Precise InSAR analysis for detection of volcanic deformation - PowerPoint PPT Presentation

Precise InSAR analysis for detection of volcanic deformation 火山性地殻変動の検出に向けた 高精度SAR干渉解析 Taku OZAWA (NIED) 小澤拓（防災科学技術研究所）

NIED Volcano Observation Network The 6th observation station of Mt. Fuji Nasu Mt.Fuji Izu-Oshima Miyake-jima Seismometer, Magnetometer, GPS, Tiltmeter, Strainmeter, Gravitmeter Iwo-tou

Deformation of 2000 Miyakejima eruption (Ueda et al., 2005)

New observation network (Plan) Mt. Usu Hokkaido Mt. Tokachi -Komagatake Mt. Kusatsu Mt. Tarumae -Shirane Mt. Unzen Mt. Iwate Kuchinoerabu Isl. Suwanose Isl. Mt. Asama Mt. Aso Expectation for utilization Mt. Kirishima of remote sensing in volcano Sakura-jima monitoring is high.

Problems of monitoring by present InSAR • Detection accuracy Noise often exceeds 5cm Uncertainty of accuracy • Temporal resolution Repeat cycle of ALOS is 46 days. This talk As a step of it, we want to make it possible to detect time-series of deformation precisely.

PALSAR data of Miyakejima 054 057 058 407 410 2006 2007 Miyakejima 2008 2009 054 057 058 407 410 2006 8/21 8/26 9/12 9/11 9/16

Interferometric pair (70 pairs) Baseline [m]

Atmospheric noise Path: 057(D,34.3), 2008/8/31 – 2009/1/16 Atmosphere induces propagation delay of radar. (Atmospheric noise) Atmospheric noise reduction: ・ Linear approximation with elevation (e.g., Fujiwara et al., 1999) ・ Simulation from numerical weather model (e.g., Shimada, 1999, Otsuka et al., 2002)

Atm.-delay simulation from weather model JMA Meso-Scale Model (MSM) Reflectivity Temperature Humidity Isobaric pressure height − × 6 ( n 1 ) 10 P P P = + + d v v K K K 1 2 3 2 T T T P d : partial pressure of dry air P v ： partial pressure of water vapor T ： temperature Convert to temperature, pressure, humidity at every layer. every 1000m Estimation of radar （ 50km ～） propagation path by ray-tracing method. every 100m （～ 10km ） Estimation of delay along propagation path

Application in Mt. Fuji No-correction Linear of elevation MSM 2006/9 – 2006/11 No-correction Linear of elevation MSM 2006/11 – 2008/8

Standard deviation No- correction Linear of elevation MSM

Interferograms subtracted sim. delay

Adjusted to GPS deformation 2006/5/26 – 2009/8/5 Assume remaining orbital fringe to be uniformly Miyake1 inclined plane. Izu Kamitsuki Estimate its plane, Miyake3 adjusting to GPS result. Ako Tsubota Miyake4 Fixed site of GPS result is Mikurajima. Miyake2 (20km south-southeast) 5km

Interferograms (adujusted to GPS)

Estimation of 2-D temporal change Quasi-UD and quasi-EW components of displace- ments are estimated from interferograms by least square analysis. g n i d n e c s e Smoothness constraint is Ascending D used for noise reduction and for interpolation. DEM error is estimated simultaneously (large error was not estimated). Horizontal direction of co-plane is almost east-west (quasi-EW), vertical direction inclines 10 degree from vertical to south (quasi-UD).

Temporal change of 2-D deformation Quasi-UD component − 0.4 − 0.2 0.0 0.2 0.4 [m] Quasi-EW component − 0.4 − 0.2 0.0 0.2 0.4 [m]

Comparison between SAR and GPS Quasi-UD [m] Kamitsuki Izu Ako Tsubota Quasi-EW [m] Quasi-UD [m] Miyake3 Miyake4 Miyake1 Miyake2 Quasi-EW [m] SAR GPS

Deformation in crater bottom Quasi-UD Change in meter Quasi-EW Furuya (2004)

Deformation around crater rim Quasi-UD Change in meter Quasi-EW

Deformation in mountainside Quasi-UD Change in meter Quasi-EW

Summary • We attempted to detect precise time-series of deformation by least-square estimation using multi-pass interferograms with smooth- ness constraint and by atmospheric delay simulation from numerical weather model. • Noise must be reduced based on the theory of least-square estimation, but ... • There is much room for improvement. • Efficient utilization of ALOS and ALOS-2 interferograms.

Acknowledgements PALSAR level 1.0 data are shared among PIXEL (PALSAR Interferometry Consor- tium to Study our Evolving Land surface), and provided from JAXA under a cooperative research contract with ERI, Univ, Tokyo The ownership of PALSAR data belongs to METI (Ministry of Economy, Trade and Industry) and JAXA. GEONET GPS data were used in this research.