volcanism may be extended to the southwest of Namibia, where the Oorlogsende Porphyry, the Kartatsaus, Langberg and Rostock rhyolites yield comparable ages (Hanson et al., 2006).

In spite of extensive work only two dykes from northern Zimbabwe have been linked with some certainty to the Umkondo LIP. They are the NNE-trending Marondera dyke north of Harare, and the NE-trending Deweras dyke in NW Zimbabwe (Allsop et al., 1989, Wilson et al., 1987) (Fig. 1). The Marondera dyke is an isolated NNE-trending dyke intruding parallel to the Popoteke fault (Wilson et al., 1987, Allsop et al., 1989). According to Allsop et al. (1989) who studied it in detail: “samples M 85-41 to 46 were taken in 1985 throughout the width of a 40 m wide, N-S trending dyke north of Marondera … which had yielded an Umkondo-type direction of magnetization”. Rb–Sr data on a large number of whole-rock samples yielded a vague linear array that when regressed produced an Umkondo-like age (Allsop et al., 1989). Based on the combination of the Rb–Sr errorchron and paleomagnetic data, Allsop et al. (1989) considered the dyke to belong to the Umkondo LIP, but it should be realized that the exact age of this dyke remains unconstrained. According to Wilson et al. (1987), the NE-trending Deweras dyke is up to 100 m wide and can be traced for ∼120 km. The Deweras dyke, which yielded an Umkondo-type magnetic remanence, was grouped with other dykes of similar trend further to the north and to the east (Wilson et al., 1987) as the Guruve swarm. The Guruve dyke swarm was further equated with a number of NE-trending dykes form the Kamativi dyke swarm in the extreme west of Zimbabwe (Wilson et al., 1987). There exists, however, a possibility that dykes within the Guruve-Kamativi swarm (Fig. 1) are representative of several magmatic events (Hahn and Steiner, 1990).

Furthermore, arching NW- to NNW-trending dykes in the Mutare dyke swarm of northeastern Zimbabwe (Fig. 1) show strong geochemical resemblance to Umkondo-related units (Hanson et al., 2006). A TIMS (Thermal ionization mass spectrometry) U–Pb baddeleyite age of 724 ± 2 Ma, however, was obtained for one dyke of the Mutare swarm (Mukwakwami, 2005). This is more consistent with the age assignment of 600–500 Ma for the swarm by earlier workers based on paleomagnetism and Rb–Sr isotopic investigations (Wilson et al., 1987). The “Mutare swarm” could also be composite in nature, comprising dykes of Umkondo-age and other younger or older generations.

Dykes from another mixed-age swarm (i.e., the giant Okavango Dyke Swarm or ODS; Fig. 1) have tentatively been linked to the Umkondo LIP (Hanson et al., 2006, Jourdan et al., 2009, Jourdan et al., 2004). From both 40Ar/39Ar (44 dykes), and geochemical analyses (77 dykes), nearly 85% of the 420 dykes identified within the ODS are thought to be Karoo in age (180–178 Ma) (Jourdan et al., 2004, Le Gall et al., 2002) as long suspected (Elburg and Goldberg, 2000), whereas the remaining dykes are Proterozoic in age (Jourdan et al., 2004, Le Gall et al., 2002). There is the possibility that some of these older dykes, referred to here as the Proterozoic Okavango Dyke Swarm (or PODS; Fig. 1), may be synchronous with the Umkondo LIP on the basis of a range of apparent Proterozoic ages obtained from the speedy step-heating 40Ar/39Ar method (Jourdan et al., 2004), and similar geochemical signatures (Jourdan et al., 2009). This approach, however, does not allow for evaluating this hypothesis with sufficient confidence.

Finally, dykes from the NE-trending Save-Limpopo Dyke Swarm (SLDS; see Fig. 1) display similar “speedy” step-heating 40Ar/39Ar results as the PODS (Jourdan et al., 2006). Out of 19 identified SLDS dykes, 14 dykes yielded ages in the 1700–700 Ma range, while eight dykes were interpreted to be Karoo-aged (Jourdan et al., 2006). The Karoo-aged SLDS, like the ODS thus also display structural inheritance from a Proterozoic and potentially Umkondo-aged dyke swarm.

Given the paucity of recognized Umkondo age dykes, the plumbing system for this LIP is largely unconstrained. Unraveling the emplacement of LIPs is pivotal for their characterization, and for understanding their origins (Ernst et al., 2008a, Ernst et al., 2005). Recent advances in the separation of magmatic baddeleyite (ZrO2) (Söderlund and Johansson, 2002) have made the U–Pb dating of even very old mafic intrusive rocks fairly routine (e.g., Gumsley et al., 2013, Söderlund et al., 2010). The precise ages yielded by U–Pb baddeleyite geochronology, coupled with the geochemistry of dykes (e.g., Olsson et al., 2010) provide the ideal tool for characterizing the plumbing system of the Umkondo LIP.

Section snippets

U–Pb geochronology

Here several dykes from southern Africa were sampled and ages of ∼1110 Ma were obtained from U–Pb dating (TIMS; see Supplementary material accompanying this paper for details on the methods used) of grains of magmatic baddeleyite. All samples yielded statistically valid ages with P-value ≥0.08. These results show the dykes to be coeval with the Umkondo-type sills and are interpreted to be part of the plumbing system for this LIP (see Fig. 2 for Concordia diagrams and Table 1 for U–Pb data).

Geochemistry of the Dibete and SLDS dykes

Despite a growing geochronological database, complete sets of published geochemical data from Umkondo-related units are still scarce (see Hanson et al., 2006 for a review of most available geochemical data). Here major and trace element results from three dykes (nine samples) of the 1110 ± 4 Ma Dibete swarm and six dykes of the Proterozoic SLDS are reported (Table 2). Analytical totals range from 97.7 to 102.6% and major oxide values were recalculated to total 100%. Original analytical totals

Comparison with the Umkondo LIP

The type Umkondo mafic rocks and related units are fractionated continental tholeiites with basaltic to basaltic andesite compositions (Fig. 3) that are enriched in LILE and LREE and depleted in HFSE (Fig. 6A). The REE compositions display unfractionated HREEs [(Gd/Yb)N = 1.1–1.5], and LREE enrichment with small negative Eu anomalies (Fig. 6B). The Umkondo dolerites are generally low-TiO2 basaltic rocks (Fig. 3, Fig. 4), but 11 samples (of which 3 analyses were reported) of the Middleburg basin

Multiple magmatic centers

The distribution of Umkondo-related dykes (mostly dated in this study) suggests at least two distinct radiating swarms, consistent with two coeval but separate Umkondo magmatic centers on the northeastern and northwestern margins of the Kalahari craton (Fig. 1). A third center is postulated along the southeastern margin of the Kalahari craton on the basis of the NW-trending Divuli Ranch dyke (this study) and the SLDS trend (assuming the older dykes in this swarm are Umkondo-aged).

The first

Acknowledgments

The authors would like to thank two anonymous reviewers, who greatly contributed to this paper. MDK thanks A. Gumsley who assisted in the baddeleyite separation from sample BHW-D and analysis thereof. The staffs of Kumba Iron Ore are thanked for access and support provided during BC's fieldwork. BC was funded by the PPM (Paleoproterozoic Mineralization Group) that is administered by NJB. AH acknowledges logistical support from the Department of Geology, University of Zimbabwe. This is

References (73)

  • C. Hawkesworth et al.

    Mantle processes during Gondwana break-up and dispersal

    J. Afr. Earth Sci.

    (1999)
  • J. Jacobs et al.

    The Kalahari Craton during the assembly and dispersal of Rodinia

    Precambrian Res.

    (2008)
  • F. Jourdan et al.

    The Karoo triple junction questioned: evidence from Jurassic and Proterozoic 40Ar/39Ar ages and geochemistry of giant Okovango dyke swarm (Botswana)

    Earth Planet. Sci. Lett.

    (2004)
  • F. Jourdan et al.

    Basement control on dyke distribution in Large Igneous Provinces: case study of the Karoo triple junction

    Earth Planet. Sci. Lett.

    (2006)
  • F. Jourdan et al.

    Lithospheric mantle evolution monitored by overlapping large igneous provinces: case study in southern Africa

    Lithos

    (2009)
  • B. Le Gall et al.

    40Ar/39Ar geochronology and structural data from the giant Okavango and related mafic dyke swarms, Karoo igneous province, northern Botswana

    Earth Planet. Sci. Lett.

    (2002)
  • B. Le Gall et al.

    The Okovango giant mafic dyke swarm (NE Botswana): its structural significance within the Karoo Large Igneous Province

    J. Struct. Geol.

    (2005)
  • Z.X. Li et al.

    Assembly, configuration, and break-up history of Rodinia: a synthesis

    Precambrian Res.

    (2008)
  • W.F. McDonough et al.

    The composition of the earth

    Chem. Geol.

    (1995)
  • H. Munyanyiwa

    Geochemical study of the Umkondo dolerites and lavas in the Chimanimani and Chipinge districts (eastern Zimbabwe) and theri regional implications

    J. Afr. Earth Sci.

    (1999)
  • J.R. Olsson et al.

    U–Pb baddeleyite ages linking major Archean dyke swarms to volcanic-rift forming events in the Kaapvaal craton (South Africa), and a precice age for the Bushveld Complex

    Precambrian Res.

    (2010)
  • V.R. Pradhan et al.

    Paleomagnetic and geochronological studies of the mafic dyke swarms of Bundelkhand craton, central India: Implications for the rectonic evolution and paleogeographic reconstructions

    Precambrian Res.

    (2012)
  • W.U. Reimold et al.

    The Anna's Rust Sheet and related gabbroic intrusions in the Vredefort Dome-Kibaran magmatic event on the Kaapvaal Craton and beyond?

    J. Afr. Earth Sci.

    (2000)
  • S.J. Singletary et al.

    Geochronology of basement rocks in the Kalahari Desert, Botswana, and implications for regional Proterozoic tectonics

    Precambrian Res.

    (2003)
  • U. Söderlund et al.

    Towards a complete magmatic barcode for the Zimbabwe craton: Baddeleyite U–Pb dating of regional dyke swarms and sill complexes

    Precambrian Res.

    (2010)
  • J.S.K. Stacey et al.

    Approximation of terrestrail lead isotope evolution by a two-stage model

    Earth Planet. Sci. Lett.

    (1975)
  • V. Van Schijndel et al.

    Baddeleyite geochronology and geochemistry of mafic cobbles from the Dwyka diamictite: new insights into the sub-Kalahari basement, South Africa

    Lithos

    (2011)
  • J.D. Vervoort et al.

    The magmatic evolution of the Midcontinent rift: new geochronologic and geochemical evidence from felsic magmatism

    Precambrian Res.

    (2007)
  • S.E. Ward et al.

    Guruve and Mutare dykes: preliminary geochemical indication of complex Mesoproterozoic mafic magmatic systems in Zimbabwe

    J. Afr. Earth Sci.

    (2000)
  • L. Xiao et al.

    Distinct mantle sources of low-Ti and high-Ti basalts from the western Emeishan large igneous province, SW China: implications for plume-lithosphere interaction

    Earth Planet. Sci. Lett.

    (2004)
  • H.L. Allsop et al.

    The age of the Umkondo Group, eastern Zimbabwe, and implications for paleomagnetic correlations

    South Afr. J. Geol.

    (1989)
  • N.T. Arndt et al.

    The role of lithospheric mantle in continental flood volcanism: thermal and geochemical constraints

    J. Geophys. Res.

    (1992)
  • G.C. Begg et al.

    Lithospheric, cratonic, and geodynamic setting of Ni–Cu–PGE sulphide deposits

    Econ. Geol.

    (2010)
  • D.S. Bullen

    The Petrology of Proterozoic mafic sills in Southern Africa, School of Earth and Environmental Sciences

    (2005)
  • M.F. Coffin et al.

    Large igneous provinces: crustal structure, dimensions, and external consequences

    Rev. Geophys.

    (1994)
  • B. Corner et al.

    Unearthing the Xade complex, Botswana

  • Cited by (63)

    View all citing articles on Scopus
    View full text