Intercomaparison of reanalyses,
JRA25, NCEP/NCAR and ERA40
with emphasis on their application to driving CTM
T. Iwasaki, H. Hamada and K. Miyazaki䟼
Tohoku University
䟼JAMSTEC
Plan for reanalysis of minor constituents
Ozone reanalysis will be performed with a CTM driven by
atmospheric reanalysis. In this system, quality of reanalyzed
minor constituents is affected by quality of the atmospheric
reanalyses used for driving the CTM.
This work
1. To compare the mean-meridional circulation of three
reanalyses, JRA-25, ERA40 and NCEP/NCAR.
2. To compare reanalyzed ozone distributions and their
mean and eddy meridional transports.
Kazutoshi ONOGI䚭et al. (JMA)
JRA-25 Overview
• Joint research project of JMA and CRIEPI
• Years䠌1979.1 - 2004.12
transitioned to JMA-CDAS (JCDAS) for after 2005
• Resolution : T106L40 with top level at 0.4hPa
• 3D-Var
• Version : JMA operational system as of April 2004
In addition, SSM/I PW, TOVS radiance level 1c(SSU) and 1d(HIRS,
MSU) were assimilated.
• JRA-25 original/firstly used observational data
TCR, SSM/I snow coverage, digitized Chinese snow depth data,
reprocessed GMS-AMV
• JRA-25 original boundary/forcing data
Daily COBE SST and sea ice (Ishii 2005, IJC), daily 3D-ozone profile
Globally-averaged Monthly Precipitation
Courtesy: H. Koide
Correlation of Monthly Precipitation with GPCPv2
SSM/I PW assimilated
Courtesy: H. Koide
Global
Temperature
Anomaly
Anomaly from
averaged temperature
of each level for each
reanalysis
Courtesy: J. Tsutsui and M. Sakamoto
㻧㻭㻩㻃㻷㼕㼒㼓㼒㼖㼓㼋㼈㼕㼈㻋㻔㻓㻓㻓㻐㻔㻓㻓㼋㻳㼄㻌
<JRA>
<ERA>
⦃ᗐ[°]
⦃ᗐ[°]
Hadley Cirulation
: ERA>JRA>NCEP
<NCEP>
Mass stream functions (*1e+11[kg/s])
500hPa㟻
JRA
ERA
NCEP
㻭㻭㻤㻃㻷㼕㼒㼓㼒㼖㼓㼋㼈㼕㼈㻋㻔㻓㻓㻓㻐㻔㻓㻓㼋㻳㼄㻌
<JRA>
<ERA>
Hadley Circulation: JRA=ERA>NCEP
⦃ᗐ[°]
⦃ᗐ[°]
<NCEP>
Mass Streamfunction (*1e+11[kg/s])
500hPa㟻
JRA
ERA
NCEP
Precipitation (1979-2001)[mm/day]
DJF
JJA
Seasonal Variations
Max and Min of Mass Streamfunction at 500 hPa
䚭JRA
e+11(kg/s)
䚭ERA
䚭NCEP
Solid䠌NH
Broken䠌SH
month
• JRA=ERA>NCEP in SH winter
• ERA>JRA>NCEP in NH winter
• SH winter > NH winter
Inter annual variability of Hadley circulation
SH-JJA
10E+11(kg/s)
NH-DJF
year
•Max 䟹Min of Mass Streamfunction at 500hPa
䚭䚭JRA
䚭䚭ERA
䚭䚭NCEP
DJF: NH Hadley circulations are gradually strengthened.
Inter-annual variability is strongly correlated among analyses.
JJA: No trend is found in SH winter.
Poor correlation of inter-annual variability among analyses.
㻧㻭㻩㻝㻃㻶㼗㼕㼄㼗㼒㼖㼓㼋㼈㼕㼈㻋㻔㻓㻓㻐㻔㻓㼋㻳㼄㻌
<JRA>
<ERA>
⦃ᗐ[°]
B-DCirculation䠌
<NCEP>
ERA>JRA=NCEP
⦃ᗐ[°]
JRA
ERA
NCEP
100hPa
Mass Streamfunction (*1e+10[kg/s])
JRA䠄䠇
NCEP䟿JRA25
DJF
ERA䟿JRA25
㻭㻭㻤㻝㻃㻶㼗㼕㼄㼗㼒㼖㼓㼋㼈㼕㼈㻋㻔㻓㻓㻐㻔㻓㼋㻳㼄㻌
<JRA>
<ERA>
B-DCirculation䠌ERA>JRA=NCEP
<NCEP>
JRA
ERA
NCEP
100hPa
Mass Streamfunction (*1e+10[kg/s])
Seasonal Variations
Max and Min of Mass Streamfunction at 100 hPa
䚭JRA
䚭ERA
䚭NCEP
Solid䠌NH
Broken䠌SH
month
䝿Downward mass flux integrated over hemispheres at 100 hPa
䝿Amplitude of Seasonality: ERA>JRA,NCEP
䝿NH winter 䠐SH winter
Inter annual variability of T-S mass exchange
NH: DJF
year
•Max 䟹Min of Mass Streamfunction at 100hPa
No significant trend.
Poor correlation among analyses.
SH: JJA
䚭䚭JRA
䚭䚭ERA
䚭䚭NCEP
Mean-Meridional Circulation
1. Hadley circulation is: ERA>JRA>NCEP. This may be due
to precipitation neat the ITCZ.
2. In NH winter, the Hadley circulation has increasing trends
and good correlations among the reanalyses.
3. Midlatitude-tropospheric circulation is almost even in all.
4. BD circulation (downward mass flux at 100hPa) is
ERA>JRA>NCEP. But we do not find a strong relationship
with EP flux divergence.
5. Stratospheric analysis does not show significant trends or
correlations but suffers from biases of satellite data.
Ozone reanalysis
with a CTM driven by the atmospheric reanalysis
(1993-1994)
CTM is driven by T42L68-AGCM and the GCM is nudged to
the reanalyses with appropriate relaxation times.
No chemical data is assimilated so as to see the transport
characteristics in detail.
Transport diagnosis is made based on mass-weighted
isentropic zonal means.
Definition of Zonal Mean State
Mass-weighted isentropic zonal means are taken for all variables, to
express the lower boundary conditions and nongeostrophic and finiteamplitude effects of waves.
Zonal Means
Eddies
A( y,* ,t )* )
A# " A ! A*
' (p
1
A
(
x
,
y
,
*
,
t
)
%
!
Lx
& (*
(p$
"dx
(* #
( AB)* = A*"
B*!( A'B')*
Correlations
Isentropic zonal mean pressure is used as the vertical coordinates.
Ԫ The thermodynamic equation is prognostic while the continuity
equation is daignostic.
Vertical Coordinate (isentropic zonal mean pressure)
In the zonal averaging, we set
p† ! p
p ( x) = ps ( x) for ! s (x) < !
Logarithmic Pressure Coordinate
z†
(
" ! H log p† / p00
)
Formulation of the zonal mean transport equation
The mass-weighting is also considered for the mixing ratio of
minor constituent as well as the mean meridional circulation.
The transport equation helps us to express the conservation
of minor constituents including the lower boundary condition.
Zonally symmetric transport equation:
Mean transport
Time derivation of
mass mixing ratio
Eddy transport
Chemical
production/loss
Formulation of the zonal mean transport equation
Mean transport flux
Eddy transport flux
The vertical component of eddy transport term can be separated
into the contributions due to adiabatic and diabatic processes.
For adiabatic processes, the direction of eddy becomes
This infers that the eddy flux is parallel to the local isentropic
surface for adiabatic processes.
DJF
Mean ozone transport
JJA
䜮䝂䝷䝙䝭䝇䜳䜽
DJF
Eddy ozone transport
JJA
Ozone mixing ratio
Contours int.: 1ppmv
UARS Clim. (JAN)
CTM JRA25 (JAN, 1994)
UARS Clim. (JUL)
CTM JRA25 (JUL,1994)
Ozone mixing ratio in Jan. 1994
Contours int.: 1ppmv
CTM: NCEP1
CTM: NCEP2
CTM: JRA25
CTM: ERA40
Ozone mixing ratio in Jul. 1994
Contours int.: 1ppmv
CTM: NCEP1
CTM: NCEP2
CTM: JRA25
CTM: ERA40
Ozone mixing ratio in Jan. 1994
DIFF(NCEP1-JRA25)
DIFF(NCEP2-JRA25)
Contours int.: 0.1ppmv
DIFF(ERA-JRA25)
L
Ozone mixing ratio in Jul. 1994
DIFF(NCEP1-JRA25)
DIFF(NCEP2-JRA25)
DIFF(ERA-JRA25)
Seasonal variation of Total ozone in 1994
OBS: UARS
CTM: JRA25
Contours int.: 20DU
Total ozone in 1994
CTM: NCEP1
CTM: JRA
Contours int.: 20DU
CTM: NCEP2
CTM: ERA40
Total ozone in 1994
CTM: JRA25
Contours int.: 20DU
DIFF(NCEP2-JRA25)
DIFF(NCEP1-JRA25)
DIFF(ERA40-JRA25)
Contours int.: 3DU
Mean meridional ozone fluxes ingerated through the atmosphere (kg/s)
CTM: JRA25
DIFF(NCEP2-JRA25)
Contours int.: 500
DIFF(NCEP1-JRA25)
DIFF(ERA40-JRA25)
Contours int.: 100
Eddy meridional ozone fluxes ingerated through the atmosphere (kg/s)
CTM: JRA25
DIFF(NCEP2-JRA25)
Contours int.: 300
DIFF(NCEP1-JRA25)
DIFF(ERA40-JRA25)
Contours int.: 100
Meridional ozone fluxes ingerated in the troposphere and stratosphere (kg/s)
Mean: Stratosphere
(Kg/s)
Mean: Troposphere
NCEP1
NCEP2
ERA
JRA
Eddy: Stratosphere
Eddy: Troposphere
DJF
Meridional ozone fluxes ingerated in the troposphere and stratosphere (kg/s)
Mean: Stratosphere
(Kg/s)
Mean: Troposphere
NCEP1
NCEP2
ERA
JRA
Eddy: Stratosphere
Eddy: Troposphere
JJA
Summary
1. Ozone mixing ratio obtained with JRA25 is greater than
those with ERA and NCEP in the lower stratosphere, while it is
less in the middle stratosphere. The mixing ratio in the lower
stratosphere may be the main reason why the extratropical
total ozone is the largest when CTM is run with JRA25.
2. Vertically integrated net poleward mean transport with
JRA25 is greater than NCEP and a little greater than ERA.
3. Vertically integrated net equatorward eddy transport with
JRA25 is greater than NCEP but a little smaller than ERA.
4. Reanalyzed ozone has some biases and requires further
improvements of CTM and the atmospheric reanalyses.
Announcement
The 3rd WCRP Reanalysis
Conference
To be held in Tokyo in January, 2008
Stream function in July & August
CTM: JRA25
DIFF(NCEP2-JRA25)
Contours int.: 500
DIFF(NCEP1-JRA25)
DIFF(ERA40-JRA25)
Contours int.: 100
Stream function in January & February
CTM: JRA25
DIFF(NCEP2-JRA25)
Contours int.: 500
DIFF(NCEP1-JRA25)
DIFF(ERA40-JRA25)
Contours int.: 100
Objective Analysis
Analysis
JRA-25
Resolution
Period
2.5°×2.5°
23 (upto 0.4hPa)
NCEP/NCAR
2.5°×2.5°
17 (upto 10hPa)
ERA-40
2.5°×2.5°
23(upto 1hPa)
1979䡐2001
JRA: 23layers:
1000 925 850 700 600 500 400 300 250
200 150 100 70 50 30 20 10 7 5 3 2 1 0.4
NCEP&NCEP2: 17layers: 1000 925 850 700 600 500 400 300 250
200 150 100 70 50 30 20 10
ERA: 23layers:
1000 925 850 775 700 600 500 400 300
250 200 150 100 70 50 30 20 10 7 5 3 2 1
2. MRI/JMA ozone reanalysis system
Meteorological fields
To reproduce the
meteorological
field observed,
Reanalysis data is
assimilated into
the GCM with a
nudging technique.
ERA 40
Data assimilation is
reanalysis
not applied to chemical
data
species.
4-D
assimilation
systemof
for GCM
Thus,
distributions
chemical species are
calculated in the CTM.
Chemical species
Chemical
processes
TOMS
ozone data
Sedimentation
& emission
Total ozone assimilation
Transports
CTM
GCM
MRI chemical transport model
An ozone reanalysis system is being developed based on the CTM
driven by MRI/JMA 98 GCM (Shibata et al., 2005).
䝿Long lived chemical species : 34 species
N2O, Ox, CH4, H2O, NOy, HNO3, N2O5, Cly, CO, CO2, AEROSOLS,
HO2NO2, NOx, ClONO2, ClOx, HOCl, Cl2, ClNO2, Cl2O2, CCl4, OClO,
H2O2, HBr, BrONO2, HOBr, HCl, BrOx, N2O, CFCl3, CF2Cl2, Bry, CH3Cl,
CH3Br, CF2ClBr, CF3Br,
䝿Short lived chemical species : 15 species
O(1D), O(3P), O3, OH, HO2, H, N, NO, NO2, NO3, Cl, ClO, Br, BrO, BrCl
䝿Chemical reactions 䠌
Photolysis䟺23䟻䚮Heterogeneous(5:PSC䚮2:SSA䟻䚮
Gas phase䟺59䟻: Ox group(5), HOx group(12), NOx group(11),
Clx group(21), Brx group(9), Hydrocarbon group(1)
䝿Resolution䠌 T42 with 68 levels䟺Surface 䡐 0.01 hPa䟻
MRI ozone reanalysis system
GCM
CTM
(Nudging)
Model run
TIME
00 obs
06 obs
12 obs
18 obs
00 obs
obs :ERA40 Reanalysis data
Data assimilation technique
A data assimilation technique
based on simple Newtonian
relaxation called the nudging
technique (Hoke and Anthes,
1976) is used to force the GCM
toward the reanalysis data. The
prognostic equation is given by
GCM forcing term
Relaxation term
Horizontal and vertical wind
䊲㻃Transport
Temperature, humidity
䊲㻃Chemistry
Meridional circulation
Mass stream function (*1e+11 [kg/s])
Hadley circulation
B-D circulation
DJF
DJF
JJA
JJA
JRA-25
ERA-40
NCEP/NCAR
zonal mean at 500hPa
zonal mean at 100hPa
Courtesy: H. Hamada & T. Iwasaki (Tohoku Univ.)
QBO and SAO
10S-10N averaged zonal wind cross section
Courtesy: H. Hatsushika
Surface temperature Trend
JRA-25 and ERA-40
Global Temperature Anomaly
JRA-25, ERA-40, CRU(Jones)
Top : monthly mean, Bottom : 5-year moving avarage
Distribution of tendency (K/decade)
Courtesy: J. Tsutsui
Comparison of ozone
and temperature of
JRA-25 and ERA-40
Ozone density is
dominant for climate in
the stratosphere.
Global averaged Total column ozone (DU)
Earth Probe
Nimbus-7 TOMS
Sudden increase from 1989
to 1991 in ERA-40
Ozone in JRA-25 is
unstable for the period
without TOMS data from
May 1993 to July 1996
El chichon
Pinatubo
Courtesy: M. Sakamoto
䜄䛮䜇
• ௑ᅂJRA-25්よᯊ䝋䞀䝃䛴Ꮔ༔㟻ᚘ⎌䛴Ẓ㍉䛴
䛥䜇よᯊ䛱p†᪁⛤ᘟ⣌䜘⏕䛊䚮NCEP/NCAR䝿ERA40්よᯊ䝋䞀䝃䛮Ẓ㍉䜘⾔䛩䛥䚯
Ꮔ༔㟻ᚘ⎌䛴Ẓ㍉
㈻㔖Ὦ⥲㛭ᩐ䛴
ᙁ䛛
ᖳ䚱ንິ
Ꮢ⟿ንິ
DJFᑊὮᅥ
(໪䝿䝓䝍䝰䞀ᚘ⎌)
ERA>JRA>NCEP
JJAᑊὮᅥ
(༞䝿䝓䝍䝰䞀ᚘ⎌)
ERA䍝JRA>NCEP ┞㛭䛵఩䛕䛰䛊
JRA䛴䜅ᘽ䜄䜑䝌䝰䝷䝍
DJFᠺᒒᅥ
(໪䝿B-Dᚘ⎌)
ERA>JRA=NCEP
ERA䍝JRA>NCEP
┞㛭䛵఩䛊䚯䛵䛩䛓䜐䛮
䛝䛥䝌䝰䝷䝍䛵ぜ䜏䜒䛰䛊
JJAᠺᒒᅥ
(༞䝿B-Dᚘ⎌)
ERA>JRA=NCEP
┞㛭䛵఩䛊䚯JRA䝿
NCEP䛵ᙁ䜄䜑䝌䝰䝷䝍
┞㛭䛵㧏䛊
ᙁ䜄䜑䝌䝰䝷䝍
ERA>JRA=NCEP
䌲
ERA>JRA=NCEP
䌳
ERA䍝JRA>NCEP
Stream function in January & February
CTM: NCEP1
CTM: JRA
Contours int.: 20DU
CTM: NCEP2
CTM: ERA40
Contours int.: 3DU
Stream function in July & August
CTM: NCEP1
CTM: JRA
Contours int.: 20DU
CTM: NCEP2
CTM: ERA40
Contours int.: 3DU
ᖈ≟ᖲᆍ䛱䛪䛊䛬
ࡌ࡬࡙ࡡንᩐ࡞ᑊࡊ࡙㈻㔖㔔ࡲ௛ࡄࡊࡒᖈ≟ᖲᆍࢅࡵࡔ࠷ࡒࠊᖈ
≟ᖲᆍࡡᏽ⩇ࡢ
ࡡࡻ࠹࡞⾪ࡈࡿࡾࠊ㕼├㟻࡞ࡢ➴Ὼన㟻࡞Ἒࡖ࡙ᖈ≟ᖲᆍࡊࡒẴ
Ὼࢅ⏕࠷ࡾࠊ
ࡐࡡ≁ᚡࡢᆀ⾦㢴㎾జࢅ⏕࠷࡝࠷࡚ἴິᖲᆍὮ┞பష⏕ࡡ⾪⌟࠿
ྊ⬗࡚࠵ࡽࠉࡱࡒୖ㒂ሾ⏲ࢅḿࡊࡂ⾪⌟࡚ࡀࡾࡆ࡛࡚࠵ࡾࠊ
䝿➴Ὼన㟻䛭᮶こᖲᆍሔ䛱䜎䜑Ꮔ༔㟻ᚘ⎌䛴䝥䝮䝇䝌
䠃䠀䝭䜴䝭䝷䜼䝩Ⓩ䛰Ꮔ༔㟻ᚘ⎌䛒ᚋ䜏䜒䜑䛥䜇䚮⇍㍲㏞䝿∸㈻
㍲㏞䜘⩻䛎䜑୕䛭ฺ᭯
䠄䠀ἴິ䛱䜎䜑㐘ິ㔖㍲㏞䛴ຝᯕ䜘㐲ว䛱᡽䛎䜑
㜾Ềฦᕱ(82ᖳJJA)
<JRA>
<NCEP>
• JRA䛒NCEP䛱Ẓ䛿㜾Ề䛒ᙁ䛕⾪⌟䛛䜒䚮䜄䛥ITCZ
䛒⾪⌟䛛䜒䛬䛊䜑䚯
㜾Ềฦᕱ(98JJA)
<JRA>
<NCEP>
• JRA䛵82ᖳ䛱Ẓ䛿ITCZ䛒ᘽ䛊
㜾Ềฦᕱ(82D~83JF)
<JRA>
<NCEP>
• JJA䛱Ẓ䛿ᘽ䛊䚯䜄䛥ฦᕱ䛵JRA䛮NCEP
䛭జ㏳䛩䛬䛊䜑䚯
㜾Ềฦᕱ(97D~98JF)
<JRA>
<NCEP>
• JRA䛮NCEP䛴ฦᕱ䛵జ㏳䛩䛬䛊䜑䚯
• 82D~83JF䛮Ẓ䛿㜾Ềᙁᗐ䛒ᙁ䛕䛰䛩䛬䛊䜑䛮䛵Ử䛝
䛬ゕ䛎䛰䛊䛑䠑
➴Ὼన㟻୕䛭䛴᮶こᖲᆍሔ
!
Ὼనθ䛴➴Ὼన㟻୕䛭䛴᮶こᖲᆍሔ䜘䚭䚭ᖲᆍሔ䛑䜏䛴೩ᕣ䜘
!'
!'= 0
䚭䚭䛭⾪䛟䛮䚭䚭䚭䚭䚭䛭ᣉᩋ䛒䛰䛊䛴䛭⇍ງᏕ䛴ᘟ䛵
*
'(*
'(*
'(* & Q #
+ v*
+ w*
=$ !
't
'y
'z % ) "
A* =
"p /"!
" p /"!
⛛Ὦ䟺✭Ẵም⮤㌗䛒ິ䛕䟻 ฆ⤎䚮ᨲᑏ➴䛱䜎䜑አ
㒂䛑䜏䛴ຊ⇍䚮෫༴
䝿➴Ὼన㟻䛭᮶こᖲᆍሔ䛱䜎䜑Ꮔ༔㟻ᚘ⎌䛴䝥䝮䝇䝌
䠃䠀䝭䜴䝭䝷䜼䝩Ⓩ䛰Ꮔ༔㟻ᚘ⎌䛒ᚋ䜏䜒䜑䛥䜇䚮⇍㍲㏞
䚭䚭∸㈻㍲㏞䜘⩻䛎䜑୕䛭ฺ᭯
䠄䠀ἴິ䛱䜎䜑㐘ິ㔖㍲㏞䛴ຝᯕ䜘㐲ว䛱᡽䛎䜑
A
୯㧏⦃ᗐ䝿ᠺᒒᅥ䛭䛴ἴິ䛱䜎䜑ᚘ⎌䛴㥉ິ
㻨㻳㻃㼉㼏㼘㼛䟺ἴິ䛒㐘䛼㐘ິ㔖䟻
&
& '+
*
F * $ ( (u )v)) ,( p$
$
% 'x
%
# #!
!!
""
᮶こ᪁ྡྷ䛴㐘ິ᪁⛤ᘟ
#u *
# u*
# u*
+ v*
+ w*
" f v* = !!F
#t
#y
#z
ἴ䛒㐘䛼㐘ິ㔖䛴
཭ᮨ䝿Ⓠᩋ
㏻⤾䟺㈻㔖ಕᏋ䟻䛴ᘟ
*
*
!v
"1 !# 0 w
+ #0
=0
!y
!z