IX-B8, 2012
sumption of the
event, potential
direct runoff to
ic manipulation
results in the
' curve number
al soil retention
d is given as
(Eq. 1)
(Eq. 2)
, S is potential
Number ranges
m annual flood
a variety of
n area (USDA-
e and soil.
combination of
Antecenedent
il retention (S)
> Unit (HRU).
osed into sub
delineated into
, 1991), which
with a unique
of their spatial
nent. An HRU
the watershed,
. The estimated
er to obtain the
it Moisture
edent moisture
the impact of
olding capacity.
is for normal
dry condition
juivalent curve
*4 CN)
10-013CN2)
id Eq. 4)
ig technique
he Muskingum
e (1969) and
al Environment
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Research Council, 1975). The Muskingum-Cunge is a viable
alternative to particularly for the cases where hydrologic data
ie. stream flow data are not available, but where hydraulic data
(cross sectional data and channel slopes) can be readily
ascertained. Ponce and Yevjevich (1978) expressed the routing
parameters of the Muskingum-Cunge method in terms of the
Courant and cell Reynolds numbers, two physically and
numerically meaningful parameters. This method involves use
of a finite difference scheme to solve the Muskingum equation
where the parameters in the Muskingum equations are
determined based on the grid spacing for the finite difference
scheme and channel geometry characteristics. The Muskinghum
equation represents the relationship between reach storage and
discharge as flood wave propagates through a reach. Assuming
that cross sectional area of the flood flow is directly
proportional to the discharge at the section, total section is:
S=K [XI + (1-X) O] (Eq. 5)
Where S is the total storage in reach ( m), I inflow in reach
(m)/s), K is a proportionality constant known as Muskingum
travel time (unit of time) and X is a weighting factor ranges
between 0 to 0.5 (dimensionless).The equation of continuity for
the reach is given as follows:
I- Q - dS/dt (Eq.6)
3. STUDY AREA DESCRIPTION
Rangagora watershed located in the catchment of Kangsawati
river is considered for the present study (Figure 1). Mainly three
rivers namely Kansai, Kumari and Tongo are contributing the
flow in Kansavati river watershed. Geographically, the study
area is located between 86° 10’ and 86° 23’ East longitude and
27° 14’ and 23° 04 North latitude. Average annual rainfall of
the study area is around 1300 mm and its elevation ranges
between 200 to 640 m. Major portions of the study area is
occupied by loamy silt soils with slope varying between 1 to
15%.
T
8520 36 30
Figure 1. Location map of the study area
3.1 Data used
3.1.1 Hydro-meteorological data
Storm rainfall data at 3 hrs interval for few rainfall events and
at daily interval for five years from Central Water Commission
(CWC), Midnapore, West Bengal were collected. Daily
discharge data at outlet of reservoir for five years was collected
from Water Commission (CWC), Midnapore, West Bengal.
3.2.2 Satellite data
IRS -1D LISS III (Linear Imaging Self Scanner) data acquired
on 23" October, 2000 was used for generation of Land
sue/cover map.
3.2.3 Spatial inputs
SCS-CN method in combination with Muskingum-Cunge
routing technique requires a detailed knowledge of several
spatially distributed parameters affecting runoff viz., soil, land
use, antecedent soil moisture conditions, channel information
etc. Hence these model parameters were derived for each
hydraulic response unit (HRU) either from remote sensing data
or conventional maps under GIS environment, which can handle
voluminous input and output data. Using the digitized contour
map of the Kansavati watershed, Digital Elevation Model
(DEM) was produced with a grid size of 23 m x 23 m, and
subsequently stream network was generated under GIS
environment that facilitated the delineation of the study
watershed into sub watersheds. The land use/land cover map
was generated using IRS 1D- LISS III sensor data. Model grid
sizes were found to be the most important factor affecting
runoff and the model parameter database was computed for grid
cell sizes of 23, 46, 92, 184, 368, 736, 1472 m resolution.
4. RESULTS AND DISCUSSION
Spatial hydrological processes were simulated using distributed
hydrological approach involving SCS-CN method and
Muskingum-Cunge technique and are validated using stream
gauging data for few storm events. In the present study overland
flow was estimated using both distributed and lumped approach
respectively. These runoff depths serve as inputs to the channel
routing model. Simulated runoff hydrographs have been
generated at main outlet of the watershed. Four storm periods
have been used to develop and validate the results of runoff
hydrograph. Runoff hydrograph for the study area was
generated by using Muskingum-Cunge flood routing technique
and is shown in figure 2. The estimated and observed
hydrographs presented in figure show good simulation for all
the storm events considered. The simulated and observed runoff
discharge rates have been plotted as 1:1 line for both distributed
and lumped approach. It was observed from these figures that
simulated discharge rates using distributed approach are
uniformly scattered around 1:1 line, while scattered away from
the 1:1 line in case of lumped approach. CMR and RMSE most
widely used statistics reported for hydrology model calibration
and validation show reasonable fit between the simulated and
measured data. For distributed approach CMR value in all the
cases is negative and ranges from -0.1 to -0.3. In lumped
approach CMR value is always positive and ranges from 0.4 to
0.6. This indicates underestimation of runoff in case of lumped
approach. Similar trend was found for all the storm events
considered. The present analysis indicated that distributed
