Volcano Deformation And Subdaily Gps Products

Abstract

Volcanic unrest is often accompanied by hours to months of deformation of the ground that is measurable with high-precision GPS. Although GPS receivers are capable of near continuous operation, positions are generally estimated for daily intervals, which I use to infer characteristics of a volcano's plumbing system. However, GPS based volcano geodesy will not be useful in early warning scenarios unless positions are estimated at high rates and in real time. Visualization and analysis of dynamic and static deformation during the 2011 Tohokuoki earthquake in Japan motivates the application of high-rate GPS from a GPS seismology perspective. I give examples of dynamic seismic signals and their evolution to the final static offset in 30 s and 1 s intervals, which demonstrates the enhancement of subtle rupture dynamics through increased temporal resolution. This stresses the importance of processing data at recording intervals to minimize signal loss. Deformation during the 2009 eruption of Redoubt Volcano, Alaska, suggested net deflation by 0.05 km³ in three distinct phases. Mid-crustal aseismic precursory inflation began in May 2008 and was detected by a single continuous GPS station about 28 km NE of Redoubt. Deflation during the explosive and effusive phases was sourced from a vertical ellipsoidal reservoir at about 7-11.5 km. From this I infer a model for the temporal evolution of a complex plumbing system of at least 2 sources during the eruption. Using subdaily GPS positioning solutions I demonstrate that plumes can be detected and localized by utilizing information on phase residuals. The GPS network at Bezymianny Volcano, Kamchatka, records network wide subsidence at rapid rates between 8 and 12 mm/yr from 2005-2010. I hypothesize this to be caused by continuous deflation of a ~30 km deep sill under Kluchevskoy Volcano. Interestingly, 1-2 explosive events per year cause little to no deformation at any site other than the summit site closest to the vent. I derive evidence for a very shallow source, likely within the edifice. This work shows that network design and individual plumbing system characteristics affect the ability to detect motion on subdaily and even weekly time scales, which stresses the importance of network scale considerations.