https://doi.org/10.1130/G45294.1 Manuscript received 20 June 2018 Revised manuscript received 18 September 2018 Manuscript accepted 20 September 2018 © 2018 Geological Society of America. For permission to copy, contact [email protected]. Published online 9 October 2018 Direct Mesoproterozoic connection of the Congo and Kalahari cratons in proto-Africa: Strange attractors across supercontinental cycles Johanna Salminen1, Richard Hanson2, David A.D. Evans3, Zheng Gong3, Tierney Larson3, Olivia Walker3, Ashley Gumsley4,5, Ulf Söderlund4,6, and Richard Ernst7, 8 1Department of Geosciences and Geography, University of Helsinki, 00014 University of Helsinki, Finland 2Department of Geological Sciences, Texas Christian University, Fort Worth, Texas 76129, USA 3Department of Geology & Geophysics, Yale University, New Haven, Connecticut 06520, USA 4Department of Geology, Lund University, Lund 223 62, Sweden 5Institute of Geophysics, Polish Academy of Sciences, Warsaw 01-452, Poland 6Department of Geosciences, Swedish Museum of Natural History, Stockholm 11418, Sweden 7Department of Earth Sciences, Carleton University, Ottawa, ON K1S 5B6, Canada 8Faculty of Geology and Geography, Tomsk State University, Tomsk 634050, Russia ABSTRACT closure of a wide ocean basin (e.g., Goscombe Mobilistic plate-tectonic interpretation of Precambrian orogens requires that two con- et al., 2018) or at most a narrow oceanic seaway joined crustal blocks may derive from distant portions of the globe. Nonetheless, many pro- (Miller, 1983; Hanson, 2003). One argument posed Precambrian cratonic juxtapositions are broadly similar to those of younger times (so- in favor of the latter hypothesis comes from called “strange attractors”), raising the specter of bias in their construction. We evaluated the recent U-Pb dating of the Huila gabbronorite possibility that the Congo and Kalahari cratons (Africa) were joined together prior to their dikes (Congo craton), which provided a tem- amalgamation along the Damara-Lufilian-Zambezi orogen in Cambrian time by studying poral match with the extensive Umkondo LIP diabase dikes of the Huila-Epembe swarm and sills in the southern part of the Congo craton (Kalahari craton) at ca. 1110 Ma (Ernst et al., in Angola and in Namibia. We present geologic, U-Pb geochronologic, and paleomagnetic 2013). Nonetheless, evidence in the Damara belt evidence showing that these two cratons were directly juxtaposed at ca. 1.1 Ga, but in a of full-fledged rifts and passive margins, and an slightly modified relative orientation compared to today. Recurring persistence in cratonic accretionary prism and orogenic foreland basin connections, with slight variations from one supercontinent to the next, may signify a style of (Hoffman and Halverson, 2008; Miller, 2008), supercontinental transition similar to the northward motion of Gondwana fragments across testify to ocean opening and closure between the Tethys-Indian oceanic tract, reuniting in Eurasia. 750 and 500 Ma. Can these observations be rec- onciled in a single tectonic model? INTRODUCTION such examples “strange attractors,” he suggested Efforts to reconstruct the paleogeographies the possibility that implicit bias might favor METHODS AND RESULTS of Precambrian landmasses have intensified in familiar reconstruction models over more exotic In order to constrain the relative positions of recent decades (summarized by Evans, 2013), possibilities. An alternative explanation might the Congo and Kalahari cratons, we conducted a although there is no consensus on the arrange- be lack of plate-tectonic mobility for most of paleomagnetic and U-Pb geochronologic inves- ments of cratons in supercontinents such as Precambrian time (e.g., Stern, 2005), but mobil- tigation of the ca. 1.1 Ga Huila-Epembe (HE) Rodinia (ca. 900–700 Ma) and Nuna (ca. 1600– istic interpretations are bolstered by many lines dikes and related mafic rocks across the Congo 1400 Ma). Methods of cratonic reconstruction of evidence. This evidence includes paleomag- craton in Angola and northernmost Namibia include tectonostratigraphic comparisons, paleo- netic support for the independent lateral motion (Fig. 1), the results of which can be compared magnetism, and matching fragments of large of lithospheric blocks, magmatic arc activity and to the well-established paleomagnetic pole from igneous provinces (LIPs) on now-separated associated ore deposits related to subduction of the coeval Umkondo LIP from the Kalahari crustal blocks (Bleeker and Ernst, 2006; Li et al., oceanic-type lithosphere, and well-documented (Swanson-Hysell et al., 2015). Methods are fully 2008). Ideally, a favored model would be sup- ophiolites and eclogites back to at least ca. 2.0 described in the GSA Data Repository1. As part ported by several independent lines of evidence. Ga (Cawood et al., 2006; Boniface et al., 2012). of a broader study on numerous dike swarms Meert (2014) noted the frequent occurrence The Ediacaran–Cambrian Damara-Lufilian- exposed in the area, we identified 25 sites (i.e., of hypothesized Precambrian cratonic connec- Zambezi orogen (DLZO) in southern Africa has separate cooling units) in mafic igneous rocks tions that approximate those of Pangea. Calling been interpreted variously to have formed by that preserve either a shallow southwest or a 1 GSA Data Repository item 2018386, methods and detailed results description, geochronology and paleomagnetic data tables, and additional data figures, is avail- able online at http://www.geosociety.org /datarepository /2018/, or on request from [email protected]. CITATION: Salminen, J., et al., 2018, Direct Mesoproterozoic connection of the Congo and Kalahari cratons in proto-Africa: Strange attractors across super- continental cycles: Geology, v. 46, p. 1011–1014, https://doi.org/10.1130/G45294.1 Geological Society of America | GEOLOGY | Volume 46 | Number 11 | www.gsapubs.org 1011 Downloaded from https://pubs.geoscienceworld.org/gsa/geology/article-pdf/46/11/1011/4527165/1011.pdf by Yale University user on 06 November 2018 less-common northeast characteristic magnetic 12°E 13°E S1656 14°E S1655 remanence vector (Fig. 2). Most of these sites S1654 are in the generally north-trending or NNW- W. S1651 AFR. trending Huila dikes, including the dike dated S1650 S1659 Lubango by Ernst et al. (2013), which yielded a U-Pb SF-CONGO S1647 craton baddeleyite age of 1110 ± 3 Ma. We obtained 15°S 15°S study ZO three new isotope dilution–thermal ionization Namibe S1646 HUMPATA area DL PLATEAU KALA. mass spectrometry (ID-TIMS) U-Pb baddeleyite ages (Fig. 3): one from our southernmost region, S1642 1110 ± 3 Ma where dolerite dikes near Epembe trend north- S12EJ west (1109 ± 10 Ma); one from a Huila dike, S12EI S1637, S1638 with a northeast-directed magnetic remanence Tombua S1629 Virei vector (ca. 1104 Ma); and one from a mafic sill intruding red beds of the Chela Plateau (1127 ± 8 Ma). The Chela sill yielded a virtual geo- S12EM, S12EN 16°S 16°S magnetic pole that is similar to, but predates, CONGO our new ca. 1.1 Ga paleomagnetic pole from CRATON the HE dikes; the latter result (34.7°S, 256.5°E, CHELA K = 12.1, A = 8.7°) was demonstrated to be PLATEAU 95 KAOKO primary via a positive baked-contact test on a S1610 100-m-wide Epembe dike intruded into host S1609 gneiss (Fig. DR3 in the Data Repository). In the S1608 1127 ± 8 Ma Iona baked-contact test, host rocks within 1–20 m of Epupa the contact show complete remagnetization at S1607 17°S 17°S Falls the time of dike intrusion, whereas host rocks ANGOLA S1602 33–130 m from the contact exhibit a distinct JS1507, JS1508S1604 S1605 S1606 stable magnetic remanence direction. Addi- BEL NAMIBIA ZEBRA MTS. S1601 ca. 1104 Ma tional baked-contact tests were inconclusive Epembe Ruacana W1516 D1601-03, L1405-07, D1607 (Fig. DR4). Sites in proximity to the Kaoko T W1517 D1606 belt, which is contiguous with the Damara belt W1518 D1604, D1605 1109± 10 Ma (Fig. 1), bear distinct WSW or northwest down- 50 km ward magnetic remanence directions, which we interpret as Ediacaran–Cambrian overprints by 18°S 18°S 12°E 13°E Opuwo 14°E comparison to the Congo craton’s apparent polar wander path from that time interval (Rapalini, Figure 1. Site locality map with Google Earth™ background. Eastern limit of Kaoko belt 2018). The new HE pole fulfills all seven of the (Namibia) allochthons is from Goscombe et al. (2018). Inset abbreviations: DLZO— Damara-Lufilian-Zambezi orogen, Kala—Kalahari craton, SF—São Francisco craton, Van der Voo (1990) reliability criteria and can be W.Afr.—West African craton. U-Pb baddeleyite age for site S1642 is from Ernst et al. called a paleomagnetic key pole (Buchan, 2014). (2013); other baddeleyite ages farther south are from this work. PALEOGEOGRAPHIC IMPLICATIONS The HE paleomagnetic pole yields a rela- N tively low latitude for the Congo craton at ca. 1.1 Ga and permits a direct connection between the Congo and Kalahari cratons (Fig. 4B), the S1601* latter of which remained in tropical low latitudes throughout late Mesoproterozoic time (Gose et S1605 al., 2013). Data presented by Ernst et al. (2013) and de Kock et al. (2014) showed that the Huila Figure 2. Equal-area pro- dikes are fractionated continental tholeiites with jection of paleomagnetic compositions similar to much of the Umkondo results from studied mafic LIP, and those workers hypothesized that the S1659 L14E06 intrusions in Angola and dikes represent a radial arm of the LIP emanat- W S12EJ E Namibia (this study). Geo- sat chronologic data for the ing from a focus in the Kalahari craton. Like the JS1507 S1608* dated dikes are listed in majority of published Umkondo paleomagnetic S1637 L14E05+D1607 Upper hemisphere Tables DR1 and DR2 (see * W1518* data, the HE remanence directions are south- S1638, S12EK S1604 D1603 Lower hemisphere footnote 1). erly and upward, favoring the same-hemisphere S1642 L14E07 D1604 W1516S1606 * Dated dike polarity option for those two suites as shown in Figure 4B. Following Swanson-Hysell et al. S1655 D1606 S1646 S1602 (2015), we propose that south-seeking mag- W1517 S12EI netic remanences represent normal polarity for S1650 the Congo and Kalahari data; this allows the S1647 ca.