how well are southern hemisphere teleconnection patterns predicted by seasonal climate models?
DESCRIPTION
How well are Southern Hemisphere teleconnection patterns predicted by seasonal climate models? The return!!. Rosmeri P. da Rocha and Tércio Ambrizzi University of São Paulo, São Paulo, Brazil. EUROBRISA 2009 – Exeter, UK. Rossby Wave Theory. Basic Theory – Rossby (1939, 1945). - PowerPoint PPT PresentationTRANSCRIPT
How well are Southern
Hemisphere teleconnection
patterns predicted by
seasonal climate models?
The return!!Rosmeri P. da Rocha and Tércio Ambrizzi
University of São Paulo, São Paulo, Brazil
EUROBRISA 2009 – Exeter, UK
Rossby Wave Theory
The barotropic vorticity equation is:
0
Vy
Vx
Ut
UUU VV
02
xxU
t
tlykxiAe Re
22 lk
kkU
22 lk
Ucx
Basic Theory – Rossby (1939, 1945)
Assuming that
And defining the perturbed streamfunction ψ, we have:
Assuming the wave solution
We get: or
Some characteristics of Rossby waves are:• They propagate to the west• They are dispersiveThe group velocity is given by:
222
22
lk
k
kkc
xg
222
2
lk
kl
lc
yg
For a stationary wave (ω=0; c=0):
2222sKU
lkK
dydKk
Krs
s2
and
Playing with the equations, it is possible to define the ray path radius of curvature which is given by the simple expression
(Hoskins e Ambrizzi 1993)
Schematic Ks profiles and ray path refraction
(a)Simple refraction
(b) Reflection from a turning latitude YTL, at which Ks = k
(c) Reflection of all wavenumbers before a latitude YB at which * = 0
(d) Refraction into a critical latitude Y CL at which U = 0
(e) waveguide effect of a Ks maximum.
(Hoskins e Ambrizzi 1993)
Main teleconnection patterns obtained from observational analysis and numerical modeling - DJF
(Hoskins e Ambrizzi 1993)
observational analysis
numerical modeling
Main teleconnection patterns obtained from observational analysis and numerical modeling - JJA
(Ambrizzi et al 1995)
observational analysis
numerical modeling
DATA AND METHODOLOGY
• Climatological Data used : ECMWF/ERA40 – period 1982 – 2001
• ECMWF Coupled GCM – Hindcast Period – 1982 – 2001 – 11 ensemble members – 6 months forecasting
• The seasons are: JFM (Summer), AMJ (Fall), JAS (Winter), and OND (Spring)
• To create the seasonal datasets it was used the third month of each six months forecasting
• Pearson linear correlation was used in some of the analyzes
•The basic variables used in this presentation is Zonal (U) and Meridional Wind (V)
• Ray tracing analysis will be presented as well
Mean Seasonal Zonal Wind Cross Section at 50ºS
ERA40
Mean Seasonal Zonal Wind Cross Section at 30ºS
ERA40
SEASONAL MERIDIONAL WIND BIAS: PREV3 – ERA40(200 hPa)
BOXES TO BE USED IN THE CORRELATION ANALYSIS
SEASONAL ZONAL WIND BIAS (PREV3-ERA40) AT RS BOXIn general the signal of bias is the same for each member ensemble
SEASONAL MERIDIONAL WIND BIAS (PREV3-ERA40) AT RS BOX
RS: Temporal Distribution, Zonal Wind ERA40-Prev3, 20 years
05
10152025303540455055
1982Jan
1983Jan
1984Jan
1985Jan
1986Jan
1987Jan
1988Jan
1989Jan
1990Jan
1991Jan
1992Jan
1993Jan
1994Jan
1995Jan
1996Jan
1997Jan
1998Jan
1999Jan
2000Jan
2001Jan
Date
Zo
na
l w
ind
(m
/s)
ERA40 Prev3
NE: Temporal Distribution, Zonal Wind ERA40-Prev3, 20 years
-10-505
10152025303540455055
1982Jan
1983Jan
1984Jan
1985Jan
1986Jan
1987Jan
1988Jan
1989Jan
1990Jan
1991Jan
1992Jan
1993Jan
1994Jan
1995Jan
1996Jan
1997Jan
1998Jan
1999Jan
2000Jan
2001Jan
Date
Zon
al w
ind
(m
/s)
ERA40 Prev3
TIME SERIES OF THE ZONAL WIND AT RS AND NE(ERA40 and PREV3)
PREV3: mean of 11 members
Bar: maximum and minimum member value
SUMMER: ZONAL WIND CORRELATION (200 hPa) BETWEEN RS BOX AND ERA40, PREV3, THE WORST AND THE BEST ENSEMBLE MEMBERS
PREV3
WORST
BEST
ERA40
11 ENSEMBLE MEMBERS
SUMMER: MERIDIONAL WIND CORRELATION (200 hPa) BETWEEN RS BOX AND ERA40, PREV3, THE WORST AND THE BEST ENSEMBLE MEMBERS
11 ENSEMBLE MEMBERS
WORST
BEST
WINTER: ZONAL WIND CORRELATION (200 hPa) BETWEEN RS BOX AND ERA40, PREV3, THE WORST AND THE BEST ENSEMBLE MEMBERS
11 ENSEMBLE MEMBERS
WORST
BEST
WINTER: MERDIONALWIND CORRELATION (200 hPa) BETWEEN RS BOX AND ERA40, PREV3, THE WORST AND THE BEST ENSEMBLE MEMBERS
11 ENSEMBLE MEMBERS
WORST
BEST
SUMMER: ZONAL WIND CORRELATION (200 hPa) BETWEEN NE BOX AND ERA40, PREV, THE WORST AND THE BEST ENSEMBLE MEMBERS
11 ENSEMBLE MEMBERS
WORST
BEST
SUMMER: MERIDIONAL WIND CORRELATION (200 hPa) BETWEEN NE BOX AND ERA40, PREV3, THE WORST AND THE BEST ENSEMBLE MEMBERS
WORST
BEST
WINTER: ZONAL WIND CORRELATION (200 hPa) BETWEEN NE BOX AND ERA40, PREV3, THE WORST AND THE BEST ENSEMBLE MEMBERS
WORST
BEST
WINTER: MERIDIONAL WIND CORRELATION (200 hPa) BETWEEN NE BOX AND ERA40, PREV, THE WORST AND THE BEST ENSEMBLE MEMBERS
WORST
BEST
SEASONAL RAY TRACING ANALYSIS FOR WAVE NUMBER=2 (WN=2) (ERA40 AND ALL 11 MEMBERS)
SEASONAL RAY TRACING ANALYSIS FOR WN=3 (ERA40 AND ALL 11 MEMBERS)
summary
• The GCM is not able to correctly represent the position of the maximum and minimum hemispheric zonal wind (large variability among the ensemble members)
• There are considerable errors in the amplitudes of the SH Rossby waves reproduced by the ensemble mean, particularly during the summer and spring seasons
• The correlation maps suggests that there some ensemble members that reproduce quite well the zonal and meridional wind spatial pattern while there are others that completely fail to do this.
• Ray tracing analyzes clearly suggest that the model is not able reproduce the expected wave trajectory because it does not represent the Southern Hemisphere zonal wind variability.
FUTURE WORK
• Analyze the seasonal forecasts taking into account the first three months of the integration
• Repeat all previous analyzes for the Meteo Office and CPTEC hindcast data.
• Select some specific years to analyze the atmospheric circulation over South America in order to determine some dynamical aspects of the model ensemble members and their deviation.
GRUPO DE ESTUDOS CLIMÁTICOS
THANK YOU FOR YOUR ATTENTIONTHANK YOU FOR YOUR ATTENTION
CLIMATE STUDIES GROUP