Qin Zhang, Jon Gottschalck, Michelle L

1
Intra-seasonal Variability in the NCEP Global
Ensemble Forecast System

• Qin Zhang, Jon Gottschalck, Michelle LHeureux,
  Peitao Peng, Song Yang, Arun Kumar, and Wayne
  Higgins

Acknowledgments Thanks Kyong-Hwan Seo, Zoltan
Toth, Yuejian Zhu, Bo Cui for discussion and
contribution.
http//www.cpc.ncep.noaa.gov/products/people/wd52q
z/mjoindex/MJO_INDEX.html
2
1. Introduction

• ? CPC is actively developing additional
  MJO-related forecast tools to aid its operational
  mission for the monitoring, assessment, and
  prediction of the MJO and its associated impacts
• ? We apply the Wheeler and Hendon (2004) MJO
  identification methodology. Although this
  approach for identification of the MJO is
  well-established, it has only been recently that
  it has been applied to dynamical model data at
  several global operational centers
• ? A description of our approach, modifications
  necessary for practical realtime operational
  considerations, and initial verification
  statistics from both the GEFS and CFS data are
  illustrated.
• ? We present products using data from the Global
  Ensemble Forecast System (GEFS) and Climate
  Forecast System (CFS)
• Wheeler and Hendon, 2004 An All-Season Realtime
  Multivariate MJO Index Development of an Index
  for Monitoring and Prediction. Monthly weather
  Review,132, 1917-1932.

3
CLIVAR MJO Experimental Prediction Project
NCEP
PSD/NCEP
UK Met
CMC
http//www.cdc.noaa.gov/MJO/Forecasts/index_phase.
html
4
2. MJO Identification

• The following steps are used to filter
  variability to that associated with the MJO
• a. Remove the first 3 harmonics of the seasonal
  cycle
• b. Remove ENSO by using the SST1 index (Bureau
  of Meteorology Research Centre, Australia)
• c. Remove mean of last 120 days
• d. Perform multivariate EOF analysis (U850, U200
  and observed OLR) and determine EOFs using
  historical data from 1979-2005
• e. Realtime OLR and NCEP GDAS/CDAS data (U850 and
  U200) to project to multivariate EOF patterns

5
Realtime MJO Monitoring and Forecast
6
3. Application to Operational Forecasts

• ? Calculated from the NCEP Global Ensemble
  Forecast System (GEFS) 20 members
• ? Use bias-corrected GEFS U850 and U200 wind data
  and CDAS climatology for anomalies (Bo Cui et al.
  2007)
• ? 120 day mean is based on observations (the last
  120 days from the reanalysis)

7
Realtime MJO Monitoring and Forecast
After Bias correction
Before Bias correction
Big jump related to model bias
8
4. Example Products
OLR
U850
forecast 15 days
9
Realtime MJO Monitoring and Forecast
OLR
850hPa wind and Chi
10
5. Verification
An Example of GEFS Forecast Verification
CFS
8 days
11 days
GEFS
11 days
12 days
11
6. Summary and Future Work

• ? Several experimental MJO tools are available
  using GFS and CFS operational data and the GEFS
  indicates greater skill than the CFS over the
  limited validation periods
• ? These tools will be used as input to the weekly
  MJO update and Experimental Global Tropics
  Benefits/Hazards Assessment routinely produced at
  CPC
• ? These tools will be incorporated as part of
  CPCs MJO overall tool consolidation effort and
  collaboration with other operational centers as
  part of the US CLIVAR MJO working group
• ? For future work OLR bias correction,
  integration longer than 15 days, add uncertainty
  information

12
7. Monsoon Indices Forecast (CPCs global
monsoon group)
Webster-Yang Monsoon Index U850-U200 (40-110ºE,
EQ-20ºN)
Australian Monsoon Index U850 (110-150ºE,
2.5-20ºS)
13
7. Monsoon Indices Forecast (CPCs global
monsoon group)
14
7. Monsoon Indices Forecast (CPCs global
monsoon group)
Dynamic Indian Monsoon Index U850 (40-80ºE,
5-15ºN) - U850 (70-90ºE, 20-30ºN)
Western North Pacific Monsoon Index U850(100-130ºE
, 5-15ºN)-U850(110-140ºE, 20-30ºN)