Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

21
Sensitivity of the position and Sensitivity of the position and variability of the eddy-driven variability of the eddy-driven jet and storm-track to different jet and storm-track to different SST profiles in an aquaplanet SST profiles in an aquaplanet general circulation model general circulation model (Arpège) (Arpège) Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France EGU Vienna - April 25, 2012

description

Sensitivity of the position and variability of the eddy-driven jet and storm-track to different SST profiles in an aquaplanet general circulation model (Arpège). Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France. EGU Vienna - April 25, 2012. - PowerPoint PPT Presentation

Transcript of Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Page 1: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Sensitivity of the position and Sensitivity of the position and variability of the eddy-driven jet and variability of the eddy-driven jet and

storm-track to different SST profiles in storm-track to different SST profiles in an aquaplanet general circulation an aquaplanet general circulation

model (Arpège)model (Arpège)

Clio Michel and Gwendal Rivière

Météo-France/CNRS, CNRM-GAME, Toulouse, France

EGU Vienna - April 25, 2012

Page 2: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Intensity of the subtropical jet (SJ).

•Eddy-driven jet (EDJ) and storm-track position and intensity controlled by:

Anchoring effect of EDJ and storm track by the SST front.

(Nakamura et al. 2004, Brayshaw et al. 2008, Ogawa et al. 2011)

(Lee and Kim 2003, Son and Lee 2005)

•EDJ climatology intensively studied but much less the EDJ variability (linked to teleconnections such as AO, NAO).

A strong SJ leads to a single jet, a weak SJ leads to two jets.

Literature and problematicLiterature and problematic

Close but on the poleward side of the SST front.

Page 3: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

ProblematicProblematic

Goal: Study of the eddy-driven jet variability using an aquaplanet AGCM

Approach: •Aquaplanet simulations using AGCM Arpège-Climat (T127L31) with prescribed axisymmetric SST at the lower boundary and equinoctial conditions.

•Systematic analysis of the storm-track and eddy-driven features (latitudinal position, intensity) relative to the SST gradient features (latitudinal position, intensity and width).

Page 4: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

SST profilesSST profiles

Idealized piecewise linear distributions to only modify midlatitudes.

3 latitudes, 2 widths, 2 SST gradient intensities 12 experiments

ERA40 SST Atl-

DJF

40°

20°

Page 5: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

The two jet regimesThe two jet regimes

One single jet Two well-separated jets

SST gradient at=30° SST gradient at=50°

Mass streamfunction (red and blue contours), zonal wind (shading)

y

p

v

Page 6: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Higher the gradient to the south, stronger the storm-track. Higher the gradient intensity (intensity and width),

stronger the strom-track.

Storm-track intensityStorm-track intensity

(Brayshaw et al. 2008, Graff and LaCasce 2012)

wide weak front

wide strong front

narrow weak front

narrow strong front

Page 7: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Baroclinicity (Meridional gradient of Baroclinicity (Meridional gradient of ))

Baroclinicity linked to SJ

Baroclinicity linked to SST

gradient

When the SST gradient is more equatorward, the two zones of baroclinic waves growth merged, the two types of baroclinicity add and the storm-track is stronger.

SST gradient at=30° SST gradient at=50°

Page 8: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Eddy-driven jet latitude Eddy-driven jet latitude (Umax at 850 hPa)(Umax at 850 hPa)

EDJ located on the poleward side of the SST gradient because of the predominance of AWB.

More poleward the gradient, more poleward the EDJ relative to the SST front.

One-to-one lineweak gradientstrong gradient

Page 9: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Wave-breaking asymmetriesWave-breaking asymmetries

anticyclonic WB

cyclonic WB

AWB = CWB AWB >> CWB

U U

SST gradient at=30° SST gradient at=50°

SST SST

WB event: reversal of the absolute vorticity gradient.

Page 10: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Interpretation in terms of refractive index Interpretation in terms of refractive index asymmetriesasymmetries

222

2

22

22 sin

cos4cos

am

NHf

cauyq

n

n2<0: evanescent waves. n2>0: waves propagation.

AWB

AWB

CWB

(Rivière 2009)Proposed interpretation:

SST gradient more poleward WB asymmetries (much more AWB) EDJ much more poleward relative to the SST gradient.

Page 11: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Eddy-driven jet variabilityEddy-driven jet variability

SST gradient at=30° SST gradient at=60°

node

U max

Only latitudinal shifting Mixed pulsing and latitudinal shifting

U regressed on PC1 (black contours) and U (shading)EOF1: first EOF of the vertically and zonally averaged zonal wind

node

U max

PC1>1PC1<-1U

PC1>1PC1<-1

U

Page 12: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

ERA40 reanalysis (JJA – Southern ERA40 reanalysis (JJA – Southern hemisphere)hemisphere)

Pacific 55°S 120°E-210°E

Indian 45°S 30°E-120°E

Meridional SST gradient

South Pacific SST gradient poleward the South Indian SST gradient

Page 13: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

ERA40 reanalysis (JJA - SH)ERA40 reanalysis (JJA - SH)

Pacific ocean 120°E-210°E

nodes

U regressed on PC1 (black contours) and U (shading)SST gradient at=55°S

Indian ocean 30°E-120°ESST gradient at=45°S

PC1>1

PC1>1 PC1<-1PC1<-1

UU

U max

U max

EOF1: first EOF of the vertically and zonally averaged zonal wind

Latitudinal shifting Mainly pulsing

Page 14: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

ConclusionsConclusions

•We find two regimes of EDJ variability: latitudinal fluctuations (middle and low latitudes) and pulsing (highlatitudes), similar to the difference between the observed SH Pacific and Indian oceans. (similar to barotropic simulations of Barnes and Hartmann 2011)

Page 15: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

ConclusionsConclusions

•We find two regimes of EDJ variability: latitudinal fluctuations (middle and low latitudes) and pulsing (highlatitudes), similar to the difference between the observed SH Pacific and Indian oceans. (similar to barotropic simulations of Barnes and Hartmann 2011)

•Intensification of the storm-track for SST gradient closer to the subtropical jet by addition of the two sources of baroclinicity.

•The position of the eddy-driven jet relatively to the SST gradient can be explained by wave-breaking asymmetries.(consistent with more idealized simulations of Rivière 2009)

Page 16: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

ConclusionsConclusions

•We find two regimes of EDJ variability: latitudinal fluctuations (middle and low latitudes) and pulsing (highlatitudes), similar to the difference between the observed SH Pacific and Indian oceans.

Stronger subtropical jet. (by increasing SST in the subtropics) Zonally asymmetric SST forcing.

(similar to barotropic simulations of Barnes and Hartmann 2011)

•Intensification of the storm-track for SST gradient closer to the subtropical jet by addition of the two sources of baroclinicity.

•The position of the eddy-driven jet relatively to the SST gradient can be explained by wave-breaking asymmetries.(consistent with more idealized simulations of Rivière 2009)

•Perspectives:

Page 17: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

End of talkThank you for your attention

Page 18: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

SST profilesSST profiles

•Idealized piecewise linear distributions to only modify midlatitudes 12 experiments:

Solid lines L=20°

Dashed lines L=10°

Blue lines weak SST gradient=-8.54 10-6 K/mBlack lines strong SST gradient=-10.79 10-6 K/m

3 latitudes: 30°,40°,50°

Page 19: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Rossby wave breakingsRossby wave breakings

RWB = large-scale and irreversible overturning of the PV gradient over an isentropic surface.

Two types (Thorncroft et al., 1993):

anticyclonic (SW-NE) cyclonic (SE-NW)

• RWB detection algorithm (based on geometrical considerations)

Page 20: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Refractive index asymetries Refractive index asymetries (Rivière 2009)

222

2

22

22 sin

cos4cos

am

NHf

cauyq

n

n2 increases with

NW-SE orientation, poleward

propagation, u’v’<0 CWB

ū

φ

φ0

n2 decreases with

φ0

φ

ū

perturbation

SW-NE orientation, equatorward

propagation, u’v’>0 AWB

AWB

AWB

CWB

Page 21: Clio Michel and Gwendal Rivière Météo-France/CNRS, CNRM-GAME, Toulouse, France

Eddy-driven jet variabilityEddy-driven jet variability

latitudinal shift

pulsing

Results partially similar to those found with a simple barotropic model. (Barnes and Hartmann 2011)

One-to-one line