Use of uncertainty estimations for decision making: Case study, the hydrological and hydraulic model predictions for the river Demer and Dijle in Belgium
Vandenberghe, V.; Pereira, F.; Huygens, M.; Vanneuville, W. (2010). Use of uncertainty estimations for decision making: Case study, the hydrological and hydraulic model predictions for the river Demer and Dijle in Belgium, in: SIMHYDRO 2010: modèles hydrauliques et incertitudes, Nice, 2-4 juin 2010 [CD-ROM]. pp. [1-8]
In: (2010). SIMHYDRO 2010: modèles hydrauliques et incertitudes, Nice, 2-4 juin 2010: proceedings. Société Hydrotechnique de France: [s.l.]. ISBN 2-906831-83-2. , more
|
| Available in | Authors |
|
Document type: Conference paper
|
| Keywords |
Decision making
Hydrologic models
Modelling
Models > Scale models > Hydraulic models
Prediction > Flood forecasting
Sensitivity analysis
Uncertainty analysis
Belgium, Demer R. [Marine Regions]; Belgium, Dijle R. [Marine Regions]
Fresh water |
| Authors | | Top |
- Vandenberghe, V.
- Pereira, F., more
- Huygens, M., more
- Vanneuville, W., more
|
|
|
| Abstract |
It is under specialists who perform model studies and work with model results well known that techniques that are used for specific case studies work only well under specific circumstances and only are justified to be used when specific conditions are fulfilled. That means for flood risk maps that the reliability and the quality of the results is often only useful to create awareness and not to present the flooding area exact.
If during a discussion about eg. Spatial organisation planning it is decided that a higher accuracy is necessary, then the model shall have to be screened, to detect where and how the uncertainty on the model results can be reduced. In this study we quantified different important uncertainties that attribute to the final uncertainty on the results of maximum flows and water heights during storms. For each of the detected uncertainties the practical consequences for decision making are discussed. Different uncertainties give different consequences. Some are on the level of the decision when to alert people. Others are about monitoring and data collection. E.g. Is it sufficient to use Thiessen coefficients to calculate precipitation over an area with the amount of precipitation gauges that are available?
Other uncertainties are important for the modellers. E.g. Is my calibrated model able to make future predictions or are the uncertainty on the prediction bounds too wide? How well is my model calibrated? The final uncertainty on the result of the maximum water height is very important for the water manager that needs to take decisions about the creation of extra flood defending structures and solutions.
This research was done for a practical case study the river Demer and Dijle in Belgium. The model is implemented in Mike11 software. The results show the implications of important uncertainties related to stage discharge relationship, uncertainty of Nam parameters and uncertainty in manning’s n coefficient. It also shows that some of the uncertainties do not have to be taken into account for this case study like the uncertainty on the bathymetry or on the spatial distribution of the rain gauges.. |
|
