Title: Seasonal Prediction Research and Development at the Australian Bureau of Meteorology
1Seasonal Prediction Research and Development at
the Australian Bureau of Meteorology
Guomin Wang With contributions from Harry
Hendon, Oscar Alves, Eun-Pa Lim and Claire
Spillman Centre for Australian Weather and
Climate ResearchA partnership between the
Bureau of Meteorology and CSIRO
2Outline
- POAMA (Predictive Ocean Atmosphere Model
for Australia) - Australian Rainfall Prediction
- Leeuwin Current Prediction
3POAMA Overview
-
- The Bureau Dynamical Seasonal Prediction System
POAMA - First version went operational in 2002
- A new version (POAMA1.5) became operational
recently and a newer version is in development - POAMA development evolves as part of Australian
Earth System Modelling project ACCESS - Webpage POAMA.BOM.GOV.AU
4POAMA Model Components
Atmospheric Model BAM T47L17 -gt T63L17 -gt
ACCESS(UKMOLSM)
3h
OASIS Coupler
time
Heat flux, P-E
Ocean Model ACOM2 lat/lon/lev0.51.5/2/25 -gt
AusCOM
5Hindcasts Design
- Control run initialized at 00Z on the first day
of each month, 1980-2006 - Extra 9 members initialized prior to control run
initial time in progressively 6 hours interval - Each hindcast is integrated for 9 months (lead
1-9)
6Skill Assessment ACC for SST and Heat Content
SST H300
1 3 5
7Skill Assessment ACC for SST Pacific Indian
Ocean Indices
Nino3.4 IOD
ACC RMS
8El Nino Vintage and Impact on Australian Rainfall
1997 2002
Correlation between Australian drought index and
SST
Wang and Hendon (2007)
9El Nino Classic vs Modoki
(from Weng et al. 2007)
- Classic El Nino
-
- Nino 3 Index SSTA over the Nino3 region
(210E-270E, 5S-5N)
- El Nino Modoki
-
- EMI SSTACentral (0.5SSTAEast
0.5SSTAWest)
10El Nino Skill Classic vs Modoki
11R 0.86 at LT1
- Classic El Nino events 82, 87, 97
R 0.83 at LT1
- El Nino Modoki events (EMI gt 0.7 STD) 86, 90,
91, 93, 94, 02, 04
12SST Forecast Composites Classic El Nino
El Nino Modoki
OBS (SON)
POAMA LT 1 (1st Sep Start) POAMA LT 3 (1st Jun
Start)
13Australian Rainfall Forecast Composites Classic
El Nino El Nino Modoki
OBS (SON)
POAMA LT 1 POAMA LT 3
14Cases 1997 vs 2002
1997
2002
15Seasonal Prediction of the Leeuwin Current
Observed Features
Freemantle sea level (FSL) is indicative of
volume transport variation of the leeuwin current
(M. Feng). Use FSL as a proxy for Leeuwin
Current strength.
16Annual Mean of SST top 300m Currents from
Reanalyses
POAMA SST and UV300 clim 1982-2003
GODAS SST and UV300 clim 1982-2003
ECOR SST and UV300 clim 1982-2003
Eq 20ºS 40ºS
90ºE 120ºE
90ºE 120ºE
90ºE 120ºE
POAMA GOGAS/NCEP
ODA/ECMWF
17Fremantle Sea Level and Ocean Heat Content
Observation vs Forecast Skill
Obs relationship between H300 and SLA at Freo
HCNW 15-25ºS,112-120ºE
H300 ACC Skill at leadtime7
18Fremantle Sea Level and SST Observation vs
Forecast Skill
N34 5ºS-5ºN 170º-120ºW
19Downscaling POAMA Forecasts to Fremantle SLA
20Skill of Fremantle SLA Prediction from
Downscaling Scheme
Nino4 NWHC Both Combined Persist
21FSLA Forecasts 1982-2006
FSLA Obs FSLA Lead 3 FSLA Lead 6 FSLA Lead 9
82 83 84 85 86 87 88 89 90
91 92 93 94 95 96 97 98 99
00 01 02 03 04 05 06 Years
22- Summary
- Introduction of The Australian Bureaus Dynamical
Seasonal Prediction System POAMA. - POAMA has higher skill for SST in Pacific and for
heat content along NW Australia. - POAMA can predict short term El Nino vintage and
respective Australian rainfall responses. - Using POAMA forecasts a downscaling scheme shows
useful seasonal forecast skill for Leeuwin
Current strength.
23Summary
- Each El Nino event has different flavour
- The impact of the central Pacific warming El Nino
(represented by El Nino Modoki Index) is as
important as the traditional eastern Pacific
warming El Nino for Australian rainfall - POAMA has good skill to predict
- - the occurrence and the detailed SST
structure of the central Pacific El Nino and the
traditional El Nino events - - the Australian rainfall difference affected
by these two types of El Nino events - - 97 and 02 El Nino events and associated
Australian rainfall ? the skill stems from the
improved atmospheric initial conditions by ALI
and the models atmosphere-ocean coupling
ability.