Partenaires et organisations internationales
(Anglais)
|
Politecnico di Milano (I), TU Wien (A), Univ. of Newcastle u.T. (UK), Universidad Polit. de Valencia (E), TH Darmstadt (D), Univ. of Bologna (I)
|
Résumé des résultats (Abstract)
(Anglais)
|
The objective of the project was to evaluate the sensitivity of flash-flood risk assessments to anthropogenic influence, there including man-induced changes of the runoff generating and propagating mechanisms and climate fluctuations. The achievement of the goal required the project to be organised in a sequential structure. First, it was necessary to develop methods for estimating flood risk that can explicitely account for anthropogenic changes, so contributing also to improve the present level of flood estimation techniques. Subsequently, the methods were applied to investigate the impact of the above mentioned changes, mainly concentrating on the effects of landuse changes on flood runoff generation.
Specifically, the project investigated flood frequency regimes by (i) building a theoretical framework for regional flood frequency estimation, that uses physically oriented methods accounting for catchment and climate characteristics to define homogeneous regions, which form the basis of regionalised flood estimates (Workpackage 1, WP1); (ii) building a methodological framework for distributed flood frequency estimation accounting for climatic fluctuations and basin modifications, based on the derivation of flood frequency distribution from storm rainfall distribution and basin characteristics (Workpackage 2, WP2); (iii) building a methodological framework to approach flood risk analysis by simulation of the Rainfall-Runoff (R-R) process based on distributed models, that can account for an explicit parameterisation of land use, watershed drainage and river engineering works (Workpackage 3, WP3). The application of these modelling techniques was carried out respectively for the whole Switzerland (regionalisation techniques) and for two test catchments (derived distribution approach and simulation techniques).
The regionalisation technique that turned out to offer satisfactory performances makes use of a dynamic definition of the homogeneous region, which the investigated site belongs to. This method, referred to as the Region of Influence (ROI) approach, is deemed to be an appropriate approach for regional flood frequency analysis, especially for a highly complex and differentiated environment, as compared to other methods, like simple- and multiscaling approach or traditional geostatistical techniques. Moreover, it can explicitely account for climate and catchment characteristics so being suitable for analysing the effects of anthropogenic forcings.
Results from WP2 showed that the derived distribution approach, which derives the probability distribution function of peak flows from the one of rainfall and from catchment charateristics, is a powerful tool that deserves further investigation, especially because of its capability to be easily applied in a distributed way and account for changes in the catchment.
Simulation techniques using an event-based model (WP3) have been finally extensively investigated, first to develop a suitable modified formulation of the Soil Conservation Service - Curve Number (CN) for alpine and prealpine areas. The CN-based model provided satisfactory performance in reproducing the main flood characteristics (peak, time to peak and volume), although further refinements are possible. The simulation of land-use changes by this model using realistic scenarios provided interesting insights on how the runoff generation is affected within the catchment and what quantitative impact one can expect from such changes.
|
