Probalilistic analysis of inundations and dike failures in Flanders
Geukens, B. (2005). Probalilistic analysis of inundations and dike failures in Flanders. MSc Thesis. Katholieke Universiteit Leuven (KUL)/Vrije Universiteit Brussel: Leuven. 102 pp. |
| Available in | Author | | Document type: Dissertation
|
| Abstract | In order to investigate the importance of dike failures in the total flood risk, an empirical analysis is made of the relative frequency of dike failures and inundations by comparing information of historical floods with load factors derived from discharge data.
Both information on dike failures and inundations from the last hundred years containing information of time, location, river, cause and number of disjoint regions was gathered from newspapers. The historical floods were geographically located within the river (sub)basins of Flanders using a GIS to correlate later on the inundation and dike failures with discharge data from the associated hydro-geographical entity.
The completeness of the historical dataset was calculated by comparing it with independent information on inundations and dike-failures from other investigations (Probabilitas, ROG-zones).
A probabilistic model was constructed for the estimation of the probability of inundation and dike failure depending on an explanatory variable or "load factor". "Peak-Over-Threshold" (POT) values selected from hourly river discharge series were considered as load factor. POT values enabled us to use the flood threshold in quanti le plots as a reference for the occurrence of inundations or overtopping. The Dender basin was used as an example to construct and test the model. From this example, it was concluded that dike failures do occur before the flood threshold was attained. For discharge values higher than the flood threshold, dike failures will have a high probability of occurrence and influence the area that is inundated and therefore the damage caused.
Also information on the occurrence of floods by external causes (not hydro- climatologically induced) was gathered and the regional amount of inundations and or dike-failures for a given load factor was calculated. |
|
