text new page (beta)
English (pdf)
Article in xml format
Article references
Automatic translation
Send this article by e-mail
Cited by SciELO 
Similars in
SciELO 
Permalink
INTRODUCTION
Gabbros and associated mafic rocks are significant constituents of many calc-alkaline granitoids and abundant in numerous Cordilleran plutons (e.g., Barbarin, 2005; Kimbrough et al., 2015). Understanding their petrogenesis provides relevant information about the role of mafic magmas in the generation and evolution of the Earth ́s crust.
The acquisition of reliable isotopic and geochronologic data of magmatic rocks is a key for determining the nature of the basement of tectonostratigraphic terranes in Mexico. In this paper, we report on gabbroic rocks found within mantle-type granite near Jilotlán, Jalisco, in the central part of the belt of Cretaceous Cordilleran plutons of Mexico (Figure 1a). This plutonic and batholitic belt extends from Baja California to Oaxaca parallel to the Pacific coast with NW-SE decreasing Late Cretaceous to early Miocene ages (~103-20 Ma) obtained by various geochronologic methods. In Jalisco, Colima and Michoacán the plutonic bodies and some associated volcanic rocks show in general Late Cretaceous to Paleocene ages (Gastil et al., 1978; Pantoja-Alor, 1983; Grajales-Nishimura and López-Infanzón, 1984; Schaaf, 1990; Schaaf et al., 1995; Morán-Zenteno et al., 2000; 2005; Ferrari et al., 2014). The predominant position of the plutonic belt close to the Pacific coast evidences a continental margin truncation that indicates intensive tectonic processes from the Late Cretaceous to the early Miocene in western Mexico, including forearc removal by subduction erosion (Malfait and Dinkelman, 1972; Schaaf et al., 1995; 2000; Schaaf, 2002; Morán-Zenteno et al., 2005).
Figure 1 a) Regional geologic map showing the distribution of plutons, volcanic and basement rocks between Puerto Vallarta and Acapulco (modified from Carta Geológica de la República Mexicana del Servicio Geológico Mexicano, 2007, escala 1: 2,000,000;); using digital elevation models from the Instituto Nacional de Estadística y Geografía (INEGI). The location of the study area is shown as well as the Tepalcaltepec and Manzanillo sampling locations. b) Location of the Guerrero Terrane, in relation to other tectonostratigraphic terranes of Mexico (from Centeno-García et al., 2008).
The study area is located in the Zihuatenejo terrane (Figure 1b), a part of the Guerrero composite terrane (Campa and Coney, 1983). The limits and evolution of the Guerrero terrane are still under discussion, because of its complex and variable stratigraphy and its large extension (Centeno-García et al., 2008; Figure 1b). In this contribution, we integrate petrographic, geochemical, isotopic, and geochronologic data in order to understand the magmatic origin of the Jilotlán gabbros and their relationship to the surrounding mantle-type granites. The results of this work contribute to the understanding of the nature of the crust in the eastern Guerrero terrane, its evolution, and tectonic setting in the late Early Cretaceous to Paleogene time span.
GEOLOGIC SETTING
Guerrero terrane
The Guerrero terrane was first defined by Campa and Coney (1983) as a composite terrane conformed by submarine volcanic and sedimentary sequences of Late Jurassic to middle Cretaceous age and a report of Upper Triassic rocks near Zacatecas. The aforementioned authors divided the Guerrero terrane into three subterranes (Teloloapan- Ixtapan, Zihuatanejo, and Huetamo), based on different stratigraphic features, metamorphic grade, and deformational style. The terrane was later named Náhuatl Terrane by Sedlock et al. (1993), characterized by deformed and metamorphosed Jurassic to Cretaceous magmatic and sedimentary rocks of continental and marine affinity with a probable occurrence of Paleozoic rocks in the eastern part of the terrane. The Náhuatl terrane corresponds to the southern part of the Guerrero terrane of Campa and Coney (1983). Referring to both nomenclatures, a main difference is that the Guerrero-Morelos platform is included in the Náhuatl Terrane, whereas it is not included in the Guerrero Terrane (Elías-Herrera et al., 2000).
Further studies, based on stratigraphic and geochemical data, subdivided the Guerrero terrane into the Tahue, Zihuatanejo, Guanajuato, Arcelia, and Teloloapan terranes (Centeno-García et al., 1993, 2003, 2008; Talavera et al., 1995; Mendoza and Suástegui, 2000; Centeno- García, 2005) with volcanic and sedimentary marine successions, and minor subaereal Upper Jurassic (Tithonian) to middle-Upper Cretaceous (Cenomanian) sequences (Centeno-García et al., 2003). Some scarce Paleozoic units were also reported (Centeno-García, 2005; Centeno-García et al., 2008).
Nature of the Guerrero terrane
Some authors emphasized the autochtonous character of the Guerrero terrane, which was predominantly built on continental crust (De Cserna, 1978; De Cserna et al., 1978). Others described it as a para- autochtonous marine volcanosedimentary sequence, polydeformed and metamorphosed to greenschist facies, with continental arc affinity and evidence of continental crust in its southern part (Elías-Herrera and Sánchez-Zavala, 1992; Elías-Herrera and Ortega-Gutiérrez, 1997; Elías-Herrera et al., 2000). Less current models proposed it as an allochthonous terrane, consisting of a distant oceanic island arc that would have collided with North America during the late Early Cretaceous (Lapierre et al., 1992) or the Late Cretaceous (Tardy et al., 1994), and would have closed an ocean basin. Another study suggested that the ocean basin had two subduction zones, verging east and west, before the easterly travelling arc finally closed the basin and was accreted to North America during the Early Cretaceous (Dickinson and Lawton, 2001).
Most authors define the Guerrero terrane as a complex system of one, two or three peripheral arcs, that developed relatively close to the continent (Campa and Coney, 1983; Elías-Herrera and Sánchez- Zavala, 1992; Centeno-García et al., 1993, 2003; Elías-Herrera et al., 2000; Mendoza and Suástegui, 2000; Centeno-García, 2005; Talavera et al., 2007), unseparated (Elías-Herrera and Sánchez-Zavala, 1992; Centeno-García et al., 1993, 2003, 2008; Elías-Herrera et al., 2000; Centeno-García, 2005; Martini et al., 2009), or separated by one or more subduction zones (Mendoza and Suástegui, 2000; Talavera et al., 2007) throughout its evolution.
The oldest basement of the Guerrero terrane is composed of Paleozoic crust, probably of oceanic arc affinity in the northwestern section (Tahue terrane), and a Triassic deep ocean basin with contemporaneous rift magmatism in the central part (Zihuatanejo terrane), which was accreted to mainland Mexico during the Early Jurassic (Centeno-García, 1994, 2005; Centeno-García and Silva, 1997; Centeno-García et al., 2003, 2008). Upon this accretionary complex, one or more middle to upper Mesozoic arc(s) evolved (Centeno-García et al., 1993; Elías-Herrera et al., 2000; Mendoza and Suástegui, 2000).
Zihuatanejo terrane
The study area is located in the southern part of the Zihuatanejo terrane, the largest Guerrero subterrane (Centeno-García et al., 2008 and references therein). This terrane was formed by the accretion of fringing terranes produced by extensional (rifting) and compressional (accretion) processes in the upper plate (Centeno-García et al., 2011) of a single long-lived subduction zone (Centeno-García et al., 1993, 2008) that dipped eastward underneath the Mexican continental mar- gin (Martini et al., 2009; Centeno-García et al., 2011). The oldest rocks found so far are located near the city of Zacatecas (Zacatecas Fm.) to the north, as well as in Huetamo and Arteaga, Michoacán (Arteaga complex) and in Las Ollas, Guerrero (Las Ollas complex) in the south (Centeno-García, 2005 and references therein; Centeno-García et al., 2008 and references therein). These sequences are composed of Triassic ocean floor assemblages, which were deformed and metamorphosed during the Early to Middle Jurassic.
Basement of the Zihuatanejo terrane in the study area
The basement rocks in the study area are probably similar to those of the Arteaga complex, which is composed of tectonically imbricated quartz-rich turbidites (Centeno-García et al., 2008 and references therein). According to Centeno-García (2005), the turbidites include siliciclastics, green volcaniclastics, pillow basalts, diabase, gabbros, black and green chert, and large recrystallized limestone olistoliths (Centeno-García, 1994; Centeno-García et al., 2003). The siliciclastics are interbedded with thick green layers, which are interpreted as metavolcaniclastics derived from a primitive arc source and/or Mid Ocean Ridge Basalts (MORB). Pillow lavas also show geochemical and isotopic compositions similar to MORB (Centeno-García et al., 1993, 2003). The whole sequence is at least ~1500 m thick and its deposition time is constrained by one Late Triassic (Ladinian-Carnian) radiolarian fossil report (Campa et al., 1982) and by the youngest detrital zircons with an age of 260 Ma (Centeno-García et al., 2003); therefore, the deposition could not be older than late Permian. One sample of the sedimentary matrix of the Arteaga complex has distinctive zircon clusters at 260 Ma and 1.0 Ga, with minor clusters at 480-650 Ma and 800 Ma. The source for these peaks is most likely the Acatlán complex of the Mixteca terrane (Centeno-García et al., 2011).
According to Centeno-García et al. (2008), deformation varies from gently folded strata to highly sheared block-in-matrix textures, and their metamorphic grade ranges up to high greenschist to amphibolite facies (Centeno-García et al., 2003). Blueschist facies rocks have been reported by Talavera (2000) in the Las Ollas locality.
The undeformed 163±3 Ma Macías granite, located close to Tumbiscatío, Michoacán, cuts the deformed and metamorphosed units (Centeno-García et al., 2011), which also indicates that deposition and deformation of the Arteaga complex is post-260 Ma and pre-163 Ma (Centeno-García et al., 2011). The 183 Ma Tizapa metagranite of the Tejupilco metamorphic suite, with continental geochemical affinity, is considered to be the limit between the ocean floor assemblages and the continental margin (Elías-Herrera et al., 2000) and would define the exact time of accretion towards the continent. Therefore, these rocks are interpreted as a part of an Upper Triassic to Middle Jurassic subduction-related accretionary complex along the Mexican continental margin (Centeno-García, 2008).
The Arteaga sediments were deposited in a deep ocean basin, probably contemporaneously with parts of the rift-related magmatic activity (Centeno-García et al., 2003; Centeno-García, 2005). The basin received deep marine, erosion-derived sediments as well as MORB- type lavas and air-fall ash deposits derived from an island arc. All this would have occurred in a marginal back arc basin or in an open ocean environment (ocean floor/continental rise setting). Deformational structures suggest that the complex was deformed during subduction processes and correspond to the upper levels of an accretionary prism (Centeno-García et al., 2003).
Basement cover
Middle and Upper Jurassic to earliest Cretaceous evolved arc rhyolitic submarine lavas, volcaniclastic rocks (163-145 Ma magmatic peak; 154-157 Ma at Cuale, Jalisco, western Zihuatanejo terrane; Bissig et al., 2008) and granitoids (163±3 Ma, Macías granite) cover and cross-cut deformation and metamorphic structures, respectively. In the Early to middle Cretaceous (Barremian-Albian) arc, two magmatic peaks at 129-123 Ma and 109 Ma were identified from detrital zircons from fluvial and shallow marine rocks in the Zihuatanejo terrane (Centeno- García, 2005; Centeno-García et al., 2008, 2011). This arc is considered as extensional, because detritus and zircons from the Arteaga complex, as well as from the Middle Jurassic to earliest Cretaceous and Early to middle Cretaceous arcs, were locally recycled into an Aptian-Albian intra-arc basin. Additional evidence for the existence of the arc is given by Early Cretaceous shallow marine to non-marine sedimentary formations that rest unconformably on Jurassic granitoids in the Arteaga- Tumbiscatio area (Centeno-García et al., 2011). During this time span, the Late Jurassic to Early Cretaceous (Tithonian-Cenomanian) Arperos basin (Sierra de Guanajuato) was formed in a back arc regime when the Guerrero terrane drifted in the paleo-Pacific domain towards the west during extension, and was subsequently accreted back to the Mexican craton (Martini et al., 2011).
The presence of recycled Santonian-Maastrichtian (85-65 Ma) arc volcanic rocks is evidenced by detrital zircons in volcanosedimentary sequences (Centeno-García et al., 2011). Undeformed granitoids of Late Cretaceous and Paleogene ages are reported by Schaaf (1990), Valencia et al. (2009) and Martini et al. (2010) in the coastal Zihuatanejo terrane. This arc is interpreted by Centeno-García et al. (2011) as compressional, because Cenomanian strata were folded together with the Early Cretaceous strata prior to the deposition of Santonian and younger sediments, which is confirmed by an angular unconformity. Therefore, this shortening must have taken place between Turonian and Santonian (93-84 Ma).
Uppermost Jurassic and Cretaceous arc successions of the study area, coastal Zihuatanejo terrane
According to Centeno-García et al. (2008), the Cretaceous arc successions include basaltic, andesitic, and some rhyolitic volcaniclastic rocks, interbedded with limestones, evaporites, and red beds (Grajales- Nishimura and López-Infanzón, 1984). These successions contain abundant fossils (rudists, gastropods, microfossils, fossil logs, and vertebrates). The uppermost Cretaceous (Santonian-Maastrichtian) rocks are red beds interbedded with andesitic volcanic rocks that rest on all previously described units (Altamira-Areyán, 2002; Benammi et al., 2005).
The study area is located within the Puerto Vallarta and Tepalcaltepec region with the following stratigraphy (Figure 2), based on the observations of Centeno-García et al. (2003) and other authors referred below.
Figure 2 Stratigraphic column with the main formations in the study area and its close su- rroundings. Rb-Sr biotite-whole rock cooling age determined by Schaaf (1990).
Jurassic limestone. The oldest rocks in the study area are limestones found near Coquimatlán, Colima, which contain early Tithonian ammonites (150-145 Ma; Michaud et al., 1987), are interbedded with volcanic and volcaniclastic rocks and are unconformably overlain by Cretaceous arc volcanic rocks.
Alberca Formation. The oldest Cretaceous rocks correspond to the lower Cretaceous Alberca Formation that outcrops near Colima (Pimentel, 1980). Its lower member consists of interbedded black shale, sandstone, limestone, and tuff. The upper member is composed of basaltic and andesitic flows, limestone and shale. This formation contains Berriasian-Hauterivian (145-130 Ma) ammonites (Cuevas, 1981).
Tecalitlán Formation. Upwards, the Alberca Formation transition- ally changes to the Tecalitlán Formation (Rodríguez, 1980; Pantoja-Alor and Estrada-Barraza, 1986), which locally records the main magmatic event with andesitic, basaltic, and some minor rhyolitic flows, intercalated with pyroclastics (intermediate tuffs and ignimbrites) and lahars, reworked tuff with limolite, and calcareous sandstone. Given that it covers the Alberca Formation and is overlain by the Tepalcaltepec Formation, a Barremian-Aptian age (130-112 Ma) is presumed. Additionally, near La Minita, in the Pihuamo region, the Tecalitlán Formation contains early Aptian rudists (125 Ma; Centeno-García et al., 2011 and references therein). A 118.5±2.5 Ma K-Ar hornblende age was reported for a porphyritic andesite, collected west of Tepalcaltepec, that is considered as the upper part of Tecalitlán Formation (Bermúdez- Santana, 1994).
Tepalcaltepec Formation. Apparently, the volcanic activity decreased during the early Albian and finally ceased in the middle to late Albian. During this time span, the Tepalcaltepec Formation was deposited (Pimentel, 1980), composed of thick limestone layers intercalated with andesitic flows and some rhythmic volcaniclastic and calcareous deposits, tuffs and rhyolites, conglomerates and sandstones, red limolites, and some evaporites (Grajales-Nishimura and López-Infanzón, 1984). A reef-type limestone with a fossil record indicating an Albian- Cenomanian age (112-93 Ma) was reported by Corona (1999). In the central Colima region, the Tepalcaltepec Formation laterally changes to the Madrid Formation (Pano, 1975), which consists of thin-bedded limestone, calcareous shale, gypsum, and few andesitic flows and tuff. The studied fossils are of Albian to early Cenomanian age (112-99 Ma, Grajales-Nishimura and López-Infanzón, 1984). According to Centeno-García et al. (2011), the Madrid and Tepalcaltepec forma- tions may correlate with the Encino (late Aptian fossils, Buitrón, 1986) and Vallecitos (early Albian fossils, Alencaster, 1986) formations, respectively (both ~112 Ma). The contact between the Tecalitlán and Tepalcaltepec formations could not be established in this study; how- ever, Centeno-García et al. (2011) stated that the Tecalitlán Formation was tilted prior to the deposition of the Encino Formation. We assume the same type of contact between the Tecalitlán and Tepalcaltepec formations in the study area.
Cerro de la Vieja Formation. The Cerro de la Vieja Formation (Parga, 1977) constitutes a conglomeratic unit composed of limestone fragments, which was folded together with the underlying Cenomanian limestone. Interlayered lava flows were dated by the K-Ar method at 80±6 and 78±6 Ma (Grajales-Nishimura and López-Infanzón, 1984). Conglomerates contain sandstone and silt layers with sedimentary structures and textures that suggest a fluvial continental depositional environment.
Granitoids and gabbros. Some of the formations listed above are intruded by granitoids with ages between 68 and 55 Ma (four K-Ar biotite and hornblende ages, three Rb-Sr biotite-whole rock ages, as well as an eight point whole rock isochron; Schaaf, 1990). As observed in the field, most of these granitoids also intrude gabbroic intrusive rocks in the study area.
GEOCHEMISTRY OF THE EARLY CRETACEOUS ARC - PREVIOUS WORKS
Most geochemical studies in the area were focused on the Early Cretaceous volcanic rocks (e.g., Centeno-García et al., 2011). Rocks of the continuous chain of plutons along the Mexican Pacific margin range from gabbros, plagiogranites, quartz monzodiorites to tonalites and from granodiorites to granites. The plutonic rocks with the highest silica and potassic feldspar content occur in the Puerto Vallarta batholith, whereas plagiogranites (M-type granites; Pitcher, 1982) and gabbros have only been observed so far between Manzanillo and Jilotlán (Schaaf, 1990; Schaaf et al., 1995).
Geochemical, isotopic, and geochronological studies on the intrusive rocks of the Mexican Pacific coast between Puerto Vallarta and Acapulco reported by Schaaf (1990), Schaaf et al. (1995, 2000), and Morán-Zenteno et al. (2000) show SiO2 variations between 47.3 2 and 76.9 wt. %, calc-alkaline character, as well as volcanic arc affinities deduced from Rb vs Yb+Nb diagrams. The alumina saturation index (molar Al2O3/Na2O+K2O+CaO) is, in general, lower than 1.1, which corresponds to I-type granites (Pitcher, 1982). These Mesozoic- Cenozoic intrusive show heterogeneous initial 87Sr/86Sr from 0.7033 to 0.713 and εNd values between +6.4 and -7. The most radiogenic Sr and least radiogenic Nd isotopic ratios are from the eastern portion of the Puerto Vallarta batholith, whereas the magmatically most primitive values are found between Manzanillo, Jilotlán, and Petatlán-Papanoa in the Colima, Jalisco, and Guerrero states, respectively. For plagiogranites intruding the Jilotlán gabbros a crystallization age of 68±12 Ma was obtained from a low-radiogenic eight point whole rock Rb-Sr isochron with an initial 87Sr/86Sr of 0.70377. The initial εNd values for the granitoids are between +3.5 and +4.3 and for a gabbro sample a +7 value was obtained (Schaaf, 1990). Gabbros have been also reported from the Manzanillo pluton and further to the NW along roadcuts between Barra de Navidad and La Huerta, Jalisco, which form a 69±3 Ma 14-point Rb-Sr isochron (Schaaf, 1990, 2002). This age is similar to the 67±1 Ma U-Pb zircon age obtained from the Compuertas mafic intrusive near Manzanillo (Panseri, 2007). Delgado-Argote et al. (1992) studied diorites, gabbros and ultramafic rocks (dunite-clinopyroxenite) in Petatlán-Papanoa and El Tamarindo, southeast of Zihuatanejo, Guerrero. They published a ~112 Ma 40Ar/39Ar hornblende age for an ultramafic body. Recently, Ferrari et al. (2014) reported U-Pb zircon ages of 47.9±1.8 Ma and 41.3±0.6 Ma, and one 40Ar/39Ar feldspar age of 36.1±0.5 Ma for gabbros from the Petatlán-Tecpan area.
The volcanic rocks of the Guerrero terrane range from basalts, andesites and dacites to rhyolites, with Th/Yb and Ta/Yb ratios indicating transitions between oceanic island arcs and continental margin arcs (Centeno-García et al., 1993, 2003; Tardy et al., 1994; Freydier et al., 1997; Mendoza and Suástegui, 2000). The basalts have REE patterns similar to those of island arc rocks, with negative Eu anomalies (Centeno-García et al., 1993; Freydier et al., 1997). Felsic lavas are calc-alkaline and show LREE enrichment compared to the basalts (Freydier et al., 1997). The initial εNd parameters are between +3 and +7.8 for all compositions (Centeno-García et al., 1993; Freydier et al., 1997), similar to modern island arcs. Neodymium TDM model ages span from 290 to 460 Ma, which excludes the participation of Precambrian crust in magma generation processes (Centeno-García et al., 1993).
GEOLOGY AND FIELD WORK
Figure 3 shows the geological map of the study area elaborated from fieldwork observations and photointerpretation. The Jilotlán pluton is composed by granitoids of tonalitic to granodioritic compositions that intrude a gabbroic body. The gabbros are located in the central part of the pluton between Jilotlán and Tecalitlán.
Figure 3 Geologic map of the study area and location of the analyzed samples. The topographic map E13B46 from INEGI (scale 1:50,000) was used as base.
The granitoids intrude the Tecalitlán Formation, which is the host rock. Locally, outcrops of this formation are highly altered. Lavas of the Tecalitlán Formation are primarily andesitic in composition and in less proportion dacitic. Few fine-grained sandstone layers alternate with the lavas. The overall formation is highly fractured and faulted (Figure 4a).
Figure 4 a) Contact of the Tecalitlán Formation with a granitic body/dike. b) Microgabbros and coarse grained gabbros intruded by a tonalitic to granodioritic body. c) Mafic porphyritic dike intruding a tonalitic to granodioritic body with mafic enclaves. d) Syntectonic alignment of mafic enclaves in the felsic body. e) Pegmatite with amphibole and pyroxene phenocrystals in a gabbro with potassic feldspar and pyroxenes. f) Several generations of dikes intruding a tonalitic body of the Jilotlán pluton.
The gabbros and the granitoids exhibit textural variations. Gabbros almost always occur as microgabbros with some subordinate coarse- grained species (Figure 4b). The gabbros were cool when they were intruded by the granitoids, which is indicated by sharp contacts.
Within the granitoids, an alignment of mafic enclaves was observed (Figure 4c, 4d), as well as potassium feldspar crystals in contact with pyroxenes, characteristic for a mineral association in disequilibrium (not shown). Cumulates of hornblende and pyroxenes are also present. All this suggests that mingling involving mafic and felsic minerals occurred within the felsic facies of the pluton, probably produced by fractional crystallization of a parental mafic magma.
Pegmatites within the pluton contain plagioclase pheno- and megacrysts, amphibole, and some pyroxenes in the gabbroic parts (Figure 4e). In the more felsic areas, phenocrysts and megacrysts of quartz and potassium feldspar were observed; some parts of the felsic bodies display aplitic textures.
Several generations of mafic dikes (Figure 4c, 4f) with aphanitic- porphyritic to aplitic phaneritic-porphyritic textures intrude all facies of the Jilotlán pluton as well as the Tecalitlán Formation (Figure 4c, 4d, and 4f). The dikes show a preferred NE-SW orientation and a thickness ranging from a few centimeters up to five meters.
In the Tepalcaltepec area, outcrops of granodiorite, gabbro, and mafic dikes were observed. A thick ignimbrite layer covers the gab- bros. Further to the west, conglomerates overlie packages of fossil-rich limestone, intercalated with sandstone and shale. At La Codiciada mine (Figure 1a) the intrusion event formed a skarn with barite and wollastonite mineralization.
Samples of the gabbros and granitoids, as well as host rocks and mafic and felsic dikes of the Jilotlán-Tecalitlán area were collected. One gabbro from Tepalcaltepec was also sampled. Additionally, two gabbro samples were obtained from the Manzanillo batholith on the Solidaridad viewpoint over Bahía de Santiago along the Manzanillo-Cihuatlán bypass road, approximately 4 kilometers NW from Manzanillo. The Manzanillo gabbros show considerable textural variations in distances of just a few meters apart, expressed by very fine to coarse-grained zones with magmatic foliations, and pyroxene and hornblende cumulates. All sampling locations are shown in Figures 1a and 3.
PETROGRAPHY
Detailed petrographic descriptions were performed on the most important units between Tecalitlán and Jilotlán. Five gabbros were selected (JLT-8A, JLT-12, JLT-30-2, JLT-32 and JLT-45), as well as one granitoid (quartz-monzodiorite JLT-23), one host rock sample (-9-), and one mafic dike (JLT-13C). Additionally, one gabbro sample from Tepalcaltepec (JLT-41) and two gabbros from Manzanillo (MZ-3, MZ-4) were analyzed.
The point counting results and corresponding classification names are given in Table 1 (except for host rock -9-). Microphotographs showing petrographic features of the samples are included as Figures A1 to A5 of the Supplementary Material.
Table 1 Point counting results and modal classification of the analyzed samples.
Plagioclase (P), potassium feldspar (FK) and quartz (Q) points normalized to 100% and reported in % units were used for the modal classification in the Streckeisen (1979) diagram (not shown). Rocks were considered microgabbros when the plagioclase size was fine- (< 1mm) to medium-grained (1 to 5 mm). The size of the mafic minerals was also determined, and is mentioned in each classification. A (-) mark appears when a mineral was altering another mineral, for example, chlori tized hornblende was counted as hornblende. All opaque minerals were considered as oxides, and few samples presented sulphides in the field or in hand sample.
In general, gabbroic samples present equigranular textures. Euhedral and/or anhedral plagioclase crystals, the most abundant mineral (Figure A1), have anorthite contents of 60 to 68% determined by the Michel-Levy method. Some crystals show antiperthitic textures (Figure A1b), evidencing lack of mixture in the rock, and saussuriti- zation, which implies a compositional change from calcic to sodic plagioclase (e.g., Figure A1d; quartz-monzodiorite). Some plagioclase crystals are also sericitized.
The second most abundant mineral is black and green hornblende that is frequently chloritized. In many cases, it is not distinguishable from clinopyroxene due to alteration and uralitization (Figure A2a). Poikilitic textures were observed in hornblende oikocrystals containing plagioclase, pyroxene and oxides (Figure A2b, A2c). Some gabbros present biotitization of hornblende and/or pyroxene (Figure A2d). Besides chlorite, epidote is another abundant alteration mineral in hornblende, probably due to hydrothermal or meteoric alteration (Figures A1a, A3a, A3b, A3c). Several clinopyroxene and pseudomorphic hornblende crystals show plagioclase inclusions (Figure A3d). This, together with the pyroxene crystals surrounded by hornblende, could indicate an excess of calcium content during hydratation of the magmatic system, leaving calcium to crystallize as plagioclase. On the other hand, rapid magma ascent could have generated a lack of mixture between clinopyroxene and plagioclase in the magmatic system.
Oxides display a skeletal texture and most of them are surrounded by titanite, (Tirich magnetite, Figures A1b, A2e, A3e). This texture results when hornblende and pyroxene require iron for formation, leaving the excess titanium to crystallize around the corroded oxides. Prehnite and potassium feldspar were observed in a few samples (Figure A3e) evidencing metasomatism related to the granitoid intrusion. The reason for the absence of potassium feldspar in some samples may lie on a local lack of fluids to achieve homogeneous potassium oversaturation or due to a sampling bias.
The basaltic dike (JLT-13C) contains mafic xenoliths with clinopyroxene, hornblende, prehnite and epidote crystals (Figure A4a, A4b). There are also mafic crystals, including actinolite, that form a cumulophyric texture (Figure A4c, A4d).
The quartz-monzodiorite (JLT-23, Figures A1c, A4e, A4f, A5a) contains plagioclase (An52), hornblende, biotite, and accessory clinopyroxene and titanite. The host rock sample from the Tecalitlán Formation (-9-, Figures A5b, A5c, A5d, A5e), collected near the contact with the granitoid, was petrographically analyzed but not point counted because many parts of the thin section are highly altered and the minerals could not be distinguished. Due to the presence of andalusite, white mica, rounded quartz and plagioclase, the rock is classified as quartz-rich sandstone metamorphosed during the granitic intrusion.
GEOCHEMISTRY
Major elements
Major elements were obtained by X ray fluorescence at the Laboratorio Universitario de Geoquímica Isotópica (LUGIS), Instituto de Geología, Universidad Nacional Autónoma de México (UNAM). All analytical procedures can be consulted in Lozano Santa-Cruz and Bernal (2005) and Villanueva-Lascurain (2011). Concentrations are expressed as weight percentage and iron is expressed as total Fe2O3. For the AFM diagram Fe2O3 and FeO were automatically calculated with the Petrograph software (Petrelli et al., 2005) and also manually calculated using Fe2O3=Fe2O3t×0.15 and FeO=Fe2O3t×0.85×0.8998 (Ragland, 1989).
Table A1 of the Supplementary Material presents the analytical results of the 10 samples described petrographically (except the host rock) together with their coordinates and altitudes. Additionally, two gabbros from Jilotlán (MS-24 and MS-25) from Schaaf (1990), a gabbro from Aquila (Aq-1), a meta-sandstone near Tecalitlán (ATG-1) and two Rivera plate basalts (D5-2Ext and D5-2Int) from Valdez-Moreno (2006) were used for comparison.
Most gabbros and the quartz-monzodiorite from Jilotlán are sub- alkaline according to the silica vs. total alkalis diagram (Figure 5a) of Cox et al. (1979), although some samples plot on the dividing line. In the silica vs. potassium diagram of Peccerrillo and Taylor (1976) (Figure 5b), the majority of Jilotlán (JLT, MS), Aquila (Aq-1) and Manzanillo (MZ) gabbros form part of the middle potassium calc-alkaline series. The quartz-monzodiorite (JLT-23) and the Tepalcaltepec gabbro (JLT- 41) are more enriched in potassium, whereas the JLT-30-2 gabbro from Jilotlán is shoshonitic. Gabbro MS-24, the basaltic dike JLT-13, and the Rivera basalt samples (D5-2) are tholeiitic.
Figure 5 a) Classification diagram of silica (SiO2) versus total alkalis (Na2O+K2O) for plutonic rocks, modified from Cox et al. (1979). b) Silica (SiO2) versus po- tassium (K2O) subclassification diagram for subalkaline plutonic rocks of Peccerillo and Taylor (1976). Sample numbers are given above each symbol (see Table 2). Two gabbros from Jilotlán (MS-24 and MS-25) from Schaaf (1990) as well as a gabbro from Aquila (Aq-1) and two Rivera plate basalts (D5-2Ext and D5-2Int) from Valdez-Moreno (2006) are included for comparison.
Sodium enrichment in basalts has been explained by hydrothermal alteration, and high potassium content in granites can be the product of contact metamorphism (Rollinson, 1993). These anomalies are probably due both to hydrothermal alteration and metasomatism caused by the intrusion of the granites. This explains the crystallization of potassium feldspar present in some gabbros (e.g., Figure A3e) by metasomatic potassium oversaturation.
Trace elements
Trace element concentrations of Jilotlán, Tepalcaltepec, and Manzanillo samples (Table A2 of the Supplementary Material) as well as those from Aquila and the Rivera plate reported by Valdez-Moreno (2006) were obtained by ICP-MS at the Laboratorio de Estudios Isotópicos (LEI), Juriquilla, UNAM (methods described in Mori et al., 2007; Villanueva, 2011). Sample MS-25 was analyzed by neutron activation analysis at the University of Mainz, Germany (Schaaf, 1990).
The trace element behavior in the investigated gabbros and the quartz monzodiorite is shown in a multielement diagram (Figure 6, normalized to N-MORB of Sun and McDonough, 1989). Nearly all rocks show enrichments in the most incompatible elements but have similar values in the less incompatible ones compared to N-MORB and to the Rivera plate sample (D5-2int). However, D5-2int also shows enrichments in the most incompatible elements (e.g., Cs, Ba) when compared to N-MORB, possibly due to secondary alteration processes.
Figure 6 Multielemental diagram normalized to N-MORB (Sun and McDonough, 1989) with increasing compatibility from left to right for the analyzed rocks of this and previous studies. *Valdez-Moreno (2006).
The gabbro from Jilotlán with the lowest overall element concentrations is JLT-32 (except for Cs, Rb and Ba) suggesting its origin from an even more depleted source; the one with the highest concentrations (except for the most incompatible elements) is JLT-30-2, being also the only sample without positive Sr and Pb spikes in Figure 6. This is probably due to the fact that plagioclase is not calcic enough to con- centrate Sr, and most Pb was removed by hydrothermal alteration. Sr is abundant in carbonate phases, concentrates in calcic plagioclase, and can be transported in aqueous solutions (e.g., derived from calcareous sediments in the subducted slab). Pb is also transported in aqueous solutions derived from marine or continental sediments of the subducted slab or the continental crust. This explains the positive Sr and Pb anomalies in most Jilotlán gabbros compared to the Rivera plate basalts (Figure 6). The negative Nb anomaly is apparently due to fractionation of ilmenite and/or rutile, given that hornblende and titanite are widely present in the gabbros.
The Tepalcaltepec (JLT-41), Aquila (Aq-1), and Manzanillo (MZ-3, MZ-4) gabbros show similar trace element patterns in comparison to those from Jilotlán, with higher Rb and Ba values for JLT-41, possibly caused by metasomatic processes. The Aquila gabbro (Aq-1) shows slightly lower concentrations in most trace elements than those of Jilotlán, indicating a more depleted source. However, high positive Sr and Pb anomalies suggest greater plagioclase accumulation and crustal participation, respectively, in comparison to Jilotlán. The Manzanillo gabbro MZ-3 reveals the lowest trace element concentrations of all studied samples, except for Sr. This is probably due to a very high Ca content in plagioclase (58% An). The source of this rock is more depleted than that of Jilotlán, Tepalcaltepec and Aquila. Interestingly, the MZ-4 sample, collected just a few meters away from MZ-3, shows the mean trace element concentrations for Jilotlán, Tepalcaltepec and Aquila.
The quartz-monzodiorite (JLT-23) has higher Cs, Rb, Ba, Th and U values in comparison to the gabbros, which indicates that its source is more enriched. Well-defined positive Sr and Pb anomalies are best explained by calcic plagioclase accumulation (52% An) and a sediment component, respectively, which is also supported by high Th and U values. Less incompatible elements such as Dy, Ho, Er, Yb, and Lu have lower values than in the gabbros, suggesting a slightly more evolved magmatic source.
The meta-sandstone (ATG-1) has overall patterns and concentrations comparable to the JLT-23 quartz-monzodiorite. The negative Ba anomaly could be caused by potassic feldspar removal, and the high Th and U values indicate a continental source, also comparable to JLT-23. The Pb contents are the highest and Sr and Eu the lowest of all samples, which emphasizes the origin from a continental source and plagioclase erosion, respectively.
REE patterns
The Rare Earth Element (REE) patterns (normalized to the CI Carbonaceous Ivuna chondrite values of McDonough and Sun, 1995) of all samples displayed in Figure 6 are graphed in Figure 7. All show negative slopes (with the exception of the Rivera plate basalt D5-2int), which result from the enrichment in light REE (LREE) with respect to the heavy REE (HREE). Two groups are distinguishable within the Jilotlán gabbros: samples JLT-30-2, JLT-45, JLT-8A and JLT-13 with stronger LREE enrichment in comparison to samples JLT-32, JLT-12 and MS-25. The Tepalcaltepec gabbro JLT-41 shows very similar REE patterns to the Jilotlán gabbros, whereas the Aquila sample (Aq-1) has lower concentrations. The Manzanillo MZ-3 sample displays a positive Eu anomaly, probably because reducing conditions with overall low oxygen fugacity caused Eu+2 substitution for Ca in accumulated plagioclase. On the other hand, the MZ-4 gabbro shows a slight negative Eu anomaly, which suggests that Eu+2 was captured by Ca-rich plagioclase during fractionation. Therefore, the opposite Eu anomaly behavior of both MZ samples suggests different processes involved in the generation of the Manzanillo gabbros from the same source, which could have been even more depleted than the sources of the Jilotlán, Tepalcaltepec and Aquila gabbros. Sample MZ-3 represents also the most depleted gabbro of all according to REEs.
Figure 7 REE distribution patterns normalized to CI chondrite values of McDonough and Sun (1995) with increasing compatibility from left to right in the analyzed rocks of this and previous studies (see legend in Figure 6).
The JLT-23 quartz-monzodiorite crosscuts both groups of Jilotlán gabbros with stronger LREE/HREE fractionation. This suggests that it originated from a more evolved source than the gabbros.
The meta-sandstone ATG-1 has generally very low REE concentrations and a pronounced negative Eu anomaly probably due to the absence of Ca-rich plagioclase.
Jilotlán, Tepalcaltepec (JLT), Aquila (Aq-1) and Manzanillo (MZ) gabbros have trace element distribution patterns (Figure 6) with typical subduction zone geochemical signatures such as high Large Ion Lithophile Element (LILE) concentrations in comparison to High Field Strength Elements (HFSE) and REE, as well as positive Pb and Sr anomalies. These signatures could be attributed to fluid release from the subducted slab/sediments to the mantle wedge (Plank and Langmuir, 1998; Winter, 2001). Rare earth element patterns (Figure 7) are LREE enriched compared to HREE, which can be accomplished by low degrees of partial melting or high degrees of fractional crystallization from a primitive mantle source (e.g., Winter, 2001). The relatively flat pattern in HREE rules out a garnet presence in the source.
Tectonomagmatic environment
To crosscheck the magmatic environment in which the Jilotlán gabbros were generated, we selected the triangular Zr/4-2Nb-Y diagram for basalts of Meschede (1986) (Figure 8). Although this diagram was constructed on the basis of 1800 basalt analyses, it can be used also for gabbros like those from this study. Most of the Jilotlán and the Tepalcaltepec samples, with the exception of MS-25, situate in the C field that corresponds to intra-plate tholeiites and volcanic arc basalts. Gabbro MS-25 together with those from Aquila (Aq-1) and Manzanillo (MZ), as well as the Rivera plate basalt (D5-2int), plot in the D field that corresponds to normal mid ocean ridge basalts (N-MORB) as well as volcanic arc basalts. The JLT-23 quartz-monzodiorite plots in the volcanic arc granite (VAG) field in the Rb vs Yb+Ta diagram of Pearce et al. (1984; not shown) for granitic rocks.
Figure 8 Discrimination diagram for the tectonic setting of basalts (Zr/4- 2Nb-Y) of Meschede (1986). Sample numbers are given above their symbols (see Table 2). *Valdez-Moreno (2006).
Models for magmatic processes
To understand which processes generated the geochemical differences between the Jilotlán gabbros and the other igneous rocks, REEs were used in partial melting and fractional crystallization models from a hypothetical source. We chose the batch melting model of Shaw (1970) (Equation 1), in which the melt is in equilibrium with the solid up to a critical amount and separates from the solid as an independent system. This simple model explains best the trace element distributions of MORB, the most adequate material to study mantle melting. The Rayleigh crystal fractionation model (Equation 2) is a continuous crystallization model in a closed reservoir in which each crystal is separated from the melt as soon as it forms; therefore there is no equilibrium between the solid and the melt. Equations for both models are shown below:
where
Both equations were solved for C
l. From the partition coefficients (
Given that a probable source for the gabbros is a garnet-free mantle, an upper mantle plagioclase peridotite was considered as the source material. The theoretical source rock was composed of 53% olivine, 28% orthopyroxene, 17% clinopyroxene, and 2% plagioclase (after Gómez-Tuena et al., 2007). The initial concentrations (C i) in ppm were La=0.192, Ce=0.550, Nd=0.581, Sm=0.239, Eu=0.096, Dy=0.505, Er=0.348, Yb=0.365, and Lu=0.058 according to the values of the average depleted mantle of Workman and Hart (2005). These values were taken mostly from White (2005) and some elements in olivine and Lu in plagioclase from Winter (2001) (Table A3 of the Supplementary Material).
Only the concentration of the elements in the melt (Cl ) was calculated, as all samples plot between the batch melting and crystallization curves obtained in Nb/Zr vs. Nb and La/Sm vs. La diagrams (not shown here) and none of them even close to the residue curves (Villanueva, 2011). On the basis of the chemical similarities observed in the previous diagrams, the Jilotlán samples were graphed separately (Figures 9a and 9b).
For a melt, small degrees of melting (F) or high degrees of crystallization (X) will greatly increase the content of the most incompatible LREE but will not significantly affect the least incompatible HREE. On the other hand, high degrees of F or low degrees of X will greatly decrease the content of the least incompatible HREE and will leave the most incompatible LREE unaffected. Between 1 and 99%, 1% increments were tested for both F and X to see which values best reproduce the patterns displayed by all samples. None of the obtained curves fully reproduces these patterns; however, low degrees of F and high degrees of X give the closest approximation.
As to the quartz-monzodiorite (JLT-23), 1% melting of the depleted average mantle partially reproduces its pattern; although the REE concentrations of the sample are slightly higher (Figure 9a). In the case of gabbros, the enriched set (JLT-30-2, 45, 8A and 13C) can be approximately matched by 97 to 99% crystallization from the complete melting of this mantle, whereas the more depleted set (JLT-12, 32 and MS-25) agrees with 94 to 97% crystallization of the complete mantle melt. The Tepalcaltepec gabbro (JLT-41) is partially reproduced by 96 to 97% melt crystallization (Figure 9b) and for the MZ-4 gabbro a melt crystallization of 98 to 99% best models its REE pattern and concentrations. As for MZ-3 gabbro, a 3 to 20% mantle melting would partially reflect the obtained patterns. In the case of the Aq-1 gabbro, 94 to 95% melt crystallization or 1% fusion of this mantle would best match the pattern and concentrations. The Rivera plate sample (D5- 2int) does not have similarities to any patterns, probably because the modeled source is the average depleted mantle (enriched + depleted mantle), whereas the source of this sample is exclusively the depleted mantle (N-MORB).
Figure 9 a) REE diagram of the gabbros and quartz-monzodiorite JLT-23 of the Jilotlán pluton and plots of the same elements for different degrees of batch melting (F) and crystal fractionation (X) for the melt of a plagioclase peridotite with the composition of the average depleted upper mantle according to Workman and Hart (2005). b) The same diagram for the Tepalcaltepec, Manzanillo, Aquila gabbros and the Rivera plate basalts. *Valdez-Moreno (2006).
This modeling is simplistic as it only considers one source (the mantle) and only one process at time. However, if an average depleted mantle was the source, most samples could be generated by high degrees of crystallization and less probably low degrees of partial melting. The quartz-monzodiorite would be generated by low degrees of mantle melting.
Rb-Sr AND Sm-Nd ISOTOPE SYSTEMATICS
Sample preparation for Rb-Sr and Sm-Nd isotope analyses was performed at the LUGIS facilities, Instituto de Geofísica, UNAM, following the procedures described by Villanueva (2011) and Pompa- Mera et al. (2013). The isotope dilution technique was applied to all samples that were analyzed by thermal ionization mass spectrometry (TIMS), except for samples Aq-1, ATG-1, D5-2Ext and D5-2Int (Valdez-Moreno, 2006), for which the concentrations determined by ICP-MS were used to calculate the 87Rb/86Sr and 147Sm/144Nd ratios (Table 2). The initial 87Sr/86Sr and 143Nd/144Nd ratios and εNd values were calculated according to the ages obtained in this study and from the literature (Table 2). For the Jilotlán gabbros (JLT-8A, JLT-12, JLT- 30-2, JLT-32, JLT-45, MS-25), an age of 114 Ma was used (see next section). For the basaltic dike (JLT-13C), the ~70 Ma eight sample Rb-Sr whole rock isochron age of the granitic rocks of the Jilotlán complex was applied (Schaaf, 1990). For the Tepalcaltepec gabbro (JLT-41), a ~100 Ma U-Pb zircon age of a granitoid from Tepalcaltepec was used (Salazar-Juárez, 2015). For the Manzanillo gabbros (MZ-3, MZ-4), a 70 Ma U-Pb zircon age was employed (Panseri, 2007). For the Aquila sample (Aq-1), a 57 Ma Rb-Sr biotite cooling age was available (Schaaf, 1990). As the age of meta-sandstone (ATG-1) is unknown, the oldest available K-Ar hornblende age of 118 Ma for the host rock (Tecalitlán Formation) was used (Bermúdez-Santana, 1994).
Table 2 Isotopic ratios of 87Rb/86Sr, 87Sr/86Sr, 147Sm/144Nd, 143Nd/144Nd, εNd and their corresponding initial ratios marked with the i suffix. Nd model ages (TDM) are also shown.
Sr, Nd and Sm isotope ratios were determined using a Finnigan MAT262 thermal ionization mass spectrometer (TIMS) equipped with eight Faraday collectors, installed at LUGIS, Instituto de Geofísica, UNAM. Isotopic measurements were made in static collection mode. Analyses of Rb were carried out in dynamic mode using a single collector NBS mass spectrometer at the Instituto de Geología, UNAM. Rb, Sr, Sm, and Nd were loaded as chlorides and measured as metallic ions. Standard deviations (1sd = ±1σabs) refer to the last two digits. Standard error of the mean (2SE(M) = 2σabs√n) values were calculated from 60 individual isotopic determinations for Sr and Nd, 40 for Rb, and 20 for Sm. The measured 87Sr/86Sr values were normalized to a 86Sr/88Sr value of 0.1194, and those of 143Nd/144Nd to a 146Nd/144Nd ratio of 0.7219. 87Sr/86Sr of the NBS987 and 143Nd/144Nd of the La Jolla standards measured throughout this study were 0.710237 ± 23 (1σabs, n = 391) and 0.511867 ± 25 (1σabs, n = 204), respectively. Rb concentrations were obtained by isotope dilution using a 87Rb spike. Sr, Nd and Sm concentrations were obtained by isotope dilution using a combined 84Sr-145Nd-149Sm spike. Relative uncertainty for 87Rb/86Sr was ±2% and for 147Sm/144Nd ±1.5% (1σ). Relative reproducibility (1σ) for Rb, Sr, Sm and Nd concentrations was ±4.5%, ±1.8%, ±3.2%, and ±2.7%, respectively. Total procedure blanks were 0.22 ng for Rb, 4.8 ng for Sr, 0.20 ng for Sm, and 0.71 ng for Nd.
The Depleted Mantle (DM) Nd values used for model age (TDM) calculations are 0.513089 and 0.2128 for 143Nd/144Nd and 147Sm/144Nd, respectively, which correspond to mean values obtained from San Luis Potosí upper mantle xenoliths (Schaaf et al., 1994). The present day Chondritic Uniform Reservoir (CHUR) 143Nd/144Nd and 147Sm/144Nd values used for (εNd)i calculations are 0.512638 and 0.1967, respectively (Faure and Mensing, 2005).
Initial 143Nd/144Nd vs. 87Sr/86Sr ratios for all samples are plotted in Figure 10 and accompanied with data fields from Cordilleran Mexico (Puerto Vallarta batholith) and from the Andean and Japan volcanic arcs (Winter, 2001). The initial 143Nd/144Nd values for the gabbros are less radiogenic than typical depleted mantle values (e.g., lherzolitic xenoliths in basanites from San Luis Potosí; Schaaf et al., 1994) but still plot within the mantle array (Figure 10). Therefore, they could be obtained a) from a mantle wedge enriched in sediment/slab components by continuous subduction (<Sm/Nd than MORB), and/ or b) the participation of crustal components (<Sm/Nd than mantle) mixed with mantle material. In this scenario, the Nd model ages (TDM) of all gabbros (averaging TDM = ~500 Ma) would indicate an average crustal residence age of all components involved in gabbro genesis.
Figure 10 143Nd/144Nd vs. 87Sr/86Sr diagram modified from Winter (2001) with initial (i) isotope ratios of the Jilotlán samples and those from previous studies. The inset shows in detail the samples within the Mantle Array field (red square). The MORB field is the average of different mid-ocean ridge basalts. The Japan field is an average of the Japanese island arc. NVZ and SVZ are the northern and southern volcanic zone of the Andes, respectively (all data from Winter (2001). The average Puerto Vallarta batholith values of Cordilleran Mexico are also displayed (Schaaf, 1990). All mentioned fields were graphed with their actual isotope ratios. *Valdez-Moreno (2006).
The initial 87Sr/86Sr values indicate very little participation of old continental crust (>Rb/Sr) that would have enriched the mantle wedge by slab components, subducted sediments, or crustal contamination. Thus, magma generation processes included moderately evolved crust/ slab components and/or sediments mixed with mantle material. As expected, the quartz-monzodiorite originated from a more evolved source than the gabbros. The Tepalcaltepec gabbro shows similar initial 143Nd/144Nd values to those of Jilotlán, but with higher initial 87Sr/86Sr, indicative for some secondary carbonate enrichment by alteration. Both Manzanillo (MZ) gabbros show initial Sr-Nd isotopic compositions very similar to Jilotlán. Therefore, their previously observed differences in trace element are probably due to different crystallization degrees, and less probably the result of different partial melting degrees (Figure 9b).
GEOCHRONOLOGY
In order to determine the age of the gabbros and compare them to the surrounding granitoids (~70 Ma), a Jilotlán gabbro sample (JLT- 12) was chosen to perform 40Ar*/39Ar geochronology on two types of hornblende (black and green). From this sample, and additionally from gabbro JLT-32, zircons were selected for U-Pb geochronology. The techniques and equipment of LUGIS at UNAM were used for mineral separation. Hornblende was isolated by handpicking and standard sepa- ration procedures (Villanueva, 2011) to guarantee inclusion-free crystals. Energy-dispersive X-ray spectroscopic analysis were performed on some crystals to test the purity of the concentrate at the Instituto de Geología, UNAM. Approximately 80% of the crystals were hornblende and 20% clinopyroxene, which are K-free or have negligible K concentrations and have thus no influence on the 40Ar*/39Ar age. 40Ar*/39Ar analyses of these minerals were performed at the University of Alaska at Fairbanks, where the samples were further purified. Analytical procedures are described in York et al. (1981), Layer et al. (1987), and Layer (2000). Black and green JLT-12 hornblende separates show similar plateau ages of 114.6±1.1 Ma and 111.0±1.1 Ma, respectively (Figure 11; Tables A4, A5 of the Supplementary Material). The plateaus include more than 80% of 39Ar released during analysis. The positively sloped Ca/K and Cl/K diagrams evidence some alteration and argon loss in both samples, although this condition is more pronounced in the green hornblende with higher Ca/K ratios (Figure 11). Given the well-defined plateau, the 114.6±1.1 Ma and the 111.0±1.1 Ma horn- blende ages are considered as cooling ages (540±40 °C; Harrison and McDougall, 1980). However, the obtained ages could be minimum ages, considering a possible reset of the Ar-Ar system as a consequence of the granitoid intrusion. To prove this, zircon U-Pb geochronology was performed on the same and an additional sample (JLT-32).
Figure 11 Step heating spectrum for JLT-12 black and green hornblende: age versus 39Ar released fraction (in red) and steps to define the plateau ages (in blue); also shown are Ca/K and Cl/K ratios versus 39Ar released for both minerals.
As zircons are generally not very abundant in gabbros and mafic rocks, only a few crystals were obtained from both samples (19 crystals for JLT-12 and 20 for JLT-32) using handpicking and standard procedures (without Wilfley table, Villanueva, 2011; Hernández et al., 2012). The zircons were analyzed by LA-MC-ICP-MS at the University of Arizona (Table A6 of the Supplementary Material; Figure 12). Analytical procedures are described in Gehrels et al. (2006; 2008) and in Pompa-Mera et al. (2013). The individual ages, ranging from 104 to 123 Ma (Table A6 of the Supplementary Material) are mostly concord- ant. Zircons are considered as magmatic since no zoning was observed in cathodoluminiscense pictures. TuffZirc ages (Ludwig and Mundil, 2002) were calculated using the Isoplot 3.6 Software (Ludwig, 2008). An age of 112.84 +1.20 -1.72 Ma was obtained for gabbro JLT-12 and of 114.61 +1.34 -0.93 Ma for gabbro JLT-32 (Figure 12).
Figure 12 206Pb*/238U vs. 207Pb*/235U Concordia diagrams for zircons from JLT-12 (a) and JLT-32 (b) gabbros. Error ellipsoids correspond to 2σ errors. To the right, the age distributions are also shown in TuffZirc diagrams with 2σ errors. Some binocular microscope photographs of the analyzed crystals are also displayed.
The 40Ar*/39Ar hornblende ages and the U-Pb zircon age of JLT- 12 are identical within standard deviations, indicating rapid cooling of the gabbroic magma. Considering the errors in the JLT-12 U-Pb age, as well the U-Pb zircon age of 114.6 +1.34/-0.93 Ma of sample JLT-32, the crystallization of the gabbros took place between 111.12 and 115.94 Ma. These ages also indicate a time span of at least 40 Ma between the gabbro formation and the emplacement of the surrounding granitoids, which were dated at 68 ± 12 Ma (eight point Rb-Sr whole rock isochron; Schaaf, 1990) and 54±10 Ma (conventional U-Pb zircon age; Salazar-Juárez, 2015).
DISCUSSION
Petrogenesis
The mineralogy of all gabbroic samples resulted from a hydrated source (plagioclase with An50-68 + Hbl + Cpx ± Opx) with hornblende being by far the most abundant mafic mineral. The presence of clino-pyroxene crystals surrounded by hornblende (Figures A2a, A3d) suggests hydratation of the magmatic system after pyroxene crystallization. However, as some hornblende crystals are in contact with well-developed clinopyroxene crystals, hornblende growth directly from the magma is also a possible mechanism. Perhaps both, mag- matic hornblende and secondary hydratation occurred: biotitization of hornblende (Figure A2d) suggests a second hydratation event, probably related to the granitoid intrusion.
All the gabbros were produced in a magmatic arc over a subduction zone, which is indicated by geochemical data (high Pb and Sr concentrations and higher LILE concentrations in comparison to HFSE or REE) and mineralogy (hornblende-rich gabbros with abundant water). On the other hand, the similar distributions and small differences in trace element concentrations of the gabbroic samples from Jilotlán, Tepalcaltepec, Aquila, and Manzanillo and their very similar isotopic ratios, evidence different magma generation processes from the same source, probably by different degrees of crystal fractionation. In comparison to the gabbros, the quartz-monzodiorite is magmatically slightly more evolved.
Geochemical and isotopic compositions of the gabbros and the quartz-monzodiorite underline the primitive magmatic character of this part of the Zihuatanejo Terrane. Nd model ages (TDM) around 500 Ma exclude the participation of old and evolved crustal components as it was observed further to the NW in the Puerto Vallarta batholith (0.6-1.6 Ga Nd (TDM) model ages; Schaaf, 1990). Subduction erosion processes (e.g., von Huene and Scholl, 1991) could have derived fragments of the volcanosedimentary cover of the Zihuatanejo Terrane (Jurassic-Cretaceous arc and sedimentary rocks) into the Wadati Benioff Zone and the mantle wedge accompanied by sediments of the oceanic lithosphere. Another possible crustal component could be the mafic lower crust. The existence of hornblende rich gabbros and the presence of plagiogranites with similar ages to the gabbros in the northern part of the Jilotlán pluton (Salazar-Juárez, 2015) support the existence of at least 20 km thick primitive crust to allow magma differentiation. On the other hand, magmatic activity involving pure mantle material indeed occurred in the ultramafic body (dunite- clinopyroxenite) of El Tamarindo in the Petatlán-Papanoa area, with an 40Ar*/39Ar hornblende age of ~112 Ma (Delgado-Argote et al., 1992), which is identical to the age of the Jilotlán gabbros. Given the geochemical features, isotopic ratios and TDM ages of all samples, the source of the gabbros is constrained to be most likely a result of partial melting of mantle material enriched by sediments/fluids from the slab with some participation of primitive crustal components. The more evolved quartz-monzodiorite JLT-23 could have been generated by partial melting of mafic crust alone or from a combination of mantle with subduction related and mafic crustal components, with later fractional crystallization in both cases. However, a greater amount of subduction or crustal components was involved in the genesis of the quartz-monzodiorite compared to the gabbros.
The modeled REE behavior using the hypothetical depleted mantle as the only source shows that high degrees of crystal fractionation partially reproduce the patterns and the different element concentrations of the gabbros. The quartz-monzodiorite is partially reproduced only by low degrees of partial melting from the same depleted source.
Geochronology
The ~40 Ma time span separating the gabbro and granitoid intrusions indicates that the gabbros in the study area were not comagmatic with the granitoids, although both were produced by subduction processes. The gabbros experienced very rapid cooling as the Ar-Ar horn- blende and the U-Pb zircon ages are identical within errors, considering their 540±40 °C and ~750 °C closure temperatures, respectively. This implies that these rocks crystallized in a shallow crustal zone and/ or were uplifted very rapidly. The plagioclase inclusions observed in some clinopyroxene and hornblende crystals could be the result of exsolution processes during rapid uplift. The granitoids intruded the gabbros ~40 Ma later and did not reheat them above ~540 °C, a fact that is also supported by field relations. These granites could be part of a batholith, also emplaced in shallow crustal levels. The 40Ar*/39Ar hornblende and U-Pb zircon crystallization ages suggest that the Jilotlán gabbros can be considered as the igneous roots of the hosting Tecalitlán Formation from which a 118 Ma K-Ar hornblende age was reported (Bermúdez-Santana, 1994).
Regional tectonics
When comparing the Sr-Nd isotopic signatures of all Jilotlán samples to those from other magmatic arcs (Figure 10), similarities to the Southern Volcanic Zone (SVZ) of the Andes are observable. The SVZ shows calc-alkaline and mid-K geochemical characteristics with relatively flat HREE patterns (Winter, 2001 and references therein). The SVZ crust is 30 to 35 km thick in average, mainly of Mesozoic and Cenozoic ages with geochemical affinities to island arcs and oceanic crust, but in an undoubtedly continental arc setting that is probably similar to the Jilotlán area ~114 Ma ago.
The Jilotlán gabbros were generated during the Early to middle Cretaceous (Barremian-Albian) arc event, considered as extensional (Centeno-García, 2005; Centeno-García et al., 2008, 2011; Martini et al., 2011). The extension could have allowed mantle magmas to ascend through the thinned and geochemically primitive basement crust to produce the gabbros. As discussed above, a pure mantle source is less likely. The gabbros could have been generated by partial melting of the mantle and emplaced along faults, whereas the plagiogranites of the same age (north of Jilotlán) would be derived from partial melting of thicker crustal blocks and/or from fractional crystallization of under- plating mafic magmas ascending through the faulted blocks. Ultramafic rocks like those of Petatlán-Papanoa were generated from pristine mantle material emplaced through deeper/wider faults without crustal contamination. How and where the basal primitive crust was created is unknown, but it was a part of or closely located to continental Mexico prior to the deposition of the Arteaga complex in the Late Permian (before ~260 Ma). The accretion of Arteaga to continental Mexico occurred in the uppermost Early to Middle Jurassic (183 Ma-163 Ma; Elías-Herrera et al., 2000; Centeno-García et al., 2011).
Early Cretaceous magmatism in an extesional regime is also documented during the opening and evolution of the Arperos basin (Martini et al., 2011; 2014) in central (Guanajuato) and southern (Santo Tomás-Arcelia, Figure 1b) Mexico. Peraluminous rhyodacitic dikes and lava flows were emplaced along the Guerrero Terrane and continental mainland margins during the Tithonian to Barremian (150-125 Ma) (Martini et al., 2011; 2014). During this time period, the Arperos basin was still floored by continental rocks or sediments, as is suggested by inherited Paleozoic and Precambrian zircons. Instead, intraplate-like and MORB basaltic flows dominated in the Aptian-Cenomanian (125-94 Ma), which suggests spreading within the Arperos basin. Its closure occurred in the late Cenomanian as the Guerrero Terrane was accreted back to continental Mexico (Martini et al., 2014). During the Tithonian-Aptian (150-112 Ma), between Guanajuato and Arcelia, the primitive calc-alkaline El Paxtle-Palmar Chico-Teloloapan magmatic arc was active in the Guerrero Terrane (Martini et al., 2011; 2014). It had a basaltic to andesitic composition, with scarce felsic volcanic activity. In this scenario, the Jilotlán gabbros would be contemporaneous with the intraplate and MORB basalts (Aptian-Cenomanian). This suggests that the locus of magmatism within the Guerrero Terrane primitive arc migrated from central and southern Mexico towards the Jilotlán area during the Aptian (125-114 Ma) during regional extensional tectonics.
The quartz-monzodiorite was formed during the Santonian- Maastrichtian arc event (Centeno-García et al., 2011) that is considered of compressional character. The compression could have thickened the crust and allowed higher crustal partial melting degrees and magma differentiation, with less direct participation of mantle material. The formation of the quartz-monzodiorite from such thickened crust ex- plains well its different evolution compared to the gabbros.
The very primitive character of the crustal gabbros studied in this work, confirms the existence of a basement that is magmatically poorly evolved without old crustal components (εNdi up to ~+7 and TDM up to ~0.7 Ga (~0.50 Ga average)). This differs substantially from the basement underneath the Puerto Vallarta batholith (~100 Ma) with εNdi parameters down to ~-7 and and TDM up to ~1.55 Ga (Figure 10, Table 2), suggesting a tectonic limit. An alternative interpretation would be the presence of a lithospheric scale fault that facilitated the emplacement of mafic magmas without significant contamination by continental crust (e.g., Kirsch et al., 2012). In this case, the fault would not have apparently affected the Puerto Vallarta batholith. However, no local or regional structural investigations were performed in this study to prove this.
Regionally, gabbros are abundant within the mid-Cretaceous Peninsular Ranges batholith (PRB) in southern and Baja California (Kimbrough et al., 2015), ranging in composition from cumulate olivine gabbro, gabbronorite to hornblende gabbro. A western zone within this batholith contains abundant gabbros whereas in the eastern zone gab- bro is rare (<1% of the outcrop area). The U-Pb zircon crystallization ages range from 130.5±1.5 Ma to 100.2±3.0 Ma for the western zone gabbros, which is consistent with the ~30 Ma time for the emplacement of the overall western zone batholith. In contrast, the eastern zone gabbros have crystallization ages from 109.4±2.9 Ma to 97.8±2.3 Ma that predate the ca. 92-98 Ma tonalite-trondhjemite-granodiorite batholith flareup of the eastern zone, and may have played a role in crustal thickening and flareup triggering. The whole rock and trace element compositions of all the gabbros closely match signatures for high-alumina olivine tholeiite, arc basalt, and basaltic andesite from modern arcs. Nearly flat REE patterns for western and eastern zone gabbros reflect derivation from a common partially contaminated depleted mantle source. This contrasts with the geochemical and isotopic data of the granitoids, which reflect progressively deeper partial melting of mafic source rocks from west to east with additional input from a radiogenic end member in the eastern side of the batholith.
In this scenario, the Jilotlán gabbros ages (~114 Ma) are coeval with both the western and eastern gabbro belts of the PRB (~130-100 Ma). The presence of some plagiogranites north of Jilotlán of the same age as the Jilotlán gabbros is also observed in the western zone gabbros in the PRB, where the crust is probably thinner than in the eastern zone (Kimbrough et al., 2015). These similarities together with paleo- reconstructions, where Baja California was part of continental Mexico, suggest a regional magmatic arc setting with prevalent gabbroic magma generation in the PRB, and, in contrast, apparently lower abundances between Puerto Vallarta and Acapulco, as documented by the Jilotlán gabbros and the Petatlán-Papanoa ultramafic rocks. This scenario fits well with the contemporaneous extensional arc documented in the Zihuatanejo terrane (Centeno-García et al., 2011) where mafic magmas were emplaced in a magmatic arc setting due to thinned crust and deep faults. The central Jilotlán plagiogranites (~70 Ma) are younger than the PRB gabbros and tonalitic-trondhjemitic granitoids (>92 Ma) and form part of the compressional arc documented in the Zihuatanejo Terrane. In this context, the scenario is similar to the eastern PRB, where the gabbros could have thickened the crust in addition to regional compression, constituting the source from which plagiogranites were generated.
This study supports the hypothesis that a long-lived subduction zone (at least since the Early Cretaceous) is responsible for the formation of continental arc rocks from very primitive sources. The primitive rocks in the Jilotlán area are the result of mantle magmas formed above a subduction zone and mixed with geochemically primitive oceanic type basement rocks. They were emplaced in a thinned continental crust due to extensional tectonics. This has been geochemically documented for the Zihuatanejo Terrane, part of the Guerrero Terrane.
CONCLUSIONS
1. The Jilotlán gabbros are part of the Guerrero Terrane, with primitive geochemical characters in the study area, and differ substantially from the to the NW neighboring Puerto Vallarta intrusives. Based on the available studies, the Jilotlán gabbros are part of the Barremian- Albian Zihuatanejo extensional arc, whereas the granitoids of Jilotlán would be part of the Santonian-Maastrichtian compressional arc, both generated by variations in the slab geometries of a single subduction zone in a continental arc setting. These tectonic scenarios could explain the geochemical features distinguishing the gabbros from the granitoids by heterogeneities in crustal thickness.
2. Trace element distributions and Sr-Nd isotopic ratios of the gabbros show that they are part of a magmatic arc generated from a homogeneous source, most probably essentially by partial melting of mantle material enriched by sediments/fluids derived from the slab with some participation of primitive crustal components. The geochemical variations could be explained by different degrees of fractional crystallization and less probably by different degrees of partial melting of this primitive source.
3. A magmatic hiatus of ~40 Ma separates the Jilotlán gabbros (114 Ma) from the intruding granodiorites and tonalites (68 Ma), which confirms that these units are not cogenetic.
4. The gabbros could be the roots of the host rock Tecalitlán Formation, emphasized by within error identical ages (~114 Ma and ~118 Ma, respectively).
5. The gabbros crystallized and cooled down very rapidly as there is no substantial differences between the Ar-Ar hornblende and U-Pb zircon ages within the same sample. This evidences a shallow emplacement and/or rapid uplift. As the gabbros were intruded by the granitoids, this implies that latter were also emplaced shallowly or exhumed rapidly.
6. Similarities in age and inferred petrotectonic settings of the Jilotlán gabbros with the gabbros of the Peninsular Ranges batholith and the ultramafic rocks from the Petatlán-Papanoa region could indicate that gabbros and mafic rocks constitute a common feature in extensional continental arc settings related to crustal thinning.
