Abstract

The Oligocene Platoro caldera complex of the San Juan volcanic locus in Colorado (USA) features numerous exposed plutons both within the caldera and outside its margins, enabling investigation of the timing and evolution of postcaldera magmatism. Intrusion whole-rock geochemistry and phenocryst and/or mineral trace element compositions coupled with new zircon U-Pb geo-chronology and zircon in situ Lu-Hf isotopes document distinct pulses of magma from beneath the caldera complex. Fourteen intrusions, the Chiquito Peak Tuff, and the dacite of Fisher Gulch were dated, showing intrusive magmatism began after the 28.8 Ma eruption of the Chiquito Peak Tuff and continued to 24 Ma. Additionally, magmatic-hydrothermal mineralization is associated with the intrusive magmatism within and around the margins of the Platoro caldera complex.

After caldera collapse, three plutons were emplaced within the subsided block between ca. 28.8 and 28.6 Ma. These have broadly similar modal miner-alogy and whole-rock geochemistry. Despite close temporal relations between the tuff and the intrusions, mineral textures and compositions indicate that the larger two intracaldera intrusions are discrete later pulses of magma. Intrusions outside the caldera are younger, ca. 28–26.3 Ma, and smaller in exposed area. They contain abundant glomerocrysts and show evidence of open-system processes such as magma mixing and crystal entrainment. The protracted magmatic history at the Platoro caldera complex documents the diversity of the multiple discrete magma pulses needed to generate large composite volcanic fields.

INTRODUCTION

Constraining conditions of magmatic storage and differentiation processes that resulted in either eruption or pluton emplacement can help constrain the larger architecture of magmatic systems (e.g., Buddington, 1959; Hamilton and Myers, 1967, 1974; Smith, 1979; de Silva, 1991; Lipman, 1984, 2007; Cashman and Giordano, 2014). However, understanding the waning stage of a magmatic system, after large-volume caldera-forming eruptions, commonly poses significant challenges because the temporally, spatially, and/or petrogenetically related plutonic and volcanic components may not be exposed or are difficult to discern. Shallow postcaldera intrusions are either later pulses of magmas or residue left behind during an eruption (e.g., Bachmann et al., 2007; Mills and Coleman, 2013; Bacon et al., 2014; Bachmann and Huber, 2016; Watts et al., 2016). Both can be possibilities in any given magmatic system (e.g., Tappa et al., 2011; Zimmerer and McIntosh, 2012; Colgan et al., 2018). Postcaldera intrusions vary in geometry from ring dikes to laccoliths to larger plutons and are commonly associated with mineralization and hydrothermal alteration (e.g., Smith, 1960; Lipman, 1984; Kennedy et al., 2012; Zimmerer and McIntosh, 2012; Colgan et al., 2018; Tomek et al., 2017).

Several Oligocene calderas of the Southern Rocky Mountain volcanic field (SRMVF) (Fig. 1) include exposed intrusions inferred to reflect the subvolcanic plutonic parts of caldera-forming magmatic systems (Lipman, 2007), offering excellent opportunities to track the evolution of a magmatic system to its endpoint. For example, the majority of plutons in the Questa-Latir volcanic locus (New Mexico, USA) postdate the caldera-forming Amalia Tuff, with only a modest component of unerupted residual magma (Lipman, 1988; Johnson et al., 1989; Tappa et al., 2011; Zimmerer and McIntosh, 2012). Likewise, in the northeastern part of the SRMVF, the Mount Princeton pluton is interpreted to reflect a period of low magma flux that occurred substantially after the ignimbrite eruption (Mills and Coleman, 2013). Elsewhere in the SRMVF, the ca. 23 Ma Lake City caldera contains postcaldera syenite intrusions that represent mushy magma that did not erupt, while monzonite intrusions represent chemically distinct, less-evolved magma emplaced after caldera collapse (Hon, 1987; Kennedy et al., 2012, 2016).

The Platoro caldera complex of the San Juan volcanic locus (SJVL) in the southeastern San Juan Mountains erupted five large-volume, crystal-rich dacite ignimbrites over 1.5 m.y., a period considerably longer than that associated with most other multicycle calderas in the SJVL. Postcaldera magmatism is preserved in compositionally diverse volcanic deposits ranging from andesite to dacite and as hypabyssal intrusions of diorite to quartz monzonite. Post-caldera intrusions occur both within the caldera, including one inferred to have caused late-stage caldera resurgence, and external to the caldera margins. Previous geochronological studies focused primarily on eruptive ages of volcanic deposits (Lipman, et al., 1970, 1996; Lipman and Zimmerer, 2019). Emplacement ages for intrusions were inferred mainly from field observations of crosscutting relations and stratigraphic position (Lipman, 1974, 1975), limiting rigorous determination of temporal and magmatic associations between postcaldera intrusions and the caldera-forming magmatic system. However, exceptional exposures and documented field relations of postcaldera volcanic and intrusive rocks (Lipman, 1974) enable geochronologic determination of intrusion ages and assessment of their petrologic affinities to eruptions from the Platoro caldera complex as monitors of the evolution of that magmatic system from explosive, silicic ignimbrite eruption to intrusion emplacement.

In this study, we examine the temporal and spatial constraints on the emplacement history of postcaldera intrusions at the Platoro caldera complex and assess the genetic relationship between the postcaldera intrusions, caldera-filling lavas, and the Chiquito Peak Tuff, the last major ignimbrite sourced from the Platoro caldera complex. Using new zircon U-Pb sensitive high-resolution ion microprobe (SHRIMP) dates and trace element data in combination with other mineral compositions, we address the timing of post-collapse pluton emplacement at Platoro in relation to compositions of the ignimbrite magma, later effusive volcanism, magma generation, and attendant shallow pluton assembly.

GEOLOGICAL SETTING

The SRMVF (Fig. 1A) is one of North America's largest mid-Cenozoic volcanic fields. As part of the larger Cordilleran ignimbrite flareup, the SRMVF is an eastern manifestation of volcanism related to subduction of the Farallon plate beneath North America. Farallon slab removal in the mid-Cenozoic likely caused asthenospheric mantle melting, resulting in the ignimbrite flareup (Coney and Reynolds, 1977; Humphreys et al., 2003; Farmer et al., 2008). The composite SRMVF sourced 26 large-volume (>100 km3) ignimbrites from multiple calderas between 37 and 23 Ma, depositing >17,000 km3 of crystal-rich, high-silica dacite and crystal-poor, low-silica rhyolite across a region of ~100,000 km2 (Lipman, 2007; Lipman and Bachmann, 2015). Estimates of total erupted volume for the SRMVF, including precursor lava flows, are approximately a factor of four greater than that of the ignimbrites, and the dominantly silicic batholithic roots of the volcanic field may constitute as much as an additional 300,000 km3 (Farmer et al., 2008; Lipman and Bachmann, 2015). Current crustal thickness in the region is estimated to vary from ~41 to 49 km (Hansen et al., 2013).

Stratigraphic and structural exposure levels within individual calderas are variable across the SRMVF owing to diverse volcanogenic controls, including syncollapse structural juxtaposition of intracaldera and extracaldera deposits, postcollapse magmatic resurgence, and presence or absence of postcaldera volcanism. More importantly, regional uplift and extension associated with the northern Rio Grande rift, and opening of the Upper Arkansas and San Luis valleys, have resulted in extensive dip-slip and oblique-slip faulting, structurally exposing deeper levels of magmatic systems adjacent to basin-bounding systems of north- and northwest-trending faults. Coupled with extensive flu-vial dissection driven by tectonic and climate influences, exposures vary from dominantly plutonic remnants of volcanic loci, such as exposed in the Sawatch Range north of the SJVL, to surface preservation of primary caldera morphology, as in the minimally dissected Creede and Cochetopa Park calderas of the central SJVL (Fig. 1A). Most calderas of the SJVL are enclosed by a large negative Bouguer gravity anomaly, interpreted as the geophysical expression of a composite upper-crustal batholith (Plouff and Pakiser, 1972; Steven and Lipman, 1976; Drenth et al., 2012). The Platoro caldera complex lies along the southeastern margin of the gravity low.

The products of postcaldera magmatism in the SJVL include intermediate lavas, andesitic to dacitic dikes, and granodiorite to granite plutons (Lipman, 2007). The subvolcanic intrusions (Fig. 1) are typically more mafic than the associated ignimbrites but overlap the compositions of the more primitive parts of some compositionally zoned ignimbrites (e.g., San Luis caldera–Nelson Mountain Tuff) (Lipman, 2007). Among the few published intrusion ages, some are within error of the ages of the associated ignimbrites; others are demonstrably younger. Intrusion textures range from equigranular and nearly aphanitic to porphyritic, and many of the intrusions have been interpreted as the cores of volcanic edifices. Within the SJVL, small- to moderate-scale hypabyssal plutons are exposed at the Bonanza, South River, San Luis, Silverton, Lake City, Uncompahgre–San Juan, San Juan, and Platoro calderas (Steven and Lipman, 1976; Lipman, 2007; Lipman and Bachmann, 2015). Published intrusion ages are limited, and only one caldera intrusion, the Sultan Mountain pluton, has a published U-Pb zircon date (Gonzales, 2015).

Precaldera rocks in the Platoro area (Figs. 1, 2), as well as generally in the San Juan Mountains, are andesitic to dacitic lavas and volcaniclastic deposits of the Conejos Formation (Lipman, 1975; Colucci et al., 1991). These deposits, volumetrically the largest component of the SJVL, were erupted from multiple centers between 35 and 30 Ma (Lipman et al., 1970). Eruptive centers for Conejos lavas in the Platoro area were interpreted as being located mainly within the area of subsequent caldera subsidence; flanks of several stratovolcanoes are preserved along caldera rims (Lipman, 1975).

The Platoro caldera complex sourced seven named ignimbrite units of the Treasure Mountain Group between 30.1 and 28.8 Ma. These ignimbrites are intercalated with andesitic lava eruptions and predate intrusion of monzonites (Fig. 2) (Lipman, 1975; Lipman et al., 1996; Tomek et al., 2019). The five largest ignimbrites associated with the Platoro caldera complex, the Black Mountain, La Jara Canyon, Ojito Creek, Ra Jadero, and Chiquito Peak Tuffs, have volumes between 100 and 1000 km3. Ignimbrites of the Treasure Mountain Group consist dominantly of compositionally crystal-rich dacites (Lipman et al., 1996).

The Black Mountain Tuff, a densely welded dacite tuff with a volume of ~300–400 km3, has an 40Ar/39Ar hornblende date of 30.19 ± 0.16 Ma (Lipman and Zimmerer, 2019). Dates for underlying and overlying units indicate that the next ignimbrite of the Treasure Mountain Group, the La Jara Canyon Tuff, erupted between 30.1 and 29.9 Ma (Lipman et al., 1996; Lipman and Zimmerer, 2019). This crystal-rich dacitic ignimbrite, which contains plagioclase, biotite, and augite phenocrysts, is widespread, thick, and has a volume >1000 km3 (Lipman, 1975). Subsequent to eruption of the La Jara Canyon Tuff, the resulting collapsed caldera was partially filled by andesite lavas of the lower member of the Summitville Andesite (Lipman, 1974; Lipman et al., 1996). The next erupted unit, the middle tuff, consists of 10–15 separate, small-volume ignimbrite sheets. The Ojito Creek, Ra Jadero, and South Fork Tuffs are also widespread dacitic ignimbrites that may have been erupted from the Summitville caldera, a tenuously postulated structure within the larger Platoro caldera complex (Fig. 1) (Lipman, 1975). Sanidine from the Ra Jadero and South Fork Tuffs yielded 40Ar/39Ar dates of 29.12 ± 0.07 Ma and 28.86 ± 0.14 Ma, respectively (Lipman and Zimmerer, 2019).

The outflow sheet of the last major ignimbrite erupted from the Platoro caldera complex, the Chiquito Peak Tuff, was originally mapped as parts of the Masonic Park and La Jara Canyon Tuffs (Lipman, 1974) but was later recognized as a separate ignimbrite based on the presence of sparse sanidine, differences in rock and mineral chemistry, and paleomagnetic data (Lipman et al., 1996). Sanidine for the Chiquito Peak Tuff yielded an 40Ar/39Ar date of 28.77 ± 0.03 Ma (Lipman and Zimmerer, 2019). The tuff is a lithic-rich, crystal-rich dacite with plagioclase, biotite, augite, and sanidine phenocrysts. Lipman et al. (1996) concluded that eruption of the Chiquito Peak Tuff created the majority of exposed caldera features, including the thick ponding of intracaldera ignimbrite (Fig. 1). The total estimated volume of ignimbrites erupted from the Platoro caldera complex is >2600 km3 (Lipman et al., 1996).

Andesitic lavas of the upper Summitville Andesite (Lipman, 1974), locally >500 m thick, filled the Platoro caldera after the eruption of the Chiquito Peak Tuff (Figs. 1, 2). Stratigraphically overlying these andesites are dacite and minor rhyolite lavas of Park Creek, Green Ridge, Silver Mountain, South Mountain, and Cropsy Mountain. The andesites and dacites at Green Ridge and Silver Mountain (Fig. 2) have been interpreted as eroded flanks of a stratovolcano on the eastern margin of the Platoro caldera complex in the Cat Creek area (Lipman, 1974, 1975).

Intrusions associated with the Platoro caldera complex (Table 1) include quartz monzonite to diorite plutons and stocks as well as andesitic to dacitic dikes (Patton, 1917; Larsen and Cross, 1956; Lipman, 1974). These intrude volcanic and volcaniclastic deposits of the Conejos Formation, ignimbrites of the Treasure Mountain Group, and the Summitville Andesite. In general, the porphyritic intrusions crosscut the equigranular monzonites (Larsen and Cross, 1956; Lipman, 1974). With the exception of a K-Ar date of 29.1 ± 1.2 Ma on biotite from the Alamosa River monzonite (Lipman et al., 1970), there have been no prior radiometric dates on these plutons.

Mineralization in the Summitville, Platoro, Stunner, and Jasper districts has been interpreted as being related to postcaldera structures and postcaldera magmatic activity of the Platoro caldera complex (Patton, 1917; Steven and Ratté, 1960; Mehnert et al., 1973; Lipman, 1975; Neuerburg, 1978; Bethke, 2011). The largest deposit, the Summitville Au-Ag-Cu deposit, is a 22.5 ± 0.5 Ma high-sulfidation, epithermal deposit associated with the quartz dacite volcanic dome at South Mountain and underlying quartz monzonite porphyry intrusion (Bethke et al., 2005). The deposit is situated at the intersection of the caldera margin and a northwest-trending fault zone that extends from southeast of Platoro village northwest to Wolf Creek Pass. Largely subeconomic mineralization in the other districts has been minimally studied, and relationships to postcaldera magmatic activity are not well established; however, mineralization in the Jasper district occurs within a postcaldera pluton at the intersection of two caldera-related faults (Lipman, 1974, 1975). Molybdenite vein stockwork and copper-lead-zinc vein mineralization (Neuerburg, 1978) and alteration in the Crater Creek area are localized around plutons and along northwest- trending faults and dikes.

INTRUSIONS, CHIQUITO PEAK TUFF, AND POSTCALDERA LAVAS

Plutons (Table 1) associated with the postcaldera magmatism at the Platoro caldera complex range from diorite to quartz monzonite and are locally intruded by small aplitic veins (Figs. 3, 4A). The dikes define a similar compositional range as the plutons but are texturally more variable, commonly porphyritic (Fig. 4B), and contain mafic enclaves locally (Fig. 4C). We distinguish two groups of the intrusions: (1) those that are within and (2) those that are outside the Platoro caldera complex. Plutons within the caldera (Fig. 2) were emplaced along structures related to subsidence and/or resurgence (Lipman, 1975). Plutons outside the Platoro caldera are less related to specific structures, but many of the dikes are radial to the western caldera boundary. In the Cat Creek area, east of the caldera, dikes are approximately radial to the Cat Creek pluton. Field, age, and compositional relations suggest that the postcaldera lavas, the dacite of Fisher Gulch and the Summitville Andesite, are also genetically associated with Platoro caldera magmatism.

Intracaldera Plutons

Alamosa River Monzonite

The Alamosa River pluton, the largest postcaldera intrusion (~3 × 7 km), was emplaced into Summitville Andesite (upper member) and the Chiquito Peak Tuff (Fig. 4D) at the northwestern margin of the hinged Cornwall Mountain block, which has been interpreted as a resurgent structure within the Platoro caldera complex. Lipman (1975) suggested that the Alamosa River pluton is the intrusive core of the volcano that sourced the compositionally similar Summitville Andesite. Magnetic fabrics in this intrusion suggest pulsed magma emplacement of a vertically extensive pluton (Tomek et al., 2019).

The pluton ranges in composition from monzonite to local quartz monzonite, 59–64 wt% SiO2 (Table 2; Table S1 in the Supplemental Material1; Fig. 3) and is equigranular fine to medium grained and locally porphyritic (Fig. 4A). The equigranular and porphyritic phases of the Alamosa River monzonite contain plagioclase, augite, orthoclase, quartz, biotite, Fe-Ti oxides, and minor orthopyroxene and accessory titanite, apatite, and zircon (Table 1; Fig. 5A). Textural relationships suggest co-crystallization of plagioclase, augite, and Fe-Ti oxides, given that these commonly occur in clusters. These mineral phases all contain apatite inclusions, suggesting apatite saturation near the liquidus. Zircon occurs both as inclusions in biotite and more commonly in interstitial quartz and orthoclase. All equigranular and several porphyritic samples include granophyric intergrowths of quartz and alkali feldspar (Fig. 5B).

Alteration within the intrusion is localized along the northwestern margin of the intrusion and is most pervasive near the contact with the Alum Creek porphyry. The dominant alteration is propylitic; chlorite and actinolite replace biotite, augite, and orthopyroxene, but localized advanced argillic alteration was also observed.

Alum Creek Porphyry

The Alum Creek porphyry intrudes the Alamosa River monzonite north of the Alamosa River (Fig. 2) and consists of fine- to medium-grained porphyritic to coarse-grained monzonite to quartz monzonite. Analyzed and dated samples are from an outcrop and a drill hole along Alum Creek. The Alum Creek porphyry ranges from 58 to 65 wt% SiO2 (Table 2; Table S1 [footnote 1]; Fig. 3). Groundmass quartz and orthoclase enclose phenocrysts of plagioclase, biotite, and augite (Table 1; Fig. 5C). Accessory minerals include Fe-Ti oxides, titanite, apatite, and zircon.

Lipman (1974, 1975) previously considered this intrusion to be a porphyritic phase of the Alamosa River pluton, but crosscutting contacts between the Alum Creek porphyry and Alamosa River monzonite are sharp, and the Alum Creek porphyry contains inclusions of equigranular Alamosa River monzonite (Calkin, 1967, 1971; Tomek et al., 2019).

Along Alum Creek, quartz-sericite-pyrite alteration is dominant but transitions to pervasive argillic alteration in places. In drill-core samples from the Alum Creek porphyry, plagioclase phenocrysts are variably sericitized and chlorite and actinolite replace biotite and augite. Locally, disseminated pyrite is spatially associated with Fe-Ti oxides, and calcite is disseminated in veinlets.

Jasper Monzonite

The Jasper monzonite intrudes the Summitville Andesite along Burnt and Jasper Creeks and along the Alamosa River (Fig. 2). The texture of the Jasper monzonite is heterogeneous, ranging from porphyritic to fine- to medium- grained equigranular. Phenocrysts consist of plagioclase, augite, and biotite, within groundmass orthoclase, Fe-Ti oxides, and quartz (Table 1; Fig. 5D). The SiO2 content ranges from 62 to 63 wt% (Table 2; Table S1 [footnote 1]; Fig. 3). Lipman (1975) suggested that much of the exposed Jasper intrusion is fine grained near its margins, but pervasive hydrothermal argillic alteration has obscured textures within much of this intrusion. Where less altered, mainly in its interior, the Jasper intrusion is chemically, mineralogically, and texturally similar to Alamosa River monzonite that crops out ~5.5 km upstream. Although the intrusion was previously undated, Lipman (1975) concluded that the Jasper monzonite might be no younger than 28 Ma based on alteration and crosscutting relations between the adjacent Summitville Andesite, which it intrudes, and the overlying Green Ridge andesite lavas.

Cornwall Mountain Quartz Monzonite Porphyry

The Cornwall Mountain quartz monzonite porphyry (Fig. 2) intrudes intracaldera Chiquito Peak Tuff within the resurgent block of the Platoro caldera (Lipman, 1974, 1975). It contains 65–66 wt% SiO2 (Table 2; Table S1 [footnote 1]; Fig. 3) and has 35%–40% phenocrysts of plagioclase, augite, and biotite in a groundmass of quartz, plagioclase, and orthoclase (Table 1; Fig. 5E). Accessory minerals include Fe-Ti oxides, titanite, apatite, and zircon. Most of the porphyry is propylitically altered, with biotite and augite replaced by chlorite and carbonate. Plagioclase phenocrysts are replaced by sericite.

Summitville Quartz Monzonite

The Summitville quartz monzonite does not crop out but was sampled from drill core. This intrusion is spatially associated with the Summitville epithermal Au-Ag deposit and mine (Fig. 2). Sample U449, from a depth of 1260–1263 m, is equigranular, contains extensive quartz-sericite-pyrite hydro-thermal alteration, and is cut by quartz-pyrite veins. Primary minerals include plagioclase, quartz, alkali feldspar, and minor biotite (Table 1). Bethke et al. (2005) suggested that mineralization at Summitville is genetically related to this intrusion. Although the Summitville quartz monzonite has not previously been dated, sanidine and biotite from the mineralization host, the dacite of South Mountain, yielded 40Ar/39Ar biotite and sanidine dates of 23.0 ± 0.1 Ma and 23.1 ± 0.1 Ma, respectively (Getahun, 1994), indistinguishable from sanidine, biotite, and alunite-alteration ages determined using the K-Ar method (Mehnert et al., 1973). Additionally, Getahun (1994) reported total-gas and plateau 40Ar/39Ar dates on sericite associated with wall-rock alteration of 22.5 ± 0.1 Ma and 22.6 ± 0.6 Ma, respectively.

Extracaldera Plutons

Cat Creek Monzonite

The Cat Creek monzonite intrusion, east of the Platoro caldera complex (Fig. 2), has been interpreted as the core of a volcano that sourced the volcanics of Green Ridge (Lipman, 1975). This unit intrudes lavas and volcaniclastic deposits of the Oligocene Conejos Formation and consists of equigranular and porphyritic phases. The fine-grained monzonite equigranular phase is 61 wt% SiO2, whereas the porphyritic phase is quartz monzonite, ranging from 66 to 67 wt% SiO2 (Table 2; Table S1 [footnote 1]; Fig. 3). Both phases contain plagioclase, orthoclase, augite, and biotite; the equigranular phase contains minor quartz, and also rare orthopyroxene largely replaced by chlorite (Table 1). Accessory phases include Fe-Ti oxides, titanite, apatite, and zircon. Alteration is pervasive in the porphyritic phase, with extensive sericite overprinting of plagioclase, zones of intense silicification, and local replacement of augite by calcite and chlorite.

Lake Annella Andesite Porphyry

The Lake Annella andesite porphyry is a small elongate east-west–trending stock that intrudes the Conejos Formation and the lower Summitville Andesite along the western margin of the caldera (Fig. 2). It ranges from 57 to 61 wt% SiO2 (Table 2; Table S1 [footnote 1]; Fig. 3) and contains plagioclase, amphibole, and biotite phenocrysts in a finer-grained matrix of plagioclase, orthoclase, biotite, quartz, and magnetite (Table 1; Fig. 5F). Accessory minerals include Fe-Ti oxides, titanite, apatite, and zircon. This intrusion, unlike most of the Platoro intrusions, lacks augite. The Lake Annella andesite porphyry contains numerous mafic enclaves that are as much as 10 cm long, contain rare plagioclase phenocrysts, and have a finer-grained groundmass than the host.

The Lake Annella andesite porphyry is propylitized locally; actinolite, epidote, and calcite replace biotite and amphibole. The intrusion was drilled during exploration for porphyry Cu-Au mineralization. One of the resulting drill holes, sampled during this study, intercepted zones of quartz-sericite-pyrite alteration.

Plutons of the Crater Creek Area

The Bear Creek, Crater Creek, Elwood Creek, and Cataract Creek intrusions, ~5 km west of the Platoro caldera margin (Fig. 2), range from diorite to monzonite and contain 56–63 wt% SiO2 (Table 2; Table S1 [footnote 1]; Fig. 3). Textures of these intrusions vary from coarsely porphyritic to coarse- to fine-grained equigranular. Accessory minerals in these intrusions include Fe-Ti oxides, titanite, apatite, and zircon.

The Bear Creek monzonite, the largest of the intrusions in the Crater Creek area, intrudes lava and volcaniclastic deposits of the Conejos Formation and the La Jara Canyon Tuff. The Bear Creek intrusion includes both porphyritic and equigranular phases. The porphyritic phase contains phenocrysts of plagioclase, amphibole, and augite in a groundmass of quartz, plagioclase, orthoclase, and minor biotite and orthopyroxene (Table 1). The equigranular phase has a similar composition but contains less amphibole (Fig. 5G). No exposed contact between the two phases was observed.

The Crater Creek and Elwood Creek monzonites intrude the Conejos Formation, Summitville Andesite, and La Jara Canyon Tuff. The Elwood Creek monzonite is coarsely porphyritic and contains phenocrysts of plagioclase, augite, and biotite (Table 1, Fig. 5H). The Cataract Creek monzonite is only ~0.5 km across and is the smallest of the Crater Creek plutons. It intrudes the La Jara Canyon Tuff, and intermediate composition dikes intrude the monzonite. Though dominantly porphyritic, it is equigranular and fine grained in places; phenocrysts include plagioclase, hornblende, biotite, and augite (Table 1, Fig. 5I).

Only the Elwood Creek intrusion has been previously dated. However, its biotite 40Ar/39Ar inverse isochron date of 26.61 ± 0.01 Ma has a mean square weighted deviation (MSWD) value of 8.76 (Lipman and Zimmerer, 2019), which suggests that this is not a statistically robust date. The large MSWD suggests an open system or substantial underestimation of the error associated with the determined date.

Patton (1917), Steven and Ratté (1960), Neuerburg et al. (1978), and Neuerburg (1978) documented porphyry-epithermal–style mineralization in the Crater Creek area. Alteration and mineralization are localized, and all the intrusions have unaltered domains.

Dikes

Andesite Dikes

Andesite dikes are widely radial to the western and southwestern margins of the caldera and locally to the Cat Creek monzonite. They predominantly intrude Conejos Formation lavas and volcanoclastic deposits, but some also cut lavas of the Summitville Andesite. Widths of these dikes range from ~0.5 m to 1.5 m. These dikes have equigranular to porphyritic textures and contain phenocrysts of plagioclase and augite or plagioclase and hornblende (Table 1). Some dikes have minor orthopyroxene. Fe-Ti oxides and apatite are the principal accessory phases.

The andesite dikes proximal to Platoro together with more distal trachy-basalt dikes have been described as the Platoro-Dulce dike swarm. 40Ar/39Ar dates of the andesite dikes range from 27.41 ± 0.03 Ma to 31.27 ± 0.05 Ma; consequently, they both predate and postdate caldera formation (Lipman and Zimmerer, 2019). The hornblende-bearing dikes and dikes with large tabular plagioclase phenocrysts are likely related to the precaldera Conejos Formation, whereas at least many of the equigranular, pyroxene-bearing dikes postdate caldera formation (Lipman, 1975; Lipman and Zimmerer, 2019). SiO2 is 61.8–62.2 wt% in three fine-grained, plagioclase-pyroxene andesite dikes (Table 2; Table S1[footnote 1]; Fig. 3) that were emplaced during or after caldera activity, one intruding the Ra Jadero Tuff and the other intruding the Summitville Andesite.

Hornblende Andesite-Dacite Dikes

Hornblende andesite-dacite dikes are radial to the Platoro caldera complex on its western side but also are exposed adjacent to the Cat Creek monzonite to the east of the caldera. Two hornblende dacite dikes were sampled within the caldera along Jasper and Ranger Creeks where they crosscut Summitville and Sheep Mountain Andesites, respectively. Another hornblende dacite dike was sampled just west of the caldera in Horsethief Park. Hornblende andesite-dacite dike widths range from a meter to 30 m wide, and individual dikes may extend for several kilometers, particularly on the western side of the caldera.

Compositions of the hornblende andesite-dacite dikes range from andesite to dacite, contain 60–68 wt% SiO2 (Table 2; Table S1[footnote 1]; Fig. 3), and are all porphyritic in texture. Plagioclase, amphibole, and biotite are the dominant phenocrysts in these dikes; sanidine and quartz are absent (Table 1). Augite is present only as phenocryst and/or glomerocryst phase in the intracaldera dike at Jasper Creek (Fig. 5J). Phenocryst abundance and size vary among dikes (Figs. 4B–4C) and also may vary considerably along strike within any particular dike. Plagioclase phenocrysts are as large as 3 cm. Many hornblende andes-ite-dacite dikes contain plagioclase-biotite-magnetite ± amphibole ± augite glomerocrysts. These crystal clusters may be as large as 0.5 cm. Accessory phases include Fe-Ti oxides, titanite, apatite, and zircon. The texture and composition of these dikes, particularly those west of the Platoro caldera complex, are similar to those of the Lake Annella andesite porphyry. A hornblende andes-ite-dacite dike sampled in Horsethief Park contains numerous mafic enclaves like those within the Lake Annella andesite porphyry. Enclaves are as much as 10 cm across. The enclaves have a finer-grained groundmass than their host and contain rare plagioclase phenocrysts (Fig. 4C). Enclave margins are commonly irregular, and some plagioclase phenocrysts straddle the boundary between the enclave and host porphyry.

Sanidine Dacite Dikes

Sanidine-quartz rhyolite to silicic dacite porphyritic dikes crop out in and around the Alamosa River monzonite and the Alum Creek monzonite porphyry but are most common west of the caldera on its western side near Crater and Elwood Creeks. These dikes intrude the Conejos Formation, Summit-ville Andesite, Alamosa River monzonite, dacite of Park Creek, Fish Canyon Tuff, Alum Creek monzonite porphyry, and Lake Annella andesite porphyry (Lipman, 1974), range from 10 to 100 m thick, and in some cases extend for several kilometers. These dikes range from 68 to 70 wt% SiO2 (Table 2; Table S1[footnote 1]; Fig. 3). The dikes are coarsely porphyritic with sanidine phenocrysts as large as 3 cm and also contain plagioclase, biotite, quartz, and hornblende phenocrysts (Table 1; Fig. 5K). Quartz phenocrysts are rounded and resorbed. Accessory minerals include Fe-Ti oxides, titanite, apatite, and zircon. Sanidine from a sanidine dacite dike near Schinzel Meadows yielded an 40Ar/39Ar date of 26.25 ± 0.04 Ma (Lipman and Zimmerer, 2019).

Chiquito Peak Tuff

Although the Platoro caldera complex sourced multiple major ignimbrites, we compare the Chiquito Peak Tuff with the postcaldera intrusions because this tuff was the last major ignimbrite and its eruption was responsible for most exposed deposits and structures of the caldera complex. This ignimbrite, having an estimated volume of ~1000 km3, is one of the most voluminous of the Platoro-associated tuffs (Lipman et al., 1996). The Chiquito Peak Tuff is a crystal-rich dacite that contains 63–65 wt% SiO2 (Table 2; Table S1[footnote 1]; Fig. 3) and as much as 45% phenocrysts as large as 2–3 mm, where unbroken, of plagioclase, biotite, augite, and minor sanidine (Table 1; Fig. 5L). Bulk chemistry for this ignimbrite suggests it is unzoned, most variation resulting from variable winnowing of ash during eruption (Lipman et al., 1996). It is locally rich in lithic clasts, mainly andesite, as much as 3 cm in size. Accessory minerals include Fe-Ti oxides, titanite, zircon, and apatite. Intracaldera Chiquito Peak Tuff is densely welded, >800 m thick without any exposed base, and widely propylitically altered; a vitrophyre is exposed only at one small site high along the southwestern caldera margin. Outflow Chiquito Peak Tuff is less welded and has a maximum thickness of ~100 m (Lipman et al., 1996). Intracaldera Chiquito Peak Tuff contains dense fiamme, whereas pumice blocks in outflow tuff are vesicular and as much as 15 cm across. Sanidine from the Chiquito Peak Tuff yielded an 40Ar/39Ar date of Ma 28.77 ± 0.03 Ma (Lipman and Zimmerman, 2019).

Dacite of Fisher Gulch

The dacite of Fisher Gulch is a 350-m-thick lava dome at the southeastern boundary of the Platoro caldera complex (Fig. 2). It overlies the intracaldera Chiquito Peak Tuff, banks against lavas of the Conejos Formation on the caldera wall, and is in turn overlain by Summitville Andesite (Lipman, 1974). The dacite contains 63–64 wt% SiO2 (Table 2; Table S1[footnote 1]; Fig. 3) and as much as 20% phenocrysts, including plagioclase, biotite, augite, and minor sanidine (Table 1; Lipman, 1975). Accessory minerals include Fe-Ti oxides, apatite, and zircon. It is similar in chemistry and petrography to the immediately underlying intracaldera ignimbrite (Tables 12; Table S1), as described previously (Lipman, 1975). Sanidine from the dacite of Fisher Gulch yielded an age of 28.74 ± 0.09 Ma, analytically indistinguishable from that of the Chiquito Peak Tuff (Lipman et al., 1996; as recalculated to Fish Canyon Tuff standard age of 28.2 Ma by Lipman and Zimmerer [2019]).

Summitville Andesite

Lavas of Summitville Andesite that fill the Platoro caldera are >500 m thick in places (Lipman, 1974, 1975). These lavas comprise two members intercalated within ignimbrites of the Treasure Mountain Group. The lower member consists of lavas and volcaniclastic sedimentary rocks deposited after the eruption of the La Jara Canyon Tuff, now exposed only along small sectors at the eastern and western margins of the caldera complex (Fig. 2); the more widespread upper member consists of lavas emplaced following the eruption of the Chiquito Peak Tuff. Summitville Andesite ranges from 57 to 60 wt% SiO2 (Table 2; Table S1[footnote 1]; Fig. 3) and is aphanitic to porphyritic with as much as 20% phenocrysts, including plagioclase, augite, and orthopyroxene (Table 1). Fe-Ti oxide clots and apatite needles are also present. Lipman (1975) interpreted the lavas to represent the lower flanks of an intracaldera strato-volcano cored by the Alamosa River monzonite.

ANALYTICAL METHODS

Whole-Rock Geochemistry

Whole-rock major and trace element analyses were determined by X-ray fluorescence and inductively coupled–plasma mass spectrometry (ICP-MS) for 49 representative samples of intrusions and volcanics from the Platoro caldera complex at AGAT Laboratories in Mississauga, Ontario, Canada (https://www.usgs.gov/media/files/contract-chemistry-method-summaries), or at the Washington State University GeoAnalytical Laboratory (samples whose names begin with SRM) in Pullman, Washington, USA, following the methods of Johnson et al. (1999). Results are summarized in Table 2 and Table S1 (footnote 1).

Scanning Electron Microscope and Electron Probe Micro-Analysis

Representative major and trace element compositions were determined for plagioclase, sanidine, biotite, augite, amphibole, and Fe-Ti oxides using a JEOL JXA-8230 electron microprobe at the University of Colorado Electron Micro-probe Laboratory (Boulder, Colorado) and at the Louisiana State University Shared Instrumentation Facility (Baton Rouge, Louisiana, USA). Prior to analysis, backscattered electron images were obtained to evaluate compositional zoning of major phases using the JEOL 5800LV scanning electron microscope (SEM). For electron probe microanalyzer analyses, a 20 kV, 10 nA beam setup with a 1 µm spot size was used to measure Si, Al, Ca, Na, K, Fe, Ti, Mg, Mn, and Cl abundances in the crystal phases (Tables S2–S4[footnote 1]). Counting times ranged from 15 to 60 s, with Na measured first and with a shorter count time to minimize Na mobility. Additionally, Ba abundances were determined in most plagioclase and sanidine crystals using a counting time between 60 and 120 s. Diopside and amphibole standards were run three times as unknowns at the beginning and end of each session.

Zircon U-Pb Geochronology and Trace Element Analysis

Zircon U-Pb dating and trace element (Ti, Fe, Y, rare earth elements [REEs], U, and Th) analyses were acquired from zircon separates from 17 samples using the SHRIMP–reverse geometry (SHRIMP-RG) instrument at the Stanford–U.S. Geological Survey SHRIMP-RG lab (Stanford, California, USA). Analyses were made in three sessions over one year and followed the analytical protocol described in Matthews et al. (2015) for combined U-Pb and trace element analyses using an O2 primary beam. Zircons were separated from bulk-rock samples using standard heavy-liquid and magnetic-separation techniques and were handpicked, mounted in epoxy, and polished to expose the interiors of the grains. Grains chosen for analysis were imaged by cathodoluminescence techniques on a JEOL 5600 SEM to assess the complexity of zircon growth histories and to choose analytical spot locations. Analysis spot dimensions were ~20–25 µm × ~2 µm depth. Complete U-Pb results and trace element data are given in Table S5 (footnote 1). The Temora-2 zircon standard was analyzed after every four unknowns. Trace element concentrations were determined using the MAD-559 standard and the reported mass fractions in Coble et al. (2018). Raw data were reduced with Squid 2.51 software (Ludwig, 2009) using the Temora-2 zircon standard (206Pb/238U date = 416.8 Ma; Black et al., 2004) and a Pb/U–UO/U calibration; ratios were derived from weighted averages of within-spot scans. Individual dates and weighted means were calculated using Isoplot 3.76 software (Ludwig, 2012). Reported 206Pb/238U dates were corrected for common Pb using a 207Pb correction (Ludwig, 2012) with 207Pb/206Pb values derived from the Stacey and Kramers (1975) evolution model. All 206Pb/238U zircon dates are reported in millions of years (Ma); 2σ uncertainties are noted in the form of ± x/y, where x is the analytical uncertainty suitable only for intramethod comparison, and y is the total uncertainty, including the decay constant and standard (Jaffey et al., 1971) suitable for intermethod comparison of dates.

To compare U-Pb dates generated in this study with previously published 40Ar/39Ar dates, systematic decay constant uncertainties of both systems and uncertainties in the standards used must be propagated into the calculated dates (e.g., Schoene et al., 2013). Changes in the accepted ages of primary standards can affect the interpreted age of an unknown determined relative to that standard (Mercer and Hodges, 2016). Consequently, it is desirable to recalculate these dates based on the current accepted value for the standard and to report uncertainties that incorporate not only analytical uncertainty but also systematic uncertainties in the decay constants. Unfortunately, given the different vintages of data and laboratories, it was not possible to incorporate the systematic uncertainties in the published 40Ar/39Ar dates, and thus cited 40Ar/39 Ar errors in the paper do not include this external uncertainty, making intramethod comparisons more challenging.

Zircon Lu-Hf Analysis by LA-MC-ICP-MS

Hafnium (Hf) isotopic compositions of zircon grains were determined by laser–ablation multicollector ICP-MS (LA-MC-ICP-MS) using a Nu Plasma II MC-ICP-MS coupled to an ESI NWR 193 excimer laser ablation system in the Mineral Isotope Laser Laboratory at Texas Tech University (Lubbock, Texas, USA). A 30 µm laser spot was focused on each zircon grain directly on top of the U-Pb analysis location, firing at a frequency of 10 Hz with an energy density of 5 J/cm2 for 300 pulses. 171Yb, 172Yb, 173Yb, 174(Hf + Yb), 175Lu, 176(Hf + Yb + Lu), 177Lu, 178Hf, 179Hf, and 180Hf were all measured simultaneously in 10 Faraday collectors. Each analysis consisted of 15 s of gas background measurement followed by 30 s of laser ablation and ending with 15 s of monitoring wash out. Isotope measurements were acquired using a 0.2 s on-peak integration time. Raw data were processed offline using an in-house spreadsheet based on calculations and corrections described by Souders et al. (2013).

For unknown zircon grains, the initial 176Hf/177Hf ratios were calculated using the measured 176Lu/177Hf, the 176Lu decay constant (λ = 1.865 × 10−11/yr) of Söder-lund et al. (2004), and the SHRIMP U-Pb date for each sample. Epsilon Hf (εHf) values were calculated using the present-day 176Hf/177Hf and 176Lu/177Hf values of 0.282785 and 0.0336, respectively for present-day chondritic uniform reservoir (CHUR) (Bouvier et al., 2008). The depleted mantle model of Griffin et al. (2000), modified to the 176Lu decay constant of Söderlund et al. (2004) and present-day CHUR Lu-Hf composition of Bouvier et al. (2008), was used as a reference. This model has a present-day 176Hf/177Hf value of 0.28325 (εHf = +16.4) at 176Lu/177Hf = 0.0388, similar to modern-day mid-ocean-ridge basalt (MORB). The results for all unknown zircon grain analyses and for all reference materials run during the analytical sessions are presented in Table S7.

RESULTS

Zircon U-Pb Dates

Zircons from ten plutons, four dikes, the Chiquito Peak Tuff, and the dacite of Fisher Gulch were dated by U-Pb geochronology (Fig. 6; Table 3). Zircons from the intrusions, the Chiquito Peak Tuff, and the dacite of Fisher Gulch are predominantly euhedral prismatic crystals with the exception of most zircons in the Alamosa River monzonite, which are subhedral to anhedral. Cathodoluminescence images of the zircons in most samples have predominantly simple oscillatory and sector growth zoning patterns, suggestive of a single period of zircon crystallization (Fig. 7).

Emplacement ages are interpreted from the weighted mean of 206Pb/238U dates for these samples (Fig. 6; Table 3; Table S2[footnote 1]). Five samples each contained a single xenocryst (Alamosa River, Jasper, Bear Creek, and Cataract Creek monzonites and the Chiquito Peak Tuff); four Proterozoic (1138–1709 Ma) and one Mesozoic (158 Ma) in age. Data for the xenocrysts are not included in Figure 6 but are presented in Table S2. The 206Pb/238U dates of some samples scatter over several million years, exceeding analytical uncertainty. This spread may be the result of magmatic processes such as protracted crystal growth prior to emplacement or eruption, the inheritance of antecrysts, or possible Pb loss.

The new 206Pb/238U dates indicate periodic emplacement and crystallization of most plutons and dikes from 28.98 to 26.92 Ma. The oldest intrusions are within the mapped Chiquito Peak caldera: the Alamosa River monzonite at (28.98 ± 0.18/0.26 Ma, data included here and in Tomek et al. [2019]), the Cornwall Mountain quartz monzonite porphyry (28.97 ± 0.25/0.28 Ma), and the Jasper monzonite (28.81 ± 0.19/0.19 Ma). The dates for these intrusions are all within error of each other and with the two dated samples of Chiquito Peak Tuff (28.92 ± 0.38/0.50 Ma and 28.31 ± 0.55/0.57 Ma) (Fig. 6; Table 3) and its 40Ar/39Ar date (28.77 ± 0.03 Ma; Lipman and Zimmerer, 2019). They are also within error of the 206Pb/238U and 40Ar/39Ar dates for the dacite of Fisher Gulch (28.89 ± 0.31/0.48 Ma and 28.74 ± 0.09 Ma [Lipman et al., 1996, recalculated to the Fish Canyon Tuff standard age of 28.2 Ma]), respectively. However, field relations indicate these intrusions are younger than the Chiquito Peak Tuff (Lipman, 1974, 1975; Tomek et al., 2019; this study).

Zircons from two samples of Alum Creek monzonite porphyry, which intrudes the Alamosa River monzonite, yielded 206Pb/238 U dates of 27.42 ± 0.35/0.57 Ma (sample DC1) and 27.32 ± 0.38/0.48 Ma (17AG45), significantly younger than the Alamosa River monzonite (data included here and in Tomek et al. [2019]). The four analyzed zircons with significantly older ages may represent antecrysts associated with the Alamosa River monzonite (Fig. 6). Intrusions outside the caldera yielded dates ca. 28 Ma or younger. Zircons from the Cat Creek monzonite yielded a date of 28.00 ± 0.19/0.19 Ma, while the Lake Annella andesite porphyry yielded a date of 27.71 ± 0.31/0.46 Ma. Three intrusions on the west side of the caldera in the Crater Creek area—the Cataract Creek monzonite, the Elwood Creek monzonite, and the Bear Creek monzonite (equigranular phase)—have indistinguishable dates within error (27.29 ± 0.21/0.36 Ma, 27.06 ± 0.22/0.25 Ma, and 26.92 ± 0.43/0.52 Ma, respectively).

Hornblende andesite-dacite dikes within the caldera along Ranger Creek (sample 17AG28) and Jasper Creek (17AG39) yield analytically indistinguishable dates of 28.27 ± 0.34/0.47 Ma and 28.25 ± 0.30/0.45 Ma. Likewise, zircons from dikes in Horsethief Park (17AG49) and along Crater Lake trail (17AG90) yielded dates of 27.62 ± 0.35/0.49 Ma and 27.45 ± 0.33/0.46 Ma that are analytically indistinguishable from the dates of the intrusions of the Crater Lake area. The youngest intrusion is the Summitville quartz monzonite (24.07 ± 0.25/0.36 Ma).

Whole-Rock Geochemistry

Silica (SiO2) contents of the Platoro intrusions vary from 56 to 70 wt% (Table 2; Table S1[footnote 1]; Fig. 3). These analyses are consistent with previously published data for some of the intrusions and for the Chiquito Peak Tuff (Lipman, 1975; Lipman et al., 1996). Variation diagrams display some compositional scatter for K2O, Na2O, MgO, and Al2O3, but for the other major elements, overall trends are linear for the intrusions (Fig. 8). Lithic fragments were removed from Chiquito Peak Tuff samples prior to analysis, although minor contamination by small lithic clasts is possible. FeOT (FeOT is total Fe as FeO), MgO, P2O5, CaO, and TiO2 concentrations decrease with increasing SiO2 (Fig. 8). Incompatible element concentrations (e.g., Na, K, Ba) correlate positively with SiO2, and K2O abundances define a high-K calc-alkaline trend (Fig. 8). Sr and Al2O3 concentrations correlate negatively, though somewhat irregularly, with increasing SiO2 (Fig. 8), with intracaldera intrusions being less enriched compared with the Chiquito Peak Tuff and the postcaldera lavas. The Chiquito Peak Tuff data are displaced slightly from the overall trends for MgO, FeOT, and Al2O3. Most elemental compositions for the Chiquito Peak Tuff are more evolved than those of the largest postcaldera intrusion, the Alamosa River monzonite, but are similar to those of the Cornwall Mountain quartz monzonite porphyry. The Chiquito Peak Tuff has higher Sr and Ba concentrations, at a given SiO2content, than the Alamosa River monzonite. Overall, the intracaldera intrusions have higher K contents than the intrusions outside the caldera.

Platoro-related samples have chondrite-normalized REE patterns that are enriched in light REEs relative to heavy REEs (Fig. 9A). These REE patterns are slightly concave up, suggestive of a clinopyroxene-amphibole fractionation (Tatsumi, 1989; Pearce and Peate, 1995; Davidson et al., 2013). Sanidine dacite dikes have the most concave patterns, suggesting even greater clinopyroxene-amphibole influence on REE pattern shape (Fig. 9A) (Davidson et al., 2013; Smith, 2014). Only the Alamosa River and Jasper monzonite REE patterns have small negative Eu anomalies (Fig. 9A). La/Yb increases with increasing silica, whereas Dy/Yb decreases with silica in the Platoro samples (Fig. 9B).

Mineral Compositions

Plagioclase

In the equigranular Alamosa River, Jasper, Cat Creek and Elwood Creek monzonites, euhedral plagioclase is commonly surrounded by granophyric quartz-orthoclase intergrowths as much as 300 µm across or by euhedral to subhedral orthoclase rims as much as 200 µm (e.g., Fig. 5B). Plagioclase grains in these units are normally zoned to unzoned with minor oscillatory zoning in some samples. Anorthite content for these equigranular monzonites shows a consistent range between units (Fig. 10; Table S2[footnote 1]). Plagioclase phenocrysts in the Alamosa River monzonite, for example, show limited zoning with anorthite (An) content ranging from An52–33 in the cores to An47–27 in the rims (Fig. 11A). By contrast, compositions of plagioclase in the Chiquito Peak Tuff range more broadly; core compositions range An75–21 and rim compositions An36–16. Its anorthite-rich cores are typically patchy and resorbed with outermost rims of An~40–30 (Fig. 11B). These cores may be either entrained crystals or relict cores in disequilibrium with subsequent melt from which the rims grew.

Hornblende andesite-dacite dikes and the Lake Annella andesite porphyry contain euhedral, normally zoned plagioclase microphenocrysts as much as 500 µm long, but some crystals have resorbed cores and sharp oscillatory zoning (Fig. 10). These textures are consistent with crystallization from more mafic magma that cooled resulting in normally zoned domains, followed by interaction with either more mafic and/or hotter pulses of magma that produced sharp jumps in anorthite content. Of the intrusions west of the Platoro caldera complex, plagioclase in the Bear Creek monzonite is the most calcic; core compositions range An80–36 and rim compositions An69–30 (Fig. 10).

Strontium concentrations in plagioclase from Platoro caldera complex rocks range from below detection limit (~150 ppm) to 3720 ppm (Table S2[footnote 1]). The intracaldera monzonite intrusions have, in general, the lowest concentrations, and the Chiquito Peak Tuff has the highest concentrations, much higher than the intracaldera Alamosa River monzonite and Cornwall Mountain quartz monzonite porphyry (Fig. 11C). The Sr content of plagioclase in the upper member of the Summitville Andesite is 1200–2300 ppm, between that of the Chiquito Peak Tuff and the intracaldera intrusions. Plagioclase compositions (both anorthite and Sr contents) for the dacite of Fisher Gulch overlap those of the Chiquito Peak Tuff, Summitville Andesite, and intracaldera intrusions.

Potassium Feldspar

Sanidine phenocrysts are as much as 3 mm long in the Chiquito Peak Tuff and 7 mm long in the sanidine dacite dikes. A single highly resorbed and embayed sanidine phenocryst (0.2 mm) was identified in the Ranger Creek hornblende dacite dike, but none were recognized in the other hornblende andesite-dacite dikes. In all other units, potassium feldspar occurs either in the groundmass, as granophyric intergrowths with quartz, or as rims on plagioclase.

Sanidine compositions range in orthosclase content from Or69–64 in the Chiquito Peak Tuff, a range similar to that reported by Lipman et al. (1996). Zoning in these phenocrysts is less well developed than that characteristic of plagioclase in the Chiquito Peak Tuff. This phase in a sanidine dacite dike is zoned with Or72–79. The single sanidine phenocryst in the Ranger Creek horn-blende dacite dike ranged Or64–68. Sanidines in the Chiquito Peak Tuff contain ~5900–17,000 ppm Ba. Rare Ba-zoned sanidines include sharp resorption surfaces, across which Ba concentrations are ~9100 ppm in the core to as much as 15,700 ppm in the rim. Ba concentrations in sanidines from the sanidine dacite dike range from ~2600 to 17,100 ppm. Two types of Ba zoning were identified: (1) oscillatory zoning, in which Ba concentrations in rims are high (10,000–17,100 ppm), and (2) step-function zoning in which resorbed low-Ba cores are surrounded by high-Ba rims (10,000–15,000 ppm). The single resorbed grain in the Ranger Creek hornblende dacite dike preserves the latter type of zoning, with a core containing 4000 ppm Ba and a rim with 15,000 ppm Ba.

Quartz

Quartz occurs predominately as anhedral groundmass grains in the large intrusions, although the sanidine-quartz dacite dikes contain resorbed and rounded phenocrysts (as much as 4 mm in diameter), many with embayed margins, consistent with decompression or mixing with a hotter magma.

Pyroxene

Augite is present in all the large Platoro intrusions and in one of the horn-blende andesite-dacite dikes. In a few samples of the Alamosa River and Jasper monzonites, fine exsolution textures in augite, a consequence of augite- pigeonite- hypersthene immiscibility, are indicative of rapid cooling. Altered pyroxenes in the Cornwall Mountain quartz monzonite porphyry were not analyzed.

Augite in the intrusions has a wide range of Mg# [Mg# = 100 × Mg2+/(Mg2+ + Fe2+)], Ti, and Mn (Figs. 12A–12B; Table S3[footnote 1]). Mg# for augite phenocrysts and groundmass crystals ranges from 60 to 84; augite from the Cat Creek monzonite is the most magnesian (Mg# 69–84), whereas that in the Bear Creek monzonite has the broadest range (Mg# 60–76). Augite in the hornblende dacite dike 17AG39 has a more restricted compositional range (Mg# 66–76). TiO2 concentrations in the augite range from below detection limit (0.01 wt%) to 1.5 wt%. Relative to augite phenocrysts in the other intrusions, those in the hornblende dacite dike at Jasper Creek have the highest mean TiO2 concentration (0.65%) and define the largest compositional range (0.33%–1.41%). Augite TiO2 concentration correlates inversely with Mg# in this dike (Fig. 12) but is uncorrelated in the other intrusions. Augite from the Alamosa River monzonite and the Bear Creek monzonite have indistinguishable mean TiO2 contents (0.51 wt%), whereas augite from the Cat Creek monzonite has a lower mean TiO2 concentration (0.33 wt%). The mean TiO2 content of augite from the fine-grained andesite dike 17AG80 is 0.48 wt%. Augite from the remaining units, including the Chiquito Peak Tuff, contains <0.32 wt% TiO2 (Fig. 12). MnO concentrations of the analyzed augites vary considerably. Augite in the Alamosa River monzonite, Alum Creek monzonite porphyry, Cat Creek monzonite, equigranular Bear Creek monzonite, andesite dike 17AG80, Summitville Andesite (upper member), and hornblende dacite dike on Jasper Creek contains 0.11–0.64 wt% MnO, whereas that in the sanidine dacite dike 17AG90, porphyritic Bear Creek monzonite, Cataract Creek monzonite, and Chiquito Peak Tuff contain >0.8 wt% (Fig. 12). Augite in the Chiquito Peak Tuff has the greatest MnO contents, as much as 1.52 wt% (Fig. 12).

Unaltered orthopyroxene was identified only in the Bear Creek monzonite and an andesite dike (sample 17AG80), although Lipman (1975) reported minor hypersthene in the Alamosa River, Cat Creek, and Jasper monzonites. Domains of serpentine ± chlorite in these units may constitute altered orthopyroxene. All analyzed orthopyroxene crystals have lower Mg# than the analyzed augite crystals. For example, augite from the Bear Creek monzonite has mean Mg# of 68, whereas that for the orthopyroxene is 53.

Biotite

Biotite is present in all Platoro-associated units except the fine-grained andesite dikes; it is most abundant in the intrusions and porphyritic dikes west of the caldera. Biotite is anhedral and interstitial to plagioclase in the equigranular monzonites (Fig. 5C), but in the porphyritic units, phenocrysts are euhedral to subhedral and form glomerocrysts with plagioclase + Fe-Ti oxides ± augite ± amphibole (Fig. 5I). Biotite commonly contains inclusions of apatite and may be intergrown with Fe-Ti oxides. Chloritized biotite in several porphyritic units could not be analyzed.

Although the biotite grains have variable Mg#, they are unzoned and have compositions that are somewhat unique to each of the units (Table S4[footnote 1]; Figs. 12C–12D). Biotite grains in the Bear Creek monzonite generally have the most iron-enriched Mg# of the analyzed biotites and the lowest TiO2 concentrations (<4 wt%). Biotite in the other units contains 1.8–7.1 wt% TiO2. Biotite grains in the Alamosa River and Cat Creek plutons are much more magnesian than those in the Chiquito Peak Tuff. TiO2 content is relatively uniform within samples, except for samples of the Alamosa River monzonite. MnO concentrations among the analyzed biotites decrease with increasing Mg#, and biotite has higher MnO concentrations in the Chiquito Peak Tuff than in the monzonite and quartz monzonite intrusions.

Amphibole

The Lake Annella andesite porphyry, Cataract Creek monzonite, Bear Creek monzonite, and hornblende andesite-dacite and sanidine dacite dikes contain euhedral to subhedral amphibole phenocrysts; amphibole grains are also glomerocryst constituents. Amphibole phenocrysts contain inclusions of apatite, plagioclase, and Fe-Ti oxides. Amphibole is absent in the intracaldera plutons and in the Cat Creek intrusion. The amphiboles are Ca and Mg rich and are composed of magnesio-hornblende, magnesio-hastingsite, and pargasite (Hawthorne et al., 2012) (Table S5[footnote 1]).

Total alkali contents of amphibole phenocrysts within each unit, and for all analyzed grains, increase with increasing AlTOT (total aluminum). Compositions of amphibole from the Platoro intrusions form two groups (Fig. 12E): (1) a high-Al group (Al2O3 = 9.4–13.3 wt%) of predominantly magnesio-hastingsite and pargasite, and (2) a low-Al group (Al2O3 = 5.9–8.4 wt%) of magnesio-horn-blende. Amphibole phenocrysts from the Lake Annella andesite porphyry, Bear Creek monzonite, and sanidine dacite dike constitute the high-Al group (although a few analyses from the sanidine dacite dike are consistent with the low-Al group), whereas amphiboles from the Cataract Creek monzonite and the hornblende dacite dike of Ranger Creek define the low-Al group (Fig. 12E).

Zircon

Abundances of trace elements in zircon from Platoro rocks vary considerably (Fig. 13; Table S6[footnote 1]). Hf concentrations in these zircons range from 6170 to 12,400 ppm (Fig. 13). Higher Hf concentrations are commonly interpreted as indicative of more evolved melts (Hoskin and Schaltegger, 2003), and whole-rock silica contents of Platoro intrusions correlate positively with Hf concentrations. The analyzed grains are unzoned with respect to trace element abundances. Zircon in the Summitville quartz monzonite contains has the highest Hf (9400–12,400 ppm). However, zircon in the Alamosa River monzonite varies most broadly and includes similarly elevated abundances (~7770–11,950 ppm). By contrast, zircon in the Bear Creek monzonite (58 wt% SiO2) has the lowest abundances and narrowest range of Hf abundances (6630–8500 ppm).

Although compatible in zircon, U and Th are incompatible in silicic melts, and thus their abundances are positively correlated with increasing Hf in zircon (Claiborne et al., 2010). Zircons in the Alamosa River monzonite define the broadest compositional range for U among all analyzed samples, include the greatest U values (110–6220 ppm, mean 1660 ppm), and have U values that are weakly positively correlated with increasing Hf abundances (Figs. 13A–13B). Zircon from the Alamosa River monzonite also contains the highest Th contents. U and Th contents of zircon from the Alamosa River monzonite are significantly greater than those of the Chiquito Peak Tuff (U, mean 170 ppm; Th, mean 135 ppm). Among zircon from intrusions younger than 28 Ma, most U contents range from 100 to 1000 ppm, but are uncorrelated with Hf content.

Chondrite-normalized zircon REE patterns for all samples are similar to well-established magmatic REE patterns and include characteristic positive Ce anomalies (Trail et al., 2012). The Alamosa River and Jasper monzonites also contain significantly greater total REE abundances than the Chiquito Peak Tuff (Table S6[footnote 1]). Although the magnitude of negative Eu anomalies in zircon is commonly negatively correlated with Hf abundance, which is indicative of feldspar crystallization (e.g., Hoskin and Schaltegger, 2003; Cooper et al., 2012; Watts et al., 2016), no such correlation is evident among zircons from the Platoro intrusions or the Chiquito Peak Tuff (Figs. 13C–13D). Importantly, Eu anomalies in zircon may also be influenced by the redox state of the magma (Dilles et al., 2015). At any given Hf content, the intracaldera monzonites have lower Eu/Eu* where Eu* is √(SmN × GdN) than the Chiquito Peak Tuff. The Bear Creek monzonite also has low Eu/Eu* (<0.2). The hornblende andesite-dacite dikes, sanidine dacite dike 17AG90, Elwood Creek monzonite, and the Summitville quartz monzonite all have Eu/Eu* >0.4.

Zircon Hf isotopic composition is commonly used as a geochemical tracer of a host magma's origin, in a similar manner to the way in which whole-rock Nd isotopes are used. Additionally, the robust nature of zircon, with regard to Lu and Hf, means that this system is less vulnerable to isotopic disturbances. Lu and Hf isotopic data for zircon from the Platoro caldera complex define εHf isotopic compositions from −9.2 (±3.2) to 1.4 (±1.7); most are generally uniform within a given unit (Fig. 14A; Table S7[footnote 1]). Single zircons in the Cataract Creek monzonite and Alamosa River monzonite have cores with outlier values of −28.3 (±1.4) and −30.4 (±3.2), respectively. These analyses may represent inherited cores and are not included in Figure 14A. Initial εHf values and SiO2 abundances are uncorrelated, although the younger intrusions appear to be characterized by progressively more negative values. Relative to crustal Lu/Hf evolution lines, zircon εHf values for Platoro-associated rocks are consistent with derivation from a lithospheric source with a similar age and composition as the 1.1–1.7 Ga crust of the southern Rocky Mountains (Fig. 14B).

Intensive Parameter Estimates

Temperature

Titanium-in-zircon thermometry estimates the temperature of zircon crystallization in magma, provided that SiO2 and TiO2 activities (aSiO2 and aTiO2) are known (Watson and Harrison, 2005; Watson and Harrison, 2006; Ferry and Watson, 2007). We assume an aSiO2 of 1, as indicated by the presence of quartz in the Platoro-associated rocks, and an aTiO2 of 0.7, as indicated by the presence of titanite in most of these rocks. Varying assumed aTiO2 between 0.6 and 0.8 results in calculated temperature variations of as much as 30 °C (Ferry and Watson, 2007). Data for the zircons from the largest intrusions (the Alamosa River, Jasper, Cat Creek, and Bear Creek monzonites) all suggest decreasing temperature with increasing Hf (Figs. 13E–13F, 15; Table S8[footnote 1]), which is typical of zircon behavior in granitic rocks (e.g., Claiborne et al., 2010; Cooper et al., 2014; Watts et al., 2016; Colgan et al., 2018). Other Platoro intrusions exhibit smaller temperature variations during crystallization, consistent with more rapid cooling. Alamosa River monzonite zircons have the highest and greatest range of Ti concentrations (5–40 ppm) and consequently exhibit the most extensive range of zircon crystallization temperatures (719–950 °C). Zircon from the Jasper, Cat Creek, and Bear Creek monzonites reflect similar crystallization temperatures (mostly >800 °C; Fig. 13). Zircon thermometry from the porphyritic intrusive units yields cooler temperatures, <800 °C in some cases, extending almost to the H2O-saturated granite solidus (Fig. 15). Zircon thermometry from the Chiquito Peak Tuff defines temperatures between 719 and 822 °C, except for one zircon that yields an anomalously high temperature of 888 °C.

Crystallization temperatures for amphibole-bearing units were calculated using the plagioclase-amphibole thermometer of Holland and Blundy (1994). Applied to in-contact amphibole-plagioclase pairs, the uncertainty associated with this method is ±40 °C (2σ). Calculated amphibole-plagioclase temperatures for Platoro system rocks range from ~820 to 970 °C (Fig. 15; Table S7[footnote 1]). Temperatures range from 822 to 829 °C for the intracaldera hornblende dacite dike on Ranger Creek, 925–966 °C for the Lake Annella andesite porphyry, 849–951 °C for the Bear Creek monzonite, and 817–877 °C for the Cataract Creek monzonite. These calculated temperatures extended to higher values than determined by the Ti-in-zircon thermometer for the same unit, confirming earlier crystallization of these phases compared with zircon.

Pressure

The validity of pressure estimations based on amphibole barometry has been increasingly questioned due to the sensitivity of amphiboles to melt composition and temperature in addition to pressure (e.g., Erdman et al., 2014; Putirka, 2016). Most amphibole barometers require buffering assemblages, and many are calibrated for temperatures <800 °C (e.g., Anderson and Smith, 1995; Mutch et al., 2016). The amphibole-bearing units at Platoro contain the necessary mineral assemblage; however, some units, most notably dikes, exhibit open-system petrologic characteristics including presence of mafic enclaves and glomerocrysts. Additionally, because most plagioclase-amphibole temperature estimates for these units are >800 °C, we do not estimate crystallization pressures for these units.

DISCUSSION

We integrate new U-Pb zircon ages with observed stratigraphic constraints, mineral assemblages, chemistry, and texture as well as whole-rock chemistry for Platoro intrusions to better constrain the magmatic history of the terminal postcaldera magmatic system, interpreted here as the waning stages of the Platoro caldera complex. All U-Pb zircon dates are interpreted as magma intrusion ages within limits of reported two-sigma errors. All intrusions postdate the terminal erupted ignimbrite from the Platoro caldera system, the Chiquito Peak Tuff, as supported by either new geochronology or mapped field relations. As such, these units may reflect a range of petrogenetic affinities associated with the youngest erupted ignimbrite, or the magmatic underpinnings of postcaldera effusive volcanism to which some intrusions are spatially associated. Alternatively, new data are also considered in light of a hypothesis whereby most, if not all, postcaldera intrusions simply reflect largely unrelated pulses of magma introduced into the upper crust from magma reservoirs at mid- to lower-crustal levels that experienced rapid rise and emplacement in the upper crust and minimum residence times at shallow levels. This would imply most intrusions need not be derived from a large integrated upper-crustal magma chamber but are potentially derived from a shared petrogenetic history at deeper crustal levels. Specifically, we evaluate petrogenetic relationships between the intracaldera intrusions, Chiquito Peak Tuff, Summitville Andesite, and postcaldera dacite of Fisher Gulch. Although not necessarily definitively, we use new data to better constrain timing of ore-deposit mineralization and hydrothermal alteration.

Temporal Constraints on Postcaldera Intrusive Magmatism and Mineralization

Intrusion emplacement ages for the Platoro caldera complex are spatially associated with proximity to mapped caldera margins of the ca. 29 Ma Chiquito Peak Tuff eruption. With two exceptions, intracaldera intrusions yield crystallization ages ranging from 28.98 ± 0.18 Ma to 28.25 ± 0.30 Ma and are indistinguishable, within analytical uncertainty, from previously reported eruption ages for the Chiquito Peak Tuff (Fig. 6; Table 3; Lipman et al., 1996). Notable exceptions include the Alum Creek monzonite porphyry (27.32 ± 0.38 Ma) and the Summitville quartz monzonite (24.07 ± 0.25 Ma). These younger intrusions either host or are inferred to have been causative to important hydrothermal mineralization. New ages on these intrusions provide additional constraints on post-emplacement alteration. Extracaldera intrusions yield crystallizations ages ranging from ca. 28.0 to 26.3 Ma with a magmatic hiatus between ca. 27.3 and 26.3 Ma. Younger volcanic activity in the Platoro area is volumetrically minor, bimodal, and associated with late Oligocene to Miocene volcanism more typically associated with early stages of Rio Grande rift extension (Lipman, 1974, 1975; Thompson and Machette, 1989; Thompson et al., 1991).

The resolution of the new U-Pb zircon geochronology is insufficient to establish discrete temporal breaks in intrusive emplacement beyond the age ranges delineated above. Consequently, quantifying the duration of pulsed activity or periods of quiescence or quantitatively estimating magma flux over the life span of a single or even multiple intrusion(s) is beyond the scope of this study. High-precision chemical abrasion isotope dilution thermal ionization mass spectrometry (CA-ID-TIMS) U-Pb zircon investigation may better resolve both the duration of emplacement for individual plutons and the spatial and temporal evolution between mapped intrusions.

All the exposed granitoid plutons are too young to have been associated with any of the pre–Chiquito Peak ignimbrites of the Platoro caldera complex. The absence of older caldera-related intrusions suggests a possible change in magma-chamber evolution during the terminal cycle of the Platoro system. Previous eruptive cycles may have been characterized by higher magma fluxes, facilitating development of focused shallow chambers subject to ignimbrite eruption, rather than smaller, isolated intrusions beyond caldera margins.

It is commonly accepted that intrusions form over protracted spans of time as a result of incremental growth from repeated pulses of magma (e.g., Coleman et al., 2004; Glazner et al. 2004; Leuthold et al., 2012; Miller et al., 2011; Frazer et al., 2014; Annen et al., 2015). The Alamosa River monzonite is the largest intrusive body exposed within the caldera (Lipman, 1974), and magnetic fabrics delineated in the Alamosa River monzonite document three domains within the intrusion that are interpreted as discrete intrusive pulses of magma emplaced over a short, but unknown, time period relative to the total intrusion emplacement interval (Tomek et al., 2019). As such, it may have had a more protracted emplacement history than that of small intrusions.

Hydrothermal alteration and mineralization are spatially associated with several postcaldera intrusions of the Platoro caldera magmatic system, most notably the Alum Creek, Alamosa River, and Summitville intrusions. Previous studies (Steven and Ratté, 1960; Mehnert et al., 1973; Lipman, 1975; Neuerburg, 1978; Bethke et al., 2005) asserted causative and/or temporal association of mineralization with individual plutons in area mineralization districts. However, the lack of geochronologic data limits determination of both spatial and temporal interpretations of genetic links to intrusion emplacement. Our new intrusion emplacement ages further constrain the timing of hydrothermal alteration and mineralization in the Platoro area. In the Stunner district, alteration and mineralization are superposed on the 27.32 ± 0.38 Ma Alum Creek monzonite porphyry. Steven and Ratté (1960) recognized two periods of hydrothermal activity in this area: (1) an older one they interpreted as associated with the northern side of the Alamosa River pluton in the area of the Alum Creek monzonite porphyry, and (2) a younger one associated with the dacite lava dome at Summitville. The new geochronologic data cannot resolve multiple episodes of alteration but do support an interpretation that it is associated with the Alum Creek monzonite porphyry. Alteration and mineralization in the Crater Creek area (Fig. 2) locally overprint all intrusions, including the sanidine dacite dike 17AG90, and consequently must be younger than ca. 26.3 Ma.

The U-Pb zircon date for the Summitville quartz monzonite (24.07 ± 0.25/0.36 Ma; Fig. 6) is older than the ca. 23 Ma date for alteration in the Summitville epithermal gold deposit and for the quartz-sanidine dacite lava dome that hosts it. However, if four anomalously old zircons (possible antecrysts, with uncertainties beyond the weighted mean) are excluded, the monzonite date becomes analytically indistinguishable from that for the mineralized dacite. Therefore, the Summitville quartz monzonite could have been the source of fluids responsible for mineralization and wall-rock alteration at Summitville, as previously inferred by Bethke et al. (2005). Most importantly, mineralization and hydrothermal alteration spatially associated with intrusions of the Platoro caldera complex occurred recurrently from ca. 29 to 23 Ma, all postdating ignimbrite eruptions and associated caldera collapse.

It is notable that ages of the exposed extracaldera Platoro plutons (28.0–26.3 Ma) closely parallel the time span of major younger ignimbrite eruptions from the central San Juan caldera cluster (ca. 28.2–27.05 Ma) (Lipman, 2007). This reflects the migration of magmatic activity to the northwest, with fluxes decreasing in the Platoro area as they increased in the area of the central San Juan caldera cluster.

Magma Storage and Differentiation

Compositions of the Platoro postcaldera rocks range from andesite to rhyolite and plot along a high-K calc-alkaline trend (Figs. 3, 8A), as is characteristic for the Southern Rocky Mountain volcanic field (SRMVF). Although the ca. 24 Ma Summitville quartz monzonite is more evolved, the long time span of intrusion and absence of systematic compositional evolution through time shows that the associated magmas cannot be related by any closed-system fractional crystallization process. Consequently, multiple batches of magma were likely responsible for the compositional range observed in the Platoro system. Increasing La/Yb and decreasing Dy/Yb with increasing silica in the Platoro magmas, indicative of amphibole (and possibly titanite) fractionation (Davidson et al., 2007), support differentiation at even deeper crustal levels for some of these magmas (Fig. 9B) prior to emplacement in shallow crust.

The variable mineralogy within the restricted bulk compositional range of the larger Platoro intrusions (monzonite to quartz monzonite) suggests cotectic or peritectic conditions with the magmas in a low variance state. The reaction-boundary crystallization buffers the bulk chemical composition, as demonstrated in reactive flow models (Jackson et al., 2018) and in phase-equilibria experiments on Mount St. Helens dacite (Washington State, USA) (Blatter et al., 2017). Phase relations within magmas of similar compositions vary with pressure, thus differentiation of magmas at different depths leads to variations in mineral assemblages and compositions, as evidenced in the Platoro intrusions. For example, larger monzonite intrusions contain augite but no amphibole, whereas chemically similar intrusions west of the caldera contain abundant amphibole.

In the smaller intrusions and dikes, mafic enclaves and abundant glomerocrysts suggest open-system behavior, likely magma mixing, particularly within the <28 Ma intrusions west of the caldera. Phenocrysts with distinct rim compositions, well-developed resorption surfaces, and compositionally diverse plagioclase crystals in single samples of the hornblende dacite dikes and Lake Annella andesite porphyry suggest that magma mixing played a prominent role in the petrogenesis of these younger intrusions. In addition, these units contain abundant glomerocrysts that are consistent with crystal entrainment where the transporting melt scavenged crystals and crystal clusters from mushy domains within the magmatic system. The larger intracaldera intrusions, by contrast, lack enclaves and glomerocrysts, and their plagioclase phenocrysts have more uniform compositions, suggesting either that reservoirs associated with the earliest intrusive magmatism were less affected by open-system behavior or that entrained crystals were completely resorbed by larger melt volumes possibly associated with higher early- postcaldera magma flux rates.

Crystals record magma-reservoir compositional evolution through time. In the Platoro magmatic system, phenocrysts in intrusions constrain how the unerupted parts of associated magma reservoirs evolved. Within the >28 Ma monzonite plutons, there are consistent mineral assemblages of plagioclase, augite, biotite, potassium feldspar, and quartz, and in the analyzed phases, compositions are broadly similar. Among these older units, plagioclase compositions are relatively uniform (Fig. 10), which suggests compositional evolution without significant mixing of crystal cargo. Negligible Eu anomalies (Eu/Eu* = 0.74–0.95) characteristic of most of these intrusions, excluding the Alamosa River and Jasper monzonites (Eu/Eu* = 0.59–0.93), suggest early suppression of plagioclase at high pH2O, probably during deep differentiation (Moore and Carmichael, 1998; Müntener et al., 2001). Euhedral augite and plagioclase grains in the larger monzonite intrusions suggest these phases crystallized early in the crystallization sequence, with augite, or plagioclase in some instances, likely being liquidus places.

Granophyric textures, characteristic of the Alamosa River, Jasper, Cat Creek, and Elwood Creek monzonites (Figs. 5B, 11A), are evidence of undercooling and rapid solidification (Barker, 1970; Morgan and London, 2012), consistent with depressurization and volatile loss during emplacement into the shallow crust from deeper reservoirs or even associated with venting during volcanic eruptions.

Experimental petrology studies focused on magmatic storage conditions pertinent to the Fish Canyon Tuff (Johnson and Rutherford, 1989; Caricchi and Blundy, 2015), erupted from the nearby La Garita caldera in the central San Juan Mountains (Fig. 1A), help constrain the intensive parameters prevailing for the Chiquito Peak magmatic reservoir. We compare the experimental run products from Caricchi and Blundy (2015) and Johnson and Rutherford (1989) with the modal mineralogy of the Chiquito Peak Tuff. Despite a similar bulk composition to the Fish Canyon Tuff, the Chiquito Peak Tuff contains no horn-blende or quartz and has only minor sanidine. Johnson and Rutherford (1989) reported experimental results for a range of pressures between 200 and 500 MPa, whereas all experiments of Caricchi and Blundy (2015) were performed at 200 MPa. Both sets of experiments showed that augite (a major mafic phenocryst mineral in the Chiquito Peak Tuff) disappears in the range of 800–820 °C for runs at 200 MPa. Experiments by Caricchi and Blundy (2015) indicated that hornblende is stable only below 800 °C in water-saturated runs, and below 850 °C for runs that contained 4 wt% H2O. Sanidine is stable below 775 °C only. Consequently, magma represented by the Chiquito Peak Tuff likely equilibrated in a reservoir in which the prevailing pressure was ~200 MPa. In addition, the lack of hornblende and the coexistence of sanidine and augite suggest that the H2O content of the Chiquito Peak magma was likely less than that of the Fish Canyon Tuff magma (<4 wt% H2O) (e.g., Johnson and Rutherford, 1989).

For the amphibole-bearing intrusions of the Platoro magmatic system— the Bear Creek monzonite, Cataract Creek monzonite, Lake Annella andesite porphyry, hornblende dacite dikes, and sanidine dacite dikes—amphibole is commonly euhedral and commonly intergrown with plagioclase, suggesting it crystallized early. The H2O contents of these magmas were likely >4–6 wt%, as required to stabilize amphibole (Naney, 1983; Richards, 2011; Richards et al., 2012; Blatter et al., 2017). Plagioclase-amphibole thermometry for these units constrains the temperature range for their magma storage to be between 960 and 850 °C for the Lake Annella and Bear Creek intrusions and between 880 and 820 °C for the Cataract Creek monzonite and Ranger Creek hornblende dacite dike.

All zircon εHf values for the Platoro intrusions and Chiquito Peak Tuff plot below those for depleted mantle and chondritic uniform reservoir (CHUR) (Fig. 14A), which helps constrain likely sources for these magmas. The data indicate a similar source but with inadequate resolution to evaluate variation in the amounts of assimilated crust. The lack of a correlation between zircon εHf and SiO2 content suggests that the isotopic signature was imparted from parental melts during differentiation with some assimilation, which is supported by the presence of zircon xenocrysts in the Alamosa River, Jasper, Bear Creek, and Cataract Creek monzonites and the Chiquito Peak Tuff, as well as the two extremely negative εHf values interpreted as inherited cores (Table S7[footnote 1]). εHf data for the Platoro zircons are similar to those for zircon from Paleogene intrusions in New Mexico and Texas (Chapman et al., 2018). The range of zircon εHf values for Platoro plots on crustal Lu/Hf evolution lines (Fig. 14B) between 1.1 and 1.7 Ga, and most are consistent with 1.4 Ga, an observation in accord with recognition that most of the San Juan volcanic locus (SJVL) is underlain by 1.4 Ga plutons and 1.7–1.6 Ga crust of the Mazatzal province (Reed et al., 1993; Shaw and Karlstrom, 1999; Whitmeyer and Karl-strom, 2007; Bickford et al., 2015). The zircon Hf isotopic data are consistent with those of other whole-rock isotopic (Sr, Nd, and Pb) studies on pre- and postcaldera Platoro rocks indicating these magmas have a lithospheric mantle origin with lower-crustal contamination (Lipman et al., 1978; Doe et al., 1979; Colucci et al., 1991; Balsley and Gregory, 1998) as well as studies in other caldera systems and volcanics in the SRMVF (Johnson et al., 1990; Riciputi et al., 1995; Lake and Farmer, 2015).

The time interval spanned by the eruption of Chiquito Peak Tuff, the intracaldera intrusions, and the youngest intrusion, the Summitville quartz monzonite, is ~5 m.y., an extensive amount of time for postcaldera magmatism. Heat-transfer models for magma delivered in pulses to the upper crust separated by long pauses show that magma pulses solidify in the interim (Annen, 2009; Schöpa and Annen, 2013; Barboni et al., 2015). Consequently, emplacement rate controls whether large-scale magma chambers form (Hardee, 1982; Gelman et al., 2013; Annen et al., 2015; Blundy and Annen, 2016). At Platoro, the volumes of magma erupted or emplaced decreased through time, suggesting a declining magma supply rate. These temporal relations in addition to variable mineral assemblages and whole-rock and mineral chemistry indicate that the Platoro postcaldera magmas consisted of multiple discrete batches emplaced into the shallow crust.

Postcollapse Evolution of Intracaldera Magmas

Field relations indicate that all Platoro intracaldera intrusions are younger than the Chiquito Peak Tuff (Fig. 4D), but the resolution of U-Pb SHRIMP zircon dates preclude temporal distinction between ignimbrite eruption and emplacement of intracaldera intrusions. Consequently, any petrogenetic relationship between Chiquito Peak Tuff magma and associated residuum, postcaldera lavas (dacite of Fisher Gulch and Summitville Andesite), and intracaldera intrusions can only be evaluated using petrologic constraints based on bulk and mineral compositions. This does not presume the possibility of a direct parental relationship between post-ignimbrite residuum and subsequent volcanic eruptions or intrusion emplacement, but does constrain the probability, or lack thereof, of related petrogenetic evolution paths within the broader Platoro magmatic system.

Whole-rock major oxides show similar abundances for a given silica content among the units, but the main intracaldera intrusions (Alamosa River and Jasper monzonites) are lower in Al2O3 at a given silica content than the postcaldera lavas and the Chiquito Peak Tuff (Fig. 8). Trace elements such as Sr and Ba are also lower in these intrusions than in the tuff and postcaldera lavas for a given silica (Fig. 8). This, along with the lower Al2O3, is opposite of what would be expected if the intrusions represented crystal residue left behind in associated eruptions because these elements are preferentially incorporated into plagioclase and alkali feldspar, principal contents of the putative crystal residue (Glazner et al., 2015). By contrast, the Cornwall Mountain quartz monzonite porphyry has Sr and Ba abundances similar to those of the Chiquito Peak Tuff and dacite of Fisher Gulch.

Mineral assemblages are similar among the units, except for the lack of biotite and potassium feldspar in the less-evolved Summitville Andesite. However, the Chiquito Peak Tuff contains larger and more abundant plagioclase and biotite phenocrysts (as much as ~3 mm) than the Alamosa River or Jasper monzonites, which are mostly fine grained and equigranular. Mineral compositions among units show more variation. Plagioclase compositions for the Chiquito Peak Tuff and the postcaldera lavas extend to higher anorthite content than do those of the younger intracaldera intrusions (Fig. 10). For example, the Chiquito Peak Tuff contains high-anorthite (An>60) plagioclase cores, absent from intracaldera-intrusion plagioclase cores but overlapping with the range of anorthite content observed in the dacite of Fisher Gulch and Summitville Andesite (Fig. 10). Additionally, the Sr contents in plagioclase in the intracaldera intrusions is distinctly lower than that of the Summitville Andesite and Chiquito Peak Tuff but do overlap somewhat with that of the dacite of Fisher Gulch (Fig. 11B).

Augite compositions have similar ranges in Mg#, TiO2, and MnO for the intracaldera intrusions, Summitville Andesite, and dacite of Fisher Gulch. The Chiquito Peak Tuff has much higher MnO and much lower TiO2 than the other units. This likely reflects earlier saturation in Fe-Ti oxides in the Chiquito Peak magma versus in the other units, with the decrease in Ti in augite the result of less Ti in the liquid due to its incorporation in ilmenite and titanomagnetite (e.g., Walker et al., 1973). Biotite compositions show more distinct groupings for these units (Fig. 12).

Zircon trace element compositions for the Alamosa River and Jasper monzonites are distinct from those of the Chiquito Peak Tuff and dacite of Fisher Gulch, as demonstrated by U content, Eu/Eu*, and Ti-in-zircon temperatures (Fig. 13). By contrast, zircon trace element compositions of and Ti-in-zircon temperatures for the Cornwall Mountain quartz monzonite porphyry overlap completely with those for the Chiquito Peak Tuff and dacite of Fisher Gulch (Fig. 13). Distinctive zircon trace element characteristics, such as those displayed by Chiquito Peak Tuff and the Alamosa River and Jasper monzonites, suggest crystallization from discrete, separate magma reservoirs (e.g., Till et al., 2019).

Thus, we conclude that the Alamosa River and Jasper monzonites are not the unerupted part of the Chiquito Peak upper-crustal magma reservoir, but rather that their petrogenetic ancestries likely diverged prior to final differentiation, emplacement, or eruption. Their εHf values overlap, as do those for most of the postcaldera intrusions; all may have differentiated from a similar parental magma but at different depths in discrete reservoirs within the crust. The whole-rock compositional affinity, modal mineralogy, and mineral chemistries for the Cornwall Mountain quartz monzonite porphyry, dacite of Fisher Gulch, and Chiquito Peak Tuff do support an upper-crustal, cogenetic differentiation.

The Alamosa River and Jasper monzonites are similar in whole-rock composition to the postcollapse Summitville Andesite but extend to more evolved whole-rock and plagioclase compositions. These intrusions were postulated to have cored volcanic edifice(s) related to eruption of the Summitville Andesite because they intrude this unit (Lipman, 1974, 1975; Lipman et al., 1996). The Summitville Andesite likely represents a large recharge component without prolonged upper-crustal storage and fractionation, similar to the precaldera andesitic lavas of the Conejos Formation. The Alamosa River and Jasper monzonites may be related to the Summitville Andesite, although the intrusions have lower Al2O3 and Sr for given silica content, likely due to plagioclase fractionation. These intrusions are likely distinct pulses of magma emplaced into the volcano that had sourced the Summitville Andesite.

Crustal Geometry of Intracaldera Intrusions

Present-day exposures of the intracaldera intrusions at Platoro are at levels close to the Oligocene paleo–land surface (Fig. 16), too shallow to provide direct constraints on the geometry of caldera intrusions deeper in the crust. Geologic mapping and the new ages demonstrate that no sizable exposed intrusions are older than the Chiquito Peak Tuff. If intrusions associated with the earlier ignimbrite eruptions exist, they must lie deeper in the composite caldera block. The lack of granitoid lithics in the Chiquito Peak Tuff provides no information about such intrusions at depth. Other lithics such as Precambrian basement, Mesozoic sediment, and earlier ignimbrites are also rare or absent. The Chiquito Peak Tuff contains only andesitic lithics, likely derived from earlier caldera-filling lavas or the precaldera Conejos Formation by shallow vent enlargement or in-sliding of adjacent oversteepened caldera walls (Lipman et al., 1996), processes well documented at other calderas (e.g., Bacon, 1983; Lipman, 1984; Hildreth and Mahood, 1986).

Eruptions of large-volume ignimbrites have been inferred by some to result from near-complete emptying of magma bodies assembled at high flux rates without concurrent formation of sizable deeper granitoid plutons, while batholithic-scale plutons are proposed to have been assembled under contrasting conditions of incremental magma accumulation at lower fluxes without major associated volcanism (e.g., Annen, 2009; Tappa et al., 2011; Zimmerer and McIntosh, 2012; Mills and Coleman, 2013; Caricchi et al., 2014; Frazer et al., 2014; Schaltegger et al., 2019). Alternatively, geologic map relations for seemingly cogenetic volcano-plutonic suites at crustal levels deeper than those at the SJVL and modeling of gravity, seismic, and other geophysical data document the presence of vertically extensive batholithic-scale granitoid magma bodies or solidified intrusions in the upper crust beneath caldera sources for large ignimbrites (Hamilton and Myers, 1967, 1974; Lipman, 1984, 2007; Bachmann et al., 2007; de Silva and Gosnold, 2007; Ward et al., 2014; Best et al., 2016).

A large negative gravity anomaly suggests the presence of such a composite upper-crustal batholith beneath much of the SJVL (Plouff and Pakiser, 1972; Steven and Lipman, 1976; Drenth et al., 2012; Lipman and Bachmann, 2015), although the timing of its assembly is unconstrained. Similar gravity anomalies are associated with ignimbrite calderas along the Sawatch trend farther north in the SRMVF (Fig. 1) and at the Questa locus farther south (Cordell et al., 1985; Lipman, 2007). The Platoro caldera complex is located along the southeastern margin of the inferred SJVL batholith, however, posing uncertainty about the possible presence of larger intrusions at greater depth than the exposed intracaldera plutons.

Ignimbrite events probably largely drain eruptible parts of a magma body, but the erupted magma may be a relatively thin upper zone of a vertically extensive and variably crystallized caldera-wide (perhaps composite) pluton. Depressurization accompanying the eruption would tend to promote crystallization in remaining portions, further decreasing the eruptibility of remaining magma. Figure 16 depicts the geometry of post-ignimbrite intrusions at Platoro caldera; the schematic extent of a possible caldera-wide composite intrusion is indicated by a red dashed line. Alternatively, if little or no intrusive remnants of the ignimbrite magmas remain as deeper intrusions and the exposed postcollapse intrusions do not widen with depth, the lower structure of the caldera complex would be dominated by the successive intracaldera accumulations of the six large Treasure Mountain ignimbrites, with a composite thickness previously estimated at 10 km or more (Lipman et al., 1996). Comparable composite fill thicknesses should initially have been present at other large multicyclic ignimbrite calderas, but at well-documented sections through volcano-plutonic assemblages elsewhere, thick caldera-related volcanic strata are truncated by underlying batholithic-scale intrusions (Lipman, 1984, 2007).

Comparison with Other Postcaldera Intrusions

Volcano-plutonic relations at Platoro are similar to those of several other mid-Cenozoic magmatic systems, based on limited available data. As at the Platoro caldera complex, plutons associated with the mid-Cenozoic Stillwater–Clan Alpine calderas in central Nevada (USA) postdate caldera-forming eruptions. Most of the intrusions have been interpreted to be subsequent magmatic pulses rather than the solidified remnants of ignimbrite magma chambers (Colgan et al., 2018). However, two small intrusions exhibit characteristics of residual magma based on zircon trace element chemistry, much like the Cornwall Mountain quartz monzonite porphyry and the Chiquito Peak Tuff in the Platoro system. The intrusions associated with the Caetano Tuff, a crystal-rich low-silica rhyolite in north-central Nevada, by contrast, show evidence of being genetically equivalent to the least-evolved part of the tuff, representing residual cumulate material (Watts et al., 2016).

Within the SRMVF, studies of the volcanic-plutonic connection have shown that most shallow-crustal intrusive activity is the result of distinct magmatic pulses and not the solidification of residual ignimbrite magma (Tappa et al., 2011; Zimmerer and McIntosh, 2012; Mills and Coleman, 2013). For example, the Questa-Latir magmatic system has a similarly protracted magmatic history (~7 m.y.) and number of exposed plutons as the Platoro system. Numerous intrusions associated with the Questa-Latir volcanic system also postdate the caldera-forming eruption of the ca. 25.4 Ma Amalia Tuff; only the peralkaline granite pluton of Virgin Canyon and small-volume dikes represent possible remnants of Amalia magma (Lipman, 1988; Johnson et al., 1989; Tappa et al., 2011; Zimmerer and McIntosh, 2012).

CONCLUSIONS

The evolution of the magmatic system that sourced the Chiquito Peak Tuff, postcaldera intrusions, and lavas of the Platoro caldera complex is summarized schematically in Figure 16. Isotopic studies (Lipman et al., 1978; Riciputi and Johnson, 1990; Riciputi et al., 1995; this study) support an origin from mantle-derived basaltic melts that assimilated lower crust for the magmas in the Platoro magmatic system. The bulk compositional similarities among the larger Platoro intrusions are consistent with the chemical character of the Platoro system having been established in the lower to mid-crust, as also inferred for other continental-arc systems (Hildreth and Moorbath, 1988; Annen et al., 2006, 2015; Jagoutz, 2010; Clemens et al., 2010; Solano et al., 2012), while the diverse textures, mineralogy, and mineral compositions of the magmas reflect the crystallization paths of distinct magma batches in the upper crust.

Eruption of the Chiquito Peak magma (Fig. 16A), ca. 28.8 Ma, was followed by eruption of the dacite of Fisher Gulch and Summitville Andesite. Emplacement of the Cornwall Mountain quartz monzonite porphyry, the Alamosa River and Jasper monzonites, and the intracaldera hornblende andesite-dacite dikes into the Chiquito Peak Tuff and Summitville Andesite (Fig. 16B) followed within analytical uncertainty of the date of the tuff eruption. Intrusive magmatism on the eastern side of the caldera is documented by the ca. 28 Ma Cat Creek monzonite, which cored the stratovolcano that sourced the volcanics of Green Ridge. On the western side of the caldera, intrusions are ca. 27.7–26.3 Ma and show evidence of open-system processes. Although pressure constraints are not available for these intrusions, characteristics of the Platoro rocks are consistent with open-system, polybaric fractionation of the associated magmas, the younger intrusions west of the caldera being derived from wetter parts of the system. The ca. 27.4 Ma Alum Creek monzonite porphyry was emplaced into the Alamosa River monzonite. The youngest granitoid intrusion by far, the ca. 24 Ma Summitville quartz monzonite as well as the dacite of South Mountain, document the protracted nature of magmatic activity and magmatic- hydrothermal mineralization at Platoro. In fact, all magmatic- hydrothermal mineralization associated with the Platoro caldera complex intrusions postdates ignimbrite eruptions and caldera collapse.

Map relations indicate that the intracaldera intrusions postdate the last-erupted ignimbrite, while our geochronologic data document a multi-million-year emplacement history but cannot quantify the intervals between eruption and pluton emplacement or between successive intrusions. With the possible exception of the Cornwall Mountain quartz monzonite porphyry and the dacite of Fisher Gulch, the intrusions are not remnants of the magma reservoir that sourced the Chiquito Peak Tuff but likely constitute distinct post-ignimbrite pulses of magma derived from the waning Platoro magmatic system.

ACKNOWLEDGMENTS

We thank Julien Allaz and Peter Horvath for help with the microprobe data acquisition and Heather Lowers and David Adams for assistance with SEM analysis. Jeremy Havens is thanked for his assistance in drafting the figures. We also thank Ryan Mills, Edward du Bray, Drew Downs, Christine Chan, two anonymous reviewers, and Science Editor Shanaka de Silva for constructive reviews of this manuscript. Amy Gilmer, Ren Thompson, and Peter Lipman were supported by the U.S. Geological Survey's National Cooperative Geologic Mapping Program, and Amy Gilmer was also supported by the U.S. Geological Survey's Mineral Resources Program. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

1Supplemental Material. Table S1: Whole-rock compositions of analyzed samples. Table S2: Major and trace element geochemistry of feldspar. Table S3: Major and trace element geochemistry of pyroxene. Table S4: Major and trace element geochemistry of biotite. Table S5: Major and trace element geochemistry of amphibole. Table S6: Zircon geochronology and trace element geochemistry. Table S7: Lutetium and hafnium isotopic compositions of zircon. Table S8: Amphibole-plagioclase thermometry. Table S9: Sample locations and lithologies. Please visit https://doi.org/10.1130/GEOS.S.13929935 to access the supplemental material, and contact editing@geosociety.org with any questions.
Science Editor: Shanaka de Silva
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.