journal of applied science and engineering.pdf

Int. J. Mech. Eng. Res. & Tech 2017
ISSN 2454 – 535X www.ijmert.com
Vol. 9 Issue. 4, Dec 2017
Flood Hydrograph Estimation Using Lumped and Distributed
Models (Kabkian Basin Case Study)
DR.M.GANESH1
, MR.R.SRINIVAS2
, MR.D.SAMPATH3
, MS.B.SRUJANA4
ABSTRACT: The mechanisms of rainfall and runoff were studied in the Kabkian basin (846.5 km2) in
Kohgilouye and Boyerahad, Iran.Since the hydrologic characteristics may change from one sub-basin to the
next, this analysis began by treating the Kabkian basin as a single entity. In this context, lumped models may be
referred to as "semi-distributed." HEC-HMS (Hydrologic Engineering Center, Hydrologic Modeling System)
and HEC-GeoHMS (Hydrologic Engineering Center, Geospatial Hydrologic Modeling Extension) are
hydrologic models. Models of rainfall and runoff were investigated, and in both situations the SCS curve number
approach (Soil conservation Service, 1972) was taken into account. Basin data was used to precisely calibrate
and verify the model. All flood episodes had determination and agreement coefficients more than 0.9, and the %
errors in peak flow and volume were within acceptable limits. The event model was then assessed using a locally
based sensitivity analysis. The sensitivity analysis focused on three factors of the event model: curve number,
initial abstraction, and lag time.There, in the Kabkian basin. The largest discrepancies between the produced
peak hydrographs and the baseline peak hydrograph were due to curve number in both the lumped and
distributed model. Semi- dispersed model performed better than Lumped model in capturing peak runoff
discharges and overall runoff volume. However, both models' overall performance was respectable.
Keywords: Key Terms: Semi-distributed model; Kabkian basin; HEC-HMS; Sensitivity analysis; Rainfall-
runoff modeling; HEC-GeoHMS; SCS; kohgilouye; boyerahmad.
INTRODUCTION
Watershed models now in use may be classified as
either being very basic, conceptual lumped models or
very advanced, physically based dispersed models.
Parameters in conceptual lumped models are
described collectively to provide an average value for
the basin as a whole. Different sub-basins within a
watershed may have distinctively different hydrologic
characteristics. In this context, lumped models may be
referred to as "semi-distributed." However, since they
use artificial means of converting precipitation into
runoff, they remain non-physically based. HEC-HMS
Version 3.2 was utilized for this analysis. The HEC
model represents a basin's hydrologic and hydraulic
components as linked systems, simulating the basin's
surface runoff response to precipitation. It works
particularly well for modeling floods. The basin
model in HEC-HMS consists of the loss, the
transform, and the base flow; all three are essential
processes. Sub-basins are smaller sections of a basin
that are modeled separately to account for their own
unique precipitation and runoff processes. Surface
runoff, a stream, or a reservoir might all be
represented by a single element. An element's unique
property and the mathematical relations describing its
physical processes are each defined by a variable. The
hydrographs of the stream flows at the basin outflow
are computed as a consequence of the modeling
procedure. For many of these issues, it would be ideal
to know the exact magnitude and the actual time of
occurrence of all stream flow events during the
construction period and economic life of the project.
ASSOCIATE PROFESSOR1,2,3,4
ELECTRONICS AND COMMUNICATION ENGINEERING
TRINITY COLLEGE OF ENGINEERING AND TECHNOLOGY, PEDDAPALLY
(mganeshtvm1719@gmail.com), ( srigana.57@gmail.com), (sampathdama402@gmail.com)
(burrasrujanagoud14@gmail.com)
Int. J. Mech. Eng. Res. & Tech 20221

The design, construction, and operation of many
hydraulic projects rely on an understanding of the
variation of the basin's runoff. If this data was
accessible at the planning and design phases of the
project, it might be used to choose the optimal design,
construction program, and operational process among
the many potential outcomes. Because it is impossible
to know the project hydrology in advance, water
resources development projects must instead use a
hypothetical set of future hydrologic conditions to
inform their planning, design, and management.
Engineering hydrologists have spent a great deal of
time trying to find acceptable simplifications of
complicated hydrologic processes and developing
sufficient models for the prediction of future
hydrologic conditions.
examines the hydraulic and hydrologic reactions of
basins to natural and man-made events. These
considerations have led to the creation of various
hydrologic models for use in flood forecasting and the
investigation of rainfall-runoff processes (Yusop and
Chan, 2007; Yener and orman,2008; Li and Jia, 2008;
Stisen and Jensen, 2008; Khakbaz and et al., 2009;
Salerno and Tartari, 2009; Amir and Emad,2010; Jang
and Kim, 2010; James and Zhi, 2010;).Similarly,
Asadi and porhemat (2012) calibrated and verified the
hydrologic parameters in the kabkian basin and the
delibajak subbasin. Due to the spatial nature of the
factors and precipitation affecting hydrologic
processes, GIS (geographic information systems) has
been more important in hydrologic research in recent
years. In the parameterization of decentralized
hydrologic models, GIS plays a crucial role. This is
done to counteract the oversimplification that occurs
when parameters are grouped together at the river
basin scale for depiction. Using a DEM (digital
elevation model), GIS applications may extract
hydrologic information including flow direction, flow
accumulation, watershed borders, and stream
networks. In this investigation, GIS was integrated
with HEC-HMS to evaluate the model's applicability
to the basins under consideration. This study uses
rainfall-runoff data from the Kabkian basin gathered
by 12 rainfall stations and 1 runoff station between
2008 and 2011 to calibrate basin characteristics (curve
number and initial abstraction).
The primary goals of this research are to (1) evaluate
the performance of the HEC-HMS hydrologic model
using statistical measures, and (2) calibrate, verify,
and analyze the sensitivity of the model for the
Kabkian basin using both lumped and distributed data.
MATERIALS AND METHODS
The Study Area
Southwest Iran is home to the kabkian basin, which
can be found to the west of Yasooj City, kohgilouye,
and boyerahmad Province. The basin is located
between the longitudes of 51° 05′ and 51° 37′ to the
east. There is a total of 846.5 km2 of basin space in
the kabkian basin, and the average channel slope is
0.014 meters per kilometer. The basin's elevation
varies from 1500 meters at the outflow to 3000 meters
at the basin boundary.Over 90% of the year's
precipitation falls between November and April,
mostly as frontal rains that causes flooding. The
average yearly temperature is about 12 c, and the
climate is damp and chilly. (Fig 1.)
Data used
Since 2000, the kogilouye and boyerahmad Regional
Water Authority has been keeping tabs on the kabkiab
basin's streamflow and precipitation. Twelve
raingauges in the central and lower regions of the
basin were used to gather precipitation data. At one-
hour intervals, stream flow measurements were
recorded at the basin's outflow (botari hydrometric
station). The local climatological station's weather
records were accessed. All simulations of hydrologic
models are run at an hourly time step.Software used
Hec-GeoHMS 5.0
Designed for engineers with little to no background in
GIS, this toolkit is a geospatial hydrological
resource[USACE-HEC, 2003]. It's an add-on for the
popular mapping program ArcMap. In this research, a
digital elevation model (DEM) of the basins is utilized
with Hec-GeoHMS to generate a river network and
divide the basins into subbasins. Kabkian basin
streamflow gages are botari in the subbasins
delineation procedure.
HEC-HMS 3.3

The Hydrologic Engineering Center of the US Army
Corps of Engineers created this program for
hydrologic modeling. Included are several of the
standard hydrologic approaches used to model river
basin rainfall-runoff dynamics. "[USACE-HEC,
2006]"
MODEL APPLICATION AND
CALIBRATION
Five flood events that occurred in the Kabkian Basin
throughout the course of the study's three-year time
frame (2009-2011) were utilized to validate the
models. A hydrologic model in HMS is identified by a
project name. Before an HMS project can be
executed, a basin model, a meteorological model, and
control parameters are required. From the data
obtained through HEC-GeoHMS for model
simulation, the basin model and basin characteristics
were constructed in the form of a backdrop map file
uploaded to HMS (Figs. 2 and 3). The user gauge
weighting technique was utilized to generate the
meteorological model from the observed precipitation
and discharge data, and the control specification
model was then generated. The simulation's temporal
structure is set by the control requirements, which
include a start date and time, an end date and time,
and a computation time step. Basin model data,
weather forecast data, and control parameters were all
brought together to make this system work. The
twelve raingauge stations, one for each sub-, that
collected historical data
model was validated using data from a single stream
gauge station in the Kabkian Basin and its
surrounding region. The time step utilized in the
simulation was one hour, which was determined by
the time span of the available observed data.
To simulate infiltration loss, the SCS curve number
approach was used. To simulate the process by which
surplus precipitation is converted into direct surface
runoff, the SCS (Soil Conservation Service) unit
hydrograph approach was used. To represent
baseflow, we used a constant monthly rate. To
simulate the stretches, we turned to the Muskingum
routing model.
In order to produce the simulated runoff hydrographs,
the values of the parameters associated with each
technique in HEC-HMS must be provided as input to
the model.Stream and basin features may be used to
estimate some of the parameters, whereas other
parameters cannot be approximated. Model
parameters are calibrated when precise estimation of
the necessary parameters is not possible; this is done
by conducting a systematic search for the parameters
that, when combined with data on rainfall and runoff,
provide the best match between the observed and
calculated runoff. Optimization refers to this kind of
methodical searching. Starting with rough estimations
of the parameters, optimization refines the model until
the simulated flow is as near as feasible to the
observed one.
To calibrate the model, a trial-and-error approach was
used, whereby the hydrologist would subjectively
modify parameter values between simulations to find
the minimum values of parameters that produce the
best match between the observed and simulated
hydrograph. Although the model was calibrated
manually, its acceptability and appropriateness for
usage in HEC-HMS was verified using the program's
in-built automated optimization technique. The need
should guide the selection of the goal function. Curve
number, starting abstraction, and percent impervious
area in the basin are the three factors utilized in the
SCS Curve Number approach for dealing with
infiltration loss in the subbasins. Since there are no
developed areas inside the subbasin, the impervious
percentage is set to zero. Consequently, the SCS curve
number method's last two parameters (curve number
and initial abstraction) were adjusted. For modeling
the change from precipitation surplus to direct surface
runoff, the SCS unit hydrograph approach
incorporates a lag time parameter. We also calibrated
this metric.
MODEL PERFORMANCE EVALUATION
METHODS:
The models are evaluated based on their ability to
forecast the timing and amplitude of hydrograph
peaks, as well as the volume of runoff, as well as the
degree of agreement between anticipated and
observed runoff discharges. Both models' accuracy in
performance throughout individual simulation periods
and as a whole were measured using the following
statistical metrics:
 Percent error in peak flow (PEPF).
The PEPF measure only considers the
magnitude of computedpeak flow and does not
account for total volume or timing of the peak:
𝑄𝑂 (𝑝𝑒𝑎𝑘) − 𝑄𝑆(𝑝𝑒𝑎𝑘)
𝑃𝐸𝑃𝐹 = 100 | |
𝑄𝑄(𝑝𝑒𝑎𝑘)
where 𝑄𝑂(𝑄𝑆) is the the observed (simulated)
flow.
 Percent error in
volume (PEV). The PEV
function only considers the
computed volumeand does not
account for the magnitude or
timing of the peak flow:

𝑉𝑂 − 𝑉𝑆
𝑃𝐸𝑉 = 100 | |
𝑉𝑂
(2)
where 𝑉𝑂(𝑉𝑆)is the volume of the observed
(simulated) hydrograph.
 Coefficient of correlation (R) . The lag-0
cross correlation coefficient was
calculated as:
∑𝑁 (𝑂𝑡 − 𝑂
) × (𝑆𝑡 − 𝑆)
𝑅 =
𝑡=1
√[∑𝑁 (𝑂𝑡 − 𝑂
)2 × ∑𝑁 (𝑆𝑡 − 𝑆)2]
𝑡=1 𝑡=1
(3)
Where 𝑂𝑡 (𝑆) is the observed (simulated) flow
at time t, and 𝑂
(𝑆)is the average observed
(simulated) flow during the calibration period.
 The relative root mean squared error,
RRMSE, were calculated as:
where N is the number of streamflow ordinates
and the meaning of the remaining symbols is the
same as inEquation (3).
SENSIVITY ANALYSIS
A sensitivity analysis is a technique for pinpointing
the model parameters that have the most significant
effect on the final model output. Specifically, it ranks
model parameters according to how much they
contribute to the total inaccuracy in model predictions.
Both regional and worldwide sensitivity analyses exist
(Haan, 2002). The event model was assessed using a
local sensitivity analysis in this research. The
sensitivity analysis focused on three factors of the
event model: curve number, initial abstraction, and lag
time. The final set of calibrated model parameters was
designated as the baseline or nominal set. The model
was then ran many times, with each parameter's
baseline value incremented by a factor of 0.7, 0.8, 0.9,
1.1, 1.2, and 1.3 while all other parameters were held
at their initial levels. After playing about with
different parameter values for the model, we
compared the generated hydrographs to the original
model's output.
RESULTS AND DISCUSSION
Each sub-basin that HEC-HMS considers represents a
different part of the basin-wide precipitation-runoff
process, as was explained in the introduction. For a
component to be represented, it must have a set of
parameters that define its unique properties and a set
of mathematical relations that characterize its
underlying physical processes. The calibrated
parameter values for the Lumped and Semi-distributed
Kabkian Basin are shown in Tables 1 and 2,
respectively. Until a satisfactory match was found
between the observed and simulated hydrographs, all
of the parameters were calibrated concurrently with
the exception of the sub-areas, which are fixed.
Basin's calibration and validation graphs are shown
below. The observed and simulated graphs correspond
well in Figs. 4 through 7. Additionally, in both the
calibration and validation basins, Tables 3 and 4
include both observed and simulated data. Table 5
provides a concise overview of the models' results. As
can be seen in the graphs above, the highest time
discrepancy between the predicted and actual peak
discharges was just one hour, making it suitable for
flood forecasting.
The absolute differences between the -30% and 0%
situations for each event model parameter are
summarized in Figures 8 and 9. The most striking
variations in both situations were caused by adjusting
the CN parameter, or Curve Number.
CONCLUSIONS
The results above
demonstrate that,
using the
aforementioned
historical flood data, the model successfully
forecasted the peak discharge. The predictions for the
size and timing of the flood were very close. This
demonstrates that HEC-HMS is appropriate for the
basin under consideration. We may infer that a
model's applicability and efficiency are not dependent
on its level of structural complexity. HEC-HMS is an
effective tool for flood forecasting despite its very
simple form. To verify HEC-HMS's usefulness for the
iran basins, its wider implementation should be
promoted. Semi-distributed model performed better
than Lumped model in capturing peak runoff
discharges and overall runoff volume. However, both
models' overall performance was respectable.In
addition, the sensitivity analysis included three event
model parameters: curve number, initial abstraction,
and lag duration. Both in the semi-distributed basin
and the lumped basin, The biggest shifts occurred
when the CN (Curve Number) parameter was altered.
We also compared the optimized hydrologic
parameters, curve number, and first abstraction. The
lumped example had a curve number of 62, an initial
abstraction of 34mm, and a lag time of 347 minutes.
The semi-distributed example has curve numbers
between 61 and 66 and starting abstractions between
33 and 40 mm. Basin slope, geologic formations, plant
cover, and land use all play a role in this variety.
REFERENCES
[1]. Applying a Conceptual Hydrologic Model to the
Classroom, Amir A. and Emad H. (2010), International
Journal of Engineering Education, Volume 26, Issue 4, Pages
1-11 [2]. Journal of Ecology and Environment and
Conservation 18(4):805-812, Asadi.A,Porhemat.J, 2012,
Calibration, verification, and sensitivity analysis of the HEC-
HMS hydrologic model (study area: Kabkian basin and
Delibajak subbasin), Iran.
Water Science and Engineering, Volume 3 Number 1 (2010):
𝑁 𝑆𝑡−𝑂𝑡
𝑅𝑅𝑀𝑆𝐸 = 100 × √
1
∑ ( )2
𝑁 𝑡=1 𝑂𝑡 (4)

Pages 14-22 James.O,Zhi-j. L,2010,Application of HEC-
HMS for flood forecasting in the Misai and Wan'an
catchments in China.
[4]. T. I. Jang and H. K. Kim. 2010.Storm hydrograph
estimation in a watershed with a variety of land uses using a
tweaked version of the TR-20 model.
Reference: Khakbaz, B., et al. Semi-distributed hydrologic
model calibration methods as a bridge between lumped and
distributed approaches. Journal of Hydrology: DOI:
10.1016/j.jhydrol.2008.10.016
[6]. 2008: Li, S. Y., R. Z. Qi, and W. W. Jia. Adjustment of
the parameters of the theoretical rainfall-runoff model. The
IAHR-ISHS 3rd Symposium and the 16th IAHR-APD
Congress Proceedings, pp. 55-59. University Press of China,
Tsinghua.
The relevance of MODIS snow cover data in validating and
calibrating a conceptual hydrologic model: Parajka J. and G.
Blosch. 2010. Journal of Hydrology, vol.
[8] Salerno, F., and Tartari, G., 2009, "A coupled approach of
surface hydrological modelling and Wavelet Analysis for
understanding the baseflow components of river discharge in
karst environments", Journal of Hydrology, Volume 376,
Issues 1-2, Pages 295-306.
Journal of Hydrology 354 (2008):131-148 cites the study of
Stisen and Jensen, "A Remote Sensing driven distributed
hydrological model of the Senegal river basin."
Hydrologic Modeling System: Technical Reference Manual,
U.S. Army Corps of Engineers (USACE), 2009b [10].
Hydrologic Engineering Center, U.S. Army Corps of
Engineers, Davis, California.
According to USACE (United States Army Corps of
Engineers) (2000a) [11],Enhancement of Spatial Modeling.
Hydrologic Engineering Center, U.S. Army Corps of
Engineers, User's Manual for HEC-GeoHMS, Davis,
California.
[12]. M. K. Yener and A. U. Orman. 2008. Hec-Hms and
Runoff Scenario Modeling Research in the Yuvacik Basin,
Turkey. Middle East Technical University, Civil Engineering
Department, 06531 Ankara, Turkey.
[13]. Runoff characteristics and use of HEC-HMS for
modeling stormflow hydrograph in an oil palm catchment, Z.
Yusop, C.H. Chan, and A. Attas. Water Science &
Technology Volume 56 Issue 8 Pages 41–48, 2007.

journal of applied science and engineering.pdf

More Related Content

Similar to journal of applied science and engineering.pdf (20)

More from nareshkotra (20)

Recently uploaded (20)

journal of applied science and engineering.pdf