We use seismic, tilt, lidar, thermal, and gravity data from 32 consecutive eruption cycles of Lone Star geyser in Yellowstone National Park to identify key subsurface processes throughout the geyser's eruption cycle. Previously, we described measurements and analyses associated with the geyser's erupting jet dynamics. Here we show that seismicity is dominated by hydrothermal tremor (~5-40 Hz) attributed to the nucleation and/or collapse of vapor bubbles. Water discharge during eruption preplay triggers high-amplitude tremor pulses from a back azimuth aligned with the geyser cone, but during the rest of the eruption cycle it is shifted to the east-northeast. Moreover,~4 min period ground surface displacements recur every 26 ± 8 min and are uncorrelated with the eruption cycle. Based on these observations, we conclude that (1) the dynamical behavior of the geyser is controlled by the thermo-mechanical coupling between the geyser conduit and a laterally offset reservoir periodically filled with a highly compressible two-phase mixture, (2) liquid and steam slugs periodically ascend into the shallow crust near the geyser system inducing detectable deformation, (3) eruptions occur when the pressure decrease associated with overflow from geyser conduit during preplay triggers an unstable feedback between vapor generation (cavitation) and mass discharge, and (4) flow choking at a constriction in the conduit arrests the runaway process and increases the saturated vapor pressure in the reservoir by a factor of~10 during eruptions.

[1] Taking advantage of large datasets of both gravity and elastic wave arrival time observations available for the Parkfield, California region, we generated an image consistent with both types of data. Among a variety of strategies, the best result was obtained from a simultaneous inversion with a stability requirement that encouraged the perturbed model to remain close to a starting model consisting of a best fit to the arrival time data. The preferred model looks essentially the same as the best-fit arrival time model in areas where ray coverage is dense, with differences being greatest at shallow depths and near the edges of the model where ray paths are few. Earthquake locations change by no more than about 100 m, the general effect being migration of the seismic zone to the northeast, closer to the surface trace of the San Andreas Fault.

We present newly compiled magnetic, gravity, and geologic datasets from the Parkfield region around the San Andreas Fault Observatory at Depth (SAFOD) pilot hole in order to help define the structure and geophysical setting of the San Andreas Fault (SAF). A 2‐D cross section of the SAF zone at SAFOD, based on new, tightly spaced magnetic and gravity observations and surface geology, shows that as drilling proceeds NE toward the SAF, it is likely that at least 2 fault bounded magnetic slivers, possibly consisting of magnetic granitic rock, serpentinite, or unusually magnetic sandstone, will be encountered. The upper 2 km of the model is constrained by an order of magnitude increase in magnetic susceptibility at 1400 m depth observed in pilot hole measurements. NE of the SAF, a flat lying, tabular body of serpentinite at 2 km depth separates two masses of Franciscan rock and truncates against the SAF.

Though shallow flow of hydrothermal fluids in Long Valley Caldera, California, has been well studied, neither the hydrothermal source reservoir nor heat source has been well characterized. Here a grid of magnetotelluric data were collected around the Long Valley volcanic system and modeled in 3‐D. The preferred electrical resistivity model suggests that the source reservoir is a narrow east‐west elongated body 4 km below the west moat. The heat source could be a zone of 2–5% partial melt 8 km below Deer Mountain. Additionally, a collection of hypersaline fluids, not connected to the shallow hydrothermal system, is found 3 km below the medial graben, which could originate from a zone of 5–10% partial melt 8 km below the south moat. Below Mammoth Mountain is a 3 km thick isolated body containing fluids and gases originating from an 8 km deep zone of 5–10% basaltic partial melt.

The Tualatin basin, west of Portland (Oregon, USA), coincides with a 110 mGal gravity low along the Puget-Willamette lowland. New gravity measurements (n = 3000) reveal a three-dimensional (3-D) subsurface geometry suggesting early development as a fault-bounded pull-apart basin. A strong northwest-trending gravity gradient coincides with the Gales Creek fault, which forms the southwestern boundary of the Tualatin basin. Faults along the northeastern margin in the Portland Hills and the northeasttrending Sherwood fault along the southeastern basin margin are also associated with gravity gradients, but of smaller magnitude. The gravity low refl ects the large density contrast between basin fi ll and the mafi c crust of the Siletz terrane composing basement. Inversions of gravity data indicate that the Tualatin basin is ~6 km deep, therefore 6 times deeper than the 1 km maximum depth of the Miocene Columba River Basalt Group (CRBG) in the basin, implying that the basin contains several kilometers of low-density pre-CRBG sediments and so formed primarily before the 15 Ma emplacement of the CRBG. The shape of the basin and the location of parallel, linear basin-bounding faults along the southwest and northeast margins suggest that the Tualatin basin originated as a pull-apart rhombochasm. Pre-CRBG extension in the Tualatin basin is consistent with an episode of late Eocene extension documented elsewhere in the Coast Ranges. The present fold and thrust geometry of the Tualatin basin, the result of Neogene compression, is superimposed on the ancestral pullapart basin. The present 3-D basin geometry may imply stronger ground shaking along basin edges, particularly along the concealed northeast edge of the Tualatin basin beneath the greater Portland area.on April 4, 2014 geosphere.gsapubs.org Downloaded from