Separating long-term deformation cycles and atmospheric signals at Mount St. Helens using PS-InSAR

Abstract:

Since its eruption in 1980, Mount St. Helens has experienced multiple inflation-deflation cycles associated with dome building eruptions. During the most recent dome-building episode, which spanned 2004 to 2008, GPS recorded the transition from pre-eruptive inflation to co-eruptive deflation and a final transition back to inflation. Such observations provide important constraints on the timing and mechanics of cyclic magma recharge and extrusion. Currently, the subtle surface deformation signal at St Helens is monitored primarily by ground based geodetic techniques like GPS. Satellite-based InSAR has the potential to substantially augment these techniques by providing spatially continuous, precise measurements of surface displacements, and may also reveal other volcanic or surficial processes too localized to be detected by ground based methods.

Traditional interferometry is challenging to apply to volcanoes in the Cascades. Widespread phase decorrelation caused by persistent snow cover and dense vegetation, combined with large, elevation dependent atmospheric phase delays, mask or make deformation signals difficult to detect. By applying StaMPS, a Persistent Scatterers (PS) technique, phase decorrelation is mitigated by utilizing only the pixels with the highest, statistically derived, signal to noise ratio. However, atmospheric water vapor, which delays the radar signal, remains problematic, particularly on the volcano edifice. To assess the bias imposed by the atmosphere, we perform a series of sensitivity tests using a suite of methods including several that rely on the linear or power-law correlation of phase delay to topography and knowledge of the spatial scale of the signal. We also apply methods that calculate wet and dry phase delay from atmospheric reanalysis datasets such as ERA-Interim provided by the ECMWF. SAR data from the ERS, Envisat, and ALOS satellites, along with newer datasets, are processed with these tools to create a time series spanning nearly two decades at Mount St Helens. Our results indicate that any pre-eruptive inflation signal must be subtle, at or below the level of atmospheric noise. We also successfully identify scatterers on the edifice showing that InSAR has potential to complement GPS for monitoring near field volcanic deformation in the Cascades.