Cenozoic Tectonic Reconstruction and the Initial Distribution of Porphyry Copper Deposits in the Sonoran Desert Region of Southwestern North America: Implications for Metallogenesis

The Sonoran Desert region of southwestern North America encompasses one of Earth’s largest porphyry copper provinces, with at least 39 deposits with >10⁵ tonnes of contained copper (production, reserves, and resources) and, in many deposits, economically significant molybdenum (Fig. 1; Cooke et al., 2005; Leveille and Stegan, 2012). Widely spaced porphyry copper deposits are present around the margins of the Pacific Ocean and at a few other locations around the world, but exceptionally rich clusters of deposits are known only in southwestern North America and in the Andes of Chile and Peru (Sillitoe, 2012). Many of the Andean deposits are associated with the intersections of linear fault sets, with the major fault set parallel to the continental margin (Farrar et al., 2023). In contrast, the Sonoran Desert deposits are geographically scattered, with three linear trends within the province but no overall order.

The Sonoran Desert porphyry copper province, encompassing the deposits shown in Figure 1A, is almost entirely Laramide in age (~50–80 Ma) (Titley, 1995; Titley and Zürcher, 2008; Greig and Barton, 2019; Seedorff et al., 2019). Most deposits are within the tectonically extended Basin and Range province, and a few are within the southern Colorado Plateau. Laramide orogenesis and porphyry copper mineralization were associated with subduction of oceanic lithosphere along the Pacific continental margin and marked the beginning of a period of low-angle subduction with magmatism east of the older coastal batholith belt (Coney and Reynolds, 1977; Severinghaus and Atwater, 1990). Laramide orogenesis was followed by middle to late Cenozoic tectonic extension that fragmented the crust and dispersed porphyry copper deposits so that their Laramide distribution is now obscure (Spencer and Reynolds, 1989; Richard, 1994; Wilkins and Heidrick, 1995; McQuarrie and Wernicke, 2005).

Our understanding of the extensional tectonic history of Arizona has improved enormously over the past 50 years. Along with much new geologic mapping and some well-placed seismic-reflection profiles, this understanding allows reconstruction of southeastern Arizona to reveal the distribution of porphyry copper deposits at the end of Laramide orogenesis. This paper outlines such a reconstruction and identifies linear trends in the reconstructed deposits that have implications for porphyry copper metallogenesis and exploration. Porphyry copper deposits form linear belts in the reconstruction, with prominent east to northeast trends that parallel foliation and lithologic layering in Paleoproterozoic Pinal schist and related igneous and metamorphic rock units of the Yavapai-Mazatzal orogenic belt (Karlstrom et al., 1987; Whitmeyer and Karlstrom, 2007). Parallelism with this fabric is interpreted to indicate the existence of E- to NE-trending zones of high copper and molybdenum fertility in underlying Paleoproterozoic deep crust and mantle lithosphere (e.g., Sillitoe, 2012; Sillitoe et al., 2014).

Crustal genesis in the Sonoran Desert region resulted from ~1.8 to 1.6 Ga subduction-related magmatism, sedimentation, and tectonic accretion. Belts of various Paleoproterozoic rock types and shear zones trend dominantly northeast to east-northeast, as does the strike of steeply dipping tectonic foliation in most metamorphic rock units. This tectonic grain reflects the trend of the Laurentian continental margin and the shortening strains resulting from plate-tectonic convergence and accretion (Karlstrom et al., 1987, 2004; Whitmeyer and Karlstrom, 2007). Later geologic events—including 1.4 Ga plutonism, 1.1 Ga diabase intrusion, and Paleozoic sedimentation on a stable continental platform—were followed by renewed orogenesis in Jurassic and Laramide time (Dickinson, 1989). Laramide magmatism was widely distributed and produced many isolated intrusions and clusters of intrusions. Laramide crustal shortening that produced thrust and reverse faults must have caused some rearrangement of Laramide mineral deposits, but later fragmentation and concealment of these faults and uncertainty regarding timing, direction, and amount of displacement inhibit reconstruction of Laramide deformation.

The overall distribution of porphyry copper deposits in the Sonoran Desert porphyry copper province does not reveal an overall geometric order, but it does include linear groups of deposits. A NW-oriented linear arrangement of deposits extends from the Cananea mine in Sonora to the Silver Bell mine northwest of Tucson (Fig. 1A). This linear belt, known as the Cananea lineament, projects southeastward toward the Nacozari mining district and northwestward toward the Bagdad and Mineral Park deposits in northwestern Arizona (Fig. 1B). This entire 700-km linear belt, from Nacozari to Mineral Park, approximately parallels the Laramide continental margin.

In contrast, east-northeast linear trends are apparent in two other areas. The four deposits centered on the Chino mine in southwestern New Mexico, forming a 90-km-long, E-NE–trending belt, are within the Colorado Plateau and have not been significantly displaced by Cenozoic normal faulting (Fig. 1A). The large Resolution and Miami-Inspiration porphyry copper deposits east of Phoenix (Fig. 1A) are part of a 30-km-long, E-NE–trending belt that includes numerous copper deposits and related mineralized breccias, shear zones, and fractures (Peterson, 1962). The east-northeast orientations of normal faults, mineralized shear zones, and dikes in many of the porphyry copper deposits in these belts support the concept that elevated east-northeast oriented compressive stress originating from Laramide subduction influenced orientations of intrusions, veins, and other conduits for magma and hydrothermal fluids (Rehrig and Heidrick, 1976; Heidrick and Titley, 1982). Lack of continuity of dikes and faults between different deposits indicates that these features are not mineralized segments of laterally extensive shear zones and dike swarms but rather reflect ambient stress conditions at each site at various times of magmatism and mineralization.

The Basin and Range tectonic province of western North America is Earth’s premier example of the geologic consequences of distributed continental extension. If one includes Eocene extension in the Pacific Northwest, the area of tectonic extension encompasses an ~3,600-km belt from central British Columbia to central Mexico (e.g., Henry and Aranda-Gomez, 1992). Most of this area is arid, resulting in typically good to excellent rock exposures and geomorphology that reflects middle to late Cenozoic magmatism and tectonic extension. Many normal faults are still active (e.g., Menges and Pearthree, 1989; Pérouse and Wernicke, 2017).

The horst and graben structure of the central Basin and Range province (the Great Basin of Nevada and Utah) initially led to the interpretation that total extension was ~5 to 15% (Stewart, 1971). Low-angle normal faults and metamorphic core complexes were identified later as products of distinctly different extensional tectonic processes that accommodated much more extension (Crittenden et al., 1980; Wernicke, 1981). With thick continental crust and elevated temperatures due to magmatism, the deep crust can flow into areas of tectonic thinning, leading to inflation and upward displacement and arching of normal-fault footwalls (e.g., Wernicke, 1992; Fig. 2). With sufficient normal-fault displacement, mylonitic fabrics that formed by distributed shearing in the middle crust down-dip from normal faults were uplifted, warped, and exposed, commonly in arched, domed, or corrugated metamorphic core complexes (Spencer, 1984, 1999). Determination of displacement direction and shear sense in mylonitic rocks supports correlation of mylonitic shear zones with tilted and warped normal faults (Davis et al., 1986). Tens of kilometers of extension are possible along normal faults known as detachment faults that are commonly gently dipping because of isostatic uplift and tilting of footwall rocks above a fluid-like deep crust (Wernicke and Axen, 1988).

Where footwall rocks were denuded by displacement on detachment faults, the distance between the footwall breakaway (the footwall side of the initial normal-fault trace) and the tip of the wedge that forms the hanging-wall block represents total extension if the block itself is unextended, and minimum extension if the hanging-wall block is internally extended. Similarly, displacement of a contact, for example the basal depositional contact of mid-Cenozoic strata on Laramide or older basement, can provide a marker for the magnitude of extension (Fig. 3). Determination of total extension by identification of footwall and hanging-wall contact cutoffs is the primary basis for the tectonic reconstructions outlined here. Four cross sections of extensional detachment faults are evaluated for extension magnitude and direction in this study. In each of these, igneous intrusions and porphyry copper deposits in the hanging-wall block are associated with intrusions in the footwall (Fig. 3), although there is evidence in only one of these cross sections (San Xavier) that hanging-wall and footwall igneous rocks are comagmatic. Extension magnitude and direction determined for each of these four cross sections are then applied to reconstruction of nearby porphyry copper deposits where the deposits can be associated with footwall or hanging-wall blocks from one or more of the four cross sections. All displacements are ultimately used for reconstruction relative to the Colorado Plateau, which is treated as a stable reference. Porphyry copper deposits far from the Colorado Plateau require reconstruction by the sum of two, three, or four displacement vectors.

Exposures of mylonitic rocks in metamorphic core complexes in the Sonoran Desert region include two clusters, one centered on the Santa Catalina Mountains north of Tucson, Arizona, and one in northern Sonora, Mexico (Fig. 4). The greater Catalina core complex is discontinuously exposed, largely because of fragmentation by younger high-angle normal faults. The core-complex phase of extension, during which mylonitic rocks were penetratively sheared, uplifted, and exhumed, occurred at ~30 to 18 Ma and was associated with widespread felsic magmatism (Spencer et al., 2022). Extension direction was dominantly east-northeast–west-southwest in the greater Tucson area, with a notable exception of north-south extension on the San Xavier fault southwest of Tucson (Fig. 4). At ~15 Ma, extension direction changed to approximately east-west, with associated basaltic rather than felsic magmatism (Spencer and Reynolds, 1989). High-angle normal faulting accommodated much less extension than the earlier phase of detachment-fault displacement and core-complex exhumation but produced steep range fronts and deep basins (Fig. 4). Some of these high-angle normal faults are still active (Suter, 2015) or have late Quaternary displacement (Pearthree and Calvo, 1987; Menges and Pearthree, 1989).

Determining the amount of tectonic extension in the Sonoran Desert region is difficult to impossible in most areas because bedrock is largely concealed beneath upper Cenozoic cover. However, exposures of detachment faults and mylonitic fabrics in metamorphic core complexes provide constraints for tectonic reconstruction, as do several seismic-reflection profiles that reveal subsurface structure. The following is an evaluation of various lithologic, structural, and seismic-reflection features along four transects that reveal approximate extensional displacements. Application of these identified displacements to nearby porphyry copper deposits allows a province-wide reconstruction. Two of these transects cross faults that meet in the subsurface beneath Tucson basin and result in complexities that require two alternative reconstructions to evaluate possibilities and uncertainties. Uncertainties in reconstructions also result from the tectonic behavior of extending crust whereby each upper crustal block floats on the plastically flowing deep crust and follows a different path from adjacent crustal fragments (i.e., “raft tectonics” of Kruger and Johnson, 1994).

The Pinaleño Mountains in southeastern Arizona consist largely of Proterozoic igneous and metamorphic rocks, with a mylonite zone at the northeast foot of the range (Fig. 5A; Thorman, 1981; Thorman and Naruk, 1987; Scoggin et al., 2021). The mylonitic fabric, characterized by a well-developed 040°-trending lineation and top-northeast, normal-sense shear (Naruk, 1986; Bailey and Eyster, 2003), overprints a 56 Ma composite granite and granodiorite pluton (Long et al., 1995; Scoggin et al., 2021). ⁴⁰Ar/³⁹Ar dates from biotite and muscovite, and fission-track dates from apatite, all from this pluton and nearby Proterozoic gneiss, indicate cooling at ~33 to 18 Ma, consistent with cooling broadly related to tectonic exhumation during displacement on a detachment fault and its downdip continuation as a mylonitic shear zone (Long et al., 1995; Chapman et al., 2024). Younger normal faulting and tectonic exhumation are inferred from (U-Th)/He dates from apatite and zircon and are attributed to displacement on a normal fault at the southeastern foot of the range (Chapman et al., 2024).

At the northwest end of the range, the scoop-shaped Eagle Pass detachment fault places mid-Cenozoic volcanic and clastic sedimentary units over Proterozoic crystalline rocks (Fig. 5A; Blacet and Miller, 1978; Bergquist, 1979; Davis and Hardy, 1981). The fault dips to the southwest and the overlying units are tilted moderately to steeply to the southwest. This geometry is consistent with the interpretation that displacement was top-northeast above an originally NE-dipping normal fault that is now tilted to the southwest due to uplift, arching, and southwestward tilting of the underlying bedrock (Spencer, 1984). Farther north, the northeast-plunging, scoop-shaped Black Rock detachment fault also displaces southwest-tilted, mid-Cenozoic volcanic and clastic sedimentary units (Blacet and Miller, 1978; Simons, 1987; Toro et al., 1990). These features are all consistent with the interpretation that the Pinaleño Mountains are a metamorphic core complex as originally proposed by Davis (1980) and Rehrig and Reynolds (1980).

A seismic-reflection profile across Eagle Pass and Safford basin revealed abundant deep crustal reflections that are arched upward beneath the northeastern Pinaleño Mountains and Safford basin in a manner consistent with isostatic uplift and arching of a detachment-fault footwall during genesis of a metamorphic core complex (Figs. 2, 3, 5B; Kruger and Johnson, 1994; Kruger et al., 1995). The subsurface depositional contact at the base of the mid-Cenozoic volcanic and clastic strata beneath Safford basin is truncated westward by the detachment fault at km 11 to 12 in seismic line 13 of Kruger and Johnson (1994; Fig. 5B). This hanging-wall cutoff of the basal depositional contact was interpreted as displaced from an equivalent footwall cutoff southwest of the Eagle Pass detachment and adjacent to the Galiuro range front where it must be located beneath the northern extension of Wilcox basin (Fig. 5A) between km 40 and 47 on the seismic line. Detachment-fault displacement of 28 to 36 km (32 ± 4 km) is indicated (Kruger and Johnson, 1994), although a small fraction of this displacement likely occurred instead on the concealed southeastward extension of the Hawk Canyon normal fault (Fig. 5A). Regardless of where exactly the extension occurred, the calculated magnitude of extension is the same.

The Dos Pobres, Lone Star (Safford), and Sanchez porphyry copper deposits in the Lone Star mining district, located in the Gila Mountains on the northeast side of Safford basin (Fig. 5A), are associated with a compositionally diverse suite of Laramide granitic and volcanic rocks (Robinson and Cook, 1966; Langton and Williams, 1982). Restorative displacement of the Pinaleño Mountains by 32 km to a position beneath the Gila Mountains—parallel to 040° mylonitic lineation in the Pinaleño Mountains or along a 060° path perpendicular to the Galiuro Range front—places the porphyry copper deposits above the 56 Ma composite granitoid pluton at the foot of the Pinaleño Mountains. This suggests that the Pinaleño pluton was originally beneath the Lone Star district deposits and that the pluton is part of an intrusive suite that was related to porphyry copper mineralization, as shown in Figure 3.

Regional displacements and rotations: The tectonic extension responsible for formation of Safford basin continues northwestward with decreasing basin width and inferred extension magnitude until terminating somewhere near the north end of Tonto basin in central Arizona (Fig. 6A). In a first-order approximation, the Galiuro Mountains and their northward continuation through the Pinal, Superstition, and Mazatzal Mountains, were all displaced southwestward relative to the Colorado Plateau by movement about a pivot point near the northern Tonto basin, with ~8.5° of clockwise rotation associated with extension. This is not an exact rotation determination because the direction of displacement of each of these ranges probably deviates slightly from adjacent ranges, and the ranges themselves have undergone slight to moderate internal extensional faulting.

Relative to the Colorado Plateau, tectonic extension in the Safford basin-Pinaleño Mountains-Wilcox basin area resulted in displacement of all the porphyry copper deposits located to the west and south. These displacements are divided into three domains in the interpretation shown in Figure 6B. In domain 1, porphyry copper deposits in the Miami-Resolution-Ray-San Manuel area near the Pinal Mountains are displaced by 8.5° of clockwise rotation about a pivot point near the edge of the Colorado Plateau. The reconstruction vectors shown as blue lines in Figure 6B in domain 1 are arcs, with 8.5° of curvature and with increasing restorative displacements in more southeasterly locations. Maximum restorative displacement is 50 km at the straight red line separating domains 1 and 2 from domain 3. The blue lines in domain 2 represent linear restorative displacements with increasing displacements in more southeasterly locations that match the total displacement in domain 1 but without curvature. In domain 3 all restoration vectors represent uniform 50 km displacement, also without rotation. Without this approximation involving three separate domains, extensional displacement in domain 2 would curve around to the northwest until eventually indicating convergence with the northwestern Colorado Plateau (northwestern Arizona is highly extended). Also, without this approximation, displacement in domain 3 would increase to the south, approaching 80 km near Nacozari where there is no evidence of such large extensional displacement. Discontinuities between these domains are consistent with the concept of raft tectonics in which each range is an individual block floating on deep viscous crust and moving somewhat independently from adjacent blocks (Kruger and Johnson, 1994).

Reconstruction of Cenozoic tectonic extension north and northwest of Tucson is feasible for a transect from the Silver Bell Mountains to the Galiuro Mountains (Figs. 7, 8). The Falcon Valley area along this transect includes numerous exposures of two generations of moderately to gently dipping normal faults. The gently dipping to subhorizontal, first-generation faults, including the Guild Wash, Suizo, Star Flat, and Cloudburst faults (Fig. 7), have all been correlated with the Catalina-San Pedro detachment-fault system to the south that accommodated tectonic exhumation and uplift of the greater Catalina core complex (Dickinson, 1991; Spencer et al., 2022). In the Falcon Valley area these gently dipping faults displace similar Oligocene to lower Miocene volcanic and clastic strata that appear to be tectonically dispersed fragments of units that were initially part of the same extensional sedimentary basin (Dickinson, 1991). Mylonitic fabrics along the southwest flank of the greater Catalina core complex, interpreted as part of a mylonitic shear zone that was the downdip continuation of the initial Catalina-San Pedro detachment-fault system, include abundant mylonitic lineations that indicate upper plate displacement toward ~244° (Fig. 8; Spencer et al., 2022). This is the inferred displacement direction for the low-angle normal faults in the Falcon Valley area that are shown with blue lines in Figure 7. A younger generation of faults, shown with green lines in Figure 7, offsets the older generation but accommodated much less extension (e.g., Lowell, 1968; Dickinson, 1991; Favorito and Seedorff, 2021).

As shown in cross section A-Aʹ in Figure 7, the subhorizontal Catalina-San Pedro detachment fault places middle Cenozoic conglomerate and volcanic rocks over crystalline rocks along a 38-km transect parallel to extension direction, with almost no pre-Cenozoic basement among the hanging-wall rocks over this distance. Restoring the hanging-wall cutoff of the basal mid-Cenozoic depositional contact in the northern Tortolita Mountains would place this feature at least 38 km back to the northeast so that it projects downdip into the equivalent footwall cutoff under San Pedro River Valley (red lines in Fig. 7; Table 1; Spencer et al., 2022). As shown in the cross section, the inferred 38-km minimum displacement includes displacement on younger normal faults that cut the detachment fault.

Upper Cenozoic cover in the San Pedro River Valley conceals the footwall cutoff for the detachment-fault system and also conceals younger normal faults. More than ~9 km of extension on the detachment fault beneath the valley, with a footwall cutoff near the Galiuro range front, is problematic because Paleozoic strata and Laramide granitoids at the foot of the Galiuro Mountains would likely be present below the hanging-wall cutoff north of the Guild Wash fault. The rocks below the hanging-wall cutoff, however, consist only of Proterozoic crystalline rocks. Less than ~3 km of extension beneath the valley due to combined displacement from both fault generations seems problematic because extensional faulting beneath the valley is indicated by normal faults on both valley flanks that project beneath the valley, and by an exposure of the top of a tilted fault block just north of the cross-section line (Fig. 7; Krieger, 1968a; Dickinson, 1991). On the basis of these considerations, I estimate 6 ± 3 km of extension beneath the San Pedro River Valley (Table 1).

Cross section A-Aʹ crosses four normal faults within the upper-plate of the detachment fault over ~12 km of exposed bedrock in the northern Tortolita Mountains (Fig. 7). Complex structural geology with moderate to steep bedding dips to the east in eastern areas and to the west in western areas indicates significant structural disruption, whereas abundant crystalline rocks without geometrically consistent lithologic markers prevent accurate structural reconstruction (Spencer et al., 2002; Ferguson et al., 2003). If each of these four faults accommodated 0.5 to 2.0 km of extension, which seems reasonable considering the high degree of structural disruption, then total extension of 5 ± 3 km is indicated within this 12-km-long transect segment (Table 1).

The cross section of Figure 7 projects southwestward for ~30 km across Avra Valley to the edge of the Silver Bell Mountains (Fig. 8). Avra Valley is part of a long north-trending basin produced by post-detachment extension that cuts across the mylonitic southwest flank of the greater Catalina core complex. Symmetry about the valley axis in bedrock-depth contours suggests that Avra Valley is a graben. In the simple case of 45°-dipping normal faults on both graben flanks, the sum of vertical displacement on the two graben flanks is equal to horizontal extension. If bedrock beneath the valley floor is 1.5-km deep (Richard et al., 2007), and the modern valley floor is 0.5 km below the paleosurface of now eroded bedrock that formed the initial graben flanks (flanking ranges are rarely higher than 500 m), then ~4 km of extension is indicated (4 ± 2 km in Table 1). This does not include extension over the ~6 km of alluvial cover along the transect in the area between the Tortolita Mountains and Desert Peak (Fig. 8). Footwall rocks could have been completely denuded or hanging-wall rocks completely unextended, but extension and structural disruption of adjacent rock units in the northern Tortolita Mountains suggests at least some extension. A nearly all-encompassing approximation is that 3 ± 2 km of extension occurred over this 6-km-long section of the transect (Table 1). These estimates lead to calculated total extension of 56 ± 10 km between the Galiuro and Silver Bell Mountains. Restoration of this extension places the Silver Bell porphyry copper deposit in a position in western Falcon Valley near and above the Laramide pluton of Chirreon Wash in the Tortolita Mountains (Fig. 8; Banks, 1980; Ferguson et al., 2003).

Cenozoic tectonic extension in the Tucson area was unusually complex. Tucson basin is surrounded by five mountain ranges, each of which was displaced away from the others in a different direction, again illustrating the concept of raft tectonics (Kruger and Johnson, 1994). The Rincon Mountains are only slightly displaced relative to the Galiuro Mountains and are used as a stable reference for extensional displacements south and southwest of the Tucson area (Spencer et al., 2022). The Galiuro Mountains in turn are used as a reference because analysis of extension in the Pinaleño Mountains area allows tectonic restoration of the Galiuro Mountains relative to the Colorado Plateau. Tectonic restoration of porphyry copper deposits in the Sierrita Mountains relative to the Rincon Mountains requires evaluation of displacements on three different fault systems that meet beneath Tucson basin, as outlined below.

Catalina-San Pedro detachment fault: Total displacement on the Catalina-San Pedro detachment fault at the southwestern foot of the Rincon Mountains was estimated at 36 ± 2 km based on correlation of an E-dipping thrust fault in the footwall block of the detachment fault with a thrust fault in the hanging-wall block (Spencer et al., 2019). An additional 1 to 3 km of extension was estimated on moderately dipping normal faults that cut the detachment-fault footwall block in the San Pedro River Valley, leading to estimated tectonic extension of ~38 km. The 40-km distance between correlated features was interpreted as greater than actual extensional displacement, however, because of 1 to 3 km of erosional retreat (eastward displacement) of the thrust contact on the east flank of the San Pedro River Valley. Spencer et al. (2019), however, did not consider that erosional retreat of the correlative thrust in the hanging-wall block would have been in the same direction and conceivably equal or greater in magnitude as erosional retreat in the footwall block. As a result of this consideration, 40 ± 2 km would be a more appropriate estimate of extension in this transect (restoration vector set 3 on Fig. 8).

Santa Rita fault: Tucson basin is bounded to the south by the Santa Rita fault at the northwest foot of the Santa Rita Mountains (Fig. 8). This fault, which has minor displacement across Quaternary alluvial fans and appears as a high-angle normal fault in an excavation (Pearthree and Calvo, 1987), has been imaged on seismic-reflection profiles with a westward apparent dip of ~20° but has a slightly steeper true dip to the northwest (Johnson and Loy, 1992; Wagner and Johnson, 2006). Seismic reflections interpreted as bedding in Cenozoic strata dip to the east, consistent with half-graben structure for the basin and with the interpretation that the basin axis marks the trailing, tapered edge of the crystalline rocks in the hanging-wall block (Wagner and Johnson, 2006, 2010).

The relationship between the Santa Rita normal fault and the Catalina-San Pedro detachment fault is not well understood and contributes most of the uncertainty in estimates of total displacement of copper deposits in the Sierrita Mountains relative to the Rincon and Galiuro Mountains. An additional ~20 km of extension was estimated for the Catalina-San Pedro detachment fault in the subsurface beneath Tucson basin, with the trailing edge of crystalline rocks in the hanging-wall block forming the basin axis (Spencer et al., 2019). The Santa Rita fault intersects the Catalina-San Pedro detachment fault in the subsurface in this general area, and the detachment fault appears to have accommodated displacement on the Santa Rita fault north of this intersection (Wagner and Johnson, 2006). In the interpretation of Spencer et al. (2019), the Santa Rita fault is an upper plate fault above the detachment fault, with displacement on the Santa Rita fault to the west-southwest as with the detachment fault. In this interpretation, normal displacement on the Santa Rita fault was oblique (west-southwest) to the northwesterly dip of the fault.

In an alternative fault model, the south-to-north increase in Tucson basin width and depth adjacent to the Santa Rita fault is due to counterclockwise rotation of the Sierrita Mountains block during normal-fault displacement, with greater fault displacement in more northerly areas. Restoration of ~20° of rotation about a pivot point in the southern Santa Rita Mountains places the trailing edge of the Sierrita fault block, which is coincident with the basin axis, near the crest of the northern Santa Rita Mountains (restoration vector set 1 on Fig. 8). As the northern Santa Rita Mountains are topographically asymmetric, with a gentle eastern slope and a steep western slope, restoration of the tapered edge of the Sierrita fault block to a position near the Santa Rita range crest would probably yield a gentle predisplacement landscape. In this rotational interpretation, the Tucson Mountains, directly north of the Sierrita Mountains, rotated clockwise so that maximum extension occurred at the latitude of the deepest point of Tucson basin due to saloon-door type rotation of the two fault blocks (Sierrita and Tucson) to the west, while minimum extension occurred in Avra and Altar valleys at about the same latitude.

In the alternative rotational reconstruction outlined here, early displacement on the Santa Rita fault was approximately perpendicular to the trace of the fault, toward the northwest rather than to the west-southwest, with a rotational component of the Sierrita Mountains as described above. Most likely, extension occurred by both types of movement, but their relative significance for displacement of the Sierrita Mountains is uncertain. Both are considered here to outline the range of possibilities for total displacement of the Sierrita Mountains.

San Xavier fault: The gently dipping San Xavier fault on the northeastern flank of the Sierrita Mountains displaced the Pima-Mission porphyry copper deposit ~12 km northward from an initial position above the Twin Buttes porphyry copper deposit (Cooper, 1960a; Jansen, 1982; Richard et al., 2003). The smaller San Xavier North copper deposit, located ~4 km north of the Pima-Mission deposit, has been interpreted as displaced northward from above the Pima-Mission deposit by a north-dipping normal fault (Jansen, 1982; King, 1982), with total indicated displacement of ~16 km on the San Xavier fault system (Spencer et al., 2022). Restoration of displacement on the San Xavier fault relative to the San Xavier North copper deposit and, farther north, the Tucson Mountains (i.e., relative to the hanging-wall block rather than the footwall block), thus requires ~12 km of northward restorative displacement of the Twin Buttes deposit, placing it beneath the Pima-Mission deposit, and then another ~4 km of northward displacement, placing the combined Twin Buttes–Pima-Mission deposit beneath the San Xavier deposit and forming the greater Pima deposit at point P in Figure 8. Each of these two restorations allow an uncertainty of 1 km in either direction, yielding a final estimate of 16 ± 2 km. Alluvial cover north of the San Xavier mine could conceal additional normal faults, but the Del Bac Hills on the north side of this alluvial cover consist largely of subhorizontal, ~26 Ma volcanic rocks (Percious, 1968) and provide no evidence of additional normal faulting.

The Ajo Road fault, located ~30 km west of the northeastern Sierrita Mountains (Fig. 8), is a north-dipping normal fault with a mylonitic footwall with north-plunging lineations and top-north shear-sense indicators (Gardulski, 1980; Davis et al., 1987; Gottardi et al., 2020). It has the same movement sense and approximate age as the San Xavier fault with which it is correlated. Farther west, W-NW–striking normal faults in the Comobabi Mountains indicate approximately north-south extension (Haxel et al., 1978). The location of the San Xavier-Ajo Road-Comobabi normal-fault system is unknown east of the Sierrita Mountains but must be north of the Empire and Santa Rita Mountains, which are unaffected by north-south extension. With this geometry, the Rosemont porphyry copper deposit would be restored northward by the same amount (16 km) as the Twin Buttes and Sierrita deposits (restoration vector set 2 on Fig. 8).

Tectonic reconstruction of the Tucson area: The red lines in Figure 9 represent the reconstruction pathways outlined by Spencer et al. (2019, 2022). The Twin Buttes, Sierrita, and Rosemont deposits were first restored 16 km toward 345°, with reassembly of the Twin Buttes–Pima-Mission–San Xavier North deposits to form the greater Pima deposit. The Sierrita deposit and the reassembled Pima deposit were then restored 60 km to the east-northeast following a slightly curved path, while Rosemont was restored by 40 km because it was not subjected to displacement by extension beneath Tucson basin. In the alternative reconstruction shown in Figure 8 and represented by the purple lines in Figure 9, the first restoration step involves a 20° clockwise rotation of the Sierrita Mountains about a pivot point to the south (Fig. 8) so that Tucson basin is closed. This restoration also changes the direction of the displacement on the San Xavier fault from 345° to 005°. After restoration of 16-km displacement on the San Xavier fault system, the Sierrita, reassembled Pima, and Rosemont deposits are then restored 40 km along the slightly curved trajectory identified by Spencer et al. (2019). The eastern ends of the red and blue lines in Figure 9 indicate possible positions of these deposits before extensional dismemberment and dispersal and reveal differences in restored positions depending on which displacement model is used.

Restoration of extension places these porphyry copper deposits ~25 to 40 km west of the Johnson Camp porphyry copper deposit, but the spatial relationship with Johnson Camp is complicated by separation of the deposits by two Laramide thrust faults, the Little Rincon thrust (Gehrels and Smith, 1991) and the Wildhorse Mountain thrust system (Fig. 9). Top-northeast mylonitic fabrics below the Wildhorse Mountain thrust, and the ~20-km extent of Proterozoic granitic rocks thrust over Phanerozoic strata along a path parallel to 067° mylonitic lineations, indicate minimum Laramide thrust displacement of the hanging-wall block ~20 km toward 067° (Spencer et al., 2019, 2022). The six dark green lines in Figure 9 represent 20-km restorations of the Sierrita-Pima-Rosemont deposits to prethrust locations relative to Johnson Camp. Thrust faulting could have been entirely premineralization, however, in which case the positions at the northeast ends of the dark green lines, rather than the southwest ends, would represent initial positions of the Sierrita-Pima-Rosemont deposits relative to Johnson Camp. As a result of uncertainty regarding relative timing of thrusting and mineralization, the green lines themselves represent a range of possible reconstructed positions and include intermediate positions where mineralization occurred during thrust faulting. Those six green lines and Johnson Camp are surrounded by a dark red ellipse in Figure 9 that includes all the uncertainties discussed above except for the possibility of >20-km displacement on the Wildhorse Mountain thrust and the uncertain significance of the Little Rincon thrust in the reconstruction.

The reconstruction vectors applied to the map location of each porphyry copper deposit are summed to produce the composite vectors shown in Figure 10. Reconstruction vectors are applied as follows:

1. Safford: Reconstruction vectors from extensional displacements in the Pinaleño Mountains area are shown in Figure 6B.

2. Falcon Valley: The 56-km reconstruction vector for the Silver Bell porphyry copper deposit (Fig. 8) is applied to deposits to the north and west, including Santa Cruz, Sacaton, and Ajo (Fig. 1). The Florence (Poston Butte) deposit is isolated by upper Cenozoic basin fill and is within the core-complex belt. The restoration vector applied to Florence is half the length for Falcon Valley displacement, which is an approximation that places the Florence deposit between deposits to the southwest and northeast, as expected without tectonic leapfrogging.

3. Rincon: Reconstruction of the 40-km displacement of bedrock at the southwest foot of the Rincon Mountains is applied to copper deposits in a belt extending from the Tucson area southward to Cananea.

4. San Xavier: All deposits south of the latitude of the Sierrita Mountains, including Rosemont, are restored northward by the inferred 16-km displacement on the San Xavier fault. Within the Sierrita Mountains, restoration reassembles the Twin Buttes, Pima-Mission, and San Xavier North deposits to form the greater Pima deposit, with total contained copper of ~10.4 Mt.

5. Bisbee and Nacozari are distant from areas with identified displacement vectors. The Rincon core-complex belt projects southeastward from the Rincon Mountains toward the Bisbee porphyry copper deposit in the Mule Mountains, which is a NW-trending range within an area of typical Basin and Range topography. The distribution of extension in this area is poorly understood, however, because range-bounding normal faults, if present, are generally concealed by upper Cenozoic strata. Basin depths, basin geometry, and range topography all suggest extension on both sides of the Mule Mountains (Fig. 4). Of the total 40 km of extension projected from the Rincon Mountains toward the Mule Mountains, it seems likely that at least 10 km of extension occurred on each side of the Mule Mountains. Displacement of Bisbee related to Rincon Mountains extension is reasonably inferred as 20 ± 10 km. Displacement related to the Rincon core complex belt is projected southward to Nacozari but with greater uncertainty (20 ± 15 km). The green vectors in Figure 10 for each deposit are 20 km long.

6. Basin and Range normal faults, active during late Cenozoic, post-detachment extension, cut the detachment faults along the Pinaleño, Falcon Valley, and Rincon transects. These normal faults contributed to extension, but their minor contributions are not separated from detachment-fault extension. Displacement of the Silver Bell deposit described above for the Falcon Valley transect is inferred to have displaced four porphyry copper deposits located to the northwest (Sacaton, Santa Cruz, Vekol, and Lakeshore; Fig. 1). These four deposits were also displaced by the Basin and Range extension that produced Picacho basin (Fig. 4), which is deeper than Avra basin and accommodated additional extensional displacement of deposits to the west (>4 ± 2 km). Minor additional extension is associated with basins closer to each deposit (<4 ± 2 km). Basin and Range displacement of 8 ± 4 km, in addition to the extension in the Falcon Valley transect, is inferred for these four deposits (Fig. 10).

7. The Ajo porphyry copper district and the New Cornelia mine (Cox et al., 2006), southwest of the four deposits mentioned above, were displaced by the same 8 ± 4-km Basin and Range extension that affected these four deposits. Additional Basin and Range extension occurred over the intervening 80-km distance between Ajo and the closest of these four deposits (Vekol), as indicated by NW-striking faults that cut mid-Cenozoic volcanic units and are associated with minor to moderate tilting (Gray et al., 1985). The amount of extension over this distance is uncertain, however. Extension of 5 to 20% in the 80-km wide area is equivalent to 9 ± 5-km displacement at Ajo. This uncertain value is possibly an underestimate, but it is used for the reconstruction in Figure 10. Also, N-S extension at Ajo (Cox et al., 2006) is not reconstructed because the Ajo deposit is within the hanging wall of the local, N-dipping detachment-fault system, as is the case at the San Xavier North mine. (These two deposits are not displaced by associated detachment faults relative to the Colorado Plateau.) Finally, Ajo is at the margin of a broad zone of late Cenozoic distributed right-lateral displacement associated with the Pacific-North American plate boundary (Richard, 1993, Faulds and Henry, 2008). Strike-slip displacements associated with this diffuse boundary likely contributed to the linear morphology and northwest trends of mountain ranges near Ajo, but the amount of hypothetical northwestward displacement of the Ajo deposit is unknown and could be zero.

Restoration of porphyry copper deposits following the composite reconstruction vectors of Figure 10 leads to the reconstruction shown in Figure 11, with ellipses included that outline linear arrangements in the pre-extension deposit distribution and surround all 39 plotted deposits. Extensional dispersal of the copper deposits can be readily visualized with the simplifying assumption of extension restricted to two linear rifts (Fig. 12). In this approximation, extension in the Rincon and Pinaleño rifts displaced the reconstructed deposits to their present locations. A third rift encompasses the Sonoran core complexes but does not displace any of the porphyry copper deposits. Core-complex rifting was not like typical continental rifting with deep subsidence and dominantly basaltic magmatism, but rather accommodated extension by lateral inflow of felsic deep crust and core-complex inflation that was so effective that rift troughs did not form (Figs. 2, 3).

Reconstruction vectors were derived from evaluation of displacements in four cross sections and then applied to groups of deposits that were treated as embedded in blocks that were not internally deformed during extension (Fig. 13). The uncertainties in the reconstruction vectors shown in Figure 13 are small compared to the vector displacements. An additional source of uncertainty is related to the large dimensions of some of the displaced deposit groups within the blocks. The Silver Bell and Rosemont blocks extend over many tens of kilometers laterally away from the tail ends of the displacement vectors plotted in Figure 13. These blocks could have rotated about vertical axes during extensional dispersal without detection in the structural analysis presented here. Because of the large size of these two blocks, minor rotation would cause significant error in the calculated displacements of copper deposits distant from the tail end of the restoration vectors. Rotational and extensional displacements of individual ranges within each block are indicated by numerous mapped faults, but these faults are generally steeply dipping with small displacements and were not used to calculate internal distortions of the blocks. In summary, numerous minor faults and rotational displacements are either known or likely but were not incorporated into this analysis because of their small displacements and poor understanding of their geometry and displacements. Rotations of reconstruction blocks could introduce significant error but are not indicated by available data.

The long Resolution-Nacozari linear belt of porphyry copper deposits in the reconstruction shown in Figure 11 is approximately parallel to the Laramide continental margin. Its orientation is most obviously attributed to Laramide subduction tectonics in which the deposits formed above descending oceanic lithosphere where conditions for porphyry copper genesis included near optimum subduction geometry. In contrast, the four other linear belts are at high angles to the Laramide continental margin, including the Chino belt in southwestern New Mexico that was unaffected by extensional tectonic rearrangement. The features and processes that were responsible for the orientation of these four belts are not well understood.

East-northeast orientations are typical for the numerous mineralized fractures, faults, dikes, veins, and elongate intrusions in many porphyry copper deposits of the Sonoran Desert porphyry copper province. The orientation of these mineralized features has been attributed to Laramide tectonic stress conditions, with maximum compressive stress oriented parallel to plate convergence direction and minimum compressive stress perpendicular to the sheet-like form of steeply dipping dikes, veins, mineralized fractures, and normal faults (Rehrig and Heidrick, 1976; Heidrick and Titley, 1982). These mineralized features are especially apparent along the ~70-km-long (reconstructed) Santa Cruz-Miami belt (Fig. 11) and include the following: (1) “Nearly all the mineral deposits of the Globe-Miami district that are of hypogene origin show the same general trend as the mineral belt,” (Peterson, 1962, p. 143). (2) Within the northeastern end of the belt, displacement on numerous NE-striking, minor normal faults “resulted in dropping downward-pointing wedges of geologically higher rocks between upward-pointing wedges of lower rocks,” (Ransome, 1903, p. 103). (3) At the Resolution-Superior porphyry copper deposit, Manske and Paul (2002, p. 205) recognized an E-NE–striking, quartz-porphyry dike swarm associated with mineralization, and observed that the “major east-northeast–striking fabric of faults and fractures… served as conduits for much of the mineralization and intrusion in the district.” (4) At the Florence (Poston Butte) deposit, underground mapping indicates a dominant east-northeast fracture orientation with preferential mineralization along this fracture set (Nason et al., 1982). (5) At the Sacaton deposit “mineralized fractures strike from N50°W to E-W and generally dip greater than 70° in either direction” (Cummings, 1982, p. 513). Similarly, generally NE-striking features also have been observed at all four deposits in the ~100-km-long Chino copper belt in southwestern New Mexico (Rose and Baltosser, 1966; Dunn, 1982; DuHamel et al., 1995; Hillesland et al., 1995).

The reconstructed ~250-km-long Ajo-Morenci belt is characterized by similar features, as follows: (1) At Morenci, “north-dipping fractures having a strike between N. 45 E. and N. 65 E. are most persistent and best mineralized.” (Moolick and Durek, 1966, p. 226). (2) At the Dos Pobres, San Juan, and Safford deposits in the Lone Star district, numerous E-NE–striking, steeply dipping shear zones over a NW-SE distance of 18 km were the principal channels for productive porphyries and ore-bearing fluids (Robinson and Cook, 1966; Langton and Williams, 1982). (3) At Silver Bell, sulfides are abundant in “systems of veinlets or seams that are usually near vertical” and strike “in the northeast quadrant” in much of the studied area (Richard and Courtright, 1966, p. 161). Such trends have also been identified in some other deposits of the Ajo-Morenci belt (Ajo, Bunker Hill [Copper Creek]) and in the Sierrita-Johnson Camp belt at Sierrita (Rehrig and Heidrick, 1972), although the measured orientations of mineralized features at Sierrita and Ajo also reflect significant tilting by younger Cenozoic extensional faulting. Finally, a dominant east-northeast fracture set characterizes many unmineralized Laramide intrusions that are located both within and outside the linear porphyry copper belts (Rehrig and Heidrick, 1972, 1976).

The generally east-northeast trends of all these features are consistent with regional stress conditions in which the least compressive stress was oriented approximately north-northwest–south-southeast, perpendicular to the Laramide dikes, veins, normal faults, and elongate intrusions. The presence of Laramide thrust faults and folds that reflect northeast-southwest shortening (e.g., Keith and Barrett, 1976; Davis, 1979; Krantz, 1989; Favorito and Seedorff, 2017, 2018, 2020; Spencer et al., 2022) further supports the interpretation that the greatest compressive stress (σ₁) was oriented approximately northeast-southwest during Laramide faulting. (Note that vertical compression [overburden] must have been greater [σ₂] than north-northwest oriented compression [σ₃], otherwise subhorizontal dikes would result.) Belts of porphyry copper mineralization conceivably formed in response to these regional stresses if diking and mineralization were continuous along the length of each belt, but this is not the case. While regional stress conditions associated with Laramide plate convergence provide a generally accepted interpretation for the orientation of many features associated with magmatism and mineralization, they do not account for the east-northeast trend of porphyry copper belts that are many tens to hundreds of kilometers long.

Another possibility, that mineralization occurred along E-NE–striking shear zones, is not supported by the geology of large expanses of Precambrian crystalline rocks within the E-NE–trending copper belts. Three examples are identified, as follows: (1) The Tyrone porphyry copper deposit at the southeast end of the Chino copper belt is within the eastern Burro Mountains, which consist largely of ~1.4 Ga granitoids with a complex and variably developed high-temperature Proterozoic foliation that is not aligned with the Chino belt (Amato et al., 2011). (2) The Lone Star mining district includes mineralized shear zones over the entire district (Langton and Williams, 1982). Restoration of extension places the Pinaleño Mountains adjacent to and beneath the Lone Star district, but there is no evidence of a steep, E-NE–striking shear zone or schistosity in the Paleoproterozoic gneiss of Pinaleño Mountains, which makes up most of the range (Thorman and Naruk, 1987). (3) Farther southwest along the Ajo-Morenci porphyry copper belt, the Silver Bell copper district is restored to a position in western Falcon Valley where it is close to, and aligned with, the San Manuel and Bunker Hill porphyry copper deposits (Figs. 7, 8, 11). Generally undeformed Proterozoic granitoids, consisting primarily of the ~1.4 Ga Oracle Granite, present no evidence of a steep, E-NE–striking shear zone. Areas underlain by these granitoids include the Black Hills north of San Manuel (Creasey, 1965; Krieger, 1968b, 1974), the northern foot of the Santa Catalina Mountains (Force, 1997; Spencer et al., 2000), and Black Mountain north of the reconstructed position of the Silver Bell district (Krieger, 1974b; Orr et al., 2004). It thus appears that there is much evidence against porphyry copper mineralization focused along E-NE–striking shear zones of regional extent in the Sonoran Desert porphyry copper province. This conclusion is consistent with a recent study by Favorito and Seedorff (2024) of the effects of Laramide reverse faulting on porphyry copper mineralization that did not identify any mineralized E-NE–striking tear faults linking Laramide reverse faults.

Another possibility for the east-northeast orientation of the porphyry copper belts is that Laramide magmatism occurred preferentially in linear zones with this orientation, and porphyry copper deposits are present in these zones simply because that is where Laramide magmatism occurred. Laramide granitoids are generally sparse directly adjacent to the deposit-encompassing, east- to northeast-trending reconstruction ellipses but are locally abundant at greater distance (>100 km). Northern Sonora contains many Laramide granitoids with similar exposed paleodepths as the copper belts, as indicated by widespread distribution of preserved Laramide volcanic rocks (Fig. 14; González-León and Moreno-Hurtado, 2021; Valencia-Moreno et al., 2017, 2021). Furthermore, the area is underlain by 1.6 to 1.7 Ga metamorphic and igneous rocks as in southern Arizona (e.g., Iriondo et al., 2004; Farmer et al., 2005). Large porphyry copper deposits are, however, absent in this area except along the Cananea lineament. Western Arizona, northwest of Phoenix, also contains a few Laramide granitoids (Fig. 14, 1) but these granitoids also are weakly affected by copper mineralization or are barren.

Two Laramide granitic intrusions in Arizona south of the western segment of the Ajo-Morenci belt consist of the granitic orthogneiss of the Alverez Mountains (Fig. 14, 2; Beikman et al., 1995) and the granodiorite of Gunsight Hills (Fig. 14, 3; Tosdal et al., 1986). Mylonitic and gneissic fabrics in both intrusions indicate that emplacement depths for at least some of the exposed granitic rocks in each area were too deep for porphyry copper mineralization (e.g., Seedorff et al., 2008). In this case, the intrusions are not a good test of whether the plutons were originally mineralized or barren.

The Chino belt projects southwestward toward the Johnson Camp porphyry copper deposit (Fig. 11). This projection crosses the Dos Cabezas Mountains in Arizona where there are four Laramide granitic stocks (Fig. 14, 4; Cooper, 1960b; Erickson, 1968; Drewes, 1986), at least two of which are associated with minor porphyry copper mineralization (Keith et al., 1983; Drewes et al., 1988). In New Mexico the Lordsburg mining district (Fig. 14, 5) is located ~20 km south of the southwestern projection of the Chino belt. Mineralization is represented primarily by polymetallic veins associated with a Paleocene biotite-hornblende granodiorite, with ~100,000 tonnes of historic copper production and significant Au and Ag (Thorman and Drewes, 1978; Richter and Lawrence, 1983; McLemore and Elston, 2000). Farther southeast, ~70 to 100 km from the Chino belt, are multiple Laramide granitic stocks in the Little Hatchet Mountains, Apache Hills, and Sierra Rica (Fig. 14, 6; Scholle, 2003). Stocks in the Little Hatchet Mountains are associated with minor polymetallic vein deposits (Lasky, 1947; Zeller, 1970; McLemore, 2000). It thus appears that mineral deposits associated with Laramide intrusions are more significant near the southwestward projection of the Chino belt and less significant at greater distance to the south. This relationship supports the concept that mineralization is associated specifically with east-northeast trending zones of elevated metallogenic fertility rather than simply an association with more abundant intrusions.

Numerous unmineralized or weakly mineralized plutons in northwestern Sonora, several plutons in western Arizona, and stocks in far southwestern New Mexico appear to be outside the area of elevated metallogenic fertility rather than representing areas that are more deeply eroded and so denuded of their porphyry copper deposits (Fig. 14). The Sonoran Desert porphyry copper province is thus bounded by geologically similar areas that lack porphyry copper deposits. Resolution of elevated metallogenic fertility at more local spatial scales, tens of kilometers rather than 100 to 200 kilometers, is problematic, however, because few barren or weakly mineralized Laramide granitic intrusions are present directly outside the margins of the E-NE–trending copper belts.

Of the four ENE-trending copper belts, three are relatively short and two of these represent deposit alignments that are obvious and long known. The reconstructed, 50-km-long, EW-trending Sierrita-Johnson Camp belt is newly identified and is similar to the other two belts in terms of number of deposits and total copper endowment. These three copper belts occupy a tiny fraction of the Sonoran Desert porphyry copper province and a small number of its granitic intrusions but contain a large fraction of its total copper endowment. The exceptional fertility of these belts likely reflects elevated metallogenic fertility of deep crustal and/or upper mantle sources with northeast to east trends rather than an abundance of Laramide intrusions. The Ajo-Morenci belt is much larger but contains some impressive characteristics. Of the 15 granitic intrusions and intrusion clusters in the belt, nine are associated with porphyry copper deposits as shown on Figure 14. Of the other six intrusions, two are mylonitized in metamorphic core complexes and represent crustal levels that were probably too deep for porphyry-copper mineralization, and one contains enough muscovite and garnet to suggest a peraluminous composition inconsistent with porphyry copper mineralization. This leads to the conclusion that nine of 12 potentially mineralized intrusions and intrusion clusters in the 300-km-long Ajo-Morenci belt are associated with porphyry copper deposits. Such a high rate of porphyry copper mineralization suggests high metallogenic fertility of the deep crust or upper mantle source region along this alignment.

The E-NE–trending copper belts in the Sonoran Desert porphyry copper province parallel the dominant lithologic and structural grain of the Laurentian paleocontinent in southwestern North America. This grain resulted from Paleoproterozoic subduction and tectonic accretion from the southeast, which occurred during 1.6 to 1.8 Ga genesis of the continental lithosphere that now forms the greater Yavapai-Mazatzal tectonic province (e.g., Karlstrom et al., 1987; Eisele and Isachsen, 2001; Whitmeyer and Karlstrom, 2007; Amato et al., 2008; Bickford et al., 2019; Holland et al., 2020). The widespread Paleoproterozoic Pinal schist, perhaps the oldest rock unit in southeastern Arizona, has been interpreted as a subduction complex (Swift and Force, 2001; Meijer, 2014, 2016). In most exposures it consists of metamorphosed siltstone and sandstone with steep, E-NE–striking bedding, cleavage, and schistosity (Table 2; Fig. 11; Keep, 1996).

This tectonic fabric parallels a metallogenic boundary defined by silver/gold ratios in ores in the American Southwest. Ore deposits of various ages that have yielded silver and gold generally have Ag/Au production ratios with contrasting values on opposite sides of the Slate Creek shear zone and its projections to the northeast and southwest (Fig. 15; Titley, 1987, 2001). Base and precious metal mines southeast of this boundary have Ag/Au production ratios that are mostly >17.5, whereas ratios in areas to the northwest and north are generally <17.5. Locations of Proterozoic tectonic province boundaries in Figure 15 are not well constrained and include broad transition zones between provinces (Karlstrom et al., 2004), but the proposed Yavapai-Mazatzal boundary approximately coincides with the metallogenic boundary for Ag/Au ratios, supporting an inference of different lithospheric compositions for the two provinces.

This interpretation of inherited metallogenic character with ENE trends is suggested by other features. The Colorado mineral belt, with dominantly Laramide and middle Cenozoic base and precious metal mineralization, also has a NE trend that has been attributed to the consequences of Proterozoic faulting and shearing along the belt (Fig. 15; Tweto and Sims, 1963; Warner, 1978). Restoration of tectonic extension in the Great Basin also suggests a pre-extension metallogenic alignment. Three porphyry copper deposits in the Great Basin were restored to approximate pre-extension locations as follows: (1) 50 km of dextral strike slip in the Walker Lane shear zone was restored, placing the Anaconda (Yerington) deposit farther southeast (Faulds and Henry, 2008). (2) Inferred approximate doubling of the width of the Great Basin during Cenozoic EW tectonic extension was reversed, although there is much uncertainty in the amount of province-wide extension (e.g., Gans, 1987; Snow and Wernicke, 2000; Bahadori et al., 2018). In this reconstruction, three Great Basin porphyry copper deposits are roughly aligned with a NE trend, like the Colorado mineral belt, the Ag/Au ratio boundary, and the four ENE-trending porphyry copper belts in Arizona (Fig. 15).

The approximate parallelism of the NE to E trends of the identified porphyry copper belts with other Paleoproterozoic lithologic, structural, and metallogenic features supports the concept that the metallogenic fertility of southwestern North America was influenced by geochemical variations in deep crustal and/or upper mantle lithosphere that were inherited from Paleoproterozoic processes of tectonic accretion and continental lithosphere genesis (e.g., Sillitoe, 2012; Sillitoe et al., 2014). This is consistent with the concept that copper and molybdenum were stored in the lithosphere from the time of Paleoproterozoic subduction and hydration under reducing conditions characteristic of Paleoproterozoic seawater, and then mobilized during Laramide subduction and influx of more oxidizing fluids characteristic of late Phanerozoic seawater (Pettke et al., 2010; Evans and Tomkins, 2011; Lee et al., 2012; Richards and Mumin, 2013). This is also consistent with the long-term porphyry copper fertility as indicated by older porphyry copper deposits at Bisbee (Jurassic) and Squaw Peak (Paleoproterozoic) (Fig. 1B; Sillitoe et al., 2014).

The areas encompassed by the reconstruction ellipses of Figure 11, and their modern dispersed fragments (Fig. 14), are obvious targets for future porphyry copper exploration. Some deposits within the rifts were uncovered by normal faulting (San Manuel, Oracle Ridge), whereas others are probably within tilted fault blocks above rift-related detachment faults (Sacaton, Florence). Also, the alignment of the Sierrita-Johnson Camp belt with the Chino belt suggests that the two are part of a single belt. Bedrock in these two areas is now separated by ~150 km of upper Cenozoic clastic basin fill and less abundant middle Cenozoic volcanic cover, with older bedrock exposed in this ~150-km gap only in the Dos Cabeza Mountains. This mountain range contains three porphyry copper mineral districts, none of which have significant production (Keith et al., 1983). This alignment suggests that the two belts are related and that the ground between them could contain concealed porphyry copper deposits.

Regardless of the origin of all five porphyry copper belts in the Sonoran Desert porphyry copper province, the projections of each of these belts into adjacent areas represent regional exploration targets. This inference should be qualified for the northeast end of the Ajo-Morenci belt. Exposure of the Morenci and Lone Star districts resulted from normal faulting and erosion through extensive mid-Cenozoic volcanic cover. Laramide and older rocks in most of the area are concealed, so it is uncertain if the Morenci and Lone Star districts are part of an alignment that resulted from a linear arrangement of mineralization and magmatism, or if they are simply the locally exhumed deposits of a larger cluster of unaligned deposits.

Displacements on low-angle normal faults and on detachment faults that uncovered mylonites in metamorphic core complexes are sufficiently well understood in southeastern Arizona that first-order geologic reconstruction to end-Laramide time (~50 Ma) is feasible. Reconstruction is facilitated by an abundance of detailed geologic maps and several seismic-reflection profiles. Reconstruction along four transects across extension belts that resemble rifts, and evaluation of generally minor extension elsewhere, leads to a reconstruction in which the 39 largest known porphyry copper deposits in the Sonoran Desert region are within five linear belts, four of which trend eastward to northeastward. The NW-trending belt is parallel to the Laramide continental margin, whereas the others are parallel to NE-striking foliation and compositional layering in accreted and deformed Paleoproterozoic rock units that make up the greater Yavapai-Mazatzal tectonic province. E-NE–trending zones of elevated copper and molybdenum fertility are inferred consequences of Paleoproterozoic subduction tectonics and associated geochemical processes.

Geologic mapping by the Arizona Geological Survey (AZGS) that formed the basis for many of the structural-geology insights outlined in this study was funded by the AZGS and the U.S. Geological Survey through the National Cooperative Geologic Mapping Program, STATEMAP awards 98HQAG2064, 00HQAG0149, 01HQAG0098, 02HQAG0016, 03HQAG0114, 06HQAG0051, 07HQAG0110, and G09AC00199. I thank the STATEMAP Mapping Advisory Committee members for their time and interest in recommending target areas for new geologic mapping, and I gratefully acknowledge the following AZGS geologists for their contributions to STATEMAP bedrock mapping in southeastern Arizona: Steve Richard, Charles Ferguson, Brad Johnson, Steve Skotnicki, Wyatt Gilbert, and Tim Orr. I also acknowledge former Arizona Geological Survey Directors Larry Fellows and Lee Allison for their consistent support of basic geologic mapping during the time of these field investigations. I thank Economic Geology Co-Editor David Cooke, Associate Editor Julie Rowland, and reviewers José Piquer and Jorge Skarmeta for many review comments that greatly improved the manuscript. I also thank Gordon Haxel for provoking an analysis of the distribution of Laramide granitoids in the Sonoran Desert porphyry copper province, and Mike Doe and Cathy Busby for comments on an early draft. Figures were drafted in Adobe Illustrator CS4. Background shaded relief images were derived from GeoMapApp.

Jon E. Spencer, originally from the San Francisco Peninsula, received a B.S. in Geology from U.C. Santa Cruz in 1977 and a Ph.D. in Geology from M.I.T. in 1981. After a one-year postdoctoral research fellowship at the U.S. Geological Survey, he worked for 33 years at the Arizona Geological Survey in Tucson, first as a Research Geologist and then as Senior Geologist. He was responsible for the bedrock component of the STATEMAP mapping program for 23 years. After retiring in 2015 he has continued researching and publishing articles on extensional tectonics and the geology of southwestern North America.