Transcript Impact of Climate Change on Flow in the Upper Mississippi

Impact of Climate Change on
Flow in the Upper Mississippi
River Basin
Eugene S. Takle
Iowa State University
Ames, IA 50011 USA
[email protected]
Project collaborators:
Manoj Jha, Zaitao Pan, Roy Gu
If we have perfect predictability
of global fields, how well can
we downscale this
predictability to stream flow at
one point?
Outline








Domain and hydrological model (SWAT)
Calibration and validation
Observations -> stream flow
NNR -> RCM -> SWAT-> stream flow
GCM -> RCM -> SWAT-> stream flow
GHG -> GCM -> RCM -> SWAT -> stream flow
Stream flow vs. precipitation
Use for policy development
For details see: Jha, M., Z. Pan, E. S. Takle, and R. Gu, 2004:
Impacts of climate change on stream flow in the Upper Mississippi
River Basin: A regional climate model perspective.
Journal of Geophysical Research.
Sub-Basins of the
Upper Mississippi
River Basin
119 sub-basins
Outflow measured
at Grafton, IL
Approximately one
observing station
per sub-basin
Approximately one
model grid point
per sub-basin
Soil Water Assessment Tool
(SWAT)




Long-term, continuous watershed
simulation model (Arnold et al,1998)
Daily time steps
Assesses impacts of climate and
management on yields of water, sediment,
and agricultural chemicals
Physically based, including hydrology, soil
temperature, plant growth, nutrients,
pesticides and land management
SWAT Output with Various Sources
of Climate Input
Calibration of SWAT:
Annual Stream Flow at Grafton, IL
Calibration of SWAT:
Monthly Stream Flow at Grafton, IL
Validation of SWAT:
Annual Stream Flow at Grafton, IL
Validation of SWAT:
Monthly Stream Flow at Grafton, IL
RegCM2 Simulation Domain
Red = global model grid point
Green/blue = regional model grid points
Annual Stream Flow Simulated by SWAT
Driven by the RegCM2 Regional Climate
Model with NNR Lateral Boundary Conditions
Seasonal Stream Flow Simulated by SWAT
Driven by the RegCM2 Regional Climate
Model with NNR Lateral Boundary Conditions
Mean Monthly Precipitation Simulated by
the RegCM2 Regional Climate Model
with NNR Lateral Boundary Conditions
120
(a)
RegCM2
Snowfall (mm)
100
SWAT
80
60
40
20
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
60
(b)
RegCM2
SWAT
50
Runoff (mm)
40
30
20
10
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec Aver.
120
(c)
RegCM2
Evapotranspiration (mm)
100
SWAT
80
60
40
20
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
120
(d)
RegCM2
100
Snow melt (mm)
SWAT
80
60
40
20
0
Jan Feb Mar
Apr May
Jun
Jul Aug Sep Oct
Nov Dec
Hydrological component comparison
between RegCM2 and SWAT
RegCM2
SWAT
Evapotranspiration
588
528
Surface runoff
151
166
Snowmelt
256
240
Note: All values are in mm per year averaged for 1980-1988 in NNR run.
Ten-Year Mean Precipitation Generated by the RegCM2
Regional Climate Model Driven with HadCM2
Global Model Results for the Contemporary and
Future Scenario (2040s) Climate
Ten-Year Mean Monthly Stream Flow Generated by the
RegCM2 Regional Climate Model Driven
with HadCM2 Global Model Results for the
Contemporary and Future Scenario (2040s) Climate
Errors in Simulated Stream Flow
and Climate Change
Comparisons
Evaluate
SWAT 1 vs. Measured
SWAT error
SWAT 2 vs. SWAT 1
RCM error
SWAT 3 vs. SWAT 2
GCM error
SWAT 3 vs. SWAT 1
GCM-RCM error
SWAT 2 vs. Measured
RCM-SWAT error
SWAT 3 vs. Measured
GCM-RCM-SWAT error
SWAT 4 vs. SWAT 3
Climate change
Comparison of Simulated Stream Flow under
Climate Change with Various Model Biases
Hydrologic Budget Components
Simulated by SWAT under Different Climates
Hydrologic budget
components
Calibration
(19891997)
Validation
(19801988)
NNR
(19801988)
CTL
(around
1990s)
SNR
(around
2040s)
% Change (SNR-CTL)
Precipitation
856
846
831
898
1082
21
Snowfall
169
103
237
249
294
18
Snowmelt
168
99
230
245
291
19
Surface runoff
151
128
151
178
268
51
GW recharge
154
160
134
179
255
43
Total water yield
273
257
253
321
481
50
Potential ET
947
977
799
787
778
-1
Actual ET
547
541
528
539
566
5
All units are mm
Yield is sum of surface runoff, lateral flow, and groundwater flow
Relation of Runoff to Precipitation
for Various Climates
Regression Analysis:
Stream Flow vs. Precipitation
Stream flow vs. precipitation
1. Measured stream flow vs. observed precipitation (1980-1997)
2. Simulated stream flow vs. observed precipitation (1980-1988)
3. Simulated stream flow vs. RCM/NNR precipitation (1980-1988)
4. Simulated stream flow vs. CTL precipitation (around 1990s)
5. Simulated stream flow vs. SNR precipitation (around 2040s)
Scenario
Observed
SWAT 1
SWAT 2
SWAT 3
SWAT 4
Slope
0.66
0.65
0.87
0.64
1.16
Summary



RCM provides meteorological detail needed by
SWAT to resolve sub-basin variability of
importance to streamflow
There is strong suggestion that climate change
introduces changes of magnitudes larger than
variation introduced by the modeling process
Relationship of streamflow to precipitation might
change in future scenario climates
Future Directions


Couple GCM, RCM, SWAT, Crop Model
and Economic Model
Evaluate policy alternatives:
Impact of introducing conservation practices
 Impact of introducing incentives


Hypothesis:
It is possible to balance profitability with
sustainability in an intensively managed
agricultural area under changing climate
through development of robust policy
GCM
OBS
NNR
RCM
Climate
Over UMRB
Crop
Model
Crop Yield
Soil
Drainage
Land-use
SWAT
Management
Choices
Economic
Model
OBS
Stream
flow
Soil
Carbon
Crop
Production
Water
Quality
Evaluate Sustainability
and Profitability
Incentives
Public
Policy