Part I. 18O/16O analyses were made on 355 samples in and around the 11 by 14 km Lake City caldera, which formed 23 m.y. ago in response to the eruption of the rhyolitic Sunshine Peak Tuff. All of the major lithologies and hydrothermal alteration facies were analyzed, and a detailed δ18O map was made of the caldera and its surroundings. Intracaldera facies Sunshine Peak Tuff consists of three members interbedded with landslide debris and megabreccias shed into the caldera during eruption and collapse. Asymmetric resurgence within the Lake City caldera followed collapse and was accompanied by intrusion of a flat-topped, granitic magma centered in the resurgent dome. Ring magmatism produced dike-like intrusions along the northern ring fault and the Red Mountain-Grassy Mountain quartz latite ring dome on the eastern caldera margin. The caldera was emplaced into older Tertiary volcanic rocks and Precambrian granitic rocks.
Based on analyses of outflow-facies samples and of the least altered intracaldera facies, we can demonstrate that the caldera-fill Sunshine Peak Tuff originally was isotopically very homogeneous, with an initial igneous δ18O value of +7.2 to +7.3. Thus, 18O depletions in the hydrothermally altered tuff could be compared without worrying about the complicating factor of different initial δ18O values. Nearly all the rocks within the caldera and outside the caldera within at least 3 km of the ring fault were altered by meteoric-hydrothermal fluids, and depleted in 18O down to values as low as δ18O = -3.1. Erosion has exposed the hydrothermally altered caldera-fill rocks and the upper contact of the altered resurgent intrusion in the western and central part of the caldera, providing about 2 km of vertical exposure. Because of post-alteration regional eastward tilting, the eastern part of the caldera has not been extensively eroded, and the original topography of the ring dome and the top of the caldera-fill rocks are locally preserved. This differential erosion from west to east furnishes a unique opportunity to study water-rock interactions in a caldera-type hydrothermal system from near-surface environments down through 3 km into the sub-volcanic intrusion that drove the hydrothermal convection.
Elevation and proximity to fractures exerted the strongest control on 18O-depletions in the Precambrian granite and the older volcanic rocks. The lowest alSo values are found in rocks from the Eureka graben, a highly-fractured and extensively altered zone that extends SW from the caldera. Low δ18O values also occur adjacent to the caldera ring fault. Those samples of the Precambrian granite and of the older volcanic rocks that are located at the greatest depths below the mid-Tertiary erosion surface have the lowest δ18O values. At present-day, constant elevations, δ18O values are lowest in the western part of the study area than in the eastern part; this is a result of the regional eastward tilting. The above effects are best interpreted as indicating higher water/rock ratios near the permeable fractures and higher temperatures at greater depth.
The δ18O values within the Lake City caldera are controlled by elevation, proximity to permeable zones, and proximity to the resurgent intrusive rocks. δ18O values decrease systematically with stratigraphic depth within the caldera. The lowest δ18O values are found along the western ring fault, along resurgence-related fractures, in the permeable megabreccia units, and along the contact of the resurgent intrusion. Mineralogic alteration facies within the caldera show complementary patterns. Intense argillization is found along fractures near the resurgent intrusion. Rocks adjacent to the resurgent intrusion have been hornfelsed but not intensely mineralogically altered. Weak argillization in stratigraphically shallow Sunshine Peak Tuff grades into both of these alteration regimes and also grades downward into chlorite-calcite alteration. These data show that the resurgent intrusion was the "heat engine" that drove the Lake City hydrothermal system. Alteration in and near the intrusion occurred at high temperatures (≈ 400°C) and intermediate water/rock ratios. Away from the resurgent intrusion, water-rock interaction in the permeable zones (megabreccia units and fractures) occurred at lower temperatures (200°C to 300°C) and high water/rock ratios. The regional eastward tilting has raised low-18O rocks in the western part of the caldera to higher elevations than stratigraphically equivalent rocks in the eastern part of the caldera. Mineralogical alteration patterns are also similarly displaced.
Near-surface solfataric alteration is centered on a brecciated zone in the Red Mountain-Grassy Mountain quartz-latite dome on the eastern caldera margin. δ18O values of hydrothermal quartz from this alteration zone are high (> +11) and decrease gradationally with depth. Vein quartz δ18O values from deeper levels within the caldera lie on the deeper projection of the solfataric quartz δ18O trend. These data can be successfully modelled using an upward-flowing, boiling, 18O-shifted meteoric water as a hydrothermal fluid. This model shows that the same fluids responsible for vein quartz precipitation also produced the shallow solfataric alteration. High δ18O values were also measured from a number of other solfatarically altered areas in the San Juan Mountains (Red Mountain district near Silverton, Calico Peak near Rico, Engineer Pass, Carson Camp, and the Summitville district). These alteration zones, some of which are economically mineralized, were also apparently produced by boiling meteoric-hydrothermal fluids.
Deeply-circulated, 18O-shifted, meteoric waters were the primary source of hydrothermal fluids in the Lake City hydrothermal system. By analogy with other deeply eroded caldera hydrothermal systems studied by other workers, such fluids probably rose along deep extensions of the fractured, permeable Lake City ring fault zone. At the present level of exposure, fluids in the Lake City hydrothermal system were apparently drawn into the central part of the resurgent dome along the permeable, outward-dipping, megabreccia units. Flow was directed upward along permeable fractures where these fractures intersected the megabreccia units. A strong thermal gradient existed around and over the resurgent intrusion. Recharge into this hydrothermal system was basically radially inward toward the caldera, but flow was greatly enhanced through the permeable, highly fractured, Eureka graben.
Part II. Oxygen isotope studies were made on 60 samples from the central Nevada caldera complex, which consists of three nested calderas that erupted from 32 to 25 m.y. ago. 18O/16O analyses were also made on 96 samples from the central San Juan caldera complex, Colorado, which contains 7 ash-flow tuffs, each erupted from separate, nested collapse structures between 28 and 26 m.y. ago. The sequence of ash-flow tuffs erupted from the earliest of the three central Nevada calderas began with the giant Tuff of Williams Ridge and Morey Peak (+ 2500 km3), followed by the Mbnotony Tuff (3000 km3), and finally by various ash flow tuffs erupted from the youngest caldera (400 km3). In the San Juan complex, the earliest ash-flow was the Fish Canyon Tuff, which is also the largest of these ash-flows (> 3000 km3). Of the six other major ash-flow tuffs erupted from the central San Juan complex, none exceeds 1000 km3 in volume.
Previous studies of other complex calderas at Yellowstone National Park and in southwest Nevada indicated that the later eruptions have δ18O values 3 per mil lower than rocks erupted early in the cycles. However, in the present study, no large negative shifts in δ18O were found. The various eruptions in both central Nevada and the central San Juans were remarkably uniform in 18O/16O, although small shifts of about -0.2 to -0.3 per mil were found in both suites of rocks in going from the early ash-flows to a later set. The indicated range of δ18O values of these quartz-latite and rhyolite magmas was 9.1 to 9.8 in the central Nevada complex and 6.6 to 7.5 in the central San Juan complex. The higher δ18O values in central Nevada probably indicate melting of sedimentary or metasedimentary country rocks at depth, whereas in Colorado, the low-18O, lower part of the craton was very likely involved in the melting process.
δ18O fractionations between coexisting phenocryst minerals in all of the ash-flow tuffs and lava flows from these two complexes are larger at the bases of the units (tops of the magma chambers) and smaller at the tops of the units (deeper levels of the magma chambers). These relationships show that temperature gradients existed in virtually all these magmas prior to eruption (cooler at the top and hotter at deeper levels).
It is not clear why some complex caldera magmas become depleted in 18O with time, and others do not. No relationship exists between the duration of caldera magmatism and low-18O rocks, nor between the size of the eruptions and these 18O depletions. However, the large 18O depletions found to date occur only in caldera complexes younger than about 15 to 20 m.y., corresponding to the initiation of Basin-Range extension in the western United States. Perhaps Basin-Range faults allow meteoric fluids to penetrate deeply into fairly high-temperature regions of the crust. These younger magmas might then be able to melt or assimilate larger amounts of altered, 18O depleted rocks during their ascent. A correlation also appears to exist between the 18O depletions and the major-element chemistry of the rocks. All of the low-18O ash-flow tuffs contain abundant high-silica rhyolites (consistently ranging up to or above 77 percent SiO2). The large silica contents of these magmas indicate very strong differentiation, suggesting prolonged assimilation-fractional crystallization in a stable magma chamber without the renewed addition much primitive, unfractionated magma having a "normal" δ18O value. It is tentatively concluded that ash-flow magmas strongly depleted in o1eo will be produced only if: (1) there is a pre-history of intense fracturing, caldera collapse, and extensive meteoric-hydrothermal activity, followed by (2) the development of a stable, strongly differentiated, zoned magma chamber, whose roofward portion is in close proximity to low-18O, hydrothermally altered roof rocks for an extended interval of time (> 100,000 years ?).
