new coupled forecast system at the japan meteorological agency
Post on 02-Jan-2016
39 Views
Preview:
DESCRIPTION
TRANSCRIPT
New Coupled Forecast System at the Japan Meteorological Agency
Yuhei Takaya1 (ytakaya@met.kishou.go.jp), Tamaki Yasuda2, Satoshi Matsumoto2, Tosiyuki Nakaegawa2,
Hirotaka Kamahori1 and Tomoaki Ose2
1. Climate Prediction Division, Japan Meteorological Agency2. Meteorological Research Institute, Japan Meteorological Agency
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, Florida
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
2
History of ENSO Monitoring and Forecast Service at JMA
1992
1996
1999
2003
2008
2010
The Second Version of JMA’s CGCM (JMA-CGCM02)
Launch of ENSO Monitoring Service
Start-up of Ocean Monitoring with the Operational Ocean Data Assimilation System (ODAS)Start-up of ENSO Forecast Service with the CGCM (JMA-CGCM)
The New Version of JMA’s Coupled Forecast System (MOVE-G/MRI.COM, JMA/MRI-CGCM)
Seasonal Forecast with a CGCM
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
3
Table of Contents
• Introduction of a new JMA’s coupled forecast system for ENSO forecast– New ocean assimilation system
(MOVE-G/MRI.COM)– New atmosphere-ocean coupled model
(JMA/MRI-CGCM)
• Forecast skill of ENSO and Asian summer monsoon
• Summary
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
JMA Forecast System for ENSO Prediction
Coupled model(JMA/MRI-CGCM)
Atmospheric model(JMA/MRI Unified AGCM)
Ocean model (MRI.COM)
Coupler
Ocean Data Assimilation System(MOVE/MRI.COM-G)
Atmospheric Analysis(JRA-25, JCDAS)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Ocean Data Assimilation System (MOVE-G/MRI.COM)
MOVE (Multivariate Ocean Variational Estimation) System
Usui et al. (2006)
• Ocean Model : MRI Community Ocean Model (MRI.COM)• 3D-VAR(T,S) Observation Data: T, S on GTS, SST (COBE-SST, Ishii et al. 2005), SSH (TOPEX/POSEIDON, JASON-1, ERS-1,2, ENVISAT)• Vertical Coupled Temperature-Salinity EOF modes Fujii and Kamachi(2003)
• Incremental Analysis Update (IAU)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Improvement of the thermocline variability
shade: dT/dzlines: isotherms of 20°C
The MOVE system represents the more realistic interannual variability of thermocline.TAO OBS.
MOVE-G CURRENT
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Comparison of Temperature Climatology
The temperature climatology of MOVE system is relatively close to the climatology of WOA01.
Forget, Heimbach, and Menemenlis (2006)CLIVAR/GODAE Meeting on Ocean Synthesis Evaluation
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Comparison of Salinity Climatology
Forget, Heimbach, and Menemenlis (2006)CLIVAR/GODAE Meeting on Ocean Synthesis Evaluation
The salinity climatology of MOVE system is relatively close to the climatology of WOA01.
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
New Coupled Model (JMA/MRI-CGCM)
AGCM JMA/MRI Unified AGCM
• TL95L40 (horizontal resol. ~ 180km)
OGCM MRI.COM
Ishikawa et al. (2005)
• 75S-75N, 0-360E
• horizontal resolution: lon 1.0° , lat 0.3-1.0°• vertical resolution : 50 levels
(23 levels in the upper 200m)
Coupler • coupling interval : 1 hour
• flux adjustment for heat and momentum flux
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of SST in NINO3.4
ACC (NINO3.4) RMSE (NINO3.4)
Period: 1979-20051 member forecast started from the end of every month(NOT ensemble forecast)
red: JMA/MRI-CGCMblue: current operational modelblack: persistence
(°C )
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of SST in WTP
ACC (WTP) RMSE (WTP)
Period: 1979-20051 member forecast started from the end of every month(NOT ensemble forecast)
red: JMA/MRI-CGCMblue: current modelblack: persistence
(°C )
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Experimental Setting of the Retrospective Seasonal Forecast
• 10-member ensemble forecasts are started from the end of January.(Two-tier AGCM forecasts are started from Feb. 10th.)
• Period of the retrospective forecast is 22 years (1984-2005).• CO2 concentration is updated during the retrospective forecast period.
• Initial data for land model are climatology.
JAN 10 membersAUG
7 months Jun.-Aug.
4-month lead
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of Precipitation (JJA, 4 months lead)
• 10-member ensemble forecast started from around the end of January.• Period of the retrospective forecast is 22 years (1984-2005).• JJA mean precipitation is verified with CMAP analysis.
The JMA/MRI-CGCM shows better skill than JMA’s two-tier operational model.
ACC (JMA/MRI-CGCM) ACC (Two-tier AGCM)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
SST-Precipitation Relationship in WTP in Boreal Summer
• Why CGCMs can predict precipitation better than AGCMs ?
Temporal correlation coefficients between JJA mean precipitation and JJA mean SST from (a) CMAP and COBE-SST analysis, (b)JMA/MRI-CGCM, (c)JMA two-tier operational model.
(a) Observation (b) JMA/MRI-CGCM (c) Two-tier AGCM
eg. Wang et al. 2004, Wang et al. 2005, Kitoh and Arakawa 1999, Kobayashi et al. 2005
Negative Correlation
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of Webster-Yang Index (JJA, 4 months lead)
JMA/MRI-CGCM
Two-tier AGCM
ACC=0.64
ACC=0.47
Webster-Yang index: U850(EQ-20N, 40-110E) –U200(EQ-20N, 40-110E)
Webster and Yang (1992)
10-member ensembleInit: the end of Jan.Period:1984-2005Ref. : JRA-25JJA mean
(m/s)
(m/s)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of East Asian Summer Monsoon (JJA, 4 months lead)
JMA/MRI-CGCM
Two-tier AGCM
ACC=0.58
ACC=0.04
DU2 index: U850(5-15N, 90-130E) -U850(22.5-32.5N, 110-140E)
Wang and Fan (1999)
10-member ensembleInit: the end of Jan.Period:1984-2005Ref. : JRA-25JJA mean
cf. Kug et al. (2007)
(m/s)
(m/s)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
ROC Score of 850 hPa Temperature (upper tercile, JJA, 4 months lead)
• 10-member ensemble forecast started from around the end of January.• Period of the retrospective forecast is 22 years (1984-2005).• JJA mean temperature at 850 hPa is verified with JRA-25.
ROC (JMA/MRI-CGCM) ROC (Two-tier AGCM)
ROC Area (tropics) : 0.59ROC Area (tropics) : 0.65
cf. Graham et al. (2005)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Summary and Future Plans
• The JMA/MRI-CGCM shows better forecast skill of Nino3.4 SST than the current operational ENSO forecast model. (The ACC is more than 0.7 at a lead time of 6 months.)
• The JMA/MRI-CGCM shows better forecast skill of the Asian summer monsoon circulation (W-Y index and DU2 index) and precipitation compared with the current operational two-tier system.
• The JMA’s new coupled forecast system for ENSO prediction will be operated in early 2008.
• JMA now plans to improve the JMA/MRI-CGCM as a seasonal forecast model and start one-tier forecast in 2010.
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Thank you for your kind attention.
JMA’s mascot “Harerun”
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
References
• Fujii, Y., and M. Kamachi (2003), Three-dimensional analysis of temperature and salinity in the equatorial Pacific using a variational method with vertical coupled temperature-salinity empirical orthogonal function modes, J. Geophys. Res., 108(C9), 3297, doi:10.1029/2002JC001745.
• Graham, R. J., M. Gordon, P. J. Mclean, S. Ineson, M. R. Huddleston, M. K. Davey, A. Brookshaw, R. T. H. Barnes (2005), A performance comparison of coupled and uncoupled versions of the Met Office seasonal prediction general circulation model, Tellus A, Volume 57, Number 3, May 2005 , pp. 320-339(20)
• Kitoh, A, O. Arakawa (1999), On overestimation of tropical precipitation by an atmospheric GCM with prescribed SST. Geophys. Res. Lett. 26:2965-2968
• Kobayashi, C., S. Maeda, A. Ito, Y. Matsushita and K. Takano (2005), Relation between SSTs and Predictability of Seasonal Mean Precipitation over the Western Tropical Pacific. JMSJ, Vol. 83, 919-929.
• Kug, J.-S., I.-S. Kang and D.-H. Choi (2007): Seasonal climate predictability with Tier-one and Tier-two prediction systems, Clim. Dyn., DOI 10.1007/s00382-007-0264-7
• Usui, N., S. Ishizaki, Y. Fujii, H. Tsujino, T. Yasuda, M. Kamachi (2006), Meteorological Research Institute Multivariate Ocean Variational Estimation (MOVE) System: Some early results. Adv. Space Res. 37, 806-822
• Wang, B. and Z. Fan (1999), Choice of South Asian summer monsoon indices. Bull. Amer. Meteor. Soc., 80, 629-638, 1999.
• Wang, B., I.-S. Kang and J. -Y. Lee (2004), Ensemble Simulations of Asian-Australian Monsoon Variability by 11 AGCMs. J. Climate, 17, 4, 803-818.
• Wang, B., Q. Ding, X. Fu, I.-S. Kang, K. Jin, J. Shukla, and F. Doblas-Reyes (2005), Fundamental challenge in simulation and prediction of summer monsoon rainfall, Geophys. Res. Lett., 32, L15711, doi:10.1029/2005GL022734.
• Webster, P. J. and S. Yang: Monsoon and ENSO (1992), Selectively interactive systems. Quart. J. Roy. Meteor. Soc., 118, 877-926.
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
22
History of ENSO Monitoring and Forecast Service at JMA
1992
1996
1999
2003
2008
2010
The Second Version of JMA’s CGCM (JMA-CGCM02)
Launch of ENSO Monitoring Service
Start-up of Ocean Monitoring with the Operational Ocean Data Assimilation System (ODAS)Start-up of ENSO Forecast Service with the CGCM (JMA-CGCM)
The New Version of JMA’s Coupled Forecast System (MOVE-G/MRI.COM, JMA/MRI-CGCM)
Seasonal Forecast with a CGCM
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
System design of MOVE/MRI.COM
Future
Past MRI.COM ( OGCM )prediction
first-guessObservation・ temperature・ salinity・ SSH
Coupled T-S EOF
Climatology
3DVAR
Gridded analysis fields
IAU
predictionMRI.COM
POpULar
Assimilation
Assimilation
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Vertical Coupled T-S EOF Modes
Domain
segmentation for
vertical coupled T-S
EOF modes
Vertical coupled
T-S EOF modes
for a red box region
Te
mp
era
ture
Sa
Iinit
y
1st 2nd 3rd 4th 5th
Fujii and Kamachi (2003)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Impact on the Warm Water Heat Content (WHC)
WHC: Heat
content in warm
water (>28 °C )WHC decreases in the thick barrier layer area
because of bad barrier layer expression.
BLT (Color) at Eq.WHC=10 (Black line)
Difference of WHC (cal/cm2) (With S correction – without S correction)
Thick barrier layer area has good correspondence with the area of large WHC.
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Salinity Impact on the Equatorial Undercurrent
With Salinity Correction 1993-2001 mean
Without Salinity Correction 1993-2001 mean
Eq-140W Sep96-Aug01 mean
Tao
With S
Without S
EUC is weak without S correction
The profile with S correction is closer to the observed profile
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of SST (JJA, started from JAN)
ACC (JMA/MRI-CGCM) ACC (Two-tier AGCM)
• 10-member ensemble forecast started from around the end of January.• Period of the retrospective forecast is 22 years (1984-2005).• JJA mean SST is verified with COBE-SST analysis.
MSSS (JMA/MRI-CGCM) MSSS (Two-tier AGCM)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of SST (DJF, started from JUL)
ACC (JMA/MRI-CGCM) ACC (Two-tier AGCM)
• 10-member ensemble forecast started from around the end of July.• Period of the retrospective forecast is 22 years (1984-2005).• DJF mean SST is verified with COBE-SST analysis.
MSSS (JMA/MRI-CGCM) MSSS (Two-tier AGCM)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of T2m (JJA, started from JAN)
ACC (JMA/MRI-CGCM) ACC (Two-tier AGCM)
• 10-member ensemble forecast started from around the end of January.• Period of the retrospective forecast is 22 years (1984-2005).• JJA mean T2m is verified with JRA-25.
MSSS (JMA/MRI-CGCM) MSSS (Two-tier AGCM)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
Forecast Skill of T2m (DJF, started from JUL)
ACC (JMA/MRI-CGCM) ACC (Two-tier AGCM)
• 10-member ensemble forecast started from around the end of July.• Period of the retrospective forecast is 22 years (1984-2005).• DJF mean T2m is verified with JRA-25.
MSSS (JMA/MRI-CGCM) MSSS (Two-tier AGCM)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
ROC Area of 2-m Temperature (JJA, upper tercile)
• 10-member ensemble forecast started from around the end of January.• Period of the retrospective forecast is 22 years (1984-2005).• JJA mean temperature at 2 m height is verified with JRA-25.
ROC Area (JMA/MRI-CGCM) ROC Area (Two-tier AGCM)
ROC Area (tropics) : 0.60ROC Area (tropics) : 0.64
cf. Graham et al. (2005)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
ROC Area of 2-m Temperature (DJF, upper tercile)
• 10-member ensemble forecast started from around the end of July.• Period of the retrospective forecast is 22 years (1984-2005).• DJF mean temperature at 2 m height is verified with JRA-25.
ROC Area (JMA/MRI-CGCM) ROC Area (Two-tier AGCM)
ROC Area (tropics) : 0.66ROC Area (tropics) : 0.68
cf. Graham et al. (2005)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
ROC Area of 850 hPa Temperature (JJA, upper tercile)
• 10-member ensemble forecast started from around the end of January.• Period of the retrospective forecast is 22 years (1984-2005).• JJA mean temperature at 850 hPa is verified with JRA-25.
ROC Area (JMA/MRI-CGCM) ROC Area (Two-tier AGCM)
ROC Area (tropics) : 0.59ROC Area (tropics) : 0.65
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
• 10-member ensemble forecast started from around the end of July.• Period of the retrospective forecast is 22 years (1984-2005).• DJF mean temperature at 850 hPa is verified with JRA-25.
ROC Area (JMA/MRI-CGCM) ROC Area (Two-tier AGCM)
ROC Area (tropics) : 0.67ROC Area (tropics) : 0.69
ROC Area of 850 hPa Temperature (DJF, upper tercile)
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
ROC Area of SLP (DJF, upper tercile)
• 10-member ensemble forecast started from around the end of July.• Period of the retrospective forecast is 22 years (1984-2005).• DJF mean temperature at 850 hPa is verified with JRA-25.
ROC Area (JMA/MRI-CGCM) ROC Area (Two-tier AGCM)
ROC Area (tropics) : 0.5ROC Area (tropics) : 0.6
Japan
NOAA 32nd Annual Climate Diagnostics and Prediction Workshop, October 22-26, 2007, Tallahassee, FloridaAgencyMeteorological
MeteorologicalResearchInstitute
ROC of T850 (JJA, upper tercile)
• 10-member ensemble forecast started from around the end of Jan.• Period of the retrospective forecast is 22 years (1984-2005).• DJF mean temperature at 850 hPa is verified with JRA-25.
ROC (JMA/MRI-CGCM) ROC (Two-tier AGCM)
top related