Towards a Near-real Time Global Flood Detection System (gfds)

The methodology for flood detection developed at Dartmouth Flood Observatory (DFO) (Brakenridge and others, 2007) on a global scale, was implemented at the Joint Research Centre (JRC) of the European Commission on an automatic operational basis. The technique is using AMSR-E passive microwave remote sensing data of the descending orbit, H polarization, 36 GHz band to detect flood events from space around the globe with a daily temporal resolution. DFO was formerly using the daily global composite images having 3 days of lag in time whereas JRC implemented the model on the near-real time swath data available about 24 hours after acquisition. A ratioing of the water/dry signals described in the article are considered as a tool to observe surface water area changes. Thresholding the calculated observation ratio allows the detection of riverine inundation events. Validation of the model results is ongoing. Nevertheless, preliminary results show a promising correlation of the increase in river discharge on-site and changes at the observed signal of the sensor. Thus following the technique the detection of flood events in ungauged and inaccessible remote river channels is feasible from space. Software was developed at the JRC to automatically acquire and process the remotely sensed data in real time on an operational basis. After the validation and calibration of the satellite based near-real time Global Flood Detection System (GFDS) the remotely observed flood events are planned to be integrated into the Global Disaster Alert and Coordination System (www.gdacs.org/floods). GDACS is running at the JRC providing near real-time alerts about natural disasters around the world and tools to facilitate response coordination, including news and maps. 1. METHODOLOGY OF THE GLOBAL FLOOD DETECTION SYSTEM (GFDS)