Report from Japan Meteorological Agencyby Tomoaki OSE (Meteorological Research Institute / JMA)

• JMA finally started the operational seasonal prediction using the output of a couple model of the JMA/MRI-CGCM3 in Feb, 2010 instead of a two-way system. The skill for NINO3.4 SST ACC is comparable to the top models shown in Jin et al (2008). High predictability for East Asia climate through the Indian Ocean SST anomaly is expected.

•  

• A 4D-Var reanalysis project (JRA-55) is started at JMA, targeting the period of 1958-2012 with TL319L60 and 0.1hPa_top. Dry bias over the Amazon basin is improved as well as cold bias in the lower stratosphere, found in JRA-25 for 1979-2004.

•  

• Monitor and verification webpage of TIGGE data (http://tparc.mri-jma.go.jp/TIGGE/), including TIGGE MJO forecast (http://tparc.mri-jma.go.jp/TIGGE/tigge_MJO.html), is operated and updated every day.

• A new high-resolution MRI-AGCM with TL959L64 (20km) improves East Asia precipitation and tropical cyclones distribution.

Kiyotoshi TakahashiClimate Prediction Division/Japan Meteorological Agency

Performance of JMA atmosphere-ocean coupled model

1. Introduction

JMA-CGCMA:T42L21

O:2.5x(0.5-2.0)M6(LAF)Aug.1999

El-Niño prediction model

JMA-CGCM02A:T42L40

O:2.5x(0.5-2.0)M6(LAF)Jul.2003

JMA/MRI-CGCMA:TL95L40

O:1.0x(0.3-1.0)M12(LAF)Feb.2008

Seasonal forecast model

AGCMT63L40M31

Mar.2003

AGCMTL95L40M31

Mar.2006

JMA/MRI-CGCMA:TL95L40

O:1.0x(0.3-1.0)M51(9m,LAF)

Feb.2010

Flow of models for seasonal forecast and El-Niño outlook

Current system

SST

JMA/MRI-CGCMA:TL95L40

O:1.0x(0.3-1.0)M30(5m,LAF)

Feb.2009

AGCMTL95L40M51

Sep.2007

Jin E. K., James L. Kinter III, B. Wang, C.-K. Park, I.-S. Kang, B. P. Kirtman, J.-S. Kug, A. Kumar, J.-J. Luo, J. Schemm, J. Shukla and T. Yamagata, 2008: Current status of ENSO prediction skill in coupled ocean–atmosphere models. Clim. Dyn., 31, 647–666.

NINO.3.4 SST ACC: dependency on lead time (quote from Fig. 8 of Jin et al. 2008)

Initial: February (1980-2001)

0 1 2 3 4 5

1.0

0.8

0.6

0.4

0.2

0.0 0 1 2 3 4 5

Lead time (month)

1.0

0.8

0.6

0.4

0.2

0.0

(JMA/MRI-CGCM) Lead time (month)

Initial: August (1980-2001)

NINO.3.4 region: 120W-170W, 5S- 5N

Impact of IOBW on world climate (JJA)

Impact of above-normal IOBW on world climate in boreal summer

When above-normal IOBW SST persists through post-El Niño summer, impacts like the right panel are expected.

Expected atmospheric responses to warmer anomaly in IOBW SST.

（ Kelvin wave ）Ref. Xie .et al(2009)

Northern Indian Ocean Tropical Western Pacific

Comparison of monthly SST ACC SST prediction by new EPS systemPrescribed SST for current 2-tier system

Anomalycorrelation

3. Comparison of performance SST between current and New systems

improvement

SST ACC for Dec.(4 month lead), Initial: the last day of July

improvement

Lead time (month) Lead time (month)

SST ACC (1084-2005)SST ACC (1084-2005)

North Western Pacific monsoon are

Indian monsoon area

(average for 4 initials)

☺☹

3. Comparison of performance precipitation ACC: dependency on lead time

Precip. ACC 3 month mean (1984-2005)

Precip. ACC 3 month mean (1984-2005)

Lead timeLead time

Ensemble method for New EPS system• 9 members per 1 initial, 13 months forecast• Execution by every 5 day→ ＬＡＦ . 　 LAF for a month• 6 initials→51members are used for one-month forecast

-25day

0day1 initial date9 members

51 members from 6 initial dates

4. Summary

Yuhei Takaya (JMA/CPD)ytakaya@met.kishou.go.jp

South China Flood in June 2010

Precipitation in June 2010 (ratio to normals)

Zhejiang 174 %

Fuzhou 157 %

Based on CLIMAT reports

June 2010 forecasts (init:May) precipitation anomaly

Beijing ECMWF Exeter 　　　　　　　　Melbourne Melbourne

Montreal Moscow Seoul 　　　　　　　　　 Tokyo Tokyo

Toulouse Washington

For details, please refer to the WMO L Ｃ websitehttp://www.wmolc.org/

Observation; OLR anomaly(original data are provided by NOAA)

(mm/day)

JRA-55the Japanese 55-year reanalysis project

- status and plan -

Ayataka Ebita, Yukinari Ota, Shinya Kobayashi*, Masami Moriya, Ryouji Kumabe, Kiyotoshi Takahashi and Kazutoshi Onogi

Japan Meteorological Agency

Specifications of data assimilation system

JRA-25 (1979~2004) JRA-55 (1958~2012)

Resolution T106L40(top layer at 0.4 hPa)

TL319L60(top layer at 0.1 hPa)

Time integration Eularian Semi-Lagrangian

Long-wave radiation

Line absorptionStatistical band model

Water vapor continuume-type

Line absorptionTable lookup + K-distribution

Water vapor continuume-type + p-type

Assimilation scheme 3D-Var 4D-Var

(with T106 inner model)

B matrix ConstantDifferent B matrices for

pre-satellite and satellite eras

Bias correction(radiosonde)

Radiation bias only(Andrae et al., 2004)

RAOBCORE v1.4(Haimberger, 2007, J. Climate)

Bias correction(radiances) Offline Variational Bias Correction

Soil wetness in the root layer(Oct 1990 ~ Sep 1991)

JRA-25 JRA-55 JRA-55 minus JRA-25

(Saturation ratio [0~1]) (Saturation ratio [0~1]) (Saturation ratio)

6-hour precipitation and surface pressure increment

(Oct 1990 ~ Sep 1991)

Difference from GPCP precipitation (mm/day)

Surface pressure increment (hPa/day)

The dry bias over the Amazon basin is likely due to an artificial anticyclonic anomaly caused by bias of surface pressure observations and the consequent lack of precipitation.

Difference from GPCP precipitation (mm/day)

Surface pressure increment (hPa/day)

JRA-25 JRA-55 Exp

Summary

• JRA-55 improves upon JRA-25 in many respects,

– a longer reanalysis period, extending back in 1958,– much better forecast performance than JRA-25,– significantly reduced cold bias in the lower stratosphere, and– reduced dry bias over the Amazon basin.

• Quality of analysis changes inevitably due to changes in observing systems, but there is a good prospect that a reasonably homogeneous analysis will be produced in the northern hemisphere troposphere.

– Quality of analysis is reasonablely high over the regions that radiosondes cover even if no satellite data is available.

– On the other hand, it is anticipated that analyses of the pre-satellite era would degrade seriously in the southern hemisphere troposphere.

Summary (cont.)

• There is a considerable possibility that quality of analysis in the pre-satellite era will improve by tuning the background error.

– An experiment with a global constant scaling factor showed a small but positive result in the southern hemisphere.

• The JRA-55 production is planned to start in Jun 2010 and expected to complete by early 2013.

Use of TIGGE data

THORPEX Interactive Grand Global Ensemble

The Observing System Research and Predictability Experiment

under WWRP

To improve the accuracy of 1-day to 2 week high-impact weather forecasts

Ensemble forecast data from 10 global NWP centers

TIGGE: THORPEX Interactive Grand Global Ensemble

http://tparc.mri-jma.go.jp/TIGGE/

TIGGE: THORPEX Interactive Grand Global Ensemble

http://tparc.mri-jma.go.jp/TIGGE/tigge_MJO.html

High-resolution models

Previous

ＪＲＡCMAP

Seasonal Change of East Asia Precip

Lines ： Height Thickness （ 200hPa-500hPa ）Shade ：　 PrecipitationArrows ：　 850hPa Wind

Asian Summer Monsoon （ JJA ）

TRMM

NEW

20km

GPCP

Variability of Tropical Precipitation (JJA) 20kmEOF1 EOF2

Previous

NEW

Present climate simulation of TC formation &tracking distribution by 20km Atmos. Model (25 years)

後期モデル

前期モデル

観測

Considerably improving

Overestimating Improving a littleImproving

Improving Improving

： Observation： Earlier model： Updated model

Tendency of formation at more eastern locations

Western North Pacific

Observation(1979-2003) 20km Updated Model (1979-2003)

20km Earlier Model(1979-2003)Blue ： January - MarchGreen ： April- JuneRed ： July-SeptemberOrange ： October - December

※TC detection is adjusted so that total global number of formations is equal to that of observed number