21st Century Water Management:
The Myth of Climate Stationarity and
Strategies for Water Resources
Management in a Rapidly Evolving
Climate
Alan F. Hamlet
•JISAO/CSES Climate Impacts Group
•Dept. of Civil and Environmental Engineering
University of Washington
Climatological Foundation of U.S. Water
Resources Planning and Management:
1) Risks are stationary in time.
2) Observed streamflow records are the best estimate of
future variability.
3) Systems and operational paradigms that are robust to
past variability are robust to future variability.
Schematic of a Typical
Water Planning Framework
Observed Streamflows
Planning Models
System Drivers
Annual streamflow reconstructions at The Dalles, OR
using tree ring growth indices derived from douglas-fir
and limber pine from SE British Columbia - Kamloops to
Banff/Jasper (1750-1964)
350000
Columbia Basin
Planning Window
300000
Observed Annual
Streamflow
Observed 5 yr mean
200000
Reconstructed Annual
Streamflow
Linear (Reconstructed
Annual Streamflow)
150000
100000
y = -22.831x + 214682
50000
1975
1950
1925
1900
1875
1850
1825
1800
1775
0
1750
Flow (cfs)
250000
Trends in Annual Streamflow at The Dalles from 1858-1998 are strongly downward.
350000
250000
Annual
200000
5 yr mean
10 yr mean
150000
Linear (Annual)
100000
50000
0
1858
1868
1878
1888
1898
1908
1918
1928
1938
1948
1958
1968
1978
1988
1998
Annual Mean Flow (cfs)
300000
16.4 MAF was considered a conservative estimate at the time of
the Compact. However, the average annual flow over the 20th
century has been only 15 MAF.
Relative to the gage record today, flows in the early 20th century appear to
be unusually high. How unusual is this period in a longer-term context?
(Figure Courtesy Connie Woodhouse)
Tree rings placed the gage record in a long-term
context
Colorado
River flow,
reconstructed
by Stockton
and Jacoby,
1976
Stockton and Jacoby 1976
“…the timing of the drafting of the Compact was
an unfortunate event, in that it did not occur during
a representative flow period.”
“The general picture of a collision between water
demand and supply in the UCRB in the not-toodistant future is all too apparent.”
Stockton and Jacoby 1976
(Figure Courtesy Connie Woodhouse)
Effects of ENSO on Cool Season Climate in the PNW
Natural Flow Columbia River at The Dalles, OR
450000
350000
300000
250000
200000
Red = warm ENSO, Green = ENSO neutral, Blue = cool ENSO
2000
1990
1980
1970
1960
1950
1940
1930
1920
1910
150000
1900
Apr-Sept Flow (cfs)
400000
The Myth of Stationarity Meets the
Death of Stationarity
Muir Glacier in Alaska
Aug, 13, 1941
Aug, 31, 2004
Image Credit: National Snow and Ice Data Center, W. O. Field, B. F. Molnia
http://nsidc.org/data/glacier_photo/special_high_res.html
Changes in Simulated April 1
Snowpack for the Canadian
and U.S. portions of the
Columbia River basin
(% change relative to current climate)
20th Century Climate
“2040s” (+1.7 C)
-3.6%
-21.4%
April 1 SWE (mm)
“2060s” (+ 2.25 C)
-11.5%
-34.8%
Simulated Changes in Natural Runoff Timing in the Naches
River Basin Associated with 2 C Warming
120
Simulated Basin Avg Runoff (mm)
100
80
Impacts:
•Increased winter flow
•Earlier and reduced peak flows
•Reduced summer flow volume
•Reduced late summer low flow
1950
60
plus2c
40
20
0
oct
nov
dec
jan
feb
mar
apr
may
jun
jul
aug
sep
2000
1996
1992
1988
1984
1980
1976
1972
1968
1964
1960
1956
1952
1948
1944
1940
CRB
1936
CA
1932
1928
3
1924
1920
1916
Std Anomalies Relative to 1961-1990
Regionally Averaged Cool Season Precipitation Anomalies
4
PNW
PRECIP
GB
2
1
0
-1
-2
-3
Annual area (ha × 106 ) affected by MPB in BC
9.0
2005
Bark Beetle Outbreak in British Columbia
8.0
2004
7.0
6.0
5.0
2003
4.0
3.0
2.0
2002
1.0
2001
2000
1999
0
1910 1930 1950 1970 1990 2010
Year
(Figure courtesy Allen Carroll)
Impact Pathways Associated with Climate
•Changes in water quantity and timing
Reductions in summer flow and water supply
Increases in drought frequency and severity
Changes in hydrologic extremes
Changing flood risk (up or down)
Summer low flows
Changes in groundwater supplies
•Changes in water quality
Increasing water temperature
Changes in sediment loading (up or down)
Changes in nutrient loadings (up or down)
•Changes in land cover via disturbance
Forest fire
Insects
Disease
Invasive species
Impact Pathways Associated with Climate
•Changes in water management practice
Hydropower production (energy demand)
Flood control operations (changing flood risk and refill
statistics)
Instream flow augmentation
Use of storage to control water temperature
•Changes in Ecosystem Protection and Recovery
Planning
Design of fish and wildlife recovery plans
Habitat restoration efforts
ESA listings (as a process)
Monitoring programs
Rebalancing Water Systems in
Response to Climate Change
Sep
Aug
Jul
Jun
May
Apr
30000
Mar
Feb
Jan
D ec
N ov
8000
7000
6000
5000
4000
3000
2000
1000
0
O ct
S to ra g e
O ct
Nov
Dec
Jan
Feb
M ar
Apr
M ay
Jun
Jul
Aug
Sep
R e s e rv o ir In flo w
Flood Control vs. Refill
Full
25000
20000
15000
: Current Climate
10000
Flood Control vs. Refill
Streamflow timing shifts can reduce the reliability of reservoir refill
8000
6000
+ 2.25 oC
5000
4000
3000
30000
Full
2000
1000
0
25000
20000
15000
: Current Climate
Sep
A ug
Ju l
Ju n
May
A pr
Mar
Feb
Ja n
D ec
10000
N ov
: + 2.25 oC No adaptation
O ct
S to ra g e
O ct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
R e s e rv o ir In flo w
7000
Flood Control vs. Refill
Streamflow timing shifts can reduce the reliability of reservoir refill
8000
7000
+ 2.25 oC
5000
4000
30000
3000
Full
2000
1000
: Current Climate
S to ra g e
Aug
25000
Sep
Jul
Jun
Apr
May
Mar
Jan
Feb
Dec
Oct
0
Nov
20000
: + 2.25 oC No adaption
15000
Sep
Aug
Jul
Jun
May
Apr
Mar
Feb
Jan
D ec
10000
N ov
: + 2.25 oC plus adaption
Oct
Reservoir Inflow
6000
Major U.S. Flood Control Checkpoints
Bonners Ferry
Columbia Falls
The Dalles
Optimization Reduces Storage Deficits without Jeopardizing Flood Protection
2,000
Storage Deficits
1,500
optimized
1,000
500
Duncan
Libby
Mica
Hungry
Horse
Dworshak
Climate change/no adaptation
600
Flood Risks
Flow at The Dalles (kcfs)
Grand
Coulee
0
Arrow
July 31 Average Storage Deficit (KAF)
Climate change/no adaptation
500
400
300
optimized
200
100
-2
-1
0
1
2
3
Extreme Value Type I Distribution Reduced Variate, Y
4
5
Adaptation Strategies
Approaches to Adaptation and Planning
•Anticipate changes. Accept that the future climate will be
substantially different than the past.
•Use scenario based planning to evaluate options rather
than the historic record.
•Expect surprises and plan for flexibility and robustness in
the face of uncertain changes rather than counting on one
approach.
•Plan for the long haul. Where possible, make adaptive
responses and agreements “self tending” to avoid repetitive
costs of intervention as impacts increase over time.
Schematic of Climate Change
Water Planning Framework
Observed Streamflows
Planning Models
Altered Streamflows
Climate Change Scenarios
System Drivers
Example of a Self-Tending Reservoir Operating System
•Improved Streamflow Forecasts Incorporating Warming and
Other Features of Altered Climate System
•Dynamic Reservoir Operating Systems Using Optimization or
Hybrid Optimization/Simulation Approaches to Rebalance the
System.
•Such systems are more flexible because they do not require a “trigger”
for a change in the operating policies, and arguably do not require as
much intervention as the climate system gradually (or not so gradually)
changes, because the system responds autonomously to improvements in
forecasts.
•These ideas are not really new:
Harvard Water Program ~1965