Pleistocene hydrothermal activity on Brokeoff volcano and in the Maidu volcanic center, Lassen Peak area, northeast California: Evolution of magmatic-hydrothermal systems on stratovolcanoes

Hydrothermal activity on stratovolcanoes in magmatic arcs is a manifestation of shallow emplacement and degassing of hydrous magma beneath these volcanoes (e.g., Henley and Ellis, 1983; Rye, 1993; Giggenbach, 1997). Volcanic hydrothermal activity forms when hot gases rise and condense in shallow ground water and flow through and react with volcanic and underlying basement rocks below volcanic edifices. The resulting hydrothermal alteration resembles shallow-level alteration associated with porphyry copper and epithermal gold-silver mineralization (e.g., Sillitoe, 1973, 1975, 2010; Sillitoe and Bonham, 1984; Hedenquist et al., 1993, 2018; Hedenquist and Lowenstern, 1994; Proffett, 2003; Simmons et al., 2005). This hydrothermal alteration also can decrease rock strength and edifice stability, which leads to potentially catastrophic debris flows (Lopez and Williams, 1993), as observed at Mount Rainier, Washington (USA) (Vallance and Scott, 1997; Finn et al., 2001; Reid et al., 2001; Zimbelman et al., 2005; John et al., 2008). However, detailed studies of hydrothermal alteration on stratovolcanoes remain uncommon. Most studies are examinations of surficial features, such as recent ash deposits, fumaroles, and crater lakes, on active volcanoes (e.g., Hedenquist et al., 1993; Ohba and Kitade, 2005; Minami et al., 2016; Caudron et al., 2017; Lowenstern et al., 2018; Takahashi and Yahata, 2018) or drill-hole studies of geothermal systems on the margins of active stratovolcanoes (e.g., Reyes, 1990; Reyes et al., 1993; Rae et al., 2003; Moore et al., 2008). Other studies examine hydrothermal alteration associated with epithermal and porphyry mineral deposits in more deeply eroded stratovolcanoes and subvolcanic intrusions where it is difficult to link volcanic and hydrothermal activity (e.g., Gustafson and Hunt, 1975; Hedenquist et al., 1998, 2018; Longo et al., 2010). Few studies directly link mineralization temporally and stratigraphically to stratovolcano evolution (Sillitoe, 1975).

Brokeoff volcano is a partly eroded, late Pleistocene andesitic stratovolcano in the Lassen volcanic center at the south end of the active Cascades arc in northeastern California (Figs. 1 and 2; Clynne and Muffler, 2010). It lies on the south side of Lassen Peak, a large, 27 ka dacite dome that last erupted in 1914–1917 (Turrin et al., 1998). Much of the interior of Brokeoff volcano is hydrothermally altered and exposed by glaciers and in altered headwalls of landslides. However, the stratigraphy and eruptive history of Brokeoff volcano are relatively simple and well defined by the remaining, readily accessible exposures (Figs. 3A and 4; Clynne and Muffler, 2010). Alteration of Brokeoff volcano is the result of superposition of multiple hydrothermal systems, including an evolving Pleistocene (ca. 410–300 ka) system and Lassen’s well-known active hydrothermal system, which is manifest by hot springs, mudpots, fumaroles, and steaming ground (Day and Allen, 1925; Anderson, 1935; Muffler et al., 1982, 1983; Clynne et al., 2003; Crowley et al., 2004; Janik and McLaren, 2010; Ingebritsen et al., 2016).

In this paper, we present new maps of hydrothermally altered rocks and hydrothermal minerals on Brokeoff volcano and characterize hydrothermal alteration in the interior of the upper kilometer of the volcano. We mapped the distribution of hydrothermal minerals, and we use these, together with ⁴⁰Ar/³⁹Ar ages, to define a series of Pleistocene hydrothermal alteration zones. Alteration mineralogy and stable-isotope data are used to infer conditions of hydrothermal alteration (temperature, acidity, and fluid composition) and fluid sources. The presence of multiple hydrothermal alteration zones and hydrothermal fluid sources is used to construct cross-sectional time slices showing development of hydrothermal systems on Brokeoff volcano as the Lassen volcanic center evolved. We also studied the Maidu volcanic center located ∼10 km from Brokeoff volcano (John et al., 2005), and we report new ⁴⁰Ar/³⁹Ar dates and stable-isotope data for early Pleistocene (ca. 1.5 Ma) magmatic-hydrothermal alteration, which we compare to Brokeoff and other stratovolcanoes. Finally, we discuss the possibility of ore deposits at Brokeoff or Maidu.

Brokeoff volcano lies at the south end of the active Cascades volcanic arc that has formed in response to subduction of the Gorda and Juan de Fuca plates beneath North America (Fig. 1; Hildreth, 2007). Brokeoff volcano is part of the Lassen volcanic center, the youngest of five long-lived (∼1 m.y.) calc-alkaline andesite-dacite volcanic centers formed since ca. 3.5 Ma that developed on a regional platform of coalescing basalt to andesite volcanoes (Fig. 2; Clynne, 1990; Clynne and Muffler, 2010; Muffler and Clynne, 2015). These volcanic centers generally consist of a central andesitic stratocone with younger flanking lava flows and domes with more silicic compositions. The other four volcanic centers, Latour, Yana, Dittmar, and Maidu, are extinct and deeply dissected. Unexposed basement rocks in the Lassen area are likely Cretaceous granitic rocks emplaced into accreted oceanic terranes (Jachens and Saltus, 1983; Berge and Stauber, 1987).

The active Lassen volcanic center was formed by three stages of magmatic activity, the Rockland caldera complex (ca. 825–600 ka), Brokeoff volcano (ca. 555–360 ka), and the Lassen dome field (ca. 300–0 ka) (Clynne and Muffler, 2010; Muffler and Clynne, 2015). The Rockland caldera complex consists of the rhyolitic Rockland tephra and a series of dacite to rhyolite lava domes that range in age from ca. 825–600 ka. This stage culminated with eruption of the ∼50 km³ Rockland tephra at 609 ± 7 ka that is inferred to have formed a small caldera subsequently filled by deposits from Brokeoff volcano.

Brokeoff volcano is an ∼80 km³, andesitic stratovolcano composed of two major sequences of deposits, the Mill Canyon and Diller sequences (Figs. 3A and 4; Clynne, 1990; Clynne and Muffler, 2010). The Mill Canyon sequence (ca. 555–445 ka; Fig. 4 and Table 1) consists of dozens of lava flows and intercalated breccias erupted from a central vent(s). The sequence is dominated by packages of thin andesite lava flows and breccias of a single lithology, although a wide variety of porphyritic olivine and pyroxene basaltic andesites and andesites, and pyroxene-hornblende dacites are present in adjacent packages. In the study area, the Mill Canyon sequence consists mostly of the andesite of Mill Canyon, which is locally overlain by thick lava flows of the dacite of Twin Meadows that forms most of the upper part of Brokeoff Mountain. The Diller sequence (ca. 445–360 ka; Fig. 4 and Table 1) overlies the Mill Canyon sequence and consists primarily of thick, large-volume, two-pyroxene andesite and silicic-andesite flows that erupted from flank vents. An east-trending series of andesite dikes (ca. 500–480 ka; Table 1) and small plugs (ca. 460–430 ka; Table 1) intrude the core of Brokeoff volcano and were emplaced into older Mill Canyon sequence rocks (Fig. 3A).

The ∼50 km³ Lassen domefield, which consists of central dacitic domes of the Bumpass (ca. 300–190 ka) and Eagle Peak (ca. 120–0 ka) sequences and the flanking Twin Lakes sequence of hybrid andesite flows, was erupted on the north flank of Brokeoff volcano (Figs. 3A and 4; Clynne and Muffler, 2010; Muffler and Clynne, 2015). Lassen Peak, which last erupted in 1914–1917, is a 27,000-yr-old dacite dome of the Eagle Peak sequence that intrudes the northeastern part of Brokeoff volcano. Current hydrothermal activity is driven by cooling of a small magma body located >7 km beneath the youngest part of the Lassen domefield (Janik and McLaren, 2010).

Lava flows and interbedded volcaniclastic rocks in the Mill Canyon sequence generally have shallow dips (≤25°). They were fed through central vents whose location changed through time (Clynne and Muffler, 2010). One vent may be filled by a large intrusion just west of Sulphur Works (Figs. 3A and 5J). Thick lava flows of the overlying Diller sequence erupted from flank vents and dip gently outward from the eroded core of Brokeoff volcano (Figs. 5A and 5D; Clynne and Muffler, 2010). The central part of the Mill Canyon sequence is intruded by a series of dikes, elongate intrusions, and small plugs (Fig. 3A). The dikes and elongate intrusions mostly strike east-northeast to west-northwest and are parallel to fractures that formed elongate “ledges” of advanced argillic and residual quartz alteration and scarce, narrow quartz veins (Figs. 5E and 5K). Few faults have been mapped in Brokeoff volcano (Clynne and Muffler, 2010). Little Hot Springs Valley and West Sulphur Creek may have formed along north-striking structures in underlying pre–Brokeoff volcano rocks that channeled upflow of hydrothermal fluids (Rose et al., 1994), but there are no mapped faults of this orientation in the Mill Canyon sequence.

Hydrothermally altered rocks and hydrothermal features were mapped in an ∼10 km² area in the core of Brokeoff volcano at scales of 1:12,000 and 1:6000 (Fig. 3B). Alteration mapping was guided in part by hydrothermal mineralogy interpreted from remote sensing data (Crowley et al., 2004). Alteration was divided into six generalized types that were distinguishable in the field (Fig. 3B and Table 2). Active hydrothermal features, including fumaroles, mudpots, and hot springs, were surveyed during the summer of 2005 (Fig. 3B). Approximately 700 rock samples were collected for mineral identification and paragenetic and geochemical studies. Hydrous silicate, sulfate, and carbonate minerals in ∼550 samples were identified using a portable shortwave infrared (SWIR) spectrometer (PIMA SP) (Table S1 in the Supplemental Files¹). For alunite-bearing samples, SWIR data were used to estimate bulk alunite K-Na compositions from position of the 1.48 µm OH-absorption feature (Thompson et al., 1999; Chang et al., 2011). For smectite- and illite-bearing samples that lacked kaolinite, position and depth of the 1.9 µm H₂O and 2.2 µm Al-OH absorption features were used to estimate Al-(Fe-Mg) composition and illite crystallinity (Thompson et al., 1999; Chang et al., 2011; Halley et al., 2015). Subsets of the rock samples were characterized in more detail by optical microscopy, scanning electron microscopy with energy dispersive spectroscopy (SEM-EDS; Table 3), X-ray diffraction (XRD; Table S2 [^{footnote 1}]), and whole-rock geochemistry (Table 4 and Table S4). Maps of hydrothermal mineral distribution were used to refine the generalized, field-based alteration maps and to help define hydrothermal alteration zones (see below; Figs. 3C–3E). Groundmass separates of unaltered rocks and alunite separates were dated by ⁴⁰Ar/³⁹Ar methods to help refine the eruptive and hydrothermal histories of Brokeoff and Maidu (Table 1 and Table S3). Isotopic analyses of sulfur, oxygen, and hydrogen were made of mineral separates from selected samples from both volcanoes to help determine hydrothermal fluid sources and alteration temperatures (Table 1, Table S4 and Fig. S1). More complete descriptions of analytical techniques are presented in the Appendix and John et al. (2019).

There is no previous detailed mapping of hydrothermal alteration or characterization of the ancient hydrothermal systems in Brokeoff volcano. Previous studies focused primarily on hydrothermal features and alteration related to Lassen’s active hydrothermal system, including characterization of Bumpass Hell, Sulphur Works, Little Hot Springs Valley, and Morgan and Growler Hot Springs (Figs. 2 and 3E; Day and Allen, 1925; Anderson, 1935; Muffler et al., 1982, 1983; Thompson, 1985; Ingebritsen and Sorey, 1985; Sorey and Ingebritsen, 1995; Janik and McLaren, 2010; Ingebritsen et al., 2016; McHenry et al., 2017). Rose et al. (1994) conducted an oxygen-isotope study of Brokeoff volcano mostly using weakly altered volcanic rock samples and showed whole-rock δ¹⁸O depletion relative to magmatic values in broad areas outside the areas affected by the modern hydrothermal system. Crowley et al. (2004) interpreted airborne visible/infrared imaging spectrometer (AVIRIS) hyperspectral data that covered much of Brokeoff volcano and presented maps showing the distribution of several hydrothermal minerals. John et al. (2005) described Pleistocene hydrothermal alteration in the nearby Maidu volcanic center and presented a model for its formation.

The mineralogy and intensity of hydrothermal alteration vary extensively, reflecting wide ranges in hydrothermal fluid composition and physical conditions of alteration and local superposition of different types of alteration, including common overprinting of ancient hydrothermal alteration by Lassen’s modern hydrothermal system. Alteration is commonly partly developed, and the interiors of massive lava flows and clasts in volcaniclastic rocks are commonly unaltered. Hydrothermally altered rocks related to ancient hydrothermal activity are commonly weathered and have undergone supergene oxidation of pyrite, resulting in widespread formation of Fe-oxide and hydroxide minerals and jarosite and ferricrete in several small drainages (Fig. 3E).

Six types of hypogene alteration were initially identified in the field, and the mineralogy of each was confirmed and modified by lab analyses (Fig. 3B and Table 2). Five alteration types are inferred to be related to Pleistocene hydrothermal activity: silicic, advanced argillic, intermediate argillic, propylitic, and zeolitic. A sixth type is actively forming and consists of steam-heated intermediate and advanced argillic alteration as described below. Figures 3C to 3E show the distribution of hydrothermal minerals as mapped and determined by petrographic, SWIR, XRD, and SEM-EDS analyses. These data were used to define four major zones (Little Hot Springs Valley, Mount Diller, Supans Springs, and West Sulphur Creek) and four smaller zones (Brokeoff Mountain, East Ridge Lakes, Pilot Pinnacle, and West Ridge Lakes) of Pleistocene hydrothermal alteration (Figs. 3C and 3D and Table 5).

Hypogene alteration types are defined here as groups of hydrothermal minerals that occur together in the same sample and appear to have formed at the same time (Table 2). We use the term “alteration association” for common groupings of minerals independent of thermodynamic equilibrium constraints. As noted in the alteration descriptions below, there is considerable overlap and variation in mineralogy among these associations, reflecting variable conditions during the evolution of hydrothermal activity and its overprinting effects. Major types of alteration described below generally follow the nomenclature of Meyer and Hemley (1967) and Simmons et al. (2005) as used in studies of hydrothermal alteration on Mount Rainier (John et al., 2008) and at Maidu (John et al., 2005).

Areas mapped as advanced argillic alteration consist of intensely altered rocks composed of mixtures of alunite, kaolinite, dickite, pyrophyllite, quartz and/or other silica phases, topaz, pyrite/marcasite, and TiO₂ minerals (anatase or rutile). Barite is present commonly in trace amounts (Fig. 6F). This alteration generally is texturally destructive with plagioclase and pyroxene phenocrysts and rock groundmasses completely replaced by hydrothermal minerals that are commonly ≤5–10 µm in maximum dimension. Alunite crystals that replace plagioclase phenocrysts or fill vugs commonly are coarser grained but seldom exceed 100 µm in maximum dimension (Fig. 6A). Alunite crystals replacing groundmass commonly have pseudocubic forms (Fig. 6C), whereas alunite filling open spaces forms prismatic crystals (Figs. 6A and 6B). Secondary Fe-oxide and hydroxide minerals and/or jarosite resulting from supergene oxidation of pyrite and marcasite are abundant in many exposures.

Advanced argillic alteration mineral associations commonly vary on a hand-specimen scale and are seldom recognizable due to the extremely fine grain size. Common associations include alunite, alunite-kaolinite, alunite-pyrophyllite, and kaolinite and/or dickite (all plus silica and TiO₂ phases and pyrite; Fig. 3C and Table 2). Alunite compositions estimated from SWIR data and limited SEM-EDS and XRD data vary from nearly pure K-alunite to natroalunite, although SEM-EDS analyses indicate considerable compositional variation within single crystals (Fig. 6C). Huangite (Ca-alunite) was identified in a few locations from SWIR analysis. Coarser alunite crystals locally contain interior zones with high P, Ca, Sr, and Ba contents (aluminum-phosphate-sulfate minerals [APS]; Fig. 6B and Table 3). Alunite in several samples analyzed by SEM-EDS contains up to ∼5 wt% fluorine (Table 3), and fluorine-rich, fine-grained topaz was identified by SWIR and SEM analyses in a few samples associated with quartz and alunite ± dickite (Fig. 6E).

Pyrophyllite ± alunite is present in several samples collected west of Sulphur Works, mostly at relatively high elevations in the western part of Supans Springs alteration zone, where it occurs with both quartz and opaline silica (Fig. 6D). Sparse pyrophyllite + alunite occurs at several other locations (Fig. 3C). Dickite is common, especially in the Supans Springs intrusion just west of Sulphur Works and in the Mount Diller system (Fig. 3C). Dickite + alunite occurs uncommonly in the Supans Springs and East Ridge Lake systems.

Most silicic alteration is residual quartz, which is a vuggy-textured rock composed almost entirely of fine-grained quartz. Other silica phases (opal-CT and cristobalite) are present in some samples and likely reflect lower-temperature deposition compared to quartz. In partly oxidized rocks, minor very fine-grained pyrite commonly is encapsulated in quartz. Residual quartz alteration mostly occurs in narrow, linear breccia zones that form low, resistant ridges (ledges) within zones of alunite-rich advanced argillic alteration (Figs. 3B and 5K). Silicification is uncommon and mostly consists of narrow (<1–2-m-wide) zones where rock has been replaced by dense fine-grained quartz with minor pyrite and is enclosed in advanced argillic alteration. Silicification is mostly exposed in the Mount Diller zone in upper West Sulphur Creek but is not shown in Figure 3B.

Andesites in the Mill Canyon sequence in the lower parts of West Sulphur Creek and Little Hot Springs Valley are extensively altered to mixtures of chlorite, mixed layer chlorite-smectite, illite or mixed layer illite-smectite, calcite, quartz, albite, pyrite, hematite, and minor K-feldspar that is classified as propylitic alteration (Figs. 3B and 5A; Table 2). This alteration generally is moderately intense with partial replacement of plagioclase phenocrysts by illite or smectite, calcite, chlorite, quartz, albite and/or kaolinite, complete replacement of pyroxenes by chlorite ± calcite + opaque minerals, and variably altered glassy groundmasses that are replaced by illite and/or smectite, albite, chlorite, quartz, calcite, and pyrite (Fig. 6J and Table 2). Primary rock textures are well preserved, and alteration is more intense in breccia units, especially in breccia matrices, than in massive lava flows. Quartz, chalcedonic quartz, calcite, chlorite, pyrite, and/or hematite locally fill vesicles and narrow (0.1–5-cm-wide), generally widely spaced veins (Figs. 5E and 6I). Minor barite and anhydrite are present locally. Propylitic alteration is gradational upward into smectite-rich intermediate argillic alteration that lacks chlorite and calcite. X-ray diffraction data indicate that smectite and chlorite are interlayered locally (Table S2 [^{footnote 1}]; Rose et al., 1994; Table 1). Chlorite contents in five samples calculated from SEM-EDS analyses using the method of Rose et al. (1994) range from ∼50%–80% and have Mg/(Mg+Fe) values of 0.58–0.74 (Table 3) that are similar to microprobe compositions reported by Rose et al. (1994). Epidote has not been identified.

Intermediate argillic alteration consists mostly of smectite + pyrite and forms most of the weak- to moderate-intensity alteration in Brokeoff volcano (Fig. 3B). Based on XRD, SEM-EDS, and SWIR analyses, smectite compositions range widely in K, Ca, Mg, Fe, Al, and H₂O contents (Table 3 and Table S1 [^{footnote 1}]). Smectite, as used in this paper, refers to clay minerals with high H₂O content by SWIR analysis and/or low K, Fe, and Mg contents and six non-interlayer cations per 11 oxygens in normalized SEM-EDS analyses. Clay that fills vesicles in sample 04-LP-144 has high Fe and Mg and low Al contents consistent with saponite (Table 3). Weakly altered rocks preserve primary rock textures and have partial to complete replacement of pyroxene phenocrysts by smectite and silica phases, partial to complete replacement of groundmass glass and microphenocrysts by smectite, silica phases, and pyrite, generally weak replacement of plagioclase phenocrysts by smectite, silica phases, and locally K-feldspar, and pyrite replacement of magnetite (Figs. 6G and 6H; Table 2). Pyrite and/or smectite commonly partly fill vesicles. Barite and anhydrite are present locally in trace amounts. In areas of more intense alteration, notably where this alteration fringes advanced argillic alteration, primary rock textures are destroyed, and the rock is replaced by quartz, smectite, mixed-layer illite/smectite, or illite, and pyrite (trace to 10 vol%). Where kaolinite is present, the alteration is gradational to kaolinite-rich advanced argillic.

A cream-colored zeolite mineral, probably Ca-clinoptilolite, forms tabular crystals as much as 0.3 mm long that coat fractures, fill vesicles, and locally replace the groundmass of lava flows of the Diller sequence along the ridgeline ∼500 m west of Bumpass Hell (Fig. 3D). The zeolite is locally intergrown with fine-grained adularia (Fig. 6K). This alteration forms the uppermost preserved part of the Little Hot Springs Valley alteration zone and is superimposed on weak intermediate argillic (smectite-pyrite) alteration. Semiquantitative chemical compositions from SEM-EDS analyses (Table 3) are similar to clinoptilolites in the ancient Ngakuru geothermal system, New Zealand (Braithwaite, 2003).

Lassen’s modern hydrothermal activity produces steam-heated alteration from acid-sulfate fluids formed where ascending H₂S-rich steam condenses in shallow, oxygenated ground water (e.g., Day and Allen, 1925; Muffler et al., 1982; Clynne et al., 2003; Janik and McLaren, 2010; McHenry et al., 2017). This alteration consists mostly of mixtures of smectite, kaolinite, alunite, opal-A, and cristobalite with accessory iron and aluminum sulfate minerals above the water table and pyrite below the water table. Bicarbonate waters resulting from condensation of CO₂-rich steam locally form along the margins of the acid-sulfate fluids and form warm springs that deposit travertine. Much of this steam-heated alteration is superimposed on previously altered rocks, some of which are in landslides. Distinguishing modern steam-heated alteration from similar older alteration is difficult in some locations. In general, modern steam-heated alteration was mapped in (1) areas containing active or recently active hydrothermal vents and (2) areas composed mostly of bleached, oxidized, highly friable rocks contiguous to active hydrothermal vents (Figs. 3B, 3E, 5A–5C, and 5F).

We define four major (Supans Springs, Mount Diller, Little Hot Springs Valley, and West Sulphur Creek) and several smaller zones of Pleistocene hydrothermal alteration (Table 5), based on the continuity of alteration, zonation of hydrothermal minerals, relative age provided by stratigraphic relations, and by alunite ⁴⁰Ar/³⁹Ar dates from Supans Springs zone (Figs. 3B–3D and Table 1).

This zone is exposed between Sulphur Works and the east side Brokeoff Mountain (Figs. 3B and 3C). Alteration is developed mostly in breccia units in the Mill Canyon sequence and in several small andesite plugs and dikes that intrude the breccias. The largest plug, the Supans Springs intrusion, may have been emplaced into a central conduit for early eruptions of Brokeoff volcano (Fig. 5J). Supans Springs is characterized by inner areas of advanced argillic alteration (commonly alunite rich) with smaller areas of residual quartz (silicic) alteration. Narrow areas of weak to intense intermediate argillic alteration (mostly quartz-illite or quartz-smectite) border the central zones of silicic and advanced argillic alteration. A large area of generally weak intermediate argillic alteration cuts across the center and southwest side of the Supans Springs intrusion. This alteration may have formed by previously partially neutralized fluids in weakly fractured, less permeable parts of the Supans Springs intrusion. Alternatively, this alteration might be related to the West Sulphur Creek zone (see below). Silica phases are zoned vertically with quartz present in deeper eastern exposures and opaline silica present in shallower western exposures.

The central alteration zones appear structurally controlled by west- to northwest-striking fractures extending west from the Supans Springs intrusion and their intersections with breccia units in the Mill Canyon sequence. Fractures are marked by narrow (mostly ∼5–20-m-wide), low (few m high), west- and northwest–trending ridges “ledges” of residual quartz and quartz-alunite alteration up to 500 m in strike length (Figs. 3B and 5K). Hydrothermal breccias are common in these ledges. Pyrite in surface exposures is mostly oxidized, and breccias in the residual quartz ledges commonly contain abundant hematite, goethite, and/or jarosite. Unoxidized pyrite generally is limited to micron-size crystals encapsulated in silica phases, although pyrite-rich intermediate argillic alteration is locally preserved in more deeply eroded gullies beside the ledges.

Field relations with dated units and ⁴⁰Ar/³⁹Ar dates of alunite from two locations indicate that the Supans Springs hydrothermal alteration formed at ca. 420–405 ka following eruption of the Mill Canyon sequence and during initial eruptions of the Diller sequence (Table 1). The undated but altered Supans Springs intrusion and/or similar shallowly emplaced magma probably were a major source of heat and volatiles (H₂O, SO₂, H₂S, HCl, and HF).

This zone is exposed in upper West Sulphur Creek just south of Pilot Pinnacle on the east and southeast flanks of Mount Diller (Fig. 3B). It is characterized by an intensely altered core of advanced argillic alteration composed mostly of Fe-sulfide-rich quartz-dickite and quartz-kaolinite locally enveloped by intense to weak quartz-illite-pyrite and smectite-pyrite (Figs. 3B and 3C). Widespread smectite-pyrite alteration is cored by another small area of intense quartz-dickite-pyrite alteration at the head of Little Hot Springs Valley (Fig. 3B). The central parts of both areas were mostly removed by glaciers and landslides and are overprinted by the modern steam-heated hydrothermal activity. Rocks in both the Mill Canyon sequence and the overlying Diller sequence were affected by the Mount Diller alteration, whereas overlying rocks of the Bumpass sequence were not altered, thereby constraining the age of alteration to between ca. 360 and 300 ka (Table 1; Clynne and Muffler, 2010).

This ∼2.5 km² altered area is the largest in Brokeoff volcano and is well exposed in the 300-m-high cliffs on both sides of the Little Hot Springs Valley west of Bumpass Hell (Figs. 3B and 5A). This zone is separated from the Mount Diller zone by areas of unaltered rocks on its west side and by the modern hydrothermal activity on its north side. Little Hot Springs Valley consists of intermediate argillic (smectite-pyrite) alteration that grades downward into propylitic alteration and is locally overlain and overprinted by zeolite alteration (Figs. 3B and 6K). Deep, propylitically altered rocks are cut by sparse banded quartz-chalcedony-calcite-pyrite veins, and vesicles in these rocks are locally filled by calcite, quartz and/or chalcedony, hematite, and chlorite (Figs. 3D and 6I). Float blocks of banded opaline silica-quartz-pyrite and/or marcasite veins locally litter the uppermost exposures of the system west of Bumpass Hell but have not been found in outcrop. Little Hot Springs Valley alteration affected rocks in both the Diller and Mill Canyon sequences. Alteration is mostly confined to volcaniclastic breccia units and brecciated margins of lava flows, whereas dense interiors of flows are largely unaltered (Fig. 5A). Hydrothermal activity did not affect and therefore likely predates emplacement of the Bumpass sequence, and thus this alteration zone has the same age constraints as Mount Diller alteration zone (i.e., between ca. 360–300 ka).

This zone consists of propylitic alteration developed in andesite of Mill Canyon exposed on both sides of West Sulphur Creek south and southeast of Sulphur Works and the Supans Springs zone (Fig. 3D). Quartz-chalcedony-calcite-pyrite-hematite veins and vesicle fillings are locally present (Figs. 3D and 5E). Smectite-rich intermediate argillic alteration locally overlies the propylitic alteration on the east side of Brokeoff Mountain, although areas of poorly exposed rock separate the two alteration types. Generally weak intermediate argillic alteration that extends north into the Supans Springs intrusion also may be part of West Sulphur Creek. Alteration mostly affects breccia units, and intercalated lava flows are only weakly altered. The age of West Sulphur Creek is uncertain, but propylitic alteration is exposed at similar elevations in the deeper parts of the ca. 360–300 ka Little Hot Springs Valley zone.

A small area in the andesite of Mill Canyon and overlying dacite of Twin Meadows ∼500 m northeast of Ridge Lakes is strongly altered to residual quartz and advanced argillic alteration containing alunite, pyrophyllite, dickite, and kaolinite narrowly bordered by smectite-rich intermediate argillic alteration (Fig. 3C). Although alteration is similar to that in Supans Springs, a distinct phase of hydrothermal activity is suggested by stable-isotope data discussed below.

The altered area northwest of Ridge Lakes (Fig. 3C) consists of alunite- and kaolinite-bearing advanced argillic alteration containing opaline silica and smectite-rich intermediate argillic alteration in rocks of the Mill Canyon sequence. Alteration does not extend into overlying rocks of the Diller sequence suggesting that the alteration age is >ca. 400 ka.

Smectite-rich intermediate argillic alteration in breccia deposits overlain by advanced argillic alteration (kaolinite-opaline silica ± alunite) occurs on the top and upper east side of Brokeoff Mountain (Fig. 3C). Alteration is developed in breccias and lavas in the andesite of Mill Canyon, the dacite of Twin Meadows, and the andesite of Mount Diller, which forms the top of Brokeoff Mountain. Therefore, this alteration is <ca. 380 ka and younger than Supans Springs alteration.

Variable intensity, smectite-rich intermediate argillic alteration of the andesite of Mount Diller is exposed in the cliffs on the NE side of Pilot Pinnacle (Fig. 3D). Most of the alteration occurs within flow breccias and brecciated units.

Areas with active hydrothermal features and associated hydrothermal alteration in Brokeoff volcano include Bumpass Hell, Little Hot Springs Valley, Pilot Pinnacle, and Sulphur Works (Figs. 3B and 3E; Table 5).

Bumpass Hell consists of superheated steam vents and acid-sulfate hot springs that form zones of advanced argillic alteration surrounded by intense to weak intermediate argillic alteration (Muffler et al., 1983; Janik and McLaren, 2010; Muffler and Clynne, 2015; Figs. 3B and 3E). Alteration is centered in a shallow basin that surrounds a zone of active steam venting (Fig. 5F). The alteration affects dacite lavas and breccias of the Bumpass sequence and underlying lava flows of the andesite of Mount Diller. Major hydrothermal minerals include kaolinite, alunite, smectite, cristobalite, opal-A, pyrite, ferric and ferrous hydroxides and sulfate minerals, and aluminum sulfates (Fig. 3E; Lee, 2008; McHenry et al., 2017). Alteration is pervasive, and most of the original host rock is completely altered with few relict textures (quartz phenocrysts in rocks of the Bumpass sequence are the notable exception). Alunite is the dominant mineral in the slopes surrounding Bumpass Hell, whereas kaolinite and opal mixtures are dominant in the central zone of hot spring vents (Fig. 3E). Potassium-rich alunite occurs in the center of Bumpass Hell, whereas more Na-rich compositions occur distally. Native sulfur and sulfate minerals occur near the active steam regions. Smectite is generally abundant in the slopes surrounding Bumpass Hell (Fig. 3E). Some of the smectite may be related to the Pleistocene hydrothermal activity.

This zone in the center of eroded Brokeoff volcano contains hydrothermal alteration related to both active and ancient hydrothermal activity. The north end of the valley encompasses an area that has both active and inactive features marked by travertine deposits (Tr, Fig. 3E; Clynne and Muffler, 2010). Zones of steam-heated advanced argillic alteration occur west of the carbonate springs and extend to Highway 89. Active steam vents and boiling mudpots occur in the central part of the valley and in landslide and slump deposits on the southern slope of a small side valley ∼1 km southwest of Bumpass Hell (Figs. 3B, 5A, and 5B). Major hydrothermal minerals in this side valley are kaolinite, alunite, smectite, and opaline silica (Fig. 3E). SWIR data indicate that much of the kaolinite is associated with smectite. Crowley et al. (2004) noted this mixture and suggested that steam-heated kaolinite alteration overprints older smectite-rich alteration that may be related to Pleistocene hydrothermal activity. Within the central part of the valley, steam-heated advanced argillic alteration is confined to active steam vents and is surrounded by smectite-rich intermediate argillic alteration.

Areas of hydrothermally altered rock related to the active hydrothermal system extend southeast from Pilot Pinnacle to California State Highway 89 and are continuous into steam-heated alteration at the north end of Little Hot Springs Valley (Fig. 3E). Steam vents are locally precipitating native sulfur (Fig. 5C), and boiling mudpots occur in glacial, talus, and rockfall avalanche deposits at the base of Pilot Pinnacle.

This zone consists of steam-heated advanced argillic alteration associated with active fumaroles and boiling springs extending north from Sulphur Works toward Pilot Pinnacle (Fig. 3E). Two large (3–4-m-diameter) travertine deposits are associated with hot springs at the south end of Sulphur Works in West Sulphur Creek (Fig. 3E; Clynne et al., 2003). Advanced argillic alteration consists of kaolinite, alunite, opal, Al-sulfates, and anatase and/or rutile. Smectite, illite, and pyrite characterize intermediate argillic alteration that lies adjacent to and surrounding the advanced argillic alteration near active vents; however, much of this alteration probably is related to the Pleistocene hydrothermal activity.

The shallow parts of an early Pleistocene magmatic-hydrothermal system are exposed in an ∼3.5 km² area within the northeast part of the Maidu volcanic center, an eroded early Pleistocene andesite-dacite stratocone, ∼15 km in diameter (Figs. 2 and 7; Clynne, 1984; John et al., 2005; Muffler and Clynne, 2015). The central stratocone is composed of two sequences of basaltic andesite to dacite lava flows and breccias, Stages 1 and 2 (Fig. 7; Muffler and Clynne, 2015). Stage 1 andesites are a heterogeneous mix of lava flows, breccias, and minor pyroclastic rocks. Within the altered area, these rocks are mostly thin lava flows of platy-jointed pyroxene andesite and andesitic breccias, which include both homogeneous flow breccias and heterogeneous debris flows. Stage 2 rocks unconformably overlie hydrothermally altered Stage 1 rocks and consist mostly of thick flows of porphyritic pyroxene dacite. New ⁴⁰Ar/³⁹Ar dates for Stage 1 (pre-alteration) lava flows are 1819.5 ± 7.6 ka to 1785.3 ± 23.3 ka, and Stage 2 (post-alteration) lava flows have dates of 1501.5 ± 36.8 ka and 1490 ± 12.3 ka (Table 1). ⁴⁰Ar/³⁹Ar dates for three alunite samples from Maidu range from 1496 ± 6 ka to 1468 ± 27 ka and indicate hydrothermal activity just prior to eruption of Stage 2 lavas.

Hydrothermally altered rocks form a blanket-like deposit exposed over a vertical range of ∼300 m that extends downward from a paleosurface marked by silica-rich sedimentary deposits (Fig. 8; John et al., 2005). Alunite- and pyrite-rich hydrothermal alteration is zoned vertically from opal-kaolinite at the highest elevations, through quartz-pyrophyllite at intermediate elevations, to residual quartz and quartz-topaz at the lowest elevations. Hydrothermal features are zoned systematically downward from hydrothermal breccias and hydrothermal sediments, probably deposited in an acid lake or hydrothermal eruption crater that formed at or near the paleosurface, through quartz-pyrite veins, to residual quartz ledges formed at the deepest exposed levels.

Chemical analyses for select elements in 200 altered rock samples at Brokeoff volcano and 14 samples of chemical precipitates at Growler Hot Springs (Table 4 and Table S4 [^{footnote 1}]) are compared with similar data for Maidu (John et al., 2005). Elemental gains and losses are relative to unaltered protoliths (Clynne, 1984). Elements tabulated in Table 4 are commonly enriched in deep fluids in active magmatic-hydrothermal systems and in the shallow, mineralized parts of ancient magmatic-hydrothermal systems, including epithermal gold-silver deposits and the upper parts of porphyry copper deposits (e.g., Tosdal et al., 2009; Saunders et al., 2014; Halley et al., 2015; Sillitoe, 2015; John and Taylor, 2016; Simmons et al., 2016). In general, gold and silver contents are low in all hydrothermal centers and in all types of alteration (only 41 and six of 214 samples had detectable gold [≥5 ppb] or silver [≥1 ppm], respectively). Copper and zinc concentrations are generally less than common unaltered andesite values. Sulfur is the most strongly enriched element, ranging up to 13.5 wt% in pyrite-rich kaolinite and/or dickite altered breccias in the Mount Diller zone. Maximum concentrations of elements enriched in altered rocks at Brokeoff and Maidu generally are low compared to concentrations in most epithermal and porphyry deposits (e.g., Saunders et al., 2014; Halley et al., 2015).

Altered rocks locally contain elevated concentrations of many elements common in epithermal and porphyry deposits, including As, Bi, Hg, Mo, Pb, Sb, Se, and Te. Of note, Au (max. 15 ppb) and Ag (max. 4 ppm) concentrations generally are below detection limits. Most anomalous element concentrations are in alunite-bearing and other types of advanced argillic alteration, although Fe-oxide-rich gossans and ferricrete derived from weathering of altered rocks in this system locally contain high concentrations of As, Cu, Se, and Te.

Altered rocks have locally elevated concentrations of the same suite of elements as Supans Springs (As, Bi, Hg, Mo, Pb, Sb, Se, and Te; Table 4). Anomalous concentrations of these elements are mostly in pyrite-rich kaolinite and/or dickite (advanced argillic) and smectite/illite (intermediate argillic) alteration.

Altered rocks are enriched in few elements of this suite. However, approximately one-third of the samples from the Little Hot Springs Valley contain detectable Au (≥5 ppb), and the highest Au content measured at Brokeoff volcano (114 ppb) is from a propylitically altered sample. Mercury and Pb also are locally enriched in argillic alteration in Little Hot Springs Valley. Zinc is not present in anomalous concentrations, although trace amounts of sphalerite were found in a quartz vein in West Sulphur Creek.

Gold, As, and Hg are the only elements enriched in more than one sample. Four of five samples of advanced argillic alteration from the West Ridge Lakes have ≥5 ppb gold (maximum 14 ppb).

Mercury is the only element that is strongly enriched in altered rocks, notably in several samples from Sulphur Works. Although not analyzed, native sulfur is present at Sulphur Works, and small quantities were mined in the late nineteenth century (National Park Service, 2016).

Growler Hot Springs discharge deeply derived chloride water (Muffler et al., 1982; Janik and McLaren, 2010; Ingebritsen et al., 2016) and deposit silica sinter, black sulfide-rich mud, yellow and orange As-Sb sulfides, and distal silica-Fe-oxide cemented sediments. These precipitates are variably enriched in As, Cs, Li, Ni, Sb, Te, Tl, and W. Notably, they have generally low Au (≤31 ppb), Ag (<1 ppm), Cu, Hg, Mo, Pb, Se, and Zn contents.

Alteration at Bumpass Hell and Sulphur Works and Growler Hot Springs are likely related to the same hydrothermal system that now underlies Brokeoff volcano (Muffler et al., 1982; Janik and McLaren, 2010; Ingebritsen et al., 2016). Alteration at Bumpass Hell and Sulphur Works formed from steam-heated acid-sulfate waters, whereas siliceous deposits at Growler Hot Springs precipitated by cooling of the residual boiled chloride water.

Light stable-isotope (O, H, and S) compositions were measured for hydrothermal minerals from Brokeoff volcano and the Maidu volcanic center to help identify fluid sources, temperatures, and environments of hydrothermal alteration. The stable-isotope data for sulfur-bearing and hydrous minerals are summarized in Tables S1 and S5 (^{footnote 1}), respectively.

Pyrites and marcasites at Brokeoff volcano have δ³⁴S values ranging from −11.9‰ to +7.2‰, although most samples are ≤−2‰ (Fig. 9; Table S1 [^{footnote 1}]). Pyrite from Supans Springs spans the complete range of δ³⁴S values, whereas all pyrite from the Mount Diller, Little Hot Springs Valley, and East Ridge Lakes has δ³⁴S values <−2‰. Pyrite from active hydrothermal areas also has a much narrower range of δ³⁴S values from −4.4‰ to +1.3‰.

Pyrite from Maidu has a more restricted range of δ³⁴S values than Brokeoff samples, varying from −11.3‰ to −6.3 ‰ (Fig. 9).

Alunites at Brokeoff and Maidu have a broad range of S and O isotopic compositions, which indicate their formation in several hydrothermal environments (Fig. 10; Rye et al., 1992; Rye, 2005). Most Brokeoff alunites fall into three groups corresponding to magmatic-hydrothermal, “high elevation” magmatic-hydrothermal samples, and steam-heated environments. The largest group has moderate to high δ³⁴S and values typical of magmatic-hydrothermal alunite. The second group collected from high elevations on and just north of Brokeoff Mountain has high and moderate δ³⁴S values that are enriched in ⁠. A third group has low δ³⁴S and variable values that are characteristic of the steam-heated environment.

Alunites at Maidu have a relatively restricted S-O isotopic range compatible with a magmatic-hydrothermal origin for all samples.

Magmatic-hydrothermal alunite is derived from H₂SO₄ generated from the disproportionation of magmatic SO₂ during condensation of magmatic vapor at temperatures below 400 °C (Rye et al., 1992). A diagnostic feature of alunites formed in this environment is their coexistence with pyrite and δ³⁴S values that reflect sulfur-isotope equilibrium between oxidized and reduced aqueous sulfur species at the temperature of disproportionation. The bulk sulfur-isotope composition of fluids lies between δ³⁴S values of pyrite and alunite and closer to the δ³⁴S values of pyrite. Pyrite-alunite sulfur-isotope fractionation for magmatic-hydrothermal alunites yields temperatures ranging from 184° to 265 °C at Brokeoff and 180° to 258 °C at Maidu using the equilibrium fractionation factors of Ohmoto and Rye (1979). Unfortunately, we were able to analyze only a few pyrite-alunite pairs at Brokeoff volcano owing to nearly complete oxidation of pyrite in alunite-rich rocks during weathering.

Several alunite samples collected at higher elevations on Brokeoff Mountain and on the ridge between Brokeoff Mountain and Mount Diller have high values (15.7‰–19.7‰) relative to other analyzed alunites (Fig. 10). This alunite alteration possibly formed near fumarolic vents high on the Brokeoff volcano edifice. In this environment, alunite may have formed where H₂SO₄ formed by disproportionation of SO₂ mixed with isotopically heavy atmospheric oxygen (δ¹⁸O = 23‰) in open vents near the paleosurface of the Brokeoff volcano edifice. Sulfate minerals with similarly high values have been measured from other stratovolcanoes and were interpreted as the result of ¹⁸O enrichment by mixing with atmospheric oxygen (Zimbelman et al., 2005; John et al., 2008).

In the shallow, steam-heated environment, aqueous sulfate is derived from atmospheric or bacterial oxidation of H₂S at ≤100 °C. The δ³⁴S value of the sulfate in this environment is typically close to the value for H₂S and H₂S-derived pyrite, because exchange is slow at temperatures ≤200 °C. The values of steam-heated alunite may vary depending on the proportions of oxygen from air (δ¹⁸O = 23‰) and ¹⁸O-depleted meteoric water incorporated in the aqueous sulfate, and the latter process likely produced the alunites on Brokeoff volcano that have lower values than the steam-heated reference composition (Rye et al., 1992). At Brokeoff, values of steam-heated alunites systematically vary by location with the highest values in Bumpass Hell and Sulphur Works alunites and significantly lower values in north Bumpass Hell and Fart Gulch samples (Fig. 10).

The H and O isotopic data for alunite, kaolinite, dickite, pyrophyllite, chlorite, illite, and smectite for samples from Brokeoff and Maidu are plotted in Figure 11 and tabulated in Table S5 (^{footnote 1}). Small grain size and intimately intergrown phases inhibited mineral separations, and several analyzed samples are mixtures of hydrous minerals as identified by X-ray diffraction. The isotopic compositions of water in parent fluids calculated using estimated formation temperatures also are plotted. These parent waters are compared to the isotopic compositions of present-day meteoric waters in the Lassen Peak area and to waters discharged at Growler Hot Springs, which are interpreted as deeply derived, neutral-pH chloride waters (Muffler et al., 1982; Janik and McLaren, 2010; Ingebritsen et al., 2016).

Most calculated parent hydrothermal fluids form a linear trend of H and O isotopic composition of water (Fig. 11), similar to trends that have long been recognized in studies of hydrothermal systems and mineral deposits and that have been explained by mixing of waters from magmatic and meteoric sources (e.g., Taylor, 1974; Giggenbach, 1992). Simple mixtures contain 20%–90% magmatic water, which is considered too great given the shallow depth, low fluid salinities, and lack of large metal anomalies. Alternatively, in geothermal systems, linear trends of water compositions with a slope of 3 are produced by non-equilibrium boiling of meteoric waters (Craig, 1961). Additionally, a trend with a slope of 0 may be explained by partial oxygen-isotope exchange of end-member meteoric water with local volcanic wall rocks resulting in a δ¹⁸O shift of 6‰–11‰ prior to mixing with magmatic fluids (e.g., Truesdell and Hulston, 1980; Field and Fifarek, 1985). This wall-rock reaction is here considered to be the likely mechanism, and the resultant calculated oxygen-isotope shift in meteoric water composition is about twice the 3‰–4‰ oxygen shift in waters discharged by the active hydrothermal system at Morgan and Growler Hot Springs (Janik and McLaren, 2010; Ingebritsen et al., 2016) but is similar to hydrothermal water δ¹⁸O compositions estimated from weakly altered lava flows with the lowest δ¹⁸O whole-rock values reported by Rose et al. (1994; Fig. 11).

Several smectite-bearing clay samples from Brokeoff volcano have calculated parent water compositions with lower δ¹⁸O values that are close to measured values of the geothermal waters discharging at Morgan and Growler Hot Springs. This suggests that these clays formed from less highly exchanged meteoric water. The range of exchanged meteoric water compositions may be modeled using variable water/rock ratios and equations for water/rock exchange (Taylor, 1979; Field and Fifarek, 1985), using a present-day meteoric water composition (δ¹⁸O = −12.5‰, δD = −90‰) and unaltered Brokeoff andesite (δ¹⁸O = +7‰, δD = −30‰; Rose et al., 1994; Feeley et al., 2008), and modeling water/rock isotopic exchange by plagioclase (An₆₀)-H₂O and muscovite- and/or smectite-H₂O fractionation for oxygen and hydrogen isotopes, respectively. At 250 °C, exchanged meteoric water compositions closely match calculated parent water compositions for clay minerals with water/rock mass ratios ranging from ∼1–0.05 (dotted curve in Fig. 11A).

Parent fluids for alunites in the Supans Springs system have the largest magmatic-water component, whereas fluids forming smectite, chlorite, and kaolinite have larger meteoric-water components. Alunite from the East Ridge Lakes system (05-LP-48, Fig. 11A) has a significantly larger meteoric-water component than the Supans Springs fluids consistent with differences in alunite-pyrite δ³⁴S values. Parental fluid compositions calculated for alunites and two clay samples from the Maidu volcanic center have generally similar isotopic compositions that are consistent with formation from mixtures of exchanged meteoric and magmatic water.

Oxygen-isotope compositions for five quartz veins in the deeper parts of the West Sulphur Creek system range from +7‰ to +11‰ (Fig. 11B). All veins have two distinct textural forms of silica: relatively coarse-grained, crystalline quartz and massive, milky to opaline, fine-grained quartz and chalcedonic quartz. Coarsely crystalline quartz generally appears to have formed later than the massive, fine-grained quartz and has measured δ¹⁸O values up to 2.4‰ less than corresponding fine-grained quartz and chalcedony (Table S5 [^{footnote 1}]). Parental water δ¹⁸O compositions calculated at 200 °C are similar to waters calculated from clay minerals, suggesting that the veins dominantly formed from exchanged meteoric water. The variation of measured δ¹⁸O values within several veins suggests either crystallization over a range of temperatures or changing δ¹⁸O fluid composition.

The wide range of mineral associations and hydrothermal-alteration types identified on Brokeoff volcano and variations in associated stable-isotope data are indicative of several distinct hydrothermal environments. These environments are commonly recognized in active and ancient hydrothermal systems (magmatic-hydrothermal, geothermal, and steam-heated) elsewhere (e.g., Schoen et al., 1974; Henley and Ellis, 1983; Reyes, 1990; Rye et al., 1992; Sillitoe, 1993; Giggenbach, 1997; Hedenquist et al., 2000; Rye, 2005; Simmons et al., 2005; John et al., 2008; Sillitoe, 2015).

Magmatic-hydrothermal acid-sulfate waters (commonly vapor and brine phases) are exsolved from shallow crystallizing hydrous magma. The low-density vapor and high-density, saline brine phases commonly separate due to buoyancy and viscosity contrasts (e.g., Henley and McNabb, 1978). The magmatic vapor, which is rich in acid-forming SO₂, H₂S, HCl, HF, and CO₂, rises buoyantly through the volcano. If the vapor condenses into shallow ground water, SO₂ will disproportionate at temperatures below 400 °C, forming H₂SO₄ and H₂S (4SO₂ + 4H₂O = 3H₂SO₄ + H₂S; e.g., Holland, 1965; Rye et al., 1992; Hedenquist and Taran, 2013). H₂SO₄, HCl, and HF in the resultant acidic fluid dissociate as the fluid cools. Hydrolysis reactions with wall rocks at high water/rock ratios (“fluid dominated”) progressively neutralize the fluid acidity, leach soluble ions, and form hydrothermal minerals that may include alunite, pyrophyllite, dickite, kaolinite, topaz, illite, quartz, anhydrite, and pyrite, whereas smectite may form at lower temperatures distal to fluid upflow zones. Supans Springs, Mount Diller, and Maidu exemplify hydrothermal alteration produced by magmatic-hydrothermal fluids.

Where H₂SO₄-rich magmatic-hydrothermal fluid mixes with atmospheric oxygen near the paleosurface, isotopically distinct alunite may form. At Brokeoff volcano, stable-isotope and field data suggest that these conditions are represented by small exposures of alunite-kaolinite-opal alteration on the top of Brokeoff Mountain and along the ridge between Brokeoff Mountain and Mount Diller.

Chloride waters in geothermal systems similar to those discharging at Growler and Morgan Hot Springs are reduced through reaction with wall rocks, have near-neutral pH, and contain up to >1 wt% Cl and 3 wt% CO₂, and tens to hundreds of ppm H₂S (e.g., Henley and Ellis, 1983; Giggenbach, 1984, 1997; Simmons et al., 2005). The concentrations of the main aqueous constituents represent equilibrium with a propylitic alteration assemblage of quartz, albite, K-feldspar, illite, chlorite, pyrite, calcite, and epidote that forms during alteration of igneous rocks (Barton et al., 1977; Giggenbach, 1997). The fluid reaches equilibrium with the rock and its principal minerals where flow is slow in a “rock-dominated” or rock-buffered environment (low water/rock; Giggenbach, 1984, 1997). The hydrothermal alteration in Little Hot Springs Valley and West Sulphur Creek formed from interaction with these chloride waters.

The steam-heated environment forms in the upper parts of both magmatic-hydrothermal and geothermal systems. Steam-heated alteration is the result of condensation of H₂S- and CO₂-rich steam released during boiling of a deep geothermal chloride-bearing hot water reservoir (Henley and Ellis, 1983; Simmons et al., 2005). These steam-heated acid-sulfate waters are close to 100 °C, have pH >2, and can alter wall rocks to an advanced argillic assemblage of opal (cristobalite), alunite, kaolinite, and minor pyrite underlain by smectite (e.g., White, 1955; Schoen et al., 1974). Bicarbonate-rich waters form by condensation of CO₂-rich steam on the margins of the acid-sulfate waters and locally are discharged as small travertine-forming hot springs fringing the acid-sulfate fluids (Fig. 3E). Where these fluids descend, mix with hot alkali-chloride water, and heat, they can precipitate carbonate minerals (e.g., Simmons and Browne, 2000). Alteration on Brokeoff Mountain (Fig. 3B) is generally similar to the modern steam-heated alteration, although alunite stable-isotope data suggest that at least part of this alteration formed from ¹⁸O-enriched magmatic-hydrothermal fluids.

A variety of features including lithology (especially breccia units), structural features, such as faults and west-northwest–trending fracture zones, intrusions, and hydrothermal breccias, location of the paleowater table, and topography influenced fluid flow in Pleistocene hydrothermal systems and resultant hydrothermal alteration at Brokeoff volcano. Zones of high permeability, produced both by primary (igneous) and secondary (hydrothermal) brecciation, provided the dominant control on the lateral extent and intensity of hydrothermal alteration. This relationship is especially evident in the Little Hot Springs Valley and West Sulphur Creek zones in which matrices of volcaniclastic rocks and brecciated flow margins are strongly altered relative to clasts in volcaniclastic rocks and dense interiors of lava flows and intrusions that are altered only along fractures (Figs. 5A and 5D). Highly porous and gently dipping volcaniclastic rocks with inferred high subhorizontal permeability were especially prone to alteration.

Fracture-controlled alteration is most evident in narrow, linear zones of intensely altered breccias in Supans Springs (Figs. 3B and 5K) and in similar structures at Maidu (Fig. 7) and in localized zones of quartz-chalcedony-carbonate veins in propylitic alteration in West Sulphur Creek (Fig. 5E). In the Supans Springs zone, the west-northwest–trending fractures that channeled hydrothermal fluids are parallel to dikes and elongate intrusions in the Mill Canyon sequence (Fig. 3B). In general, quartz veins at Brokeoff are sparse, narrow (≤10 cm wide), and discontinuous, and no zones of stockwork veining have been found, suggesting that permeability in volcaniclastic rocks was controlled by interconnected intergranular porosity. These observations are consistent with fluid flow dominantly within the subhorizontal permeable zones of the volcaniclastic rocks outlined above.

The large vertical and horizontal extent of Little Hot Springs Valley alteration probably resulted from lateral continuity of gently dipping, water-saturated volcaniclastic units. Contoured whole-rock δ¹⁸O data presented by Rose et al. (1994) suggest several possible centers of hydrothermal activity marked by low whole-rock δ¹⁸O values (Fig. 3F). One of their contoured δ¹⁸O lows (“bullseyes”) corresponds to the most intense propylitic alteration in Little Hot Springs Valley, and a similar δ¹⁸O low is centered on propylitic alteration in West Sulphur Creek.

Hydrothermal minerals and mineral associations at Brokeoff volcano are indicative of wide variations in hydrothermal fluid composition and temperatures of alteration. Figure 12 shows temperature ranges of hydrothermal minerals sampled globally in drill holes in active hydrothermal systems including alteration minerals identified at Brokeoff volcano (e.g., Browne, 1978; Henley and Ellis, 1983; Reyes, 1990). The presence of clinoptilolite and absence of analcime indicate temperatures ∼≤150 °C for the uppermost preserved parts of Little Hot Springs Valley, whereas the absence of epidote in deeper exposures and in West Sulphur Creek indicates that temperatures did not exceed ∼240 °C. Pyrophyllite + opal-CT alteration in the upper parts of Supans Springs probably formed at ≤150 °C due to high silica activity, whereas pyrophyllite + quartz assemblages in deeper parts of Supans Springs likely formed at higher temperatures (>225–270 °C; Hemley et al., 1980). The presence of quartz plus dickite or kaolinite and absence of pyrophyllite in Mount Diller suggests lower maximum temperatures than in Supans Springs. Temperatures estimated using sulfur-isotope fractionation for six alunite-pyrite pairs range from 184 to 265 °C (Table S1 [^{footnote 1}]). These data suggest maximum temperatures of ∼275–300 °C for Supans Springs, 225–270 °C for Mount Diller, and 240 °C for Little Hot Springs Valley and West Sulphur Creek.

Most Pleistocene alteration formed from water-dominated systems at or below the paleowater table in permeable breccia units on the basis of stable-isotope data that indicate dominantly magmatic-hydrothermal alunite compositions; the lack of preserved steam-heated alunite outside of modern hydrothermal areas; the presence of quartz ± calcite veins and vug fill and hydrothermal breccias; and the presence of zeolites in the upper part of the Little Hot Springs Valley zone. The absence of paleosurface features, such as sinter and opal and/or chalcedony blankets, in the Pleistocene altered exposures of Little Hot Springs Valley and West Sulphur Creek also suggest formation below the paleowater table (e.g., Henley and Ellis, 1983; Hedenquist et al., 2000; Sillitoe, 2015).

Reconstruction of the Brokeoff volcano edifice at ca. 400 ka and 300 ka based on radiating dips of units in the Mill Canyon and Diller sequences provides estimates of paleodepths of present exposures during hydrothermal activity (Figs. 13A and 13B; also see Williams, 1932; Rose et al., 1994). Inferred paleodepths range from near the paleosurface for several samples collected on the ridge between Brokeoff Mountain and Mount Diller that may have been altered on the volcanic edifice to ∼1 km for the lowest exposures of propylitic alteration in Little Hot Springs Valley and West Sulphur Creek. Most of the preserved parts of Supans Springs probably formed at ∼500–1000 m depth. Estimated paleodepth of preserved exposures of Mount Diller, which is not in the plane of these reconstructions, is ≤150 to ∼300 m. The shallow depths are consistent with ∼225–300 °C maximum temperature estimates for magmatic-hydrothermal alteration at hydrostatic pressures.

Some steam-heated alteration probably formed above Supans Springs, Little Hot Springs Valley, and other zones of Pleistocene hydrothermal activity as schematically depicted in Figure 13, but most steam-heated alteration has been removed by erosion. Small exposures of opal-kaolinite-alunite-pyrite alteration on the ridgeline between Mount Diller and Pilot Pinnacle possibly formed in a steam-heated environment above the Mount Diller zone.

Hydrothermal alteration mineralogy of the four major Pleistocene hydrothermal zones indicate two fundamentally different types: (1) Supans Springs and Mount Diller formed by acid (low-pH) fluids and (2) Little Hot Springs Valley and West Sulphur Creek formed by near-neutral pH to weakly alkaline fluids. The cores of the acid alteration zones are further distinguished by two predominant mineral associations: alunite-rich alteration in Supans Springs and alunite-poor, pyrite-kaolinite and/or dickite-rich alteration in Mount Diller. Maidu has characteristics of both Supans Springs and Mount Diller (i.e., alunite- and pyrite-rich). The variable sulfur-containing minerals suggest Supans Springs formed from more oxidized, sulfate-rich fluids than the H₂S-rich fluids that formed the Mount Diller system. In addition, topaz is limited to Supans Springs and Maidu, suggesting higher HF contents of the causative fluids.

Three different types of wall-rock alteration, all derived from similar magmatic fluid sources, may result from variable depths of magma emplacement and degassing and varying interaction of magmatic vapor with surrounding rocks and ground water (variable water/rock ratios, Fig. 14; Giggenbach, 1992, 1997; Einaudi et al., 2003; Chambefort et al., 2017). Figures 14A–14C represent possible paths for fluids forming the Supans Springs, Mount Diller, and Little Hot Springs Valley and West Sulphur Creek alteration zones, respectively. In Supans Springs (path A), shallowly emplaced magma (as suggested by numerous small, approximately coeval intrusions in the system) released magmatic vapor that ascended with little interaction with wall rocks and condensed into shallow ground water at a high water/rock ratio. The resultant highly acidic fluids characterized by high m_HCl/m_KCl values were stable with pyrophyllite or kaolinite (depending on temperature and silica activity) and alunite and had a large HF content that locally formed topaz and F-rich alunite. In Mount Diller (path B), which lacks recognized intrusions, magma may have emplaced at somewhat greater depth, and vapor released during magma crystallization was partly condensed and neutralized by reaction with wall rocks during its ascent. The resultant, less acidic fluid, characterized by intermediate m_HCl/m_KCl, produced kaolinite or illite at moderate water/rock ratios but not pyrophyllite or alunite. In Little Hot Springs Valley and West Sulphur Creek (path C), magma was emplaced at still greater depths, and magmatic vapor released during crystallization mixed with or was absorbed into deep ground water and maintained equilibrium with wall rocks at low water/rock ratios as it slowly ascended. This low m_HCl/m_KCl and near-neutral pH fluid was similar to mature geothermal waters and is inferred to have formed deeper propylitic and shallower intermediate argillic alteration. These fluids precipitated calcite in deeper exposures, indicating appreciable dissolved CO₂ (Fig. 3D). The absence of calcite in shallower intermediate argillic and zeolite alteration may indicate a boiling interface at the transition from propylitic to intermediate argillic alteration and loss of dissolved CO₂ at shallow depths.

Zeolite (clinoptilolite) ± K-feldspar alteration overlies and is partly superimposed on smectite-rich intermediate argillic alteration in Little Hot Springs Valley but has not been identified on other stratovolcanoes in the Cascades or elsewhere to our knowledge. However, zeolite alteration is relatively common in geothermal systems developed in calderas and rift zones, such as in the Taupo volcanic zone (Henneberger and Browne, 1988; Yang et al., 2000, 2001; Braithwaite, 2003), at Yellowstone (Bargar and Beeson, 1981; Bargar and Keith, 1995), and in Valles caldera (Chipera et al., 2008). Based on geothermal systems, the clinoptilolite-adularia alteration probably formed at shallow depths (<100 m) by mixing of cool ground water with ascending dilute, near-neutral pH, alkali-chloride water that likely had already lost CO₂ and H₂S due to slightly deeper boiling (e.g., Henneberger and Browne, 1988; Braithwaite, 2003).

The differences in sulfur speciation of the magmatic-hydrothermal fluids also may be due to different depths and pressures of magma degassing. Supans Springs vapor likely was released at shallower depths and pressures than Mount Diller vapor, which for a given temperature and magmatic oxidation state would result in higher SO₂/H₂S for Supans Springs fluids relative to Mount Diller fluids (Giggenbach, 1987; Symonds et al., 1994; Einaudi et al., 2003; Burgisser and Scaillet, 2007; Burgisser et al., 2008). Still deeper magma emplacement and degassing for Little Hot Springs Valley and West Sulphur Creek would lead to H₂S-dominant fluids in these systems. The change from shallow structural control on flow of acid fluids to inferred deep upwelling of neutral pH fluids along basement structures also suggests increasing depth of magmatic-hydrothermal fluids and interaction with wall rocks.

Geologic and geochemical studies described above define multiple phases of Pleistocene hydrothermal activity that were an integral part of the development and erosion of Brokeoff volcano (Fig. 4). Figure 13 shows a series of reconstructed, approximately west-east cross sections through Brokeoff volcano at 400 ka, 300 ka, and present day; these sections integrate volcanic and hydrothermal histories of the volcano and are outlined here.

1. Eruption of the andesites of Mill Canyon, a sequence of interbedded thin pyroxene andesite lava flows and volcaniclastic rocks, capped by thick flows of the dacite of Twin Meadows and formation of the Brokeoff volcano stratocone (Fig. 13A). A central vent may have been located just west of Sulphur Works, as previously suggested by Williams (1932). New Ar-Ar dates suggest eruption of the Mill Canyon sequence between ca. 555–445 ka.

2. Emplacement of small andesitic dikes and plugs into rocks of the Mill Canyon sequence. The largest plug, located just west of Sulphur Works, is the ∼0.3 km² Supans Springs plug, which may intrude a central vent for the Mill Canyon sequence rocks. New Ar-Ar dates for these intrusions range from ca. 505–435 ka (the plugs are ca. 460–435 ka). The dikes are dominantly approximately west trending (Fig. 3A).

3. Eruption of the andesite of Bluff Falls (ca. 445 ka) along the south edge of the map area, which marks the beginning of the Diller sequence.

4. Onset of hydrothermal activity at Supans Springs (Fig. 13A). Alunite Ar-Ar dates suggest hydrothermal activity at ca. 420–405 ka coinciding with early volcanism of the Diller sequence. Field relations suggest Supans Springs may be related to the undated intrusion (or underlying crystallizing magma) just west of Sulphur Works, from which numerous narrow quartz-alunite ledges extend westward (Fig. 3B). Several small andesite intrusions farther west dated at ca. 460 and 435 ka appear unaltered but are ringed by strongly altered rocks. These intrusions either were impermeable to Supans Springs hydrothermal fluids or indicate that Supans Springs alteration formed from several pulses of hydrothermal activity.

5. Formation of small hydrothermal alteration zones near Ridge Lakes. East Ridge Lakes postdates emplacement of the dacite of Twin Meadows, and both alteration zones likely predate eruption of the andesite of Mount Diller. West Ridge Lakes might be an upper remnant of Supans Springs: it includes isotopically distinct “high-elevation” alunite suggesting formation at very shallow depth. East Ridge Lakes alteration apparently formed from a fluid source different than Supans Springs based on stable-isotope data.

6. Eruption of the andesites of Mount Diller and Glassburner Meadows from vents on flanks of Brokeoff volcano (Fig. 13B). These units consist of thick flows of silicic pyroxene andesite. New Ar-Ar dates suggest eruption of the andesite of Mount Diller between ca. 405–390 ka and the andesite of Glassburner Meadows at ca. 360 ka (Fig. 4 and Table 1).

7. Formation of large hydrothermal alteration zones near Mount Diller, in Little Hot Springs Valley, in West Sulphur Creek, and on Brokeoff Mountain (Fig. 13B). The Mount Diller and Little Hot Springs Valley zones formed between ca. 380–300 ka as they affected both the Diller and Mill Canyon sequences but not the younger (≤300 ka) Bumpass sequence. At West Sulphur Creek, only the Mill Canyon sequence rocks are hydrothermally altered. Brokeoff Mountain affected both the Diller and Mill Canyon sequences. Intrusive rocks are not evident in any of these zones, and alteration is inferred to be related to magma degassing at greater depths than Supans Springs. Acid alteration in Brokeoff Mountain probably formed near the paleosurface, whereas alteration in West Sulphur Creek and Little Hot Springs Valley extends over ∼1000 m paleodepth.

8. Emplacement of the Lassen domefield, which consists of central pyroxene-hornblende dacitic domes of the Bumpass (ca. 300–190 ka) and Eagle Peak (ca. 120–0 ka) sequences and the flanking Twin Lakes sequence of hybrid andesite flows. No hydrothermal activity related to the Bumpass sequence has been recognized in Brokeoff volcano.

9. Modern hydrothermal activity at Lassen, including Bumpass Hell, Little Hot Springs Valley, and Sulphur Works, and at Morgan and Growler Hot Springs, developed after glaciation ended at ca. 8,000–12,000 yr B.P. (Clynne and Muffler, 2010; Fig. 13C). The heat and volatile source is an inferred magma body at >7 km depth that last erupted in 1914–1917 from Lassen Peak (Janik and McLaren, 2010). In Brokeoff volcano, the surface effects of the active hydrothermal system vary seasonally and consist of steam-heated acid-sulfate and bicarbonate fluids that form fumaroles, mudpots, steaming ground, and small travertine mounds. Deep, boiling alkali-chloride waters discharge at Morgan and Growler Hot Springs ∼8 km south of Bumpass Hell and form silica sinter deposits.

Vertically zoned magmatic-hydrothermal alteration at Maidu is exposed from near the paleosurface to ∼300 m deep and is broadly similar to the Supans Springs and Mount Diller alteration zones at Brokeoff (Table 5). All three zones are primarily hosted by sequences of andesite breccias with fracture-controlled alteration evident in deeper exposures of Maidu and Supans Springs (paleodepths probably ∼500–1000 m). Alteration at Maidu has a blanket-like morphology with abrupt lateral transitions into weakly altered or unaltered rock similar to the core of Mount Diller. Hydrothermal fluids at Maidu formed extensive amounts of alunite alteration with local residual quartz and topaz alteration in the deepest exposures, similar to Supans Springs. Stable-isotope relations indicate that Maidu formed from mixed magmatic and meteoric waters in proportions similar to Supans Springs.

The differences in magmatic-hydrothermal activity at Brokeoff and Maidu may be the result of (1) paleodepths of present exposures; (2) depths of magma emplacement; (3) fluid flow as influenced by paleotopography; and (4) centers of hydrothermal activity (Fig. 13A). The deepest exposures of Maidu are at paleodepths similar to Mount Diller (∼150–300 m) but shallower than most Supans Springs exposures. The blanket-like morphology of the Maidu and Mount Diller alteration zones suggests upward flow of hot acidic fluids along high-angle structures into shallow, water-saturated breccia sequences followed by lateral dispersion. Ledges of brecciated, residual quartz alteration in the deepest parts of Maidu and Supans Springs likely mark structurally controlled fluid conduits. The presence of topaz in Maidu and Supans Springs suggests significant HF and magmatic-gas components in these hydrothermal fluids. Maidu hydrothermal alteration formed along the outer flank of the volcanic center (Fig. 2), whereas the Supans Springs and Mount Diller alteration zones formed closer to the center of Brokeoff volcano. Maidu has a much larger altered area (∼3.5 km²) than the acid zones at Brokeoff, which may be due to (1) larger supply of SO₂-rich magmatic gas, (2) longer sustained hydrothermal activity, and/or (3) more efficient condensation of magmatic gases into an aquifer that may have been larger or thicker. Hydrothermal sediments exposed at the top of Maidu indicate that the water table was at the paleosurface at least during the late stages of hydrothermal activity, whereas it was at unknown depths at Supans Springs and Mount Diller.

Due to greater depths of erosion, the shallow levels of Pleistocene hydrothermal activity on Brokeoff and Maidu are better exposed than on other Quaternary stratovolcanoes in the Cascades arc, and therefore, provide a depth model for the types, environments, and timing of hydrothermal activity on stratovolcanoes elsewhere.

Mount Rainier is the only other Cascades stratovolcano on which similar magmatic-hydrothermal alteration is exposed (Fig. 1; Zimbelman, 1996; John et al., 2008). At Mount Rainier, eruptive and hydrothermal activity has been episodic with periods of hydrothermal activity following major effusive fluxes (Sisson et al., 2001; John et al., 2008). The top and upper 1000–1500 m of the north side of the volcano collapsed at ca. 5600 yr B.P., forming the ∼3.5 km³ Osceola Mudflow that flowed >100 km downstream (Vallance and Scott, 1997). The Osceola Mudflow contains abundant clasts of hydrothermally altered rock in a clay- and pyrite-rich matrix. All mineral associations identified at Brokeoff volcano except zeolite-adularia were found in clasts in the Osceola Mudflow, although weak smectite-pyrite alteration is dominant. Higher-temperature, magmatic-hydrothermal advanced argillic alteration is exposed in Sunset Amphitheater on the upper west side of the volcano, and this area poses the greatest risk for future debris flows (Zimbelman, 1996; Finn et al., 2001; Reid et al., 2001; John et al., 2008).

At Mount Adams, steam-heated, alunite-bearing advanced argillic and smectite-rich intermediate argillic alteration and native sulfur deposits are exposed near the summit and in debris flows derived from the upper edifice (Fig. 1; Hildreth et al., 1983; Vallance, 1999; Finn et al., 2007). Geophysical modeling suggests that the upper interior of Mount Adams is hydrothermally altered as much as 1.5 km deep beneath ice and shallow steam-heated alteration (Finn et al., 2007). This alteration is likely similar to deeper alteration sampled by the Osceola Mudflow at Mount Rainier (John et al., 2008).

At Mount Baker, Frank (1983) reported hydrothermal minerals that suggest higher-temperature magmatic-hydrothermal alteration in recent ejecta from a fumarole field on the flanks of the volcano and in the clasts and matrices of Holocene debris flows. Finn et al. (2018) modeled geophysical data and suggested that hydrothermally altered rocks likely similar to those at Mount Rainier are limited to two small (500-m-diameter) and thin (250–500 m) zones beneath active fumarole fields.

Hydrothermal alteration on other stratovolcanoes in the Quaternary Cascades arc generally is not well studied, and high-temperature hydrothermal-alteration assemblages have not been described on volcanoes that have been studied in more detail (e.g., Mounts Hood and Shasta, Bargar et al., 1993; Crowley et al., 2003; Zimbelman et al., 2005). Especially noteworthy is the near absence of hydrothermally altered rock in the ∼45 km³ debris-avalanche deposit formed by collapse of ancestral Mount Shasta (Crandell, 1989). The eroded Pliocene Yana and Dittmar stratovolcanoes near Lassen Peak (Fig. 2) also appear only weakly altered and lack the acid alteration present at Brokeoff and Maidu.

Detailed studies of hydrothermal alteration on active stratovolcanoes elsewhere are scarce. Based largely on geothermal drill holes, Reyes et al. (1993) and Rae et al. (2003) describe alteration on the flanks of active volcanoes in the Philippines, and Moore et al. (2008) present similar data for Galunggung volcano, Indonesia. Hedenquist et al. (2018) summarize alteration and present a composite hydrothermal model for several active stratovolcanoes in Japan. All these studies show the complexity and evolving nature of magmatic-hydrothermal systems on the flanks of stratovolcanoes and indicate that these systems are characteristic features of arc-related stratovolcanoes.

Our studies at Brokeoff and Maidu provide knowledge of the upper 1000 m of an arc-related geothermal system and highlight very diverse alteration at Brokeoff, likely the result of variable water/rock, pH, and S content of magmatic-hydrothermal fluids and corresponding variations in magma depth.

Hydrothermally altered rocks at Brokeoff and Maidu locally are enriched in many “pathfinder” elements present in epithermal gold-silver deposits and in the upper parts (lithocaps) of porphyry copper systems (Table 4; Saunders et al., 2014; Halley et al., 2015). Alteration types and hydrothermal features, such as hydrothermal breccias, residual quartz ledges, and banded quartz ± carbonate veins, are similar to characteristics of these deposits (e.g., Hedenquist et al., 2000; Simmons et al., 2005; Sillitoe, 2010; Hedenquist et al., 2018).

At Brokeoff and Maidu, exposed residual quartz alteration and the matrices of hydrothermal breccias formed at paleodepths between ∼200–800 m are not infilled by Cu-As-Sb sulfide or sulfosalt minerals or gold, which characterize mineralization in high-sulfidation Au-Ag deposits and in some lithocaps. Quartz-carbonate veins exposed at ≥∼800 m paleodepth in West Sulphur Creek might represent parts of an epithermal vein system. However, these veins lack textures suggesting formation from boiling fluids (e.g., platy calcite or adularia), contain only trace amounts of sulfide minerals, and are not enriched in gold or silver. Zones of stockwork quartz veins are not exposed.

The Cascades arc in the Lassen area (∼40.4°N latitude) lacks known mineral deposits. Miocene arc rocks farther north in Washington and northern Oregon (north of 44.5°N latitude) host several porphyry copper systems (du Bray and John, 2011), whereas many Miocene and Pliocene epithermal gold-silver deposits are present in the arc south of 40°N latitude (du Bray et al., 2014; John et al., 2015). Formation of arc-related mineral deposits may be affected by: (1) magma composition, (2) rate and angle of subduction, (3) tectonic stress regime, and (4) thickness and composition of underlying crust. du Bray and John (2011) concluded that porphyry copper systems were limited to the northern part of the arc that was undergoing transpression to mild compression and underlain by thicker, lower-density crust along the western edge of the North American plate. John et al. (2015) concluded that large epithermal gold-silver deposits are limited to the southern part of the arc overlying the North America craton margin or accreted terranes with continental affinity and that these deposits are mostly related to intermediate to silicic dome complexes in long-lived volcanic centers formed in mildly extensional or transtensional stress regimes.

du Bray et al. (2014) showed that magmas from the northern and southern parts of the arc have fundamental compositional differences that reflect magma generation beneath different types of crust. Magma compositions in the Lassen area are similar to other parts of the southern half of the arc that contain significant epithermal deposits. However, Brokeoff and Maidu volcanic rocks are mostly pyroxene andesites that likely had lower water and sulfur(?) contents than hornblende-(biotite) dacites in the younger Bumpass and Eagle Peak sequence magmas. The latter are similar to more fertile arc magmas (higher water content) globally (e.g., Tosdal et al., 2009; Richards, 2011).

Brokeoff and Maidu formed in a stress regime transitional from northwest-directed dextral shear related to northward propagation of the Walker Lane belt to east-west extension related to westward expansion of the Basin and Range province onto the Cascades arc (Guffanti et al., 1990; Faulds and Henry, 2008; Janik and McLaren, 2010). This stress environment appears more conducive for formation of epithermal deposits than porphyry deposits, as provided by western Nevada examples.

Unlike mineralized dome complexes farther south in the Miocene arc, large silicic domes in the Lassen area are mostly unaltered and appear unmineralized. Steam-heated alteration at Bumpass Hell is forming in Bumpass sequence domes on the north side of Brokeoff volcano, but mercury and sulfur are the only elements commonly enriched in this alteration. Alkali-chloride waters from the deep hot-water (∼240 °C) reservoir under Bumpass Hell laterally discharges at Growler and Morgan Hot Springs 7–8 km south (Fig. 2). This fluid precipitates silica sinter that locally contains As and Sb sulfides with trace amounts of gold (Table 4), which might indicate deposition of epithermal mineralization in favorable structural locations at depth beneath Brokeoff volcano (e.g., boiling upwelling zones, Rowland and Simmons, 2012).

In summary, outcrop-scale features indicate that hydrothermal processes characteristic of epithermal and the upper parts of porphyry systems were operative at Brokeoff and Maidu, but they are generally less extensive and smaller volume than those related to ore deposits and do not provide strong evidence for the presence of mineralized rock. Brokeoff and Maidu magmas mostly formed pyroxene andesites that likely were less water rich than arc-related magmas that form these deposit types. Further, regional metallogeny suggests that Cascades arc magmatism at Lassen is more likely to form epithermal gold-silver mineralization than porphyry mineralization, but this part of the arc lacks known mineral deposits. The deep hot-water reservoir presently beneath eroded Brokeoff volcano potentially is forming epithermal Au-Ag mineralization at ≥1 km depth.

Brokeoff and Maidu provide exceptional exposures of hydrothermal alteration formed by evolving magmatic-hydrothermal systems in the upper 1 km of andesitic stratovolcanoes that are seldom preserved. At least four major phases of late Pleistocene hydrothermal activity occurred within ∼100 k.y. in the core of Brokeoff volcano, and a large, ca. 1.5 Ma phase of activity affected the northeast flank of the Maidu volcanic center. Modern hydrothermal activity is producing new steam-heated alteration that overprints the Pleistocene alteration at Brokeoff.

On Brokeoff volcano, both acid and neutral pH alteration zones formed in close proximity. The earliest alunite-rich acid alteration zone is spatially and temporally related to shallow andesite intrusions, whereas younger, smaller acid and larger neutral-pH alteration zones lack temporally associated intrusions. Especially noteworthy are exposures of neutral-pH alteration that are zoned from shallow zeolite-K-feldspar through intermediate argillic to deep propylitic alteration. Variations in types of hydrothermal alteration likely result from increased interaction with, and neutralization by, wall rocks during ascent of magmatic gas (decreasing water/rock), which may reflect increasing depth of magma emplacement through time that resulted in increasing H₂S/SO₂ ratios of gases exsolved from crystallizing magma. Alunite-rich alteration at Maidu is similar to early alteration at Brokeoff, but preserved exposures formed at shallower depths (∼0–300 m) and are more laterally extensive.

Fluid flow and hydrothermal alteration were controlled by hydraulic gradients and rock types, and by steeply dipping fractures, where numerous narrow ledges of quartz-alunite and residual quartz alteration probably mark vertical fluid feeders. Deep fluid upflow zones for neutral pH alteration are marked by whole-rock δ¹⁸O lows (Rose et al., 1994) along inferred north- to northwest-striking structures in underlying basement rocks.

Stable-isotope data indicate hydrothermal fluids were mixtures of exchanged meteoric and magmatic waters; alunites have larger magmatic water components than clay minerals. Variations in alunite S-O isotopic compositions indicate formation in both near-surface and deeper environments consistent with their locations on reconstructions of Brokeoff volcano, whereas modern alunite alteration is forming in a surficial steam-heated environment.

Mounts Rainier and Adams in the Cascades and several active volcanoes in the western Pacific show similar hydrothermal activity, suggesting that magmatic-hydrothermal systems similar to those on Brokeoff and Maidu are common on arc-related stratovolcanoes worldwide; however, these systems remain understudied. The lack of exposed ore at Brokeoff and Maidu may reflect relatively small amounts of magmatic acid, chlorine, and sulfur consistent with degassing of small volumes of pyroxene andesite compared to large dacite and/or granodiorite batholiths that form significant epithermal and porphyry deposits.

The U.S. Geological Survey (USGS) Mineral Resources and Volcano Hazards programs supported part of this research, and USGS Mineral Resources External Research Grant 04HQGR0164 supported Lee and Dilles. We thank the U.S. National Park Service for permission to work within Lassen Volcanic National Park and to collect samples there. Peter Seward allowed us access to sample Growler Hot Springs. Jim Rytuba, Kathryn Flynn, Deb Bergfeld, and Geoff Plumlee helped with field work. Rhonda Driscoll and Kathryn Flynn prepared samples for mineralogy and X-ray diffraction studies. Cynthia Kester, Cayce Gulbransen, William Christiansen, and Pamela Gemery performed stable-isotope analyses, and Cyrus Berry conducted many of the sulfur speciation analyses. Stuart Simmons, Allen Anderson, Isabelle Chambefort, and an anonymous reviewer provided helpful reviews of an earlier version of this manuscript. Shan de Silva and Peter Larson provided helpful editorial guidance. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Approximately 700 variably altered rock samples were collected from Brokeoff volcano. The constituent minerals were identified using a combination of optical microscopy, powder X-ray diffraction (XRD), scanning electron microscopy (SEM), and shortwave infrared (SWIR) spectroscopy. Polished thin sections of ∼100 altered rocks were examined in transmitted and reflected light. XRD was applied to ground bulk samples and hand-picked grains using CuKα radiation from 4° to 64° 2Ø with step size of 0.02° and a count time of 2 s/step. Clay-sized separates (<1 μm) from 30 samples were separated by centrifugation. Oriented mounts prepared from the ≤1 μm separates were scanned using CuKα radiation from 2° to 40° 2Ø as air-dried, saturated with ethylene glycol, heated to 375 °C, and heated to 550 °C. Interpretative mineralogy of the XRD scans are tabulated in Table S2 (see text ^{footnote 1}). About 35 polished thin sections were examined in backscatter and secondary electron modes using an SEM equipped with an energy dispersive analyzer. Most altered rocks were scanned in SWIR wavelengths (1.3–2.5 µm) using a portable infrared mineral analyzer (PIMA SP), which allows identification of hydrous phases (e.g., clay minerals, pyrophyllite, micas, and chlorite), sulfates (alunite, jarosite, and gypsum), carbonates, and topaz (Thompson et al., 1999). Illite composition and crystallinity were estimated from position of the 2.2 µm Al-OH absorption feature and Hull-normalized depths of the 1.9 µm H₂O and 2.2 µm Al-OH absorption features (Thompson et al., 1999; Chang et al., 2011). Similarly, Na-K contents of alunites were estimated from position of the ∼1.48 µm absorption feature (Thompson et al., 1999; Chang et al., 2011). Interpretative mineralogy from the SWIR analyses is presented in Table S1. Complete analytical data for XRD and SWIR spectroscopic scans are available as a U.S. Geological Survey data release (John et al., 2019).

Whole-rock chemical analyses of 214 strongly altered rocks and 62 weakly or unaltered rocks are summarized in Table 4 and reported in Table S3 (see text ^{footnote 1}). Analyses were performed by SGS Minerals under contract to the U.S. Geological Survey. Samples were dissolved using a hydrochloric, nitric, perchloric, and hydrofluoric acid digestion and analyzed for 42 elements using inductively coupled plasma–mass spectrometry (ICP-MS) and atomic emission spectrometry (ICP-AES) methods. In addition, gold was analyzed by fire assay and mercury by cold-vapor atomic absorption techniques. Some samples were analyzed for Se by hydride generation–atomic absorption spectrometry. See Taggart (2002) for analytical techniques.

Sulfur-bearing phases were sequentially extracted using chemical reagents to estimate phase abundance and to recover sulfur and sulfate for isotopic analysis (Berry and Breit, 2007). Chemical extraction was required, because the sulfur-bearing phases were generally fine grained (<100 μm) and sufficiently intergrown such that mechanical separation was impractical. The extraction procedure separately recovers elemental sulfur, water-soluble sulfates (anhydrite and/or gypsum), jarosite, pyrite, and alunite and/or barite. Replicate extraction analyses were within 10 relative percent for each phase measured. All phases were analyzed to determine their δ³⁴S composition, and the anhydrite and/or gypsum, jarosite, and alunite and/or barite fractions were analyzed to determine δ¹⁸O. Purified mineral separates using standard physical separation techniques were used for isotopic analysis of minerals in some samples. Oxygen and hydrogen isotopes were analyzed for clay-sized separates from 12 altered rocks, as well as for 12 alunite separates. Oxygen-isotope compositions were determined for five quartz vein samples from Brokeoff volcano.

Most sulfur-isotope analyses were made by continuous-flow mass spectrometry similar to that described by Giesemann et al. (1994) using an elemental analyzer on an Optima MicroMass mass spectrometer. Hydrogen-isotope analyses of hydrous minerals and oxygen-isotope analyses of sulfates were made by continuous-flow techniques through a temperature-conversion elemental analyzer (TC EA) coupled to a ThermoFinnigan Delta XL mass spectrometer using techniques similar to those described by Sharp et al. (2001). Oxygen-isotope analyses of hydrous minerals were made on CO₂ prepared by the BrF₅ technique of Clayton and Mayeda (1963) followed by conventional analyses on a Finnigan MAT 252 mass spectrometer. Precision is ±0.2‰ for sulfur isotope data, ±2‰ for hydrogen, ±0.5‰ for δ¹⁸O of sulfates, and ±0.2 for hydrous minerals. Results of the isotopic analyses are listed in Tables S1 and S5 (^{footnote 1}).