< Back to previous page

Publication

Scientific assistance towards a probabilistic formulation of hydraulic boundary conditions: test case for the Yser basin

Book - Report

A methodology for the probabilistic formulation of hydraulic boundary conditions is tested within a test case for the Yser basin. This navigable waterway in the Flanders Region of Belgium has its origin in Northern France and mouths in the North Sea in Nieuwpoort. A hydrodynamic model in the Mike11 software is made available by Flanders Hydraulics Research. The model has a downstream boundary in Nieuwpoort and seven upstream boundaries, i.e. the Yser in Roesbrugge and six tributaries of the Yser. The boundaries coincide with gauging stations. The time series available from the stations range from 20 to 40 years. On the basis of the time series, synthetic boundaries are generated for each of the boundaries. The time series have been validated. To reduce the uncertainty of an extreme value analysis, the length of the long-term time series should be maximized while taking care of the homogeneity of the time series in order to reduce bias. Correlation analysis indicates that the occurrence of extreme discharge events at the upstream boundary of the Yser (Roesbrugge) and high as well as low water levels at Nieuwpoort are independent. On the other hand it indicates that the occurrence of extreme discharge events at the upstream boundary of the Yser (Roesbrugge) and at a representative tributary of the Yser, i.e. the Poperingevaart, is dependent. Hence two dependency tests are considered: partial and full dependency of discharge at the Yser and the tributaries. In the first test a bivariate extreme value Copula analysis is applied to determine joint probabilities. In the second test a univariate extreme value analysis is applied to determine probabilities. On the basis of the probabilities, sets of synthetic events with known frequencies have been dressed up of for both tests. This yields in a considerably larger number of events in case of the Copula analysis. The two sets of synthetic events have been simulated with the hydrodynamic model. On the other hand a long-term simulation of 18 years of gauged data is executed as part of a validation. The simulation results are analysed in return level-return period plots for 12 check points along the Yser. The water levels and discharges based on the synthetic boundary conditions accord well with the levels and discharges on the basis of historic events in the long-term simulation. The synthetic events from the two dependency tests produce corresponding water levels and discharges in the case of the Yser basin.

Series: FHR reports
Number of pages: 145
Publication year:2019
Keywords:Extreme values, Flooding, Flood risks