Floods have the greatest damage potential of all natural
disasters worldwide and affect the greatest number of people due to
their being widely distributed in nature. Mitigating the effects of
floods, which is intended to reduce the loss of life and property, can
be accomplished by flood forecasting and warning systems. Most of the
flood forecasting models embedded in such systems are either physically
based or conceptual. The great disadvantage of using such flood forecasting
models is that they require a huge amount of data to describe the catchments
where it is intended to make forecasts. However, there exists another
important class of models called data driven models based on computational
intelligence techniques and which are gaining popularity nowadays in
flood forecasting.
In spite of the scientific and technological advances in the last few
decades, accurate flood forecasting is still the challenge for hydrologists
and engineers. The accuracy of flood forecasting may be a significant
factor influencing decision makers in the context of flood warning systems
in real time. The accuracy can be enhanced by using computational intelligence
methods such as artificial neural networks, decision and model trees,
fuzzy logic, support vector machines, non-linear dynamic and chaos theory,
committee machines, and including novel and innovative approaches of
computational intelligence methods such as committee machines, hybrid
modelling, and composite model construction approaches.
Since flooding is a complex and inherently uncertain phenomenon; and
whatever model is used for forecasting it is just a simplified representation
of the real world, forecasts are inherently uncertain in nature. However,
incorporating the uncertainty estimation in the forecast can help the
decision maker within the flood warning system and thus enhance the
reliability and credibility of both the forecast and the warning system.
This research aims to develop a framework for flood forecasting and
uncertainty modelling using computational intelligence methods. Thus,
the primary goal of this research is to develop a framework and methodology
for an uncertainty modelling in flood forecasting using computational
intelligence methods. The secondary goal is to develop a framework and
methodology for flood forecasting using computational intelligence methods
to improve the accuracy of forecast. The methodologies and tools developed
in this research will be applied to at least two pilot sites of the FLOODsite project and the Bagmati
catchment from Nepal. The proposed research can contribute to society especially in developing
countries by improving the accuracy of forecasting with uncertainty
assessment using computational intelligence methods.
