Intra-seasonal and inter-annual variability in the BMRC atmospheric GCM

1
Intra-seasonal and inter-annual variability in
the BMRC atmospheric GCM
- Andrew Marshall
2
Acknowledgements
Dr. Oscar Alves
Supervisor, BMRC
Prof. David Karoly
Supervisor, Monash
BMRC
(Monash Departmental PhD Scholarship)
NCEP reanalysis data
(NOAA-CIRES Climate Diagnostics Centre, Boulder,
Colorado) http//www.cdc.noaa.gov/
3
Plan
? Madden Julian Oscillation and the TOGA COARE IOP
? El Niño Southern Oscillation and the 1997/98
extreme event
? Is there a link between the MJO and ENSO?
? How do GCMs perform with MJO simulations and El
Niño forecasts?
? The BAM T47L17 model - does it capture the MJO?
? Summary
4
Madden-Julian Oscillation (intra-seasonal
variability)
5
What is the Madden-Julian Oscillation?
1971 Roland Madden and Paul Julian analysed 10
years of station pressure and upper air data from
stations in the tropics.
Detection of intra-seaonsonal variability in the
40-50 day period range at stations in the tropics
(Madden and Julian, 1972)
6
? Significant peaks in the pressure and zonal
wind spectra were found in the 40-50 day period
range, revealing an eastward-moving disturbance
whose characteristics change with time.
? Analysis of the pressure data found the
disturbance to originate in the Indian Ocean,
propagate eastward into the central Pacific
Ocean, and decay near the dateline (180
longitude) over colder sea surface temperatures
(SSTs)
? We conclude that the 40-50 day oscillation is
not a temporary phenomenon in the equatorial
Pacific occurring only in the late 1950s and the
1960s but one that has appeared at other times.
Moreover, the oscillation apparently is not
highly tuned to a specific frequency. (Madden
and Julian, 1972)
7
? The oscillation manifests itself as a slow
eastward propagation of atmospheric disturbances
with maximum amplitudes in the eastern hemisphere
(Hendon and Salby, 1994)
? The oscillation has also been related to
precipitation fluctuations in the Indian Ocean
(Raghavan et al., 1975) and to active and break
periods of the Indian, Australian and Asian
summer monsoons (Yasanari, 1979 Wang and Rui,
1990)
? The MJO is the strongest signal so far found in
the intra-seasonal variability of the tropical
atmosphere.
8
Observations of the MJO -
TOGA COARE (Nov 1992 - Feb 1993)
Two pronounced MJO events associated with super
cloud clusters and westerly wind bursts were
observed (Yanai et al, 2000).
Hovmöller plot of ERA zonal wind at 850 hPa
averaged between 5S and 5N. The 170 Wm-2
contours of OLR are also shown. (Yanai et al.,
2000)
9
El-Niño Southern Oscillation (inter-annual
variability)
10
The El Niño Southern Oscillation
The ENSO phenomenon is known to be the single
most prominent signal in the inter-annual
variability of the earths climate.
Pacific Ocean El Niño versus normal conditions
11
Past ENSO events
SST anomalies in Nino 3 region, 1982 - 2001
12
The 1997/98 El Niño event
1997/98 El niño (Nino 3 - eastern Pacific Ocean)
13
Is there a link between the MJO and ENSO?
14
MJO and ENSO
Do the two phenomena have a definite link?
Onset of an El Niño event - triggered by MJO?
Need to determine the relationship between MJO
and El Niño
Traditionally they had been considered as
separate events
Kelvin waves play an important role in MJO
dynamics
New suggestions that oceanic Kelvin waves may be
related to ENSO
Oscillation may act to intensify ENSO through
air-sea interaction
Westerly wind bursts - do they help to shift
convection eastward?
15
westerly wind bursts produce Kelvin waves
Observed westerly wind bursts prior to the 97/98
El Niño event
16
Intra-seasonal and interannual variability in GCMs
17
Simulating the MJO in GCMs
Uncoupled AGCMs
? Simulations with prescribed SSTs typically
produce MJOs that move eastward too fast, are too
weak, and have incorrect seasonality
? SST fluctuations are coherent with the MJO,
suggesting that air-sea coupling is important
(Hendon, 2000)
Coupled GCMs
? Coupling has alleviated some problems
associated with uncoupled simulations of the MJO,
but the ability of a GCM to simulate the MJO
dynamics appears more to be determined by the
model physics than the presence/absence of
air-sea coupling in the model.
18
Current ENSO prediction skill
BMRC new atmosphere model SST forecasts
? El Niño prediction important - impacts worldwide
? Dynamical models currently unable to generate
significant warming episodes (ie. 97/98 El Niño)
19
The BAM T47L17 atmospheric GCM
? BMRC uncoupled atmosphere model
17 levels
Moisture-convergence convection closure
Fixed SSTs
? AMIP-type run
1982
2000
Daily OLR, u850
20
Inter-annual variability
Standard Deviation (variance1/2) of OLR over
Pacific Ocean, 1982-2000
21
Power spectrum of OLR averaged 160W - 140W, 10N -
10S.
22
with intra-seasonal variability
Standard Deviation (variance1/2) of OLR over
Pacific Ocean, 1982-2000
23
Power spectrum of OLR averaged 60E - 80E, 10N -
10S.
24
Time series of OLR anomaly averaged 60E - 80E
lon, 10N - 10S lat.
25
Hovmöller plot of equatorial OLR anomaly, 1982-83
26
Power spectrum of surface u wind averaged 60E -
80E, 10N - 10S.
27
Time series of surface u wind anomaly averaged
60E - 80E lon, 10N - 10S lat.
28
Hovmöller plot of equatorial surface u wind
anomaly, 1982-83
29
Future research
How well does BAM T47L17 capture the MJO?
How well does the coupled model represent the MJO?
How to use the MJO for seasonal forecasting?
What role does the MJO play in ENSO?
30
Summary
BAM T47L17 model - high variance in daily OLR
data over 19 year period compared with the NCEP
re-analysis record (approx. twice as much
variability).
Bad NCEP OLR data?
(Variability comparable in surface u wind data)
BAM OLR and surface u wind power spectra - tend
to peak around 30 and 60 days on intra-seasonal
timescales, whereas NCEP spectra peak around 45
days.
Further work must be done to establish whether
BAM captures the MJO...
Extend u wind analysis and investigate other
equatorial regions ie Niño 3 and Niño 4
31
The End!
32
Figure 6 First principal components of OLR over
Pacific Ocean, 1982-2000
33
Figure 7 Time series of first principal
component, 1982-2000