Mapping Tsunami Disaster Impact of by Satellite Remote Sensing

Shunichi Koshimura, Erick Mas, Luis Moya, (International Research Institute of Disaster Science (IRIDeS), Tohoku Univ., ) Naoto Yokoya, Bruno Adriano (RIKEN AIP Center, Japan) Masashi Matsuoka (Tokyo Tech., Japan)

Rokhis Khomarudin (Lembaga Penerbangan dan Antariksa Nasional – LAPAN, Indonesia) Abdul Muhari (Ministry of Marine Affairs and Fisheries, Indonesia) The 28 September Tsunami of , Indonesia M7.5, 10:02:45 (UTC), 18:04:44 (Local time)

BMKG Digital Globe 2 Objectives

• Developing satellite remote sensing methods and mapping technology based on emergency earth observation for identifying impacts and vulnerabilities as emergency response efforts in future major disasters. • Understanding building vulnerability of Palu, as a form of “Tsunami Fragility Curve”.

3 Remote Sensing Approach for Mapping Damage

(a) (b) (d) (e) 0 400 800 ¹ m

(d)

(a) 0 400 800 0 490 980 0 400 800 (e) m m m

Legend 0 600 1,200 m (b) Destroyed

Damaged

Possibly Damaged

Note: (f) - The background image was provided by Planet Lab Inc.

(c)

Data Sources - Pre-event image: Sentinel-2 (c) ESA (2018) (acquired on 17/09/2018), Sentinel-1 (c) ESA (2018) (acquired on 07/06/2018) 0 500 1,000 - Post-event: Sentinel-2 (c) ESA (2018) (acquired on 02/10/2018), PlanetScope (c) Planet Labs Inc. (2018) (acquired on 01/10/2018), ALOS-2 (c) JAXA (2018) (acquired on 01/10/2018) m - Building vector layer: OpenStreetMap (c) OpenStreetMap contributors - Digital Elevation Model: SRTM (30 m) (NASA/USGS) - Building damage annotatio: Copernicus Emergency Management Service - Mapping (EMSR317: Earthquake 0 2 4 in Indonesia) (c) European Union

km Disclaimer The information contained in this map is for general information purposes only. The information is provided by Geoinformatics Unit and while we endeavor to keep the information up to date and correct, we make no (c) (f) representations or warranties of any kind, express or implied, about the completeness, accuracy, reliability, © Tohoku Univ., RIKEN AIP, Planet Lab Inc. suitability or availability with respect to the map or the information, products, services, or related graphics contained on it for any purpose. Any reliance you place on such information is therefore strictly at your own risk. 4 Disaster Impact Consequences of the interaction among hazards and exposure

VULNERA IMPACT HAZARD EXPOSURE BILITY

Remote sensing Observation Building data Numerical model Population data Fragility curves

5 6 Dataset

7 Dataset

8 Work Flow

9 Results

10 Measurement of Inundation Extent

500m 11 Measurement of Flow Depth

500m 12 Structural Damage

Copernicus Damage Data (8,236)

923

3268

4044

Destroyed Damaged Possibly damaged

https://emergency.copernicus.eu/mapping/list-of-components/EMSR317

13 Tsunami Fragility Curve (Preliminary)

Banda Aceh, Palu, Sulawesi (Koshimura et al., 2009)

14 Summary • Ensemble learning classifiers using multi-temporal and multi-sensor data would work for building damage recognition. • 1-3m flow depths were measured in central Palu coast. The maximum flow depth was 4.6 m at the west. However, the spatial distribution of tsunami flow depths are scattered. • Major impact was concentrated within about 200 m from the shoreline. • Relationships between tsunami hydrodynamics and structural damage is determined as a form of tsunami fragility curve. 15