GEOSCIENCE CANADA Volume 31 Number 3 September 2004 103 SERIES have been linked to regional-scale uplift, Dans l’Archéen, les meilleurs candidats continental rifting and breakup, and sont représentés par les bandes de climatic crises. They can be used as roches vertes à komatiites. De precisely dated time markers in the nombreuses GPI ont été associées à des stratigraphic record, and are key targets épisodes de soulèvement régionaux, de for Ni-Cu-PGE exploration. LIPs have dérives ou de fragmentations also become a focus in the debate on the continentales, ainsi qu’à des crises existence and nature of mantle plumes. climatiques. Elles peuvent servir de Canada has a rich record of marqueurs temporels stratigraphiques et LIPs. At least 80 candidates are recog- sont des cibles de première importance nized in Canada and adjacent regions, dans l’exploration de gisements de Cu- with ages ranging from 3100 to 17 Ma. Ni-ÉGP. Les GPI sont aussi devenues Igneous Rock Associa- We review proposed links between the des arguments très considérés dans le tions in Canada 3. LIP record of Canada and mantle débat sur l’existence et la nature des plumes, continental breakup, regional panaches mantelliques. Large Igneous Provinces uplift, and ore deposits. However, given Le Canada possède de riches (LIPs) in Canada and that many mafic units in Canada remain archives de GPI, et au moins 80 candi- Adjacent Regions: 3 Ga poorly characterized, a concerted datures ont été isolées sur le territoire geochronology campaign with integrated canadien et dans les régions adjacentes, to Present paleomagnetism and geochemistry would leur âge délimitant une fourchette allant be invaluable in expanding the applica- de 3 100 Ma à 17 Ma. Nous passons R.E. Ernst and K.L. Buchan tion of the Canadian LIP record to en revue les liens proposés entre la suite Geological Survey of Canada solving major geological problems. des GDI canadiennes d’une part, et celle 601 Booth Street des panaches mantelliques, des Ottawa, Ontario K1A 0E8 RÉSUMÉ fragmentations continentales, des [email protected] L’histoire de la Terre est ponctuée de soulèvements régionaux, et des nombreuses périodes de mise en place gisements minéraux, d’autre part. SUMMARY de forts volumes de magma mafiques. Toutefois, vu le piètre état de Earth history is punctuated by numerous De tels magmas qui ne sont pas issus de caractérisation des unités mafiques au periods during which large volumes of zones d’expansion « normale » ou de Canada, une campagne de mafic magma were emplaced. Such subduction sont appelés Grandes prov- caractérisation géochronologique, magmas not generated by a ‘normal’ inces ignées (GPI), et celles-ci sont paléomagnétique et géochimique serait spreading ridge or by subduction are constituées de basaltes d’épanchements d’une valeur inestimable pour favoriser termed Large Igneous Provinces (LIPs), continentaux, de marges de fosse l’utilisation des GDI canadiennes pour and consist of continental flood basalts, volcaniques, de plateaux océaniques, nous aider à solutionner de grands volcanic rifted margins, oceanic pla- d’épanchements de basaltes de bassins problèmes géologiques. teaus, ocean basin flood basalts, subma- océaniques, de crêtes sous-marines, et rine ridges, and seamount chains. de chaînes de monts sous-marines. INTRODUCTION Associated felsic rocks may also be Peuvent également y être associées des Large Igneous Provinces (LIPs) repre- present. LIPs of Mesozoic and Cenozoic suites de roches felsiques. Générale- sent voluminous magmatic events that age are typically the best preserved. ment, les GPI du Mésozoïque et du were not generated by a ‘normal’ spread- Those of Paleozoic and Proterozoic age Cénozoïque sont les mieux préservées. ing ridge or by subduction (Coffin and are usually more deeply eroded, and Celles du Protérozoïque et du Eldholm, 1994; 2001; Ernst et al., consist of flood basalt remnants and a Paléozoïque sont généralement plus 2004). They may be emplaced as often deep-level plumbing system (of giant fortement érodées et sont constituées de as once every 10 Ma (e.g. Coffin and dyke swarms, sill provinces and layered vestiges de basaltes d’épanchement et Eldholm, 2001), and time series analysis intrusions). In the Archean the most des réseaux de conduits d’origine of the LIP record for the past 3.5 Ga promising LIP candidates are greenstone (réseaux géants de dykes, provinces de suggests weak cyclicity (Isley and belts containing komatiites. Many LIPs filons-couches et d’intrusifs stratifiées). Abbott, 2002; Prokoph et al., 2004). 104 The most dramatic LIPs are emplaced The extrapolation of the LIP The compilation includes rapidly (within <10 Ma and often within record into the Archean is more specu- information on tectonic setting. Our only a few Ma). These include continen- lative. There are erosional remnants of criteria for determining setting rely tal flood basalts, seaward-dipping typical flood basalt provinces, namely heavily on dyke swarm geometry and its reflector sequences, oceanic plateaus, the Fortescue sequence of the Pilbara relationship to cratonic margins and ocean basin flood basalts. Conti- craton of Australia and the Ventersdorp (Fig. 3, Table 3). Events are inferred to nental flood basalts can be as large as sequence of the Kaapvaal craton of have a mantle plume origin if a giant several million cubic km (e.g. the southern Africa (Eriksson et al., 2002). radiating dyke swarm is present. Giant Siberian Traps; Reichow et al., 2002). However, most Archean volcanic rocks linear dyke swarms that extend into a The largest LIP is the Ontong Java occur as deformed and fault-fragmented craton (i.e. trend perpendicular to a oceanic plateau, which has a volume of packages termed greenstone belts. cratonic margin) are inferred to repre- 44.4 million cubic km (Coffin and Among these, the best candidates for sent an aulacogen-type swarm (‘failed- Eldholm, 2001) for combined extrusive LIPs are thick tholeiite sequences that arm’ type in Fahrig, 1987), and can also (6 million cubic km) and intrusive contain komatiites. The nature of be used to infer a plume origin with the components (Courtillot and Renne, Archean LIPs is discussed in greater plume centre situated at the edge of a 2003). The initial large-volume short- detail below. craton. By contrast, linear swarms that duration stage of magmatism of some LIPs are important 1) for testing parallel the edge of a craton may simply LIPs has been linked to the arrival of a plume and non-plume models for the be rift/breakup related (Ernst and mantle plume (e.g. White and generation of LIPs; 2) as precise time Buchan, 1997) or may possibly repre- McKenzie, 1989; Campbell and markers for stratigraphic correlations; sent a back arc rifting setting (e.g. Griffiths, 1990; Coffin and Eldholm, 3) as an aid in reconstructing conti- Rivers and Corrigan, 2000), or overrid- 1994; 2001; Campbell, 1998, 2001; nents; and 4) as the hosts of major PGE ing of a spreading ridge (Gower and Ernst and Buchan, 2001; Courtillot et deposits and as a potential tool in Krogh, 2002). al., 2003). Subsequent rifting/breakup diamond exploration. In addition, they In addition (Table 3), those is often associated with a second burst can be helpful in studying 5) climatic Archean greenstone belts containing of volcanism (Campbell, 1998) by effects, and 6) regional uplift. We return komatiites are inferred to be plume- decompression melting (White and to these topics after a review of the LIP related on the basis of the elevated McKenzie, 1989). In addition, LIP record of Canada and adjacent regions. temperatures required for generation of magmatism can continue for prolonged komatiites (e.g. Campbell, 1998, 2001; periods after the initial outburst (or Arndt et al., 1998; Condie, 2001). outbursts), in the form of seamount PRELIMINARY LIP HISTORY OF Finally, small, intraplate events not chains and ridges, which are usually CANADA AND ADJACENT REGIONS obviously linked to a cratonic boundary explained as hotspot tracks associated Methodology are categorized as ‘hotspots’. with a plume tail. Other models invoke Our compilation is based on a recent Below (and in Tables 1 and 2) we plate fracturing and ‘edge convection’ summary of the global LIP distribution summarize the main events, their age (upper mantle convection between thick (Ernst and Buchan, 2001) and a newly distribution and tectonic setting. It and adjacent thin lithosphere), and have published compilation of dyke swarms should be noted that referencing has been suggested as an alternative to and related magmatic units in Canada been minimized in the text below plume models for LIPs and hotspot and adjacent regions (Buchan and Ernst, because detailed referencing is available chains (e.g. Anderson, 2001; Foulger 2004). We currently recognize at least through Tables 1 and 2. Also note that and Natland, 2003). 80 LIPs and possible LIP remnants in we have included number-labels of the The volcanic portion of older and adjacent to Canada. The form [#14a] in order to facilitate easy continental LIPs is largely removed by Proterozoic and Phanerozoic mafic cross-correlation with entries in Tables 1 erosion and deformed during continen- magmatic record is reviewed first since and 2, and with the distribution of main tal collision, whereas older oceanic LIPs its links with LIPs are better defined units in Figure 1. are mostly lost during subduction and (Table 1, Fig. 1). The more speculative deformed during ocean closure. There- Archean LIP history follows (Table 2, Proterozoic to Present fore, in the Paleozoic and Proterozoic Fig. 1). 2.51–2.41 Ga: The earliest Proterozoic record, continental LIPs are typically The Proterozoic record relies LIPs consist of dykes, layered intrusions recognized by their exposed plumbing heavily on diabase dyke swarms and sills and volcanic rocks and are mainly system of giant dyke swarms, sill (Fig. 2). Dykes injected laterally into associated with the eastern and southern provinces, large layered intrusions, and the interior of continents have a preser- margin of Laurentia.