Hybrid modeling framework for flood prediction

A hybrid computational framework for short-term flood prediction in urban watersheds (characterized by overland runoff) will be developed to improve prediction accuracy. The framework aims to accurately predict an event, e.g. flood or no-flood, as opposed to traditional methods which estimate water flow characteristics, e.g. 6 feet above flood stage. Successful early prediction of these events can help authorities to take appropriate mitigation measures and to minimize losses from the flooding. The framework requires only current and historical data on water levels and precipitation in the area of interest (such as those collected by TRCA’s flood monitoring gauge network). The development of the new framework will complement existing hydrologic simulation systems to improve and enhance services provided by TRCA to local agencies and public.

Nathan Rose
Faculty Supervisor: 
Drs. Marina G. Erechtchoukova, Stephen Chen & Peter Khai