Geodetic evidence for lower crustal magma withdrawal during the 2009 eruption of Redoubt Volcano, Alaska

Redoubt volcano, on the western side of Cook Inlet about 100 miles WSW of Anchorage, Alaska began erupting in March 2009. The eruption continued for nearly 3 months, and slow dome growth may still persist. No continuously recording GPS instrumentation existed with 25 km of Redoubt at the beginning of major precursory unrest in January 2009. The closest CPGS instrument at that time was the Plate Boundary Observatory (PBO) backbone station AC17, about 27 km northeast of the volcano's summit. A small GPS campaign network, consisting of about 15 benchmarks, had been established at Redoubt in 2001 and had been partially reoccupied in 2008. In response to the precursory unrest, the Alaska Volcano Observatory deployed continuously recording GPS instruments at five of the campaign benchmarks, though only one of these was telemetered. Several distinct signals appear in the GPS time series, suggesting an interplay of at least two sources ranging in depth from the lower crust to within the volcanic edifice. The most remarkable of these signals, measured more than 25 km from Redoubt at AC17, shows a movement down and toward the volcano coincident in time with the initial onset of extrusion in late March, but ending well before the emplacement of the large, 70 million cubic meter lava dome through mid-April to mid-May that culminated the eruption. Closer stations show an exponentially decaying pattern of deflation that seems to follow the temporal pattern of dome growth. These contrasting styles and scales of deformation almost certainly indicate multiple sources operating over a range of depths. The rapid augmentation of the Redoubt geophysical network with CGPS proved quite useful, not just from the standpoint of engendering scientific research, but also from the perspective of providing short-term forecasts of volcanic hazard. As demonstrated during the recent eruption of Redoubt, as well as at other volcanoes in Alaska and elsewhere, we argue that routine use of CGPS on stratovolcanoes is an investment well worth making. The initial outlay of funding, logistics, and general effort involved with building CGPS instrumentation and telemetry infrastructure has now paid off handsomely at three volcanoes in the last five years: Augustine in 2006, Okmok in 2008, and now Redoubt in 2009. Of course, experience has shown that deploying CGPS instrumentation before unrest, as at Augustine and Okmok, is vastly preferable to a hasty after-the-fact deployment, which is inevitably more dangerous to install, subject to much more inflexible logistical constraints, and is likely to image only a fraction of the total deformation signal in evidence over the entirety of an eruption.