A New Model for Episodic Caldera Deformation at Yellowstone

For nearly 90 years, geodetic measurements at Yellowstone have shown recurring episodes of uplift and subsidence confined mostly to the caldera but also extending into the Norris Geyser Basin. The most recent such episode began in late 2004 with the onset of caldera-wide uplift that continued for about 5 years before switching to subsidence in late 2009. The physical mechanism driving the deformation is unknown, though several researchers have proposed kinematic models that can reproduce the observed data. The "Lake" earthquake swarm, which occurred in the northern part of Yellowstone Lake from December 2008 through January 2009, provides a new constraint on caldera deformation models. The timing of the swarm correlates with an abrupt change in local deformation, which preceded the gradual transition from uplift to subsidence in late 2009. Thus, caldera deformation, at least in the vicinity of Yellowstone Lake, consists of two (or more) distinct parts, implying the existence of two (or more) distinct deformation sources. This fresh information leads us to propose a new kinematic model for deformation at Yellowstone, which we develop from the last 15 years of continuous GPS and InSAR data. Our new model consists of three deformation sources: (1) a cauldron block source that is subject to a constant displacement at its base while its surrounding ring fault remains locked; (2) a pressurizing (or depressurizing) spherical cavity near the Norris Geyser Basin, which is known to deform separately from the caldera; and (3) a pressurizing (or depressurizing) spherical cavity at the Sour Creek Dome, which we infer from the abrupt change in deformation rate after the Lake Swarm. We use the GPS and InSAR data from the period of strongest signal, summer 2005 through summer 2007, to optimize the geometry of the three sources: the locations and depths of the spherical cavity, and the perimeter of the cauldron block. We then, while holding their geometry fixed, estimate the strengths of these sources as a function of time. We show that our inferred model can reproduce the observed data well from 1996 through the present. Although our kinematic model does not entail a particular physical interpretation, it does fit with a simple mechanical account of the deformation cycle: a steady supply of fresh magma that underplates the silicic magma body brings with it abundant volatiles that make their way upward to the base of the cauldron block. Pressure increases there, creating elastic deformation of the overlying block, which does not behave as a simple piston because its sides are not free to slip. Eventually enough pressure builds to create a fracture, an earthquake swarm ensues, pressure is released abruptly at the fracture's tip, and then slowly bleeds off from the top of the magma body. Uplift then gives way to subsidence, the fracture heals via precipitation of hydrothermal minerals, and the cycle begins anew.