-
Content of LecturesLecture 1: Current status of Climate
modelsLecture 2: Improvement of AGCM focused on MJOLecture 3:
Multi-model Seasonal PredictionLecture 4: Seasonal
PreditabilityClimate Modeling and Prediction
In-Sik KangSeoul National University
-
Current Status of Climate Models In-Sik Kang
Climate Environment System Research CenterSeoul National
UniversityLecture 1Climate Environment System Research Center
-
Procedure What is the climate model?Part : AGCM General
performance of state-of-the-art AGCMs Inherent limitation of
two-tier strategy using AGCM Part : CGCM Current status of CGCMs
Efforts for development of CGCMPart : Climate System Model Future
perspective on the climate model
-
What is the Climate Model ? The general circulation model (AGCM)
is the model close to the real atmospheric state of the whole
Earth, which has been developed since middle of the 20th
century.
As the AGCM can reproduce the real atmospheric condition in the
planetary scale, it is the most useful equipment of experiment and
climate prediction. Recently, the concept of global climate model
considering the condition of ocean and vegetation as well as
atmosphere, has been established.Integrated Climate and Environment
Model
-
Structure of Atmospheric General Circulation ModelDynamics
Three-dimension hydrostatic primitive equations on sphere with
sigma coordinate
-
General Performance of State-of-the-art AGCMsClimate Environment
System Research CenterLecture 1: Current status of climate models
Global Atmospheric Anomalies associated with ENSO Climatological
Monsoon Variabilities Monsoon Variabilities during 97/98 El Nio
Inherent Limitation of Two-tier Strategy using AGCM
-
Experimental Design andParticipated Models CLIVAR
Asian-Australian Monsoon Atmospheric GCM Intercomparison Project
The AGCM intercomparison program was initiated by the
CLIVAR/AsianAustralian Monsoon Panel to evaluate a number of
current atmospheric GCMs in simulating the global climate anomalies
associated with the recent El Nio. Experimental Design Models
Participated
-
Monsoon Predictability: Climatological JJA Precipitation
-
Two Categories of AGCMs following to Basic State 10N-20N
Latitudinal Mean of Rainfall VariabilityIndian MonsoonregionWestern
North Pacific Monsoon regionRed SeriesBlue SeriesJJA Precipitation
(shading )and 850 hPa Streamfunction (contour)(c) Composite (DNM,
IAP, MRI, NCAR)(b) Composite (COLA, GEOS, IITM,SNU)(a) CMAP
Observation
-
1st Mode of EOF for Climatological MJJAS Precipitation
-
Pattern correlation for each EOF mode for MJJAS precipitation
The pattern correlations between the eigenvectors of individual
models and the observed counter parts
All correlation values of the model composite are quite
high.
But most of the models have a large value of correlation only
for the first eigenvector but not for the higher modes.
u850
0.88425610.7254650.5043399-0.2965269
0.67491210.32100470.0504237410.6542835
0.84416160.7872080.58498660.404967
0.87755420.86704130.563264-0.049364682
0.86198410.64080720.50917880.2217971
0.81585820.81928450.47214310.094350524
-0.1173572-0.37162240.0357588420.2612103
0.75565030.44869980.63875090.1157582
0.80494360.90493360.75579040.645509
0.82039130.6663390.38659470.4122365
0.89017890.88512570.76951010.6930253
1st mode
2nd mode
3rd mode
4th mode
Pattern correlation for each mode of SVD for U850
prcp
0.57957050.44997780.19698510.3471215
0.38592230.38680920.1520170.1898449
0.68981130.24826970.37772130.1842896
0.68713930.28527910.13440360.1845249
0.45462360.46947710.0162162090.2503924
0.72243580.45472230.0810204970.1462904
0.22144850.43495490.40457480.074436501
0.11474890.38852840.3861360.045623373
0.61850390.65645060.60929610.4031093
0.65258990.70295270.18531530.1966341
0.76049990.69304710.62309130.4669239
1st mode
2nd mode
3rd mode
4th mode
Sheet1
prcp1st2nd3rd4thCISOCISO1st
CMAP0.9999999110.999999911
COLA0.57957050.44997780.34712150.19698510.10783320.3120939
DNM0.38592233.87E-011.90E-011.52E-010.2598015-5.69E-02
GEOS0.68981130.24826971.84E-010.3777213-0.2731366-0.1484626
GFDL0.68713930.28527910.18452491.34E-010.2344090.2150892
IAP0.45462360.46947710.25039241.62E-02-0.17591934.72E-02
IITM0.72243580.45472230.14629048.10E-02-0.2193118-0.2603764
MRI0.22144854.35E-017.44E-020.40457480.2005076-1.93E-03
NCAR0.11474893.89E-014.56E-023.86E-012.78E-02-0.2764597
SNU0.61850390.65645060.40310930.6092961-0.23509260.5032701
SUNY0.65258990.70295270.19663410.1853153-0.1077461-0.2677587
Comp.0.76049990.69304710.46692390.6230913-0.10474190.1522722
u8501st2nd3rd4th
CMAP1111
COLA0.88425610.7254650.5043399-0.2965269
DNM0.67491210.32100475.04E-020.6542835
GEOS0.84416160.7872080.58498660.404967
GFDL0.87755420.86704130.563264-4.94E-02
IAP0.86198410.64080720.50917880.2217971vi
IITM0.81585820.81928450.47214319.44E-02
MRI-0.1173572-0.37162243.58E-020.2612103
NCAR0.75565030.44869980.63875090.1157582
SNU0.80494360.90493360.75579040.645509
SUNY0.82039130.6663390.38659470.4122365
Comp.0.89017890.88512570.76951010.6930253
Sheet2
Sheet3
-
SOI = SLP anomaly difference over two regions [145oW-155oW,
5oS-5oN] [125oE-135oE, 5oS-5oN]Evolution of1997-98 El Nio and SOI
Indices (a) NINO3.4 INDEX(b) SST anomaly DJF97/98(c) Observed and
Simulated SOI indices
-
Precipitation Anomalies for Each Summer and Winter Model
CompositeCMAP Observation
-
Fig. 6. Distribution of precipitation anomaly during the
97/98winter. (a) is for the CMAP observation, and the rest of the
figures are the ensemble mean of each model.
-
Current Predictability: Pattern Correlation and RMS of
Rainfall(b) Root-mean-square(a) Pattern CorrelationMonsoon-ENSO
region:60oE-90oW, 30oS-30oN
-
DJF97-98 200hPa Geopotential Height AnomaliesPrecipitation200hPa
Geopotential heightPNA CorrelationPNA Normalized RMSPNA region:
180oE-60oW, 20-80oNCorrelation vs. RMSPrecipitation vs.
Circulation
-
Tropical SST AnomalyImprovement of physical parameterization :
PBL, Convection.Advances in the computing power : High
resolutionImprovement of Predictability following to ENSO
Simulation
-
Current Monsoon Predictability: Pattern CorrelationEl-Nino
region (160oE-80oW, 30oS-30oN)Monsoon region (40-160oE,
30oS-30oN)Southeast Asian and Western North Pacific region
(80-150oE, 5-30oN)Correlation between CMAP and models for
JJA97/98
-
(a) JJA (b) JJA (c) DJF (d) DJF Observation5 Model
CompositeCause of Low Predictability: Atmosphere-Ocean
InteractionCorrelation between JJA SST and Precipitation during
1979-1999
InstituteModelResolutionExperiment TypeEnsemble
MemberJMAJMAT63L40SMIP)10KMAGDAPST106L21SMIP10NCEP
NCEPT62L28SMIP10NASA/NSIPPNSIPP2ox2.5o
L43AMIP9SNUGCPST63L21SMIP10
-
(a) Observation (1979-2001)(b) AGCM (1979-2001)(c) Mixed layer
model (16 years)(d) CGCM (50 years) No ENSO Only local air-sea
interactionCorrelation between JJA SST and PrecipitationImproved
Simulation using Coupled System over WNP
-
Precipitation Climatology During Boreal Summer Observation
(CMAP)CGCM(Ver.2)AGCM
-
Current Status of CGCMsClimate Environment System Research
CenterLecture 1: Current status of climate models Present the
problem of state-of-the-art CGCMs through CGCM Intercomparison
Project (CMIP)
-
Coupled Model Intercomparison Project (CMIP) Participating Model
Under the auspices of the Working Group on Coupled Modeling (WGCM)
The PCMDI supports CMIP by helping WGCM to determine the scope of
the project. CMIP has received model output from the pre-industrial
climate simulations ("control runs") and 1% per year increasing-CO2
simulations.
Sheet1
Atmospheric modelOceanic modelAtmospheric resolutionOCEAN
resolutionFlux adjust
MRI2MRI/JMA98Bryan-Cox Primitive eq.
codeT42(2.8X2.8),L302.0X2.5,L23H,W,M
GFDL_R30GFDLGFDL MOM 1.1R30(2.25X3.75),L141.875X2.25,L18H,W
CSIRO Mk2CSIRO 9-level agcmBryan-Cox Primitive eq.
codeR21(3.2X5.6),L93.2X5.6,L21H,W,M
HadCM3Unified modelBryan-Cox Primitive eq.
code2.5X3.75,L191.25X1.25,L20.
HadCM2Unified modelBryan-Cox Primitive eq.
code2.5X3.75,L192.5X3.75,L20H,W
CCCma CGCM1GCM2GFDL MOM1.1T32(3.8X3.8),L101.8X1.8,L29H,W
DOE PCMCCM3LANL POPT42(2.8X2.8),L180.67X0.67,L32.
CSM 1.0CCM3.0NCOM1.1T42(2.8X2.8),L182.0X2.4,L45.
ECHO-gECHAMHOPE-gT30(3.75X3.75),L19T42(2.8X2.8),L20H,W
ECHAM4/OPYC3ECHAMOcean isoPYCnal
GCMT42(2.8X2.8),L192.8X2.8,L11H,W
Sheet2
Sheet3
-
CMIP: SST Climatology Warm Bias at Eastern Edge of the
Equatorial Pacific Too strong Cold tongue Kuroshio Extension region
Common Problems in CGCM Simulations
-
CMIP: Precipitation Climatology-
-
CMIP: Vertical Structure of Zonal Current along the Equator
Common Problems in CGCM Simulations Mostly simulate weak equatorial
undercurrents Strong easterly surface currents Some models have a
critical problem to simulate oceanic vertical structure
-
CMIP: Interannual SST Variability Weak Interannual variability
in the eastern Pacific Relatively strong in the central-western
Pacific. Better interannual variability seems to be connected to
better vertical ocean structure simulation except BCM case Common
Problems in CGCM Simulations
-
Development of CES Coupled GCMMixed Layer ModelVertical Eddy
Viscosity:Vertical Eddy Diffusivity:: empirical Constantwhere: TKEl
: the length scale of turbulenceNoh and Kim (1999) To simulate
correct vertical ocean structure
Coupled GCMAGCMOGCMCoupling StrategyCES CGCM (Ver. 1)CES
AGCMT31, 21 levels (3.75X3.75)MOM3 OGCMUneven Grid(3 lon. X 1 lat.
near equator)1-day Mean Exchange(SST, Heat Flux, Wind stress, Fresh
Water Flux)No Flux CorrectionCES CGCM(Ver. 2)CES AGCMT42, 21 levels
(2.8125X2.8125)MOM2.2 OGCM + Ocean mixed layer modelUneven Grid(1
lon. X 1/3 lat. near equator)1-day Mean Exchange(SST, Heat Flux,
Wind stress, Fresh Water Flux)No Flux Correction
-
SST ClimatologyObservationCGCM with MLMCGCM without MLM
-
a) Observationb) CGCM without MLMVertical Structure of Ocean
Temperature1oS-1oN meanb) CGCM with MLM
-
Vertical Structure of Zonal Current along the Equator1oS-1oN
meana) Observationb) CGCM without MLMc) CGCM with MLM
-
ObservationInterannual SST VariabilityCGCM with MLMCGCM without
MLM
-
Effect of Horizontal Diffusiona) Observationb) Strong
Diffusionc) Weak DiffusionEXP_strong (CNTL)EXP_weakHorizontal
Mixing for MomentumNotes When horizontal diffusion is strong Weak
Equatorial Undercurrent Strong Equatorial Surface Current Westward
extension of cold tongue Weak SST zonal gradient Weak Interannual
Variability
-
Effect of Horizontal DiffusionStrong DiffusionWeak DiffusionSST
ClimatologyInterannual Variability
-
ENSO Variability in the CGCM with MLMYear NINO3.4 SST Linear
Regression with respect to NINO3.4 SST SST Anomalies along the
Equator
Correlation of rainfall and geopotential heightEach model
precipitaiton anomalyComparison of climate models having different
system also back up theses results.These are all SNU AGCM and this
is coupled with slab ocean and this is fully coupled CGCM.Coupled
systems mimic the realistic negative relationship clearly different
from AMIP.Even in the slab ocean model case only having local
air-sea interaction without any advection or ocean dynamics,the
negative relation is shown so clearly as the seasonal
characteristics in the summer hemisphere.