US SCS

1
A Study on the Improvement of the Applicability
of the Distributed River Simulation
Model(2005/08/08)
Mitsuo Yamashita Ph.D. Candidate Energy and
Environmental System Engineering, Graduate School
of Fukuoka University, Kitakyuushuu, Japan
2
My Profile

• Mitsuo Yamashita
• Born and bred in Fukuoka City
• Graduate of Japan National Defense Academy,
• Majoring Aeronautical Engineering
• Flying officer of Japan Air Self-Defense Force
  (-1994)
• Engineer of civil engineering consultant (-2002)
• Master of Recycling and Environmental
  Engineering.
• 
  (Fukuoka U. 2004)
• Ph.D. candidate of Energy and Environmental
• System
  Engineering(2004-)

3
Introduction
To solve the conflicts about water, the sharing
of detail information about water is important.
It is to be desired that the information should
be obtained by measuring of actual conditions.
But it is impossible to do the perfect
observation, and rainfall events have
uncertainties.
4
Then the estimation by the hydrological model is
needed as another approach.
5
The lumped model has widely been used to estimate
the peak flow rate or hydrograph
at the downstream end of the basins.
Whole Basin
Estimation of hydrograph at the downstream end
with the lumped model
Upstream
Cross section No.1
Downstream
The character of large basin is described by one
set of parameters.
But detailed information has been required in
these days.The lumped model cannot fully reply
on stakeholders demand.
6
Instead of the lumped model, the distributed
model has been proposed.
Upstream
Basin 3
Basin 1
Main river Channel and Tributaries
Basin 5
Basin 2
Basin 4
Basin 6
Downstream
It consists of sub-basins of tributaries and main
river channel.
rainfall-runoff model (lumped model)
sub-basin
river water simulation model
main river channel
Flow rate and flow stage can be obtained
at any point of main river channel
continuously.
7
As for the river water simulation model, there
are several commercial models, such as MIKE11,
HEC and Inforworks RS, with the progress of
mathematical techniques for hydraulic calculation.
one dimensional river channel model based on
the Saint-Venant equations.
The distributed hydrological modeling is getting
easier recently.
8
There are several problem still left for
application of distributed model.

• What is the appropriate size of sub-basins ?

2. How to provide the necessary information to
sub-basins individually, such as hyetograph ?
3. How to model the main river channel?
4. Which is the suitable rainfall-runoff model
for sub-basins?
9
Integrated Analysisfrom forested mountain to sea
Forested area is current object
Forest
Water Supply
Conduit
Urbanised Area
Paddy Field
Wetland
Sewage collection
Sea
10
Our Countermeasures
11
Objective Basin for the exact evaluation of our
model
Cacthment of Shimouke Artificial Reservoir
The catchment of Shimouke Artificial Reservoir,
Chikugo River, Oita pref. Kyuushuu Island, West
of Japan.
Area 182km2
Annual rain (2001) 2,289mm (Shimouke)
2,727mm (Taio)
Almost all of this basin is covered with
artificial forest of Cedar and Cypress.
Soil textures Layer-A CLSiL (-0.5cm-23cm)
Layer-B CL
(-23cm )
?CL Clay Loam, SiL Silty Loam
The exact inflow volume is obtained at a dam
site, from MLIT of Japan Qin?S Qout
Qin Inflow, Qout Outfall, S Reserved volume
12
Countermeasures
13
In this study, we would like to divide the basin
into sub-basins as small as possible.
In the distributed model, every sub-basin must
abut on the main river channel.
Upstream
Basin 3
Basin 1
Basin 5
Basin 2
Basin 4
Basin 6
Downstream
14
We used InfoWorks RS as river simulation
model. In this software, there is the limitation
that
Maximum number of node for hydraulic calculation
must be less than 10,000 for one time simulation.
Under this limitation, we designated the main
river channels as long as possible. And it
governed the size of sub-basin based on tributary.
15
Division of the basin
Modelling of main river channels
Shimouke Artificial Reservoir
Divide the basin into 129 sub-basins based on
tributaries (averaged area 1.4km2)
Estimate the hydrographs of every sub-basin with
the lumped model
Input the hydrographs into the river simulation
model
16
Countermeasures
17
Rainfall data must be provided for each sub-basin
individually
Matsubara
Ohno
shimouke
Shakadake
Tochino
Nakabaru
Taio
Kijiya
Tsuetate
Anagawa
Tatekado
Uchida
Rain gauge station
But there are only 12 rain gauge stations for 129
sub-basins, and the data from them are hourly
data.
18
We approximated the smooth surface of rainfall
intensity over the basin on the GIS,
Then the rainfall intensities at each centroid of
sub-basin were estimated from this surface.
This procedure was routinized, and the
hyetographs of 5-minute intervals for 11days were
estimated for 129 sub-basins.
19
Countermeasures
20
Shimouke Artificial Reservoir
Example of long section of a main river channel
in the objective basin
As for the steep river, more than
3-slope, Infoworks RS requests that the more
cross sections than usual.
Especially in Japanese mountainous area, surveyed
data can hardly be obtained at such
intervals. And the river cross sections change at
every rainfall event.
21
Cross Sections
Infoworks RS requires the intervals of cross
sections should be generally less than
(m)
Let S be the mean slope of the River
22
Adoption of Presumed Cross Section Like a Conduit
? Divide the river channel at an interval of 1km
length
? Assume the slopes of which are constant
? Apply the ideal triangular cross sections to
them
a
As the slope goes steep, the base angle goes
acute.
a 150(for 0)
a 30 (for 5 or above)
S mean slope of the river()
23
Countermeasures
24
Tentative lumped rainfall-runoff model for
sub-basins
Application of SCS-CN method
Surface runoff Pe
New original model
Soil water retention
Channel flow
25
Interflow model
(Presumed partial runoff model from soil water
retention)
( P-Pe ) Input to the soil retention
S Potential maximum retention
0 lt k lt 1 Interflow division factor
We calculate the effective rainfall for interflow
with this model.
26
Basic Equations of the tentative model
(1) Surface Flow
P Accumulated rainfall amount (mm), Pe Excess
rainfall for surface flow (mm), S Potential
maximum retention (mm)
(2) Retention Capacity
P05 Total 5-day antecedent rainfall (mm), So
Retention capacity for driest condition (mm), Sc
Retention capacity for wettest condition (mm),
Retention capacity compensation factor
There are 4 parameters to be
identified.
(3) Interflow
dm(t) Effective rainfall for interflow (mm), k
Interflow division factor (mm)
27
Identification of parameters
(First approximation)
The parameters should be identified from analysis
of observed rainfall and outflow data of each
sub-basin, but we could not obtained the flow
data except for the data at the downstream end
(Shimouke Artificial Reservoir).
Therefore, the parameters of whole basin were
estimated from analysis of averaged rainfall data
and flow data at the reservoir, as 1st
approximation.
And they were applied to all sub-basins.
28
Simulation Results With the parameters obtained
by the whole basin
Inflow hydrograph at the Shimouke Artificial
Reservoir
P05 is depicted as indication of soil moisture.
The peak flow rates of the rain on 07/21 were
calculated excessively and the recession after
the peak was so rapidly than that of observed
one. The problem was assumed to be in the
estimation of retention capacities of the soil
from the sensitivity analysis of parameters.
29
Results of trial simulation With the doubled
values of retention capacities
Inflow hydrograph at the Shimouke Artificial
Reservoir
Averaged Rainfall Intensity
Rainfall Intensity
Observed hydrograph
Simulated hydrograph
Inflow to Resorvoir(m3/s)
The tendency of simulated hydrograph was fairy
agree with observed one. It means that the
estimation of retention capacities of each
sub-basin is most important for a rainfall runoff
model for forested area.
30
Conclusion
It is definitely shown by this study that a
long-term simulation with the distributed model
is easy for public people with commercial
simulation model and ordinary PC.
Retention capacities (So, Sc) are dominant
parameters in our model and they might be related
with the volume of soil pores. This year, we
would like to do the field research about them.
Other parameters( , k) had better to be
identified from analysis of response of
hydrograph and total water volume.
31
For further investigation
The parameters must be identified from analysis
of observed rainfall data and outflow data of
small sub-basins, at first. To do so, many
observed data about rainfall, outflow data of
short intervals and conditions of soil are
needed. For further improvement of distributed
model, the long term observation of various types
of small basins at short time intervals such as 5
minutes should be done.
Thank you for your kind attention.