MONITORING OF NISSYROS VOLCANO (GREECE) USING ERS REPEAT-PASS INTERFEROMETRY

Transcription

1 FINAL REPORT ESA PROJECT AO MONITORING OF NISSYROS VOLCANO (GREECE) USING ERS REPEAT-PASS INTERFEROMETRY Dr Athanassios Ganas ATHENS May 8, 2002

2 Principal Investigator s Addresses: Dr Athanassios Ganas, Institute of Geodynamics, Lofos Nymfon, Thission, PO Box 20048, Athens , Greece. Tel: / 195 Fax: aganas@gein.noa.gr Table of Contents ABSTRACT... 3 A. INTRODUCTION... 4 Principles of Interferometry... 7 B. DIFFERENTIAL INTERFEROMETRIC SAR METHODOLOGY... 9 C. DIFFERENTIAL INTERFEROMETRY RESULTS...13 D. DIFFERENTIAL SAR: DISCUSSION-CONCLUSIONS...15 E. DEM EXTRACTION FROM TANDEM ERS PAIRS...18 ACKNOWLEDGEMENTS...24 REFERENCES...25 APPENDIX A APPENDIX B ERS1/ERS ERS2/ERS ERS1/ERS ERS2/ERS APPENDIX C APPENDIX D

3 ABSTRACT Nissyros is an active volcano in the south Aegean Sea, Greece. Deformation maps at a resolution of 20 metres were produced using ERS C-band SAR image pairs acquired in 1995/96, over a time period of 279 days. The interferograms show one, well-developed fringe trending E-W and a second less developed fringe in the northwestern part of the island. This is the earlier deformation signal ever mapped on Nissyros. These results clearly indicate significant off-caldera ground deformation which is interpreted as indicating a minor inflationary phase of the volcanic centre of Yali - Nissyros. Surface uplift occurred in Mandraki area (NW Nissyros), and the affected area coincides with a particular geologic discontinuity, a normal fault that moved during the August Ms earthquake. Derived uplift rates are supported by the results of geodetic measurements (GPS) during the years Successful extraction of interferometric DEMs from 1995 tandem pairs is also reported. 3

4 A. INTRODUCTION The Nissyros volcano is located at the southeastern end of the South Aegean Volcanic Arc at about 36.5 degrees North and 27.1 degrees East (Georgalas, 1962; Figure 1). Although the last volcanic activity on Nissyros dates back at least years, the present geodynamic activity encompasses high seismic unrest, widespread fumarolic activity, and recent gas and hydrothermal explosions (Papadopoulos et al., 1998; Dietrich et al., 1998). Earthquakes and steam blasts accompanied the most recent hydrothermal eruptions in and Mudflows and hydrothermal vapors rich in CO 2 and H 2 S were emitted from fracture zones that cut the caldera of the volcano and extend towards NNW through the vicinity of the village of Mandraki. The seismic unrest of the years prompted us to initiate an interferometric study of the ground deformation using spaceborne data from the European satellites ERS1 and ERS2. Differential Interferometry has been an effective tool in mapping ground deformation of active volcanoes (e.g. Massonnet et al., 1995; Lanari et al., 1998; Lu et al., 2000). We used ERS data originated in the I-PAF facility (Fucino, Italy) in CEOS- SLCI (Single Look Complex) format with dimensions 4900 pixels (range) by approximately rows (azimuth) resulting in ground resolution 7.9 metres in range metres in azimuth. The orbital characteristics of the data are shown in Table 1. The Pulse Repetition Frequency (PRF), a critical factor in data processing (Gens, 2000) for both satellites is Hz. The radar wavelength is 5.65 cm (frequency 5.3 GHz) and the antenna depression angle is 67 degrees during data 4

5 acquisitions. The ERS repeat interval is 35 days except for the tandem missions (see section E below). The radar data were processed by the software EVInSAR, (1.2.1 version) made by ATLANTIS (van der Kooij, et al., 1997) on a Pentium III computer with 392 Mb RAM. EVInSAR 1 requires large amounts of RAM to run so 1 Gb of virtual memory had to be reserved. The software can be operated interactively through a Graphical User Interface (GUI) or on a command line to allow batch processing. Most data processing was performed through the GUI. Operating system environments used were Windows NT and Windows Table 1. The ERS SAR scenes used in the project. ID refers to CD production by the data provider. GS is ground station, PAF is Processing-Archiving Facility, FS is Fucino, Italy. The tools used to select the orbits below are shown in Appendix A. All orbits considered for InSAR are shown in Appendix B. ID Satellite Sensor Prod GS PAF Orbit Frame Date ERS1 SAR SLCI FS IP ERS1 SAR SLCI FS IP ERS1 SAR SLCI FS IP ERS1 SAR SLCI FS IP ERS1 SAR SLCI FS IP ERS1 SAR SLCI FS IP ERS2 SAR SLCI FS IP ERS2 SAR SLCI FS IP ERS2 SAR SLCI FS IP ERS2 SAR SLCI FS IP ERS2 SAR SLCI FS IP ERS2 SAR SLCI FS IP The software is described through a series of on-line publications in: 5

6 Figure 1. Location map of the Nissyros volcano, Aegean Sea. Inset in the left shows the study area. The image to the right is a black-and-white space photograph of the Space Shuttle. The image shows in 2D the circular geometry of Nissyros and the caldera in the middle. Nissyros is imaged by the ERS satellites in the descending mode, i.e. the satellites travel approximately from north to south and look to the west, inclined 23 down from vertical. 6

7 Principles of Interferometry The SAR imaging system uses a long, straight antenna mounted on a satellite platform with its long-axis parallel to the orbit path. The antenna emits pulses of electromagnetic energy downward to the surface of the planet, at right angles to the orbital path. The radar backscatter carries amplitude and a phase. Phase refers to the starting point of the wave relative to some time zero. Thus, the phases of images with a difference of time (i.e. one SAR acquires images at two separate orbits) can be compared after careful image registration. The resulting phase difference is a new image called interferogram. Negative phase values indicate movement towards the sensor, positive phase values away from the sensor. One phase cycle equals 2π radians or 28 mm of displacement (C band) and defines a fringe. The overall fringe pattern contains all the information on relative geometry between the two overpasses, provided a number of noisy contributions to the phase are accounted for. In particular, to map ground movements along the line of sight between the SAR and the earth s surface (Figure 2) we need to eliminate the topographic, ionospheric and orbital contributions to the phase signal. In addition, the organized fringe pattern can vanish if phase coherence between the two SAR acquisitions is lost. This is most commonly observed on dynamic surfaces, such as the sea or agricultural fields. Phase coherence is a measure of similarity between phase information acquired at different times. Coherence depends on sensor parameters (wavelength, noise), image geometry (baseline and local incidence angle) and target characteristics. For instance, in tandem acquisitions forests show low coherence because the target scatterers 7

8 (leafs, etc) change their moisture level or facing directions, whereas bare soil is stable (see Figure 6 below). The surface with the greatest variability is water. In general, the longer the time interval between phase acquisitions the less coherent the phase returns are. This is called temporal decorrelation. Figure 2. The imaging geometry for SAR interferometry. A1 and A2 represent two antennas viewing the same surface simultaneously, or a single antenna viewing the same surface on two separate passes. From trigonometry we can derive z (y) = h -?cos? where h is platform height and? is the look (incidence) angle of the radar. B is perpendicular baseline between A 1 and A 2, ρ is range (distance) to ground. 8

9 B. DIFFERENTIAL INTERFEROMETRIC SAR METHODOLOGY The ODISSEO web Server of ESA 2 (was searched for the selection of suitable pairs for InSAR. The search was completed using the DESCW software available from the Eurimage Web Site (Appendix A). Ten (10) images were selected from the baseline search (Appendix B) and were delivered as SLCI products (Table 1). Two optimal scenes were selected for further analysis, namely orbits 5973 (11 June 1996) and 1965 (5 September 1995) of the ERS 2 satellite (Table 2). These scenes had temporal separation of 279 days which span the time period of increased seismicity rates while preserving adequate phase coherence. Locally there is a high level of correlation between the two ERS scenes; however, over most of the island the coherence remains at relatively low levels. The orbital baselines (B) 3 were: perpendicular B = m, parallel B = m. The orbital parameters of the two scenes were refined by more accurate estimations of the state vector positions as calculated by the DEOS 4 (Delft Orbits; Scharroo et al., 1998). We used the 2-pass method (Massonnet et al., 1993) by eliminating topographic phase using an external DEM with a resolution of 2 metres. First we produced a 20-m DEM (see Appendix C) that represents overall topography satisfactorily, however, it was found to locally introduce many artifacts in spaces between contour intervals. The finer DEM was produced by digitizing of the 1:5000 topographic maps of Nissyros provided by HMGS (Hellenic Military Geographical Service). The DEM was edited and corrected with field measurements by the Baseline is the spatial displacement of the antenna between the two overpasses

10 Institute of Cartography, ETHZ at a final resolution of two (2) metres (Vassilopoulou and Hurni, 2001). This good quality had two implications: a) it provided ground control for the co-registration between master SAR scene - slant range projected DEM between 1-4 m and b) brought down the magnitude of the topographic artifact to 4 m / 117 m (altitude of ambiguity 5 ha) or of a cycle or 1mm using Pair 1. All datasets were reprojected to the WGS84 datum, UTM zone 35. Table 2. Best Interferometric Pair Validation results. The first stage of SAR image processing after data input. Spectral overlap between two SAR scenes indicates amount of noise introduced to the phase difference of the two interferograms because of their bandwidths centered at different frequencies. SAR Image Sensor-Orbit Date Spatial Overlap Range Spectral Azimuth Spectral Pair 1 Master ERS % 92.71% 98.99% Pair 1 Slave ERS This is the height sensitivity (in metres) of the pair or Dh/Df = {2 * B * 1 / l * r * sinq) -1 See Figure 2 for definitions. 10

11 Figure 3. The interferogram of Nissyros volcano spanning the period between orbits 5973 (11 June 1996) and 1965 (5 September 1995) of the ERS 2 satellite. The sensitivity to topography of this pair (ha) is 114 m. Note the E-W strike of the first fringe (black arrow - blue to red colour 28 mm change), that is across topography. A second fringe can be seen at about 3 km to the North near the Mandraki Loutra coastal area. The interferogram depicts line of sight displacements between the two SAR acquisitions and is geocoded. SAR data processing starts with image co-registration analysis. This includes: a) Resampling of DEM to master SAR image 11

12 b) Processing to distort the re-sampled DEM into the master SAR space creating foreshortening and layover distortions c) Fine coregistration of DEM to master SAR image using ground control points at well-defined bright pixels (spots) and d) Calculation and subtraction of topographic phase from the interferogram prior to interferogram enhancement. e) Calculate coherence image f) Unwrap the phase g) Convert unwrapped phase to height h) Correct for terrain distortions and geocode In step c) an advanced SAR image backscatter simulator is used to generate a simulated SAR image for use in fine coregistration of DEM to master. For example, the fine coregistration for the Pair 1 was done using 23 ground control points with root mean square errors along x=0.54 and along y=0.66 pixels. The flat earth phase is removed during interferogram generation. The phase difference (Figure 3) is unwrapped using an iterative disk masking algorithm with controlled error propagation and error correction. The algorithm uses multiple tiling and seeding techniques, providing automatic masking of low coherence areas and automatic connection of residues. An unwrapping, interactive editor was also used to correct remaining errors. Data resampling to 20 m (azimuth) 20m (range) is applied to reduce the amount of data to be phase unwrapped and give square dimensioned pixels in ground range. 12

13 C. DIFFERENTIAL INTERFEROMETRY RESULTS The interferometric phase is shown in Figure 3. Because of the cyclical nature of phase, colour is used to represent it. A value of zero is represented by red and as the phase increases it is represented by smooth transition from red to green to blue and then back to red as it reaches 2π (i.e. one cycle). Independent phase effects, which may be present in such an interferogram, are due to topography, groundwater content, surface change and atmospheric anomalies. Because the baseline of Pair 1 is relatively small (80 m) for ERS the phase is not highly sensitive to the topography (one cycle of phase corresponds to about 117 m in height variation). Despite that the maximum height variation in the region is about 700 m the contribution to the phase due to topography was considered negligible relative to other effects, such as deformation, in the interferogram. This is because topography follows a conic pattern (Figure 1 right; Appendix C) whereas the main fringe trends approximately eastwest. Other possible causes such as atmospheric perturbation or groundwater table fluctuations are also excluded because of the geometry of the main fringe. Therefore, there is a significant indication of surface deformation (uplift) at the study area. A nearly circular phase pattern can be seen in the top near the middle of the scene. Considering the baseline for this InSAR pair this pattern cannot be accounted for by the topography known to be present. By unwrapping the phase and applying a straightforward conversion a height change map is produced. Note that this conversion assumes that the phase effect measured is due to height change only (no lateral movement). It is important to recognize that the SAR measures only the component of surface motion along the 13

14 viewing direction of the antenna. The actual surface motion must be inferred from additional knowledge about the area of interest. In the case of the Nissyros volcano, given the relatively conic topography (Figure 4) and the nature of the hydrothermal activities, it is assumed that the surface motion is height change. Figure 4. Field photograph of the Nissyros caldera showing the Stefanos crater (circular feature at the bottom) and the steep slopes of the Profitis Elias where layover occurs in ERS SAR images during daytime, descending orbits. Lithology is mainly rhyodacites. The highest peak is 698 m high. The Stefanos crater is at 120 m above mean sea level. View to the west. 14

15 D. DIFFERENTIAL SAR: DISCUSSION-CONCLUSIONS The nearly circular feature that can be clearly seen in the Interferogram (Figure 3) is most likely due to uplift. The uplift is due to volcanic activity and not to tectonic uplift by many earthquakes (Figure 5) occurred around Nissyros, as none of them had enough magnitude to generate surface deformation (see Appendix D). Note that no height change measurements were recovered for most of the bottom of the caldera. This was because the phase unwrapping failed due to a) prohibitively low coherence as the central-northern areas are primarily agricultural fields and b) geometrical effects (layover of the N-S slopes, i.e. perpendicular to the radar beam during descending orbits). The above results were compared to the pattern of deformation measured by local GPS surveys from 1997 onwards. During the last four years, a series of GPS measurements were made of the heights of 14 fixed points distributed throughout Nissyros (Lagios et al., 1998; Lagios, 2000). Analysis of the results provided a preliminary estimate of surface uplift with millimeter accuracy. Comparison of the survey results to those of the 1995/96 InSAR deformation map is favourable, as the GPS data indicate up to 70 mm of uplift. Therefore, areas affected by uplift are clearly evident from the interferogram. 15

16 Figure 5. Seismicity map of the Nissyros area (Data from the NOA National Network). Notice the 2 strong earthquakes of Magnitude 5 near the west coast. The cause of the uplift observed by differential InSAR is most probably upward magma movement at the eastern edge of the south Aegean volcanic arc. This movement was accompanied by extensive seismic activity (32 events - Figure 5) which includes two (2) intermediate-depth earthquakes of magnitude 5.5 Ms on 30/11/1995 and 12/4/1996, respectively. This correlation between volcanic and seismic activity has been recently established worldwide by Marzocchi (2002). In the Nissyros case, the triggering mechanism seems to be stress transfer that increased the pressure on the magma chamber which penetrates 16

17 through the upper crust up to depths of 1-2 km (Nikolova et al., 2002). However, most of the earthquakes are shallow (see Appendix D) and are concentrated along a NE-WS to E-W general direction which suggests that they occurred within the Kos neotectonic graben (Papanikolaou and Nomikou, 2001). It is worth pointing out that the area of greatest uplift seen in Figure 3 is the northwestern part of the island (which probably includes a significant offshore part) where the destructive, August 1997 earthquake sequence occurred. Our master SAR scene ( ) was also used by Parcharidis and Lagios (2001) in their study of the surface deformation of Nissyros. The 3-5 cm vertical deformation mapped by our interferogram (Figure 3) was followed during the years by mostly horizontal movements to the east and southwest, respectively. This change in the mode of deformation could be due to the activation of surface fault zones that cross-cut the volcano. 17

18 E. DEM EXTRACTION FROM TANDEM ERS PAIRS Interferometric SAR can be also used to extract topography with reasonable errors (10-30 m; Kenyi and Raggam, 1997). To test this capability we produced an interferometric DEM from a tandem ERS pair, orbit of ERS 1 (31 July 1995; master image) and orbit 1464 of ERS 2 (1 August 1995; slave image). Their scene parameters are reported in Tables 4 and 5, respectively. The spatial overlap of this pair was 98.3%, the spectral overlap in range is 96.3% and in azimuth 76.2%. The baselines were relatively small (perpendicular B = 40 m, parallel B = 26 m) which implied a rather long height difference per one phase cycle (2π) or m. Practically, one fringe represents about 240 metres in elevation. The master to slave SAR image registration (along range) was achieved using 37 valid tiepoints with an RMS error of 0.09 pixels. On the other hand, the coherence image (Figure 6) shows that a good correlation exists which was not the case for other tandem pairs. The wrapped phase image is shown in Figure 7 (bottom) in slant-range view. The interferogram is quite clear with a well developed pattern of fringes except in the layover region, inside the caldera. Phase unwrapping was done using the iterative disk masking method with the coherence threshold set to The extracted DEM is shown in Figure 7 (top) with intensity values increasing towards the highest elevations. The model has a pixel size of 20 metres. For comparison, a digitized DEM (20-m) of Nissyros is shown in Appendix C. 18

19 To check the vertical accuracy of the DEM we measured heights from 5 ground control points against those of the tandem DEM (Table 3). The points are located at topographic highs on the 1:10,000 map (see Figure 6 for locations). The RMS error of the height difference is 15.6 metres which is satisfactory given the rough terrain of Nissyros. Given that the orbital parameters used in processing were those of CEOS and not any refined ones we estimate that our RMS error is an overestimation. Table 3. Quality control of the ERS tandem DEM Number of elevation control points: 5. The location of points is shown in Figure 7 (top). All figures refer to WGS84 datum, UTM projection zone Point Feature GCP GCP GCP GCP DEM GCP Height Error Row Col Lat Long Height Height (DEM Height - GCP Height) (degrees) (degrees) (m) (m) (m) (I) Yes Yes (I) Yes (I) Yes (I) Yes

20 Figure 6. Phase coherence image of the ERS tandem DEM This product is an 8-bit image (see side scale), therefore a perfectly coherent value would be 255 (bright) while a zero coherence would be 0 (dark). Image is shown in slantrange and is flipped vertically. Notice that coherence is lost on water and on steep slopes facing the radar beam. 20

21 Figure 7. Interferometric DEM of Nissyros (top) and wrapped phase (bottom). The GUI of EVInSAR is also shown. Bottom image is slant-range and flipped vertically. The colour wheel defines one fringe or approximately 250 m of topography. 21

22 Table 4. Tandem pair 31 July 1995 (ERS-1) scene parameters report. Read_CEOS_data_from_KV = 0 CEOS_DATA_SET_SUMMARY. Processing_system_identifier = VMP CEOS_DATA_SET_SUMMARY. Processing_facility_identifier = I-PAF CEOS_DATA_SET_SUMMARY. pixel_spacing = CEOS_DATA_SET_SUMMARY. Pulse_repetition_Frequency = CEOS_DATA_SET_SUMMARY. Line_spacing = CEOS_DATA_SET_SUMMARY. Scene_centre_time = CEOS_DATA_SET_SUMMARY. Total_processor_bandwidth_in_range = CEOS_DATA_SET_SUMMARY. Total_processor_bandwidth_in_azimuth = CEOS_DATA_SET_SUMMARY. Range_sampling_rate = CEOS_DATA_SET_SUMMARY. Nominal_resolution_in_range = CEOS_DATA_SET_SUMMARY. Nominal_resolution_in_azimuth = CEOS_DATA_SET_SUMMARY.Zero_doppler_range_time_of_first_valid_range_pixel = CEOS_DATA_SET_SUMMARY.Zero_doppler_azimuth_time_of_first_azimuth_pixel = 31-JUL :53: CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_centroid_const = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_centroid_linear = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_centroid_quad = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_rate_const = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_rate_linear = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_rate_quad = CEOS_PLATFORM_POSITION_DATA. Year_of_data_point = 1995 CEOS_PLATFORM_POSITION_DATA. Month_of_data_point = 7 CEOS_PLATFORM_POSITION_DATA. Day_of_data_point = 31 CEOS_PLATFORM_POSITION_DATA. Number_of_data_point = 5 CEOS_PLATFORM_POSITION_DATA. Seconds_of_day_of_data = CEOS_PLATFORM_POSITION_DATA. Time_interval_between_data_points = CEOS_PLATFORM_POSITION_DATA.Position_vector_X0 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y0 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z0 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X0 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y0 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z0 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X1 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y1 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z1 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X1 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y1 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z1 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X2 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y2 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z2 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X2 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y2 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z2 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X3 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y3 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z3 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X3 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y3 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z3 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X4 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y4 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z4 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X4 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y4 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z4 = CEOS_IMOP_FILE_DESCRIPTOR.Number_of_left_border_pixels_per_line = 0 CEOS_IMOP_FILE_DESCRIPTOR.Number_of_right_border_pixels_per_line = 0 CEOS_IMOP_FILE_DESCRIPTOR.Total_number_of_data_groups_per_line_per_SAR_channel = 4900 CEOS_IMOP_FILE_DESCRIPTOR.Number_of_lines_per_data_set =

23 Table 5. Tandem pair 1 August 1995 (ERS-2) scene parameters report. Read_CEOS_data_from_KV = 0 CEOS_DATA_SET_SUMMARY. Processing_system_identifier = VMP CEOS_DATA_SET_SUMMARY. Processing_facility_identifier = I-PAF CEOS_DATA_SET_SUMMARY. pixel_spacing = CEOS_DATA_SET_SUMMARY. Pulse_repetition_Frequency = CEOS_DATA_SET_SUMMARY. Line_spacing = CEOS_DATA_SET_SUMMARY. Scene_centre_time = CEOS_DATA_SET_SUMMARY. Total_processor_bandwidth_in_range = CEOS_DATA_SET_SUMMARY. Total_processor_bandwidth_in_azimuth = CEOS_DATA_SET_SUMMARY. Range_sampling_rate = CEOS_DATA_SET_SUMMARY. Nominal_resolution_in_range = CEOS_DATA_SET_SUMMARY. Nominal_resolution_in_azimuth = CEOS_DATA_SET_SUMMARY.Zero_doppler_range_time_of_first_valid_range_pixel = CEOS_DATA_SET_SUMMARY.Zero_doppler_azimuth_time_of_first_azimuth_pixel = 01-AUG :53: CEOS_DATA_SET_SUMMARY.Cross_track_Doppler_frequency_centroid_const = CEOS_DATA_SET_SUMMARY.Cross_track_Doppler_frequency_centroid_linear = CEOS_DATA_SET_SUMMARY.Cross_track_Doppler_frequency_centroid_quad = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_rate_const = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_rate_linear = CEOS_DATA_SET_SUMMARY. Cross_track_Doppler_frequency_rate_quad = CEOS_PLATFORM_POSITION_DATA. Year_of_data_point = 1995 CEOS_PLATFORM_POSITION_DATA. Month_of_data_point = 8 CEOS_PLATFORM_POSITION_DATA. Day_of_data_point = 1 CEOS_PLATFORM_POSITION_DATA. Number_of_data_point = 5 CEOS_PLATFORM_POSITION_DATA. Seconds_of_day_of_data = CEOS_PLATFORM_POSITION_DATA. Time_interval_between_data_points = CEOS_PLATFORM_POSITION_DATA.Position_vector_X0 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y0 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z0 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X0 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y0 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z0 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X1 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y1 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z1 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X1 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y1 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z1 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X2 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y2 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z2 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X2 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y2 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z2 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X3 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y3 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z3 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X3 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y3 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z3 = CEOS_PLATFORM_POSITION_DATA.Position_vector_X4 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Y4 = CEOS_PLATFORM_POSITION_DATA.Position_vector_Z4 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_X4 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Y4 = CEOS_PLATFORM_POSITION_DATA.velocity_vector_Z4 = CEOS_IMOP_FILE_DESCRIPTOR.Number_of_left_border_pixels_per_line = 0 CEOS_IMOP_FILE_DESCRIPTOR.Number_of_right_border_pixels_per_line = 0 CEOS_IMOP_FILE_DESCRIPTOR.Total_number_of_data_groups_per_line_per_SAR_channel = 4900 CEOS_IMOP_FILE_DESCRIPTOR.Number_of_lines_per_data_set =

24 ACKNOWLEDGEMENTS The ERS data were provided to the author free of charge as part of AO3 series of ESA contracts awarded during 1998 (Proposal AO-3 201). I am grateful to the ERS Help desk at ESRIN and the ATLANTIS Support team in Ottawa, CA for valuable help. I am indebted to Prof. Evangelos Lagios (NKUA, Athens), Prof. Volker (Wumme) Dietrich (ETH Zurich) and Dr George Stavrakakis, Director of NOAGI, for their continuous support. I also thank Alexis Rigo (IPGP), Alain Arnaud (ESRIN), Prof Geoff Wadge (ESSC), James Ikkers (ATLANTIS), George Vouyoukalakis (IGME), Pepi Vassilopoulou (NKUA), Lorenz Hurni (ETH), Isaak Parcharidis (NKUA), Vassilis Karastathis (NOAGI) and Vassilis Sakkas (NKUA) for many useful discussions. Mr. Yannis Cobopoulos and IIS SA provided resources and space during the first stages of this project. The production of the fine DEM was funded by the GEOWARN project (IST ). The InSAR Processor is 1996 Atlantis Scientific. 24

25 REFERENCES Dietrich, V. J., Kipfer, R. and Schwandner, F Mantle-derived noble gases in the South Aegean volcanic arc: Indicators for incipient magmatic activity and deep crustal movements. Newsletter of the European Centre on Prevention and Forecasting of Earthquakes (Council of Europe) 2 (Sept. 1998), Gens, R., The influence of input parameters on SAR interferometric processing and its implication on the calibration of SAR interferometric data. Int. J. Remote Sensing, 21 (8), Georgalas, G.C., 1962, Catalogue of the active volcanoes of the world including solfatara fields; Part XII Greece: International Association of Volcanology, Rome, Italy, 40 p. Kenyi, L., and Raggam, H., Accuracy assessment of interferometrically derived DTMs. In Proceedings of the Fringe 96 Workshop on ERS SAR Interferometry, Zurich, 30 September 2 October 1996 (ESA SP-406). Lagios E., Intense crustal deformation rates on Nissyros Island (Greece), deduced from GPS studies, may foreshadow a forthcoming volcanic event. Proc. 2 nd Intern. Conf. On Earthquake Hazard and Seismic Risk Reduction, Sept ,1998, Yerevan, Armenia, S. Balassanian et al. (Eds.), Kluwer Academic Publishers, Lagios E., Chailas S., Giannopoulos J. & Sotiropoulos P., Surveillance of Nissyros Volcano: Establishment and remeasurement of Rn & GPS networks. Proc. 8th Intern. Congress Geol. Soc. Greece, Patras, May 27-29, Vol. 32/4, Lanari, R., Lundgren, P., and Sansosti, E., Dynamic deformation of Etna volcano observed by satellite radar interferometry. Geophysical Research Letters, 25 (10),

26 Lu, Z., Mann, D., Freymueller, T., and Meyer, D. J., Synthetic aperture radar interferometry of Okmok volcano, Alaska: Radar observations. Journal of Geophysical Research, Marzocchi, W., Remote reismic influence on large explosive eruptions. Journal of Geophysical Research, 107 (B1), Massonnet, D., Rossi, M., Carmona, C., Adragna, F., Peltzer, G., Feigl, K., and Rabaute, T., The displacement field of the Landers earthquake mapped by radar Interferometry. Nature, 364, Massonnet, D., Briole, P., and Arnaud, A., Deflation of mount Etna monitored by spaceborne radar interferometry. Nature, 375, Nikolova, S., Ilinski, D., Makris, J., Chonia, T., and Stavrakakis, G., D velocity tomography of the Kos - Nissyros volcanic area - east Aegean sea. EGS Abstracts Nice, April 2002 (EGS02-A-06157). Papadopoulos, G. A., Sachpazi, M., Panopoulou, G., and Stavrakakis, G., The volcanoseismic crisis of in Nissyros, SE Aegean Sea, Greece. Terra Nova, 10, Papanikolaou, D., and Nomikou, P., Tectonic structure and volcanic centers at the eastern edge of the Aegean volcanic arc around Nissyros Island. Bulletin of the Geological Society of Greece, 34(1), Scharroo, R. and P. N. A. M. Visser, and G. J. Mets, Precise orbit determination and gravity field improvement for the ERS satellites, Journal Geophysical Research, 103 (C4), Parcharidis, I., and Lagios, E., Deformation in Nissyros volcano (Greece) using Differential radar Interferometry. Bulletin of the Geological Society of Greece, 34(4),

27 Van der Kooij, M.W.A., Armour, B., Ehrismann, J., Farris-Manning, P., and S. Sato, S., The EarthView software for Spaceborne Interferometric SAR processing. Proceedings 3 rd ERS Symposium, (ESA) March 1997 Florence (Italy). Vassilopoulou, S., and Hurni, L The Use of Digital Elevation Models in Emergency and Socio-Economic Planning: A Case Study at Kos - Yali - Nisyros - Tilos Islands, Greece. Proc. 20th International Cartographic Conference, Beijing, China, pp

28 APPENDIX A. 1. Working window of the DESCW software used to locate suitable baselines for differential InSAR in Nissyros (white circle). The list of the available 1997 ERS 2 scenes is shown. Note that frames 711 and 729 indicate ascending (night) and 2871 descending (day) orbits, respectively. Highlighted is orbit of 30 March,

29 2. Working window of the ODISSEO software tool used to locate suitable baselines for differential InSAR in Nissyros. The list for the April 22, 1997 ERS 2 scene is shown. Suitable pairs form when frame Id=2871 and baseline length is less than 1000 metres. 29

30 APPENDIX B. The Nissyros Baselines. Yellow highlights indicate suitable pairs. FROM TRACK 107 FRAME 2871 ERS1/ERS1 ORBIT APRIL 1996 BASELINE PARAL BASELINE PERP JAN SEPT JUL JUN MAY ORBIT JAN 1996 BASELINE PARAL BASELINE PERP SEPT JUL JUN MAY ORBIT SEPT 1995 BASELINE PARAL BASELINE PERP JUL JUN MAY ORBIT JUL 1995 BASELINE PARAL BASELINE PERP JUN MAY ORBIT JUN 1995 BASELINE PARAL BASELINE PERP MAY ERS2/ERS2 ORBIT AUG 1996 BASELINE PARAL BASELINE PERP JUN SEPT AUG JUN ORBIT JUN 1996 BASELINE PARAL BASELINE PERP SEPT AUG JUN ORBIT SEPT 1995 BASELINE PARAL BASELINE PERP AUG

31 JUN ORBIT AUG 1995 BASELINE PARAL BASELINE PERP JUN 1995 ERS1/ERS2 ORBIT APRIL 1996 BASELINE PARAL BASELINE PERP E SEPT E AUG E JUN ORBIT JAN 1996 BASELINE PARAL BASELINE PERP E SEPT E AUG E JUN ORBIT SEPT 1995 BASELINE PARAL BASELINE PERP E SEPT E AUG E JUN ORBIT JUL 1995 BASELINE PARAL BASELINE PERP E AUG E JUN ORBIT JUN 1995 BASELINE PARAL BASELINE PERP E JUN ERS2/ERS1 ORBIT AUG 1996 BASELINE PARAL BASELINE PERP E APRIL E JAN E SEPT E JUL E JUN E MAY ORBIT JUN 1996 BASELINE PARAL BASELINE PERP E APRIL E JAN E SEPT E JUL E JUN E MAY

32 ORBIT SEPT 1995 BASELINE PARAL BASELINE PERP E SEPT E JUL E JUN E MAY ORBIT AUG 1995 BASELINE PARAL BASELINE PERP E JUL E JUN E MAY ORBIT JUN 1995 BASELINE PARAL BASELINE PERP E JUN E MAY

33 APPENDIX C. (Top) Nadir view of Nissyros topography using a 20-m DEM. Also showing is the overlay of 20-m contours that were digitized from the 1:10,000 map provided by IGME (G. Vouyoukalakis, 1999). A linear histogram stretch has been applied using the EASI PACE software package. North is towards the top. (Bottom). Image fit showing ground control points used to georeference the 1:10,000 topographic map. 33

34 APPENDIX D. Earthquakes recorded by the National Observatory of Athens (NOA) near Nissyros between 8/3/95 and 31/8/1996 (32 events). YEAR M D HR MIN SEC LAT LON DEP MAG 1995 MAR AUG SEP NOV NOV NOV NOV NOV DEC DEC JAN JAN JAN FEB FEB APR APR APR APR APR APR APR APR MAY JUN JUN AUG AUG AUG AUG AUG AUG