AGENT BASED MODELLING FOR FLOOD INCIDENT MANAGEMENT IN CANVEY ISLAND, UNITED KINGDOM
Canvey Island is an island on the Thames Estuary, East of England. The 1953 North Sea storm surge, which is regarded as the worst natural disaster in Northern Europe over the past two centuries, is the most recent tidal flood disaster that hit the island. This led to the construction of the Thames Barrier which became operational in 1982. Canvey Island is now ringed by 24km of 4.66m high concrete walls and an extensive system of internal drainage. However, increasing population (over 200% increase from 1953), land-use changes, rising sea-levels, and continued subsidence of southeast corner of England act to increase the risk of storm surge flooding. The development of a sustainable flood incident management plan with emphasis on the people directly exposed to floods calls for a paradigm shift from the traditional classical engineering approach to problem solving (e.g. increasing the height of the flood defences).
In this study, an agent based model (ABM) to simulate a flood incident was developed which takes into consideration: demographics, road networks, building types and locations, travel patterns and flood water depth at various points on the island. The key research interest is to identify from the model how timing of flood incidents, location of evacuation centres and timing of flood warning will affect the overall flood risk.
The results show that the evacuation time for a given number of agents varies for different time of the day. It is minimum during the early hours of the day and maximum late at night. This can be attributed to the large population located in the residential areas in the night which is far away from the exit routes. It was also observed that the combined effect of using two new evacuation centres and a new road suggested in this study to reduce traffic congestion, reduced the evacuation time by 44% compared to the “do nothing” scenario. Flood warning was found to be
a major factor that affects how people are exposed to flood water. When only 50% of the population received flood warning and acted on it with the other half only acting 30 minutes later, it was found out that over a 1000 people were exposed to flood water depth above 25cm and this was concentrated in a given area of the island.