A stochastic simulation framework for flood engineering

A. Efstratiadis, A. Koukouvinos, P. Dimitriadis, A. Tegos, N. Mamassis, and D. Koutsoyiannis, A stochastic simulation framework for flood engineering, Facets of Uncertainty: 5th EGU Leonardo Conference – Hydrofractals 2013 – STAHY 2013, Kos Island, Greece, doi:10.13140/RG.2.2.16848.51201, European Geosciences Union, International Association of Hydrological Sciences, International Union of Geodesy and Geophysics, 2013.

[doc_id=1384]

[English]

Flood engineering is typically tackled as a sequential application of formulas and models, with specific assumptions and parameter values, thus providing fully deterministic outputs. In this procedure, the unique probabilistic concept is the return period of rainfall, which is set a priori, to represent the acceptable risk of all design variables of interest (peak flows, flood hydrographs, flow depths and velocities, inundated areas, etc.). Yet, a more consistent approach would require estimating the risks by integrating the uncertainties of all individual variables. This option can be offered by stochastic simulation, which is the most effective and powerful technique for analysing systems of high complexity and uncertainty. This presupposes to recognize which of the modelling components represent time-varying processes and which ones represent unknown, thus uncertain, parameters. In the proposed framework both should be handled as random variables. The following computational steps are envisaged: (a) generation of synthetic time series of areal rainfall, through multivariate stochastic disaggregation models; (b) generation of random sets of initial soil moisture conditions; (c) run of hydrological and hydraulic simulation models with random sets of parameter values, picked from suitable distributions; (d) statistical analysis of the model outputs and determination of empirical pdfs; and (e) selection of the design value, which corresponds to the acceptable risk. This approach allows for estimating the full probability distribution of the output variables, instead of a unique value, as resulted by the deterministic procedure. In this context, stochastic simulation also offers the means to introduce the missing culture of uncertainty appreciation in flood engineering.

PDF Full text (1860 KB)

See also: http://dx.doi.org/10.13140/RG.2.2.16848.51201

Tagged under: Determinism vs. stochasticity, Floods, Uncertainty