Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
WATER RESOURCES,
COASTAL AND
ENVIRONMENTAL
ENGINEERING
Editors
Dr. H. L. Tiwari,
Dr. S. Suresh,
Er. R. K. Jaiswal
Editors
Dr. H. L. Tiwari
Dr. S. Suresh
Er. R. K. Jaiswal
DISCLAIMER
The authors are solely responsible for the contents of the papers compiled in this volume. The
publishers or editors do not take any responsibility for the same in any manner. Errors, if any, are
purely unintentional and readers are requested to communicate such errors to the editors or
publishers to avoid discrepancies in future
Published by:
Excellent Publishing House
Kishangarh, Vasant Kunj, New Delhi 110070
Tel : 9910948516, 9958167102
E mail : exlpubservices@gmail.com
Website : www.excellent-publshing.com
Typeset by:
Excellent Publishing Services
Kishangarh, Vasant Kunj, New Delhi 110070
Tel : 9910948516, 9958167102
E mail : expubservices@gmail.com
Website : www.excellent-publshing.com
Preface
In the process of development, quality and quantity of the resources are generally depleted
day by day unless they are replenished by natural or artificial process. Water resource
which is an important resource to sustain the life on earth is under tremendous pressure all
over the world due to climate change, population growth and socioeconomic development.
Hence effective management of water resources with use of latest available technologies
and scientific research have become very crucial for water resources planners and
engineers. Aiming with this HYDRO 2104 INTERNATIONAL CONFERENCE on Hydraulics,
Water Resources, Coastal and Environmental Engineering jointly organized by MANIT
Bhopal and ISH in association with NIH Roorkee, IIT Bombay, VNIT Nagpur, SVNIT
Surat, Peoples University Bhopal during December 18-20,2014. HYDRO conference is
organized every year by ISH in association with Institutions/organizations.
Analysis and Design, Soft Computing Techniques, Water and Wastewater Management,
Water Quality Assessment and Modelling, Water Resource and Hydrology.
ii
Contents
Chapter 1.
Preface
Chapter 2.
Surge Protection Design for Water Conveyance System for the Case of Power
Failure To Pumps in Lift Irrigation Scheme Using SAP2
Chapter 3.
21
R.B. Katiyar, Shashank Tiwari, Piyush Pratap Singh, Sanjay Singh, Shakti Nath Das, S. Suresh
Chapter 4.
28
Chapter 5.
Flood Plain Mapping of Shivnath River By Using Gis and HEC RAS
Sanjeev Kumar Bhraria, Ishtiyaq Ahmad , M. K. Verma
42
Chapter 6.
52
Chapter 7.
Flood Disasters, River Training & Flood Control Measures in River Ganga
and Its Two Tributaries
61
S.K. Mazumder
Chapter 8.
70
Chapter 9.
83
Chapter 10.
96
Chapter 11.
109
Chapter 12.
124
Chapter 13.
139
iii
Chapter 14.
149
Chapter 15.
162
Chapter 16.
171
Chapter 17.
185
S. K. Debbarm, S. K. Biswal
Chapter 18.
196
Chapter 19.
205
M. Athar, M Aamir
Chapter 20.
217
Shri Ram
Chapter 21.
226
Chapter 22.
Evaluation of Existing Equations for Maximum Scour Depth Near Spur Dikes
236
Chapter 23.
248
Chapter 24.
255
Chapter 25.
265
S. K. Verma
Chapter 26.
274
Chapter 27.
Groundwater Storage Analysis in Changing Land Use / Land Cover for Four
Districts of Upper Ganga Canal Command (1972 2011)
283
iv
Chapter 28.
295
Chapter 29.
305
Chapter 30.
Model Studies on Oscillating Water Column Based Simple Wave Energy Buoy
316
Chapter 31.
327
Chapter 32.
337
Chapter 33.
Use of Artificial intelligence for Sediment Rating and Gauge Discharge Curve
347
Chapter 34.
Event Based Stream Flow Estimation and Validation Using Semi Distributed
Hydrological Model in Netravati River Basin, Karnataka State, India
362
B. P. Ganasri, Neeraj Rajasekar, Tasneem Ashraf , Suruchi Sah, Raju A., Dwarakish G. S.,
Chapter 35.
Assessing Swat for Discharge and Sediment Yield Estimation From Satluj
Basin in Indian Himalayas
373
Chapter 36.
388
Chapter 37.
Rain fall Runoff Model Development Under Regulated River Flow Condition
399
Chapter 38.
413
V. D. Loliyana, P. L. Patel
Chapter 39.
430
Chapter 40.
441
Chapter 41.
453
Chapter 42.
463
S. Kumar, Z. Ahmad
Chapter 43.
475
Chapter 44.
Gaps and Scope of Turbulence Study Near Piano Key Weir (PKW)
486
\
Chapter 45.
493
Chapter 46.
502
Chapter 47.
512
Chapter 48.
522
Chapter 49.
531
Chapter 50.
542
Chapter 51.
552
R. Singh, D. Roy
Chapter 52.
557
Vishnu Arya
Chapter 53.
561
Chapter 54.
569
Chapter 55.
581
vi
Chapter 56.
598
Chapter 57.
604
Sangeeta Kumari
Chapter 58.
612
Chapter 59.
622
Chapter 60
637
Chapter 61.
646
Prakash Bhamare, Ravindra Shrigiriwar, Deepak kumar Meshram, Sanjay Pande, Anita Morkar
Chapter 62.
653
Chapter 63.
663
Chapter 64.
676
Chapter 65.
688
Chapter 66.
699
Chapter 67.
711
Chapter 68.
723
vii
Chapter 69.
737
Chapter 70.
747
Chapter 71.
759
Arun Goel
Chapter 72.
768
Chapter 73.
777
Chapter 74.
787
Chapter 75.
800
P. R. Dixit, S. N. Londhe
Chapter 76.
810
Chapter 77.
820
Chapter 78.
Estimation of Runoff and Flood Risk in the Narmada River Basin Using
Hydrological Time Series Data Mining
829
Chapter 79.
841
Chapter 80.
857
Chapter 81.
868
Chapter 82.
875
Chapter 83.
880
viii
Chapter 84.
891
Chapter 85.
901
Chapter 86.
911
S. Lata , S. R. Samadder
Chapter 87.
923
Chapter 88.
940
Chapter 89.
950
Chapter 90.
964
Chapter 91.
975
Chapter 92.
985
Chapter 93.
996
Chapter 94.
1006
Chapter 95.
1015
Chapter 96.
1030
Chapter 97.
1037
ix
Chapter 98.
1044
Chapter 99.
1053
Chapter 100.
1069
Chapter 101.
1080
M. A. Mohamed Ansari
Chapter 102.
1091
Chapter 103.
1097
Chapter 104.
1107
Chapter 105.
1116
P. Krishnendu, R. Balaji
Chapter 106.
1124
Chapter 107.
1136
Chapter 108.
1143
Chapter 109.
1153
Chapter 110.
Sea Water intrusion in the Coastal Area of Navsari District and Its Control
By Direct Surface Method
1165
Chapter 111.
1176
Chapter 112.
1188
Chapter 113.
1195
Chapter 114.
1206
Chapter 115.
1217
Chapter 116.
1226
Chapter 117.
1243
Chapter 118.
1254
Chapter 119.
1266
Chapter 120.
1278
Chapter 121.
1292
Chapter 122.
1306
Chapter 123.
1317
Chapter 124.
1327
xi
Chapter 125.
1340
Chapter 126.
1350
Chapter 127.
1359
Chapter 128.
1367
Chapter 129.
1378
Chapter 130.
1387
Chapter 131.
1397
Chapter 132.
1408
xii
Chapter - 63
Hydrological Modelling of Upper and Middle Narmada
River Basin, India Using Geospatial Tools
A.Gupta1, P.K. Thakur2 , B.R. Nikam3, A. Chouksey4
1
Abstract : The need for assessment of water resources availability in large and ungauged river basin is
frequent topic of discussion. It is now becoming increasingly important for water resources evaluation in
India. Water resources development activities have focused attention on development and application of
physically based hydrological models, which was used to simulate the impact of land and water use on
water resources. The main objective of this study was to test the performance and feasibility of the SWAT
(Soil and Water Assessment Tool) model for water balance study and prediction of stream flowin the
Upper and Middle Narmada River Basin of India, which can be used for understanding the effects of
future development and management actions. To simulate these impacts, long-term daily meteorological
data was used.The Sequential Uncertainty domain parameter Fitting algorithm (SUFI-2) of SWAT CUP
(Calibration and Uncertainty Program) with multiple sets of parametervalues is used for calibration and
validation, over the entire basin.This calibration and validation was done based on the observed daily
discharge data from India-WRIS (India Water Resources Information System).The goal was to bracket
most of the estimated data within the 95% prediction uncertainty (95PPU), by getting a significant
coefficient of determination (R2) and coefficient of efficiency (NS) between observed and estimated data.
The results after post calibration and validation indicates decrease in average annual water yield from
44.83 mm to 36.67 mm and R2 calculated before calibration 0.86 and increases to 0.88 after calibration.
The simulation results indicate that relatively small parts of the total basin area have a high impact on the
water balance in the catchment. It also indicates considerable reduction in surface runoff from 346.42 mm
to 320.91 mm during 1979 to 1987.The results of present work also indicates that the parameter
uncertainty is not the sole source of uncertainty; the model structure uncertainty is also important. These
processes are mainly associated with the existing large reservoirs regulating the runoff of the River
Narmada.
Keywords:Water Balance in Narmada River Basin, Hydrological Modelling, Soil and Water Assessment
Tool (SWAT).
1. INTRODUCTION
Presently high inhabitants expansion, fast urbanization and climate change along with the irregular
frequency and intensity of rainfall cause difficulty in appropriate water management and storage
plans. Therefore, there is an urgent need of evaluation of water resources at various scales, as it
plays a primary role in the sustainability of livelihood and regional economics throughout the
world. It is the primary safeguard against drought and plays a central role in food security at local
Hydraulics, Water Resources, Coastal and Environmental Engineering (HYDRO 2014)
97893 84935047
663
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
and national as well as global levels. Modern researches using satellite based data and GIS
techniques have created a very promising research tool for hydrological investigation and
interpretation of landscape. Surface hydrological indications are one of the promising scientific
tools for assessment and management of water resources.
SWAT, is acontinuous-time, semi-distributed, process based river basin model, developed
toevaluate the effects of alternative managementdecisions on water resources and nonpoint-source
pollutionin large river basins (Arnold et al., 2012). Arnold and Fohrer (2005) described the
expanding global use of SWAT as well asseveral subsequent releases of the model. Gassman et al.
(2007) provided further description of SWAT, including SWAT version 2005,and also presented
an in-depth overview of over 250SWAT-related applications that were performed worldwide. It
was developed to predict impact ofland management practices on water, sediment yield,
andagricultural chemical yield such as nitrogen, phosphorus andbiological oxygen demand,
chemical oxygen demand, runoffmodeling, water balances modeling of large basin. For
thecalibration analysis of this model Sequential Uncertainty Fitting (SUFI-2)program, linked with
ArcGIS i.e. SWAT CUP is being usednow a days (Manaswi et al., 2014).
Many other methods were usedin past to simulate hydrology and soils, land use and management,
also several models were developed to simulate single storm events using a square grid
representation of spatial variability (Young et al., 1987; Beasley et al., 1980). However, many of
these models did not consider subsurface flow, ET or plant growth. Continuous models were also
developed (Johansen et al., 1984; Arnold et al., 1990) but generally lacked sufficient spatial detail.
Narmada River, generally known as Life line of Madhya Pradesh is a fifth longest river in the
Indian subcontinent and it is the third longest river that flows entirely within India. It also longest
westwardflowing river that drains in toArabian Sea at 30 km west of Bharuch, Gujarat after
running for 1,312 km thorough Madhya Pradesh, Maharashtra, Gujarat. Narmada Basin. The
Narmada Basin, extends over an area of 98,796 km2most of which lies in Madhya Pradesh (86%)
Gujarat (14%) and a comparatively smaller area (2%) in Maharashtra. The Narmada River is
traditionally considered to be originated from Amarkantak in Anuppur district of Madhya Pradesh,
at an elevation of 1037 m above mean sea level. In Basin Atlas, CWC 2014, theNarmada Basin is
divided into 3 Sub-basins viz. Narmada Upper, Narmada Middle and Narmada Lower Sub-basin.
The drainage network of Narmada River consists of 19 major tributaries. The Upper and Middle
Narmada River Basin has an elongated shape with a maximum length of 844.86 km. from east to
west and a maximum width of 234 km from north to south. The hilly regions are in the upper part
of the basin, and lower middle reaches are broad and fertile areas well suited for cultivation. It has
been noted from the elevation map of the basin (Figure 2)that the maximum area of the basin falls
in the 300-500m elevation range. Maximum elevation is observed in the uppermost region of the
basin. The highest elevation in the entire basin is around 1,328 m.
There is a need for hydrological research of the Narmada Basin to support improved catchment
management programs that safeguard the degradation of soil and water resources in various
governing states. The lack of decision support tools and limitation of data concerning weather,
Hydraulics, Water Resources, Coastal and Environmental Engineering (HYDRO 2014)
664
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
hydrological, topographic, soil and land use are factors that significantly hinder research and
development in the area.The main objective of this study was to test the performance and
feasibility of the SWAT (Soil and Water Assessment Tool) model for water balance study and
prediction of stream flowin the Upper and Middle Narmada River Basin of India, which can be
used for understanding the effects of future development and management actions.
1.1 Use of GIS & Remote Sensing in Hydrological Modeling:
Hydrological modeling is the mathematical representation of the major components of hydrological
cycle in which components are derived with the help of various empirical and complex physical
based mathematical formulae. Hydrological models in spatial domain are mainly of two types:
Lumped models and Distributed models. In lumped model, spatial heterogeneity is not
considered i.e. it considers watershed as single entity with single rainfall input as a whole. It
assumes that whole grid is homogenous and physical property such as soil, land cover, climate, etc.
are same everywhere. These models do not use physical formulas to derive water balance
components. Also variations in meteorological, hydrological and geological parameters are
considered as one aggregated value. Whereas in distributed model, grid heterogeneity is considered
by dividing whole area into number of homogenous units and all the properties lying in the area are
given equal weightage(Krysanova et al., 1999; Singh and Frevert, 2006).
The synoptic and temporal coverage of an area or phenomenon by satellite based remote sensing
sensors has a potential advantage in distributed hydrological modelling of various scales.
Parameters such as runoff cannot be directly measured from remote sensing but can be estimated
with the help of hydrological modelling in which remote sensing data goes as major dynamic input.
Remote sensing has emerged as a powerful tool for cost effective data acquisition in shorter time at
periodic intervals (temporal), at different wavelength bands (spectral) and covering large area
(spatial).The availability of GIS tools and more powerful computing facilities makes it possible to
overcome many difficulties and limitations and to develop distributed continuous time models,
based on available regional information(Sahoo, 2013). Geographic Information System (GIS) helps
in generating various hydrological properties from Digital Elevation Model (DEM) such as
drainage network, flow direction map, flow accumulation map, aspect map, stream order, etc. It
also helps in satellite data storing, processing, interpreting and analysing.
1.2 Soil and Water Assessment Tool:
SWAT (Soil and Water Assessment Tool) is used for analysing the impact of land management
practices on water, sediment, and agricultural chemical yields in large complex watersheds(Setegn
et al., 2008). SWAT uses a modified formulation of the soil conservation service (SCS) curve
number (CN) technique to calculate surface runoff. The CN technique relates runoff to soil type,
land use and management practices and is computationally efficient (Arnold et al., 1995a). The
computational components of SWAT can be placed into eight major divisions: hydrology, weather,
sedimentation, soil temperature, crop growth, nutrients, pesticides, and agricultural
management(Sahoo, 2013). Although the model operates on a daily time step and is efficient to run
Hydraulics, Water Resources, Coastal and Environmental Engineering (HYDRO 2014)
665
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
for many years, it is intended as a long term yield model and is not capable of detailed, single-event
flood routing (Arnold et al., 1998). ArcSWAT represents both pre-processor and user interface to
SWAT model (SWAT User's guide, 2012).
In this study we focus on calibration, evaluation and application of SWAT 2012 model for
simulation of the hydrology of Upper and Middle Narmada River Basin. The main objective of this
study was to test the performance and feasibility of the SWAT 2012 model for prediction of stream
flow in Narmada Basin.
1.3 Calibration and Parameter Uncertainty Analysis:
Model calibration is the modification or adjustment of model parameters, within recommended
ranges, to optimize the model output so that it matches with the observed set of data and Sensitivity
analysis is the determination of the most influential independent parameter of the model in
predicting the flow (Khare et al., 2014).Vandenberghe et al. (2002) have proposed that a sensitivity
analysis should beperformed before model calibration to identify the most sensitive
parameters.Uncertainty Analysis of distributed model is based on generalized likelihood measures.
For which Sequential Uncertainty domain parameter Fitting algorithm (SUFI-2) of SWAT CUP
has been used, with some limitation in model structure (Singh, 2013).
In the new version of SWAT-CUP a more powerful SWAT edit program is available where all
SWAT parameter arehandled, including different soil layers and managementrotation-operation,
precipitation data etc.(Manaswi et al., 2014). Performance of the SUFI-2 techniques was evaluated
using five objective functions, namely P-factor, R-factor, coefficient of determination R2, Nash
Sutcliffe (NS) and coefficient of determination divided by coefficient of regression bR2 calculated
on daily and monthly time-steps(Singh, 2013).
2. STUDY AREA
Catchment area of the Upper and Middle Narmada basin, with outlet at Garudeshwar in Gujrat
state, extends over an area of 87,581.33 km2 and bounded on the north by the Vindhyas, on the east
by the Maikal range, on the south by the Sapuras and on the west by the Coastal Alluvial Plain.
Lying in the northern extremity of the Deccan plateau, the basin covers large areas in the States of
Madhya Pradesh, Gujarat and a comparatively smaller area in Maharashtra and Chhattisgarh. Study
area lies between the geographical extent of north 21.25 to 23.875 latitude and east 73.625 to
81.8125 longitude, as shown in Figure 1.
The Tropic of Cancer crosses the Narmada basin in the upper plains area and a major part of the
basin lies just below this line. The climate of the basin is humid and tropical, although at places
extremes of heat and cold are often encountered.Rainfall is heavy in the upper hilly and upper
plains areas of the basin. It gradually decreases towards the lower plains and the lower hilly areas
and again increases towards the coast and south-western portions of the basin. In the upper hilly
areas, the annual rainfall, in general, is more than 1400 mm but it goes up to 1650 mm in some
Hydraulics, Water Resources, Coastal and Environmental Engineering (HYDRO 2014)
666
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
parts.The average annual water potential of the basin is 45.64 BCM. The utilizable surface water in
the basin accounts to 34.50 BCM(Basin Atlas, CWC 2014; WRIS 2012).
667
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
Model. These characteristics include division of watershed into subwatershed, location of main
outlet, creating HRUs from Land Use Land Cover and Soil map. Next step for simulation in
modeling is defining climate of the catchment. This has been done by reading or generating
weather data of daily precipitation, maximum/minimum temperature, wind speed, solar radiation,
relative humidity and long term weather data (Aggarwal et al., 2007).
Automatic watershed delineation has been done by using SRTM DEM. The drainage map of the
basin isshown in Figure 6. The impact of hydrologic response units (HRU) definition on stream
flow has also been studiedin which the subwatershed obtained from a watershed is further
subdivided into landuse and soil characteristics. For this, SWAT requires the landuse, soil, weather
and terrain data sets for assessment of water yield at the desired outlet of the basin. The SWAT
Simulation menu allows to finalize the setup of all inputs for the SWAT model and perform
sensitivity analysis and auto simulation.
668
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
Calibration of watershed model is a challenging task because of input data uncertainties, model
structure and algorithms, parameterization and output. This can be accomplished manually or using
auto calibration tools in SWAT-CUP (Abbaspour et al., 2007). Using SUFI-2 provides advance
option in hydrological modelling and create control environment between large amounts of data
sets during parameter sensitivity analysis. The long time-series data of discharge is available for the
Garudeshwar gauging station and this was utilized to simulate the model parameters and calibrate
stream-flow correlation between simulated and observed data.A converged solution has been
reached when the objective functions such as coefficient of determination (R2) and coefficient of
efficiency (NS), reaches constant values.
4. RESULTS AND ANALYSIS
The Upper and Middle Narmada catchment has been divided into 51 sub-basins and 106
Hydrological Response Units (HRUs). The HRUs of this catchment have been categorized into
different classes mainly on the basis of landuse, soil and slope. It has been clearly observed from
LULC map that the Upper and Middle Narmada River Basin is a deciduous forest dominated area
followed by an agricultural land which contribute to the significant economic importance of the
area. The slope of the catchment has been divided into five classes, viz. 13 (Very Gentle), 35
(Gentle), 510 (Moderate), 1035 (Moderate to Steep) and > 35 (Very Steep),as shown in
Figure 7. It has been found that most of the catchment area has general smooth slope and it covers
about 6070% of the total catchment area but the rest of region especially near the origin of river
the area falls under steep slope category. This high-altitude area contributes to a significant amount
of soil erosion as well as high run-off, especially during monsoon periods may be partly due to
inadequate management practices. Clay and Loamy are the most dominating soil categories found
in this catchment, shown in Figure 5.
The result of initial simulations have clearly shown that the hydrology of the basin has not been
well represented in this setup of model, making calibration inevitable. Based on the calibration
results, the hydrology of the selected subbasins and the entire Upper and Middle Narmada River
Basin has been validated. After simulation, the default values of parameter sets has been used for
calibration for the year 1979 1983 asshown in Figure 8, which results in new values of
calibrating parameter, which has been used to validate for the year 1984 1987 as shown in
Figure 9.
The goal of this parameter fitting procedure was to bracket most of the estimated data within the
95% prediction uncertainty (95PPU). If upon reaching this goal a significant R2 and NS exits
between the observed and measured runoff data then the model can be referred to as calibrated.
However, Schuol and Abbaspour(2005) have suggested the practically it is sufficient to bracket 80
percent of measured data within the 95PPU. The 95PPU represents also the parameter uncertainty
resulting from the non-uniqueness of effective model parameters.
Five parameters were included in the calibration procedure: CN2 value i.e. SCS Runoff Curve
Number [unit less] and some groundwater parameters like ALPHA_BF i.e. Base Flow Factor
[days], GW_DELAY i.e. Groundwater Delay [days], GWQMN i.e. Threshold depth of water in the
669
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
shallow aquifer required for return flow to occur [m] and GW_REVAP i.e. Groundwater "revap"
coefficient [unit less].
The fitted values of each parameter shown in Table 1,has been used in first iteration of calibration
for the sensitivity analysis and varied by replacing the values,obtained at the end of each iteration,
within their recommended range.In each iteration, previous parameter ranges were updated by
calculatingthe sensitivity matrix and the equivalent of a Hessianmatrix (Magnus and Neudecker,
1988),followed by the calculation matrix. Parameters were then updated in such a way that new
ranges were always smaller than previous ranges and were centered on the best
simulation(Abbaspouret al., 2007).Table 2 & 3 shows the fitted values, obtained after ninth
iterations of calibration and values obtained after validation, respectively.
670
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
Fitted value
Minimum value
Maximum value
78.54
0.0
100.0
v__ALPHA_BF.gw
0.550000
0.0
1.0
v__GW_DELAY.gw
135.000000
30.0
450.0
v__GWQMN.gw
1.700000
0.0
2.0
v__GW_REVAP.gw
0.137000
0.02
0.2
Fitted value
r__CN2.mgt
Minimum value
Maximum value
69.92
35.0
98.0
v__ALPHA_BF.gw
0.163513
0.065973
0.243319
v__GW_DELAY.gw
178.5368642
25.2
420.0
v__GWQMN.gw
2.720288
2.604801
3.066749
v__GW_REVAP.gw
0.075097
0.063544
0.096552
Fitted value
Minimum value
Maximum value
66.47
35.00
98.00
v__ALPHA_BF.gw
0.145779
0.065973
0.243319
v__GW_DELAY.gw
147.1221125
25.2
420.0
v__GWQMN.gw
2.674093
2.604801
3.066749
v__GW_REVAP.gw
0.075097
0.063544
0.096552
For the evaluation of the calibration (and validation) performance of the model, two statistical
parameters, namely R, the squared correlation coefficient between the observed and simulated
output, which in SWAT is usually the stream flow, and NS, Nash-Sutcliffe efficiency parameter,
have been evaluated. Values of R > 0.6 and NS > 0.5 for the calibration of the daily and monthly
simulated stream flow are usually considered as adequate for an acceptable calibration (Santhi et
al., 2001).According to Norusis (1999) when the R2value is equal to 1, the model is considered to
Hydraulics, Water Resources, Coastal and Environmental Engineering (HYDRO 2014)
671
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
be good meanwhile if the R2 is lower than 0.5 (near to zero), the model would be considered as not
suitable.
Abbaspour et al. (2004, 2007)suggest to use two more measures i.e. p-factor and r-factor, in case
the comparison of R2 and NS is not adequate. The p-factor is the percentage of the measured data
bracketed by the 95PPU, which value should ideally be 1 and on the other hand, the r-factor, is a
measure of the quality of the calibration and indicates the thickness of the95PPU, which value
should ideally be near zero (Arnold et al., 2012) Table 4 shows the valuesof accuracy parameters
before calibration, after calibration (obtained after ninth iterations) and values obtained after
validation.
Table 4. Accuracy parameters for sensitivity analysis
VARIABLE
p_factor
r-factor
R2
NS
bR2
BEFORE
CALIBRATION
0.55
0.61
0.86
0.63
0.7009
AFTER
CALIBRATION
0.13
0.07
0.88
0.26
0.5638
AFTER
VALIDATION
0.06
0.05
0.93
0.74
0.7054
At first glance, the results appears to be diverse, however a closer look reveals the emergence of
clear patterns. The information gained from ninth iteration of calibration for the year 1979 1983,
results in new values of parameter sets, which has been used to validate the hydrological processes
over the entire basin for the year 1984 1987, as shown in Figure 10&11 respectively.
The mismatch in observed and best estimate of model in calibration and validation phase (i.e.
Figure 10&11) may have occurred due to error in measured input data e.g., rainfall and
temperature, error in measured data used in calibration e.g., river discharge or error in model
Hydraulics, Water Resources, Coastal and Environmental Engineering (HYDRO 2014)
672
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
parameters e.g., hydrologic processes. Thus it reveals that calibration must always be accompanied
by an assessment of the goodness of the calibration,taking intoaccount of all modeling errors. Apart
from error in parameters, it seems that not all processes were included in the model, especially
some that are important in case of large river basins. These processes are mainly associated with
the existing large reservoirs regulating the runoff of the River Narmada. It would be ideal to
include reservoirs and water use in the model, but readily available, detailed information on the
management of the reservoirs and on stored water in the wetlands are almost non-existent. Thus it
emphasizes that the parameter uncertainity is not the sole source of uncertainty, the model structure
uncertainity is also important.
Figure 12. Average Water Balance Components Figure 13. Rainfall and Runoff (1979 1983)
673
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
674
Hydrological Modelling Of Upper And Middle Narmada River Basin, India Using Geospatial Tools
Arnold JG, Srinivasan R,MuttiahRS, Williams JR (1998) Large Area Hydrological Modeling and
Assessment part1. Water Resources Assoc., 34; 73-89. DOI:10.1111/j.1752-1688.1988.1998.tb05961.x
Arnold JG and Fohrer N (2005) SWAT2000: Current capabilities and research opportunities in applied
watershed modeling. Hydrol. Proc. 19(3): 563-572. DOI:10.1002/hyp.5611
Arnold JG, Moriasi DN, Gassman PW, Abbaspour KC, White MJ, Srinivasan R, Santhi C, Harmel RD,
Griensven A van, Van Liew MW, Kannan N, Jha MK (2012) SWAT: Model Use, Calibration, and
Validation. Transactions of the ASABE. Vol. 55(4): 1491-1508
Basin Atlas (2014) Central Water Commission, Ministry of Water Resources, Government of India
Beasley AP, Huggins LF, Monke EJ (1980) ANSWERS: A Model for Watershed Planning. Transaction of
the ASAE. 23(4):938-944
Gassman PW, Reyes M, Green CH, Arnold JG(2007) The Soil and Water Assessment Tool: Historical
development, applications, and future directions. Trans. ASABE 50(4): 1211-1250
Johansen NB, Imhoof JC, Kittle JL, Donigian AS (1984) Hydrological Simulation Program-Fortran (HSPF):
Users Manual. EPA-600/3-84-066, U.S. EPA, Athens, Georgia
Khare D, Singh R, Shukla R (2014) Hydrological Modelling of Barinallah Watershed using Arc-Swat Model.
International Journal of Geology, Earth & Environmental Sciences. ISSN: 2277-2081, Vol. 4 (1) JanuaryApril, pp. 224-235
Krysanova V, Bronstert A, Mller-Wohlfeil DI (1999) Modelling river discharge for large drainage basins:
from lumped to distributed approach. Hydrol. Sci. J. 44, 313331
Magnus JR and Neudecker H (1988) MatrixDifferential Calculus with Applications in Statistics and
Econometrics. 2nd Edition. John Wiley, USA.ISBN-10: 047198633X, pp: 395
Manaswi CM, Thawait AK (2014) Application of Soil and Water Assessment Tool for Runoff Modeling of
Karam River Basin in Madhya Pradesh. International Journal of Scientific Engineering and Technology
(ISSN: 2277-1581), Volume No.3 Issue No.5, pp: 529-532
Sahoo S(2013)Assessing Hydrological Modeling of Bandu River Basin, West Bengal (India). International
Journal of Remote Sensing & Geoscience. Vol. 2, Issue 5
Santhi C, Arnold JG, Williams JR, Dugas, WA, Srinivasan R, Hauck LM (2001) Validation of the SWAT
model on a large river basin with point and nonpoint sources:Journal of the American Water Resources
Association, 37(5), 1169-1188
SchuolJ and Abbaspour K(2005) Limitations, Problems and Solutions in the Setup of SWAT for a LargeScale Hydrological Application. 3rd International SWAT Conference, Jul 11-15, 2005, EAWAG, ETH,
Zurich, Switzerland
Setegn GS, Srinivasan R, Dargahi B(2008) Hydrological Modelling in the Lake Tana Basin, Ethiopia Using
SWAT Model. The Open Hydrol. J, 2008, 2, 4962
Singh V,Bankar NK, Salunkhe SS,BeraAK, Sharma JR(2013)Hydrological stream flow modelling on
Tungabhadra catchment: parameterization and uncertainty analysis using SWAT CUP. Current Science, Vol.
104, No. 9: 1187-1199
Singh VP and FrevertD (2006) Watershed Models
Vandenberghe V, Van Griensven A, Bauwens W (2002) Detection of the mostoptimal measuring points for
water quality variables: Application to the river waterquality model of the river Dender in ESWAT. Water
Science and Technology, 46 (3), 1-7
WRIS (2012). River Basin atlas of India, Publication of Ministry of Water Resources, Government of India,
jointly created by Central Water Commission (CWC) and Indian Space Research
Organization (ISRO) under India WRIS project, pp. 144
Young RA, Onstad CA, Bosch DD, Anderson WR (1987) AGNPS, Agricultural Non-Point Source Pollution
Model. A Watershed Analysis Tool. USDA Conservation Research Report 35
675