384883 - Defining real-time flood alerts with multicriteria analysis
Monday, June 4
2:00 PM - 3:30 PM
Location: Greenway AB
Flavio Conde, Sao Paulo – FCTH; Sara Pion, São Paulo – FCTH; Sandra Uemura, São Paulo – FCTH
User reliance in flood forecasting is based on the accuracy of model outputs that can generate correct alerts. There are problems related to model imprecisions on forecasted rainfall and mathematical models. If events are observed but not forecasted or forecasted but not observed, the alert is considered a miss. In many cases citizens are not well informed about how flood forecasting works and the imprecisions that may occur. So once an error occurs, citizen reliance may be reduced for subsequent events. To avoid this reliance problem, this study objective is to develop a methodology to create specific flood maps that unite flood risk and intensity, with location site characteristics. The main objective of this article is to explain how a multi criteria analysis can help reduce misses, and provide specific real time flood alerts that prevent dispersing wrong information to a large number of citizens.
The case study is held on São Paulo’s highly urbanized basin named Jaguaré, chosen because of its diverse floodplain occupation.
Real time flood forecast alerts relies on creating thresholds that represent the situation about to happen. They will be described thru flood maps that unite the main parameters on urban events. The study methodology is based on using depth, velocity, flood probability and edification type/vulnerability as decision parameters to draw precise alert boundaries that are deployed only when necessary.
Deploying flood alerts is a very responsible task and by doing this judicious analysis, it’s possible to clarify and reduce misses, building up citizens reliance.