PRECURSORY SLOPE DEFORMATION AROUND LANDSLIDE AREA DETECTED BY INSAR THROUGHOUT JAPAN

Interferometric Synthetic Aperture Radar (InSAR) technique is able to detect a slope deformation around landslide (e.g., Singhroy et al., 2004; Une et al., 2008; Riedel and Walther, 2008; Sato et al., 2014). Geospatial Information Authority (GSI) of Japan has been performing the InSAR analysis regularly by using ALOS/PALSAR data and ALOS-2/PALSAR-2 data throughout Japan. There are a lot of small phase change sites except for crustal deformation with earthquake or volcano activity in the InSAR imagery. Most of the phase change sites are located in landslide area. We conducted field survey at the 10 sites of those phase change sites. As a result, we identified deformation of artificial structures or linear depressions caused by mass movement at the 9 sites. This result indicates that InSAR technique can detect on the continual deformation of landslide block for several years. GSI of Japan will continue to perform the InSAR analysis throughout Japan. Therefore, we will be able to observe and monitor precursory slope deformation around landslide areas throughout Japan. * Corresponding author. + Present affiliation: Ministry of Education, Culture, Sports, Science and Technology, Japan ⵐ Present affiliation: Japan Digital Road Map Association


INTRODUCTION
Synthetic Aperture Radar (SAR) data is able to provide detailed and spatially comprehensive ground information. Interferometric SAR (InSAR) enables us to detect ground deformation with high precision by InSAR imagery (e.g., Massonnet and Feigl, 1998;Bürgmann et al., 2000). The InSAR imageries show the range change between the ground and the satellite, that is, the line-of-sight (LOS) displacement from phase difference. Satellite InSAR is able to detect ground deformation of several millimetres to several centimetres with several meters to dozens of meters resolution typically. Geospatial Information Authority (GSI) of Japan has been performing the InSAR analysis regularly by using ALOS/PALSAR data and ALOS-2/PALSAR-2 data to detect crustal deformation induced by large earthquake, volcanic activity and land subsidence, etc. (e.g. Amagai et al., 2007;Suzuki et al., 2008;Tobita et al., 2011;Kobayashi et al., 2011;Kobayashi et al., 2015). ALOS was operated from January 2006 to May 2011, and ALOS-2 has been operated since May 2014. Both have L-band radar. In Addition, InSAR technique is able to detect a slope deformation around landslide (e.g. Singhroy et al., 2004;Une et al., 2008;Riedel and Walther, 2008;Sato et al., 2012). InSAR analysis of GSI of Japan by using ALOS/PALSAR data and ALOS-2/PALSAR-2 data cover whole of Japan. There are a lot of small phase change sites except for crustal deformation with earthquake or volcano activity in the InSAR imagery. Most of the phase change sites are located in landslide area. Therefore, it is supposed that these phase changes are generated by slight slope deformation like precursory landslide. So, we conducted field survey at the 10 sites of those phase change sites and investigated deformation of artificial structures or topography with gravity sliding.

Data and Analysis
We used ALOS/PALSAR data from January 2006 to May 2011 and ALOS-2/PALSAR-2 data from May 2014 to October 2015. The ALOS/PALSAR data was observed from descending (D) and ascending (A) orbits. The ALOS-2/PALSAR-2 data was observed from right side of descending orbit (DR), left side of descending orbit (DL), right side of ascending orbit (AR) and left side of ascending orbit (AL). We used InSAR imageries throughout Japan which analysed regularly. The InSAR analysis was conducted by using unique "GSISAR" software Tobita et Tobita, 2003). The specification of InSAR imageries referred for field survey on 10 sites in this paper are shown in Table 1.

Interpretation of signal of slope deformation
We interpreted and extracted the signal (phase change) of slope deformation on the InSAR imageries. The main rule of interpretation is as follows. 1) Line-of-sight (LOS) displacement is approximately 3 cm over. 2) Size of phase change is approximately 150 m by 150 m over.
3) Coherence of InSAR imagery is good. 4) Phase change zone has continuity and high density. 5) Not water and marsh area. 6) Not quarry area. 7) Similar phase change is not appeared on the same direction slope generally. 8) Phase change is no correlation with elevation and topography. 9) Boundary of phase change zone is inconsistent with boundary of land-use or vegetation. 10) Phase change is able to identify around same zone on the plural InSAR imageries of difference pair. (in case of ALOS-2/PALSAR-2) 11) Phase change is located on mountain or hill slope. 12) The direction of phase change is consistent with direction of slope. 13) The location of phase change is consistent with area of landslide topography and vulnerable geology.

FINDINGS
As a result of interpretation of the signal of slope deformation, we extracted 117 sites and 79 sites from ALOS/PALSAR interferograms and ALOS-2/PALSAR-2 interferograms respectively. We conducted field survey at the 10 sites of those from 2013 to 2016 (Figure 1). The 7 sites were extracted from ALOS/PALSAR interferograms, and the 3 sites were extracted from ALOS-2/PALSAR-2 interferograms (Table 1).
In this paper, we introduce the 3 cases (2 cases of ALOS/PALSAR and 1 case of ALOS-2/PALSAR-2) of those.  Figure 2). There is landslide block around the phase change area. We surveyed on site in June 25, 2013, and we identified collapse of the landslide (Figure 3). Because the landslide didn't collapse in aerial photo on November 2012, it is estimated that the landslide collapsed between November 2012 and June 2013. Therefore, it can be said that InSAR technique was able to detect the precursory slope deformation of landslide block before the final landslide collapse of 2013.

CONCLUSIONS
We conducted field survey at the 10 sites of phase changes detected by ALOS/PALSAR and ALOS-2/PALSAR-2 interferograms. As a result, we identified deformation of artificial structures or linear depressions caused by mass movement at 9 sites. These results indicate that InSAR technique is able to detect on the continual slope deformation of landslide block for several years. There are exceptions to it, however. It is necessary that our rule of phase change interpretation of slope deformation is improved. In parallel with it, GSI of Japan will continue to perform the InSAR analysis throughout Japan. Therefore, we will be able to observe and monitor precursory slope deformation around landslide areas throughout Japan.

ACKNOWLEDGEMENTS
PALSAR data are provided from JAXA (Japan Aerospace Exploration Agency) through "Joint Cooperative Agreement between GSI and JAXA for observation of geographic information using Advanced Land Observing Satellite (ALOS) data." The ownership of PALSAR data belongs to JAXA and METI (Ministry of Economy, Trade and Industry). PALSAR-2 data are provided from JAXA through joint cooperative agreement between GSI and JAXA. The ownership of PALSAR-2 data belongs to JAXA. The products of the numerical weather model were provided by JMA (Japan Meteorological Agency) under the agreement between GSI and JMA.