Using Synthetic Aperture Radar (SAR) interferometry, we detect several centimeters of uplift that accumulated during two years (1993-1995) around the vent of the New Trident volcano in Alaska's Katmai National Park. The areas of image coherence correspond to fresh, blocky lavas, while coherence is lost in ashcovered areas. From the uplift gradient we estimate the depth of a pressure source under New Trident volcano to be approximately 0.8-2.0 km. Our results show that in spite of the difficult sub-arctic environment of southern Alaska, strain build-up can be monitored over a twoyear period, showing the potential for global monitoring of volcano deformation using SAR interferometry. Introduction SAR interferometry measures the phase difference between two images taken on different satellite passes, corresponding to the change in the round-trip path length of radar waves to the same ground point. The phase differences are displayed in an interferogram, effectively a contour map of the change in distance to the ground surface along the look direction of the satellite. Each cycle of phase, or fringe, in the resulting interferogram (after removal of the "fiat-earth" effects) corresponds to a change in range distance from the satellite to the ground surface equal to one-half of the radar wavelength (28.3 mm for ERS-1). We represent fringes in the interferograms hown here as red-green-blue (RGB) color bands. Each RGB band corresponds to a range change of 28.3 mm. SAR interferometry has been used to map Many volcanic eruptions are preceded by pronounced displacements caused by earthquakes [e.g., Massonnet ground deformation in response to increasing pressure et al., 1993; Peltzer and Rosen, 1995], and volcanic dein subsurface magma chambers, or the upward intrufiation following the eruption of Etna volcano [Global sion of magma beneath the volcano. Therefore, sysVolcano Monitoring, 1994; Massonnet et al., 1995]. tematic geodetic surveillance and monitoring might enThe interferometric phase difference is controlled by: able eruptionsto be forecast [e.g., Swanson et al., 1985]. a) topography, b) ground-surface deformation, c) atThe logistical difficulties of surface-based measurements mospheric path delays, and d) noise sources including on Alaskan volcanoes make the application of satelliteradar receiver noise, and environmental changes such based monitoring techniques uch as SAR interferomas those caused by vegetation, snow accumulation and etry highly desirable. We chose the Katmai volcano melting, and rain events. The topographic component group, located on Katmai National Park of the Alaska of the interferogram scales with the baseline separation Peninsula, as the site for our investigation because sev(the perpendicular distance between two satellite passes eral major eruptions have happened in this area this at the times of the radar observations), while the comcentury, including the largest eruption of this century ponent due to ground surface movement is independent from Novarupta Dome in 1912 [$imkin and $iebert, of the baseline. When the spatial baseline approaches 1994]. New Trident volcano (Figure 1) is part of the zero, the fringes in an interferogram are completely conKatmai group of volcanoes, and is located 5 km from trolled by ground movement and contain no information Novarupta Dome. The vent of New Trident volcano proabout the topography if noise sources are negligible, the duced eruptions from 1953 to 1968 [Ray, 1967; $imkin ideal case for measuring surface deformation. When the and $iebert, 1994]. These eruptions have Volcanic Exbaseline separation is not zero, the component of the plosivity Indices of 2 to 3, and are termed as "moderphase signal contributed by ground movement can be ate" to "moderate-large" eruptions [$imkin and $iebert, obtained by removing the topographic component us1994]. The lavas produced 'by these eruptions have ing a pre-existing DEM or another interferogram [e.g., rugged surfaces comprised of 1-3 m blocks, and have Massonnet et al., 1993, 1995]. not yet undergone significant weathering. •Now at Hughes STX Co., EROS Data Center, Sioux Falls, South Dakota Copyright 1997 by the American Geophysical Union. Paper number 97GL00539. 0094-8534/97/97GL-00539505.00 Analysis In order to measure the long-term deformation of volcanoes using SAR interferometry, it is necessary that phase coherence be maintained over a long period of time. This requires that the scattering characteristics of
[1]
P. Rosen,et al.
Surface Displacement of the 17 May 1993 Eureka Valley, California, Earthquake Observed by SAR Interferometry
,
1995,
Science.
[2]
D. Massonnet,et al.
Deflation of Mount Etna monitored by spaceborne radar interferometry
,
1995,
Nature.
[3]
K. Feigl,et al.
The displacement field of the Landers earthquake mapped by radar interferometry
,
1993,
Nature.
[4]
Howard A. Zebker,et al.
Decorrelation in interferometric radar echoes
,
1992,
IEEE Trans. Geosci. Remote. Sens..
[5]
E. Rodríguez,et al.
Theory and design of interferometric synthetic aperture radars
,
1992
.
[6]
P. Ward,et al.
Seismic evidence for magma in the vicinity of Mt
,
1991
.
[7]
C. Newhall,et al.
Forecasts and predictions of eruptive activity at Mount St. Helens, USA: 1975-1984
,
1985
.
[8]
Giovanna Berrino,et al.
Ground deformation and gravity changes accompanying the 1982 Pozzuoli uplift
,
1984
.