Ctl | Meridian | 11. Lecture and Exercise: Uncertainty Check of the Results

11. Lecture and Exercise:
Uncertainty Check of the Results
Hirokazu Murai
Climate Prediction Division
Japan Meteorological Agency
TCC Training Seminar on , Global Warming Projection Information
27 January 2015 (Tokyo, Japan)
1
Uncertainties in global warming projection
• Climate models have improved since the AR4. Models reproduce observed
continental-scale surface temperature patterns and trends over many
decades, including the more rapid warming since the mid-20th century and the
cooling immediately following large volcanic eruptions. (WG1 AR5 SPM)
2
Uncertainties in global warming projection
• However, the global warming projection contains many uncertainties.
• Therefore, we cannot say the results are correct projections without
considering the uncertainties.
It is necessary to consider the uncertainties !!
3
Sample (Global warming projection for Japan)
Check of uncertainty(*)
▶ Computer ▶ Local Disk (C:) ▶ TCC_2015 ▶ Doc ▶ Lecture ▶ 12 ▶ GWP_Sample.pptx
4
Uncertainties in global warming projection
Check of uncertainty(*)
Element and period
Uncertainty
• Temperature in January
• Natural climate variability
• Precipitation in January
• Regional scale
• Temperature in June
• Incompleteness of climate
model
• Short period for calculation
• Precipitation in June
5
Uncertainties in global warming projection
Uncertainty
• Natural climate variability
• Regional scale
• Incompleteness of climate
model
• Short period for calculation
Low uncertainty
Check !
How to check ?
• Wide scale
• AR5 used multi model
Feature of projection results
1. Features are similar or not.
2. Can you explain reason of the
difference ?
• Synoptic scale and other
element
Check consistency
• Model bias
in results.
6
Preparation
1. Copy sample scripts to your working folder.
▶ Computer ▶ Local Disk (C:) ▶ TCC_2015 ▶ Doc ▶ Lecture ▶ 11 ▶ Script
▶ Computer ▶ Local Disk (C:) ▶ TCC_2015 ▶ User ▶ 11 ▶ Script
2. Click icon of “Notepad” and “OpenGrADS” on your desktop.
3. Move to working folder on OpenGrADS console.
ga-> cd /cygdrive/c/TCC_2015/User/11/Script/
7
Sample 1 (Temperature in January)
Temperature(Future-Present) (Jan)
(deg C)
 Temperature around Japan is projected to
increase between 3 to 5 deg C.
 The region at high latitude will warm more
rapidly.
8
Sample 1 (Temperature in January)
Check wide scale
1. Open “Temp_January.gs” by Notepad.
function main (args)
;* Temperature
'open ../../../Data/AGCM/ta-P.ctl'
'open ../../../Data/AGCM/ta-F.ctl'
……
'printim ta_fp_jan.png white'
return
2. Run script by OpenGrADS.
ga-> Temp_January
3. Open image file.
9
Sample 1 (Temperature in January)
 Temperature around Japan is projected to increase
between 3 to 5 deg C.
 The region at high latitude will warm more rapidly.
1. Features are similar or not.
2. Can you explain reason of the
difference ?
Check wide scale
• These two features are similar.
• In addition, land area will warm more than ocean.
• Therefore, uncertainty of small area is small.
10
Sample 1 (Temperature in January)
Check AR5 by multi model
1. Open AR5 image file.
▶ Computer ▶ Local Disk (C:) ▶ TCC_2015 ▶
Doc ▶ IPCC_AR5 ▶ WG1 ▶ graphics ▶
SPM ▶ WGI_AR5_FigSPM-8.jpg
• These features are consistent.
• So, uncertainty of single model is small.
WG1 AR5 Figure SPM.8
11
Sample 1 (Temperature in January)
2m Temperature (℃)
Check model bias
The result of AGCM
is very colder than
that of JRA55.
Negative bias
Sea Level Pressure (hPa)
Model bias
Bias is offset !
Observed
Present
Future
H
L
F-P
• Pressure gradient of winter is reproduced.
• Negative bias is offset.
12
Sample 1 (Temperature in January)
Conclusion
 Temperature around Japan is projected to
increase between 3 to 5 deg C.
 The region at high latitude will warm more
rapidly.
 Therefore, it is considered that the two features show future
climate change.
13
Sample 2 (Precipitation in January)
Precipitation (Future-Present) (January)
(mm)
 Precipitation on land of Japan is
projected to increase in the future.
14
Sample 2 (Precipitation in January)
Check synoptic scale
1. Open “Prec_January.gs” by Notepad.
function main (args)
;* Sea Level Pressure
'open ../../../Data/AGCM/slp-P.ctl'
'open ../../../Data/AGCM/slp-F.ctl'
;* Surface Meridian Velocity
;* 'open ../../../Data/AGCM/va-P.ctl'
;* 'open ../../../Data/AGCM/va-F.ctl'
……
return
Comment out
2. Run script by OpenGrADS.
ga-> reinit
ga-> Prec_January
3. Open image file.
15
Sample 2 (Precipitation in January)
Check synoptic scale
1. Edit “Prec_January.gs”
function main (args)
;* Sea Level Pressure
;* 'open ../../../Data/AGCM/slp-P.ctl'
;* 'open ../../../Data/AGCM/slp-F.ctl'
;* Surface Meridian Velocity
'open ../../../Data/AGCM/va-P.ctl'
'open ../../../Data/AGCM/va-F.ctl'
……
return
Comment out
2. Run script by OpenGrADS.
ga-> Prec_January
3. Open image file.
16
Sample 2 (Precipitation in January)
Precipitation (Future-Present) (January)  Precipitation on land of Japan is
projected to increase in the future.
Sea Level Pressure
(F-P) (hPa)
H
L
weaken
weaken
Surface Meridian Velocity
(F-P) (m/s)
+
-
Check synoptic scale
• Sea level pressure gradient weakens in future, and surface meridian velocity
strengthens in future.
• These indicate that moisture supply decrease to atmosphere at Sea of Japan
(A). This is different from the projection results.
17
Sample 2 (Precipitation in January)
Check other element
Precipitation
(F-P) (Jan)
Temperature
(F-P) (Jan)
(deg C)
• On the other hand, increasing surface temperature suggest that precipitation
increases because saturated vapor pressure increases (B). This is consistent
with the prediction results.
18
Sample 2 (Precipitation in January)
Check model bias
200hPa Wind (m/s)
Cold air
P
F
Weaker
Positive
bias
(A)
Enough moisture
supply
Stronger
More
Warm
North wind
(B)
sea
• Jet stream is located more equatorward, which means cold air from Siberia is
stronger.
• This strengthen effect of (B) than (A), and might increase precipitation.
•
•
Moisture supply decrease to atmosphere at Sea of Japan (A). This is different from the
prediction results.
Increasing surface temperature suggest that precipitation increases since saturated vapor
pressure increases (B). This is consistent with the prediction results.
19
Sample 2 (Precipitation in January)
Conclusion
 Precipitation on land of Japan is projected to
increase in the future.
• Therefore, it is difficult to consider that the feature show
future climate change.
• Uncertainty of Precipitation is larger than temperature.
• This feature may be the influence of the bias.
20
Sample 3 (Temperature in June)
Temperature(Future-Present) (June)
(deg C)
 Temperature around Japan is projected to
increase between 2 to 4 deg C.
 Warming in June will be smaller than in January.
• “Temp_June.gs”
• Uncertainty of area is small.
• Uncertainty of single model is small.
• Model bias is acceptable.
• Therefore, it is considered that the two features show future
climate change.
21
Sample 4 (Precipitation in June)
Precipitation
(F-P) (June)
(mm)
 Precipitation is projected to increase on the
pacific. On the other hand, precipitation is
projected to decrease on the Sea of Japan side
of northern Japan.
• “Prec_June.gs”
•
•
Baiu front will be delayed moving to north. This is
consistent with the prediction result.
Model bias is acceptable.
• Therefore, it is considered that the two features show future
climate change.
22
Sample 5 (Summary)
[ Reproducibility ]
• Although there are little biases, these are acceptable.
[ Future climate change and Reliability]
• Temperature in January … risk of avalanche
• Temperature around Japan is projected to increase between 3 to 5 deg C.
• The region at high latitude will warm more rapidly.
• [OK] It is considered that the two features show future climate change.
• Precipitation in January … risk of snow depth change
• [NG] It is difficult to consider that the feature show future climate change.
• Temperature in June … risk of heat stroke
• Temperature around Japan is projected to increase between 2 to 4 deg C.
• [OK] It is considered that the feature show future climate change.
• Precipitation in June … risk of flood and drought
• Precipitation is projected to increase on the pacific. On the other hand,
precipitation is projected to decrease on the Sea of Japan side of northern
Japan.
• [OK] It is considered that the two features show future climate change.
23
That’s all. Thank you!
24
Download PDF