Sar Interferometry Applications In The Philippines Using Ers-1, Ers-2 And Jers-1: Case Studies In Mayon And Taal Volcanoes
Salvador, J.H.G.1, Kanbara, H.2, Abundo, R.1, Tsukada, M.3,
Hirose, K.3, Dayao, A. 4, Corpuz, E.5and Baloloy, A. 5
1Senior Science Research Specialist, Mines and Geosciences Bureau
North Avenue, Diliman, Q.C. 1104 Philippines.
Email:bernards@pacific.net.ph
2Senior Geologist, Nippon Mining and Metals Co. Ltd.
7-10 Toranomon 2-chome, Minatoku-Tokyo, Japan.
Email: kanbara@tankai.co.jp
3Earth Remote Sensing Data Analysis Center
Forefront Tower 3-12-1 Kachidoki Chuo-ku Tokyo 104, Japan.
Email: hirose@ersdac.or.jp
4Mines and Geosciences Bureau Regional Office No. 5 Daraga, Albay, Philippines
5Philippine Institute of Volcanology and Seismology
C.P. Garcia Avenue, U.P Diliman Campus, Diliman, Q.C. 1104, Philippines
Keywords: Interferometry, Mayon Volcano, Taal Volcano
Abstract
SAR Interferometry studies were conducted in Mayon and Taal Volcanoes in the Philippines. In Mayon, we used the ERS-1 and ERS-2 data. In Taal, we used the JERS-1 SAR data. In both cases, differential SAR interferograms were successfully generated.
For Mayon Volcano, we applied the three-pass method. Three ERS-1 datasets dated 1993/11/17, 1993/10/13 and 1997/10/10 were utilized. The baseline perpendiculars are -21m for first pair and -156m for the second pair. Two coherence maps and two interferograms were used. Further processing produced a differential interferogram. The result was refined using a "helper phase". The measured displacement along the southwest flank of the volcano amounted to 6cm and roughly coincides with the tiltmeter record in the Mayon Observatory.
We also applied the two-pass method to the available pair of ERS-2 data dated 1999/03/20 and 1999/08/07. The baseline perpendicular is 105m. A coherence map and interferogram was generated. A DEM was integrated to produce a differential interferogram. The measured displacement along the southwest flank of Mayon Volcano is up to 30cm. Later in February 2000, the volcano erupted.
For Taal Volcano, the three-pass method was applied to the JERS-1 data dated 1995/11/15, 1995/02/24 and 1995/12/29. The baseline perpendiculars are 171m for first pair and 741m for the second pair. The resulting differential interferogram shows displacement of more than 10cm along the southwest flank of Taal Volcano.
1. Introduction
Recent technological advances in radar image analysis show those interferometry techniques using SAR data is able to detect changes in the earth's surface in the order of one to several mm. The principle of SAR interferometry is explained in many documents such as Dixon (ed., 1995), Gen and Van Genderen (1996), Fujiwara and Tobita (1999), and Tanaka and Nakano (1997). The technique was used to detect surface changes due to earthquakes (Massonnet et al., 1993), volcanic activity (Massonnet et al., 1995) and glacier movements (Goldstein et al., 1993) among others. The application of SAR interferometry to study deformations in the Mayon and Taal Volcanoes in the Philippines is the subject of this paper. In Mayon, we used the ERS-1 and ERS-2 data. In Taal, we used the JERS-1 SAR data. In both cases, differential SAR interferograms were successfully generated.
This paper presents some results of the three-year Joint Study called RP-ERSDAC Remote Sensing Project. This collaboration project involves the Earth Remote Sensing Data Analysis Center (ERSDAC) of Japan, the Philippine Mines and Geosciences Bureau (MGB), the Philippine Institute of Volcanology and Seismology (PHIVOLCS) and the Philippine National Oil Company-Energy Development Corporation (PNOC-EDC). The project is part of ERSDAC's "Technology Development for Phased Array type L-band Synthetic Aperture Radar (PALSAR) data": ERSDAC designated Nippon Mining & Metals Co., Ltd. (NMM) as the contractor to implement the work.
2. Brief Geologic Setting Of The Study Areas
Taal Volcano near Manila and Mayon Volcano in Albay belong to the Philippine Mobile Belt described by Gervacio (1966, 1971). To the east of the study areas, we find the west subducting Philippine Trench. To the west, we find the east subducting Manila Trench. In between, we find the 1,200-km-long left-lateral strike-slip Philippine Fault cutting across the Philippine Archipelago (Aurelio, 1997). Two volcanic rows, the western and the eastern belts, correspond to these sinking plates. The western belt runs from Mindoro Island to the Verde Island Passage. This belt includes Taal and Pinatubo Volcanoes. The eastern belt corresponds to the Philippine Sea Plate. It runs from the Bicol Peninsula in Luzon down to Mindanao. It includes Mayon and Bulusan Volcanoes.
2.1 Mayon volcano
Mayon Volcano is an active volcano with highest elevation of about 2,462masl. It has a long history of eruptions. The first recorded was in 1616 and the latest in February 2000. It has erupted 45 times. The most destructive took place on February 14, 1814, killing at least 1,500 people (Moore and Melson, 1969). The eruption cycle is from 10 to 12 years. It erupted in the years 1928, 1938, 1947, 1968, 1978, 1987,1993. During the last eruption, more than 5,000 people were evacuated.
2.2 Taal Volcano
Taal Volcano, located 40-km southeast of Manila, is one of the most active volcanoes in the world. It is composed of between two and four coalesced lake-filled calderas with a central active volcanic center called the Volcano Island. The volcanic center has erupted about 40 times since 1540. Major eruptions occurred in 1574, 1911, and 1965 resulting in several thousand deaths. Taal Volcano has been restless since early 1991, with earthquake swarms, new steaming, and rise in temperature of the main Crater Lake.
In 1994, an intrusion of magma was observed underneath Taal near the Crater Lake. The Crater Lake's water temperature rose and high temperature volcanic gas were emitted. The gas contained much SO
2 and the water in the crater lake became strongly acidic. The earthquake swarm concentrated in the Crater Lake and in the Calawit district in the vicinity. By this activity, the southeast part of the Volcano Island upheaved by about 15cm. Presently, Taal volcano is Alert Level Zero, meaning there is no danger of the eruption.
3.0 Differential SAR Interferometry
For this study, we used the repeat pass method that utilized two spatially close radar observations of the same area separated in time and using only one antenna or two different sensors having nearly identical radar system parameters. The repeat pass method can be "two-pass" or "three-pass". The former needs a digital elevation model (DEM) data to simulate the topographic fringe. The latter has the advantage of removing the topographic fringe element by using two interferograms.
Four JERS-1 SAR images for Mayon and eight for Taal were accumulated during the six and a half years period from 11 February 1992 (JERS-1 satellite launch) until 12 October 1998 (JERS-1 stopped). Since then, the ERS SAR was used to continue observing the study area. The ERS data is being received at the Chung-Ii station (Space and Remote Sensing Research Center, National Central University) in Taiwan.
Generally, the suitable pairs for SAR interferometry are those where the baseline perpendicular (Bp) distance and the acquisition interval are both short. These data were prioritized and selected. From them, we generated the initial interferograms. We processed the data using the VEXCEL 3D SAR Processing Software Ver. 2.27 in the Unix Platform. The DEM in each study area was digitized from existing topographical maps published by National Mapping & Resource Information Authority (NAMRIA) of the Philippines.
For Mayon Volcano, we applied the three-pass method on the first set of data pairs. Three ERS-1 datasets dated 1993/11/17, 1993/10/13 and 1997/10/10 were utilized. We also applied the two-pass method to the second set of ERS-2 data pair dated 1999/03/20 and 1999/08/07 and a DEM. For Taal Volcano, the three-pass method was applied to the JERS-1 data dated 1995/11/15, 1995/02/24 and 1995/12/29. Global Positioning System (GPS) measurements were made over Mayon Volcano during 1997 to 1999. However, time and space does not allow presentation of the GPS measurement data and results in this occasion.
