Gravity wave characteristics derived from quasi

Gravity wave characteristics
derived from quasi-Lagrangian balloon flights
in the stratosphere
A. Hertzog (LMD)
R. Plougonven (LMD)
V. Jewtoukoff (LMD)
R. A. Vincent (U. of Adelaide)
[email protected]
http://www.lmd.polytechnique.fr/VORCORE/McMurdo.htm
OGOA meeting, Lyon, 2013
Motivations




In the atmosphere, gravity waves transport energy and
momentum from their source regions (mainly the troposphere)
to the middle atmosphere
Wave breaking in the stratosphere and mesosphere contributes
to the driving of the global-scale Brewer-Dobson circulation
GW scales (10 – 1,000 km in the horizontal, 100 m – 10 km in
the vertical) are such that they are only marginally resolved in
AGCMs that are used to study climate change
Their global effects are parameterized in AGCMs, but these
parameterizations are based on simplifying assumptions and
lack observational constraints
−
In particular, a significant part of the momentum flux is put in a generic
''non-orographic'' GW drag part, for which there is no link with underlying
source processes: a constant and homogeneous source is assumed
OGOA meeting, Lyon, 2013
Long-duration stratospheric balloons
Closed and non-deformable
Fly on constant-density surfaces
For 2-3 months
Advected by the wind →
OGOA meeting, Lyon, 2013
Long-duration stratospheric balloons
Closed and non-deformable
Fly on constant-density surfaces
For 2-3 months
Advected by the wind →
Vorcore campaign
Sep. 2005 – Feb. 2006
27 balloons, 60 and 80 hPa
In-situ measurements of u, v, P every 15 minutes
and
from wavelet analysis
Hertzog et al. (JAOT 2007, JAS 2008)
OGOA meeting, Lyon, 2013
150,000 obs
Long-duration stratospheric balloons
Closed and non-deformable
Fly on constant-density surfaces
For 2-3 months
Advected by the wind →
Concordiasi campaign
Sep. 2010 – Jan. 2011
19 balloons, 60 hPa
In-situ measurements of u, v, P every 30 s
and
from wavelet analysis
Rabier et al., BAMS, (2010)
OGOA meeting, Lyon, 2013
2,600,000 obs
Gravity-wave momentum flux
(with Concordiasi observations)
The long-duration balloon obs. provide us with:


high time resolution (30 s) along the balloon trajectories →
whole gravity-wave spectrum is resolved
high precision of GPS altitude
and pressure measurements
→ Eulerian P disturbance
-2
u v
and wave parameters:
Wavelet analysis of timeseries
Use of standard wave relations
(Boccara
et al., JGR, 2008)
OGOA meeting, Lyon, 2013
f
N
Map of absolute momentum flux
(Sep.-Jan. average)
2.5° x 2.5° boxes
Mean = 9.0 mPa
Enhanced activity over Peninsula, Drake passage and Transantarctic mountains,
as well as along the continental coast
Higher activity above Austral Ocean than above the Plateau
OGOA meeting, Lyon, 2013
Map of absolute momentum flux
(Sep.-Jan. average)
2.5° x 2.5° boxes
Mean = 9.0 mPa
160 mPa
0.6 mPa
Enhanced activity over Peninsula, Drake passage and Transantarctic mountains,
as well as along the continental coast
Higher activity above Austral Ocean than above the Plateau
OGOA meeting, Lyon, 2013
Zonal average
of absolute momentum flux
Total
Mountain
Flat
GW momentum fluxes maximize between 55S and 75S
Zonally averaged non-orographic gravity-wave activity above the Austral ocean
is as important as orographic activity above the continents
OGOA meeting, Lyon, 2013
Zonal and meridional momentum fluxes
Zonal momentum fluxes are almost everywhere negative (mountains & ocean),
whereas both positive and negatives meridional fluxes are found.
The campaign-averaged net fluxes are significantly smaller than the absolute fluxes:
OGOA meeting, Lyon, 2013
Phase speed distribution
of zonal momentum fluxes
Most of the westward flux is found between 0-40 m/s, while eastward fluxes are in 0-20 m/s
A secondary maximum in the intrinsic phase speed distribution is found between 20-30 m/s,
and corresponds to mountain waves. It is shifted toward ground-based c < 10 m/s.
The ground-based phase speed distribution exhibits “intro waves”,
i.e. waves with
OGOA meeting, Lyon, 2013
distributions
of momentum fluxes and kinetic energy
N
f
Momentum fluxes and kinetic energy maximize for vertical wavelengths between 2-30 km.
Yet, the momentum flux distribution is broader than the kinetic energy one:
in particular higher frequency waves contribute more significantly to the flux than to the energy.
Mountain waves induce the enhancement observed between 2-5 km, and 2-4 hr.
OGOA meeting, Lyon, 2013
Gravity-wave intermittency
(with Vorcore observations)
To study GW intermittency,
we directly look into the GW momentum-flux PDFs
Comparison with momentum flux
derived from HIRDLS observations
OGOA meeting, Lyon, 2013
Balloons/HIRDLS
October, PDF of
Good agreement between both kind of observations.
The major part of the momentum flux is due to rare, large amplitude events.
OGOA meeting, Lyon, 2013
Mountains/flat areas
Balloons, PDF of ρ<u'//w'>
Wave momentum fluxes are as expected higher over mountain than over flat areas,
but wave intermittency is also increased.
Similar ''background'' wave activity (@ small momentum fluxes)
OGOA meeting, Lyon, 2013
Monte-Carlo simulations of GW propagation

Assume a non-intermittent source
Constant momentum flux: F0=1
Invariant spectral distribution <u'w'(m)>, for m in [m0, m1] ; k is fixed →
range of c.



Propagate upward that spectrum in a random horizontal wind
field: U(z). Assume that U is homogeneous in the horizontal,
and stationary → Ground-based c is conserved.
At each level, discard the spectral elements that have
encountered a critical level (c=u). Propagate conservatively
the others to the next level (cf. Hines 1993, Souprayen et al.,
2001)
Do many realizations, and look at the pdf of wave momentum
flux as a function of z, obtained with all these realizations.
OGOA meeting, Lyon, 2013
Simulated PDF evolution
OGOA meeting, Lyon, 2013
Simulated PDF evolution
OGOA meeting, Lyon, 2013
Simulated PDF evolution
OGOA meeting, Lyon, 2013
Simulated PDF evolution
OGOA meeting, Lyon, 2013
Simulated PDF evolution
OGOA meeting, Lyon, 2013
Simulated PDF evolution
In this very simple approach, the lognormal distribution arises as
F(zn) = an x an-1 x … x a1 x F(z0)
where the ais are comprised between 0 and 1, and represent that part of the spectrum that
can
freely propagate from level i-1 to level i (Central limit theorem for multiplicative processes)
OGOA meeting, Lyon, 2013
Conclusions


The full GW characteristics over Antarctica should hopefully be
diagnosed with the Concordiasi dataset.
Gravity-wave momentum flux distributions exhibit long pdf tails
in the lower stratosphere
−
A major fraction of the total GW flux is due to only a small fraction of the
wave events → GW activity is very intermittent
−
This is a robust feature, observed in different observational datasets, as
well as in numerical simulations
−
The intermittency is higher over mountains than over flat areas
OGOA meeting, Lyon, 2013