Chapter 6
Assessment of Climate/Change Impacts
Chap,6 Climate/Change Impact
methods
? Index,variables
? Statistical methods
? GCMs with downscaling
Chapter 6
Assessment of Climate/Change Impacts
6.1 Index
1,Variables for element characteristics
Precipitation regime
Temperature
Humidity/evaporation
Wind
Hazards,drought,flood,frost,etc,
2,Variables for element changes
Difference
Ratio
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
1,The value of statistics
Can not provide definitive answers,but can be of great help
in establishing relationships and quantifying uncertainties
within climatic data
2,Some typical problems
? Mean state
? Correlation,linear,auto-,cross-(spatial)
? Stationary/cyclo-stationary/non-stationary
? Quality of forecasts
? Time and spatial characteristics
? Pairs of characteristic patterns
? Model evaluation
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
3,Mathematical models
? Certain relationships
? Random relationships
? Stationary– stable mean
? Non-stationary– varying mean
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
4,Main aspects for analysis
? Probability and distributions
? Confirmation/test,validation
? Fitting statistical models
? Time series analysis
scale,length,time step,duration
stationary/non-stationary
? Spatial analysis
? Specific topics
EOF/PC/CCA/SVD/POP
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis methods,
(1) Graphical depiction,trends/cycles
Different functions,
linear
exp
ln
inverse
poly-
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis methods
(2) Autocorrelation,temporal autocorrelation
? e.g,today’s wind speed is similar to yesterday’s wind speed
this is the persistence or inertia of the climate system
? Diurnal/annual cycles should be removed before estimating
autocorrelation
? X-average(X)/ fitting and removing an analytical function,
such as a series of sinusoids or polynomials
? The most natural way to visualize autocorrelation in a time
series is by plotting the autocorrelation as a function of lag
time,that is the autocorrelation function
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis models
Component decomposition
(1) A linear trend
(2) an annual cycle
(3) a diurnal cycle
(4) autocorrelation
(5) a random component
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis models
Spectrum analysis
(1) Discrete Fourier Transform
(2) The power spectrum
(3) Cross spectrum analysis
(4) Filtering
(5) Wavelets
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? spatial analysis
Grid
Interpolation
Relationships between related fields
Geostatistics,Kriging
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? Some problems
correlation coefficient– linear,independent
relationship functions– independent variables
samples size
test
Chapter 6
Assessment of Climate/Change Impacts
GCMs-General Circulation Models
? Purpose
Modeling general circulation
Projecting climate change for different scenarios
Data supplied for assessment of climate change
impacts
Linkage of assessment between climate change and
its impact
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Necessary
?Output of GCMs
Spatial resolution,>200km
Temporary resolution,monthly
?Needs of assessment
Spatial resolution,1-200km
Temporary resolution,daily or hourly
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Objectives
? Establishment of relationship between
local variables and output variables of
GCMs
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Assumptions
? Best prediction from GCMs
? Best relationship between predictors
and predictands
? Stable relationship between predictors
and predictands
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Methods
? Statistical,monthly
? Stochastic,daily
? Dynamic,daily
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,
year
Titles
Methods
Predictors
predictands
Better
predictors
Tereza
Cavazos and
Bruce
Hewitson,
Relative Performance of
Empirical Predictors of
Daily Precipitation
RPCA to get top 10 PCs
ANNs for downscaling
29 variables
Best,
For midlatitude,
Winter (W),P = ? (z5,q7,slp);
P = ? (z7,q7,slp)
Summer (S),P = ? (z7,q7,u8);
P = ? (z7,q7,v8)
For Tropical,
Winter (W),P = ? (th1,q7,z7,
v0)
Summer (S),P = ? (z7,q7,th1,
u8)
Daily rainfall
Dynamic1
Nested RCM
Hydrolocgical cycle
Statistic1
R,Cano,A,
S,Cofino,
M.A,
Rodruez and
J.M,
Gutierrez
Self-Organizing Maps for
Statistical Downscaling
in Short-Range Weather
Forecast
PCA,
Self-organized map
daily Temperature (T),relative
Humidity (H),Geopotential (Z)
and U,V wind components at
six pressure levels (300,500,
700,850,
925,and 1000 mb) at 00,06,
12,18,and 24 UTC
daily precipitation (Pp),
maximum wind speed
(Rx),and insolation (In)
Statistic2
D.J,Sailor,
T,Hu,X,Li,
J.N,Rosen,
Renewable
Energy 19
(2000) 359-
378
A neural network approach
to local downscaling of
GCM output for assessing
wind power implications of
climate change
ANNs for downscaling
upper level winds,temperature,
sea level pressure
Daily mean wind speed
Power,cut-in/shut-down
threshold
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Statistic3
J,Olsson,C,B,Uvo
and K,Jinno
Phys,Chem,Earth (B),
Vol,26,No,9,pp,695-
700,200l
Statistical Atmospheric
Downscaling of Short-
Term Extreme Rainfall By
Neural
Networks
correlation
coefficients (cc)
Neural networks
Precipitable water,u,v at
850hpa
Mean 12-hour rainfall
Statistic4
Reiner Schnur,
Dennis P,
Lettenmaier
Journal of Hydrology
212–213 (1998) 362–
379
A case study of statistical
downscaling in Australia
using weather classification
by recursive partitioning
Classification tree
analysis
SLP-circulation pattern,
EOF time coefficients of
SLP anomalies
Daily rainfall
occurrence (wet/dry)
Daily rainfall amount
Distribution of storm
interarrival time
Statistic5
Heinz-thco
Mengelkamp,
Hartmut Kapitza,
Ulrich Pfluger,
J,of wind
engineering and
industrial
aerodynamics 67&68
(1997) 449-457
Statistical dynamical
downscaling of wind
climatologies
Classification,
clusters
Dynamics,non-
hydrostatic
Mesoscale model
Geostrophic wind
Vertical temperature
gradient
Radiosonde data at
850hpa(wind)
temperature gradient (100m-
1.5km)
Mean wind speed field
and wind speed
distribution and
direction
No time step
Statistic6
D,Conway,P.D,
Jones
Journal of Hydrology
212–213 (1998) 348–
361
The use of weather types
and air flow indices for
GCM downscaling
Weather type by
using 3
indices(lamb)
probabilities
between 3 indices
and rainfall
weather generation
3 indices of weather types
Daily rainfall
characteristics,
Probability and amount
Statistic7
R.L,Wilby,H,
Hassan,K,Hanaki
Journal of hydrology
205(1998)1-19
Statistical downscaling of
hydrometeorological
variables using general
circulation model output
Regression transfer
functions for
different elements
3 indices calculated from
daily mean SLP
Daily data for 7
elements
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Statistic8
Jiri Stehlik,Andras
Bardossy,Journal of
hydrology
256(2002)120-141
Multivariate stochastic
downscaling model for
generating daily precipitation
series based on atmospheric
circulation
Classification by fuzzy
rules
Geopotential height in
500 and 850hap
Daily rainfall
characteristics
Statistic9
Burkhard Oelschlagel,
Ecological modeling
82(1995)199-204
A method for downscaling
global climate model
calculations by a statistical
weather generator
Spatial interpolation
Large scale t,p,global
radiation,
Monthly t,p,global
radiation for W GEN
parameters
Then generate daily
values
Statistic10
Andras Bardossy
Journal of
Environmental
Management (1997)
49,7–17
Downscaling from GCMs to
Local Climate through
Stochastic Linkages
Fuzzy rule classification
Stochastic transfer fun,
Daily mean pressure to
estimate circulation
patterns
Daily p by using
stochastic model
Statistic11
R.L,Wilby,L.E,Hay,
G.H,Leavesley
Journal of Hydrology 225
(1999) 67–91
A comparison of downscaled
and raw GCM output,
implications
for climate change scenarios
in the San Juan River basin,
Colorado
Stepwise multiple
regression,3/15
Mean SLP,h in 500hpa,
2mT and 0.995sigma
RH for SH(specific
humidity)
Daily tmax,tmin,p
Statistic12
A.Bardossy,J,
Stehlik and H.J,
Caspary
Phys,Chem,Earth
(B),Vol,26,No,9,
pp,683487,2001
Generating of Area1
Precipitation Series in the
Upper Neckar Catchment
fuzzy classification
Stochastic transfer fun,
Fourier series of annual
cycle
Spatial interpolation,
external-drift-kriging
Areal precipitation
Statistic13
András Bárdossy
Keynote Paper 1 of a
workshop
Stochastic Downscaling
Methods to Assess the
Hydrological
Impacts of Climate Change
on River Basin Hydrology
Fuzzy rule classification
Stochastic transfer fun,
Daily H500,H750,SLP
Daily p and t
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Sta-dyn1
W.J,Gutowski,Jr.,R,
Wilby,L,E,Hay,C.J,
Anderson,R.W,
Arritt,
M,P,Clark,G,H,
Leavesley,Z,Pan,R,
Silva,E,S,Takle
Statistical versus Dynamical
Downscaling for Hydrologic
Analysis
Stepwise multiple
regression,3/15
RCM
Different regressions
for different seasons
Daily tmax,tmin,p
Statistic15
E.J,F?rland,R.E,
Benestad,I,Hanssen-
Bauer,K.A,Iden &
O.E,Tveito
Local climate scenarios for
Norway based on empirical
downscaling
CCA,EOF
SLP,2mT,SST,H500,
T500,sea ice extension
Monthly T,P
Copy
David J,Sailor and
Xiangshang Li
Journal of climate
12(1999)103-114
A semiempirical
downscaling approach for
predicting regional
temperature impacts
associated with climatic
change
Stepwise multiple
regression
Four seasons separately
21 predictors
Monthly t
Copy
James Murphy
Journal of climate
12(1999)2256-2284
An evaluation of statistical
and dynamical techniques for
downscaling local climate
EOF,PC,multiple
regressions
RCM for daily
8 predictors derived
from large-scale
variables in different
levels
Monthly t,p,daily t,p
Copy
John W,Kidson and
Craig S,Thompson
Journal of
climate11(1998)735-
735
A comparison of statistical
and model-based
downscaling techniques for
estimating local climate
variations
EOF,screening
regressions
Data transform for daily
analysis by using Fourier
RCM
5 predictors chosen
form large-scale
variables in different
levels by screening
regression
Daily and monthly t,p
Copy
R.E,Benestad
Climate
research21(2002)105
-125
Empirically downscaled
temperature scenarios for
northern Europe based on a
multi-model ensemble
EOF,multiple regression
GCM by interpolation
2mT
Scenario T
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Copy
Radan Huth
Climate research 13
(1999) 91-101
Statistical downscaling in
central Europe,evaluation of
methods and potential
predictors
PCA,CCA,SVD,MLR
Normalized anomalies
z500,SLP,T850,
TH1000-500
Daily mean
temperature for winter
DJF
Copy
Clare M,Goodess
and Jean P,Palutikof
International journal
of climatology
10(1998)1051-1083
Development of daily rainfall
scenarios for southeast Spain
using a circulation-type
approach to downscaling
Circulation types
Relation between
circulation types and
daily rainfall
Weather generator
SLP
daily precipitation
Copy
Deliang Chen and
Youmin Chen
Development and
verification of a multiple
regression downscaling
model for monthly
temperature in Sweden
EOFs
Multiple regressions
Station and month
separately
Circulation indices,
geostrophic wind and
total vorticity derived
from SLP at 16 grid
points
22 predictors derived
from EOFs
Monthly temperature
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Statistical methods
? Step 1,Selection of predictands– local
variables
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
? Step 3,Establishment of the relationship
between local variables and output
variables from GCMs
? Step 4,validation of the relationship
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
Location,Ansai in the Loess
Plateau
Objectives,Assessment of climate
change on crop productivities
? Step 1,Selection of predictands– local
variables
(1) monthly temperature,1958.1-1998.12
(2) Monthly rainfall,1958.1-1998.12
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(1)Types of GCMs
HadCM2/HadCM3,from England
CCMA (GCMI/GCMII),from Canada
GFDL/NCARI/DOE-PCM,from USA
CSIRO,from Australia
CCSR/NIES,from Japan
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(2)General Circulation Background
Sea level,
700hPa/500hPa,
Output variables,for T– p,t,h,
for P- p,t,q,h
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(3)Spatial scale
Data format,GRIB(binary),GZIP(ASCII)
Scale,Lat.—25-45o
Lon.—95-115o
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(3) Normalization of data
(Xi-average)/average
(4) Selection of predictors
EOF analysis
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 3,Establishment of the relationship
between predictands(local variables) and
predictors(output variables from GCMs)
Step-wise regression
Multiple regression
Y(t)=a*slp + b*t2m + c*h500 + d*t500
? Step 4,validation of the relationship
Cross-validation
Chapter 6
Assessment of Climate/Change Impacts
100 102 104 106 108 110 112 114
32
34
36
38
40
42
44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30 31
32
33
34
35
36
37
38
39
40
Chapter 6
Assessment of Climate/Change Impacts
Thanks!
Assessment of Climate/Change Impacts
Chap,6 Climate/Change Impact
methods
? Index,variables
? Statistical methods
? GCMs with downscaling
Chapter 6
Assessment of Climate/Change Impacts
6.1 Index
1,Variables for element characteristics
Precipitation regime
Temperature
Humidity/evaporation
Wind
Hazards,drought,flood,frost,etc,
2,Variables for element changes
Difference
Ratio
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
1,The value of statistics
Can not provide definitive answers,but can be of great help
in establishing relationships and quantifying uncertainties
within climatic data
2,Some typical problems
? Mean state
? Correlation,linear,auto-,cross-(spatial)
? Stationary/cyclo-stationary/non-stationary
? Quality of forecasts
? Time and spatial characteristics
? Pairs of characteristic patterns
? Model evaluation
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
3,Mathematical models
? Certain relationships
? Random relationships
? Stationary– stable mean
? Non-stationary– varying mean
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
4,Main aspects for analysis
? Probability and distributions
? Confirmation/test,validation
? Fitting statistical models
? Time series analysis
scale,length,time step,duration
stationary/non-stationary
? Spatial analysis
? Specific topics
EOF/PC/CCA/SVD/POP
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis methods,
(1) Graphical depiction,trends/cycles
Different functions,
linear
exp
ln
inverse
poly-
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis methods
(2) Autocorrelation,temporal autocorrelation
? e.g,today’s wind speed is similar to yesterday’s wind speed
this is the persistence or inertia of the climate system
? Diurnal/annual cycles should be removed before estimating
autocorrelation
? X-average(X)/ fitting and removing an analytical function,
such as a series of sinusoids or polynomials
? The most natural way to visualize autocorrelation in a time
series is by plotting the autocorrelation as a function of lag
time,that is the autocorrelation function
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis models
Component decomposition
(1) A linear trend
(2) an annual cycle
(3) a diurnal cycle
(4) autocorrelation
(5) a random component
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? time series analysis models
Spectrum analysis
(1) Discrete Fourier Transform
(2) The power spectrum
(3) Cross spectrum analysis
(4) Filtering
(5) Wavelets
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? spatial analysis
Grid
Interpolation
Relationships between related fields
Geostatistics,Kriging
Chapter 6
Assessment of Climate/Change Impacts
6.2 Statistical Methods
? Some problems
correlation coefficient– linear,independent
relationship functions– independent variables
samples size
test
Chapter 6
Assessment of Climate/Change Impacts
GCMs-General Circulation Models
? Purpose
Modeling general circulation
Projecting climate change for different scenarios
Data supplied for assessment of climate change
impacts
Linkage of assessment between climate change and
its impact
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Necessary
?Output of GCMs
Spatial resolution,>200km
Temporary resolution,monthly
?Needs of assessment
Spatial resolution,1-200km
Temporary resolution,daily or hourly
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Objectives
? Establishment of relationship between
local variables and output variables of
GCMs
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Assumptions
? Best prediction from GCMs
? Best relationship between predictors
and predictands
? Stable relationship between predictors
and predictands
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Methods
? Statistical,monthly
? Stochastic,daily
? Dynamic,daily
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,
year
Titles
Methods
Predictors
predictands
Better
predictors
Tereza
Cavazos and
Bruce
Hewitson,
Relative Performance of
Empirical Predictors of
Daily Precipitation
RPCA to get top 10 PCs
ANNs for downscaling
29 variables
Best,
For midlatitude,
Winter (W),P = ? (z5,q7,slp);
P = ? (z7,q7,slp)
Summer (S),P = ? (z7,q7,u8);
P = ? (z7,q7,v8)
For Tropical,
Winter (W),P = ? (th1,q7,z7,
v0)
Summer (S),P = ? (z7,q7,th1,
u8)
Daily rainfall
Dynamic1
Nested RCM
Hydrolocgical cycle
Statistic1
R,Cano,A,
S,Cofino,
M.A,
Rodruez and
J.M,
Gutierrez
Self-Organizing Maps for
Statistical Downscaling
in Short-Range Weather
Forecast
PCA,
Self-organized map
daily Temperature (T),relative
Humidity (H),Geopotential (Z)
and U,V wind components at
six pressure levels (300,500,
700,850,
925,and 1000 mb) at 00,06,
12,18,and 24 UTC
daily precipitation (Pp),
maximum wind speed
(Rx),and insolation (In)
Statistic2
D.J,Sailor,
T,Hu,X,Li,
J.N,Rosen,
Renewable
Energy 19
(2000) 359-
378
A neural network approach
to local downscaling of
GCM output for assessing
wind power implications of
climate change
ANNs for downscaling
upper level winds,temperature,
sea level pressure
Daily mean wind speed
Power,cut-in/shut-down
threshold
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Statistic3
J,Olsson,C,B,Uvo
and K,Jinno
Phys,Chem,Earth (B),
Vol,26,No,9,pp,695-
700,200l
Statistical Atmospheric
Downscaling of Short-
Term Extreme Rainfall By
Neural
Networks
correlation
coefficients (cc)
Neural networks
Precipitable water,u,v at
850hpa
Mean 12-hour rainfall
Statistic4
Reiner Schnur,
Dennis P,
Lettenmaier
Journal of Hydrology
212–213 (1998) 362–
379
A case study of statistical
downscaling in Australia
using weather classification
by recursive partitioning
Classification tree
analysis
SLP-circulation pattern,
EOF time coefficients of
SLP anomalies
Daily rainfall
occurrence (wet/dry)
Daily rainfall amount
Distribution of storm
interarrival time
Statistic5
Heinz-thco
Mengelkamp,
Hartmut Kapitza,
Ulrich Pfluger,
J,of wind
engineering and
industrial
aerodynamics 67&68
(1997) 449-457
Statistical dynamical
downscaling of wind
climatologies
Classification,
clusters
Dynamics,non-
hydrostatic
Mesoscale model
Geostrophic wind
Vertical temperature
gradient
Radiosonde data at
850hpa(wind)
temperature gradient (100m-
1.5km)
Mean wind speed field
and wind speed
distribution and
direction
No time step
Statistic6
D,Conway,P.D,
Jones
Journal of Hydrology
212–213 (1998) 348–
361
The use of weather types
and air flow indices for
GCM downscaling
Weather type by
using 3
indices(lamb)
probabilities
between 3 indices
and rainfall
weather generation
3 indices of weather types
Daily rainfall
characteristics,
Probability and amount
Statistic7
R.L,Wilby,H,
Hassan,K,Hanaki
Journal of hydrology
205(1998)1-19
Statistical downscaling of
hydrometeorological
variables using general
circulation model output
Regression transfer
functions for
different elements
3 indices calculated from
daily mean SLP
Daily data for 7
elements
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Statistic8
Jiri Stehlik,Andras
Bardossy,Journal of
hydrology
256(2002)120-141
Multivariate stochastic
downscaling model for
generating daily precipitation
series based on atmospheric
circulation
Classification by fuzzy
rules
Geopotential height in
500 and 850hap
Daily rainfall
characteristics
Statistic9
Burkhard Oelschlagel,
Ecological modeling
82(1995)199-204
A method for downscaling
global climate model
calculations by a statistical
weather generator
Spatial interpolation
Large scale t,p,global
radiation,
Monthly t,p,global
radiation for W GEN
parameters
Then generate daily
values
Statistic10
Andras Bardossy
Journal of
Environmental
Management (1997)
49,7–17
Downscaling from GCMs to
Local Climate through
Stochastic Linkages
Fuzzy rule classification
Stochastic transfer fun,
Daily mean pressure to
estimate circulation
patterns
Daily p by using
stochastic model
Statistic11
R.L,Wilby,L.E,Hay,
G.H,Leavesley
Journal of Hydrology 225
(1999) 67–91
A comparison of downscaled
and raw GCM output,
implications
for climate change scenarios
in the San Juan River basin,
Colorado
Stepwise multiple
regression,3/15
Mean SLP,h in 500hpa,
2mT and 0.995sigma
RH for SH(specific
humidity)
Daily tmax,tmin,p
Statistic12
A.Bardossy,J,
Stehlik and H.J,
Caspary
Phys,Chem,Earth
(B),Vol,26,No,9,
pp,683487,2001
Generating of Area1
Precipitation Series in the
Upper Neckar Catchment
fuzzy classification
Stochastic transfer fun,
Fourier series of annual
cycle
Spatial interpolation,
external-drift-kriging
Areal precipitation
Statistic13
András Bárdossy
Keynote Paper 1 of a
workshop
Stochastic Downscaling
Methods to Assess the
Hydrological
Impacts of Climate Change
on River Basin Hydrology
Fuzzy rule classification
Stochastic transfer fun,
Daily H500,H750,SLP
Daily p and t
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Sta-dyn1
W.J,Gutowski,Jr.,R,
Wilby,L,E,Hay,C.J,
Anderson,R.W,
Arritt,
M,P,Clark,G,H,
Leavesley,Z,Pan,R,
Silva,E,S,Takle
Statistical versus Dynamical
Downscaling for Hydrologic
Analysis
Stepwise multiple
regression,3/15
RCM
Different regressions
for different seasons
Daily tmax,tmin,p
Statistic15
E.J,F?rland,R.E,
Benestad,I,Hanssen-
Bauer,K.A,Iden &
O.E,Tveito
Local climate scenarios for
Norway based on empirical
downscaling
CCA,EOF
SLP,2mT,SST,H500,
T500,sea ice extension
Monthly T,P
Copy
David J,Sailor and
Xiangshang Li
Journal of climate
12(1999)103-114
A semiempirical
downscaling approach for
predicting regional
temperature impacts
associated with climatic
change
Stepwise multiple
regression
Four seasons separately
21 predictors
Monthly t
Copy
James Murphy
Journal of climate
12(1999)2256-2284
An evaluation of statistical
and dynamical techniques for
downscaling local climate
EOF,PC,multiple
regressions
RCM for daily
8 predictors derived
from large-scale
variables in different
levels
Monthly t,p,daily t,p
Copy
John W,Kidson and
Craig S,Thompson
Journal of
climate11(1998)735-
735
A comparison of statistical
and model-based
downscaling techniques for
estimating local climate
variations
EOF,screening
regressions
Data transform for daily
analysis by using Fourier
RCM
5 predictors chosen
form large-scale
variables in different
levels by screening
regression
Daily and monthly t,p
Copy
R.E,Benestad
Climate
research21(2002)105
-125
Empirically downscaled
temperature scenarios for
northern Europe based on a
multi-model ensemble
EOF,multiple regression
GCM by interpolation
2mT
Scenario T
Chapter 6
Assessment of Climate/Change Impacts
File name
Authors,year
Titles
Methods
Predictors
predictands
Copy
Radan Huth
Climate research 13
(1999) 91-101
Statistical downscaling in
central Europe,evaluation of
methods and potential
predictors
PCA,CCA,SVD,MLR
Normalized anomalies
z500,SLP,T850,
TH1000-500
Daily mean
temperature for winter
DJF
Copy
Clare M,Goodess
and Jean P,Palutikof
International journal
of climatology
10(1998)1051-1083
Development of daily rainfall
scenarios for southeast Spain
using a circulation-type
approach to downscaling
Circulation types
Relation between
circulation types and
daily rainfall
Weather generator
SLP
daily precipitation
Copy
Deliang Chen and
Youmin Chen
Development and
verification of a multiple
regression downscaling
model for monthly
temperature in Sweden
EOFs
Multiple regressions
Station and month
separately
Circulation indices,
geostrophic wind and
total vorticity derived
from SLP at 16 grid
points
22 predictors derived
from EOFs
Monthly temperature
Chapter 6
Assessment of Climate/Change Impacts
GCMs-Downscaling
-Statistical methods
? Step 1,Selection of predictands– local
variables
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
? Step 3,Establishment of the relationship
between local variables and output
variables from GCMs
? Step 4,validation of the relationship
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
Location,Ansai in the Loess
Plateau
Objectives,Assessment of climate
change on crop productivities
? Step 1,Selection of predictands– local
variables
(1) monthly temperature,1958.1-1998.12
(2) Monthly rainfall,1958.1-1998.12
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(1)Types of GCMs
HadCM2/HadCM3,from England
CCMA (GCMI/GCMII),from Canada
GFDL/NCARI/DOE-PCM,from USA
CSIRO,from Australia
CCSR/NIES,from Japan
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(2)General Circulation Background
Sea level,
700hPa/500hPa,
Output variables,for T– p,t,h,
for P- p,t,q,h
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(3)Spatial scale
Data format,GRIB(binary),GZIP(ASCII)
Scale,Lat.—25-45o
Lon.—95-115o
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 2,Selection of predictors-- output
variables from GCMs both in time and
spatial scales
(3) Normalization of data
(Xi-average)/average
(4) Selection of predictors
EOF analysis
Chapter 6
Assessment of Climate/Change Impacts
Statistical Downscaling
- An example
? Step 3,Establishment of the relationship
between predictands(local variables) and
predictors(output variables from GCMs)
Step-wise regression
Multiple regression
Y(t)=a*slp + b*t2m + c*h500 + d*t500
? Step 4,validation of the relationship
Cross-validation
Chapter 6
Assessment of Climate/Change Impacts
100 102 104 106 108 110 112 114
32
34
36
38
40
42
44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30 31
32
33
34
35
36
37
38
39
40
Chapter 6
Assessment of Climate/Change Impacts
Thanks!
