Circumglobal Wave Packets and
Middle East precipitation: Dynamics
and Predictability
Steven B. Feldstein
Department of Meteorology, The Pennsylvania State University,
University Park, Pennsylvania, U.S.A.
Presented at Tel Aviv University, Tel Aviv, Israel on May 11, 2010
Middle Eastern Precipitation

Is Middle East precipitation associated with the
variability of a particular teleconnection pattern?
Middle Eastern precipitation
Data and methodology

Data: daily precipitation data averaged over 12 sites in Israel.
(Ziv et al. 2006, Quart. J. Roy. Meteorl. Soc.)

Calculate composite 300-hPa geopotential height field for dates with extreme
precipitation

Daily SL (Southern Levant) index obtained by projecting the daily 300-hPa
geopotential height field onto composite pattern
Feldstein and Dayan (2008)
Middle Eastern precipitation
Composite 300-hPa geopotential height field:
Southern Levant (SL) pattern
H
L
Feldstein and Dayan (2008)
300-hPa geopotential evolution - Middle Eastern precipitation
-6 days
0 days
+5 days
-4 days
+2 days
+7days
-2 days
+4 days
+9 days
Feldstein and Dayan (2008)
Circumglobal Teleconnection Pattern
EOF1
Wave packets associated with SL precip
wet
dry
Time-averaged
V 300 over
persistent event
(lag -6 to lag +9 days)
Correlation with EOF1 =0.83
Correlation with EOF1 =-0.72
Wave packet evolution & potential vorticity gradient
-6 days
0 days
+5 days
-4 days
+2 days
+7days
-2 days
+4 days
+9 days
Feldstein and Dayan (2008)
Wave Packet & Middle East precip

Wave packet first observed in the northeast Pacific. The packet travels 3/4 of the
distance around the earth before decaying over the northwest Pacific

Wave packet amplifies as it passes over western Europe and the Middle East.
This coincides with enhanced precipitation over the Israel.

Wave packets closely associated with east Asian monsoon (e.g., Ding and
Wang 2005).

These wave packets related to the circumglobal teleconnection pattern
(Branstator 2002).

Questions: What processes account for the formation and decay of
circumglobal wave packets (CWPs)?

Why do the wave packets have an eastward group velocity with a near zero
phase velocity? Why are the wave packets dominated by zonal wavenumber
5?
Numerical Model
Numerical Model
 Dissipated by surface friction and 8th order hyperdiffusion
-A dynamic core of GFDL GCM (Gordon and Stern 1982)
-Driven
by relaxing
towardnumber
Te with15
timescale of 30 days, R30

R30L10
but zonalT wave
-Dissipated by surface friction and 8th order hyperdiffusion ipated by
surface friction and 8th order
Te(C,H) = Tbase + ΔTe(C,H); Te(C,H) is independent of longitude
Te(
→ control the baroclinic zone
→ control the strength of STJ : highlatitude cooling (K/day) H : tropical heating (K/day)
Zonal wind response to C and H
A
B
250-hPa [u]
A
250-hPa [u]
B
Eddy-driven jet
Subtropical jet
Son and Lee (2005)
EOFs From Idealized Climate Model Runs
Anomalous 300-hPa meridional wind (CS1 Run)
Phase Speed of Model Runs
CS1-CS6 experiments yield wrong latitude for CWP, sometimes the wrong
zonal wavenumber (k=4,6), and the phase speed is too large.
Reduce midlatitude baroclinicity, which weakens the eddy-driven jet, and strengthens
the subtropical jet. (MODIFIED CS1 RUN)
MODIFIED CS1 EOF1 & EOF2 (300-hPa meridional wind
\
Composite Methodology
A Circumglobal Teleconnection Pattern (CTP) Event:
10-day, low-pass filtered, CTP amplitude must (a) exceed a one standard
deviation threshold for (b) a minimum number of 15 consecutive days
Motivated by the autocorrelation function for the CTP amplitude
time series
MODIFIED CS1: 300-hPa Meridional Wind Composites
Cg eastward
Cp near zero
K=5
k=5
CTP near 30N
CTP and non-CTP contributions to eddy kinetic
energy during CTP event
Energy fluctuation mostly associated with CTP
Non-CTP 300-hPa meridional wind
Larger Cp
7-day-period
relative to
CTP
Anomalous 300-hPa eddy momentum flux [u’v’]
Anomalous 300-hPa total flux [u’v’]
CTP/nonCTP flux:
Constructive
Interference
Total
flux
Anomalous 300-hPa total flux [up’vp’]
CTP
flux
Anomalous 300-hPa total flux [unp’vp’]
Anomalous 300-hPa total flux [unp’vnp’]
Non-CTP
flux
7-day CTP amplitude fluctuation due to interaction between CTP & non-CTP waves
Anomalous 850-hPa eddy heat flux [v’T’]
Anomalous 300-hPa total flux [vp’Tp’]
Anomalous 300-hPa total flux [vnp’Tnp’]
CTP
flux
Non-CTP
flux
Anomalous 300-hPa total flux [v’T’]
Total
flux
Slow steady growth
of CTP due to CTP
eddy heat fluxes
Anomalous E-P flux cross-sections
Lag -4 days
Lag -2 days
Lag -3 days
Lag -1 days
Anomalous 300-hPa zonal-mean zonal wind
Anomalous heat flux maxima coincide the zonal wind maxima
Anomalous zonal wind driven by the CTP/non-CTP eddy momentum flux
EOF1 and Composite PC1 (zonal wind)
EOF1 Annular Mode
Composite PC1
Eddy heat flux strongest when PC1 most negative
Overall Picture
Interaction between CTP and non-CTP eddies drives fluctuations in
zonal mean zonal wind (between the negative NAM and the climatology).
When the subtropical/eddy-driven jet is strengthened and displaced
equatorward (negative NAM), the CTP grows baroclinically
(baroclinic instability with Cp=0, Cg>0, k=5?) When the jet is near
its climatology, the baroclinic growth ceases.
CTP decay coincides with an increase in the lower tropospheric
zonal wind shear which suppresses subsequent baroclinic growth
Barotropic governor (James and Gray 1986; Moon and Feldstein 2009)?
WHAT ARE THE IMPLICATIONS FOR THE PRECIPATION IN ISRAEL?
Implications for precipitation in Israel?
Interaction between small amplitude disturbance on the subtropical jet over
North Africa and extratropical eddies over Europe intensifyies the
subtropical jet and drive the subtropical jet equatorward.
The above process results in a background flow which is favourable for
circumglobal wave packet growth via baroclinic instability. This
process can be examined with 3-D wave activity flux vectors.
The decay of the circumglobal wave packet coincides with an increase in the
lower tropospheric zonal wind shear .
16 North Pacific sea level pressure cluster patterns
Tropical Convection Associated with
the Madden-Julian Oscillation (MJO)
Phase 1




Dominant intraseaonal
oscillation in the tropics
Phase 2
MJO cycle: 30-60 days
Phase 3
Shading OLR
Phase 4
Time between phases ~
6 days
Phase 5
Phase 6
Time between Phases ~
6 days
Phase 7
Phase 8
From Wheeler and Hendon (2004)
20۫°E
180۫°
From Wheeler and Hendon (2004)
60۫°W
Frequency of occurrence for each cluster pattern and MJO phase
1-7 day Forecast of Anomalous Precipitation in Israel
Phase Number = location in
Israel
Lag = 1 to 7 days (Feldstein
and Dayan (2008)
Pattern Number = cluster pattern
Color denotes anomalous
precipitation determined
from composites for each
pattern number
Conclusions
Precipitation in Israel strongly influenced by circumglobal wave packets
Circumglobal wave packet growth triggered by the interaction between
CTP and non-CTP eddies which alters the subtropical jet toward a structure
that favors baroclinic instability. Circumglobal wave packet decays via the
barotropic governor?
Based on ideas of the CTP, and the continuum perspective, one may be able to
develop a probabilistic 7- day forecast model of precipitation in Israel
The forecast model can be extended to a multimodel ensemble which
includes the observation features of the cluster model (a Bayesian approach).
F = (w1F1 + w2F2 + w3F3)/(w1 + w2 + w3)
Fi are the model forecasts
B= model verification (observations)
A= model forecast
The weights are determined by wi = P(B|A) =P(A|B)*P(B)/P(A)