1. GPC-Seoul: Status and future plans
WMO CBS ET-OPSLS, Exeter, UK, March 2014
GPC-Seoul: Status and future plans
Suhee Park
Korea Meteorological Administration

2. Contents Current seasonal forecasting system and products
New seasonal forecasting system: GloSea5
Description of GloSea5
General features of operational system
Results of hindcast and forecast
Summary and future plans

3. Current seasonal forecasting system & products
The operational LRF system of GPC-Seoul is a 2-tier system.
GPC-Seoul provides forecasts of 6 variables to WMO LC-LRFMME and verification results to WMO LC-LRFSVS.
Global SST Prediction Model
LRF Global Dynamic model
Initial Condition
monthly : OISST, UV850, UV200, T850,Z500, SLP, Prcp.
6 hourly : U, V, T, GPH, Psfc, MSLP, RH
El-Nino Prediction Model
CPPM(Coupled Pattern Projection Model)
LLRM(Lagged Linear Regression Model)
Persistence Model
GDPAS(GCM T106L21)
230 day forecast
Ensemble size : 20 members (fcst&hcst)
Hindcast period :
Boundary Condition
Nino 3.4 Index
Global SST forecast
Monthly forecast : T, GPH, RH, Precip., MSLP, U, V
Verification : T2m, Z500, Z850, MSLP, Precip.
 2m surface air temperature
 Precipitation
 Mean sea level pressure
 850hPa temperature
 500hPa geopotential height
 Sea surface temperature

4. New seasonal forecasting system: Joint system
June 2010, the collaboration agreement was made between KMA and UK Met Office for a joint seasonal forecasting system.
To establish a joint seasonal forecasting system with the joint aim of developing and providing operational seasonal forecasting products.
Joint System
Products sharing
Joint products
KMA product
UKMO product
Major advantages are to share ensemble members as many as possible.
The Joint system will be launched by December 2014.

5. ORCA tri-polar grid at 0.25°
Joint seasonal forecasting system : GloSea5
Joint system, GlosSea5, is the fifth version of the Met Office ensemble prediction system for seasonal forecasting, which has been built based on the latest version of HadGEM3. UM
(Atmosphere)
NEMO-CICE
(Ocean & Sea Ice)
OASIS3
(coupler)
Atmosphere
Ocean & Sea Ice
Model UM
(UM 8.0)
NEMO-CICE
(NEMO 3.2, CICE 4.1)
Horizontal resolution
N216 (0.83° x 0.56°)
ORCA tri-polar grid at 0.25°
Vertical resolution
85 levels (~85km)
75 levels
UM (Met Office Unified Model) for Atmosphere
JULES (Joint UK Land Environment Simulator) for Land Surface
NEMO (Nucleus for European Modeling of the Ocean) for Ocean
CICE (Los Alamos National Laboratory) for Sea-ice
OASIS (CERFACS) for coupling between component models

6. GloSea5: Initialization of the system
Forecast (initialized daily):
- Atmosphere & land surf: KMA NWP analysis (4d-Var)
UKMO system uses Met Office NWP analysis!
- Ocean & sea-ice: Met Office NEMOVAR analysis (3d- Var joint system for ocean, med-range, monthly and seasonal)
Hindcast (initialized weekly):
- Atmosphere & land surf: ERA-Interim reanalysis
- Ocean & sea-ice: Seasonal ODA reanalysis
- Fixed start dates of 1st, 9th, 17th, 25th of each month
- Hindcast period : 14 years ( )
Period will be extended for 25 years ( )

7. GloSea5: Generation of ensemble member
Seasonal Forecast:
- 2 members run each day.
- Seasonal forecast updated monthly by pulling together
last 3 weeks (i.e. 42 members)
Monthly Forecast:
- 2 additional members run each day.
- Monthly Forecast updated weekly by pulling together
last 7 days (i.e. 28 members)
Hindcast (for monthly-seasonal):
- 3 members run each week in real time (14 years per member)
- Hindcast for seasonal forecast updated monthly by pulling together last 4 weeks (i.e. 12 members)
Hindcast for measonal forecast updated monthly by pulling
together last 2 weeks (i.e. 6 members)

8. Hindcast result : GPC-Seoul vs GPC-Exeter
Bias and correlation of 500hPa geopotential height
KMA
UKMO
Bias
Corr

9. Forecast result : GPC-Seoul vs GPC-Exeter
Mean and anomaly of 500hPa geopotential height
KMA
Observation
UKMO

10. Summary
KMA and UKMO are trying to establish a joint seasonal forecasting system (GloSea5) with the joint aim of developing and providing operational seasonal forecasting products.
Although both GPC-Seoul and GPC-Exeter are using same forecasting model, they use different initial conditions (ICs) for the atmospheric and land surface model. Also, a forecasting model utilize a stochastic physics.
Different ICs and a stochastic physics lead to the significant regional-scale diversity in products between GPC-Exeter and GPC-Seoul.

11. Future plans
From June 2014 the operational forecast model of KMA will be replaced by the GloSea5, a join seasonal forecasting system with Met Office.
All products of KMA will be provided with GloSea5 to WMO LC-LRFMME for LRF, ECMWF for S2S.

12. Thank you very much!