APPLICATION: FLOOD FORECASTING MODEL FOR KLODZKO CATCHMENT (POLAND) - Modelling Uncertainty in Flood Forecasting Systems - page 82

Chapter 5
APPLICATION: FLOOD FORECASTING
MODEL FOR KLODZKO
TCHMENT (POLAND) CA

Summary of Chapter 5
This chapter presents the application of the uncertainty assessment
methodology using temporal disaggregation (developed in Section 4.1)
to rainfall time series. The methodology uses temporal disaggregation
of the time series into subperiods and uses the fuzzy Extension
Principle for the propagation of the uncertainty through the forecasting
model. Genetic algorithms are used to determine the maxima and
minima, which is an essential part of the Extension Principle. A
rainfall-runoff model of the Klodzko catchment (Poland) built with
HEC-1 and HEC-HMS is used for the application. The results show
that the output uncertainty arising from the uncertain temporal
distribution of the precipitation can be significantly dominant over the
uncertainty arising from the uncertainty in the magnitude of the
precipitation. The results also show the potential applicability of
genetic algorithms in combination with the Extension Principle for
uncertainty assessment. The description of the Klodzko catchment and
the flood forecasting model are presented in Section 5.1. The
implementation of the methodology is described in Section 5.2. The
principle and the versions of genetic algorithms used in the present
application are presented in Section 5.3. The results of the application
presented in Section 5.4 are analysed as follows: (i) uncertainty in the
forecast discharges with reconstructed precipitation (with
disaggregation) and with uniform precipitation (without
disaggregation), (ii) comparison of results with 3 and 6 subperiods,
(iii) influence of time series correlation between subbasins, and (iv)
influence of the type of genetic algorithms used for the determination
of the minimum and maximum. The results are presented in the form
of upper and lower bounds on the forecasts and in the form of a
complete membership function for some selected results. The
conclusions and discussion are presented in Section 5.5.

5.1 Klodzko catchment flood forecasitng model

A methodology for the treatment of uncertainty using temporal disaggregation is
presented in Chapter 4 (Section 4.1). This chapter presents the application of the
methodology to the flood forecasting model of the Klodzko catchment in Poland. The