View
2
Download
0
Category
Preview:
Citation preview
Assessing the impact of climate change forMahanadi basin using SWAT model
P. C. Nayak1, B. Venkatesh, T. Thomas and Y. R. Satyaji Rao
1Deltaic Regional Centre,National Institute of Hydrology,
Kakinada-533 003, AP
Objectives:
•Assessment of change in hydrometerological andhydrological data by employing statisticalsignificance test.•To downscale the GCMs output•To downscale the GCMs output•Hydrological assessment using SWAT model
Data collected
S. No. Gauaging site
Period Years Catchment area (km2)
1 Salebhatta 1971-2009 39 4650
2 Sundergarh 1978-2009 32 58702 Sundergarh 1978-2009 32 5870
3 Kesinga 1978-2009 32 11,960
4 Kantamal 1971-2009 39 19,600
IMD gridded rainfall data (10 X 10) from 1901 to 2004
GCM DataEnvironment Canada, Canadian Centre forClimate Modelling and Analysis, CanESM2,R1, R2, R3, R4, R5 (256 X 192)
Historical data (1951 – 2005)Historical data (1951 – 2005)
Future projection : 2006 - 2100 (eg., RCP2.6, 4.5, 8.5 Scenario)
MK Sign Test Statistics for Discharge data
S. No.Gauaging
site H alpha Z Inference
1Salebhatta 1 0.05 -3.989Decreasing trend in the Pre-whitened series1Salebhatta 1 0.05 -3.989whitened series
2Sundergarh 0 0.05 -0.2377NO trend in the Pre-whitened series
3Kesinga 1 0.05 35.79Increasing trend in the Pre-whitened series
4Kantamal 1 0.05 30.11Increasing trend in the Pre-whitened series
Monthly Total Precipitation [1961-2004]
400
500
600
Mon
thly
Pre
cipi
tatio
n (m
m)
Kantamal Kesinga Salebhata Sundergarh
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan
0
100
200
300
Mon
thly
Pre
cipi
tatio
n (m
m)
Months
Plot showing number of rainy days for different
sub-basin in Mahanadi river
1960 1970 1980 1990 2000 2010
105
120
135
150
165
180
135
150
165
180
195
210
225
Num
ber
of r
ainy
day
s
Kantamal Rainfall Trend line
Num
ber
of
rain
y d
ays Kesinga Rainfall
Trend line
1960 1970 1980 1990 2000 2010
1960 1970 1980 1990 2000 2010
105
120
135
150
165
180
1960 1970 1980 1990 2000 2010
90
105
120
135
150
165
180
195
Num
ber
of
rain
y d
ays
Salebhata Rainfall Trend Line
Nu
mbe
r o
f ra
iny
day
s
Years [1961-2004]
Sundergarh Rainfall Trend line
Plot showingnumber of rainydays for 1 daymaximumrainfall forKantamal 2
4
Rainfall more than 75mm Trend line
Num
ber
of r
ainy
da
ys
0
2
4 Rainfall more than 100mm Trend line
Num
ber
of r
ainy
da
ys
Kantamalsubbasin inMahanadi river
1960 1970 1980 1990 2000 2010
0
2
4
6
8
10 Rainfall more than 50mm Trend line
Year from 1961-2004
Num
ber
of
rain
y da
ys
0
Num
ber
of r
ainy
da
ys
Plot showingnumber of rainydays for 1 daymaximumrainfall for 0
2
4 Rainfall more than 75mm Trend line
Nu
mb
er o
f ra
iny
day
s
0.0
0.5
1.0 Rainfall more than 100mm Trend line
Num
ber
of r
ainy
day
s
rainfall forKesinga sub –basin inMahanadi river
1960 1970 1980 1990 2000 2010
0
2
4
6
8 Rainfall more than 50mm Trend line
Year from 1961-2004
Num
ber
of r
ainy
da
ys
0
Nu
mb
er o
f ra
iny
day
s
Plotshowingnumber ofrainy daysfor 1 daymaximumrainfall for
0
2
Rainfall more than 75mm Trend line
Nu
mb
er o
f ra
iny
days
0
2
Rainfall more than 100mm Trend line
Nu
mbe
r of
rai
ny d
ays
rainfall forSalebhatasub basin inMahanadiriver 1960 1970 1980 1990 2000 2010
0
2
4
6
8 Rainfall more than 50mm Trend line
Year from 1961-2004
Num
ber
of
rain
y da
ysN
um
ber
of
rain
y da
ys
Plot showingnumber of rainydays 1 daymaximum rainfallfor Ib sub-basin in
2
4
6
8 Rainfall more than 75mm Trend line
Nu
mb
er o
f ra
iny
days
0
2
4 Rainfall more than 100mm Trend line
Nu
mb
er o
f ra
iny
days
for Ib sub-basin inMahanadi river
1960 1970 1980 1990 2000 2010
0
2
4 Rainfall more than 50mm Trend line
Year from 1961-2004
Num
ber
of
rain
y d
ays
0
2
Nu
mb
er o
f ra
iny
days
Regime Shift Detection TestWavelet analysis is becoming a common toolfor analyzing localized variations of powerwithin a time series. By decomposing a timeseries into time–frequency space, one is ableseries into time–frequency space, one is ableto determine both the dominant modes ofvariability and how those modes vary intime.
Kantamal Sub-basin
1965 1970 1975 1980 1985 1990 1995 2000-2
0
2
4
Time (year)
Rai
nfal
l (m
m)
a) Rainfall in mm (seasonal)
b) Rainfall Wavelet Power Spectrum
0.5
1
c) Global Wavelet Spectrum
Time (year)
Per
iod
(yea
rs)
1965 1970 1975 1980 1985 1990 1995 2000
1
2
4
8
16
320 5 10 15
x 104Power (mm2)
1965 1970 1975 1980 1985 1990 1995 20000
5
10
15x 10
4
Time (year)
Avg
var
ianc
e (m
m2 )
d) 2-8 yr Scale-average Time Series
Kesinga Sub-basin
1965 1970 1975 1980 1985 1990 1995 2000-2
0
2
4
Time (year)
Rai
nfal
l (m
m)
a) Rainfall in mm (seasonal)
b) Rainfall Wavelet Power Spectrum c) Global Wavelet Spectrum
Time (year)
Per
iod
(yea
rs)
1965 1970 1975 1980 1985 1990 1995 2000
0.5
1
2
4
8
16
320 2 4 6 8 10 12
x 104Power (mm2)
1965 1970 1975 1980 1985 1990 1995 20000
5
10
15x 10
4
Time (year)
Avg
var
ianc
e (m
m2 )
d) 2-8 yr Scale-average Time Series
Salebhata Sub-basin
1965 1970 1975 1980 1985 1990 1995 2000-2
0
2
4
Time (year)
Rai
nfal
l (m
m)
a) Rainfall in mm (seasonal)
b) Rainfall Wavelet Power Spectrum
0.5
c) Global Wavelet Spectrum
Time (year)
Per
iod
(yea
rs)
1965 1970 1975 1980 1985 1990 1995 2000
0.5
1
2
4
8
16
320 5 10 15
x 104Power (mm2)
1965 1970 1975 1980 1985 1990 1995 20000
5
10
15x 10
4
Time (year)
Avg
var
ianc
e (m
m2 )
d) 2-8 yr Scale-average Time Series
Ib Sub-basin
1965 1970 1975 1980 1985 1990 1995 2000-2
0
2
4
Time (year)
Rai
nfal
l (m
m)
a) Rainfall in mm (seasonal)
b) Rainfall Wavelet Power Spectrum
0.5
c) Global Wavelet Spectrum
Time (year)
Per
iod
(yea
rs)
1965 1970 1975 1980 1985 1990 1995 2000
0.5
1
2
4
8
16
320 2 4 6 8 10
x 104Power (mm2)
1965 1970 1975 1980 1985 1990 1995 20000
0.5
1
1.5
2x 10
5
Time (year)
Avg
var
ianc
e (m
m2 )
d) 2-8 yr Scale-average Time Series
Downscaling PrecipitationThe area of typical GCM grid cells range between 10,000km2 and 90,0002 km, but watershed area is less than GCMGrid area, there is difference/ mismatch in spatial scales.
To overcome mismatched spatial scale:
– (1) based on analogies with different climatic zones – (1) based on analogies with different climatic zones
– (2) Downscaling methodologies (statistical/dynamic).
There are three types of statistical downscaling, namely weather classification methods, weather generators, and transfer functions methods.
Change FactorMethodology
Anandhi, A., A. Frei, D. C. Pierson, E. M. Schneiderman, M. S. Zion, D. Lounsbury, and A. H. Matonse (2011), Examination of change factor methodologies for climate change impact assessment, Water Resour. Res., 47, W03501, doi:10.1029/2010WR009104.
Downscaling Precipitation • CanESM2_1960_2005_R1 & CanESM2_2006_2100_R1_RCP26 • CanESM2_1960_2005_R1 & CanESM2_2006_2100_R1_RCP45• CanESM2_1960_2005_R1 & CanESM2_2006_2100_R1_RCP85• CanESM2_1960_2005_R1 & CanESM2_2006_2100_R2_RCP26 • CanESM2_1960_2005_R1 & CanESM2_2006_2100_R2_RCP45• CanESM2_1960_2005_R1 & CanESM2_2006_2100_R2_RCP85• CanESM2_1960_2005_R1 & CanESM2_2006_2100_R2_RCP85• CanESM2_1960_2005_R1 & CanESM2_2006_2100_R3_RCP26• .• .• .• .
• Total 75 projections are generated using CFM for each sub-basin)
Uncertainty analysis
• To establish confidence in the projected precipitation downscaled from GCMscenario outputs, it was important that the downscaled outputs represented thescenario outputs, it was important that the downscaled outputs represented thecurrent state of the precipitation regimes reasonably well. In fact, the confidenceon the reliability of the climate change anomalies computed from the scenariosrun relied on the downscaled outputs’ ability to represent the baseline Climate.
Non-parametric Uncertainty Test• The Wilcoxon Signed Rank test (Wilcoxon,
1945) is one of the best nonparametricmethods for conducting hypothesis tests(Conover, 1980) and widely used in(Conover, 1980) and widely used inuncertainty analysis of downscaled climateparameters provided with predictor scenariosof the GCMs (Dibike et al., 2008; Khan et al.,2006a,b).
Uncertainty Test• Analysis carried out using daily, monthly,
seasonal and annual GCM downscaledprecipitation data. If more than 20% data isoutside the 95% confidence level, is rejected.outside the 95% confidence level, is rejected.
• Monthwise (Jan, Feb, Mar, …time series)uncertainty analysis is carried out to check thevariation in projected future rainfall.
Decadal Rainfall variation
1300.0
1400.0
1500.0
1600.0
Kantamal
Kesinga
Salebhatta
Sungergarh
1000.0
1100.0
1200.0
0 1 2 3 4 5 6
Linear (Kantamal)
Linear (Kesinga)
Linear (Salebhatta)
Linear (Sungergarh)
CanESM2_1960_2005_R1 & CanESM2_2006_2100_R1_RCP26
Decadal Rainfall variation
400.0
450.0
500.0
550.0
Kantamal
Kesinga
Salebhatta
Sungergarh
Linear (Kantamal)
300.0
350.0
400.0
0 1 2 3 4 5 6
Linear (Kantamal)
Linear (Kesinga)
Linear (Salebhatta)
Linear (Sungergarh)
CanESM2_1960_2005_R1 & CanESM2_2006_2100_R1_RCP45
Decadal Rainfall variation
400.0
450.0
500.0
550.0
Kantamal
Kesinga
Salebhatta
Sungergarh
Linear (Kantamal)
300.0
350.0
400.0
0 1 2 3 4 5 6
Linear (Kantamal)
Linear (Kesinga)
Linear (Salebhatta)
Linear (Sungergarh)
CanESM2_1960_2005_R1 & CanESM2_2006_2100_R1_RCP85
Rainfall-runoff model development using SWAT Model
• 5 years data for warming upCalibration of Data
– Kantamal 1978 -2000– Kantamal 1978 -2000– Kesinga 1983-1990 – Slaebhata 1978-1985 – Sundergarh 1983-1992
• Rest of the data used for Validation
SWAT model development
Topographic information were derived usingDigital Elevation Model (DEM) data (the SRTMDEM have been used).
The soil data available in NBSS&LUP, whichcontains soil maps at a 1:250,000 scale.
Calibration of the model was done by adoptingthe manual calibration procedure.
Calibrated Model ParametersParameter Kantamal Kesinga Sundargarh Salebhata
Calibrated Rank Calibrated Rank Calibrated
Rank Calibrated
values
Rank
values values values Alpha_Bf 0.95 6 6.84E-01 6 0.18 5 0.03 7
Cn2 76.40 1 81.70 1 76.10 2 85.10 1Epco 0.42 9 0.82 9 0.61 1 0.00 10Esco 0.04 4 0.77 4 0.65 4 0.12 4Esco 0.04 4 0.77 4 0.65 4 0.12 4Gw_Delay 3.00 10 7.38 10 5.33 6 0.00 9
Gwqmn 402.00 3 820.00 3 745.00 7 0.27 3
Rchrg_Dp 0.76 2 0.16 2 0.70 10 0.66 2
Sol_Awc 12.00 5 21.50 5 20.50 3 0.10 5
Sol_K 16.60 7 19.90 7 19.80 9 0.04 6Surlag 0.92 8 0.00 8 0.25 8 0.02 8
NSE 0.78 0.82 0.76 0.63
Observed and Computed plots during calibration period
7000
14000
7000
14000 Observed
Computed)
Dis
char
ge [
m3 /s
]
Kesinga Observed
Computed)
Kantamal
1/1/1983 1/1/1985 1/1/1987 1/1/1989 1/1/1991
0
6/19/1980 6/19/1985 6/19/1990 6/19/1995 6/19/2000
0
6/17/1976 6/17/1978 6/17/1980 6/17/1982 6/17/1984
0
2000
4000
6000
6/16/1981 6/16/1984 6/16/1987 6/16/1990
0
2000
4000 Observed
Computed)
Dis
char
ge [
m3
/s]
Period [Daily]
Salebhata
Observed
Computed)
Period [Daily]
Sundergarh
FUTURE WATERRESOURCESSCENARIO –DEPENDABLE
FLOWS AT KESINGA
0 20 40 60 80 100
0
200
400
600
800
1000
1200
1400
RCP 2.6RCP 4.5 RCP 8.5
De
pend
able
flo
ws
(cum
ecs)
0 20 40 60 80 100
0
200
400
600
800
1000
1200
1400
RCP 2.6RCP 4.5 RCP 8.5
400
600
800
1000
1200 RCP 2.6RCP 4.5 RCP 8.5
De
pend
able
flo
ws
(cum
ecs)
400
600
800
1000
1200
1400
RCP 2.6RCP 4.5 RCP 8.5
0 20 40 60 80 100
0
200
400
De
pend
able
flo
ws
(cum
ecs)
I0 20 40 60 80 100
0
200
400
M
0 20 40 60 80 100
0
200
400
600
800
1000
1200
1400
RCP 2.6RCP 4.5 RCP 8.5
De
pend
able
flo
ws
(cum
ecs)
Probability of exceedance (%)
0 20 40 60 80 100
0
100
200
300
400
500
600
700
800
900
Probability of exceedance (%)
Observed
FUTURE WATERRESOURCESSCENARIO –DEPENDABLE
FLOWS AT
0 20 40 60 80 100
0
500
1000
1500
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flo
ws
(cum
ecs)
0 20 40 60 80 100
0
500
1000
1500
RCP 2.6 RCP 4.5 RCP 8.5
1000
1500
RCP 2.6 RCP 4.5 RCP 8.5
De
pend
abl
e flo
ws
(cu
me
cs)
1000
1500
RCP 2.6 RCP 4.5 RCP 8.5F
KANTAMAL
0 20 40 60 80 100
0
500
De
pend
abl
e flo
ws
(cu
me
cs)
I
0 20 40 60 80 100
0
500
M
0 20 40 60 80 100
0
500
1000
1500
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flo
ws
(cum
ecs)
Probability of exceedance (%)
0 20 40 60 80 100
0
500
1000
1500
Probability of exceedance (%)
Observed
FUTURE WATERRESOURCES SCENARIO
– DEPENDABLEFLOWS AT SALEBHATA 0 20 40 60 80 100
0
50
100
150
200
250
300
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flow
s (c
umec
s)
0 20 40 60 80 100
0
50
100
150
200
250
300
RCP 2.6 RCP 4.5 RCP 8.5
150
200
250
300
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flow
s (c
umec
s)
150
200
250
300
RCP 2.6 RCP 4.5 RCP 8.5
0 20 40 60 80 100
0
50
100
Dep
enda
ble
flow
s (c
umec
s)
I
0 20 40 60 80 100
0
50
100
M
0 20 40 60 80 100
0
50
100
150
200
250
300
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flow
s (c
umec
s)
Probability of exceedance (%)
0 20 40 60 80 100
0
50
100
150
200
250
300
Probability of exceedance (%)
Observed
FUTURE WATERRESOURCES SCENARIO –DEPENDABLE FLOWS AT
SUNDERGARH0 20 40 60 80 100
0
200
400
600
800
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flow
s (c
umec
s)
0 20 40 60 80 100
0
200
400
600
800
RCP 2.6 RCP 4.5 RCP 8.5
400
600
800
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flow
s (c
umec
s)
400
600
800
RCP 2.6 RCP 4.5 RCP 8.5
0 20 40 60 80 100
0
200
Dep
enda
ble
flow
s (c
umec
s)
I
0 20 40 60 80 100
0
200
M
0 20 40 60 80 100
0
200
400
600
800
RCP 2.6 RCP 4.5 RCP 8.5
Dep
enda
ble
flow
s (c
umec
s)
Probability of exceedance (%)
0 20 40 60 80 100
0
100
200
300
400
500
Probability of exceedance (%)
Observed
Conclusions• Sign test indicates that discharge is increasing except Ong
tributary of Mahanadi River.• Number of rainy days is decreasing for all sub-basins.• Severe rainfall events are increasing for most of the basin
except Ib tributary.• Wavelet analysis indicated that there is change in rainfall• Wavelet analysis indicated that there is change in rainfall
pattern after 1980• Decadal analysis of projected rainfall shows that there is
decrease in rainfall for some scenarios 10 to 35 %.• All the RCP scenarios show that there is an increase in flood
in the future. It is shown that the RCP-2.6 and RCP-4.5 showmoderate increase in the runoff, whereas RCP-8.5 showssignificant increase.
Recommended