The Muskingum scheme is one of the most popular and widely used methods in flood routing. The usual procedure in using this model is to optimize the parameters that represent physical characteristic of river. These parameters are gained from the upstream and downstream historical data and used to estimate the downstream hydrograph without any updating kernel. Improved method which is coupled with updating kernel like transfer function method needs upstream hydrograph (as Muskingum input) to predict downstream hydrograph while upstream flood data is not available at same time. Predicting the upstream hydrograph as the model input in the prediction period, causes the difference between the structure of this hydrograph and the upstream hydrograph in the calibration period. as a result, the optimized parameters in the calibration period do not provide proper performance in the prediction period. In this research, instead of using the observed upstream hydrograph in the calibration period, the predicted upstream hydrograph in the calibration period is used to optimize the parameters. In the light of this, parameters obtained via genetic algorithm provide better performance for predicting downstream hydrographs. This process lead to unify the structure of the upstream hydrographs in the calibration and forecast period. Also in this paper, constant parameter Muskingum stage (CPMS method) is presented and the performance of combining CPMS with transfer function method has been investigated and compared with the results obtained by combining Saint-Venant equation with transfer function method. The mentioned structural modifications show the improvement of downstream flood prediction which is applied by Muskingum model.
The
Muskingum
scheme is one of the most popular and
widely
used
methods
in flood routing. The usual procedure in using this model is to optimize the
parameters
that represent physical characteristic of river. These
parameters
are gained
from the
upstream
and
downstream
historical data and
used
to estimate the
downstream
hydrograph
without any updating kernel.
Improved
method
which
is coupled
with updating kernel like transfer function
method
needs
upstream
hydrograph
(as
Muskingum
input) to predict
downstream
hydrograph
while
upstream
flood data is not available at same time. Predicting the
upstream
hydrograph
as the model input in the prediction
period
, causes the difference between the structure of this
hydrograph
and the
upstream
hydrograph
in the
calibration
period
.
as
a result, the optimized
parameters
in the
calibration
period
do not provide proper performance in the prediction
period
. In this research,
instead
of using the observed
upstream
hydrograph
in the
calibration
period
, the predicted
upstream
hydrograph
in the
calibration
period
is
used
to optimize the
parameters
. In the light of this,
parameters
obtained via genetic algorithm provide better performance for predicting
downstream
hydrographs
. This process lead to unify the structure of the
upstream
hydrographs
in the
calibration
and forecast
period
.
Also
in this paper, constant
parameter
Muskingum
stage (
CPMS
method)
is presented
and the performance of combining
CPMS
with transfer function
method
has
been investigated
and compared with the results obtained by combining
Saint-Venant
equation with transfer function
method
. The mentioned structural modifications
show
the improvement of
downstream
flood prediction which
is applied
by
Muskingum
model.