Improving Multiscale Structural Targeting Precision for Large Ni-Cu (±PGE, Co) Sulfide Discoveries from Global Case Studies

To sustain the clean energy transition, society needs to increase the rate of discovery and development of new green and critical mineral ore deposits containing metals such as copper (Cu), lithium (Li), nickel (Ni), cobalt (Co), rare earth elements (REEs), and platinum group elements (PGEs). The global resource inventory for Ni and Cu has been sustained by adding resources of lower quality and/or increased development risk, more commonly by resource expansions at lower cutoff grades than by new discoveries. Discovery rates for new polymetallic Ni sulfide deposits have declined over the past 25 years while unit discovery costs have sharply increased (Schodde, 2017). One challenge for discovery is the low occurrence frequency of large Ni ore deposits. Many well-endowed provinces have yielded only one or rarely two mineable deposits, and most new Ni sulfide discoveries were found in provinces that had no previously known Ni sulfide resources or mines (e.g., Julimar, Nova-Bollinger, Enterprise, Xiarihamu, Kingash, and West Musgraves). These factors favor a first-mover Ni exploration strategy. To support such a strategy, the minerals industry needs mineral system models that more accurately predict the location of new districts (camps) with large high-quality deposits in search spaces that were historically unrecognized or difficult to access and test. Improved prediction accuracy is especially required where traditional belt-scale detection methods (e.g., ground-based geophysical and geochemical surveys) are challenged by deeper and less penetrable cover, the very small footprint of Ni deposits, and by challenges for land access arising from greater environmental, social, and governance restrictions.

To help improve greenfield discovery rates, this study aims to address two key questions:

1. What is the role of lithospheric faults in localizing the ascent and emplacement of Ni-Cu ore-forming magmas?

2. How can targeting translithospheric faults improve the accuracy and precision of area selection for large new greenfield discoveries?

Although the ascent of mafic to ultramafic magmas through the lithosphere is widely interpreted to be controlled by networked dike-sill complexes (Magee et al., 2016; Cruden and Weinberg, 2018; Lesher, 2019), few magmatic models explicitly invoke a role for translithospheric faults in focusing magma ascent and emplacement (e.g., Barnes et al., 2016). Furthermore, the degree of lateral transport of fertile magmas through dike-sill complexes and implications for the lateral spread of magmatic Ni-Cu deposits are uncertain. Opinions range from deposits forming within regional strike-slip faults at craton margins (Lightfoot and Evans-Lamswood, 2015) to forming potentially hundreds of kilometers distal to transcrustal fault conduits and plume impact centers due to long-distance lateral transport of magmas through lava sheets, interconnected sill and dike complexes, or plume-related radial dike swarms (Ernst et al., 2019; Lesher, 2019).

This study tests the efficacy of lithospheric fault targeting at three common exploration scales:

1. Subprovince to corridor scale (10⁴–10⁵ km²), comprising the 100-km-long prospective belts along craton margins within much broader large igneous provinces;

2. District or camp scale (10²–10³ km²), comprising discrete clusters of prospective mafic-ultramafic intrusions or volcanic units, deposits, and prospects located along corridors;

3. Prospect scale (10¹ km²), comprising individual mineralized intrusions to deposits.

At the subprovince scale, Begg et al. (2018) provided the first quantitative analysis of the proximity of magmatic Ni-Cu sulfide deposits to cratonic boundaries mapped in the upper mantle. From 67 deposits studied they observed that 29 occur ≤25 km from a craton margin, 26 occur within 120 km of craton margins, and seven occur on margins of microcontinental blocks in Proterozoic or Phanerozoic orogenic belts. At the prospect scale, Lightfoot and Evans-Lamswood (2015) speculated that emplacement of some mineralized intrusions (Voisey’s Bay and several deposits in China) was controlled by local releasing bends and jogs along transtensional strike-slip faults, which they suggested directly controlled the ascent of dense sulfide-laden magmas by seismic pumping. Few studies, however, have addressed fault controls on ore localization along structural corridors at the district (camp) scale, and fewer still have attempted to quantify these spatial relationships in terms of endowment potential. The two best documented examples of a transverse translithospheric fault association include Voisey’s Bay (Myers et al., 2008) and West Musgraves (Smithies et al., 2015); however, there are very few published maps that demonstrate this relationship for other large Ni ore deposits. The lack of case studies is a critical knowledge gap for explorationists. Targeting district-scale deposit clusters in first-mover exploration campaigns has advantages over targeting individual deposits, because the district target footprint is larger and the opportunity of discovering and controlling a new world-class district that can be mined from a central operation offers the biggest reward for explorers and their stakeholders.

This study reports results from 72 district-scale case studies conducted globally that provide new and original map interpretations and insights. It is the first to quantitatively analyze structural controls on Ni-Cu (±PGE, Co) ore localization in terms of endowment at the district scale. Implications for exploration targeting and models for the transport and emplacement of sulfide-bearing magmas are also discussed. To set the context, the range of deposit types and settings studied here are first briefly reviewed.

Magmatic Ni-Cu (±PGE, Co) sulfide deposit subtypes can be classified based on parental magma MgO content (Fig. 1; Barnes et al., 2016; Begg et al., 2018), forming a continuum between end-member compositions of high-MgO (18–30% MgO, komatiitic) to low-MgO (5–12% MgO, tholeiitic basalt) that brackets medium-MgO compositions (12–18% MgO; ferropicrite, komatiitic basalt). This compositional range is associated with three orders of magnitude difference in magma viscosity, which directly influences magma ascent rates, emplacement styles, and the degree of assimilation and fractional crystallization; and indirectly influences mineralization styles, commodity mix, and metal tenor. High-MgO komatiitic magmas have very low viscosity (0.1–1.0 Pascal-second [Pa·s], like olive oil; Lesher, 2019) and ascended rapidly to very shallow crustal environments with little fractionation where they formed sulfide-undersaturated extrusive lava sheets and subvolcanic sills. Because the degree of sulfide undersaturation increases during magma ascent (due to the inverse relationship between S solubility and pressure [Mavrogenes and O’Neill, 1999]), komatiite flows only attained sulfide saturation by local country rock sulfur assimilation to produce Ni (±PGE) sulfide deposits with little Cu (e.g., Barnes et al., 2016). In contrast, low-MgO mafic magmas have one to two orders of magnitude higher viscosity (5–100 Pa·s for tholeiitic basalt; Lesher, 2019) and were emplaced predominantly at mid- to upper-crust depths as sills, dikes, and plutons. They exhibit greater degrees of assimilation and fractional crystallization, so are more contaminated and hydrous and commonly include variably textured to pegmatoidal taxite units at intrusion margins. Deposits from low-MgO suites tend to be more Cu rich and include supergiant districts like Norilsk-Talnakh. Intermediate- to high-MgO (picritic) parental magmas yielded deposit types with characteristics transitional between these two end members (e.g., Raglan, Pechenga; Barnes et al., 2001). Low-MgO parental magmas in continental arc settings have higher H₂O and S contents and typically host smaller, lower tenor, PGE-depleted Ni-Cu-Co sulfide deposits (e.g., Huangshan), whereas low-MgO parental magmas emplaced into intracratonic settings with thick lithosphere (Fig. 1) tend to be sulfide poor (0.5–5%) and host disseminated PGE (±Ni, Cu) sulfide reefs in large oxide-rich mafic layered intrusions (e.g., Bushveld) or mega dikes (e.g., Great Dyke). PGE-dominated magmatic deposits and the supergiant Sudbury deposits (the result of an exceptional bolide impact) were excluded in this study. Sulfide mineralization in both extrusive and intrusive units typically accumulated along subhorizontal floors in channelled flow regimes comprising trough-like lava channels, tube-like channels in sills described as chonoliths, or blade-shaped keels in dikes. Gravity-driven infiltration and melting of floor rocks by hot cumulus sulfide pools commonly also yielded footwall sulfide lenses, veins, and breccias (Barnes et al., 2016).

Many magmatic Ni-Cu (±PGE, Co) sulfide deposits are associated with the roots of mafic large igneous provinces (LIPs) derived from mantle plume impacts. Plumes melt by decompression (adiabatic) at shallow mantle depths (mostly <5 GPa, <150 km) to yield voluminous, S-undersaturated, mafic-ultramafic magmas. They only produce large melt volumes where they impact or are deflected to areas with a shallow lithosphere-asthenosphere boundary (Fig. 1; Niu, 2021). Plume impacts have three components: plume head, plume tail, and plume (head) deflection. Plume head impact centers may be centered among circumferential dike swarms where radial dikes converge (Ernst et al., 2019). Plume tails may contribute to komatiitic magmatism after mafic magmatism (Barnes et al., 2016). Where a plume head deflects to shallow lithosphere-asthenosphere boundary depths (<120 km) at a craton margin, it undergoes high-degree partial decompression melting at lower pressures to yield high volumes of more primitive (higher-MgO) magmas (Sleep, 1997). This scenario favors formation of magmatic Ni-Cu deposits at or close to craton margins. Favorable craton margins include the edges of older intracontinental paleocratons sutured together to form larger cratons and supercontinents, particularly those bordering cratons with preserved thick, strong, and relatively buoyant Archean subcontinental lithospheric mantle (SCLM; Begg et al., 2010). Magma fertility at paleocraton margins may be enhanced by assimilation of SCLM that was refertilized by prior subduction-related metasomatism and locally altered to phlogopite, pyroxene, sulfide, and carbonate assemblages with variable Ni, Cu, PGEs, sulfur, and volatile enrichment (Ezad et al., 2024).

LIPs associated with supercontinent breakup make up more than half of the global LIP inventory yet have little significant known Ni-Cu endowment. Examples include the Karoo LIP in southern Africa and the Paraná-Etendeka LIP in eastern South America and West Africa. This observation seems paradoxical given that many Ni sulfide deposits, especially komatiitic Ni deposits, formed in intracontinental rifts, pericratonic marginal rift basins, and intra-/back-arc rifts. However, the mineralized rift settings were localized, failed early, and did not lead to continental breakup. Most magmatic Ni-Cu deposits formed within LIPs emplaced during supercontinent assembly in convergent tectonic settings rather than the main supercontinent breakup stages (Begg et al., 2018).

The case studies reported here cover the global inventory of magmatic Ni-Cu sulfide deposits that contained premine resources >50 kt Ni metal. Three levels of detail are reported: (1) detailed case studies for six districts that host giant deposits, (2) summary maps and table for another 11 districts, and (3) charted statistics for the rest (App. 1). The deposit size classifications (supergiant, giant, major, moderate) follow the schema of Schodde (2017) for total premine Ni-equivalent metal tonnes. All deposit locations were verified with Google Earth satellite imagery.

Primary data types used for the new structure interpretations included regional geology, global and national aeromagnetic imagery (reduction to pole, first vertical, horizontal, and vertical integral derivatives), global satellite gravity and national ground gravity imagery (isostatic residual gravity and total horizontal gradient derivatives), shuttle radar topography, and landsat imagery. Most of the primary data sets for these map layers are publicly available for little to no cost. Verification and refinement of structure interpretations at district scale used georeferenced published geology maps.

For each case study, upper crustal lithospheric domain boundaries, paleocraton edges, and regional faults were mapped at scales of 1:1,000,000 to 1:200,000. Three common approaches to delineating paleocraton edges are: (1) as upper lithospheric domain boundaries mapped at nominal 80 to 120 km depth using multidisciplinary data (Begg et al., 2010), (2) as lithosphere-asthenosphere boundary steps mapped at nominal 170 km depth using seismic tomographic data (Hoggard et al., 2020), and (3) as associated translithospheric faults and suture zones mapped in the upper crust from multidisciplinary data (this study). Mapping upper crustal translithospheric fault positions is important where translithospheric faults are inclined, potentially yielding a shallow crustal position that is offset laterally up to several tens of kilometers from the upper mantle position. Whereas lithospheric-scale faults may be mapped as 1- to 20-km-wide brittle-ductile shear zones, for this study a fault zone center line was delineated for representation clarity and spatial analyses. Three classes of lithospheric-scale faults are defined: (1) primary edge faults at paleocraton edges, (2) secondary edge-parallel faults located near and parallel to paleocraton edges, and (3) transverse faults that strike obliquely to paleocraton edges. Shallow-dipping regional thrusts and detachment faults were omitted because they are typically not translithospheric. Although the depth extents of faults are uncertain, for simplicity all moderate- to steep-dipping craton-scale faults (nominally >200-km long) are classified as translithospheric faults, (sensu lato) on the assumption that they penetrate the crust to reach SCLM depths, which is very likely for faults positioned along lithospheric domain boundaries. However, some translithospheric faults might only extend to the Moho or terminate in listric ductile lower crust shear zones (transcrustal, sensu stricto).

Many transverse translithospheric faults mapped in this study have cryptic expression and are not widely recognized or previously documented. Cryptic faults and lineaments are likely inherited from older basement shear zones that were buried under younger cover sequences, and which may have been only partially propagated vertically through cover as weak fracture zones or inflections. To address the uncertainty in mapping cryptic translithospheric faults, a simple qualitative confidence rating of high (>5 layers of evidence including geology and widely recognized), medium (2–5 layers of evidence), and low (1–2 layers of evidence excluding geology) was applied to each fault interpretation.

The minimum distance between a fault zone center line and the largest deposit in each district was measured in the ArcPro geographic information system. Given that a center line approximates a fault zone, reported distance-to-fault measurements commonly have a precision around ±3 km, rarely up to ±10 km. Statistical analyses of the spatial relationships between translithospheric faults and Ni-Cu deposits in this study considered fault accuracy, precision, and causality. Accuracy is related to the true positive and true negative capture rates and quantified here as the spatial proximity of known deposits to target faults or fault intersections (true positives). True negatives generally cannot be assessed due to the absence of effective drill testing in many prospective areas (i.e., high false negative rate). Precision is inversely related to the false positive rate and was assessed here qualitatively. Because total deposit areas are very small (10¹ km²) relative to a permissive subprovince search area (10⁵ km²), the false positive rate is approximately equivalent to the permissive search space selected by the buffered extents of targeted structures, most of which will be barren. Causality assessment considers evidence for a genetic link between deposit emplacement and a possible feeder fault, such as whether the target structure could have been physically linked to an active magma conduit or staging chamber at the time of mineralization, and whether the district, host intrusion, and/or deposit show geometrical alignment and/or zonation vectors consistent with being derived from an inferred feeder structure.

The following section presents six case studies on the relationship between translithospheric faults and giant Ni deposits, with five from intracontinental settings and one from a pericratonic setting.

The giant Jinchuan Ni-Cu-Co deposit in China occurs in the 600-km-long, ~50-km-wide Longshoushan belt along the southwest margin of the Alxa block, which is a reworked paleocraton joined to the western North China craton (Fig. 2). Sparse Neoproterozoic mafic-ultramafic intrusions found throughout the Longshoushan belt may be correlated with remnants of the ~825 Ma Guibei LIP in the northern Qaidam and Yangtze cratons (Wang et al., 2007). The Jinchuan ultramafic intrusion was emplaced at 827 ± 8 Ma (Li et al., 2005) proximal to the NW-striking Longshoushan fault, which likely originated as a Paleoproterozoic suture during the 2.0 to 1.95 Ga Longshoushan orogeny (Porter, 2016) and was reworked several times since. The craton margin rifted apart postmineralization during Neoproterozoic Rodinia breakup and subsequently collided with another microcraton in the Silurian (430–422 Ma) to form the adjacent Qilian orogen. The Longshoushan belt was further deformed during opening and closure of an early Paleozoic marine back-arc basin (Song et al., 2013), then later exhumed with >4 km of Cenozoic uplift by inversion of inward dipping (60°–70°) reverse faults to the northeast and southwest. The Jinchuan intrusion was originally interpreted to be a single subvertical funnel-shaped dike (Chai and Naldrett, 1992), which implied the steeply dipping ore lenses were deposited in steep-dipping conduits. However, Lehmann et al. (2007) demonstrated that the intrusion was emplaced as a shallow dipping and semiconcordant sill complex among marbles and paragneiss before tilting to a steep dip with N-verging thrusting. Song et al. (2012) subsequently interpreted the Jinchuan intrusion complex to comprise two adjacent sill-like intrusions separated by the EW F16-1 fault.

New mapping of regional faults in the Alxa block (Fig. 2) reveals that Jinchuan occurs in the hanging wall ~11 km north of the Longshoushan fault, very close (<1 km) to intersections of both NE- and EW-striking transverse translithospheric faults. The northeast lineament (here named the Yabulaishan lineament) is a cryptic, craton-scale translithospheric fault delineated with medium confidence by regional strike changes in outcrop geology, which switch from predominantly west-northwest across the western Alxa block to north-northeast on the east side (Fig. 2A), and truncations to regional gravity and magnetic features (Fig. 2B). At Jinchuan, the Yabulaishan lineament is exposed as a NE-striking reverse fault at the eastern end of the Jinchuan intrusive complex, which may downthrow extensions of the intrusive complex beneath cover to the east. Another mapped extension of this fault traverses the north Qilian orogen to the south of Jinchuan (white pointer in Fig. 2A), suggesting further postmineralization fault growth.

The E-W-striking Cha-Gu fault has ~15 km of apparent sinistral displacement. It is part of a regional set of east-west reverse-sinistral faults mapped with high confidence from outcrops that resulted from fault reactivation during Neogene to Quaternary compression related to the distal Indo-Eurasia collision (Fig. 2A; Zhang et al., 2009). The combined displacements of both the Yabulaishan and Cha-Gu faults account for the abrupt juxtaposition of Proterozoic metamorphic rocks of the Longshoushan belt against Paleozoic granitic rocks immediately southeast of Jinchuan. The Jinchuan intrusion is similarly cut by east-west sinistral-reverse faults (F16-1, F8, and F23 in fig. 2 of Song et al., 2012) that are splays of the Cha-Gu fault system. Although bulk displacements along the east-west fault set are clearly young, major east-west faults occur very close to several other magmatic Ni sulfide prospects and mafic-ultramafic intrusion clusters in the region, such as the Neoproterozoic Zhangbutai and Qingshuyao ultramafic complexes northwest of Jinchuan (prospects 1 and 2 in Fig. 2A) and the Permian Yejili mafic-ultramafic complex north of Jinchuan (prospect 3 in Fig. 2A; Barnes and Zhong-Li, 1999). These spatial correlations suggest that the suggest that the EW fault set may have reactivated fault set may have reactivated older basement faults that localized emplacement of the Neoproterozoic and Permian mafic-ultramafic intrusions. The intersection of the northwest Longshoushan edge translithospheric fault with both the northeast Yabulaishan and east-west Cha-Gu transverse translithospheric faults is unique along the Longshoushan belt and has high accuracy, moderate precision, moderate causality, and high confidence as a district-scale targeting element. If the origins of these translithospheric faults can be confirmed to be premineralization, mafic-ultramafic intrusions located near other east-west and/or northeast translithospheric fault intersections along the Longshoushan belt may be prospective for new Ni discoveries.

The giant Kabanga Ni-Cu-Co district in Tanzania is located within a >350-km-long, 20-km-wide, linear belt of Mesoproterozoic mafic-ultramafic intrusions (the Kabanga-Musongati Alignment) along the west edge of the Tanzania craton (Fig. 3). This intrusive suite is part of the 1405 to 1360 Ma Kabanga-Kunene LIP, which also hosts the Lubalisi (née Kapalagulu) Ni-Cu-Co sulfide deposit (Tanzania) and Musongati Ni-Co laterite deposit (Burundi). A plume impact center was interpreted by Mäkitie et al. (2014) to occur ~165 km north-northwest of Kabanga, based on a core of 1380 to 1360 Ma S-type granites partially bordered by the coeval Lake Victoria dike swarm, which they suggested has a circumferential distribution (Fig. 3C). Dates of 1403 ± 14 Ma for the Kabanga Gabbronorite and 1392 ± 26 Ma for the Lubalisi Gabbro (Maier et al., 2007) indicate that Ni mineralization was deposited at the start of the Kabanga-Kunene LIP and up to ~25 m.y. before the Lake Victoria LIP plume impact, although the dating error ranges are large and overlapping. Intrusions of the Kabanga-Musongati alignment were emplaced into a failed intracontinental rift among Mesoproterozoic metasediments of the Karagwe-Ankole belt (Fig. 3A), which unconformably overlie a narrow belt of reworked Archean lithosphere between the Tanzania and Congo Archaean cratons to the east and west, respectively. The intrusions closely follow the Kibaran shear zone, which is interpreted in Figure 3 to be the upper crust extension of a buried Archean lithospheric domain boundary (edge translithospheric fault) that dips moderately west-northwest with a center line located ~15 km east of Kabanga.

The Tanzania craton is divided into several W-NW-striking (super) terranes (Fig. 3B; Kabete et al., 2012; Sun et al., 2018). The positions of these terrane boundaries under cover in the west Tanzania craton are remapped in Figure 3A and C from regional magnetic and gravity imagery. The most prominent internal terrane boundary lies between the Lake Nanza superterrane, comprising several Neoarchean granitoid-greenstone belts that host orogenic Au deposits, and the Moyowosi-Manyoni superterrane to the south which comprises deeply exhumed Meso- to Neoarchean granitoid-gneiss (greenstone) and granitoid-migmatite belts devoid of Au deposits (Fig. 3B). This central terrane boundary is one of the most prominent transverse translithospheric faults (here named the Dodoma shear zone), and likely originated during closure of an Archean continental back-arc basin into which the greenstone belt stratigraphy was deposited. The Dodoma shear zone largely terminates at the craton edge (Kibaran shear zone) although subtle fault extensions can be traced further west in the Kabanga mining district (Fig. 3A). Further south, the Musongati Ni laterite deposit is similarly located along trend of another W-NW-striking terrane boundary (Fig. 3). As district targeting elements, the translithospheric faults associated with these west-northwest terrane boundaries in the Tanzania craton have moderate accuracy, high precision, medium confidence, and medium causality ratings.

The west-northwest translithospheric faults appear kinked with up to 20 km sinistral shear along a newly identified NE-striking mega-shear, here named the central Tanzania lineament. This lineament is defined with medium confidence by regional, long-wavelength magnetic, gravity, and geology discontinuities (Fig. 3A–C). It juxtaposes the Dodoma basement block against Cenozoic sediments in the southwest part of the Tanzania craton and bounds the western domain of the Lake Nyanza superterrane, where the most abundant greenstone belts and Au deposits are preserved, indicating a component of bulk W-block-down throw (dip uncertain). This transverse translithospheric fault is closely associated with the Lubalisi Ni deposit, where it intersects the Rukwa shear zone along the southwest edge of the Tanzania craton, and the Dutwa Ni deposit, where it intersects the boundary between the East Lake Victoria and Lake Nyanza superterranes (Fig. 3B, C). Accordingly, the subset of mafic-ultramafic intrusions that are located near the most prominent west-northwest and west-northwest translithospheric fault intersections with Tanzania craton terrane boundaries are predicted to be the most prospective.

The Norilsk-Talnakh district is located near the northwest corner of the Siberian craton in Russia and comprises two supergiant Ni-Cu PGE Co sulfide deposits ~25 km apart (Fig. 4). Mineralization here was deposited at 251 ±0.5 Ma (Burgess and Bowring, 2015) in differentiated gabbro-dolerite to leucogabbro chonolithic sills emplaced into the Norilsk-Kharaelakh trough; a deep, N-NE–striking graben filled with thick Neoproterozoic to Permian sediments and overlain by up to 3.5 km of mafic volcanics of the Permo-Triassic Siberian traps. The trough sequence unconformably overlies the Angara orogenic belt that wraps around the west margin of the Siberian craton, comprising composite Paleoproterozoic terranes of microcratonic blocks, island arcs, and ophiolites that form a distinct gravity high (Fig. 4A, C). The overlying Siberian traps comprise one of the world’s most voluminous and short-lived LIP events (251–252 Ma), which triggered the world’s most severe Phanerozoic mass extinction with extremely rapid emplacement of high volumes of mafic sills among sulfate and carbonate-rich sediments, which led to massive outgassing of greenhouse gases (SO₂, CO₂, CH₄) and forced catastrophic climate change (Burgess and Bowring, 2015; Le Vaillant et al., 2017). Ore deposition coincided with an ~90º change in the general strike of volcanic thickness contours and dolerite dike swarms along the Norilsk-Kharaelakh trough, from (1) premineralization north-northeast strike (present day coordinates) for faults, dikes, and the Ivakinsky-Gudchikhinsky and Khakanchansky-Nadezhdinsky volcanic assemblages, to (2) dominantly west-northwest to northwest strike for the Ebekhaya and Maimecha dike swarms and early syn- to postmineralization Morongovsky-Samoedsky volcanic assemblages (Fedorenko, 1994). The strike change reflects a 90º switch in the σ3 horizontal stress axes related to either a far-field tectonic change (Begg et al., 2018) or plume-generated transition from radiating to circumferential dolerite dike swarms (Ernst et al., 2024).

The Norilsk-Kharaelakh fault is well documented as an ~500-km-long, narrow, subvertical to steep, W-NW–dipping fault with relatively small, normal displacements (<500 m; Diakov et al., 2002; Krivolutskaya et al., 2018). It is part of a set of N-NE–striking faults that includes the Boganida, Imandga, and Keta-Irbo faults (Diakov et al., 2002; Fig. 4B), which likely originated as Paleoproterozoic (Angaran) basement faults that were reactivated during both Permo-Triassic west-northwest to east-southeast (premineralization) and northeast-southwest (synmineralization) extension events. Begg et al. (2018) speculated that the Norilsk-Kharaelakh fault may have originated as a suture (edge translithospheric fault) between the Siberian craton and a microcontinental block to the west although supporting evidence was not detailed. Based on new interpretations of the lithospheric architecture (Fig. 4B, C), the Imangda-Letniya fault to the east appears as a well-defined lithospheric domain boundary with sharp geophysical gradients at the boundary between the Angara mobile belt and Siberian craton core, so the Imangda-Letniya fault is defined here as the principal craton edge translithospheric fault. Geophysical models show that both the Norilsk-Kharaelakh and Imangda-Letniya edge translithospheric faults localized midcrustal mafic intrusions and Ni mineralization (Diakov et al., 2002). The Norilsk-Kharaelakh translithospheric fault is widely considered to be the primary feeder for mineralizing intrusions at Norilsk and Talnakh, since these deposits show outward zonation in host mineralogy, thickness, and ore metal ratios from the fault over distances <5 km. Although this fault likely localized feeder dikes at least transiently, no remnants of feeder dikes have been mapped in the fault zone in the district. Overall, definition of the Norilsk-Kharaelakh translithospheric fault has high accuracy, moderate precision, high causality, and high confidence for Ni targeting, although it is not clear from lithospheric fault interpretations alone why it should be more prospective than the Imangda-Letniya edge translithospheric fault.

A possible role for transverse translithospheric faults in localizing the Norilsk-Talnakh district has received little attention because basement faults in this region are deeply buried under several-kilometers-thick Neoproterozoic to Paleozoic cover. Yakubchuk and Nikishin (2004) speculated that a NW-striking transfer fault that coincides with the Pyasina Uplift between Talnakh and Norilsk was the principal magma conduit. Alternatively, Diakov et al. (2002) highlighted a >300-km-long, W-NW–striking, magnetic lineament ~55 km northeast of Talnakh. Reinterpretation of regional gravity and magnetic imagery in this study shows evidence for other W-NW magnetic lineaments, including one (L3 in Fig. 4) that is projected to intersect near Talnakh and bounds the south margin of a regional gravity low. The L2 and L3 lineaments are elements of a craton-wide linear corridor, here named the Talnakh-Dyupkun Lineament, that bounds high amplitude gravity highs (Fig. 4C) and forms part of an extensive set of basement faults expressed by subtle linear magnetic and gravity discontinuities spaced ~70 to 150 km apart (Fig. 4). This lineament/fault set was likely activated during the Siberian LIP event as weakly extensional faults that influenced emplacement of the Maimecha dike swarm and the early syn- to postmineralization Morongovsky-Samoedsky volcanic assemblages. They likely originated as Neoproterozoic intracratonic rift faults, similar to and partly coincident with the E-W to W-NW–striking paleorift zones identified by Krivolutskaya et al. (2019), which includes the Lower Tunguska paleorift margin that coincides with the L7 lineament in Figure 4C. Furthermore, the Talnakh-Dyupkun lineament is approximately aligned with the axis of Cambrian basin deepening, where extensive Cambrian platform evaporites transition northward to carbonate shelf and deep-water shale facies (Fig. 4D). The Norilsk-Talnakh district also coincides with the intersection of regional E-NE–trending faults that define the northern margin of the Syvermin gravity low (Diakov et al., 2002). As a targeting ingredient, the Talnakh-Dyupkun lineament (transverse translithospheric fault) has high accuracy with the world’s largest Ni-Cu sulfide deposit located at its intersection with the Norilsk-Kharaelakh fault, but moderate precision, low confidence, and low causality ratings. Other prominent west-northwest and east-northeast translithospheric fault intersection zones along the Norilsk-Kharaelakh and Imangda-Letniya edge translithospheric faults may also be prospective, depending on the exposure depths of productive stratigraphic horizons.

The giant Voisey’s Bay Ni-Cu-Co deposit in eastern Canada is one of the best documented examples of lithospheric- to deposit-scale structural controls on Ni-Cu deposit emplacement (Myers et al., 2008; Lightfoot and Evans-Lamswood, 2015; Saumur et al., 2015). Voisey’s Bay is hosted by troctolite-gabbronorite units of the 1363 to 1322 Ma Nain Plutonic Suite, which comprises mostly anorthosite and leuconorite intrusions and is correlated with the 1350 to 1300 Ma Gardar alkaline mafic complexes in western Greenland (Myers et al., 2008). The small areal extent (~20,000 km²), long duration, and highly evolved nature of the Nain Plutonic Suite does not qualify it as a typical LIP. The Nain Plutonic Suite was emplaced as mid-crustal batholiths along the Paleoproterozoic Torngat orogen, which comprises a narrow belt of thin, weak lithosphere (Myers et al., 2008) sandwiched between two reworked Archean cratons, the Nain Province (east) and Churchill Province (west). The Torngat orogen is bound on the east margin by the >750-km-long Abloviak shear zone, which appears to dip steeply west (Ryan, 2000). The Abloviak shear zone is not mapped at surface where it is intruded by the Nain Plutonic Suite, but it is delineated here (Fig. 5A) as an edge translithospheric fault (suture) with moderate confidence by steep magnetic and gravity gradients.

The Voisey’s Bay district directly overlies the Abloviak shear zone center line within the intersection of the E-W-striking Voisey’s Bay-Gardar fault zone (Myers et al., 2008). As the largest and most prominent transverse translithospheric fault to crosscut the Torngat orogen, the Voisey’s Bay-Gardar fault zone intersection zone broadly defines the most prospective district-scale position along the entire magmatic province. In detail, the fault zone comprises an ~60-km-wide fault/fracture zone between the Fraser Canyon and Kogaluk lineaments that envelopes many subsidiary W-NW- to E-W-striking faults that are spaced ~2- to 4-km apart and show relatively minor (mostly <1 km) bulk sinistral throw (Fig. 5). A critical question for deposit-scale structural targeting here is whether any of the subsidiary faults within the Voisey’s Bay-Gardar fault zone intersection zone can be distinguished as more prospective than others?

From reinterpretation of magnetic and gravity imagery, four prominent W-NW-striking fault zones are distinguished from numerous lesser faults by their larger displacements and segmentation of magnetic and gravity domains (Fig. 5B). One of the most prominent west-northwest fault zones (L2, Fig. 5B) intersects the Abloviak zone next to the Voisey’s Bay deposit, where it appears to largely terminate or transition westward to offset minor faults. Whereas most ~east-west fault displacements postdate the Nain Plutonic Suite, a few have evidence of pre- to synmagmatic emplacement control on Nain Plutonic Suite intrusions, based on pluton elongation, syndeformation pluton borders, colinear dike trends, and magmatic shear foliations (Myers et al., 2008; Saumur et al., 2015). The L2 shear zone likely influenced emplacement of the W-NW–elongate Mushuau Intrusion for which it defines the northern border (Fig. 5C). At Voisey’s Bay, a west-northwest fault colinear with L2 also defines the synmagmatic North Wall fault that borders the northern margin of the Eastern Deeps chamber (Saumur et al., 2015). The Voisey’s Bay deposit cluster stretches ~5 km east-west along the Discovery Hill dike parallel to the western portion of the North Wall fault. Lightfoot and Evans-Lamswood (2015) proposed that transtensional dextral displacement along the Discovery Hill dike produced dilational jogs into which the Discovery Hill and ovoid chambers were emplaced. However, Saumur et al. (2015) interpreted that the EW faults show only pre-, syn-, and postmineralization normal-sinistral throw, and acted as largely passive host rock anisotropies during dike emplacement. From a greenfield exploration targeting perspective, it appears possible to narrow the prospective search space along the Abloviak shear zone to the four most prominent transverse fault intersections, and L2 in particular, but it is very difficult to structurally predict mineralized chamber positions along these faults without very detailed prospect mapping.

The giant Nebo-Babel Ni-Cu-Co deposit and nearby moderate-sized Succoth Cu deposit occur in the Mesoproterozoic Musgrave Province of central Australia (Fig. 6). These deposits were emplaced during the extensive Warakurna LIP event (1078–1070 Ma; Wingate et al., 2004), with Nebo-Babel deposited at 1068.0 ± 4.3 Ma (Seat et al., 2011) and Succoth deposited at 1078 ± 4.3 Ma (Grguric et al., 2018). Both deposits occur central to a 70 × 30 km cluster of mafic intrusions (Alcurra Suite) that overlies a high amplitude gravity anomaly (Alghamdi et al., 2018) at the structurally complex nexus of the Yilgarn, Mawson, and Arunta cratonic blocks (Begg et al., 2010, 2018; Smithies et al., 2015). Final collision of the three adjoining cratons was recorded by the Musgrave orogeny (1220–1120 Ma; Howard et al., 2015). The Musgrave Province is characterized by a ribbon of thin and exceptionally weak lithosphere that follows a moderately shallow S-dipping suture (edge translithospheric fault) represented by the Mann shear zone (Fig. 6; Korsch et al., 2013). The lithospheric edge was repeatedly tectonothermally reworked over >1.6 b.y. Subsequent orogenic collapse and lithospheric delamination produced an EW-striking failed intracontinental rift (the Ngaanyatjarra Rift; Evins et al., 2010) with voluminous, high-temperature, A-type granitoids of the Giles Magmatic event (1085–1040 Ma, Smithies et al., 2015). It was into this ultra-hot rift setting that the Warakurna plume impacted, and the West Musgrave area became a site of extreme magma flux producing very large layered mafic intrusions (e.g., Mantamaru and Jameson complexes) and the rhyolitic Talbot supervolcano (Smithies et al., 2015). Further tectonothermal reworking of the province occurred during the Paleozoic Petermann and Alice Springs orogenies (Howard et al., 2015).

The interpreted lithospheric fault map of Figure 6 shows that the West Musgrave Ni-Cu district occurs in the hanging wall of the Mann shear zone precisely where it is intersected by the N-striking Mundrabilla shear zone, a >750-km-long, subvertical transverse translithospheric fault with reverse-sinistral throw. Both translithospheric faults are defined with high confidence from outcrop maps and high-resolution aeromagnetic imagery. Although no plume impact center has been identified, the West Musgrave district stands out as the most anomalous tectonomagmatic center in the entire Warakurna LIP in terms of triple-edge translithospheric fault intersections, high amplitude geophysical responses, and exceptional intrusion volume. Accordingly, this Ni-Cu district can be targeted with moderate accuracy, high precision, and high confidence.

At its northern terminus, the Mundrabilla shear zone branches into a major splay fault that bounds a 50-km-wide fracture zone (Fig. 6). The location of the Nebo-Babel and Succoth Ni-Cu deposits within the broad Mundrabilla fracture zone is analogous to the Voisey’s Bay district setting and similarly raises the question of whether any particular fault, among a high frequency set of subsidiary faults and fractures, might be predicted to be more prospective at the deposit scale? All mineralized intrusions in this district are distributed along a narrow 30-km-long corridor in the hanging wall of the NE-striking Cavenagh fault (Fig. 6; Grguric et al., 2018). The Cavenagh fault is a curvilinear eastward fault extension of the northern edge of the subjacent Yilgarn craton, which underlies the Gunbarrel Basin, and is interpreted to be a NW-dipping normal fault (Joly et al., 2014). The spatial alignment and proximity of deposits to this fault suggests that it controlled emplacement of multiple host intrusions. The Cavenagh fault was likely inverted by postmineralization ductile deformation since Succoth is tightly folded and Nebo-Babel may be overturned (Seat et al., 2007). Although there are several other NE-striking faults in the province, the Cavenagh fault coincides with a geological domain boundary (Joly et al., 2014) and is geophysically distinguished as one of the most prominent cross faults within the Mundrabilla fracture zone, which supports the notion that the most prominent pre- to synmineralization fault internal to a broad translithospheric fault intersection zone may be predicted to be the most prospective.

The giant Raglan district in northeast Canada formed from moderately high MgO komatiitic basalt magmas (Lesher, 2007) in association with the very extensive ~1880 Ma Circum-Superior LIP event (Ciborowski et al., 2017). This LIP event likely had multiple plume impacts and plume deflection melt centers around the edges of the Superior craton that resulted in diverse magmatic Ni subprovinces including the Cape Smith belt, the Thompson belt some 1,500 km southwest, and Labrador trough to the southeast. Raglan formed in a narrow oceanic marginal rift basin in a pericratonic setting, which contrasts with the foregoing case studies all from intracontinental settings. Mineralization here formed ~10 m.y. before basin closure and inversion within the 1870 to 1850 Ma Ungava orogeny (St-Onge et al., 2002). The northern edge of the Superior craton comprises the Bergeron suture, which is well defined by steep potential field gradients under overthrust Povungnituk Group volcanosedimentary units.

The Raglan district deposit cluster covers an area ~100 km long and 30 km wide that coincides with a well-defined regional gravity high. The Ni-Cu deposits follow two strike-parallel belts located between 2 and 30 km north of the Bergeron suture (Fig. 7A). The northern Raglan belt comprises volcanic Ni-Cu (PGE) lava channels and sheet flows emplaced among sulfidic siliciclastic metasediments of the Nuvilic Formation. The Southern Expo-Ungava belt comprises Cu-Ni (PGE) mineralization in bladed dikes and sills emplaced at lower structural/stratigraphic levels among complexly folded sulfidic siliciclastic, carbonate, and banded iron formation metasediments of the Nuvilic and Beauparlant Formations. All Raglan district deposits occur ≤2 km from major strike-parallel, N-NW–dipping reverse faults/thrusts, some of which may have originated as synrift faults. Although these mineralized corridors are well defined by geology mapping and geochemistry, it is unclear which, if any, of the numerous strike-parallel faults could be selectively targeted as the most prospective potential magmatic feeder structures based on structure alone. Furthermore, although several craton-scale NW- and NS-trending translithospheric faults in the Superior craton intersect the Bergeron suture, no major transverse structures are mapped within the Cape Smith belt (Fig. 7A). From additional case studies (e.g., Albany Fraser orogen, Fig. 7B; Thompson belt), well-defined transverse translithospheric faults are lacking in most highly deformed pericratonic settings.

Table 1 summarizes key observations for the previous six case studies together with 11 more case study map interpretations shown in Figures 7 and 8. The additional case studies highlight the general proximity (mostly <40 km) of magmatic Ni-Cu sulfide deposits to paleocraton edge and edge-parallel translithospheric faults in each case, as well as close proximity (≤30 km) to transverse translithospheric faults for deposits located in intracontinental settings (Fig. 8).

In addition to the six detailed and 11 brief case studies presented above, new regional- to district-scale map interpretations of lithospheric faults were compiled for another 55 magmatic Ni-Cu (±PGE, Co) deposits (see App. 1). Distance measurements between center lines of the nearest paleocraton edge fault, edge-parallel fault (if present), transverse fault, and their intersection center points were statistically analyzed for spatial proximity to the Ni-Cu districts and deposits, as well as deposit emplacement controls (structural vs. stratigraphic). Results are summarized below at scales relevant to area selection in exploration, from subprovince to corridor, district, and deposit.

Subprovinces that are prospective for magmatic Ni-Cu sulfide exploration comprise linear belts of differentiated mafic-ultramafic intrusions or komatiite flows containing olivine-bearing mafic and ultramafic cumulates along paleocraton margins. The most prospective paleocraton margins are associated with unusually large and shallow lithosphere-asthenosphere boundary steps (<120 km deep) that border large resilient cratons, and which favored high-temperature, high-degree partial melting of deflected plume heads. As exemplified by the preceding case studies, subprovinces hosting intracontinental low-MgO Ni-Cu systems typically coincide with narrow (<150 km wide), nascent or failed intracontinental extension (rift) zones that nucleated on the cores of older suprasubduction orogenic belts and have relatively thin, weak, refertilized SCLM.

The most prospective structural corridors are defined by proximity to translithospheric faults along paleocraton edges. In this study, the proximity relationships reported by Begg et al. (2018)—in which most magmatic Ni-Cu deposits occur <120 km from an upper mantle domain boundary—are independently verified and further refined with higher resolution maps of edge and edge-parallel translithospheric faults at their upper crust position. Figure 9A shows a general decrease in deposit frequency with increasing distance from edge translithospheric faults. Proximity is tighter for low- to moderate-MgO suites than for komatiitic suites. The priority exploration search space can be defined by an inflection in the cumulative frequency curve where ~75% of deposits occur ≤40 km from edge translithospheric fault center lines. Greater distances out from edge translithospheric faults are permissive because the supergiant Norilsk and Talnakh deposits are located 65-km distal to the principal edge translithospheric fault interpreted in this study, albeit also <2 km from a prominent edge-parallel translithospheric fault. Alternatively, Begg et al. (2018) speculated that the Norilsk-Kharaelakh fault could be a primary edge translithospheric fault originating from a buried microcraton suture. If their interpretation is correct, close proximity to an edge translithospheric fault is indicated and the distribution in Figure 10B would define a lognormal trend.

The spatial distribution between magmatic Ni deposits and the next largest edge-parallel translithospheric fault is tighter at the corridor scale than for edge translithospheric faults. About half of the prospective corridors studied include one to two edge-parallel translithospheric faults in addition to the primary edge translithospheric fault. Figures 9C and D show that 90% of deposits and ~97% of metals occur within 30 km of the most prominent edge-parallel translithospheric fault, including ~54% of deposits and 75% of metals located within just 5 km of an edge-parallel translithospheric fault. This strong proximity relationship holds true across the spectrum of high- to low-MgO suites.

Deposit locations were compared to the translithospheric fault dip direction. For structures not well mapped in outcrop, dip direction may be inferred from 2D forward models and 3D inversion models of potential field data, and/or seismic reflection and magnetotelluric profiles. In the case studies where the translithospheric fault dip direction is known or reasonably inferred, it is observed that at least 70% of deposits are located either preferentially in the hanging wall of an inclined edge or edge-parallel translithospheric fault, or on either side of a subvertical edge-parallel translithospheric fault. Exceptions to these generalizations include the Kalatongke, Yangliuping, and Radio Hill deposits located in footwall positions to mapped edge (parallel-)translithospheric faults. And whereas Julimar (Fig. 8H) is in the footwall of the transverse Meckering suture zone, the dip and precise location of the original craton edge translithospheric fault, now represented by the Darling Fault zone, remains unknown because of extensive later deformation along the latter fault during Rodinia breakup.

Magmatic Ni-Cu district extents are highly variable. High-MgO (>17% MgO) magmatic systems commonly form large Ni districts, 30 to 100 km long, 5 to 15 km wide, with clusters of numerous smaller deposits distributed along and parallel to major rift-parallel fault corridors. Low-MgO magmatic systems typically form smaller districts (5 30 km long) with one main deposit and few or no satellite deposits. Notable exceptions to this generalization include the Norilsk-Talnakh and Duluth-Mesaba districts, which are up to 60 km long. In many but not all cases, a cluster of mafic to ultramafic intrusions is associated with a district-scale gravity high associated with accumulations of high-density mafic intrusions in the mid to lower crust.

As highlighted by the preceding case studies, deposit clusters commonly occur near the intersection of a prominent transverse translithospheric fault with an edge or edge-parallel fault. Most low-MgO intrusion-hosted Ni deposits in intracontinental settings occur proximal to transverse translithospheric faults comprising faults or lineaments with strike lengths between 250 and 1,000 km. The most common intersection geometry is T-shaped where an older transverse translithospheric fault terminates against a younger edge or edge-parallel translithospheric fault at the paleocraton margin. Distances to both the nearest transverse translithospheric fault center line and translithospheric fault intersection center point were analyzed. Figures 9E and F show a tight correlation, with ~82% of deposits and ~90% of metal occurring within 25 km of the nearest major transverse translithospheric fault. Notably, ~43% of deposits and ~63% of metal occurs within just 5 km of transverse translithospheric faults. Proximity to a translithospheric fault intersection center point is less tight. Figure 9G shows that ~67% deposits and ~86% of metals occur within 25 km of intersection center points, whereas ~87% of deposits and ~91% of metals occur within 55 km of center points. Proximity to translithospheric fault intersection center points decreases proportionally as the acute fault intersection angle decreases. The Voisey’s Bay and West Musgraves districts are atypical in being located central to broad (~50 km wide) transverse fracture zones. The supergiant Platreef PGE Ni-Cu deposit is also atypical in that it has no proximal (<100 km) transverse translithospheric fault mapped.

The case studies showed that although some well-endowed districts are located within a few hundred kilometers of mapped plume impact centers (e.g., Kabanga, Fig. 3C; Norilsk-Talnakh, Fig. 4B; Thompson, Ciborowski et al., 2017), those centers have no consistent or predictable relationship for targeting magmatic Ni districts. The radiating mega-dike swarms with which they are associated typically also lack major Ni-Cu sulfide deposits, although some dikes, e.g., the Jimberlana dike in Western Australia, may host small and discontinuous Ni-Cu prospects (Pirajno and Hoatson, 2012).

Deposit size vs. translithospheric fault proximity was analyzed. Figure 9H shows that supergiant deposits are located closest to both edge-parallel translithospheric faults (≤5 km) and transverse translithospheric faults (≤22 km). Giant-sized deposits are the next most proximal group to both edge-parallel translithospheric faults (≤46 km) and transverse translithospheric faults (≤33 km), whereas major- to moderate-sized deposits occur up to 95-km distal to translithospheric faults. These measurements indicate that deposit size potential is roughly inversely proportional to translithospheric fault distance, which has important implications for selectively targeting giant discoveries.

Spatial relationships with transverse translithospheric faults are generally not apparent in pericratonic and Archaean greenstone Ni deposit settings. As noted in the Cape Smith belt and Albany Fraser orogen case studies (Fig. 7A, B), transverse translithospheric faults in accreted pericratonic marginal basins and microcontinent arcs rarely extend from adjacent cratons across sutures into these accreted terranes. Fortunately, the reduced efficacy of district-scale structural targeting in these settings may be partly countered by the larger district-scale footprints of high-MgO systems. In Archean greenstone settings, early formed transverse translithospheric faults, if present, were likely obscured or obliterated by the large postmineralization granitoid batholiths emplaced between narrow greenstone belt remnants. That said, a detailed study by Perring (2016) in the Agnew-Wiluna belt of Western Australia demonstrated that small west-northwest cross faults were a local control on the distribution of komatiitic Ni deposits in that belt (Fig. 7D). The extents of those cross faults were enhanced by later sinistral strike-slip reactivation. The Agnew-Wiluna belt Ni deposit clusters have a semi-regular spatial periodicity along strike with a mean spacing of ~22 km (n = 6, σ = 3.9, coefficient of variation = 17.8%, as defined by Hayward et al., 2018). The most likely control on this periodicity was self-organization of the west-northwest fault set into semiregular spacing.

The very close spatial relationship between the supergiant Norilsk-Talnakh deposits and the Norilsk-Kharaelakh fault suggests that proximity to mantle-tapping translithospheric faults could be an economically important targeting element at the deposit scale. It is widely accepted that this fault behaved as a dike-like feeder to the mineralized sills located immediately adjacent to both sides, despite the lack of evidence for preservation of local feeder dikes (Zen’ko and Czamanske, 1994; Diakov et al., 2002). Similarly, the Voisey’s Bay deposits occur within a structurally controlled dike-sill complex immediately adjacent and parallel to an EW-striking synmagmatic shear zone (Fig. 5C). Lightfoot and Evans-Lamswood (2015) interpreted that Voisey’s Bay, together with some Ni-Cu deposits in China that are hosted by rhomboid-shaped chonolith intrusions, were emplaced directly into dilational jogs along strike-slip faults. However, these examples appear to be the exception rather than the norm. From this study, it was observed that few mineralized sills, chonoliths, or flows show compelling evidence for direct structural controls on their emplacement at the deposit scale. And for the exceptions, identification of the active synmineralization fault control on chonolith emplacement can be difficult to distinguish from postmineralization deformation and difficult to predict ahead of discovery. For example, the Kangguer fault, into which the Huangshan and Huangshandon Ni chonoliths were emplaced, was inferred to be a synmineralization sinistral shear by Lightfoot and Evans-Lamswood (2012), but was mapped as a late syn- to postmineralization dextral shear zone by Branquet et al. (2012). Similarly, synmineralization movement along the Fuyun Fault adjacent to the Kalatongke Ni-Cu deposit was alternatively interpreted to be dextral (Lightfoot and Evans-Lamswood, 2012) or reverse-sinistral (Li et al., 2015).

Local structures at the deposit scale could alternately be interpreted to have a passive role comprising highly anisotropic and rheologically weak zones exploited by magma. For example, premineralization fold hinges and shear zones frequently influenced thickening and thinning of sills and dikes emplaced among them (Zen’ko and Czamanske, 1994; Saumur et al., 2015). And host rock rheology appears to have been the dominant control on emplacement of mineralized intrusions at the deposit scale, where mineralization was typically deposited in subhorizontal sulfide traps that are semi-conformable with rheologically weak beds, contacts, or foliations. Of 59 intrusion-hosted Ni-Cu sulfide deposits with well documented host lithologies (Fig. 10), 35% are hosted in carbonaceous metapelites, 20% in mixed carbonaceous metapelite and carbonate sequences, 12% in carbonates, and 12% are in metasedimentary paragneiss. Collectively, ~80% of deposits were emplaced among metasediment host units. In detail, many mineralized intrusions have irregular geometries and margins that locally truncate host layering as a result of wall-rock melting and assimilation (Barnes et al., 2016).

Wall-rock assemblages also commonly contain minor to significant amounts of sulfur (5–10%) as either anhydrite with carbonate or pyrite or pyrrhotite within graphitic metapelites. The role of local host rock sulfur assimilation in driving sulfide saturation is critical for high-MgO Ni deposits and likely important for low-MgO Ni-Cu deposits (Barnes et al., 2016; Lesher, 2019). High-MgO deposits are typically confined to a thin S-rich stratigraphic interval defined as the productive stratigraphic horizon (e.g., Raglan, Thompson, Pechenga). The host sequence for the supergiant Talnakh-Kharaelakh deposit also has an exceptional abundance of semi-massive anhydrite. There are, however, significant exceptions where little to no local S-sources are identified and distal (deeper) sulfide saturation is inferred, including Jinchuan, West Musgrave, Voisey’s Bay, Eagle’s Nest, and Selebi-Phikwe (Barnes et al., 2016).

Exploration for magmatic Ni-Cu (±PGE, Co) ore deposits is like searching for a needle among many haystacks. Only a small fraction of the ≥300 known continental mafic LIPs (Ernst et al., 2021) are endowed with at least one to two well-mineralized Ni districts. Whereas each LIP may extend over 300,000 to 30,000,000 km², Ni districts typically only cover 100 to 1,500 km² and individual Ni deposits are typically 0.1 to 10 km² in area and 10- to 100-m thick. This vast scale range requires targeted area reduction over seven orders of magnitude. Furthermore, emplacement depths for low-MgO intrusions may range from 1 to 30 km, whereas greenfield explorable depths are <2 km below surface. The subhorizontal geometry of undeformed deposits, lack of broad alteration haloes, and very small mineralization footprints make detection of buried Ni-Cu deposits extremely challenging. Consequently, any significant improvements to target area selection accuracy and precision can yield enormous benefits to explorationists.

Targeting accuracy was assessed as the true positive capture rate of proximal lithospheric structures and is scale dependent. At the deposit scale, proximity to the mapped translithospheric faults and their intersections is very inaccurate and of limited use beyond the few noted exceptions. At subprovince to district scales, however, targeting positions proximal to translithospheric faults in the upper crust should provide a major improvement in accuracy over previous approaches that only targeted lithospheric domain boundaries in the upper mantle. Shallow lithosphere-asthenosphere boundary steps are typically resolved spatially to ±100 km using seismic tomography and long-period magnetotelluric transects, and challenges arise where postmineralization tectonothermal reworking has significantly modified lithosphere-asthenosphere boundary positions. Results from this study show potential to capture 90% of magmatic Ni deposits in relatively narrow (≤30 km) corridors centered on one- to two-edge or edge-parallel translithospheric faults (Fig. 9) while prioritizing the hanging wall side. Corridor areas can be further reduced with >80% capture rate to a few high priority district-scale areas by also targeting proximity (≤30 km) to prominent transverse translithospheric faults (Fig. 9). Targeting proximity to both edge-parallel and transverse translithospheric faults is more accurate than targeting either fault type alone or their intersection center points.

In theory, targeting precision is inverse to the false-positive capture rate, which can be determined by measuring the total area of buffered translithospheric faults not associated with Ni deposits. In practice, several issues render precision assessment subjective and qualitative. For the scope of this study, case study areas generally did not extend to the full extents of each subprovince, so targeted fault extents may be incomplete. Whereas most edge-parallel translithospheric faults are mapped with high confidence, there are uncertainties in mapping cryptic transverse translithospheric faults, including: (1) the position and representativeness of center lines to fault zones and lineaments, (2) whether some cryptic fault interpretations with little or no surface expression may be false, (3) the inferred pre- to synmineralization timing of translithospheric faults, and (4) dip direction. Furthermore, the distance measures reported here are maximums because they reference center lines rather than fault zone edges. For wide fault zones (5–20 km), distance to the fault zone edge could be significantly less than distance to the center line.

Prioritization of fault zone prominence was based on subjective scale ratings that also introduced uncertainty. Each edge-parallel and transverse translithospheric fault is part of a set of faults with assumed power law size-frequency scaling. In other words, there are many more small faults than large ones in each set. The thesis developed here is that targeting precision increases by selecting the largest faults in each set, assuming that they are likely to more efficiently tap mantle source reservoirs. Fault prioritization criteria are based on subjective ratings of translithospheric fault length scale, continuity, geophysical gradient magnitude, and discontinuity in crustal geology and basement isotopic maps. This approach underscores the importance of understanding fault hierarchy in terms of scale, from first-order master faults to subsidiary second-, third-, and fourth-order faults.

At face value, the case study results indicate that targeting proximity to mapped translithospheric fault intersections has moderate precision because there are only a few (typically 2–5) prominent, craton-scale, transverse translithospheric faults identified for each subprovince, typically at spacings >100-km apart. Their total area with nominal 30-km proximity buffer widths covers a relatively small portion (typically <10%) of each prospective subprovince area (Fig. 11A). Targeting giant discoveries can be further focused by prioritizing areas ≤10 km from edge-parallel and transverse translithospheric faults. The presence of multiple untested translithospheric fault intersection targets may pave the way for additional discoveries in known mineralized subprovinces. There is no a priori reason why endowment should be limited to one to two ore districts if the prospective search space is sufficiently preserved, explorable, and accessible.

Two end-member applications can be considered for greenfield area selection: (1) target proximity to the most prominent predicted translithospheric fault intersection(s) then follow up with surveys to detect evidence of prospective mafic-ultramafic intrusions with sulfide mineralization, or (2) extensively map prospective mafic-ultramafic intrusive suites and prioritize intrusions by proximity to prominent translithospheric fault intersections. Prospective intrusions may be characterized by proxies for the highest degree partial melt (e.g., Mg#), least fractionation, most adcumulate rocks, magmatic sulfide textures, and geochemical anomalism in target metal and pathfinder elements (e.g., Barnes, 2023). The former approach suits extensively covered and hard to access terranes, whereas the latter approach suits more exposed and accessible terranes. A combined approach would also work for shallow covered terranes with high resolution gravity and magnetic surveys.

A causal relationship can be interpreted for the few mineralized intrusions that are located immediately adjacent to presumed feeder faults (e.g., Norilsk-Talnakh, Voisey’s Bay, Huangshan; Lightfoot and Evans-Lamswood, 2015). Analyses of magnetic susceptibility anisotropy in intrusions at Norilsk, along with sulfide-metal fractionation trends at Oktyabrsky, indicate that magmas in the Norilsk-Talnakh district flowed outward from the Norilsk-Kharaelakh fault (Latyshev et al., 2023). In other examples, the Nkomati chonolith occurs in the immediate hanging wall and elongate parallel to the Laersdrift fault (Fig. 8B), as is the Munali chonolith in the hanging wall of the adjacent Munali fault (Blanks et al., 2022). However, the fact that many other mineralized intrusions were emplaced several kilometers distal to translithospheric or regional faults and lack of evidence for connecting pathways makes it difficult to prove causal relationships. Accordingly, many proximal transverse translithospheric faults in the case studies have low causality ratings. Proximity could be coincidental or reflect hindsight interpretation bias after discovery. Given that LIPs extend far distal to paleocraton edges, it is expected that much melt ascent occurs independent of edge translithospheric faults and migrates laterally through regionally extensive dike swarms and sills. However, the observation from multiple case studies that the most primitive and fertile intrusion clusters with Ni deposits consistently occur near prominent but infrequent translithospheric fault intersections is highly unlikely to be coincidental, and rather supports a role for translithospheric faults in focusing the ascent of fertile magmas through the crust.

Similar proximity relationships between deposits, translithospheric faults, and translithospheric fault intersections are documented for other magmatic-hydrothermal ore systems. Orogenic Au deposits are widely accepted to be controlled by first- to second-order, orogen-parallel translithospheric faults even though most Au deposits occur along subsidiary faults, antiformal hinges, and adjacent lithological contacts with strong rheological contrast, mostly <4 km but up to 10 km distal to a primary fault center line (Gardoll, 2005). Many large porphyry-related Cu-Au deposits and districts are also distributed along and proximal to arc-parallel translithospheric faults, mostly <15 km but up to 30 km distal to a primary fault zone center line (e.g., Sillitoe, 2010; Hayward et al., 2018; Farrar et al., 2023). Moreover, many porphyry districts occur proximal to transverse translithospheric fault intersections (Sillitoe, 2010; Farrar et al., 2023; Wiemer et al., 2023). For both orogenic Au and porphyry Cu-Au, translithospheric faults are assumed to be the primary conduits that localized high fluid flux of fertile magmas (± aqueous fluids) from deep crust and mantle sources.

The driving force for magma ascent is buoyancy due to the lower density of mafic to ultramafic magmas relative to lower crustal and upper mantle rocks. Magma ascends principally via subvertical dikes that open perpendicular to the least compressive stress direction (σ3). Rupture occurs when P_l + P_e ≥ σ₃ + T₀, in which P_l = reservoir magmatic pressure at lithostatic load, P_e = reservoir excess fluid (over)pressure, σ₃ = minimum compressive stress, and T₀ = tensile strength at rupture site. Increases in P_e decrease the effective normal stress and can drive rupture of misoriented faults (Sibson, 1996). Increased tensile stress also lowers the effective stress and promotes vertical dike propagation. For low- to moderate-MgO magma suites, which are much more viscous than komatiitic magmas, magmas tend to pond at major rheological boundaries such as the lithosphere-asthenosphere boundary, Moho, and brittle-ductile transition, where sill-like reservoirs become overpressured relative to lithostatic pressure gradients (Watanabe et al., 1999). Magma channels typically propagate in these overpressured systems by fluid-induced failure at high fluid pressures rather than by shear-induced failure.

The trajectories of fluid-driven fractures are likely to be strongly influenced by rheology. Boundaries between domains of contrasting rheology impact magma conduits across all scales, from the edges of thick rigid Archean SCLM domains to crustal terrane boundaries and the lithospheric boundary layers where staging chambers accumulate, including the base of sedimentary basins. Sills are preferentially emplaced along contacts defined by (1) high rigidity contrast where stiff strata overlie soft strata, and (2) high rheology anisotropy comprising weak ductile zones (Sili et al., 2019). Sedimentary basins in general, and productive horizons in particular, are favorable hosts for sills not just for the availability of sulfur, but because of their abundance of rigidity contrasts, weak interfaces, and lower lithostatic load pressure. Sills may also develop in volcanic successions and crystalline rock masses with low rigidity contrast where weak interfaces (lithologic contacts, faults) are subjected to higher compressive stresses (Barnes et al., 2001; Krivolutskaya et al., 2018; Sili et al., 2019). The dominant controls on emplacement of mineralized chonoliths and bladed dikes, where dikes transition to sills, are country rock interfaces characterized by weak rheology, high anisotropy, and high fusibility, all of which favor thermomechanical magma injection laterally along the dominant foliation with partial host rock assimilation.

If translithospheric faults typically do not directly host mineralized intrusions and have low causality ratings at the deposit scale, what is the role of translithospheric faults in localizing the ascent and emplacement of mineralized intrusions? Four possible mechanisms are considered below and summarized in Figure 11.

1. Translithospheric fault root zones tap mantle melt reservoirs: Many edge-parallel and transverse translithospheric faults are likely rooted in discrete mantle shear zones, which may be a few tens of kilometers wide or perhaps linked to broad 100-km-wide zones of distributed low strain (Vauchez et al., 2012). Craton-edge translithospheric faults commonly originated as listric rift faults and subduction margins with moderately dipping and laterally extensive mantle shear zones, though many such shear zones were likely steepened as reactivated transpressional shears during subduction, collision, orogenic inversion, and lithospheric thickening. Magma may initially percolate through channelled flow networks in mantle shear zone roots to feed Moho sills and crustal dike arrays (Cruden and Weinberg, 2018). Long-lived shear heating and melt transfer around shear zone roots in the upper mantle and lower crust leads to local shallowing of isotherms which, in turn, makes weakened shear zone roots strong attractors for melt and strain localization during plume impact events. Depending on fault zone dip, curvature, width, and crust thickness, mantle shear zones could be offset laterally relative to upper crust fault positions by up to 45 km (e.g., 45° dip) in the downdip direction. This distance range matches the proximity spread observed in Ni deposit locations.

2. Exceptional fluid flux and permeability of intersecting translithospheric fault damage zones: Exceptionally high flux of fertile mantle melts is required to form magmatic Ni-Cu sulfide ore deposits. High flux rates with turbulent flow allow for vertical transport of dense sulfide emulsions in cases where sulfide saturation occurs at deeper levels, which seems common to many low-MgO systems (e.g., Lightfoot and Evans-Lamswood, 2015; Barnes et al., 2016). High flux rates also allow high metal tenors to form with high R-factors, where immiscible sulfide droplets scavenge metals from a large mass of silicate melt before entrapment (Naldrett, 1999). Evidence of high flux in mineralized systems includes an abundance of olivine normative mafic and ultramafic cumulates in flowthrough conduits and unusually large thermal aureoles relative to conduit volume. Tectonically active translithospheric fault intersections conceptually seem best suited to sustain high magma flux rates by providing the most permeable steeply-plunging, transcrustal conduits with vertical permeability vectors. Vectors for permeability and fluid flow in fault zones that are subject to stress-induced rupture preferentially develop along dilated structures oriented parallel to σ₂ and orthogonal to the fault slip vector (Sibson, 1996). Permeable channels in reverse and normal faults are subhorizontal, which favors lateral fluid flow, whereas dilational jogs in strike-slip faults and relays in reverse and normal faults align with the vertical hydraulic gradient and maximize vertical fluid flow. High-angle translithospheric fault intersections, particularly T-shaped terminations, would have a dilational component under most contractional and extensional stress regimes. Furthermore, each translithospheric fault is typically enveloped in a broad damage zone comprising subsidiary faults and fractured protolith where the bulk permeability is two- to three-orders of magnitude greater than fault cores (Caine et al., 1996). Preferential fluid flow through fault damage zones may explain why magmatic Ni-Cu deposits occur lateral to translithospheric fault center lines, similar to the distribution of orogenic Au deposits in faults subsidiary to their parental translithospheric faults (Gardoll, 2005). Total damage zone widths scale with fault length and displacement but remain poorly constrained for most translithospheric faults because they are rarely mapped in detail. They may be only hundreds of meters wide around individual large faults (Faulkner et al., 2010) or link up in complex translithospheric fault zones with multiple splay and secondary faults extending over widths of up to several kilometers.

3. Lateral flow through dike-sill-dike networks: Mafic magmas ascend and disperse laterally through networks of interconnected dikes and sills, commonly following complex “stair-step” paths. Lateral transport of magmas through mega-dikes and well-connected dike-sill-dike complexes can potentially disperse magmas tens to hundreds of kilometers lateral to master lithospheric fault conduits (Kavanagh et al., 2015; Magee et al., 2016). Despite the potential for such long-range lateral transport, the distribution of Ni-Cu deposits indicates that sulfide saturation and mineralization generally only occur close to translithospheric faults along paleocraton edges, either because magmas can only transport dense magmas and sulfide emulsions for short distances away from translithospheric fault conduits before gravitational settling, and/or because crustal S-sources are readily available near edge-parallel translithospheric faults.

4. Translithospheric fault hanging-wall shortcut pathways: The strong preference for Ni deposits to occur in the hanging wall of inclined translithospheric faults may be explained by a process in which overpressured magmas seek not just the path of least resistance, but also the shortest path along the hydraulic gradient from reservoir to surface. Steep hanging-wall shortcuts dissipate energy more efficiently and favor higher fluid flux than shallow dipping or convoluted dike-sill-dike paths. Consequently, magmas that ascend from the roots of inclined translithospheric faults are likely diverted into steep dipping hanging-wall shortcut faults and their damage zones for both edge-parallel and transverse translithospheric faults. In this scenario, deposits emplaced deeper in the crust should conceptually occur closer to inclined master translithospheric faults than deposits emplaced at shallow crustal levels, though this spatial relationship has yet to be confirmed. This relationship implies that explorers should take a bottom-up targeting approach that prioritizes proximity to the lower crustal root of a translithospheric fault, rather than its uppermost crustal position, while considering fault dip and lateral spread to hanging-wall shortcuts.

These four scenarios are not mutually exclusive and likely operated in combination. Figure 11B presents a model where the steep-plunging root zones of edge translithospheric faults at transverse translithospheric fault intersections initially focus ascent of buoyant mantle magmas, then fluid-induced failure of steep-dipping shortcut hanging-wall faults (subsidiary edge-parallel translithospheric faults), fault damage zones, and rheologically weak contacts create dike-sill-dike complexes that laterally disperse fertile magmas 1- to 30-km away from the cores of translithospheric fault conduits. In cases of extreme flux of high temperature magmas, mineralizing intrusions create and preserve their own channel ways (chonoliths) through the rheologically weakest units as well as their own sulfide traps, through wall-rock brecciation, melting, and assimilation (Barnes et al., 2016). The synergistic combination of these processes likely led to self-organization of the structural-magmatic system.

Self-organization in complex systems leads to the emergence of (multi-)fractal size-frequency distributions and pattern recurrence at certain scales. Ordered patterns emerge as the most efficient way to minimize large potential energy gradients, such as those arising from plume impacts, through organized growth of energy dissipative structures and complex feedback loops (e.g., Hayward et al., 2018). One emergent pattern is the lognormal size frequency distribution of Ni endowment, which is common to many other mineral deposit types (e.g., Singer, 2013). Another is the broad 600 to 900 m.y. periodicity in global Ni mineralizing events linked to supercontinent and superplume cycles (Begg et al., 2018). At a much smaller scale, the semiregular periodicity of vent-proximal komatiite Ni deposits in the Agnew-Wiluna belt resulted from self-organization of a synrift extensional fault set (Fig. 7D).

The extreme magma flux required to form large magmatic Ni-Cu sulfide deposits likely emerged from positive feedback loops that led to nonlinear and far from equilibrium conduit growth. Magma infiltration into a lithospheric fault zone potentially reduces its mechanical strength by an order of magnitude. Positive feedbacks that grow and sustain high flux magma channels include melt-enhanced strain localization and thermal weakening that led to increased strain rate, wall-rock dilatancy and permeability which, in turn, promote further magma infiltration and fault zone growth while sustaining low viscosity and above-solidus temperatures in the magma (Klepeis et al., 2022). These processes can be cyclical like seismic pumping of fluids. Positive feedbacks (above) likely accumulated in high energy environments around active translithospheric faults that were physically linked to mantle melt zones. Conversely, negative feedbacks that suppress conduit growth include compression that narrows or closes dike apertures, stress shadows near dike openings, stiff rheological barriers, thick sills, and slower ascent rates that allow cooling, fractionation, viscosity increases, and solidification. Negative feedbacks likely accumulated in lower energy environments for translithospheric faults not linked to zones of high degree mantle melting. These feedbacks drove bottom-up self-organization of the dynamic structural-magmatic system (Cruden and Weinberg, 2018).

The results of this study lead to several conclusions with important implications for discovering new magmatic Ni-Cu (±PGE, Co) sulfide deposits. The results mostly benefit district-scale targeting for giant ore deposits under shallow cover and addresses a critical gap in current Ni mineral system models. More generally, assumptions about the role of translithospheric faults in localizing magmatic ore deposits were tested and developed into a new model for structurally controlled high flux of fertile mafic-ultramafic magmas.

1. The extreme magma flux rates required for Ni sulfide ore formation emerge from positive feedbacks and bottom-up self-organization centered around mantle-tapping translithospheric faults linked to lithospheric domain boundaries.

2. Targeting proximity to upper crustal translithospheric faults should provide greater accuracy and precision than targeting mantle steps, but many translithospheric faults are cryptic and have some map uncertainty. Their interpretation requires use of multidisciplinary data supported by objective analysis of accuracy, precision, and causality.

3. At the subprovince to corridor scale, the most prospective target areas are ≤30 km from paleocraton edge- or edge-parallel translithospheric faults and preferentially in the hanging wall of inclined translithospheric faults. This generalization applies to all magmatic Ni-Cu (±PGE, Co) sulfide deposit settings. Relatively limited lateral dispersion arises from dike-sill-dike networks developed in shortcut hanging-wall fault splays, their damage zones, and along rheologically weak contacts. There is no evidence for long-range (100s km) lateral dispersion of ore-forming magmas through radiating dike swarms, sill complexes, or lava flows.

4. At the district scale, the most prospective target areas are ≤30 km from where the most prominent transverse translithospheric faults intersect edge-parallel translithospheric faults. This generalization applies to intracontinental settings but is generally not applicable to inverted pericratonic, accreted microcontinent arc, or greenstone komatiite settings.

5. The most prospective fault zones in each subprovince are typically the largest, and the largest deposits are generally found closest to translithospheric faults. This means that area selection criteria can be optimized for giant discoveries using a hierarchal approach to large-scale structural targeting.

6. At the deposit scale, many deposits lack compelling evidence for direct structural control on their emplacement, although the supergiant Norilsk-Talnakh district is an important exception. Rather, they were emplaced as semi-concordant and subhorizontal flow channels, sills, chonoliths, and bladed dike extensions among rheologically weak metasedimentary or gneissic units lateral to feeder faults. Mineralizing intrusions can create and preserve their own sulfide traps sites through wall-rock melting and assimilation.

It is anticipated that these targeting criteria could potentially increase greenfield discovery rates to help meet society’s demand for green energy metals.

Part of this research was initiated while the author was employed by Teck Resources Limited with the support of Stuart McCraken and collaboration of Teck exploration staff, especially Jamie Kraft. Subsequent independent research benefitted from discussions with Steve Beresford, Graham Begg, Sandy Cruden, and Daniel Wiemer. Review comments provided by Richard Ernst and Peter Lightfoot helped significantly improve the manuscript.

Nick Hayward is currently the Director and Principal Consultant for PredictOre Pty Ltd and an Adjunct Senior Research Fellow with the Centre of Exploration Targeting, UWA. Nick has 35 years global technical leadership, project generation, and project management experience for base and precious metals (Cu, Ni, Zn, Au, Fe). Recent industry roles include Director, Global Project Generation (Exploration), Regional Chief Geoscientist, and Global Zinc Specialist for Teck Resources; also, Global Practice Leader–Structural Geology and Resource Targeting for BHP Billiton. Nick obtained a Ph.D. in Structural Geology and Tectonics in 1993 at James Cook University (Queensland).