High-Resolution Radar Damage Assessment After the Earthquake in Haiti on 12 January 2010

After a disaster such as an earthquake or tsunami it is important to know quickly the dimensions of the damage to human relief work and crisis management. To study the utility of high-resolution radar damage-assessment for this purpose, pre- and post-seismic TerraSAR-X and Cosmo-SkyMed data of the earthquake in Haiti on 12 January 2010 were compared with the results of visual damage assessment, carried out in the framework of the Post Disaster Needs Assessment and Recovery Framework (PDNA). Studied were grid-based methods estimating the average damage, and detailed methods up to building level. On a grid-scale of 200 m, regression analysis of the average normalized difference between pre- and post-seismic data, and the PDNA, results in an average correlation coefficient of up to 0.49 for medium density built-up in TerraSAR-X. Correlation of the shanty areas is lower than for high density built-up. Industrial areas are excluded because of the continuous activity not related to the earthquake. Detailed damage assessment at building level is less hopeful because the correlation per building is low, though improvement is expected from adding information on the boundaries of buildings, if available. To summarize, grid-based average radar damage assessment in the first days after an earthquake event looks promising, but the type of built-up and the period of time between the acquisitions have to be taken into account. For a more detailed assessment on the long term, in working on reconstruction, it is better to rely on other sources.