Day 4

Detailed paper information

Back to list

Paper title Harnessing Sentinel-1 co-seismic DInSAR products and Geospatial Intelligence as an operational tool on Earthquake Impact Assessment: The case of the Arkalochori (Crete, Greece) on September 27, 2021
Authors
  1. Triantafyllos Falaras Department of Geography. Harokopio University of Athens Speaker
  2. Andreas Karavias Department of Geography. Harokopio University of Athens
  3. Pavlos Krassakis Department of Geography. Harokopio University of Athens
  4. Olga Markogiannaki University of Western macedonia
  5. Issaak Parcharidis Department of Geography. Harokopio University of Athens
Form of presentation Poster
Topics
  • D1. Managing Risks
    • D1.01 Satellite EO for Geohazard Risks
Abstract text Earthquake risk is a global-scale phenomenon that exposes human life to danger and can cause significant damages to the urban environment increasing globally more or less in direct proportion to exposure in terms of the population and the human-built environment. Earth observation (EO) science, plays an important role in operational damage management, showing the most affected areas on a large scale and in a very short time helping the decision making processes more effectively. This can be done by combining the geodetic information with geospatial data to generate a Geospatial Intelligence (GEOINT) product which is the organization of all the available geographical information on the area of interest.

In the morning (09:17 EEST) of September 27, 2021, a strong M=5.8 with a 10 Km focal depth (35.1430 N, 25.2690 E) earthquake struck in the area of Arkalochori town, Crete ~22 Km at the Southeast of the city of Heraklion. Several aftershocks followed for the next few days with the strongest to be that of September 28, 2021 (07:48 EEST) M=5.3 with an 11 Km focal depth (35.1457 N, 25.2232 E) according to the Institute of Geodynamics of the National Observatory of Athens (http://www.gein.noa.gr/en/). The main earthquake caused extensive damages to numerous buildings including homes and schools constituting many of them unsafe to use in the impacted region. Some people were injured, one lost his life, and others became homeless.

This study that was performed operationally at the time of the earthquake in Arkalochori, Crete, aims at developing a useful Geospatial Intelligence operational tool for the impact assessment of that earthquake. This was carried out by retrieving the ground deformation information from co-seismic Differential SAR Interferometry (DInSAR) products and then combining it with infrastructure-related geospatial data. The developed tool was available the following days after the event to be used by the stakeholders (e.g. emergency responders, scientists, civil protection, etc)

For the geodetic purposes of this study were used, (i) two Sentinel-1 SAR SLC (Single-Look-Complex) IW (Interferometric Wide Swath) images in ascending (master 24/09-slave 29/09) and descending (master 18/09-slave 30/09) geometry before and after the earthquake event in order to generate co-seismic interferometric pairs, and (ii) a Digital elevation model (DEM) SRTM-3 sec (90m) of the study area. ESA Copernicus Sentinel-1 SLC satellite images are openly available within a few hours since their acquisition from the Copernicus Open Access Hub platform (URL: https://scihub.copernicus.eu/). The processing of Sentinel-1 SLC images was performed on ENVI SARscape software.

Regarding the generation of geodetic products is separated into three main steps: The first step is the pre-processing of Sentinel-1 SAR SLC images that include the orbit correction, the burst selection, and the co-registration of master and slave image in ascending and descending geometry, respectively. The second step in the main processing of the interferometric pairs in every geometry is performed by the coherence and wrapped interferogram generation, interferogram flattening using a DEM SRTM-3 sec, adapted filtering, phase unwrapping using MCF (Minimum Cost Flow), and finally the phase to displacement in Line-Of-Sight (LOS) and geocoding. DInSAR displacement map in LOS, only measures the path length difference between the earth surface and the satellite. In order to estimate the vertical (up-down) and horizontal (east-west) deformation, the third step of displacement decomposition was carried out. In this step, the ascending and descending LOS displacement products were used to recover the true movements in vertical and horizontal axes. The final products are exported in GeoTiff format for further analysis of ground deformation and damage estimation in correlation with urban fabric and infrastructures on the GIS environment.

Other datasets regarding infrastructure are vector points, polylines, and polygons derived from ready-to-use from various open sources or information for digitization. These include Airports, Hospitals – Health Centers, Schools, Cultural-Archaeological sites, Urban Fabric, Roads, Bridges, and Dams. The utilized software for the GIS processing is the commercial ESRI ArcGIS Pro 2.8 and for the development of the Geospatial Intelligence application via a Web App the ESRI ArcGIS Online and its WebApp Builder. After the import of the ground deformation products to the GIS software then the mining of that information to the already prepared vector datasets was performed with the proper tools leading to the creation of the new vector Geospatial Intelligence products. These products were then uploaded to the cloud-based ESRI ArcGIS online to create a web map needed for the operational tool. With the WebApp Builder, the app was developed and then combined with the web map.

Finally, the results of this study have shown that there was a subsidence up to -20 cm regarding the vertical (Up-Down) displacement while eastward movements up to 13 cm and westward movements up to 6 cm exist according to the horizontal (East-West) displacement. Generally, in the area around the town of Arkalochori, there is subsidence which reaches the maximum negative values. Regarding the web app tool, the integration of co-seismic deformation maps and geospatial data including the exposure datasets into a tool for post-disaster infrastructure assessment can be very useful. It contributes to the identification of the most severely impacted areas and the prioritization of in-situ inspections. The use of the proposed tool for on-site inspections in the affected area around Arkalochori showed a good match of the “red” area of the co-seismic deformation map with the location of the identified large number of extensive damages on structures. It also contributed to the effective and quick inspection of roadway networks focusing on the identified bridges in the geospatial intelligence tool. However, the inspected bridges were found in good condition and seismic damages were not developed. In conclusion, this Geospatial Intelligence web app can be used further for more analytic research, decision making, and other uses while it can also be enhanced with more datasets and special information integration.

The ESRI ArcGIS Online Web App that was developed is open and accessible from every portable device or pc in the following link via any web browser: https://learn-students.maps.arcgis.com/apps/webappviewer/index.html?id=339cd0b5020f40cb93607d4c4d519cea

Acknowledgments
We would like to thank Harris Geospatial local dealer Inforest Research o.c. for the access to ENVI SARscape as well as ESRI for the Learn ArcGIS Student Program license.