The Sonobari Complex in northwestern Mexico preserves evidence of the consolidation of Pangaea and the Cordilleran orogenic cycle. Six Mesozoic magmatic pulses extending from the Early Triassic to the Paleocene are recognized in this complex. The volumetrically predominant rocks are calc-alkaline metaluminous and peraluminous granitoids. Mafic rocks are mainly tholeiitic gabbros. All studied rocks show high concentrations of large ion lithophile elements (LILE) and negative Nb, Ta, and Ti anomalies. Initial Nd and Sr isotopic ratios in granitoids (εNd(i) = –5.7 to –0.5; 87Sr/86Sr(i) = 0.70630–0.71302) point to evolved continental sources, while radiogenic Pb isotopes (206Pb/204Pb(i) = 16.292–19.17; 207Pb/204Pb(i) = 15.503–15.666; 208Pb/204Pb(i) = 35.257–38.984) indicate a heterogeneous basement. Initial Nd and Sr isotope ratios in mafic rocks (εNd(i) = –1.9 to +5.0; 87Sr/86Sr(i) = 0.70384–0.70626) point to mantle sources with crustal assimilation, which is also supported by the radiogenic Pb values (206Pb/204Pb(i) = 18.412–19.081; 207Pb/204Pb(i) = 15.595–15.672; 208Pb/204Pb(i) = 38.147–38.988). Geochemical and isotopic signatures suggest that magmatic rocks in the complex originated from fractional crystallization with assimilation of a heterogeneous basement isotopically similar to the Grenville orogen of Mexico. Whole-rock compositions are compatible with volcanic-arc followed by back-arc tectonic settings, where subduction and extensional processes occurred. Therefore, the main granitic pulses in the Sonobari Complex originated after the late Paleozoic Gondwana-Laurentia collision, by subduction of oceanic plates or microplates along the western border of Pangea. Also, two extension-related magmatic pulses occurred after the Late Jurassic and in the Cenomanian, separated by a collisional orogenic event that is recorded by regional metamorphism. The continental arc setting of the Sonobari Complex differs from the oceanic arc context of the Guerrero-Alisitos superterrane, indicating that there is no genetic relation between these blocks as previously proposed; rather, a relationship with the eastern Peninsular Ranges batholith is proposed.


The Sonobari Complex of northwestern Mexico is an enigmatic igneous-metamorphic block consisting of Ordovician to Late Cretaceous rocks, and whose petrogenesis and magmatic history are poorly understood. This complex is composed of metasedimentary rocks with North American and South American provenance (Vega-Granillo et al., 2008, 2013), and therefore played a role in the interaction between Laurentia and Gondwana during Pangea consolidation. The complex includes magmatic pulses of 249–241 Ma (Early Triassic), 213–203 Ma (Late Triassic), 161–150 Ma (Late Jurassic), post-162 Ma–pre-99 Ma (latest Jurassic–Early Creteaceous), 99–97 Ma (Cenomanian), 83–80 Ma (Campanian), and 64 Ma (Paleocene) ages (Keppie et al., 2006; Vega-Granillo et al., 2012, 2013; Sarmiento-Villagrana et al., 2016), which have been described in the southern North American Cordillera, particularly in the Peninsular Ranges batholith of Baja California. Some authors have considered the Sonobari Complex as part of the basement of the composite Guerrero terrane of western Mexico (e.g., Campa and Coney, 1983; Centeno-García et al., 2008). Others, pointing to significant differences, have regarded it as a separate terrane in its own right (Tahue terrane; Sedlock et al., 1993). In order to define the petrogenesis and tectonostratigraphic setting of the complex, we performed an extensive geochemical and Nd, Sr, and Pb isotopic study of magmatic and metaigneous rocks. Our data, coupled with previously published geochemical data, provide a framework for a discussion of the magmatic and tectonic evolution of the complex, as well as the nature of its basement, in order to establish its link with the magmatic provinces of the southern Cordilleran orogenic belt.


The Sonobari Complex is a ~60-km-long and ~100-km-wide crustal block cropping out across the boundary between Sonora and Sinaloa states in northwestern Mexico (Fig. 1). Recent studies demonstrated that the complex is composed of two distinctive assemblages: the Eastern Sonobari Complex of Gondwanan provenance and the Western Sonobari Complex of Laurentian affinity, which were juxtaposed in late Paleozoic time (Vega-Granillo et al., 2008, 2013).

Eastern Sonobari Complex

The Río Fuerte Formation is the oldest unit of the Eastern Sonobari Complex (ESC). This foliated metasedimentary unit contains Middle–Late Ordovician conodonts (Mullan, 1978; Poole et al., 2005, 2010). Neoproterozoic detrital zircon age groups in this formation point to a Gondwanan provenance (Vega-Granillo et al., 2008). The Río Fuerte is intruded by the Realito Gabbro, Cubampo Granite, and Capomos Granodiorite, and is thrust over the Topaco Formation. The Realito Gabbro includes a pluton of coarse-grained gabbro intruded by mafic dikes. The Cubampo Granite is composed by a larger body of two-mica granite and related sills and dikes, all characterized by large spheroidal quartz aggregates. These rocks yielded 155–151 Ma (Kimmeridgian) U-Pb zircon ages (Vega-Granillo et al., 2008). The Topaco Formation is a foliated volcano-sedimentary sequence made of metamorphosed epiclastic breccia intruded by mafic dikes, one of the later dated at 155 ± 3.5 Ma (U-Pb zircon; Vega-Granillo et al., 2012). The breccia includes clasts of volcanic, metamorphic, and granitic rocks. One granitic clast was dated at 151 ± 1 Ma (U-Pb zircon; Vega-Granillo et al., 2011), indicating derivation from the Cubampo Granite previous to the tectonic juxtaposition of this unit against the Río Fuerte Formation. A heterogeneous mylonitic foliation and greenschist facies metamorphism overprint both the Topaco Formation and the Cubampo Granite. The Capomos pluton is a homogeneous coarse-grained hornblende biotite granodiorite (de Cserna et al., 1962). This rock was originally dated at 57.2 Ma (K-Ar; Damon et al., 1983), although a 97 Ma age is considered in this work based on U-Pb geochronology (sample SFO-14; Supplemental Data1, Table S1). The Guamuchil Formation is a volcanic unit dated at 73 Ma (Campanian) by U-Pb in zircon (Vega-Granillo et al., 2012). This unit overlies the Topaco and Río Fuerte Formations in angular unconformity, and in turn, is overlain by Los Amoles Formation.

Western Sonobari Complex

The Western Sonobari Complex (WSC) is composed of metamorphic rocks, post-tectonic gabbro, and ultramafic dikes, all transected by aplitic-pegmatitic dikes. The complex geological relations prevent mapping of discrete units at the scale of Figure 1. The metamorphic rocks, known as the Francisco Gneiss (Mullan, 1978), consist of metasedimentary rocks, orthogneisses, and amphibolites. U-Pb detrital zircon ages in quartzite yield a maximum depositional age of 509 Ma, and main Paleo- and Mesoproterozoic age groups are interpreted as derived from Laurentian sources in northwestern Mexico and the southwestern USA (Vega-Granillo et al., 2013).

The oldest orthogneisses of the WSC yield U-Pb zircon ages of 249–241 Ma (Sarmiento-Villagrana et al., 2016). Younger biotite orthogneisses, originally dated at 220 Ma by U-Pb zircon (Anderson and Schmidt, 1983), yield younger U-Pb zircon ages ranging from 213 to 203 Ma (Keppie et al., 2006; Sarmiento-Villagrana et al., 2016). Two-mica quartz-feldspar orthogneisses yielded U-Pb zircon ages ranging from 163 to 159 Ma (Sarmiento-Villagrana et al., 2016). Orthogneisses and paragneisses are intercalated with tabular bodies of foliated amphibolite, interpreted as mafic dikes intruding the granitic and sedimentary protoliths. Crosscutting relations indicate that amphibolites postdate Late Jurassic orthogneiss and predate undeformed diorite dated at 100 Ma (Sarmiento-Villagrana et al., 2016). A regional metamorphism and related foliation overprint all units described above excluding the Cenomanian diorite. Thermobarometry yields average pressure (P) and temperature (T) of 8.0 ± 0.9 kb (kilobar) and 699 ± 42 °C, respectively, indicating a medium P/T amphibolite facies metamorphism (Vega-Granillo et al., 2017). Migmatization is common in orthogneisses and metasedimentary rocks mainly as stromatic bands, patches of leucosome, and veins, all indicating low melt volume. Metamorphism was dated at 91 Ma on recrystallized zircon rims (Sarmiento-Villagrana et al., 2016; Vega-Granillo et al., 2017).

Coarse-grained gabbro, diorite, hornblendite, and pyroxenite intrude the metamorphic rocks. A diorite dike transecting an undeformed gabbro yielded a 100 Ma zircon age (Sarmiento-Villagrana et al., 2016). This age is coeval with that obtained from a foliated diorite (99 Ma [Sarmiento-Villagrana et al., 2016]). Accordingly, the deformation event producing the main foliation in the area is regarded to have culminated 100 Ma. Otherwise, the undeformed Macochin Gabbro yielded a 54 Ma Ar-Ar age in hornblende (Vega-Granillo et al., 2013), which is considered as a cooling age; therefore, its crystallization age remains undefined. Non-foliated leucocratic aplitic and pegmatitic dikes transecting the tectonic foliation were dated at 82.0 to 80.6 Ma (U-Pb zircon; Sarmiento-Villagrana et al., 2016). Finally, the Los Parajes Granodiorite that intrudes the Sonobari Complex was dated at 64 Ma (U-Pb zircon; Vega-Granillo et al., 2013).


Petrographic studies were performed on 130 rocks from the Sonobari Complex; locations and brief descriptions are given in the Supplemental Data (Table S2 [footnote 1]). From these, 44 samples were analyzed for major and trace element concentrations. Geochemical analyses were performed at the ALS Chemex laboratory in Vancouver, Canada. Major elements were analyzed by inductively coupled plasma–atomic emission spectroscopy (ICP-AES), while trace elements and rare earth elements (REEs) were analyzed with ICP–mass spectrometry (ICP-MS). Additionally, 42 samples were studied for lead isotopes, 21 for strontium isotopes, and 16 for neodymium isotopes (Table 1). Digestion, chromatography, and isotopic analyses were performed at the University of Arizona, Tucson, Arizona, USA. Between 100 and 200 mg of whole rock was digested for Pb and Sr isotopes, and between 43 and 55 mg for Nd, all from the same whole-rock aliquots. Digestion was done in Parr bombs with a mixture of ultrapure 29N HF and 16N HNO3 and HClO4 at 150 °C. Pb and Sr purification was done in columns with Sr-Spec resin (Eichrom Industries, Darien, Illinois, USA) using twice-distilled HCl and HNO3 with different molarity. Lead isotope analyses were done according to procedures described by Thibodeau et al. (2013), on a GV Instruments multicollector–inductively coupled plasma–mass spectrometer (MC-ICP-MS). All samples were diluted below 50 ppb, and a Tl spike was added to each sample. For calculation of initial ratios, concentrations of U, Th, and Pb were obtained with ICP-MS in the Chemex laboratory. Then, 238U/204Pb, 235U/204Pb, and 232Th/204Pb were calculated using these concentrations and their respective atomic weights and isotopic abundances. Once these ratios were obtained, the initial Pb ratios were calculated using the age of the rock in the general formulae (Faure and Mensing, 2005). Isotopic fractionation of lead was monitored by analyzing the NBS-981 standard (206Pb/204Pb = 16.9405, 207Pb/204Pb = 15.4963, and 208Pb/204Pb = 36.7219) reported by Galer and Abouchami (1998). Errors for Pb isotope ratios were calculated from the external reproducibility of the NBS-981 standard, and the external errors ranged between 0.0022% and 0.0081% for 206Pb/204Pb, between 0.0029% and 0.0081% for 207Pb/204Pb, and between 0.0077% and 0.020% for 208Pb/204Pb (2σ). Relative errors for the samples are between 0.011% and 0.052% (2σ). Strontium isotopes were measured at the University of Arizona using a VG Sector 54 thermal ionization mass spectrometer (TIMS) following the protocol described by Thibodeau et al. (2015). The 87Sr/86Sr ratios were corrected for mass fractionation using 86Sr/88Sr = 0.1194. The average value of the NBS-987 standard over all sample runs was 0.7102594 ± 0.0014 (n = 5, 1σ).

Whole-rock Nd isotopes were analyzed following procedures described by Ducea et al. (2002). Isotopic analyses were performed on a VG Sector 354 multicollector TIMS at the University of Arizona. Concentrations of Sm and Nd were determined by isotope dilution. The average 143Nd/144Nd of the La Jolla Nd standard measured during the analytical session was 0.511864 ± 2 (n = 4). Nd was normalized to 146Nd/144Nd = 0.7219. The estimated analytical ±2σ uncertainties for samples in this study are 143Nd/144Nd = 0.0007%–0.0018%.


Major and Trace Geochemistry

Petrographic descriptions are given in the Supplemental Data (Table S2 [footnote 1]). Major and trace elemental data are available for all samples in the Supplemental Data (Tables S3 and S4 [footnote 1]).

Triassic to Late Jurassic Orthogneisses and Granites (249–150 Ma)

Triassic to Late Jurassic orthogneisses and granites have SiO2 content from 49.1 to 80.5 wt%, implying a compositional variation from gabbro to granite (Fig. 2A). Considering that classification with major elements could be altered by metamorphism, a classification diagram using immobile elements is included in Figure 2B. Magmatic differentiation is also indicated by the magnesium number (Mg#), which varies from 8.6 to 43. The TiO2 contents in felsic rocks are mostly <1.0 wt%. An age-related compositional variation is not evident from either REE or multi-element diagrams. In general, primitive mantle–normalized multi-element diagrams display enrichment of large ion lithophile elements (LILE) with respect to the high field strength elements (HFSE), which is indicated by (Rb/Yb)N (N—normalized values) from 2.28 to 43.29. Patterns display negative anomalies in Nb, Ta, P, and Ti, and relative enrichment in K, Rb, Ba, and Th (Figs. 3A, 3C). In the REE diagrams (Figs. 3B, 3D), most samples are enriched in light REEs (LREEs) with respect to the heavy REEs (HREEs), except for one mafic sample, which displays a flatter pattern. This is evidenced by (La/Yb)N from 1.9 to 19.29. Many Late Jurassic samples display negative Eu anomalies. The Late Jurassic Cubampo granite has a peraluminous composition (Fig. 2C).

Late Jurassic Mafic Rocks (161–151 Ma)

The Late Jurassic Realito Gabbro and Topaco basic dikes have SiO2 contents from 44 to 53 wt% (Fig. 2A) and Mg# from 65.6 to 37.3. In the Nb/Y versus Zr/Ti diagram, all samples have basic composition (Fig. 2B), and in the alkali-FeOt(total)-MgO (AFM) diagram display tholeiitic affinity (Fig. 2D). The multi-element and the REE patterns of the Topaco dikes indicate that they are more enriched than the Realito Gabbro. Compared with Triassic–Jurassic felsic to intermediate rocks, the Nb-Ta anomaly in mafic rocks is less marked, the P and Ti negative anomalies are not present, and instead a positive anomaly in Sr is common (Fig. 3E). In the REE diagram, the Topaco and Realito rocks have a smooth slope with (La/Yb)N varying from 2.2 to 5.1, and two samples display positive Eu anomalies (Fig. 3F).

Post–Late Jurassic Amphibolites and Cenomanian Intrusions (Post 162–98 Ma)

In the WSC, post–Late Jurassic amphibolites and Cenomanian intrusions have 40–58.6 wt% SiO2 (Fig. 2A), with average of 49.75 wt% in amphibolites and 45.5 wt% in later rocks. These basic compositions are also evident in the Nb/Y versus Zr/Ti diagram (Fig. 2B), while in the AFM diagram a tholeiitic affinity is apparent (Fig. 2D). TiO2 ranges from 0.9 to 2.9 wt%, with average of 1.6 wt% in amphibolites and 1.2 wt% in gabbros. In the multi-element diagram, these two units have similar normalized abundances, although amphibolites are slightly more enriched. The pattern has a mild slope with (Rb/Yb)N from 0.9 to 8.6. Most samples have negative anomalies in Nb and Ta. Amphibolites have positive anomalies in Th, while in the Cenomanian gabbro the anomaly is negative (Fig. 3G). Both pulses are enriched in REEs, although amphibolites have somewhat larger values and gabbros are slightly depleted in LREEs (Fig. 3H). The Capomos Granodiorite has 69.2 wt% SiO2, and a moderately low Mg# of 33.6. The multi-element diagram for this rock displays enrichment in LILEs and HFSEs with (Rb/Yb)N of 71.60, negative anomalies in Nb, Ta, Sm, and Ti, and positive anomalies in K and Sr (Fig. 3G). The REE pattern displays high LREEs and HREEs with (La/Yb)N of 15.54 (Fig. 3H). The Shand index indicates that this rock is metaluminous (Fig. 2C).

Campanian Volcanic and Plutonic Rocks (80–73 Ma)

The Campanian Guamuchil Formation has 48.9–58.4 wt% SiO2, but the rocks are mostly basic in composition, which is confirmed in the Nb/Y versus Zr/Ti diagram (Fig. 2B). In the AFM diagram, this unit shows a tholeiitic affinity (Fig. 2D). The TiO2 average in these rocks is 0.96 wt%. The multi-element patterns are very similar to those of the Cenomanian rocks, excepting the Th negative anomaly that is not present on this unit (Fig. 3I). The Campanian felsic leucogranites intruding the WSC display similar behavior to coeval mafic rocks, but with marked negative anomalies in Nb, Ta, P, and Ti, and positive anomalies in K and Sr (Fig. 3I). The Campanian rocks have a very similar REE pattern to the post–Late Jurassic to Cenomanian rocks (Fig. 3J). The leucogranites have peraluminous compositions (Fig. 2C).

Paleocene Los Parajes Granodiorite (64 Ma)

The Paleocene Los Parajes Granodiorite has 58 wt% SiO2 (Fig. 2A), and a low Mg# of 31.6 indicating derivation from an evolved magma. The multi-element pattern indicates enrichment of LILEs with respect to HFSEs with (Rb/Yb)N of 13.52, displaying slight negative anomalies in Nb, Ta, and Ti, and positive anomalies in Kr and Sr (Fig. 3I). The REE patterns are enriched in LREEs with respect to HREEs with (La/Yb)N of 3.10 (Fig. 3J). In the Shand index diagram (Fig. 2C), this unit has a metaluminous character.

Nd, Sr, and Pb Isotopes

A summarized data set of isotopic values is included in Table 1 and a more detailed one in the Supplemental Data (Table S5 [footnote 1]). The Nd isotopic contents vary depending on the age and composition of the rocks. The Triassic and Jurassic intermediate and felsic orthogneisses and metagranites have εNd(i) values ranging from –5.7 to –0.73. Post–Late Jurassic amphibolites and Cenomanian gabbros have values from –1.9 to +4.97. The εNd(i) in the Late Jurassic Realito Gabbro is –0.25. In the Campanian Guamuchil Formation the εNd(i) is 0.36, while a negative value of –3.7 was reported from the Campanian leucogranites (Zurcher, 2002).

Three Triassic orthogneisses have initial Sr ratios ranging from 0.7063 to 0.7064, similar to the 0.7071 value obtained by Valencia-Moreno et al. (2001) in the same area. A low value of 0.7029 obtained in similar rocks can be ascribed to metamorphism as suggested by the low content in alkalis of that sample (SFO-154). The Late Jurassic Realito Gabbro has initial Sr ratios from 0.7044 to 0.7051, while coeval orthogneisses and the Cubampo Granite have ratios from 0.7065 to 0.7130. Post–Late Jurassic amphibolites have an initial Sr value of 0.7060 analogous to a previously obtained value of 0.7063 (Valencia-Moreno et al., 2001). Cenomanian gabbro, diorite, and granodiorite have initial Sr from 0.7038 to 0.7055. The Campanian Guamuchil Formation yields a value of 0.7050, while the Paleocene Los Parajes Granodiorite has initial Sr of 0.7056. In a 87Sr/86Sr(i) versus εNd(i) correlation diagram (Fig. 4), the Triassic and Jurassic rocks and the Campanian leucogranites plot below the chondritic uniform reservoir (CHUR) in εNd(i) and toward higher 87Sr/86Sr(i). Amphibolites and Cenomanian rocks have positive and negative values of εNd(i) following the mantle array evolution trend. The Realito, Capomos, and Guamuchil units have εNd(i) values near those of bulk Earth.

In the initial 207Pb/204Pb versus 206Pb/204Pb diagram (Fig. 5A), our values are arranged along the crustal evolution line of Stacey and Kramers (1975). The larger isotopic ratios are displayed by the Early Triassic, post-Late Jurassic, pre-Cenomanian, and Cenomanian rocks. The majority of the rocks in the ESC show higher values in 207Pb/204Pb than rocks of the WSC. Initial 208Pb/204Pb versus 206Pb/204Pb from all samples are placed in a narrow band with 208Pb/204Pb from 35 to 39 and 206Pb/204Pb from 16 to 19.5 (Fig. 5B).


Tectonic Setting

Diagrams by Pearce et al. (1984) and Whalen et al. (1987) were used to discriminate the tectonic setting of orthogneisses and igneous plutons. Most samples fall in the volcanic arc granite (VAG) field (Fig. 6A), excepting the Late Jurassic Cubampo Granite that crosses the boundary between VAG and within-plate granite (WPG). In the Zr versus Ga/Al diagram of Whalen et al. (1987), most granitoids are of I or S type (Fig. 6B), excepting the Late Triassic diorite, one Cubampo dike, and the Capomos Granodiorite, which fall in the field of anorogenic granitoids (A type).

Diagrams by Shervais (1982) and Cabanis and Thieblemont (1988) were used to distinguish the tectonic setting of volcanic and plutonic mafic rocks. In the Shervais (1982) diagram, samples of the Realito Gabbro fall in the low-Ti island arc tholeiite (IAT) field (Fig. 6C). Most of the Late Jurassic Topaco rocks plot in the IAT and MORB-BAB fields. Amphibolites and Cenomanian mafic rocks plot in mid-ocean ridge basalt—back-arc basin basalt (MORB-BAB) fields. In the diagram by Cabanis and Thieblemont (1988), the Realito, Topaco, and Guamuchil rocks mostly fall in the orogenic subdivision, with some Realito dikes lying in the continental basalt section. Amphibolites and Cenomanian gabbro and diorite are mostly located in or near the BAB field (Fig. 6D), with some samples plotting in the IAT and normal MORB (N-MORB) fields.

The Early Triassic to Late Jurassic orthogneisses and metagranites have TiO2 contents mostly <1.0 wt%, similar to the average contents in modern island or continental arcs (Kovalenko et al., 2010). In contrast, amphibolites and Cenomanian mafic rocks have TiO2 averaging 1.63 and 1.23 wt% respectively, the former value similar to the average 1.68 wt% of TiO2 reported for MORB by Gale et al. (2013). TiO2 as an immobile element suggests an arc setting for the Early Triassic to Late Jurassic magmatism, in spite of any metamorphic overprint, and a trend to MORB or BAB compositions for amphibolites and Cenomanian mafic rocks.

Igneous Petrogenesis

Triassic and Jurassic intermediate to felsic rocks have high LILE and HFSE, high LREE and HREE, negative anomalies in Nb, Ta, P, and Ti, negative εNd(i), and relatively high 87Sr/86Sr(i), which indicate derivation from evolved crustal sources in a continental arc setting (Fig. 4). High content of REEs and negative anomalies in Eu, P, and Ti reflect fractional crystallization of plagioclase + apatite + sphene. For the majority of the magmatic events, inherited zircon data (Fig. 7), as well as crosscutting relationships, indicate melting and/or assimilation of older rocks.

The Realito Gabbro textural characteristics suggest a cumulate process, which is supported by >20 wt% Al2O3, >50 ppm Sc, and >200 ppm Ni (Pearce, 1996) compositions, and by positive anomalies in Sr and Eu. In turn, the Realito and Topaco dikes with lower Mg# and higher REE content are interpreted as derived by fractional crystallization, which is also indicated by slight negative anomalies in Eu and P. Negative εNd(i) and radiogenic 87Sr/86Sr(i) values in both the Realito and Topaco units suggest crustal contamination during magma ascent or emplacement.

Post–Late Jurassic pre-Cenomanian amphibolites emplaced as dikes in older rocks have REE patterns similar to those of MORB, suggesting a depleted mantle source (Fig. 3H). However, high LILEs and negative anomalies in Nb and Ta indicate a subduction influence in the source. These geochemical characteristics have been associated with back-arc basins (Pearce and Stern, 2006). This setting is also supported by the geological relationships and tectonic diagrams (Figs. 6C, 6D). High La/Nb (1.80–3.92) and Nb/Ta (24.8–76) values, compared with depleted mantle values of 1.1 and 15.19 respectively (Salter and Stracke, 2004), suggest assimilation of continental crust. The εNd(i) varying from positive to slightly negative values and the variable 87Sr/86Sr(i) radiogenic values support crustal assimilation. Large variations in Ba (27–495 ppm) and Sr (18–671 ppm) indicate mobility of these elements, probably during metamorphism.

TiO2 and MgO compositions in post-tectonic Cenomanian mafic rocks suggest that these rocks are not cogenetic with amphibolites. Textural characteristics, high content of Al2O3, Sc, Cr, and Ni, and high Mg# in Cenomanian gabbros point to crystal fractionation processes from a mantle-derived magma. This interpretation is also supported by positive εNd(i) values and low 87Sr/86Sr(i) (Fig. 4). Cenomanian diorites have εNd(i) and 87Sr/86Sr(i) similar to those of the host gabbro, indicating a common mantle source but greater differentiation. The Macochin Gabbro in the Sierra Sonobari, with negative εNd(i) and more radiogenic 87Sr/86Sr(i), must have experienced greater contamination by continental crust. Variation in Mg# and REE contents, as well as negative anomalies in Eu, P, and Ti in that unit, support such a fractionation-assimilation process. The Cenomanian Capomos Granodiorite has very similar REE content as coeval diorite of the WSC, although less positive εNd(i) values and more radiogenic 87Sr/86Sr(i) point to a larger evolved source.

In a Nb/Th versus Th/Yb diagram (Fig. 8A), mafic rocks of the ESC have chemical affinity with the lower continental crust (LCC), with a slight tendency to upper continental crustal (UCC). Amphibolites and Cenomanian gabbros display a trend from depleted mantle (DM) to LCC, although amphibolites have slightly larger influence of the UCC. In a Hf/Yb versus Lu/Hf diagram (Fig. 8B), the majority of the samples display variable degrees of lower crust assimilation. The contamination trend is differentiated in the Th/Yb versus Nb/Yb diagram (Fig. 8C), where amphibolites and mafic rocks of the ESC tend to the UCC composition, while Cenomanian gabbros have a larger influence of the LCC.

One of the more conspicuous characteristics of the WSC is the widespread occurrence of peraluminous leucocratic dikes emplaced during Campanian time. The variations in LILEs, REEs, and trace elements indicate a complex origin of markedly evolved magmas. Clear negative anomalies in P and Ti indicate fractionation of apatite and sphene. Negative values in εNd(i) and radiogenic 87Sr/86Sr(i) indicate a crustal source. Coeval with these magmas, tholeiitic basaltic rocks of the Guamuchil Formation were emplaced in the ESC. The Mg# and REE content are interpreted as resulting from fractional crystallization, which is also supported by negative anomalies in Eu and P. Basalts have positive εNd(i) and 87Sr/86Sr(i) similar to those of the mantle array, indicating their mantle derivation. Contamination by continental crust in these magmas is indicated by ages of inherited zircons and the mixing diagrams of Figure 8. The Paleocene Los Parajes Granodiorite displays moderate REE contents and 87Sr/86Sr(i) of 0.706, indicative of a crustal influence. Negative anomalies in Nb and Ta reflect the influence of subduction-derived fluids in the origin of magmas.

Implication for Basement Configuration

In the study area, a geographic pattern in the variation of εNd(i) isotope ratios is not clear (Fig. 9); instead, εNd(i) is more radiogenic in mafic and Cenomanian rocks. Felsic rocks in the ESC have more radiogenic Sr and are richer in 207Pb than felsic rocks of the WSC (Fig. 9). Variation in Pb isotopes seems to be controlled by the age and interaction with a heterogeneous basement, which may have been remnants of old blocks left behind after Pangea breakup. Rocks of the ESC have higher 207Pb/204Pb(i) and more limited range in 206Pb/204Pb(i) than rocks of the WSC (Fig. 5A), reinforcing the separation of these two zones previously proposed on geochronological grounds (Vega-Granillo et al., 2013). The Pb isotopic values of rocks in the ESC are very similar, suggesting a common homogeneous basement and/or assimilation of the Río Fuerte Formation, which is intruded by the Realito Gabbro, Topaco dikes, Cubampo Granite, and Capomos Granodiorite. In the 207Pb/204Pb versus 206Pb/204Pb diagram (Fig. 5A), ratios mostly coincide with those of the Grenville rocks of Mexico. Isotopic data in those rocks overlap with or are similar to data from the Colombian Santa Marta and Garzon Mesoproterozoic massifs (Lopez et al., 2001). Some Pb isotopic ratios in the Sonobari Complex fall in the range of the Arequipa-Antofalla craton, which is a Paleoproterozoic to Grenville basement terrane of eastern South America. However, our data clearly differ from those of the Grenville of Texas, USA. This suggests that the region may be underlain by a Grenvillian heterogeneous basement similar to the Oaxaquia microcontinent (Ortega-Gutierrez et al., 1995) of South American affinity (Fig. 10). However, the Cretaceous rocks have Pb isotopic signatures partially comparable to those of the Guerrero terrane and the Peninsular Ranges batholith (Figs. 5A, 10) (Potra et al., 2014; Shaw et al., 2003), suggesting that similar basements underlie these terranes, or at least did so in the time when the studied samples were formed. In the 208Pb/204Pb versus 206Pb/204Pb diagram (Fig. 5B), our values partially fit with data from the Grenville of Mexico represented by the Guchicovi Complex, and with those from the Colombian Garzon massif (Ruiz et al., 1999). Otherwise, our data differ from those obtained from the Laurentian Mojave and Mazatzal provinces (Fig. 5B) and from the basement of the Caborca area (Iriondo et al., 2004) but partially coincide with data of the Paleoproterozoic Yavapai province. Location of these provinces is displayed in Figure 10.

Neodymium and strontium isotopic values obtained in this work and previously published data were used to build a map encompassing different terranes of Mexico. The εNd(i) data (Fig. 10) show a clear isotopic variation from northern to southern Mexico, with more positive εNd(i) values in the south. That variation is supported by the Triassic rocks in the Sonobari Complex, which have less negative εNd(i) values than Triassic plutonic rocks in northern Sonora (Arvizu et al., 2009), indicating that the basement of the latter region is older than the basement in the Sonobari region. The εNd(i) values in the Sonobari Complex are similar to those of the Permo-Triassic granitoids of southern Mexico, which intrude the Oaxacan and Acatlan complexes (basements of the Oaxaca and Mixteco terranes, respectively; Fig. 10) (Torres et al., 1999), suggesting similar influences and probable relationships before the breakup of Pangea. Jurassic rocks in the Sonobari are isotopically similar to coeval rocks on Maria Madre Island (Fig. 10) (Pompa-Mera, 2014), whereas further south, the Jurassic Guerrero superterrane displays less radiogenic values. This suggests that the Sonobari Complex is located in the overlapping zone between older and younger basement terranes. The 87Sr/86Sr(i) data show a clear isotopic variation from northern to southern Mexico (Fig. 10), with ratios being more radiogenic in north-central Sonora. In the study area, mixed values are found. The Triassic and Jurassic rocks tend to be more radiogenic than Cenomanian rocks, the latter having similar values to those in the Guerrero terrane, Maria Madre Island, Los Cabos block, and Alisitos–Santiago Peak terranes (Fig. 10), suggesting a common basement. Our initial Nd and Sr data are consistent with those of Valencia-Moreno et al. (2001), who ascribed the north-south variation to changes in the basements of the respective regions. Mahar et al. (2016) found that the Hf isotopes in the border between Chihuahua and Sinaloa, east of the study area, display analogous N-S variation. The 207Pb/204Pb(i) values in the study area are similar to values reported in Chihuahua, and both are less radiogenic than those in northern Sonora (Housh and McDowell, 2005). The similarity of the Pb isotope ratios between the study area and Chihuahua suggest that the Mesoproterozoic basement of that region continues into southern Sonora and northern Sinaloa.

Tectonic Implications

In paleogeographic models and tectonic maps, the Sonobari Complex is located in the overlapping zone between Laurentia and Gondwana (Dickinson and Lawton, 2001; Poole et al., 2005; Vega-Granillo et al., 2008). Most authors have placed the northern limit of the composite Guerrero terrane to the north of this region (e.g., Campa and Coney, 1983; Centeno-García et al., 2008). However, surface boundaries between crustal blocks have been obliterated by superposed deformation, intrusion, and Cenozoic volcanic and sedimentary cover.

The earliest tectonic event in the area occurred in the ESC with amalgamation of the southern Laurentian passive margin with Gondwanan blocks approaching from the south (Poole et al., 2005; Vega-Granillo et al., 2008). It is interpreted that this continental collision culminated during the late Permian with metamorphism of the Río Fuerte Formation in the study area (Vega-Granillo et al., 2008) and the thrusting of the Laurentian slope and basin assemblages over coeval platform sequences in central Sonora (Poole et al., 2005). From the Early Triassic to Late Jurassic (Oxfordian), the region was intruded by plutonic rocks derived from the subduction of oceanic lithosphere beneath the continental margin. These magmatic rocks evolved by differentiation of basic magmas and were contaminated by assimilation during their emplacement through previously amalgamated Paleozoic blocks. This scenario is consistent with tectonic models proposing the initiation of subduction along the western side of Pangea in Permo-Triassic times (Dickinson and Lawton, 2001; Dickinson, 2004; Arvizu et al., 2009). Therefore, Triassic rocks in the area are interpreted as the southward continuation of the Permo-Triassic magmatic belt occurring from southwestern USA (e.g., Barth et al., 1990, 1997; Barth and Wooden, 2006; Miller et al., 1995), eastern Peninsular Ranges batholith (Kimbrough et al., 2015; Schmidt et al., 2014), northern Sonora (Arvizu et al., 2009; Arvizu and Iriondo, 2015), eastern and southern Mexico (Torres et al., 1999; Weber et al., 2007; Ducea et al., 2004), and perhaps extending to South America (e.g., Cardona et al., 2010; Maksaev et al., 2014). Jurassic rocks also occur along a continental magmatic arc extending from southwestern USA through the Mojave and Sonora Deserts (Busby-Spera et al., 1990; Barth et al., 2017), northern Sonora (Anderson et al., 2005; Haxel et al., 2005), southern Sonora and northern Sinaloa (this work), central Sinaloa (Cuéllar-Cárdenas et al., 2012), and Maria Madre Island (Pompa-Mera et al., 2013). These rocks are also coeval with rocks in El Arco, Baja California, although the rocks from this region have a geochemical signature indicating an island arc setting (Weber and López-Martínez, 2006). The Late Jurassic magmatism in the eastern region seems to reflect initiation of an extensional phase with emplacement of bimodal plutons including the Realito-Topaco mafic rocks and the Cubampo Granite. However, felsic and mafic rocks display negative Nb and Ta anomalies typical of subduction-related magmatism, which are also found in basalts that underwent crustal contamination (Wilson, 2007) (Figs. 3 and 8). The Cubampo Granite may be related to the thermal input caused by intrusion of the Realito Gabbro into Paleozoic units (e.g., Huppert and Sparks, 1988; Meade et al., 2014). Emplacement of mafic dikes (amphibolites) in the WSC postdating the Kimmeridgian and predating the Cenomanian suggests a western migration of the extensional phase. Mixed MORB and subduction-related characteristics of these magmas have been associated with back-arc basins (e.g., Pearce and Stern, 2006), but an intra-arc setting cannot be precluded. This extensional back-arc or intra-arc setting is consistent with geologic relationships of mafic dikes intruding continental granitic rocks. An Early Cretaceous back-arc basin in northern Sonora–southern Arizona has been also proposed to form the depocenter of the Bisbee Group (Bilodeau, 1982; Dickinson and Lawton, 2001). The tectonic setting of the post–Late Jurassic pre-Cenomanian magmatism in the Sonobari is different from the oceanic arc setting of the partially coeval Alisitos arc of western Baja California (Busby et al., 2006).

Following emplacement of the mafic dikes, a collision event occurred, which caused deformation and orogenic metamorphism. Considering the age of the involved rocks, this event could be related to the collision of the Alisitos oceanic arc against the continental border (Sarmiento-Villagrana et al., 2016; Vega-Granillo et al., 2017). The final collisional event has been dated in Baja California at between 111 and 103 Ma (Johnson et al., 1999; Schmidt et al., 2014). This event closed the back-arc or intra-arc basin in the study area and caused amphibolite facies metamorphism and deformation in rocks older than ca. 100 Ma. Thermobarometry of the metamorphic rocks indicates that the region underwent burial to a depth of ~30 km (Vega-Granillo et al., 2017). During the latest stages of this event, the Río Fuerte Formation was thrust over the Topaco Formation causing heterogeneous mylonitic deformation and greenschist facies metamorphism in both units and in the Cubampo Granite (Vega-Granillo et al., 2011, 2017). In the WSC, thermal readjustment caused a temperature peak and migmatization at ca. 90 Ma (Sarmiento-Villagrana et al., 2016; Vega-Granillo et al., 2017), ca.10 Ma after cessation of the deformation event. That thermal peak was enhanced by Cenomanian post-tectonic gabbro plutons, and diorite, hornblendite, and pyroxenite dikes. Positive εNd(i) and low 87Sr/86Sr(i) of these rocks suggest mantle sources, negative anomalies in Nb-Ta evidence continued influence of subduction-related fluids, while negative Th anomalies indicate little to no influence of the upper crust consistent with their emplacement deep within an overthickened crust. The MORB-BAB tectonic setting of these rocks (Fig. 6) and their isotopic signatures similar to that of oceanic-island basalt suggest that an extensional phase followed the main contractional event. Cenomanian magmatism occurred in a transitional period between the cessation of subduction underneath the Cordilleran margin, collision of the Alisitos arc, and inception of the Farallon plate subduction in the western side of Baja California. In the ESC, the Capomos Granodiorite originated at that time. Some Cenomanian rocks in the area have geochemical and isotopic compositions similar to those of post-tectonic mafic rocks in central Sinaloa (Henry et al., 2003), coeval gabbros in Baja California (Schmidt et al., 2014; Kimbrough et al., 2015), the Los Cabos block (Schaaf et al., 2000), and gabbros in the Guerrero terrane (Villanueva-Lascurain et al., 2016). Peraluminous magmatism similar to that of the 99–92 Ma La Posta suite reported in the Peninsular Ranges batholith has not been recognized in the study area.

Campanian (83–80 Ma) magmatism in the WSC represented by numerous post-tectonic leucocratic dikes is ascribed to partial melting of an overthickened crust during an active subduction process that generated the Laramide batholiths of Sonora and Sinaloa. Coeval magmatic rocks have been reported from coastal and central Sonora (e.g., McDowell et al., 2001; Ramos-Velázquez et al., 2008; Roldán-Quintana et al., 2009), Chihuahua-Sinaloa (Mahar et al., 2016), and Maria Madre Island (Pompa-Mera et al., 2013). In the ESC, the Guamuchil rocks have REE and trace element compositions very similar to those of the Topaco Formation. This suggests that the Guamuchil rocks originated from subduction-related magmatism in an extensional intra-arc environment. The Paleocene Los Parajes Granodiorite has a continental arc geochemical signature and is coeval with the extensively studied Laramide belt of Sonora and Sinaloa (e.g., Damon et al., 1983; Valencia-Moreno et al., 2001; Dickinson and Lawton, 2001; Henry et al., 2003).


The Sonobari Complex records an extended magmatic history starting in the Early Triassic and culminating in the Late Cretaceous (Campanian), which includes six main pulses related to or influenced by subduction processes. Intrusion of mafic rocks after the Late Jurassic and before the Campanian is interpreted as recording two extensional phases with an intervening compressional phase. The interpreted tectonic scenario of that period is complex, implying rifting behind a continental arc (i.e., back-arc or intra-arc), closing of a marginal sea by collision of the Alisitos arc against mainland Mexico, and rifting behind a continental arc when Farallon plate subduction initiated under western Baja California. Our geochemical and isotopic data indicate that the Sonobari Complex rocks were formed and evolved in a continental arc setting similar to that of the eastern Peninsular Ranges batholith and distinct from the oceanic arc setting of the Guerrero-Alisitos arc. Consequently, we suggest that the northern limit of the Guerrero terrane must be reappraised and may be located further south than the study area. Recent geophysical data suggest that the Alisitos arc runs parallel to continental rocks of the eastern Peninsular Ranges batholith (Langenheim et al., 2014) and does not continue into mainland Mexico as previous models proposed. The similarity of Pb isotopic compositions of the rocks in the study area and those of the Grenville rocks of Mexico, as well as the presence of inherited zircons in the magmatic rocks, indicate the existence of old basement blocks under the study area that have remained since Pangea breakup. These blocks are distinct from those in the Paleoproterozoic crust of northern Sonora.


This work was supported by the Consejo Nacional de Ciencia y Tecnología (CONACYT), grant number 177668 to Ricardo Vega-Granillo. Special thanks to Mark Baker, laboratory manager in the Geosciences Department at the University of Arizona. Thanks to Frederick Arroyo and Analine Vázquez from the Universidad Autónoma de Guerrero for assistance with isotope chromatography. We appreciate the thorough and useful reviews by Andrew Barth, Peter Schaaf, and associate editor Nancy Riggs. The English version of the manuscript benefitted from the detailed review by Travis Ashworth, which is greatly appreciated.

1Supplemental Data. Table S1: Geochronology; Table S2: Petrographic description; Tables S3 and S4: Geochemistry tables; and Table S5: Detailed isotopic analyses. Please visit http://doi.org/10.1130/GES01540.S1 or the full-text article on www.gsapubs.org to view the Supplemental Data.
Science Editor: Shanaka de Silva
Associate Editor: Nancy Riggs
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.

Supplementary data