The Geodesy Advancing Geosciences and EarthScope (GAGE) Facility Global Positioning System (GPS) Data Analysis Centers produce position time series, velocities, and other parameters for approximately 2000 continuously operating GPS receivers spanning a quadrant of Earth's surface encompassing the high Arctic, North America, and Caribbean. The purpose of this review is to document the methodology for generating station positions and their evolution over time and to describe the requisite trade‐offs involved with combination of results. GAGE GPS analysis involves formal merging within a Kalman filter of two independent, loosely constrained solutions: one is based on precise point positioning produced with the GIPSY/OASIS software at Central Washington University and the other is a network solution based on phase and range double‐differencing produced with the GAMIT software at New Mexico Institute of Mining and Technology. The primary products generated are the position time series that show motions relative to a North America reference frame and secular motions of the stations represented in the velocity field. The position time series themselves contain a multitude of signals in addition to the secular motions. Coseismic and postseismic signals, seasonal signals from hydrology, and transient events, some understood and others not yet fully explained, are all evident in the time series and ready for further analysis and interpretation. We explore the impact of analysis assumptions on the reference frame realization and on the final solutions, and we compare within the GAGE solutions and with others.

The Yellowstone caldera began a rapid episode of ground uplift in mid-2004, revealed by Global Positioning System and interferometric synthetic aperture radar measurements, at rates up to 7 centimeters per year, which is over three times faster than previously observed inflation rates. Source modeling of the deformation data suggests an expanding volcanic sill of approximately 1200 square kilometers at a 10-kilometer depth beneath the caldera, coincident with the top of a seismically imaged crustal magma chamber. The modeled rate of source volume increase is 0.1 cubic kilometer per year, similar to the amount of magma intrusion required to supply the observed high heat flow of the caldera. This evidence suggests magma recharge as the main mechanism for the accelerated uplift, although pressurization of magmatic fluids cannot be ruled out.

The Yellowstone hotspot resulted from interaction of a mantle plume with the overriding North America plate highly modifying the lithosphere by magmatic-tectonic processes and producing the 17 Ma Yellowstone-Snake River Plain (YSRP) volcanic system. The accessibility of the YSRP has allowed largescale geophysical experiments to seismically image the hotspot and to evaluate its kinematic and dynamic properties using geodetic measurements. Tomography reveals a Yellowstone crustal magma body with 8-15% melt that is fed by an upper-mantle plume extending from 80 km to 660 km deep and tilting 60º west. Contemporary deformation of the Yellowstone caldera is dominated by SW-extension at up to ~3 mm/yr, a fourth of the total Basin-Range opening rate, but with superimposed volcanic uplift and subsidence at decade scales, averaging ~2 cm/yr and unprecedented caldera uplift from 2004-2008 at up to 7 cm/yr. Convection models reveal eastward upper-mantle flow beneath Yellowstone at relatively high rates of 5 cm/yr and opposite in direction to the overriding N. American Plate. This strong flow deflects the ascending plume melt into a tilted configuration, i.e., the plume is caught in a mantle "wind". Dynamic models of the Yellowstone plume revealed relatively low excess temperatures, up to 120°K, with up to 1.5% melt, properties consistent with a weak buoyancy flux of ~0.25 Mg/s. The flux is several times smaller than for oceanic plumes, but it produced a ~600-km wide topographic ~300-m high swell. Employing the plume-geometry we extrapolated the location of the Yellowstone mantle-source southwestward to its initial position at 17 million years beneath eastern Oregon and the southern edge of the LIP Columbia Plateau basalt field suggesting a common origin. Our model suggests that the original plume head rose vertically behind the subducting Juan de Fuca plate, but at ~12 Ma it lost the protection of the subducting plate and encountered cooler, thicker continental lithosphere and became affected by the eastward upper-mantle flow. Regionally, excess gravitation potential energy of the swell drives the SW motion of the YSRP lithosphere that becomes part of a general clockwise rotation pattern of intraplate western U.S. tectonism. Our models thus demonstrate that plume-plate processes of the YSRP have "continentalized" oceanic lithosphere enhancing intraplate extension and highly modifying topography, deep into the continental interior. Our results demonstrate that the dynamic properties of the Yellowstone hotspot deserved its recognition as a "window into the Earth's interior". JVGR

Geodetic measurements of Yellowstone ground deformation from 2006 to June 2010 reveal deceleration of the recent uplift of the Yellowstone caldera following an unprecedented period of uplift that began in 2004. In 2006–2008 uplift rates decreased from 7 to 5 cm/yr and 4 to 2 cm/yr in the northern and southwest caldera, respectively, and in 2009 rates further reduced to 2 cm/yr and 0.5 cm/yr in the same areas. Elastic‐dislocation modeling of the deformation data robustly indicates an expanding sill at ∼7–10 km depth near the top of a seismically imaged, crystallizing magma reservoir, with a 60% decrease in the volumetric expansion rate between 2006 and 2009. Reduction of hydrothermal‐volcanic recharge from beneath the northeast caldera and seismic moment release of the 2008 and 2010 large earthquake swarms are plausible mechanisms for decelerating the caldera uplift and may have influenced the change in recent caldera motion from uplift to subsidence.