Page 1
Assessing the Water Cycle Assessing the Water Cycle in in
Regional Climate SimulationRegional Climate Simulation
W. J. Gutowski, Jr.W. J. Gutowski, Jr.Iowa State UniversityIowa State University
SASAS (September 2001)
with thanks to: with thanks to: F. OtienoF. Otieno, Z. Pan, R. W. Arritt, E. S. Takle, Z. Pan, R. W. Arritt, E. S. Takle
Page 2
SASAS (September 2001)
OutlineOutline
10-yr RCM simulations10-yr RCM simulations
Precipitation analysisPrecipitation analysis
Error evaluationError evaluation
Conclusions: Sources of error?Conclusions: Sources of error?
Page 3
Simulations
Model Observed GCM-control GCM-Scenario
RegCM2 NCEPReanalysis(1979-1988)
HadleyCentre(~1990’s)
HadleyCentre(2040-2050)
HIRHAM(DMI)
“ “ “
Page 4
Simulations
Model Observed GCM-control GCM-Scenario
RegCM2 NCEPReanalysis(1979-1988)
HadleyCentre(~1990’s)
HadleyCentre(2040-2050)
HIRHAM(DMI)
“ “ “
Page 5
Simulation DomainSimulation Domain
Page 6
PrecipitationPrecipitation
– VEMAP - monthly; 0.5˚ x 0.5˚VEMAP - monthly; 0.5˚ x 0.5˚
– Higgins et al. - hourly; 2˚ x 2.5˚Higgins et al. - hourly; 2˚ x 2.5˚
Circulation – NCEP/NCAR reanalysisCirculation – NCEP/NCAR reanalysis
Precipitable water – NCEP/NCAR reanalysisPrecipitable water – NCEP/NCAR reanalysis
Runoff - UNH 0.5˚ climatologyRunoff - UNH 0.5˚ climatology
Observations
Page 7
SASAS (September 2001)
OutlineOutline
10-yr RCM simulations10-yr RCM simulations
Precipitation analysisPrecipitation analysis
Error evaluationError evaluation
Conclusions: Sources of error?Conclusions: Sources of error?
Page 8
Bias: average differenceBias: average difference
Bias Score: intensity “spectrum”Bias Score: intensity “spectrum”
SOM: SOM: pattern differencepattern difference
Diagnoses
Page 9
Precip.Bias by Month & Location
-2 0 2 4 [mm/d]
Page 10
-40
0
1979 1981 1983 1985 1987Year
WINTER SPRING
SUMMER FALL
Season’s Bias by Year~ south-central US ~
(“Fall” = Sep-Oct-Nov, etc.)
Page 11
(“Score” ~ relative exceeding of threshold)
Bias Score by LocationThreshold = 1 mm/d
Page 12
(“Score” ~ relative exceeding of threshold)
Bias Score by LocationThreshold = 2 mm/d
Page 13
(“Score” ~ relative exceeding of threshold)
Bias Score by LocationThreshold = 4 mm/d
Page 14
(“Score” ~ relative exceeding of threshold)
Bias Score by Month~ south-central US ~
Page 15
Self-Organizing Maps
Set of maps
• Show characteristic data structures
• Trained to distribution of data
• Give 2-D projection of higher order map space
• Are approximately continuous
Page 16
SOM: RegCM2 & VEMAP Precipitation
Page 17
1
2
SOM: RegCM2 & VEMAP Precipitation
Page 18
1
2
SOM: Major Dimensions
0 [mm/mo] 100 200
“warm”
“cold”
Page 19
SOM: RegCM2 & VEMAP Precipitation
Page 20
SOM: RegCM2 & VEMAP Precipitation
Page 21
SOM Trajectories
RegCM
VEMAP
J-J-A
Page 22
SOM Trajectories
RegCM
D-J-F
VEMAP
D-J-F
RegCM
VEMAP
J-J-A
Page 23
2
SOM: Major Dimensions
0 [mm/mo] 100 200
“cold”
Page 24
Trajectory Separation by Month
Page 25
SASAS (September 2001)
OutlineOutline
10-yr RCM simulations10-yr RCM simulations
Precipitation analysisPrecipitation analysis
Error evaluationError evaluation
Conclusions: Sources of error?Conclusions: Sources of error?
Page 26
500 hPa Heights & BiasSep-Oct-Nov
[m]
Page 27
[m2]
500 hPa Bandpass Variance & Bias
Sep-Oct-Nov
Page 28
[kg-m-2]
Precipitable Water & Bias
Sep-Oct-Nov
Page 29
Simulation DomainSimulation Domain
Page 30
Sep-Oct-Nov 1984
-25
0
25
50
1 31 61 91Time (days)
RegCM2Higgins
Page 31
Sep-Oct-Nov 1987
-25
0
25
50
1 31 61 91Time [days]
RegCM2
Higgins
Page 32
Water Balance AnalysisWater Balance Analysis
ΔE =ΔP −(P −E)m +Ro
Observed runoff
Model output
Precipitation error
Evapotranspiration error
From terrestrial balance:
Page 33
Water Balance AnalysisWater Balance Analysis
ΔE=ΔP −(P −E)m +Ro +dWdt o
Observed runoff
Model output
Precipitation error
Evapotranspiration error
From terrestrial balance:
Subsurface storage
Page 34
Water Balance AnalysisWater Balance Analysis
ΔC=ΔP −ΔE
Precipitation error
Evapotranspiration error
From atmospheric balance:
Vapor convergence error
Page 35
Error: Ten-year averageError: Ten-year average
-1.5
-1
-0.5
0
B C DΔ ΔΔP E C
Page 36
Water Balance AnalysisWater Balance Analysis
ΔE=ΔP −(P −E)m +Ro +dWdt o
For Sep-Oct-Nov:
Page 37
Water Balance AnalysisWater Balance Analysis
ΔE=ΔP −(P −E)m +Ro +dWdt o
For Sep-Oct-Nov:
Plausible values?• model’s root-zone storage: => -0.5 mm/d• (P-E)m - Ro: => -0.1 mm/d
Page 38
Error: S-O-N averageError: S-O-N average
-3
-2
-1
0
1
-1 -0.5 0 0.5 1
P E C
Storage [mm/d]
Δ ΔΔP E C
Page 39
SASAS (September 2001)
OutlineOutline
10-yr RCM simulations10-yr RCM simulations
Precipitation analysisPrecipitation analysis
Error evaluationError evaluation
Conclusions: Sources of error?Conclusions: Sources of error?
Page 40
SASAS (September 2001)
Evapotranspiration error from:Evapotranspiration error from:
Temperature error?Temperature error?
Page 41
Daily Max. TemperatureRegCM2 Difference VEMAP
-12.5 - 2.5 2.5 12.5 22.5 [oC]
Page 42
Daily Min. TemperatureRegCM2 Difference VEMAP
-12.5 - 2.5 2.5 12.5 22.5 [oC]
Page 43
SASAS (September 2001)
Evapotranspiration error from:Evapotranspiration error from:
Temperature error? Apparently no.Temperature error? Apparently no.
Surface moisture error?Surface moisture error?
Page 44
SASAS (September 2001)
The region has substantial wetlands:The region has substantial wetlands:
Page 45
SASAS (September 2001)
Not resolved by Not resolved by
60 km grid.60 km grid.
Large source of Large source of
water?water?
The region has substantial wetlands:The region has substantial wetlands:
Page 46
AcknowledgmentsAcknowledgments
Primary Funding: Primary Funding: U.S. National Oceanic and Atmospheric U.S. National Oceanic and Atmospheric
Administration (NOAA)Administration (NOAA)
Additional Support: Additional Support: Electric Power Research Institute (EPRI)Electric Power Research Institute (EPRI)
SASAS (September 2001)
Page 48
RegCM2 Bias VEMAP
JAN
JUL
0-2 +2 [mm/d]+4 +6- 4
Page 53
SOM TrajectoriesRegCM
VEMAP
J-J-A
Page 54
SOM Trajectories
RegCM
D-J-F
VEMAP
D-J-F
RegCM
VEMAP
J-J-A
Page 55
500 hPa Bandpass Variance &
Bias
(a) Sep-Oct-Nov
(b) Dec-Jan-Feb
[m2]
Page 56
[m2]
500 hPa Bandpass Variance & Bias
Sep-Oct-Nov
Page 57
[m]
500 hPa SOM-weighted ave. & bias
October
Page 58
SASAS (September 2001)
The region has substantial wetlands:The region has substantial wetlands:
Page 59
SASAS (September 2001)
……and substantial irrigation:and substantial irrigation:
Page 60
SASAS (September 2001)
Hydrological units boundaries:Hydrological units boundaries:
Page 61
Hydrological units boundaries:Hydrological units boundaries: