Transcript 2.3 Climate Scenarios

Chapter 2 Projection of Future
Climate Scenarios
Contents
2.1 General Circulation Model
2.2 Downscaling
2.3 Climate Scenarios
2.4 Uncertainty of Climate Scenarios
Purposes
• What are climate scenarios?
• Why do we need climate scenarios?
• How can we setup our climate scenarios?
– Current climate scenario?
– Future climate scenario?
What are climate scenarios?
• Climate scenarios represent possible weather statistics,
which may consist of
– monthly mean rainfall,
– Probability of wet day (rainy day)
– monthly mean temperature,
– Standard deviation of temperature for each month
…
Why?
• Those studies related to climate require weather data
as inputs.
• However, future weather data are not available,
unless you have a time machine.
• Thus, we need to project possible future climate
scenarios. Then, possible future weather can be
generated based on these scenarios.
Model
Daily Rainfall
Q=ϕ✕R
Q: Daily stream flow
R: Daily rainfall
ϕ : stream flow index
Daily Stream flow
How?
• Future climate scenarios can be derived based on the
outputs of GCMs and current climate scenarios.
• Past and current weather data are recorded by weather
stations, and thus current climate scenarios can be
determined.
• Current climate scenarios are modified based on the
climate projections of GCMs to form future climate
scenarios.
• General Circulation Models (GCMs) can produce
daily weather data, but the daily data are not used
directly. Instead, they are used to setup climate
scenarios.
• GCMs have better resolutions in recent years, but
there is still space for further improvement on the
ability of climate projection for a local area.
TO11
SiChou
#
KY05
KY06
#TouChen
TO09
KY03
KY04
TO10
TO08
TO07
KY02
KY01
ChuLin
#
TO05
TO04
#
RuiFeng
TO03
N
TO06
TO02
W
E
S
0
10
全球環流模式
GCM Projections
Risk Information for
Decision Making
降尺度分析
Downscaling
評估模式
Assessment Model
氣候情境
Climate Scenarios
流量增減模擬個數
The Risk Assessment Procedure
6
5
4
3
2
1
0
-1 短期 中期 長期
-2
-3
-4
-5
-6
氣候預設情境
氣象資料合成
Weather Generation
TO01
20 Km
9
How to Set up Climate Scenarios
• The most import step on an impact study is to setup
climate change scenarios. There are four methods to
define climate scenarios.
1. Based on GCMs’ projections
2. Assumptions
• T=+2 oC; +4 oC
• P= 0%, 10%, 20%
3. Spatial Analog
• Such as future climate in Taiwan may be similar to current climate
in Philippines.
4. Temporal Analog
• Assuming future climate may be repeated as climate in a specified
period in the past.
• Only the first type of scenario can reflect the physical
characteristics of enhanced greenhouse effects and
man-induced global warming .
• Different seasons may have different changes in
climate. Moreover, there is difference between night
and day time. Such difference can only be reasonably
provided by GCMs.
2.1 General Circulation Models
General Circulation Model
• A general circulation model (GCM) is a mathematical model
of the general circulation of a planetary atmosphere or ocean
and based on the Navier–Stokes equations on a rotating sphere
with thermodynamic terms for various energy sources
(radiation, latent heat). These equations are the basis for
complex computer programs commonly used for simulating
the atmosphere or ocean of the Earth. Atmospheric and
oceanic GCMs (AGCM and OGCM) are key components of
global climate models along with sea ice and land-surface
components. GCMs and global climate models are widely
applied for weather forecasting, understanding the climate, and
projecting climate change.
---From Wikipedia
Global Atmospheric Model
• Climate models are systems of
differential equations based on the
basic laws of physics, fluid motion,
and chemistry. To “run” a model,
scientists divide the planet into a 3dimensional grid, apply the basic
equations, and evaluate the results.
Atmospheric models calculate winds,
heat transfer, radiation, relative
humidity, and surface hydrology
within each grid and evaluate
interactions with neighboring points. From Wikipedia
How GCM works?
GCM Grids
GCM Resolutions
Model
Country
AGCM
OGCM
CCCM
Canada
T32L10
1.8o 1.8o L29
GFDL
USA
R30L14
2.0o2.0o L18
GISS
USA
4o5o L9
4o5o L13
UKMO
UK
2.5o3.8o L19
2.5o3.8o L20
Spatial Scale of Taiwan Main
Island
基隆25008 121.44
Taiwan：3o1.2o
梧棲
GISS ：4o5o
花蓮121.36
高雄 120.18
恆春 220
（台灣大學吳明進教授提供）
• GCMs provide projections for
each grid point, which represent
average values for a grid.
• The risk studies for water
resources or ecosystems often
require climate projections for a
smaller area and may need daily
weather data. Thus, as
mentioned, spatial and temporal
downscaling processes are
necessary.
2.2 Downscaling
Definition of Downscaling
• Downscaling is a technique to obtain information for
a finer scale from information for a larger scale.
• For example, temporal downscaling can provide daily
data based on monthly means. Weather generation is
kind of temporal downscaling.
• On the other hand, spatial downscaling can produce
data for a local area, such as a watershed, from data
for a regional area, such as an island or a state.
Known areal average to find local
characteristics
250 km
25 km
25 km
250 km
• GCM Scale
– 250 km250km
• Upstream Watershed
Scale
– 25 km25km
• Ecosystem
– 1 km1 km
Downscaling
How to find its climate?
How to Apply GCMs’ Outputs?
Grids that GCMs provide projections
Spatial Downscaling Methods
• Simple Downscaling (Delta Method)
– Climate changes of a local area are assumed the same as
the nearest grid point
• Modifying recorded weather data by imposing the
predicted climate changes of the nearest grid.
• Modifying historic weather statistics based on climate
change forecasts of the nearest grid.
• Statistical Downscaling
– Finding the statistical relationships between regional
climate and local climate.
• Physical Downscaling
– Taking GCMs’ forecasts as boundary conditions for a
regional climate model.
Method 1.1
• Modifying recorded weather data by imposing the
predicted climate changes
– It assumes there is uniform climate changes within a grid
and observed weather sequence is repeatable.
– The days in the same month are modified by the same
changes even though they are in different years.
Tt,m  Tt ,m  Tm
t  m & m  1 to 12
Pt,m  Pt ,m  RPm
t  m & m  1 to 12
1. T1ocal = Tregional
2. RPprecip-1ocal = RPprecip- regional
Method 1.2
• Modifying historic weather statistics based on
climate change projections and then generating
weather data.
– It also assumes there is uniform climate changes within a grid,
e.g. the change of a weather station is the same as the change
predicted by the nearest grid point.
– Monthly changes predicted by GCMs are used to modify
historical monthly weather statistics.
  Tm  Tm m  1 to 12
Tm
   Pm  RPm m  1 to 12
 Pm
– Modified monthly statistics are future climate scenarios and are
applied to generate future weather data.
Method 2
Statistical Downscaling
• First, finding the relationships between regional
climate pattern and local climate.
• Then, monthly changes predicted by GCMs are
projected to a local station based on the identified
relationships.
  Tm  Trans(Tm ) m  1 to 12
Tm
   Pm  Trans( RPm ) m  1 to 12
 Pm
Recorded Rainfall
FROM :21-OCT-2005 00:00
TO
:21-OCT-2005 08:30
Rank Rainfall (mm)
station
Code
Location
1
40.0
冬山
C1U68
宜蘭縣冬山鄉(冬山國中)
2
28.5
竹子湖
46693
台北市陽明山(氣象站)
3
27.0
竹子湖
01A42
台北市陽明山(十河局)
4
24.5
泰平
C0A55
台北縣雙溪鄉
5
24.0
寒溪
C1U67
宜蘭縣冬山鄉(大進國小)
6
23.5
玉蘭
C0U65
宜蘭縣大同鄉
7
23.5
太平
L1A84
台北縣雙溪鄉(翡翠水庫)
8
18.0
三星
C1U66
宜蘭縣三星鄉(三星鄉運動公園)
9
13.5
牛鬥
C1U50
宜蘭縣大同鄉
10
13.0
北投國小
A1A9V
台北市北投區(養工處)
Historical
Weather Data
Regional Climate
Regional Weather
Pattern
GCMs’ Predicted
Regional Weather
Pattern
Downscaling
Local Weather
Data
Relationship
Local Climate
Projected
Local Weather
Data
Similar Pattern, but different
Weather
10.21.2005 Rainy Day
10.28.2005 Sunny Day
Method 2 Bias Correction
• Assume the probability of observed data is the same as
the probability of GCM projections.
X GCM   GCM
 GCM

X OBS   OBS
 OBS
XGCM - mGCM
s GCM
Xunbias = (
CDF
=
Xunbias - mOBS
s OBS
XGCM - mGCM
s GCM
) ´ s OBS + mOBS
CDF
OBS
GCM
37
Example of Bias Correction
GFCM21_Jan
38
Bias Correction of GCM projections
Xunbias,m = (
XGCM ,m - mGCM ,m
Xunbias,F,m = (
s GCM ,m
) ´ s observed,m + mobserved,m
XGCM ,F,m - mGCM ,m
s GCM ,m
) ´ s observed,m + m observed,m
munbias,m = mobserved,m
mGCM ,F,m - mGCM ,m
munbias,F,m = (
) ´ s observed,m + m observed,m
s GCM ,m
Bias Correction for Temperature
munbias,m = mobserved,m
mGCM ,F,m - mGCM ,m
)´ s observed,m + mobserved,m
munbias,F,m = (
s GCM ,m
mGCM ,F,m - mGCM,m
)´ s observed,m
Dmunbias,m = munbias,F,m - munbias,m = (
s GCM ,m
m Local,F,m = mobserved,m + Dmunbias,m
mGCM ,F,m - mGCM,m
)´ s observed,m + mobserved,m
m Local,F,m = (
s GCM ,m
Bias Correction for Precipitation
munbias,m = mobserved,m
mGCM ,F,m - mGCM ,m
munbias,F,m = (
) ´ s observed,m + m observed,m
s GCM ,m
munbias,F,m
mGCM ,F,m - mGCM ,m s observed,m
RX F,m =
=(
)´
+1
munbias,m
s GCM ,m
mobserved,m
m Local,F,m = mobserved,m ´ RX F,m
mGCM ,F,m - mGCM ,m
m Local,F,m = (
) ´ s observed,m + m observed,m
s GCM ,m
Bias Correction for Climate Variable
mGCM ,F,m - mGCM ,m
m Local,F,m = (
)´ s observed,m + mobserved,m
s GCM,m
•where μLocal,F,m is the bias-corrected future mean value
of month m for a local area, μGCM,F,m and μGCM,m are
GCM projected future and current mean values, σGCM,m
is the GCM projected standard deviation of month m
under current climate condition, μobserved,m and σobserved,m
are the observed mean and standard deviation of month
m under current climate condition for a local area.
Example of Bias Correction
43
Generalized Bias-Correction
Method
• Climate variable X belongs to distribution F(X, α, β)
where α and β are parameters of distribution function.
F(X, αGCM, βGCM)
F(X, αobs, βobs)
Prob
Prob
CDF
CDF
Xunbias
XGCM
Xunbias,m = F -1 (Prob, a observed , bobserved )
Xunbias,m = F -1 (F(XGCM , aGCM , bGCM ), aobserved , bobserved )
Method 3 physical downscaling
• Taking GCMs’ forecasts as boundary conditions for
a regional climate model.
– Prof. Wu in AS of NTU uses the outputs from Global
Spectral Model (GSM)_CCM3 as boundary conditions to
drive Regional Spectral Model (RSM).
– Other regional models, including Purdue Model and MM5,
are used in Taiwan to produce a local climate scenario
matrix.
資料來源：tccip.ncdr.nat.gov.tw
資料來源：台大吳明進教授
GSM & RSM Resolutions
Model
Resolutions
GSM or RSM0
280km 280km
RSM1
50km 50km
RSM2
15km 15km
GSM: Global Spectral Model
RSM: Regional Spectral Model
Weather Research and
Forecasting (WRF) Model
• The Weather Research and Forecasting (WRF)
Model is a next-generation mesoscale
numerical weather prediction system designed
for both atmospheric research and operational
forecasting needs.
• TCCIP project used WRF model to downscale
MRI projections to the scale of 5km.
2.3 Climate Scenarios
Procedure of Risk
Assessment
Emission Scenarios vs Climate Scenarios
Outputs:
Daily & Monthly Data
Stabilization
at 550 ppm
Emission Scenarios
GCMs
Climate Scenarios:
• Current Scenarios
• Future Scenarios
Scenarios
• Emission Scenarios （溫室氣體排放情境）
– SRES Scenarios （Special Report on Emissions
Scenarios, 2000）
– RCPs Scenarios (Representative Concentration
Pathways)
• radiative forcing values in the year 2100 relative to preindustrial values (+2.6, +4.5, +6.0, and +8.5 W/m2,
respectively)
• Climate Scenarios （氣候情境）
– Current Climate Scenario （Baseline）
– Future Climate Scenario
Experiments of GCMs
• Equilibrium Experiment
1.
2.
3.
4.
Setup initial atmospheric conditions
Running model to reach equilibrium states
Change climate parameters, e.g.2×CO2
Re-running model to reach another equilibrium
states
5. Comparisons between two equilibrium states
• Transition Experiment
1.
2.
3.
4.
Respond to instant forcing
Respond to gradual change of CO2 concentration
Gradually increase CO2 concentration
Simulate climate
SRES Scenarios
• SRES is defined based
on future economic
growth toward
B1
– global or regional
– Economic or
environmental
• SRES classifies four
scenarios, including A1,
A2, B1,and B2.
IPCC AR4 Popular Scenarios
• A2
• A1B
• B1
A2
A1B
B1
Stabilization
at 550 ppm
RCPs
• RCP8.5
– Rising radiative forcing pathway leading
to 8.5 W/m2 in 2100.
• RCP6.0
– Stabilization without overshoot pathway
to 6 W/m2 at stabilization after 2100
• RCP4.5
– Stabilization without overshoot pathway
to 4.5 W/m2 at stabilization after 2100
• RCP2.6
– Peak in radiative forcing at ~ 3 W/m2
before 2100 and decline
Database
• IPCC Data Distribution Center
– http://www.ipcc-data.org/index.html
• TCCIP also provides scenarios for Taiwan.
– http://tccip.ncdr.nat.gov.tw/NCDR/main/index.aspx
AR5採用GCMs - 由CMIP5專案彙整
↑
來自28個單位，共計61個GCMs
TCCIP Climate Scenarios
• TCCIP [Taiwan Climate Change Projection and
Information Platform Project: 臺灣氣候變遷推估與
資訊平台計畫] is a core project funded by National
Science Council, which is in charge of preparing
climate scenarios for climate change impact study in
Taiwan. Climate scenarios are developed for four
periods.
–
–
–
–
Current Climate (Baseline, 基期)
Short-term Future Climate （短期）
Mid-term Future Climate （中期）
Long-term Future Climate （長期）
: 1986~2005
: 2020~2039
: 2050~2069
: 2070~2099
GCMs對應各RCP情境整理表(1/2)
Modeling ACenter
Model
BCC-CSM1.1
BCC
BCC-CSM1.1(m)
CanCM4
CCCma
CanESM2
CMCC-CESM
CMCC
CMCC-CM
CMCC-CMS
CNRM-CERFACS CNRM-CM5
CNRM-CERFACS CNRM-CM5-2
ACCESS1.0
CSIRO-BOM
ACCESS1.3
CSIRO-QCCCE
CSIRO-Mk3.6.0
EC-EARTH
EC-EARTH
FIO
FIO-ESM
GCESS
BNU-ESM
INM
INM-CM4
IPSL-CM5A-LR
IPSL
IPSL-CM5A-MR
IPSL-CM5B-LR
LASG-CESS
FGOALS-g2
LASG-IAP
FGOALS-s2
MIROC4h
MIROC
AMIROC5
MIROC-ESM
MIROC
MIROC-ESM-CHEM
HadCM3
HadGEM2-A
MOHC
HadGEM2-CC
HadGEM2-ES
RCP2.6
RCP4.5
RCP6.0
RCP8.5
Baseline
GCMs對應各RCP情境整理表(2/2)
Modeling ACenter
Model
MPI-ESM-LR
MPI-M
MPI-ESM-MR
MPI-ESM-P
MRI
MRI-CGCM3
MRI-ESM1
GISS-E2-H
NASA GISS
GISS-E2-H-CC
GISS-E2-R
GISS-E2-R-CC
NCAR
NCC
NIMR/KMA
CCSM4
NorESM1-M
NorESM1-ME
HadGEM2-AO
GFDL-CM2.1
NOAA GFDL
GFDL-CM3
GFDL-ESM2G
GFDL-ESM2M
CESM1(BGC)
CESM1(CAM5)
NSF-DOE-NCAR
CESM1(CAM5.1, FV2)
CESM1(FASTCHEM)
CESM1(WACCM)
RCP2.6 RCP4.5 RCP6.0 RCP8.5
Baseline
How many GCMs should be chosen?
How to choose GCMs?
Procedure to Choose GCMs
Climate
Zonation
Analysis
Performance
of GCMs for
each station
Ranking
GCMs
Choice of
GCMs
Weighted
performance
for a zone
Climate Zonation
• Those which GCMs could provide reasonable
baseline climate have higher priority for risk
study.
• However, it is not realistic that nearby weather
stations choose different GCMs, especially
within the same watershed.
• Therefore, climate zonation is determined first.
The weather stations in the same zone should
use the same GCMs.
CWB weather stations
站名
淡水
鞍部
臺北
竹子湖
基隆
彭佳嶼
花蓮
蘇澳
宜蘭
東吉島
澎湖
臺南
高雄
嘉義
臺中
阿里山
大武
玉山
新竹
恆春
成功
蘭嶼
日月潭
臺東
梧棲
經度
121.83
121.73
121.51
121.54
121.73
122.07
121.61
121.86
121.75
119.66
119.56
120.23
120.31
120.42
120.68
120.81
120.90
120.95
121.01
120.74
121.37
121.55
120.90
121.15
120.52
緯度
25.17
25.19
25.04
25.17
25.13
25.63
23.98
24.60
24.77
23.26
23.57
23.01
22.57
23.50
24.15
23.51
22.37
23.49
24.83
22.01
23.10
22.04
23.88
22.75
24.26
共25站
林嘉佑（2014）
劉子明（2010）
吳明進（1993）
北部
台北、淡水、新竹、
梧棲、台中、
北部
台北、淡水、新
竹
北部
台北、新竹、
彭佳嶼
北海岸
基隆
東北海岸
基隆
北海岸
基隆
東部
宜蘭、花蓮
成功、台東
東北部
宜蘭
東北部
宜蘭
東部
花蓮、成功、台
東
東部
花蓮、成功、
台東、大武、
恆春
南部
大武、恆春
恆春半島
大武、恆春
西南部
嘉義、台南、高雄
西部
台中、高雄、台
南
北部山區
鞍部、竹子湖、蘇澳
北部山區
竹子湖、鞍部
中部山區
日月潭、玉山
中部山區
日月潭、玉山
中部
台中、日月
潭
南部山區
阿里山
南部山區
阿里山
山地
阿里山
北部外島
彭佳嶼
西部外島
澎湖、東吉島
東部外島
蘭嶼
西南
澎湖、台南、
高雄
針對氣象站挑選GCMs鄰近格點
GCM
採用站點
bcc-csm1-1-m 119.25、22.991
MRI-CGCM3 119.25、24.112
120.375、21.869
120.375、22.991
120.375、24.112
121.5、21.869
121.5、22.991
121.5、24.112
121.5、25.234
bcc-csm1-1
122.625、25.234
120.938、23.72
120.938、26.511
對應測站
東吉島
澎湖
大武、恆春
台南、高雄、嘉義、阿里山
台中、日月潭、梧棲
蘭嶼
玉山、成功、台東
花蓮、蘇澳
淡水、鞍部、台北、竹子湖、基隆、宜蘭、
新竹
彭佳嶼
臺北、花蓮、蘇澳、宜蘭、東吉島、澎湖、
臺南、高雄、嘉義、臺中、阿里山、大武、
玉山、新竹、恆春、成功、蘭嶼、日月潭、
臺東、梧棲
淡水、鞍部、竹子湖、基隆、彭佳嶼
Performance Indicator
• Indicator
– Correlation of mean monthly rainfall（越高越好）
– RMSE of mean monthly rainfall in dry season (越低越好)
– RMSE of mean monthly rainfall in wet season (越低越好)
• Rank based on the three indicators
1. Rank based on each indicator
2. Sum of ranks of three indicators
3. Determine final rank
R
Dry RMSR
Wet RMSE
Total
Final Rank
GCM A
2
1
3
6
2
GCM B
1
2
1
4
1
GCM C
3
4
2
9
3
GCM D
4
3
4
11
4
Performance for a zone
• Sum of final ranks of all stations in the zone
淡水
台北
台中
新竹
梧棲
• Rank again
bcc-csm1-1-m
14
12
3
11
3
bcc-csm1-1
CCSM4
CESM1-CAM5
CSIRO-Mk3-6-0
FIO-ESM
GFDL-CM3
GFDL-ESM2G
GFDL-ESM2M
GISS-E2-H
GISS-E2-R
HadGEM2-AO
IPSL-CM5A-LR
IPSL-CM5A-MR
MIROC-ESM-CHEM
MIROC-ESM
MIROC5
MRI-CGCM3
NorESM1-M
NorESM1-ME
1
5
4
11
6
19
17
20
15
10
2
18
16
12
9
13
8
7
3
8
2
6
4
5
18
15
19
14
7
1
20
13
16
17
11
9
10
3
20
10
16
11
17
2
15
12
14
13
1
4
5
18
19
6
7
8
9
10
3
9
4
6
14
15
16
12
5
1
13
18
20
19
2
17
7
8
19
11
12
2
17
4
16
13
15
14
1
7
6
20
18
5
8
9
10
Sum
43
58
31
47
32
51
57
78
80
70
49
6
62
58
86
82
37
49
41
33
Final
11
10
3
9
4
6
14
15
16
12
5
1
13
18
20
19
2
17
7
8
Choice of GCMs
rank
1
2
3
4
5
西北部
東部
恆春半島
HadGEM2-AO CESM1-CAM5 MIROC5
CCSM4
GISS-E2-R
GISS-E2-R
CSIRO-Mk3-6-0 CCSM4
CCSM4
NorESM1-ME bcc-csm1-1
CSIRO-Mk3-6-0
MIROC5
CSIRO-Mk3-6-0 HadGEM2-AO
南部
HadGEM2-AO
MIROC5
bcc-csm1-1-m
CCSM4
CESM1-CAM5
北部山區
中部山區
bcc-csm1-1
MIROC5
CESM1-CAM5 CCSM4
NorESM1-ME HadGEM2-AO
HadGEM2-AO CESM1-CAM5
MRI-CGCM3 MRI-CGCM3
西部離島
台灣
HadGEM2-AO HadGEM2-AO
MIROC5
CESM1-CAM5
CESM1-CAM5 CCSM4
bcc-csm1-1-m MIROC5
CCSM4
GISS-E2-R
• 部分氣候分區僅包含單一測站，排序結果與單站
分析結果相同
• 為確保區域氣象站之推薦模式亦可反映出台灣整
體之氣候特性，部分僅於分區內表現較佳之模式
已被剔除
2.4 Uncertainty of GCM projections
10.00
35.0
台北
30.0
CGCM1
CSIRO
6.00
HADCM3
4.00
CCCM
氣溫(oC)
降雨(cm)
8.00
GFDL
2.00
GISS
0.00
1
2
3
4
5
6
7
8
9
10 11 12
台北
25.0
CGCM1
20.0
CSIRO
15.0
HADCM3
10.0
CCCM
5.0
GFDL
0.0
GISS
-5.0
1
2
3
4
5
6
月
6.00
4.00
2.00
0.00
5
6
7
月
11 12
CGCM1
30.0
台中
CSIRO
25.0
CGCM1
20.0
CSIRO
15.0
HADCM3
GFDL
10.0
CCCM
GISS
5.0
GFDL
0.0
GISS
CCCM
4
10
35.0
HADCM3
3
9
台中
8
9 10 11 12
氣溫(oC)
降雨(cm)
8.00
2
8
月
10.00
1
7
1
2
3
4
5
6
7
月
8
9
10 11 12
Relationships between records of a
local weather station and predictions
for the nearest grid point
SRES
CGCM1
CSIRO
Country Study Program
HADCM3
CCCM
GFDL
GISS
Taipei
Rainfall
0.68
-0.68
0.69
0.72
-0.36
0.91
Temp.
0.98
0.98
0.99
0.98
0.97
0.97
TaiChong
Rainfall
0.74
-0.80
0.86
0.76
-0.21
0.87
Temp.
1.00
0.95
0.98
0.99
0.99
0.98
Tainan
Rainfall
0.85
-0.81
0.59
0.86
-0.26
0.72
Temp.
1.00
0.93
0.98
0.99
1.00
0.95
TaiDong
Rainfall
0.80
-0.57
0.69
0.84
-0.60
0.39
Temp.
0.99
0.97
0.99
0.99
0.99
0.94
Comparisons between different
grids’Projections
N24E124中
N20E120中
N24E120短
N24E124長
N20E120長
N24E120中
2.5
2.5
2.5
2
2
2
1.5
1
1.5
降雨比值
降雨比值
降雨比值
N24E124短
N20E120短
SRES-CGCM2 A2
1
N24E120長
1.5
1
0.5
0.5
0.5
0
1 2 3 4 5 6 7 8 9 10 11 12
0
1
2
3
4
5
6
7
8
9 10 11 12
N24E124中
N20E120中
N24E120短
N24E120中
5
5
5
4
4
4
1
0
-1
1
2
3
4
5
6
7
月份
8
9 10 11 12
氣溫差值
6
2
3
2
0
0
3
4
5
6
7
月份
8
5
6
7
8
9 10 11 12
9 10 11 12
N24E120長
2
1
2
4
3
1
1
3
N24E124長
N20E120長
6
3
2
月份
6
氣溫差值
氣溫差值
1
月份
月份
N24E124短
N20E120短
0
1
2
3
4
5
6
7
月份
8
9 10 11 12
Summary
• Climate scenarios, in fact, are weather statistics.
Current climate scenarios can be determined based on
historical weather data.
• Climate change scenarios may be derived from
GCMs or just simple assumptions. Then, future
climate scenarios can be designed.
• Future weather data could be generated based on
climate statistics or simply impose changes on current
records.
• Using more than one GCM is recommended to avoid
the effects of model bias.
What should you know?
•
•
•
•
What are SRES and RCPs scenarios?
Why should you downscale GCMs’ outputs?
What are climate scenarios?
How can you downscale GCMs’ outputs to setup
climate change scenarios?
• How can you prepare your weather data for your risk
assessment model?