European Space Agency

Use of SAR Data to Study ActiveVolcanoes in Alaska

K. G. Dean, K. Engle, Z. Lu, J. Eichelberger

Geophysicallnstitute, Univ. of Alaska
Fairbanks, P.O. Box 757320, Alaska 99775-7320, USA.
Tel: 1-907-474-7364; Fax: 1-907-474-7290; E-mail: kdean@gi.alaska.edu

T. Neal & M. Doukas

Alaska Volcano Observatory, US
Geological Survey, 4200 University Drive, Anchorage, Alaska 99508, USA

Synthetic Aperture Radar (SAR) data of the Westdahl, Veniaminof, and Novarupta volcanoes in the Aleutian Arc of Alaska were analysed to investigate recent surface volcanic processes. These studies support ongoing monitoring and research by the Alaska Volcano Observatory (AVO) in the North Pacific Ocean Region. Landforms and possible crustal deformation before, during, or after eruptions were detected and analysed using data from the European Remote Sensing Satellites (ERS), the Japanese Earth Resources Satellite (JERS) and the US Seasat platforms. Field observations collected by scientists from the AVO were used to verify the results from the analysis of SAR data.

The SAR data were recorded on multiple dates, from various platforms, at different frequencies and different look directions. Radiometric and geometric processing was required to analyse surface conditions in mountainous terrains on these disparate datasets. The data were filtered and re-sampled to 30 m and 90 m pixel spacing to remove most of the system noise and the large back- scatter variability which results in the 'salt-and pepper' appearance, typical of full-resolution SAR images. The data were also merged with a digital elevation model (DEM) to correct for terrain distortions.

To test the effects of resolution on our ability to detect eruption-induced landform changes, we compared analyses derived from both 30 m and 90 m data. Landforms observed on the 30 m data were also detected on the 90 m data with minimal loss of information in most cases. We conclude that SAR data with 90-m pixel size are adequate to investigate gross geomorphic features. However, when fine-scale detail is required,30 m data are preferred. Use of the 90 m data reduces processing time, optimises the use of disk space, and makes network transfers of data significantly more efficient.

At Westdahl Volcano, we examined modifications of the land resulting from the 1978 and the 1991 -92 eruptions. The source crater of the 1978 eruption was observed on a Seasat image (Fig. 1) as well as three linear structures (prongs) radiating from the crater. Evidence of mudflows observed by field parties who visited the eruption site on 6-7 August 1978 (Krafft, 1978) was not observed on the image. By happenstance, the 1991 -92 eruption was recorded on ERS-1 SAR image (Fig. 2).

Westdahl volcano 1978
Figure 1. Westdahl Volcano 1978. An explosive eruption in February formed a crater 500 m deep through glacial ice. This Seasat SAR image was recorded on 7 August 1978, and shows the crater and three prongs radiating from the crater. The 1964 lava flow is clearly visible. The image has been filtered, and resampled to a 90-m pixel spacing. Westdahl Volcano is located on Unimak Island approximately 1000 km southwest of Anchorage, Alaska.

Westdahl volcano erupted Nov. 1991
Figure 2. Westdahl Volcano erupted in November 1991. This ERS-1 SAR image was recorded on 19 January 1992, and shows the fissure that resulted from the eruption and a new lava flow. The caldera-end of the fissure is approximately 2 km north of the 1978 vent. The 1978 vent, clearly seen in Figure 1, is buried by snow but is still detectable on the ERS-1 image. The prongs observed on the 1978 image cannot be seen on this SAR image. This image has been filtered, resampled to a 90-m pixel spacing and terrain-corrected.

A new cinder cone and lava flow emanating from a fissure northeast of the 1978 cone are clearly visible. The 1978 crater, now buried by snow, was barely detectable on the 1992 image.

At Veniaminof Volcano, melt pits (Neal et al., 1996; Yount et al., 1985) that formed in an intra-caldera ice cap from eruptions in 1983-84 and again in 1993-95 were detected. In addition, we observed changes in the radar signatures of the ice on SAR images recorded during the 1993 eruption. Typically, glaciers have a dark signature in the summer and a bright signature in the winter depending upon the presence of liquid water at the surface. Anomalous radar signatures in the ice cap in November 1992, August 1993, and October 1994 may be related to melting from volcanic activity. On one of the images, ash deposits radiating away from the source vent, an intracaldera cinder cone, are evident (Fig. 3).

Veniaminof volcano erupted 23 July 1993
Figure 3. Veniaminof Volcano erupted on 23 July 1993 and continued to be active into 1995 with only periodic observations of hot spots seen on AVHRR data during the later part of this period. A series of ERS and JERS SAR images were used to monitor the eruption and to assess changes in surface morphology. All of the images were filtered, resampled to 90-m pixels and terrain-corrected. The ice cap that overlies the caldera has a dark radar signature during the summer months when liquid water is present in the ice cover. During the winter, water is frozen and the ice-radar signature is bright. This ERS SAR image, from 24 August 1993, shows the dark summer ice cap. The 1983 and 1993 eruptions emanated from a cinder cone west of the caldera centre. The ice cap adjacent to the south side of the cone melted, forming an ice pit in 1983. During the 1993 eruption additional ice melted along the east side of the cone, enlarging the ice pit in that direction. A V-shaped light area with its apex at the cone and radiating to the southeast may be ash on the snow or roughening of the ice surface from fallen ash. Veniaminof Volcano is located on the Alaska Peninsula 800 km southwest of Anchorage.

At Novarupta Volcano, experimental interferometric SAR (IFSAR) processing techniques were used to measure ground deformation and to generate DEMS. Results to date suggest a ring pattern of deformation over a 5-km region northeast of the Novarupta vent. This surface displacement may be related to approximately 5 cm of upward vertical movement over a three-year period (Lu et al, in press), however these results have yet to be verified by other ground or GPS-based geodetic techniques. A new DEM with 10-m vertical resolution and 30-m spatial resolution has also been generated from the SAR data for this area. Our intention is to refine EFSAR techniques so they may be used routinely by AVO to monitor and study volcanoes, determine the effects of eruptions, and measure crustal deformation that may precede eruptions in this remote region.

References

Kraffl M, 1989. Global Volcanism 1975-1985, The First Decade of Reports from the Smithsonian Institution's Scientific Alert Network (SEAN), L McClelland, T Simpkin, M Summers, E Nielsen, T Stein (eds.), Prentice Hall, Englewood Cliffs, NJ, Am. Geophys. Union, Washinqton DC, DD. 329-332.

Lu Z, R Fatland, M Wyss, S Li, J Eichelberger & K Dean, 1996. Deformation of Volcanic Vents Detected by ERS-1 SAR Interferometry, Katmai National Park, Alaska (Abstr. & Figs.). Prepared for Geophys. Res. Letters. Neal C A, R G McGimsev & M Doukas, 1996. Volcanic Activity in Alaska: Summary of Events and Response of the Alaska Volcano Observatory, US Geol. Surv., Open File Rep. 96-24, pp. 7-10.

Yount M E, T P Miller, R P Emanuel, R & F H Wilson, 1985. Eruption in an ice-filled caldera, Mount Veniaminof, Alaska Peninsula, in Bartsch Winider, Susan and Reed K.M., eds., The US Geol. Surv. in Alaska, Accomplishments during 1983: US Geol. Survey Circular 945, pp. 58-60.


About | Search | Feedback

Right Left Up Home ESA EOQ Nr. 53
Published September 1996.