[S-TT39] Synthetic Aperture Radar and its application

[STT39-06] Development and implementation of flood mapping algorithm using ALOS-2

★Invited Papers

*Masato Ohki¹, Kosuke Yamamoto¹, Takeo Tadono¹ (1.Japan Aerospace Exploration Agency)

Keywords:Synthetic Aperture Radar, PALSAR-2, flooding

Increasing human and economic losses from urban floods demand rapid flood monitoring using synthetic aperture radar (SAR), which has all-weather, day-and-night observation capability. Although many studies on flood monitoring have proposed SAR-based flood detection algorithms, only a limited number of methods mainly using C- and X-band SARs are actually used in operational disaster response systems. We developed the first L-band SAR-based algorithm for operational flood monitoring using the Advanced Land Observing Satellite-2 (ALOS-2). ALOS-2 offers the advantages of (1) SAR observation at midnight, which is particularly important for responding to flooding at night, and (2) highly frequent observation (more than once a day) thanks to the satellite’s wide-ranging incidence angles. Our algorithm combines SAR data by ALOS-2, near-real-time flood simulation data from the Today’s Earth system (TE), and other geographical information (hazard maps and terrain models). The algorithm estimates the flood extent from the flood probability calculated by Bayesian inference using ALOS-2-derived input data (multitemporal amplitude and coherence images) and prior probability based on the flood fraction data from TE-Japan. The validation results showed that the accuracies of the estimated flood extent were about 60 to 90%, depending on the flood events. The incidence angles used in the evaluation range from 15° to 59° and proved the robustness of the algorithm. The algorithm was implemented in an operational disaster response system and can provide flood extent less than an hour after the input data are entered.