Augmenting the Persistent Scatterer Method with Models of Atmospheric Phase Lag to Image Pre-eruptive Deformation at Mount St Helens

In October of 2004, Mount St Helens erupted, ending the period of quiescence that followed its famous eruption in 1980. The 2004 eruption was much smaller than in 1980, and initiated by gas and steam explosions which continued intermittently through 2005. The primary style of activity, from 2004 to 2008 was a slow, viscous extrusion of lava which formed new lava domes within the crater of Mount St Helens. Indicators of unrest, including increased seismicity, and deformation of the earth’s surface usually precede volcanic eruptions, as was the case in 1980. While a swarm of earthquakes was detected beneath Mount St Helens weeks before the 2004 eruption, no surface deformation on the spatial scale of tens of kilometers was detected in the decade prior to its onset. If pre-eruptive deformation did in fact occur on a smaller spatial scale, localized to the edifice or crater, the resulting displacement of the ground surface should be measureable by InSAR techniques. The timing, magnitude, location, and shape of pre-eruptive deformation at a volcano lend insight into the dynamics of magma at depth and also the temporal relationship between magma recharge and the eruptive cycle. The ability to observe and study volcanic deformation at Mount St Helens is important answering various volcanological questions and also to volcano monitoring and hazard mitigation.

There are, however, significant barriers to applying InSAR methods to Mount St Helens or any volcano in the Pacific Northwest. Any potential signal in interferograms due to ground displacement is often obscured by phase decorrelation caused by snow and trees, or affected heavily by atmospheric phase lags. In an effort to detect deformation at St Helens, I have implemented a relatively new InSAR processing technique called StaMPS (Stanford Method for Persistent Scatterers), which helps to overcome the problem of phase decorrelation due to trees and snow. I have applied this technique to 8 SAR scenes from the ERS-2 satellite. The chosen scenes are those with the best perpendicular baselines from the summer and early fall (to avoid snow), and are spread over a large portion of the pre-eruptive period (1996-2002).

StaMPS processing appears to do a good job of overcoming phase decorrelation, but the resulting maps of surface velocities and displacements are clearly affected by atmospheric phase lag, which is known to correlate strongly with surface elevation. The atmospheric signal, which masks potentially observable ground displacement, is largely due to temporal changes in the water content of the atmosphere, and possibly unaccounted for masses of water bearing air which arise from topographic corrections. In order to make more confident conclusions about the existence of pre-eruptive deformation at Mount St Helens, I plan to estimate the spatial and temporal contribution of the atmosphere to interferometric phase. This will be done using simple models which incorporate digital elevation models and remote-sensing measurements of various atmospheric characteristics, including vertical profiles of water content and temperature. Sub-daily atmospheric data from NASA’s Moderate Resolution Imaging Spectroradiometer (MODIS) are available for the area surrounding Mount St Helens from 2000 to the present. Although atmospheric data are not available over the entire study period, a generalized atmospheric phase model can be applied within StaMPS processing and can provide valuable constraints and insight into the effect of atmosphere on Interferometric products.