Identification of mine collapses, explosions and earthquakes using INSAR: a preliminary investigation

Interferograms constmcted from satellite-borne synthetic aperture radar images have the capability of mapping sub-cm ground surface deformation over areas on the order of 100 x 100 km with a spatial resolution on the order of 10 meters. We investigate the utility of synthetic aperture radar interferomehy (InSAR) used in conjunction with regional seismic methods in detecting and discriminating different types of seismic events in the context of special event analysis for the CTBT. For this initial study, we carried out elastic dislocation modeling of underground explosions, mine collapses and small (M<5.5) shallow earthquakes to produce synthetic interferograms and then analyzed satellite radar data for a large mine collapse. The synthetic modeling shows that, for a given magnitude each type of event produces a distinctive pattern of ground deformation that can be recognized in, and recovered from, the corresponding interferogram. These diagnostic characteristics include not only differences in the polarities of surface displacements but also differences in displacement amplitudes from the different sources. The technique is especially sensitive to source depth, a parameter that is crucial in discriminating earthquakes from the other event types but is often very poorly constrained by regional seismic data alone. The ERS radar data analyzed is from a M_L 5.2 seismic event that occurred in southwestern Wyoming on February 3,1995. Although seismic data from the event have some characteristics of an underground explosion, based on seismological and geodetic data it has been identified as being caused by a large underground collapse in the Solvay Mine. Several pairs of before-collapse and after-collapse radar images were phase processed to obtain interferograms. The minimum time separation for a before-collapse and after-collapse pair was 548 days. Even with this long time separation, phase coherence between the image pairs was acceptable and a deformation map was successfully obtained. Two images, separated by 1 day and occurring after the mine collapse, were used to form a digital elevation map (DEM) that was used to correct for topography. The interferograms identify the large deformation at the Solvay Mine as well as some areas of lesser deformation near other mines in the area. The large amount of deformation at the Solvay Mine was identified, but (as predicted by our dislocation modeling) could not be quantified absolutely because of the incoherent interference pattern it produce