FRAMEWORK - Flash-flood Risk Assessment under the iMpacts of land use changes and river Engineering WORKs

Motivation

Because of the erratic behaviour of flash-flood processes and of the lack of extended data with fine resolution in space and time, flashy streams are highly sensitive to the present uncertainties in risk assessment. Flash floods actually produce large social and economic impacts as shown by recent figures on flood disasters in Europe, the United States and Latin America. Also, flash-flood prone areas in Europe have experienced in the last decade more severe floods than those predictable before, even based on long-term data series. The prediction of and prevention from flood hazard has been long a major task in both physical and engineering sciences. Nevertheless, current methods of evaluating flood flows and designing accordingly the structural and non-structural measures for flood mitigation is still far from a satisfactory assessment. This is because of complexity of physical processes controlling flood occurrence and severity, including nonstationarity produced by climate fluctuations and man-induced changes in the drainage basin and throughout its river network.

Objective

The main focus of FRAMEWORK is on the explanation of nonstationary effects on the occurrence and intensity of extreme floods in flashy streams. These are produced by climate forcing and anthropogenic changes of the river basin system, which involve land use, watershed management practices and river engineering works. FRAMEWORK aims at providing a revision of the current methods for flood prediction and prevention in flashy streams under the improved knowledge of the river basin system and its interaction with climate, land use and watershed management practices.

Methodology

FRAMEWORK provides enhanced basic understanding of flood generating processes, and substantial innovation in the methods and techniques for the development of the combined stochastic and deterministic approach to flood frequency estimation and associated risk assessment:

Regionalisation methods FRAMEWORK provides an assessment of the methods used to describe flood frequency regimes at the regional scale using a physically-based approach, so yielding a new multi-level technique to flood regionalisation. This is based on seasonality and scaling of storm and flood statistical properties, as the mirror image of the intrinsic complexity of flood generation processes. These are jointly used with revised statistical inference of extreme values in order to provide robust technique for regional analysis. Other significant results deal with methodological advances in the evaluation of the uncertainty affecting flood frequency estimates for extreme events.

Derived distribution methods FRAMEWORK has investigated the capabilities of a number of different analytical methods for the derivation of the flood frequency curve in a river site from deterministic basin characterisation and stochastic storm representation. Several advances in this area were achieved, but this approach is shown to be incapable of providing reliable estimates for statistical predictions in ungauged catchments if not supplemented by regional analysis. However, derived distribution methods provide a substantial diagnostic tool, as they provide a deeper insight of the parent flood frequency distribution, including its sensitivity to anthropogenic and climate forcing.

Simulation methods FRAMEWORK developed physically-based and spatially-distributed models to account for the effects of system nonstationarity on flood estimation throughout the river network. The conception, fabrication, validation and demonstration of simulation models specially oriented towards flood frequency estimation and risk assessment shows this approach is capable of reducing the uncertainties of traditional estimation methods in describing the river system. The increasing availability of computational resources supports improved modelling facilities, but effective improvements only emerge from the enhanced physical background in describing both the natural and man-modified river and riparian systems. The combination of stochastic and deterministic models shows that Monte Carlo simulation schemes can provide scenarios useful for decision makers, including the engineering design of structural flood mitigation schemes.

Regional and basin case studies The combination of the regional approach with the basin approach yields comprehensive criteria for evaluating the sensitivity of flood risk assessments to the investigated natural and anthropogenic forcing. The effect of land use changes and urban development of earlier rural areas has a decreasing effect with increasing basin area and increasing return period of extreme flows. FRAMEWORK is also designed to provide innovative methods for flood risk assessment under the requirement of accomplishing a common background to the practices for evaluation of flood risk in Europe. The results indicate the route towards the development of a common policy for mitigation of flood hazards, which includes land and urban planning strategies in order to evaluate economic and social impacts of flood hazards within a sustainable development perspective.

Conclusion

FRAMEWORK provides methodological advances including (1) the development of methods for physically-based regionalization of flood flows in the probability domain: these are useful to approach the flood hazard at the regional scale; (2) the assessment of the capabilities of derived distribution methods as a useful diagnostic tool for flood frequency analysis at the basin scale, and (3) the development of simulation methods for flood frequency estimation throughout the river network, which provide a comprehensive technology to investigate anthropogenic effects on the flood hazard using Monte Carlo experiments under a physically based skeleton. The demonstration of these methodologies was performed through the application of the methods and models at different and complementary spatial scales, that is, through (4a) regional case studies including Austria, Switzerland and North-Western Italy, and (4b) basin case studies, including selected flashy streams of Germany, Italy, Spain, Switzerland and the United Kingdom.

Various methods are presently used in European countries and in different regions of a country. Therefore, it is somewhat difficult to compare the different policies for the mitigation of flood hazard, because the effective level of acceptable risk in different geographic areas is not assessed homogeneously in space. Conversely, the optimal allocation of available resources for mitigation of flood risk requires a unified assessment of the methods to estimate flood occurrence, severity, and impacts. Accordingly, FRAMEWORK provides methodologies that can reduce these uncertainties, also initiating the scientific basis for a unified approach to risk assessment in the European areas subject to increasing vulnerability to flood disasters. This involves the search for a deeper insight of the unsolved complexity of flash flood processes, jointly with an appropriate framework to include the river basin system in the analysis of extreme hydrological events.