Mapping severe damage to land cover following the 2004 Indian Ocean tsunami using moderate spatial resolution satellite imagery

  • Authors:
  • A. S. Belward;H. -J. Stibig;H. Eva;F. Rembold;T. Bucha;A. Hartley;R. Beuchle;D. Khudhairy;M. Michielon;D. Mollicone

  • Affiliations:
  • European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy;European Commission, Joint Research Centre, 21020 Ispra (VA), Italy

  • Venue:
  • International Journal of Remote Sensing - Satellite Observations Related to Sumatra Tsunami and Earthquake of 26 December 2004
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Visual analysis of Moderate Resolution Imaging Spectroradiometer (MODIS) 250m before-and-after event imagery along 11 335 km of South East Asia and East Africa's coastline identified major changes to land cover along 1 220 km of coast caused by the tsunami of 26 December 2004. Over 81 000 ha of land were mapped as lost or severely damaged. High spatial resolution images (10-32 m) from the Terra satellite's Advanced Spaceborne Thermal Emission and Reflection Radiometer, Disaster Monitoring Constellation's Surrey Linear Imager and the Landsat Enhanced Thematic Mapper (ETM) provided verification. Area estimates of damaged land from MODIS were within 17% of the estimates made using high spatial resolution systems. Errors of commission and omission were estimated at 14% and 31%, respectively. The correlation between damage estimates from MODIS and high spatial resolution systems resulted in an R 2 of 0.73 at the 99.5% confidence level, with MODIS proving especially effective at mapping tsunami-damaged areas greater in size than 150 ha. Correcting the original area estimate to account for errors of omission and commission using the regression model gave a revised figure of 103 854 ha of land damaged. Land cover types could not be assigned from the MODIS pre-and post-disaster image pairs, but interpretation of pre-disaster ETM imagery for the damaged areas indicated that around 5% was previously classed as barren, 11% as urban, 14% forest and 70% rural. The study provides confirmation that moderate resolution satellite imagery can provide rapid assessments of severe damage to land resources (though not cover type), as well as confirmation of non-affected areas, over very large geographical regions in the aftermath of natural disasters such as the tsunami.