16 Rood Forecasting and Control
365
the parameters have been changed in such a way that the difference between observed and computed flood hydrographs becomes minimal as can be seen in Fig.
16.4 (diagram bottom right - last trial). This optimization procedure can be done
manually by trial and error or by an automatic optimization technique.
16.3 Real-time Flood Control with the Aid of Flood Forecasts
Based on Remote Sensing Data - an Example
There are many ways to use remote sensing data for real-time flood forecasting
and control. In order to show, how this can be done in the real world an example
will be presented here. This example is not better than any other - it is shown here
only because the author is more familiar with this technique than with others.
16.3.1Basic Principle
In the field of water management real-time flood forecasts are used basically in
two different ways:
for flood warning purposes,
for flood control purposes.
Flood warning can help to become active long before the arrival of a flood peak,
thus preventing losses of lives and reducing flood damages to a certain extent.
Despite the benefits of such flood warnings flood forecasts do not prevent the
occurance of extreme floods and corresponding damages. These can be alleviated
significantly by efficient operation of flood protection reservoirs based on realtime flood forecasts, if provided at an early stage. The following example shows
the procedure to generate real-time flood forecasts followed by the presentation of
optimum flood control by reservoirs on the basis of such real-time flood forecasts.
The procedure is as follows:
(1) precipitation is observed by groundbased weather radar and the observed data
sets are transferred to a forecasting center,
(2) observed radar data during the last time intervalls are used for real-time quantitative precipitation forecasts (QPF) in the catchment area under consideration,
(3) a rainfall-runoff model is applied in order to transform the precipitation input
(consisting of observed and forecast rainfall) in order to generate a real-time
forecast of the expected flood hydro graph. The initial values at start of the
model run have to be either obtained from measurements of the relevant parameters in real-time, or the model has to be started at a very early stage of the
flood in order to allow the model parameters to adjust during a "warm-up period" of the model simulation run.
365
the parameters have been changed in such a way that the difference between observed and computed flood hydrographs becomes minimal as can be seen in Fig.
16.4 (diagram bottom right - last trial). This optimization procedure can be done
manually by trial and error or by an automatic optimization technique.
16.3 Real-time Flood Control with the Aid of Flood Forecasts
Based on Remote Sensing Data - an Example
There are many ways to use remote sensing data for real-time flood forecasting
and control. In order to show, how this can be done in the real world an example
will be presented here. This example is not better than any other - it is shown here
only because the author is more familiar with this technique than with others.
16.3.1Basic Principle
In the field of water management real-time flood forecasts are used basically in
two different ways:
for flood warning purposes,
for flood control purposes.
Flood warning can help to become active long before the arrival of a flood peak,
thus preventing losses of lives and reducing flood damages to a certain extent.
Despite the benefits of such flood warnings flood forecasts do not prevent the
occurance of extreme floods and corresponding damages. These can be alleviated
significantly by efficient operation of flood protection reservoirs based on realtime flood forecasts, if provided at an early stage. The following example shows
the procedure to generate real-time flood forecasts followed by the presentation of
optimum flood control by reservoirs on the basis of such real-time flood forecasts.
The procedure is as follows:
(1) precipitation is observed by groundbased weather radar and the observed data
sets are transferred to a forecasting center,
(2) observed radar data during the last time intervalls are used for real-time quantitative precipitation forecasts (QPF) in the catchment area under consideration,
(3) a rainfall-runoff model is applied in order to transform the precipitation input
(consisting of observed and forecast rainfall) in order to generate a real-time
forecast of the expected flood hydro graph. The initial values at start of the
model run have to be either obtained from measurements of the relevant parameters in real-time, or the model has to be started at a very early stage of the
flood in order to allow the model parameters to adjust during a "warm-up period" of the model simulation run.