Zinc on the edge—isotopic and geophysical evidence that cratonic edges control world-class shale-hosted zinc-lead deposits

The North Australian Zinc Belt is the largest zinc-lead province in the world, containing three of the ten largest known individual deposits (HYC, Hilton-George Fisher, and Mount Isa). The Northern Cordillera in North America is the second largest zinc-lead province, containing a further two of the world’s top ten deposits (Red Dog and Howards Pass). Despite this world-class endowment, exploration in both mineral provinces during the past 2 decades has not been particularly successful, yielding only two significant discoveries (Teena, Australia, and Boundary, Canada). One of the most important aspects of exploration is to choose mineral provinces and districts within geological belts that have the greatest potential for discovery. Here, we present results from these two zinc belts that highlight previously unused datasets for area selection and targeting. Lead isotope mapping using analyses of mineralized material has identified gradients in μ (238U/204Pb) that coincide closely with many major deposits. Locations of these deposits also coincide with a gradient in the depth of the lithosphere-asthenosphere boundary determined from calibrated surface wave tomography models converted to temperature. Furthermore, gradients in upward-continued gravity anomalies and a step in Moho depth correspond to a pre-existing major crustal boundary in both zinc belts. A spatial association of deposits with a linear mid- to lower-crustal resistivity anomaly from magnetotelluric data is also observed in the North Australian Zinc Belt. The change from thicker to thinner lithosphere is interpreted to localize prospective basins for zinc-lead mineralization and to control the gradient in lead isotope and geophysical data. These data, when combined with data indicative of paleoenvironment and changes in plate motion at the time of mineralization, provide new exploration criteria that can be used to identify prospective mineralized basins and define the most favorable parts of these basins.


Introduction
Sediment-dominated basins are by far the largest global source of zinc and lead, containing 54% and 68%, respectively, of the world's endowment of these metals . The giant size of some of these deposits makes them attractive exploration targets. However, they are difficult to find as they are rarer than other zinc-lead deposits such as volcanic-hosted massive sulfide (VHMS) deposits. Despite their attractiveness, very few major sediment-hosted zinc-lead deposits have been discovered in the past 2 decades. One possible reason has been a poor understanding of fundamental, large-scale controls on mineralization. Given that hundreds of sedimentary basins exist in the world, the following are key questions addressed in this study: (1) are there ways to screen more prospective from less prospective basins, and (2) are there ways to determine the most prospective parts of favorable basins?
The North Australian Zinc Belt, which is hosted by the Paleo-to Mesoproterozoic North Australian Basin System (Southgate et al. 2000(Southgate et al. , 2013, is the richest zinc province in the world, containing a total endowment of 89 Mt Zn and 1 3 41 Mt Pb-10% and 13%, respectively, of global pre-mining resources of these metals . The Northern Cordillera of western North America (Siberling et al. 1992;Colpron et al. 2006;Nelson et al. 2013), which extends from the northwestern conterminous USA through western Canada and into Alaska, is another major zinc province, containing a total endowment of 72 Mt Zn and 33 Mt Pb (Huston et al.2022). A striking feature of both zinc-lead provinces is that each had several distinct mineralizing events that occurred over a period of many tens to hundreds of millions of years. In both provinces metallogenic lightning has struck twice, actually many times, raising the possibility that some fundamental feature of these metallogenic provinces has enhanced the metal endowment.
In this contribution, we build on previous work by Hobbs et al. (2000), Southgate et al. (2013), Gibson et al. (2017), Hoggard et al. (2020), Huston et al. (2020), and Hayward and Paradis (2021). Hobbs et al. (2000) presented evidence that the major zinc-lead deposits of the North Australian Zinc Belt are associated with a prominent gravity "worm," which marks a horizontal gradient in upward-continued gravity data. Southgate et al. (2013) and Gibson et al. (2017) also presented evidence that zinc-lead deposits in this belt and parts of the Northern Cordillera are associated with the edges of continental shelves or platforms. Here, we present evidence that the distribution of basin-hosted zinc-lead deposits in these provinces is fundamentally controlled by cratonic edges that are visible in multiple datasets, including lead isotope, potential field, surface-wave tomographic, reflection seismic, and magnetotelluric data. Although lead isotope, reflection seismic, and magnetotelluric data are restricted in coverage due to acquisition logistics, potential field data (e.g., gravity and magnetic) are available globally, albeit at different resolutions, in addition to a range of global tomographic models (c.f. Hoggard et al. 2020). The continuity of these latter datasets over large parts of the world facilitates their first-order use to identify prospective basins at the global, continental, and province scales (Lawley et al. 2022).
All of the above datasets are used herein to answer the two key questions identified above. We demonstrate that basin-hosted zinc-lead deposits are associated with cratonic edges and that the resolutions of different datasets enable their use at different stages of exploration. Moreover, as the datasets presented in this study reflect different geochemical and geophysical properties of the host provinces, using all in combination provides new and more robust insights into geological processes that control basin-hosted zinc-lead deposits and districts. Although we have concentrated on zinc, it is possible or even likely that some of the same datasets can be used in understanding the genesis of, and exploration for, other types of basin-hosted mineral deposits, as demonstrated by Hoggard et al. (2020) for sediment-hosted copper deposits.

Basin-hosted zinc-lead deposits-timing of mineralization and terminology
Sedimentary basins host a wide range of zinc-lead deposits, some clearly epigenetic, but others formed either during sedimentation or early diagenesis. Although Mississippi Valley-type deposits are generally accepted products of as epigenetic mineralization that formed after lithification of the host rocks (Leach et al. 2005;Wilkinson 2014), the timing of shale-hosted 1 zinc-lead mineralization has been the topic of significant debate during the last half-century. Some authors (Williams 1978;Kelley et al. 2004a,b;Gadd et al. 2016Gadd et al. , 2017Johnson et al. 2015Johnson et al. , 2018Slack et al. 2017;Spinks et al. 2021) have advocated an early diagenetic timing, whereas others have argued for deposition during sedimentation (Carne and Cathro 1982;Goodfellow et al. 1993;Large et al. 2005) or during late diagenesis or basin inversion (Broadbent et al. 1998;Gibson et al. 2017). Advocates of the syn-sedimentary timing have termed these shalehosted deposits "SEDEX" (sedimentary exhalative: Carne and Cathro 1982), historically the most commonly used term for these deposits. Given the significant debate as to the timing of mineralization, we prefer to use a purely descriptive term such as "shale-hosted" rather than SEDEX. Leach et al. (2005) proposed the term "clastic-dominated" (CD) as an alternative non-genetic term.
At about the same time that the term SEDEX was coined, Morganti (1981) proposed a classification of basin-hosted, stratiform deposits by basin type, recognizing three types: (1) intracratonic basin-type deposits, (2) flysch basin-type deposits, and (3) platformal-marginal-type deposits. We have used the philosophy of assessing basin environment for deposit classification and adapted the Morganti (1981) approach along the lines used by Barrie and Hannington (1999) for VHMS deposits. In this classification of VHMS deposits, classifications are based on the lithological succession that hosts the deposit. For shale-hosted zinc-lead deposits, we recognize two types, siliciclastic-mafic and siliciclastic-carbonate (see also Hofstra et al. 2021). The former includes deposits hosted by siliciclastic-dominated basins with significant syn-depositional mafic igneous rocks (basaltic lavas and volcaniclastic rocks or high-level mafic sills) but without volumetrically major carbonate units. The siliciclastic-carbonate type, in contrast, refers to deposits hosted by basins with abundant siliciclastic and carbonate units but that mostly lack coeval mafic igneous rocks. The siliciclastic-mafic type corresponds to the flysch basin type of Morganti (1991), whereas the siliciclastic-carbonate type corresponds to the intracratonic basin and platformalmarginal types.
Our classification is also broadly consistent with the classification of Cooke et al. (2000), which is also based indirectly on the characteristics of basin fill. Cooke et al. (2000) inferred that their McArthur-type deposits were formed from oxidized brines produced by oxidized basins, whereas their Selwyn-type deposits were formed from reduced brines produced by reduced basins. The McArthur-type deposits correspond broadly to our siliciclastic-carbonate type, and the Selwyn type corresponds broadly to our siliciclastic-mafic types.

The North Australian Zinc Belt
The North Australian Zinc Belt (Fig. 1) contains three of the ten largest zinc-lead deposits in the world (McArthur River [HYC-Here's-Your-Chance], Hilton-George Fisher, and Mount Isa), as well as several smaller, yet still important, deposits (Century, Dugald River, Teena, Lady Loretta, and Cannington). Among these, only Teena is a recent (2013) discovery (Taylor et al. 2017). In addition, significant copper-cobalt and zinc-lead resources are being defined at the Walford Creek deposit (Valenta 2020). Figure 1a shows the surface geology of the North Australian Zinc Belt and the location of sediment-hosted zinc-lead, sediment-hosted copper-cobalt, iron oxide-copper-gold, and orogenic gold deposits. The North Australian Basin System, which hosts the North Australian Zinc Belt, has been subdivided into three superbasins: the 1790-1740 Ma Leichhardt, the 1730-1640 Ma Calvert, and the 1640-1595 Ma Isa superbasins (Southgate et al. 2013;Gibson et al. 2016). The Leichhardt Superbasin consists of a rift, filled with continental tholeiites, fluvial to lacustrine siliciclastic rocks, and minor carbonate rocks Gibson et al. 2018). Figure 2 shows a schematic east-west cross-section illustrating sedimentary facies assemblages and spatial and temporal relationships of the Calvert Superbasin and the locations of basin-hosted deposits within the various facies assemblages. The Calvert Superbasin consists of shallowmarine siliciclastic and carbonate rocks in the west, with deeper-marine siliciclastic rocks in the east, that mostly lack carbonates but contain coeval mafic sills and possible lavas (Fig. 2;Jackson et al. 2000;Southgate et al. 2013;Withnall and Hutton 2013;Gibson et al. 2018). Deposition of the siliciclastic-dominated succession in the east coincided with erosion and development of the Gun unconformity and emplacement of granitic rocks to the west (Neumann et al. 2009). With the conclusion of rifting and rift-related magmatism toward the end of Calvert time, passive margin conditions became widely established across northern Australia. The uppermost Isa Superbasin postdates passive margin formation and consists of fluviatile to deep-marine sandstone, siltstone, and dolostone (Southgate et al. 2000).
The North Australian Basin System has been compartmentalized into third-and fourth-order sub-basins by long-lived, north-northeast-and northwest-striking faults (Fig. 1a). The former are steep to subvertical and thought to have been inherited from the underlying ≥ 1840 Ma crystalline basement (Gibson et al. 2017;Hejrani et al. 2020). The northwest-striking faults were mainly developed during formation of the Calvert Superbasin as a right-stepping en echelon array of crustal-scale normal faults along the western margin of the Lawn Hill and Mount Isa regions, where the majority of zinc deposits are located (Fig. 1a). This fault array originated during northeast-southwestdirected extension and broadly marks the western limits of bimodal magmatism and lithospheric thinning that took place during and before formation of the Calvert Superbasin (possibly indicating a craton edge). Together with the older basement faults, these Calvert-age faults were reactivated during later basin-forming events and strongly influenced the location and distribution of younger, more easterly trending sub-basins of the Isa Superbasin. Calvert-age faulting and rift-related basaltic magmatism concluded at or before ca. 1655 Ma, to be followed by thermal subsidence, basin inversion, and orogenesis from ca. 1650 to ca. 1640 Ma (Riversleigh Event; Withnall and Hutton 2013;Gibson et al. 2017). Subsequent to the Riversleigh Event, extension resumed in a north-south-directed orientation and continued until ca. 1620 Ma when terminated by the onset of the Isan Orogeny (e.g. Gibson et al. 2017). Orogenesis and sedimentation in the Isa Superbasin ended around 1575 Ma. No basaltic rocks are present in this basin with the exception of minor, but locally abundant, tuff units (e.g., Davidson and Dashlooty 1993) and late ca. 1620 Ma rhyolite sills; the basin was largely amagmatic (Gibson et al. 2018 Southgate et al. (2013) and Gibson et al. (2017) indicated that these deposits are associated with the transition between shallow-water shelf and deeperwater basinal environments. These settings bring together the essential mineral system components of saline basinal brines, source rocks, and the deposition of mixed reductants and carbonates necessary for the deposition of zinc and lead sulfide minerals ).

The Northern Cordillera of western North America
The northern segment of the Cordilleran Orogen of western North America (Northern Cordillera) has a large and diverse metallogenic endowment that reflects its diverse geology and a protracted tectonic history that occurred along and offshore of the western and northern Laurentian continental margin since Mesoproterozoic time (Fig. 3a). The Red Dog (Alaska) and Howards Pass (Yukon-Northwest Territories) siliciclastic-carbonate and Sullivan (British Columbia) siliciclastic-mafic-deposits are among the largest shale-hosted, zinc-lead deposits in the world (Leach et al. 2005). Smaller deposits of similar type in proximity to these include the Anarraaq, Lik, and Aktigiruq deposits in the Red Dog district; the Drenchwater deposit in the central Brooks Range (all siliciclastic-carbonate deposits); the Tom, Jason, and recently discovered Boundary deposits in the Macmillan Pass district (Yukon); and the Faro, Dy, Grum, Swim, and Vangorda deposits in the Anvil district (Yukon). Barite deposits are spatially and/or genetically related to many of these Zn-Pb deposits (Moore et al. 1986;Goodfellow et al. 1993;Fernandes et al. 2017;Magnall et al. 2020), and in the Brooks Range of Alaska, some of the barite bodies are extraordinarily large (~ 1 Gt: Kelley et al. 2004a;Kelley and Jennings 2004;Johnson et al. 2004). Sub-seafloor replacement of carbonate (e.g., Kelley et al. 2004a) or barite (Kelley et al. 2004b;Magnall et al. 2020) by metal-bearing fluids has been documented as an important factor for sulfide deposition in many of these deposits. Gibson et al. (2017) indicated that in the Selwyn Basin, like the North Australian Zinc Belt, many of the deposits are spatially associated with the transition from shallower-water shelf into deeper-water basinal environments.
The Northern Cordillera includes autochthonous and parautochthonous strata of the western North American margin and allochthonous, mobile crustal fragments (terranes) that formerly resided in the northeastern Pacific Ocean basin. The tectonic collage was built up through the amalgamation of terranes and their accretion to the continental margin between late Paleozoic and Cenozoic time (Nelson et al. 2013). The geologic terranes have diverse origins and histories and are generally grouped by cratonic associations, lithotectonic setting, and tectonic evolution (e.g., Coney et al. 1980;Silberling et al. 1992;Colpron et al. 2006).
The eastern part of the Northern Cordillera consists of autochthonous and parautochthonous continental margin rocks deposited along the western margin of Laurentia, or the North American Craton, together with overlying Proterozoic to Triassic cover rocks. The North American Craton extends beneath part of the Cordillera (Ross 1991;Cook et al. 2012;Nelson et al. 2013;Hayward 2015). During the early Paleozoic, shallow-water platformal strata and adjacent deeper water basinal strata were deposited along the margin.
In the southern part of the Northern Cordillera, the North American Craton is dominated by the Belt-Purcell Basin. This a Mesoproterozoic, clastic-dominated basin that contains the giant Zn-Pb-Ag Sullivan deposit as well as the Ag-Pb-Zn Coeur d'Alene Cordilleran vein district in Idaho. At ca. 1475 ± 4 Ma (Slack et al. 2020) and ca. 1500 Ma (Ramos and Rosenberg 2012), respectively, these are the oldest zinc-lead deposits that formed along the western margin of the North American Craton.
West of the craton and continental margin rocks, the Northern Cordillera consists of allochthonous terranes, most of which accreted to the margin beginning in the Permian and continuing into Cretaceous time. These terranes include volcanic, plutonic, sedimentary, and metamorphic assemblages of the peri-Laurentian, Arctic-northeastern Pacific, and Coastal "realms" that originated as magmatic arcs, accretionary complexes, microcontinents, and floors of ocean basins (Nelson et al. 2013). The Intermontane terranes of the peri-Laurentian "realm" (Yukon-Tanana, Quesnel, Stikine) have affinities to western Laurentia (North American Craton) and evolved west of the margin following establishment of outboard subduction zones in the Devonian. The Intermontane terranes were progressively accreted to the margin from Permian closure of the Slide Mountain marginal ocean to terminal amalgamation in the Jurassic .
Terranes of Arctic-northeastern Pacific affinity lie west and north of the peri-Laurentian terranes. The Arctic terrane is a large, composite allochthonous crustal block that originated in the northeastern Arctic region near Baltica  or from Laurentia (Canadian Arctic or Atlantic margin: Strauss et al. 2013). These rocks constitute Neoproterozoic to Paleozoic continental shelf-slope strata and younger peri-cratonic rocks of the southern Brooks Range  (Lane et al. 2020), and e conductivity model at a depth of 36 km using data from the AusLAMP magnetotelluric survey (Duan et al. 2021). Locations of significant mineral deposits are overlain as different symbol types. Traces of reflection seismic lines 94MTI-01 and 07GA-IG1 are shown in part (c) ◂ (Moore et al. 1994). The continental fragments that make up the Arctic terrane were derived from both Laurentia and Baltica, which were juxtaposed against northern Laurentia during the Devonian to Mississippian (Strauss et al. 2013) and then underwent collisional orogenesis in the Brooks Range during the Mesozoic (Till 2016). The Arctic terrane also includes the Seward Peninsula, which is primarily underlain by metamorphosed Neoproterozoic to middle Paleozoic rocks with Baltica affinity (Till et al. 2014).
The rest of Alaska, south of the Brooks Range and outboard of the Laurentian and peri-Laurentian terranes, is made up of a diverse assemblage of pericratonic, arc, and oceanic terranes (Western Alaska in Fig. 3a) that were predominately accreted during the Mesozoic (e.g., Ridgway et al. 2002;Trop and Ridgway 2007;Hampton et al. 2010;Box et al. 2019).
Post-Caledonian (Devonian-Mississippian) extension, which affected both Arctic Alaska and the western margin of the North America Craton , provided the setting for formation of some of the major sedimenthosted sulfide and barite deposits. The westward drift of the North American Craton during Cretaceous and early Paleogene time caused intense compressional deformation in the Cordillera and focused in two plutonic-metamorphic belts (Monger and Gibson 2019), the Coastal plutonic complex and the Omineca Belt (not shown in Fig. 3a), which separate the three main divisions of the Cordillera: the outer Arctic-northeastern Pacific (Outboard and Insular) terranes, the central peri-Laurentian (Intermontane terranes), and the North American Craton.

Methods and data sources
Although a relationship between continental margins and the distribution of shale-hosted zinc-lead deposits has been inferred by Nelson et al. (2002) and Leach et al. (2005), identification of such margins in complexly deformed terranes is difficult. In this contribution, we show that a range  (Colpron and Nelson 2020), b variations in μ as determined from lead isotope analyses from mineral deposits and occurrences (Church et al. 1987;Thorpe 2008: Appendix E2), c variations in the depth of lithosphere-asthenosphere boundary as determined from surface wave tomography , and d 30-km upward-continued Bouguer anomaly map (Phillips et al. 1993;Saltus et al. 2008; Geological Survey of Canada 2017). The locations of significant mineral deposits are overlain as different symbols. The trace of Southern Canadian Cordillera reflection seismic traverse is shown in part (c) of geochemical and geophysical datasets can be used to map these margins and can, when combined with basin characteristics, identify prospective provinces at global to district scales.

Lead isotope data
The distribution of lead isotope data is irregular as these were collected almost exclusively from mineral deposits and occurrences. The North Australian Zinc Belt and the Northern Cordillera have among the spatially densest lead isotope coverages in the world. For the North Australian Zinc Belt, we have used the dataset of Carr et al. (2001) supplemented by new data collected as part of this study (Appendix E1). For the Northern Cordillera, our combined dataset is mostly from Church et al. (1987) and Thorpe (2008) All lead isotope analyses were obtained from galena and other ore minerals such as pyrite or from gossans. Analyses were divided into two groups, lead-rich (samples with Pb ≥ 1000 ppm) and lead-poor (< 1000 ppm Pb). For leadrich samples, measured ratios were inferred to approximate initial ratios. For lead-poor samples, initial ratios were calculated by removing ingrown lead where lead, uranium, and thorium concentration data were available and an estimate of the age of mineralization could be made. The initial ratios were used to calculate isotopic ratios using the lead isotope evolution model of Stacey and Kramers (1975).
As many of the deposits/occurrences have yielded multiple analyses, the most precise and least radiogenic (with oldest model age) analysis was used for data contouring. In addition, to guard against extreme ingrowth, which can significantly affect μ, the dataset was culled to remove analyses with young model ages, which can result in anomalously high μ. This process took into account varying ages of deposits and is summarized in Appendix E3. In general, lead-poor samples with model ages 100 Myr younger than the deposit age range in the region were culled from the dataset. Lead-rich analyses were not culled except for samples hosted by younger basins (these samples commonly have anomalously radiogenic J-type lead). Contouring was undertaken using the Esri ArcGIS v10.8.1 average nearestneighbor tool following the methods of Champion and Huston (2016).

Surface wave tomography
Surface waves are a form of seismic energy that travels along the interface between the ground and atmosphere. These seismic waves have typical periods of 10 to 500 s and exhibit dispersive behavior, resulting in particularly high vertical resolution (~ 25 km) of seismic velocity structure in the upper ~ 300 km of the Earth. Since the seismic velocity of silicate rocks is strongly dependent on temperature, this velocity can therefore be used to map variations in the thermal structure of the conducting lithosphere, which is related to lateral changes in plate thickness. In this study, we adopt the scheme of Richards et al. (2020), which uses a recent parameterization for anelastic deformation of mantle rocks at seismic frequencies that is calibrated using independent estimates of mantle temperature structure from oceanic plate cooling models and continental geotherms derived from xenocryst thermobarometry.
We constructed a background upper mantle seismic velocity structure using the global tomographic model SL2013sv (Schaeffer and Lebedev 2013), locally enhanced with higher resolution regional models FR12 in Australia (Fishwick and Rawlinson 2012) and SL2013NA in North America (Schaeffer and Lebedev 2014). Having converted these velocities to temperature, lithospheric thickness is estimated using the depth of the 1175 °C isotherm. Additional details, including data and maps, can be found in Hoggard et al. (2020).

Gravity data
Gravity data measure changes in the gravitational field related to the distribution of mass, including lateral changes in rock density from the Earth's surface to the core. The grids of Bouguer gravity were calculated from national databases available for Australia (Lane et al. 2020), the conterminous USA (Phillips et al. 1993), Alaska (Saltus et al. 2008), and Canada (Geological Survey of Canada 2017). For the Northern Cordillera, 2-km grids of the Bouguer gravity anomalies were merged from Canada, Alaska, and the conterminous USA to create a seamless Bouguer anomaly grid across the study area. The Bouguer gravity grids for Australia and the Northern Cordillera were then upward continued 2 to 30, 50, and 100 km to evaluate how density sources and their edges, as reflected in gravity gradients, change with progressively deeper parts of the lithosphere.

Magnetotelluric data
The magnetotelluric method utilizes variations in Earth's naturally occurring magnetic field and induced electrical fields to derive resistivity structures of the crust and upper mantle. Long-period (10-10,000 s) magnetotelluric data were acquired in northern Australian as part of the Australian Lithospheric Architecture Magnetotelluric Project (AusLAMP) under the Geoscience Australia's Exploring for the Future (EFTF) program. Longer-period data are sensitive to features far from the observation site, including the upper mantle, whereas shorter-period data can be used to image closer features in the middle to upper crust. Resulting 3D resistivity models at depths of ~ 10 km to ~ 200 km give insights into the distribution of conductive rocks through the lithosphere. Further details are found in Duan et al. (2021). A similar study has not been undertaken across the Northern Cordillera, precluding consideration of this dataset for that region.

Seismic reflection data
Since 1994, Geoscience Australia, in collaboration with the Geological Survey of Queensland, acquired a series of deep seismic reflection transects in northern Queensland (Korsch et al. 2012;Korsch and Doublier 2016). Over 3500 km of data were acquired on 16 traverses. Since the early 2000s, the seismic data has been acquired using three Hemi-60 vibrators with data collected to a 20 s two-way travel time. The data were processed using the Disco/Focus seismic processing package. The processed images were interpreted by a team of geologists and geophysicists from Geoscience Australia, the Geological Survey of Queensland, and collaborators from Australian universities. Korsch et al. (2012) provide information on the surveys and final interpretations; Jones et al. (2009) reported on processing methods.
Between 1984 and 1988, Lithoprobe, Canada's multidisciplinary research program for earth sciences, in collaboration with the Geological Survey of Canada, acquired seismic reflection data along a traverse in the southern part of the Northern Cordillera (location shown in Fig. 3c), just to the north of the US border (at ~ 50°N latitude). This transect extended from the North American Craton in the east to the west coast (Cook 1995;Hammer et al. 2010) and involved 19 discrete seismic lines that were merged to form the Cordilleran transect. The data were collected using four 20,000 kg vibrators to 18 s two-way travel time. Data were processed using non-standard approaches as described by Cook et al. (1992). The data were interpreted by geophysicists and geologists from Canadian universities and the Geological Survey of Canada. Clowes (1990) and Cook et al. (1992) document data acquisition, processing, and interpretation.

Variations in lead isotope data
Previous studies (Huston et al. 2014;Hollis et al. 2019) have shown that variations in parameters such as μ ( 238 U/ 204 Pb), as determined from lead isotope data from mineral deposits, can reveal spatial controls and indicate fertility of mineral provinces. Figure 1b shows spatial variations in μ in the North Australian Zinc Belt. Along the eastern margin of this image, μ increases from northeast to southwest. Significantly, most siliciclastic-carbonate zinc-lead deposits and the Tick Hill orogenic gold deposits are localized along a north-northwest-trending break in μ values. Deposits occurring along the break are relatively consistently spaced about 140 km apart. Iron oxidecopper-gold and siliciclastic-mafic zinc-lead deposits are located to the east of this break and are characterized by lower μ. Other parameters determined from lead isotope data, including κ ( 232 Th/ 238 U) and ω ( 232 Th/ 204 Pb), define a similar break (not shown).
In detail, the prominent north-northwest-trending break in μ appears to be en echelon, broadly following the trend of the northwest-trending Calvert-age (1730-1640 Ma) structures, with individual offsets or steps controlled by the position of the underlying north-northeast basement structures (Fig. 1b). As both sets of structures are thought to have been active during formation of the Calvert and Isa superbasins and related sub-basins, it would appear that variations in μ mimic development of these sub-basins along the isotopic break. Although the break in μ appears to be controlled by basement structures, the break itself cuts across the north-south grain of the surface geology and across trends in total magnetic intensity, particularly in the east. Figure 3b shows spatial variations in μ for the Northern Cordillera. Like the North Australian Zinc Belt, all major shale-hosted zinc-lead deposits, except those of the Red Dog district, are spatially associated with a break in μ values from higher values inboard to lower values outboard of the boundary of the North American Craton and the allochthonous terranes. Further outboard of the North American Craton, the μ continues to decrease with major VHMS districts associated with zones of lowest μ, similar to the relationship noted for Archean and Paleoproterozoic VHMS deposits elsewhere in the world (Huston et al. 2014). In detail, major shalehosted zinc-lead districts are associated with higher-order complexities in the μ distribution patterns such as re-entrants along the ancient continental margin.
Of all the datasets considered, the spatial relationship of deposits with μ gradients is spatially the most precise, but this is only possible due to the density of data in both provinces. Very few zinc-lead provinces globally have a similar density of lead isotope data. The only province with a comparable data density is the Irish Midlands, where deposits are also associated with a gradient in μ (Hollis et al. 2019). Collectively, the distribution patterns in the North Australian Zinc Belt, the Northern Cordillera, and the Irish Midlands suggest that local zones (i.e., districts) with complex variations in μ (e.g., re-entrants) may define more prospective areas within basins. As such, lead isotope datasets may be useful in identifying new plays at the district-scale, even in brownfield provinces. Figure 1c shows that the locations of major deposits in the North Australian Zinc Belt coincide with the transition from thicker to thinner lithosphere, mapped using surface wave tomography . All deposits in this beltincluding stratiform zinc-lead deposits hosted by carbonaterich and siliciclastic-dominated successions and epigenetic iron oxide-copper-gold deposits-are within 100 km of the 170-km thickness contour of the lithosphere-asthenosphere boundary (LAB). Major deposits are unknown outside of this corridor, even though the North Australian Basin System extends well beyond it. Similar to the lead isotope pattern, the edge of thick lithosphere cuts across the regional geological and total magnetic intensity grains to the east and north of McArthur River (compare Fig. 1b and c with Fig. 1a), suggesting a fundamental, deep-seated control on mineralization. Figure 3c shows that there is a broad correlation of basin-hosted zinc-lead deposits in the Northern Cordillera with the 170-km LAB contour, although in detail some of these deposits are located outboard of this contour (e.g., Macmillan Pass and Howards Pass districts). Hayward and Paradis (2021) suggested that the host basin may have been displaced 100 km to the northeast relative to basement.

Lithosphere-asthenosphere boundary as defined by surface wave tomography
Analysis by Hoggard et al. (2020) of a global correlation between LAB depth and the distribution of sediment-hosted deposits showed that 85% of zinc resources and 85% of copper resources in basins are located within 200 km of the 170-km contour. The correlation with the LAB, originally identified in the North Australian Zinc Belt, thus applies globally.

Gravity data
Indications that sediment-hosted zinc-lead deposits were associated with regional-scale breaks in geophysical data were originally documented by Hobbs et al. (2000). This study reported that major zinc-lead deposits in the North Australian Zinc Belt are aligned along a composite gravity "worm" or horizontal gradient in Bouguer anomaly gravity data. Persistence of this worm through upward continuation of the data indicates a major crustal discontinuity at depth. Figure 1d illustrates variations in the Bouguer gravity anomaly for the North Australian Zinc Belt upward-continued to 30 km. The image defines a broadly north-northwest-trending gravity high in Queensland that turns to the northwest in the Northern Territory. In Queensland, this gravity high is flanked by low-gravity anomaly zones to the west and east and is transected by several northeast-trending discontinuities that translate the southern part of the highgravity zone to the southwest (i.e., dextral displacement). The western margin of the high gravity anomaly zone is the Rufus Fault, the eastern margin corresponds to the Gidyea Structure (Figs. 4a,b and 5), and the main discontinuity corresponds to the Quamby-Fountain Range Fault at surface. In the far south, the gravity high is truncated by the Cork Fault, which juxtaposes the North Australian Craton with the Thomson Province of the Phanerozoic Tasman Element to the southeast. Similar, though less well-defined, patterns are present in the 15-, 50-, and 100-km upward-continued images (not shown). Using the basin architecture model of Southgate et al. (2013: Fig . 2), the gravity high is associated with the outboard, deeper-water part of the basin, whereas the gravity low is associated with the inboard, shallow water part of the basin that existed during Calvert and Isa time. The horizontal gravity gradient appears to mark the continental margin during evolution of the North Australian Basin System.
Most siliciclastic-carbonate shale-hosted zinc-lead deposits are localized along the gravity gradient that defines the western margin of this high gravity anomaly, which largely coincides with the 170-km LAB contour. The McArthur and Walford Creek districts are spatially associated with reentrants on the western margin of the gravity high. Siliciclastic-mafic shale-hosted zinc-lead deposits (e.g., Cannington) and iron-oxide copper-gold deposits (e.g., Ernest Henry) are restricted to the eastern edge of the high gravity anomaly zone. The western margin of the high gravity anomaly and the 170-km LAB contour broadly coincide to the north of Mount Isa. In the vicinity of Mount Isa, however, the gravity gradient and 170-km LAB contour diverge (compare Fig. 1c and d). The significance of this divergence is not known and requires further analysis.
Upward continuation of gravity is often used to distinguish shallow from deeper density sources (Jacobsen 1987). Although shallow sources can be uniquely identified, longerwavelength gravity anomalies can result from either deeply buried sources or from extensively distributed shallow sources such as basins. As shallow geological features in this area comprise Phanerozoic basins without significant volcanism (Fig. 1a), the high gravity anomaly likely reflects a deep, dense feature. In upward-continued gravity anomalies at a height z, the depth of the sources is approximately z/2 or deeper (Jacobsen 1987). As the gravity gradient is present in the 30-50-, and 100-km upward continuations, the change in density structure persists from the mid-crust into the lithospheric mantle. The gravity gradient associated with the siliciclastic-carbonate shale-hosted deposits separates low-density rocks to the southwest from higherdensity rocks to the northeast. The Leichhardt River Trough is positioned over these denser rocks and contains a large thickness (to 8 km) of mafic igneous rocks (Bain et al. 1992).
Major shale-hosted deposits in the Northern Cordillera of western North America are also associated with a (smaller amplitude) gradient in the 30-km upward-continued gravity image (Fig. 3d) as well as with gradients in the 15-, 50-, and 100-km upward-continued images (not shown), as also noted by Hayward and Paradis (2021). The relationship in this province differs from that in the Northern Australian Zinc Belt in two important ways. First, the gradient is not as strong as in the North Australian Zinc Belt, and second, the gradient has the opposite sense: the  ) and seismic transects (Cooke et al. 1992;Drummond et al. 1997;Goleby et al. 1998;Korsch et al. 2012): a east-west transect across the southern part of the North Australian Basin System (based on reinterpretation by R Korsch and M Doublier (unpublished) of 94MTI-01 (Drummond et al. 1997;Goleby et al. 1998)), b northeast-trending transect across the eastern margin of the North Australian Basin System showing the geometry of the Gidyea Structure (based on seismic line 07GA-IG1 (Korsch et al. 2012) in the northwest and extended to the south-west based on extrapolation of 94MTI-01 (Drummond et al. 1997;Goleby et al. 1998)), and c east-west transect across the southern part of the Northern Cordillera of western North America. Crosssection (a) equates to the southwestern part of 07GA-IG1 as extended to the southwest as a dashed line in Fig. 1c. Cross sections (b) and (c) equate to 94MTI-01 and Lithoprobe seismic traverses shown in Figs. 1c and 3c, respectively. Approximate location of significant nearby deposits have projected onto the transects. The Sullivan deposit projects just to the east of the eastern extent of (c). Symbols for deposit types are after Fig. 1 low-gravity zone is outboard relative to margin of the North American Craton. Figure 1e shows the 36-km depth slice of the Northern Australia magnetotelluric electrical resistivity model (Duan et al. 2021). Unlike other mineral systems, in which deposits are associated with conductors (Heinson et al. 2018 for IOCG;Kirkby et al. 2022 for orogenic gold), the siliciclastic-carbonate zinc-lead deposits appear to be associated with a resistive zone between two more conductive zones. In the vicinity of Mount Isa, this resistivity anomaly trends north-south and broadly corresponds to the outcrop expression of the Kalkadoon-Leichhardt Belt, which is dominated by granites and gneisses. To the south of Mount Isa, this resistive zone trends toward the Cork Fault, although the lack of data in this region precludes determination of the southern extension of this zone. To the north of Mount Isa, the resistive corridor turns to the northwest and tracks toward the McArthur district, with all major siliciclastic-carbonate deposits, except Dugald River, located above the high resistivity anomaly.

Magnetotelluric data
To the east (northeast) and the west (southwest), the middle crust is significantly more conductive (Fig. 1e). Iron-oxide copper-gold and related epigenetic deposits of the Cloncurry district are spatially associated with the eastern high conductivity zone (Duan et al. 2021) as seen in other IOCG provinces such as Olympic Dam (Heinson et al. 2018). Siliciclastic-mafic zinc-lead deposits, such as Cannington and Pegmont, are also associated with this zone. The moderate to high conductive zone to the west of Mount Isa, however, is not associated with significant known mineral deposits of any type.
At the time of writing, insufficient magnetotelluric data have been collected in the Northern Cordillera to assess whether similar relationships between crustal electrical structure and the distribution of sediment-hosted zinc-lead deposits are present in that province.  (Korsch et al. 2012) and 94MTI-01 of the 1994 survey (Goleby et al. 1998). These transects were chosen as they illustrate the crustal architecture of the North Australian Zinc Belt and cross gradients in μ, LAB depth, and gravity and magnetotelluric data as described above (locations shown in Fig. 1c). Figure 4a and b have been extended downwards to show the base of the lithosphere using depth contours of the LAB .
As shown in Fig. 4a and b, the southern part of the North Australian Zinc Belt comprises three crustal blocks that are divided by two north-south-trending, crustal-scale structures, the west-dipping Gidyea Structure 3 in the east and the steeply west-dipping Rufus Fault in the west. These structures are offset by the northwest-trending Quamby-Fountain Range Fault (Fig. 5), which offsets the gravity high in the 30-km upward-continued gravity image (Fig. 1d), as described above. Both the Gidyea Structure and the Rufus Fault are associated with offsets of the Moho, and both are considered major crustal boundaries (Korsch and Doublier 2016). Figure 4c shows a Northern Cordillera cross-section from the western margin of the North American Craton in the east to Fraser River in the west (location shown in Fig. 3c) using the interpretation of Cook et al. (1992) for the composite Lithoprobe Northern Cordillera transect and the LAB depth model of Hoggard et al. (2020). Like the North Australian Zinc Belt transect, the Northern Cordillera section also indicates major crustal changes away from the margin of the North American Craton. The crust thins from east to west, accompanied by a major step in the Moho depth that also approximates the 170-km LAB depth contour. Similar to the North Australian Zinc Belt, the gradients seen in the lead isotope data and upward-continued gravity data, and distribution of shale-hosted zinc-lead deposits, appear to coincide with an edge, in this case the thinning of the North America Craton.

Discussion
Disparate datasets suggest that basin-hosted zinc-lead deposits in the North Australian Zinc Belt, the Northern Cordillera of western North America, and elsewhere in the world are associated with changes in the isotopic character, thickness, and physical properties (density and resistivity) of the crust and lithosphere. Data from the North Australian Zinc Belt suggest that these changes correspond to pre-existing boundaries between crustal blocks, whereas data from western North America suggest that deposits spatially coincide with the western boundary of the North American Craton.

North Australian Zinc Belt
The present-day architecture of the southern part of the North Australian Zinc Belt consists of four major crustal blocks separated by three major structures (Figs. 4a and b and 5). The Gidyea Structure juxtaposes thicker, conductive crust of the Mount Isa Province to the west with thinner, resistive crust of the Kowanyama/Numil Seismic Province to the east (Fig. 4a). Within the Mount Isa Province, the eastdipping Pilgrim Fault juxtaposes conductive crust in the east from resistive crust to the west. Iron-oxide copper-gold and related deposits, and siliciclastic-mafic shale-hosted zinclead deposits, are mostly restricted to conductive Mount Isa crust, whereas siliciclastic-carbonate shale-hosted zinc-lead deposits are localized in the resistive corridor to the west. The Rufus Fault juxtaposes Mount Isa crust with thinner Aileron Province crust to the west (Fig. 4a). Lithospheric thickness decreases from west to east, with the 170-km LAB contour between the surface projection of the Gidyea Structure and the Rufus Fault.
In detail, trends determined from isotopic and gravity data, and the trend of the deposits themselves, cut at a broad scale the crustal-scale architecture described above. Moreover, the change in orientation to the northwest of the lead isotope, gravity, LAB-depth, and deposit trends and the resistive corridor are not obvious in the trends determined from surface geology and magnetic anomalies. In Fig. 5, we present a structural model that can account for the observed trends in multiple datasets and the distribution of mineral deposits using existing tectonic and basin architecture models for the North Australian Basin System (Neumann et al. 2009;Southgate et al. 2013;Gibson et al. 2016Gibson et al. , 2018Gibson et al. , 2020. This interpretation assumes that initiation of the Gidyea Structure and the Rufus Fault predated the formation of the North Australian Basin System and that their geometry strongly influenced the architecture of the North Australian Basin System. Gibson et al. (2008Gibson et al. ( , 2012 argued that the evolution of the basin system between ca. 1790 and ca. 1620 Ma was strongly influenced by eastnortheast-west-southwest-to northeast-southwest-directed extension associated with the rifting of Laurentia from the North Australian Craton. Korsch et al. (2012) argued that the Gidyea Structure formed before ca. 1850 Ma.
Our interpretation (Fig. 5) conceptually involves the formation of a basin during northeast-southwest extension and is based mostly on gravity data but uses all datasets described herein. Three sets of faults controlled the geometry of the basin: (1) newly formed, northwest-striking extensional faults; (2) north-trending trans-tensional faults that reactivated pre-existing crustal-scale faults such as the Gidyea Structure and Rufus Fault; and (3) deep-seated northeast-striking transfer faults. In this interpretation, the edges created by the extensional and trans-tensional faults controlled the sedimentary facies architecture of the North Australian Basin System (cf. Southgate et al. 2013; Fig. 2). Smaller-scale extensional and transfer faults, developed on this large-scale architecture, likely compartmentalized local sub-basin development (Gibson et al. 2016(Gibson et al. , 2018. The complex, small-scale en echelon distribution pattern of lead isotope data, reflected by μ values (Fig. 1b), likely reflects the development of such basins, and the re-entrants in gravity anomalies in the McArthur River and Walford Creek districts (Fig. 1d) may similarly reflect sub-basin development.

Northern Cordillera
The present-day Northern Cordillera consists of a series of allochthonous and para-autochthonous terranes accreted onto the western margin of the North American Craton. This margin is irregular with many sub-basin embayments and basement-cored promontories controlled by local extensional and transform faults. Changes in the width and thickness of sedimentary rocks within these sub-basins reflect asymmetric extension (Thomas 1977;Lund 2008;Colpron and Nelson 2021). The broad-scale breaks in both upwardcontinued gravity (Fig. 3c) and μ data (Fig. 3b) likely reflect this margin, but the complex, smaller-scale patterns in μ reflect the development of sub-basins along this margin.
Shale-hosted zinc-lead deposits in the Northern Cordillera occur along linear trends that may be related to basement structures that originated during rifting along the margin of the North American Craton (Lund 2008;McMechan 2012). These structures influenced the geometry of the rift system and provided conduits for magmas and/or hydrothermal fluids. Northwest-striking asymmetric extensional segments and crustal thinning in the region are subdivided by northeast-striking transform and transfer segments.
Asymmetric extension produced margins, or edges, that mark changes in the character of both basin and basement. Lower-plate margins (i.e., Selwyn Basin and Mackenzie Platform in Yukon) are marked by the thinning of continental crust, subsidence (producing thick sag basins), and rotated crustal blocks. Upper-plate margins (i.e., south of the Laird line up to the Vulcan Low-St. Marie-Moyie transform zone; Fig. 1 of Lund et al. 2010) are characterized by limited crustal thinning and narrower continental margins (Thomas 1993(Thomas , 2006. Most of the largest shale-hosted zinc-lead deposits in the Northern Cordillera are located in the segment associated with the lower plate margin, including those in the Selwyn Basin and Mackenzie Platform. Another group of important deposits, including Sullivan, is in southern British Columbia where the Vulcan Low-St. Marie-Moyie transform zone, a deep and old crustal structure within the North American Craton (Hoffman 1988), defines an upper-plate margin. Northeast-striking transform and transfer faults played important roles on sedimentation, intrusion, deformation, and mineralization (Lund et al. 2010), and many sedimenthosted deposits are located along these faults (e.g., Sullivan located along the Vulcan Low-St. Marie-Moyie structure).
Like the North Australian Zinc Belt, major shale-hosted zinc-lead deposits appear to be associated with cratonic edges, determined both geologically and using lead isotope data, the depth of the LAB and gravity data. It also appears that local-scale features, such as basement structures, also were an important control on the distribution of the deposits. Similar to the North Australian Zinc Belt, some of the finescale complexity in μ variations may reflect these structures.

Edges-fundamental controls on the distribution of shale-hosted zinc-lead deposits
Continental or cratonic edges, which can be outlined using a range of isotopic and geophysical datasets, appear to have exerted a large influence on the distribution of shale-hosted zinc-lead deposits. In the North Australian Zinc Belt, the older (ca. 1680 Ma) siliciclastic-mafic deposits occur outboard of the platform edge in deep-water turbiditic successions that formed during the early rift phase of basin development. In contrast, the younger (ca. 1665-1575 Ma), siliciclastic-carbonate deposits formed along or inboard of the edge as the basin continues to develop. This distribution of both deposit types is shown schematically in Fig. 2. In the Northern Cordillera, the relationship is broadly similar. Figure 6 presents an evolutionary model for the North Australian Zinc Belt illustrating how edges have controlled the metallogeny of this province. Figure 7 shows the general relationships of deposits to edges and geochemical and geophysical proxies of edges. Figure 6 shows conceptually how a basin-hosted tectonometallogenic system may have evolved, based on geometric relationships in the southern part of the North Australian Zinc Belt. In this model, the basement architecture of the showing relationship of these systems to crustal edges: a stage 1, preconditioning; b stage 2, initial rifting; c stage 3, formation of rift basin; and d formation of sag basin ◂ zinc belt is inherited from pre-existing major crustal boundaries and/or structures, including the Gidyea Structure and the Rufus and Pilgrim faults (Fig. 6a). These faults separate domains with different physical (i.e., density, seismic character, and resistivity structure) and isotopic properties, and, as such, represent the edges of major cratonic blocks that controlled later basin evolution and metallogenesis.

Architecture of the North Australian Zinc Belt
At ca. 1790 Ma, extension initiated, leading to deposition of the Leichhardt Superbasin and the basal part of the Calvert Superbasin over the following 140 Myr (Fig. 6b). Although shown simply as an extensional basin with edges defined by pre-existing boundary faults, in detail this early extension may have involved formation of core complexes and other more complex extensional geometries (e.g., Gibson et al. 2008). Rocks deposited at this time are mainly tholeiitic mafic volcanic rocks, lacustrine and shallow marine sandstone and other siliciclastic rocks, and minor carbonate rocks. This deposition continued to ca. 1680 Ma (Fig. 6c), at which time a deeper rift basin formed in the eastern part of the Calvert Superbasin, partly controlled by the early formed structural architecture, particularly the Pilgrim Fault. This depositional system was dominated by turbiditic, siliciclastic rocks but included mafic volcanic rocks. Siliciclastic-mafic deposits such as Cannington formed at this time. The deposition of this rift basin corresponded in time with erosion and the development of the regional Gun unconformity in the western part of the North Australian Basin System (Neumann et al. 2009;Southgate et al. 2013).
Rifting ceased at ca. 1670-1665 Ma, and the basin system went into a sag phase (Fig. 6d). At this time, magmatism became very minor and sedimentation involved deposition of siliciclastic-and carbonate-dominated facies that changed laterally from the shelf outward and eastward into deeperwater sediments (Southgate et al. 2013: Figs. 2 and 6d). The transition from the shelf into deeper water in this sedimentary system was likely controlled by the edges formed during earlier extension. This transition from carbonate-rich successions on the platform to calcareous/dolomitic shale successions in deeper water was important for the development of mineral systems that form siliciclastic-carbonate shalehosted zinc-lead deposits (e.g., Emsbo et al. 2016). These deposits formed throughout much of the sedimentary system, from carbonate-rich succession on the platform (e.g., McArthur River and Lady Loretta) to successions where carbonate is not as dominant (e.g., Mount Isa, Dugald River).
Although the model presented in Fig. 6 does not apply directly to the Northern Cordillera, the presence of an edge, as indicated in μ variations and variations in the LAB and in gravity data, had a strong control on the distribution of most shale-hosted zinc-lead deposits in this orogen (Fig. 3).  Huston et al. 2016). The relative resolution of different datasets discussed in the text in defining prospective basins and zones within basins for these deposits is shown. The diagram is generalized, and the characteristics and utility of the datasets presented vary between basins. Other factors that should be considered in basin analysis include age, paleoenvironmental characteristics of the basin (e.g., paleolatitude and/or evidence of evaporites), tectonic triggers (e.g., out-of-area structural and/or tectonic events as indicated by bends in paleomagnetic apparent polar wander paths or local structural events), the presence of and evidence for leaching of source rocks (e.g., mafic volcanic rocks and certain turbiditic sedimentary rocks), and the presence of trap rocks (mainly carbonaceous and dolomitic shale/siltstone, but in some cases, dolostone breccia) Moreover, the range in age of the deposits suggest that this margin was long-lived, with mineralizing events occurring at ca. 1475 Ma,ca. 560 Ma,ca. 520 Ma,ca. 440 Ma, Ma.

Controls on siliciclastic-mafic and siliciclastic-carbonate shale-hosted deposits
The spatial association of the Cannington deposit with amphibolite (Wright et al. 2017), interpreted as a mafic volcanic rock or sill, and the presence of mafic sills and dikes within the host Soldiers Cap Group (Derrick et al. 1976;Beardsmore et al. 1988;Gibson et al. 2018), indicates significant coeval magmatism in a relatively deep rift basin. It is likely that the edges, as discussed above, controlled the development of this basin. Moreover, lithospheric thinning and mafic magmatism associated with rifting likely resulted in a high heat flow, which drove the circulation of relatively hot, reduced ore fluids that deposited metals in the shallow subsurface below the seafloor (Cooke et al. 2000;Huston et al. 2006;Wright et al. 2017). The metal source was likely mafic volcanic rocks or immature, turbiditic siliciclastic rocks within the rift basin.
In contrast, siliciclastic-carbonate zinc-lead deposits are thought to have formed from lower temperature, oxidized ore fluids (Cooke et al. 2000;Huston et al. 2006). Modeling by Hoggard et al. (2020) has suggested that thicker subcontinental lithospheric mantle and crust insulates overlying basins from hot asthenosphere mantle during mild to moderate rifting, resulting in lower thermal gradients within rift basins. In addition, thicker basins develop during rifting of the thick and buoyant lithosphere. These two factors are combined to produce voluminous sediments in basins having lower-temperature brines, which, if oxidized, can be highly efficient at transporting base metals (Cooke et al. 2000;Huston et al. 2016). Moreover, the mafic volcanic rocks deeper in the basin are ideal source rocks for zinc (Cooke et al. 1998(Cooke et al. , 2000. Cratonic edges juxtapose contrasting geological and tectonic environments in space and time-environments that provide essential components (architecture, fluids, fluid pathways, source rocks, and traps) for both siliciclastic-mafic and siliciclastic-carbonate zinc mineral systems (Fig. 7). In this study, we have demonstrated that cratonic edges can be mapped using diverse datasets-lead isotopes, surface-wave tomography, gravity, seismic reflection, and magnetotelluric data-and that the locations of deposits correspond closely with the mapped edges. Moreover, these edges have persisted and localized deposits for long periods of geological time. In the North Australian Zinc Belts, deposits formed along the edge for over 100 Myr, and in the Northern Cordillera, deposits formed along the edge at ca. 1475 Ma and for over 100 Myr in the Paleozoic.
Although we emphasized the relationship of shale-hosted zinc-lead deposits with the large-scale edges (Fig. 7), in detail, the locations of many deposits seem to be controlled by second-order features along these large-scale boundaries. In particular, the Walford Creek deposit and those in the McArthur district in the North Australian Zinc Belt are associated with embayments in the 30-km upward-continued gravity image (Fig. 1d); some deposits in the Northern Cordillera are associated with complexities in μ patterns (Fig. 3b). The McMillan Pass and Howards Pass districts are associated with an embayment in the gravity data (see also Hayward and Paradis 2021). These relationships suggest that anomalous, smaller-scale features in isotopic, gravity, and other datasets may be useful in defining more prospective zones along the larger-scale margins, although more detailed data are required to test this concept.
In the North Australian Zinc Belt and to a lesser extent in the Northern Cordillera, most deposits are localized where edges identified by all datasets coincide (Fig. 7). Outside of these areas of coincidence, major deposits are not known, although much of the area is covered by younger basins, at least in the North Australian Zinc Belt. Outside of the areas of known mineralization, locations of edges, as determined from the different datasets, diverge, which raises the question: which dataset provides the best vector for targeting at different scales? More data and examples are required to resolve this question.
Lastly, evidence presented by Young (2004) for the Red Dog tectonic plate, which hosts the Red Dog district, and Hayward and Paradis (2021) for the Selwyn Basin, which hosts the MacMillan Pass and Howards Pass districts, suggests that some basins that contain shale-hosted zinc deposits have been transported significant distances relative to basement by post-basin thrusting. This inferred post-ore tectonic displacement may account for local divergence of deposits from geophysical features identified in the basement and underlying mantle (e.g., the LAB 170-km contour: Hayward and Paradis 2021) and the location of deposits. Although these deeper features provide criteria to identify mineralized provinces or basins at the global scale, more precise location information may be provided by datasets that directly map the characteristics of the host basin and are independent of post-mineral translation. These datasets include lead isotope data and more detailed gravity data.

Other criteria bearing on basin fertility
Global reviews on the characteristics and temporal distribution of basin-hosted zinc-lead deposits (Leach et al. 2005(Leach et al. , 2010Wilkinson 2014) provide additional insights on important controls of these deposits. The oldest major basin-hosted zinc-lead deposit has an age of ca. 1800 Ma (Rampura-Agucha in India; Deb and Thorpe 2004) and was succeeded by a major period of deposit formation, including deposits from the North Australian Zinc Belt and the Northern Cordillera, at ca. 1800-1475 Ma. Formation of this cluster followed the development of an oxidized, upper layer of the ocean (Farquhar et al. 2010), which allowed, for the first time, the development of oxidized fluids essential for hydrothermal transport of metals in cool basins that formed over thickened lithosphere. Hence, a basin age younger than ca. 1800 Ma can be used as an additional criterion to define prospective basins.
Information on paleoenvironments at the time of mineralization can also refine the search space. Based on data for Phanerozoic deposits, Leach et al. (2005) suggested that virtually all basin-hosted zinc-lead deposits formed within 30° (mostly 10-30°) from the equator, the equivalent of modern desert belts. Paleomagnetic data suggest that the North Australian Zinc Belt was within 40° of the equator at the time of zinc-lead mineralization (Idnurm 2000). These data indicate that an arid littoral environment, critical for the formation of evaporative brines, is essential to form basin-hosted zinclead deposits. In addition to paleomagnetic data, the presence of evaporites, evaporitic textures, evaporitic minerals, or mineral or textural pseudomorphs of evaporitic minerals within the host succession may also indicate favorable basins (Davidson and Dashlooty 1993;Czarnota et al. 2020). Idnurm (2000) showed that many of the major siliciclastic-carbonate zinc-lead deposits in the North Australian Zinc Belt are temporally associated with bends in the apparent polar wander path of the North Australian Craton. The 1635-1645 Ma McArthur River and Lady Loretta deposits correspond in time not only to a major bend in this path, but also to the Liebig Orogeny (Scrimgeour et al. 2005) on the southern margin of the North Australian Craton and the Riversleigh Tectonic Event within the North Australian Basin System. Hence, the trigger of fluid flow in the North Australian Zinc Belt may be regional or out-of-area structuraltectonic events that are apparent in the paleomagnetic and/or structural history of the North Australian Craton. Similarly, the formation of the Sullivan deposit (~ 1475 Ma; Slack et al. 2020) coincides in time with a major bend on the North American apparent polar wander path (Elston et al. 2002;Leach et al. 2010), and the Howards Pass (~ 442 Ma; Kelley et al. 2017) and Red Dog (~ 338 Ma; Morelli et al. 2004) deposits are broadly associated with flexures in the Paleozoic North American polar wander path (Cocks and Torsvik 2011). This information may provide time constraints that can further refine the exploration search space in those regions. Cooke et al. (1998) and  have demonstrated that zinc and copper in many basaltic rocks from the Leichhardt and Calvert superbasins were depleted in association with regional chlorite-, K-feldspar-, and/or hematite-bearing alteration assemblages. Preliminary data suggest that copper leaching may be associated with demagnetization of the basalt . The presence of such altered rocks is an additional characteristic of a fertile basin. Insufficient data are available to test this concept in basins developed at various times along the margin of the North American Craton.
All significant siliciclastic-carbonate zinc-lead deposits in the North Australian Zinc Belt are hosted by carbonaceous and dolomitic siltstone lenses within carbonate-and/ or sandstone-rich successions Czarnota et al. 2020). These fine-grained rocks most likely formed in restricted, deeper-water depocenters and acted as redox traps to metals deposited from oxidized ore fluids, either epigenetically (Broadbent et al. 1998;Spinks et al. 2021), diagenetically in the shallow subsurface (Williams 1978;Kelley et al. 2004a,b;Gadd et al. 2016Gadd et al. , 2017, or syngenetically on the sea floor (Carne and Cathro 1982;Goodfellow et al. 1993;Large et al. 2005). The presence of these finegrained rocks enhances the fertility of basins for siliciclasticcarbonate zinc-lead deposits (although not for some other basin-hosted deposits); hence, the identification of spatial and temporal positions of these depocenters can further constrain the search space.
Our results indicate that other disparate datasets, including stratigraphic, geochemical, and paleomagnetic data, may further refine spatial and/or temporal domains within basin systems that are most prospective. Moreover, it is possible that potential field, facies architecture, and, possibly, lead isotope data can be used to identify prospective domains at the district scale.

Conclusions
Although cartoon-like diagrams of basin-hosted mineral systems (e.g., Large 1980;Leach et al. 2005Leach et al. , 2010Huston et al. 2016) have long suggested that these deposits form along the edges of continental blocks on passive margins, this study shows that data such as ore lead isotope analyses, surface-wave tomography converted to temperature, upwardcontinued gravity, and, possibly, low-to mid-crustal magnetotelluric data can map the distribution of these cratonic margins (Fig. 7), even if the margins have been involved in later collisional events. The surface-wave tomography data, in particular, suggest that continental margins are associated with thicker lithosphere; seismic-reflection data further suggest that the formation of passive margins localized along pre-existing crustal boundaries are more prospective for major and world-class shale-hosted zinc deposits. These features appear to be a first-order control on basins that are highly fertile for basin-hosted mineral deposits and can be used as exploration guides in area selection. Importantly, there are ways to screen more prospective from less prospective basins and to determine the most prospective parts of prospective basins. Our results and previous data suggest that more prospective basins are (1) no older than ca. 1800 Ma, (2) formed within 30-40° of the equator, (3) overlap in age with bends in apparent polar wander paths, and (4) overlap in space with cratonic edges identified using variations in radiogenic isotopes and geophysical data (Fig. 7). The most prospective parts of basins are indicated by the location of cratonic edges, perturbations in these edges, sedimentary facies changes in time and in space, evidence for the presence of evaporites in the sedimentary succession, and existence of extensive syn-basin faults.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.