Hydrological Modelling of Narmada basin in Central India using Soil and Water Assessment Tool (SWAT)
T. Thomas, N. C. Ghosh, K. P. Sudheer
National Institute of Hydrology, Roorkee (A Govt. of India Society under Ministry of Water Resources, RD & GR, Govt. of India)
Presentation Outline
• Introduction
• Objectives
• Study area
• Methodology
• Climate change detection in historical time horizons
• Hydrologic Modelling
• Impact assessment for future time horizons
Historical and Present time horizon:
Warming is Unequivocal and Unprecedented (IPCC).
Water is one of the sectors going to be impacted.
Future time horizons:
Global temperature change is likely to exceed 1.5°C for the end
of the 21st century relative to 1850.
Expected to have significant impacts on the hydrological
cycle.
Introduction
• Application of climate change science remains a challenge.
– Heavy analytical procedures
– Limited attention to implementation
– High level of uncertainty
– Political dimensions
• Indian sub-continent: combination of
– climate change
– rapid population growth
– land use change
– adaptation mechanisms
• Hydrologic modelling can be a useful tool for scenario analysis for
water resources planning
Identification of climate change signals in Narmada basin through statistical analysis of historical climate variables.
Modelling the watershed hydrology in Narmada basin by a
physically based large scale hydrologic model.
Assessment of the impact of climate change using projected
climate scenarios (AR5) with emphasis on future water
availability including extreme events.
Objectives
Objectives
• Catchment area: 98796 sq. km.
• Non-snowfed river basin
• Sustained by base flows
• West flowing river
• Lies in the States of MP, Gujarat,
• Maharashtra and Chhattisgarh
• Average annual rainfall: 1178 mm
Study Area
• 30 major & 150 medium projects planned in Narmada basin.
• Sustainability of the existing and upcoming projects uncertain.
• Hydrological impact assessments are much more imperative
now.
– Non snow-fed perennial basin
– Sustained only by base flow
– Change in water balance components
• Predictions of hydrologic variables in future under climate
change.
• Research gaps:
– impact assessment in data scarce areas
– Uncertainty issues,
– Non-stationarity issues.
Methodology
Observed Climatic Variables AND Hydrologic Variables (Precipitation, Temperature, Stream flows)
AR5 RCM Climate Variables & Downscaled Climate Data for 2 RCP’s and 4 time horizons
Multi-site Calibration & Validation SWAT Model after incorporating existing dams
Climate change impact assessments (existing dams only & with proposed and existing dams)
SWAT Run with default parameter values
LH-OAT for Sensitivity Analysis
Establish Initial Parameter Ranges
LH-sampling to assign 1000 groups of parameter values
B Behavioural simulations > 80% ?
Select parameter values with corresponding NS greater than 0.50
Use parameter range of last iteration as input range of next iteration
Yes
No
Detection of Climate Change Signals in Present Climate
TESTED SWAT MODEL
Forecasted hydrological variables with existing and proposed dams
Water availability, Flood, Drought Assessments (Indices based)
Addition of proposed dams in the basin
Bias Correction of the Downscaled Data (QM)
Forecasted hydrological variables with existing dams
Extreme rainfall: Increased in recent times
• 1-day maximum P: Significant positive trend (z = 3.66)
Particulars 1951-70 1989-08
1-day rainfall > 300 mm 1 9
1-day rainfall > 200 mm 8 17
Average 1-day maximum rainfall 203.9 mm 275.9 mm
Temporal variation of amount of 1-day maximum rainfall in the basin (MK test z statistic = 3.66).
Frequency analysis: Increase in n-yr return period rainfall
Grid cells with significant increasing trends in 5-year return period rainfall
Droughts: Increased in recent times
• Drought frequency more pronounced in the middle zone.
• Increase of drought in upper zone a cause of concern.
SPI based drought duration for various zones
Extreme Temperatures: Increasing continuously
• 1-day maximum temperature : increasing @ 1.10°C/100 yr.
(more than global average rate).
• 1-day minimum temperature: increasing @ 3.20°C/100 yr.
S.
No.
Zones of
basin
Mann-Kendall test statistic & inference
1969-88 1990-2009
1. Upper zone +1.72 (not significant) +1.01 (not significant)
2. Middle zone +1.24 (not significant) +1.29 (not significant)
3. Lower zone +2.31 (significant) +2.11 (significant)
Hydrologic Modelling using SWAT
Features of SWAT
• Physically based distributed model
• Continuous time model
• Long term yield model
• Uses readily available data
• Can be used for long term impact studies
Processes Modelled
• Weather, Hydrology, Sedimentation
• Plant growth, Nutrient cycling
• Pesticide dynamics, Bacteria
• Management
SWAT setup for the Narmada up to Barmanghat
• SWAT model setup for the Narmada upto Barmanghat (26000 sq. km).
• One Major Multi-purpose Project (Bargi Dam) for irrigation, power
generation, industrial and domestic use.
• 10 projects being planned in the catchment : Rosra, Basania, Halon,
Upper Narmada, Upper Burhner, Raghavpur, Atariya, Machhrewa,
Sher and Chinki.
• Virgin simulation of the model has been carried out before setting up of
reservoirs.
• Default run of model Bargi Reservoir (properties and inflows).
.
DEFAULT RUN: VIRGIN AND BARGI DAM
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
Jan
-88
Jun
-88
No
v-8
8
Ap
r-8
9
Sep
-89
Feb
-90
Jul-
90
De
c-9
0
May
-91
Oct
-91
Mar
-92
Au
g-9
2
Jan
-93
Jun
-93
No
v-9
3
Ap
r-9
4
Sep
-94
Feb
-95
Jul-
95
De
c-9
5
May
-96
Oct
-96
Mar
-97
Au
g-9
7
Jan
-98
Jun
-98
No
v-9
8
Ap
r-9
9
Sep
-99
Feb
-00
Jul-
00
De
c-0
0
Mo
nth
ly d
isch
arge
(cu
me
cs)
Calibrated (1988-00)
Virgin (1988-00)
• Sources of irrigation allocated: based on the actual water use of
reservoir water in command areas & groundwater in non-command
areas.
SWAT Parameters for Stream flow simulation
1. Base flow recession coefficient: ALPHA_BF
2. GW Delay time : GW_DELAY
3. Revap coefficient : GW_REVAP
4. Soil evaporation coefficient : ESCO
5. Available Water Content : SOL_AWC
6. Saturated hydr. Conductivity : SOL_K
7. Manning’s n : OV-N
7. Curve number : CN_f
8. Slope : SLOPE
9. Manning’s n for channel : CH_N
10. Water depth (shallow aq.) :GWQMN
11. Slope length of main channel : CH_S
12. Snowfall temperature : SFTMP
13. Slope of sub-basin : SLSUBBSN
14. Surface lag : SURLAG
• One at a time (LH-OAT) Sensitivity Analysis has been performed
for the parameters responsible for stream flow simulation.
Multi-site SWAT Calibration
• Multi-site calibration of SWAT model using 6 gauging sites
(Dindori, Mohgaon, Manot, Belkheri, Patan and Barmanghat)
• Calibration: 1988-00, Validation: 2001-05
• Sensitivity Analysis has been performed for the parameters
responsible for stream flow simulation.
• CN2, SOL_AWC, ESCO, GW_REVAP, ALPHA_BF and
GW_DELAY.
0
100
200
300
400
500
600
700
800
900
Jan
-88
Jun
-88
No
v-8
8
Ap
r-8
9
Sep
-89
Feb
-90
Jul-
90
Dec
-90
May
-91
Oct
-91
Mar
-92
Au
g-9
2
Jan
-93
Jun
-93
No
v-9
3
Ap
r-9
4
Sep
-94
Feb
-95
Jul-
95
Dec
-95
May
-96
Oct
-96
Mar
-97
Au
g-9
7
Jan
-98
Jun
-98
No
v-9
8
Ap
r-9
9
Sep
-99
Feb
-00
Jul-
00
Dec
-00
Mo
nth
ly d
isch
arge
(cu
me
cs)
Month/Year
Observed_1980-00
Calibrated (1984-00)
0
50
100
150
200
250
300
350
400
450
500
Jan
-88
Jun
-88
No
v-8
8
Ap
r-8
9
Sep
-89
Feb
-90
Jul-
90
Dec
-90
May
-91
Oct
-91
Mar
-92
Au
g-9
2
Jan
-93
Jun
-93
No
v-9
3
Ap
r-9
4
Sep
-94
Feb
-95
Jul-
95
Dec
-95
May
-96
Oct
-96
Mar
-97
Au
g-9
7
Jan
-98
Jun
-98
No
v-9
8
Ap
r-9
9
Sep
-99
Feb
-00
Jul-
00
Dec
-00
Mo
nth
ly d
isch
arge
(cu
me
cs)
Month/Year
Observed_1980-00
Calibrated (1984-00)
Simulated flows at Patan
Simulated flows at Dindori
Model calibration
0
50
100
150
200
250
300
Jan
-88
Jun
-88
No
v-8
8
Ap
r-8
9
Sep
-89
Feb
-90
Jul-
90
Dec
-90
May
-91
Oct
-91
Mar
-92
Au
g-9
2
Jan
-93
Jun
-93
No
v-9
3
Ap
r-9
4
Sep
-94
Feb
-95
Jul-
95
Dec
-95
May
-96
Oct
-96
Mar
-97
Au
g-9
7
Jan
-98
Jun
-98
No
v-9
8
Ap
r-9
9
Sep
-99
Feb
-00
Jul-
00
Dec
-00
Mo
nth
ly d
isch
arge
(cu
me
cs)
Month/Year
Observed_1980-00
Calibrated (1984-00)
0
500
1000
1500
2000
2500
3000
3500
4000
4500
Jan
-88
Jun
-88
No
v-8
8
Ap
r-8
9
Sep
-89
Feb
-90
Jul-
90
Dec
-90
May
-91
Oct
-91
Mar
-92
Au
g-9
2
Jan
-93
Jun
-93
No
v-9
3
Ap
r-9
4
Sep
-94
Feb
-95
Jul-
95
Dec
-95
May
-96
Oct
-96
Mar
-97
Au
g-9
7
Jan
-98
Jun
-98
No
v-9
8
Ap
r-9
9
Sep
-99
Feb
-00
Jul-
00
Dec
-00
Mo
nth
ly d
isch
arge
(cu
me
cs)
Month/Year
Observed_1980-00
Calibrated (1988-00)
Simulated flows at at
Belkheri
Simulated flows at
Barmanghat
Observed & Simulated Hydrographs at Barmanghat
0
500
1000
1500
2000
2500
3000
3500
4000
4500
Ja
n-8
8
Ju
l-88
Ja
n-8
9
Ju
l-89
Ja
n-9
0
Jul-90
Ja
n-9
1
Ju
l-91
Ja
n-9
2
Ju
l-92
Ja
n-9
3
Ju
l-93
Ja
n-9
4
Ju
l-94
Ja
n-9
5
Ju
l-95
Ja
n-9
6
Ju
l-96
Jan-9
7
Ju
l-97
Ja
n-9
8
Ju
l-98
Ja
n-9
9
Ju
l-99
Ja
n-0
0
Ju
l-00
Dis
charg
e (
cum
ecs) Observed
Simulated
0
500
1000
1500
2000
2500
3000
3500
Ja
n-0
1
Ap
r-01
Ju
l-01
Oct-
01
Ja
n-0
2
Ap
r-02
Ju
l-02
Oct-
02
Ja
n-0
3
Ap
r-03
Ju
l-03
Oct-
03
Ja
n-0
4
Ap
r-04
Ju
l-04
Oct-
04
Ja
n-0
5
Ap
r-05
Ju
l-05
Oct-
05
Dis
ch
arg
e (
cu
me
cs) Observed
Simulated
Model performance
S.
No.
Name of the
gauging site
Calibration Validation
NSE RSR PBias NSE RSR PBias
1. Dindori 0.85 0.39 -7.48 0.85 0.32 -4.56
2. Mohgaon 0.35 0.81 -37.60 0.90 0.28 -5.05
3. Manot 0.95 0.23 -8.91 0.97 0.14 -1.57
4. Patan 0.63 0.61 -50.45 0.68 0.51 -42.59
5. Belkheri 0.90 0.32 -5.80 0.70 0.48 -18.61
6. Barmanghat 0.79 0.46 13.93 0.79 0.28 8.33
Climate Change Impact Assessments
Global Climate Models
(GCM)
Representation Concentration
Pathways (RCPs)
Downscaling
Regional Climate
Models (RCMs)
Bias Correction
(Quantile mapping)
Calibrated and
Validated SWAT Model
Multiple outputs of
streamflow
Climate Models, RCPs and Extreme Events Indices
• 6 RCMs (0.50o x 0.50o)
• CCSM4
• CNRM-CM5
• GFDL-CM3
• ACCESS1.0
• MPI-ESM-L
• NOR-ESM-M
4 time-horizons
1970-05 (historical)
2006-40 (near term)
2041-70 (mid term)
2071-99 (end term)
2 RCP Scenarios
RCP4.5
RCP8.5
S. Index Rainfall range
1. Wet day > 2.50 mm
2. Heavy rainfall day 100 mm - 200 mm
3. Very heavy rainfall day > 200 mm
Source: European Climate Assessment & Dataset (ECA&D)
Sl. # Index Range
1. Extremely hot days MaxT > 45oC
2. Summer days MaxT > 40oC
3. Warmer Days MaxT > 35oC
4. Summer nights MinT > 25oC
Scenario Analysis: Higher rainfall variability
0 200 400 600 800 1000 1200 1400 1600 1800
2006
2009
2012
2015
2018
2021
2024
2027
2030
2033
2036
2039
RCM-Mean average annual rainfall (mm)
Year
0 200 400 600 800 1000 1200 1400 1600 1800
2041
2044
2047
2050
2053
2056
2059
2062
2065
2068
RCM-Mean average annual rainfall (mm)
Year
0 200 400 600 800 1000 1200 1400 1600 1800
2071
2074
2077
2080
2083
2086
2089
2092
2095
2098
RCM-Mean average annual rainfall (mm)
Year
RCM mean average annual rainfall under RCP4.5
Future scenario : Higher extreme rainfall events
0
50
100
150
200
250
300
350
400
450
ACCESS1.0 CCSM4 CNRM-CM5 GFDL-CM3 MPI-ESM-LR Nor-ESM-M RCM-Mean
1-d
ay m
axim
um
rain
fall
(mm
)
2006-40 2041-70 2071-99
Future 1-day maximum rainfall
Future scenario: Higher extreme temperatures
The summer days, warmer days as well as the summer nights are projected
to increase in all future time horizons
RCM Time
horizon
RCP4.5 RCP8.5
RCM
mean
2006-40 1 in 3.50 1 in 2.69
2041-70 1 in 2.73 1 in 2.61
2071-99 1 in 2.60 1 in 2.26
Extremely hot days
Time
horizon
Maximum of MaxT
(oC)
Minimum of MaxT
(oC)
Scenario RCP4.5 RCP8.5 RCP4.5 RCP8.5
2006-40 45.67 46.06 21.84 22.40
2041-70 45.64 46.20 21.13 21.43
2071-99 46.22 46.65 22.28 22.78
Average daily maximum temperature : CCSM4
Future scenario: Lower water availability
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
5 10 20 50 75 90 95
Annu
al de
pend
able
flow
(cum
ecs)
Probability of Exceedance (%)
HISTORICAL SIMULATEDACCESS1.0CCSM4CNRM-CM5GFDL-CM3MPI-ESM-LRNOR-ESM-M
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
5 10 20 50 75 90 95
Annu
al de
pend
able
flow
(cum
ecs)
Probability of exceedance (%)
HISTORICAL SIMULATEDACCESS1.0CCSM4CNRM-CM5GFDL-CM3MPI-ESM-LRNOR-ESM-M
0.0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
800.0
5 10 20 50 75 90 95
Annu
al de
pend
able
flow
(cum
ecs)
Probability of exceedance (%)
HISTORICAL SIMULATEDACCESS1.0CCSM4CNRM-CM5GFDL-CM3MPI-ESM-LRNOR-ESM-M
Future scenario: More surface water droughts
-1.50
-1.00
-0.50
0.00
0.50
1.00
20
06
20
10
20
14
20
18
20
22
20
26
20
30
20
34
20
38
20
42
20
46
20
50
20
54
20
58
20
62
20
66
20
70
20
74
20
78
20
82
20
86
20
90
20
94
20
98
3m
onth
-SD
I (J
une t
o A
ugu
st)
Hydrological drought scenario at Barmanghat (RCP4.5)
Future scenario: More groundwater droughts
-3
-2
-1
0
1
2
3
4 Ju
n-7
1
Ju
l-72
Aug-7
3
Sep-7
4
Oct-
75
No
v-7
6
De
c-7
7
Jan-7
9
Fe
b-8
0
Ma
r-8
1
Apr-
82
Ma
y-8
3
Ju
n-8
4
Ju
l-85
Aug-8
6
Sep-8
7
Oct-
88
No
v-8
9
Dec-9
0
Ja
n-9
2
Fe
b-9
3
Ma
r-9
4
Apr-
95
Ma
y-9
6
Ju
n-9
7
Ju
l-98
Aug-9
9
12
m-S
PI
Groundwater Drought Scenario RCP8.5
Increased frequency of dry and wet events
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1988-05 2006-40 2041-70 2071-99
# of
eve
nts/
year
Extremely dry
Severely dry
Moderately dry
(a) dry events
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1988-05 2006-40 2041-70 2071-99
# of
eve
nts/
year
Extremely wet
Severely wet
Moderately wet
(b) wet events RCP8.5 RCP8.5
Conclusions SWAT can be effectively used for hydrologic modelling of large river basins
and used effectively for scenario analysis.
Increasing maximum and minimum temperature during the historical and
all future time-horizons.
Higher extreme rainfall events in the recent and future time-horizons.
Lower water availability during all future time-horizons.
Magnitude of low flows and high flows to decrease in future time-horizons.
More frequent droughts in all future time-horizons.
Simultaneous occurrences of droughts and wet events with changing
rainfall pattern with frequent dry spells calls for optimal water utilisation.
Supply-side management of the available water coupled with demand-side
management can be considered as a viable climate change adaptation tool
Top Related