Inundation risk for embanked rivers
- 1Institute of Geophysics, Polish Academy of Sciences, Ksiecia Janusza 64, 01-452 Warsaw, Poland
- 2Institute of Meteorology and Water Management, Podlesna 61, 01-673 Warsaw, Poland
Abstract. The Flood Frequency Analysis (FFA) concentrates on probability distribution of peak flows of flood hydrographs. However, examination of floods that haunted and devastated the large parts of Poland lead us to revision of the views on the assessment of flood risk of Polish rivers. It turned out that flooding is caused not only by the overflow of the levee crest but also due to the prolonged exposure to high water on levees structure causing dangerous leaks and breaches that threaten their total destruction. This is because the levees are weakened by long-lasting water pressure and as a matter of fact their damage usually occurs after the culmination has passed the affected location. The probability of inundation is the total of probabilities of exceeding embankment crest by flood peak and the probability of washout of levees. Therefore, in addition to the maximum flow one should also consider the duration of high waters in a river channel.
In the paper the new two-component model of flood dynamics: "Duration of high waters–Discharge Threshold–Probability of non-exceedance" (DqF), with the methodology of its parameter estimation was proposed as a completion to the classical FFA methods. Such a model can estimate the duration of stages (flows) of an assumed magnitude with a given probability of exceedance. The model combined with the technical evaluation of the probability of levee breaches due to the duration (d) of flow above alarm stage gives the annual probability of inundation caused by the embankment breaking.
The results of theoretical investigation were illustrated by a practical example of the model implementation to the series of daily flow of the Vistula River at Szczucin. Regardless of promising results, the method of risk assessment due to prolonged exposure of levees to high water is still in its infancy despite its great cognitive potential and practical importance. Therefore, we would like to point out the need for and usefulness of the DqF model as complementary to the analysis of the flood peak flows, as in classical FFA. The presented two-component model combined with the routine flood frequency model constitutes a new direction in FFA for embanked rivers