Diapositiva 1

1
The propagation and dissipative filtering of
gravity waves from deep convection in the
thermosphere and F region at equatorial latitudes
Sharon Vadas Colorado Research Assoc./NWRA
Atmospheric coupling processes in the equatorial
region IAGA 11th Scientific Assembly in Sopron,
Hungary, DC.01 WED, Aug 26, 400 PM, 2009
2
The Earth's Atmosphere
Ionosphere (plasma, strongly ionized)? z90-1000
km
Thermosphere (neutrals)? z90-1000 km
Fluid stable, gravity waves can propagate
Mesosphere z50-90 km
Stratosphere z10-50 km
Troposphere, z0-10 km
Fluid unable to support GW propagation during
deep convection
900 K
300 K
Earth's surface
Temperature
3
In a stable fluid, there are only 2 linear
responses to a disturbance (Hines, 1960) (e.g.,
wind flow over mountains, convective
overshoot)? Sound Waves Atmospheric Gravity
Waves - these waves carry nearly ALL of the
momentum flux from typical disturbances
High frequency wave, steep propagation angle
Low frequency wave, steep propagation angle
altitude
Gravity wave moves up, phase moves down
Horizontal plane
(Vadas and Fritts, 2001)?
4
Dissipation altitudes for white noise GWs
(Vadas, 2007)

• GWs with ?H 100-600 km, ?z50-300 km and
  ??20-100 minutes propagate well into the F
  region before dissipating
• Dissipation altitude is much higher when
  thermosphere is hotter

(Vadas, 2007)?
5
Wind filtering is very important Those GWs which
survive critical level filtering propagate
against the background wind
Vertical velocity
Wind filtering from the easterly phase of the
QBO
Piani etal, JAS, 2000
6
White noise GWs
Those GWs propagating against the wind propagate
to the highest altitudes in the thermosphere with
?H?100-400 km, ?z?100-300 km, and observed
periods of 10-40 min Observed
??10,20(bold),30, and 60 min
(Fritts and Vadas, 2008)?
7
Properties of mediun-scale TIDs observed at
Leicester, U.K.

• Periods of 15-40 min
• Phase speeds of 100-250 m/s
• Most TIDs tend to propagate opposite to the
  thermospheric winds

(Waldock and Jones, JATP, 1986)?
8
Gravity waves from deep convection propagate to
the OH layer at z87 km
Yucca Ridge imager, Colorado 11 May, 2004 OH
airglow images from 0338-0458 UT, every 2
minutes
These concentric rings were centered on 2
convective plumes separated by ??100 km 1 hour
earlier
Note outward apparent motion of wave fronts,
although waves are actually moving upwards and
outwards
(Yue et al, 2009)?
9
Convective plume model and ray tracing
updraft of air is modelled as a vertical body
force neglect small-scale structure retain
large-scale envelope of updrafts
Spectrum of GWs excited from a convective plume
with a 15 km envelope
Convective plume
Vertical wavelengths
w'0
ground
image force for wave reflection
Vadas and Fritts (2009)?
Horizontal wavelengths
10
Model results using zonal mean April
winds Scales, periods and amplitudes agree well
with data (Vadas et al, 2009) Note
Easward-propagating waves disappear at late times
because of wave reflection---observed in the data
as well

Intensity at 402 UT
t25-125 min, every 4 min
(Vadas et al, 2009)?
11
Modeled GWs excited by a deep convective plume in
Brazil on 01 October, 2005 density perturbations
in the mesosphere and thermosphere
z90 km (mesopause)?

?
????????
????????
??????
latitude
Follow a wave packet upwards into the
thermosphere. Waves filtered by tidal winds
which rotate anticlockwise with altitude Note
missing waves which have dissipated!!
???????
??????
??????
???????
(Vadas and Liu, 2009)?
??????
??????
longitude
12
We have now seen that primary GWs (excited
directly by deep convection) reach the
thermosphere.
Are there any thermospheric sources of GWs from
deep convection??? When the primary GWs dissipate
in the thermosphere, they accelerate the fluid
horizontally. This process generates secondary
GWs (e.g., Zhu and Holton, Luo and Fritts,
Fritts and Luo, Vadas et al, 2003).
The thermosphere contains a mix of primary and
secondary GWs excited (directly and indirectly)
from deep convection
(Zhu and Holton, 1987)?
13
Tropospheric source of GWs
Thermospheric source of GWs
dissipation altitude - - - maximum ?z ____ Period
????????????
T600 K
Because H is larger, GWs excited in
thermosphere propagate to MUCH higher altitudes
than those excited in lower atmosphere (zlt110
km), because GWs have smaller periods for a
fixed ?H
T1000 K
_____
T1500 K
????
T2000 K
(Vadas, 2007)?
14
Case Study on 01 October, 2005, Brazil Excited
GWs with ?Hlt150 km dissipate in the thermosphere
at z120-200 km, creating horizontal body forces
Convective plume
Body force in this case is southward because the
winds are Nward in the lower thermosphere.
Maximum amplitude is 1 m/s2 at z?180 km.
Duration is 40 min

(Vadas and Liu, 2009)?
15
These body forces excite large-scale upward and
downward-propagating secondary GWs Secondary
waves with scales less than 2000 km cannot be
resolved by this study
Times measured from 2200 UT
TIME-GCM study Temperature perturbations are ?T?
20-30o
latitude
(Vadas and Liu, 2009)?
longitude
Amplitudes agree well with DE2 and CHAMP
satellite data of O and density perturbations
(of an unknown origin) at z300 and 400 km
16
Recent evidence of secondary GWs with medium
scales GOES satellite image of the mid Atlantic
american coast during a rocket experiment on
Wallops Island in October, 2007. Image is
color-coded for temperature. Only those plum and
red convective plumes overshot the tropopause

Deep convection with convective overshoot is
located at 65-77o W and 15-23o N
Wallops Island
Tropical Storm Noel dominates the image with
intense convection. Storm located 2000 km S/SE
of Wallops
(Vadas and Crowley, in preparation)
17
Density perturbations of primary GWs excited by
Tropical Storm Noel As a function of time at
z140 km, with parameterized GW breaking
(Lindzen, 1981 Miyoshi and Fujiwara Yigit et.
al, 2009). Note GW breaking in thermosphere has
been verified recently by Lund and Fritts
?????34
?????23
?????18
latitude
?????17
?????23
?????23
(Vadas and Crowley)
longitude
18
Thermospheric body forces created from
dissipating small-scale GWs with ?Hlt150 km from
Tropical Storm Noel
Thermospheric body forces extend up to 35o, which
is only a few degrees south of Wallops Total
horizontal extent of body force is 62-80o W and
20-33o N
(Vadas and Crowley)
19
In non-viscous fluids, horizontal body forces
create a spectrum of secondary GWs are a
spectrum, with maxima at ?H 2 x full width,
?z? 2 x full depth ?? characteristic time
scale
(Vadas et al, 2003)?
Given the horizontal variability of the
thermospheric body forces (1-3 degrees), we
expect excited secondary GWs with medium scales
of ?Hgt100 km
20
Waves observed at Wallops from 0400-1000
UT with a new Doppler radar system called TIDDBIT
(Crowley et al, 2009) which utilizes three
transmitters and one receiver to measure the
propagation characteristics of travelling
ionospheric disturbances (TIDs) at the bottomside
of the F-region. Most of the waves were
propagating N/NWward from TS Noel
?H100-2000 km CH100-500 m/s ??15-90 min
Are the N/NWward waves from TS Noel? And if so,
are they primary GWs (directly from convection),
or secondary GWs (from thermospheric body
forces), or both?
(Vadas and Crowley)
21
Distribution of TIDDBIT sounder waves
GW theory predicts that ?Hgt100 km ?zgt50
km cHgt100 m/s ?Ir10 to 90 min log10????
-2.1 The wavelengths, periods, phase speeds,
and dissipative factors observed by TIDDBIT agree
very well with GW dissipative theory (Vadas,
2007 Fritts and Vadas, 2008)
(Vadas and Crowley)
??Dissipation factor ????if wave mom flux is max,
????if wave is strongly dissipating
22
Possible sources of TIDDBIT waves
2 main gravity wave sources from deep
convection tropospheric thermospheric
Thermospheric source of secondary GWs (from
small-scale wave dissipation and wave breaking)
Tropospheric source of primary Gws (from
convective overshoot)
(Vadas and Crowley)
23
Source locations for N/NWward propagating waves
via reverse ray tracing
GWS possibly from tropopause have ?Hlt235 km and
cHlt105 m/s.
These locations are 500 km north of deep
convection (i.e., at 15-23o N)!! Waves likely
cannot be primary GWs!
GWs with ?Hgt235 km are likely initially upward
or downward-propagating secondary GWs
GWs with ?Hlt235 km likely secondary GWs!
Downward-propagating secondary GWs
Upward-propagating secondary GWs
(Vadas and Crowley)

24
Explanation as to why none of the GWs which
reverse ray trace to the tropopause are primary
waves. Note These waves all have periods of
15-25 min, and reach the tropopause 500 km north
of Noel.
z?250 km (reflection altitude for TIDDBIT)
Higher-frequency GW-- propagates 750-1250 km
horizontally
lower-frequency GW-- propagates 2000 km
horizontally
?? 40-60 min
?? 15-25 min
(e.g., Hines, 1967 Walock and Jones, 1987)
Overshooting convective plume
25
Therefore, via comparing the reverse ray trace
results with the locations of convective
overshoot (65-77o W and 15-23o N) and the
location of the thermospheric body forces
(62-80o W and 20-33o N), none of the GWs were
identified as primary waves, and 27 out of 33 of
the N/NWward propagating TIDDBIT GWs were
identified as secondary GWs. Secondary GW
characteristics 21 initially downward-propagatin
g (higher frequency) 6 were initially
upward-propagating (lower frequency)
26
Histograms of NW/Nward propagating secondary GWs
identified by ray tracing
Init dw, ?Hlt235km Init dw, ?Hgt235km Init up,
?Hgt235km All secondary GWs
Secondary GW spectra peaks at ?H100-300 km,
?z100-200 km, cH100-300 m/s, and ?r15-30 min !
(Vadas and Crowley)
27
Spectra of secondary GWs
(Vadas and Crowley)
Secondary GW spectra with dissipative filtering
(from z140 to 290 km) peaks at horizontal scales
of ?H100-300 km, with a tail out to ?H1000-2000
km
primary GWs
(Vadas and Crowley)
Therefore, the peak of the secondary GW spectrum
overlaps with the peak of the primary GW spectrum
at the same altitudes. Further modeling work is
needed to determine the secondary GW spectra
excited from horizontal body forces in a viscous
fluid.

28
Neutral Horizontal Wind Perturbations from
Secondary GWs at z400 km simulated from 6
continuous hours of deep convection in Brazil on
October 1, 2005, over 20x20 degrees
z375 km, 2230 UT
DIFFERENCE neutral horizontal winds at z375 km
are max(U,V) ???? m/s.
These are the wind perturbations from SECONDARY
GRAVITY WAVES!!
Vadas and Liu, in preparation
29
Evidence in support of the existence of
secondary GWs 1) DE2 satellite measured O
long-wavelength perturbations of 2 at z300 km
near the equator during geomagnetically quiet
times (Hedin and Mayr, 1987)--agrees well with
our results 2) CHAMP satellite measured RMS
density perturbations of 4-6 at z400 km near
the equator during geomagnetically quiet times
during solar minimum---agrees well with our
results 3) The TIDDBIT ionospheric sounder
observed mostly NW/Nward propagating GWs at
Wallops Island at the bottomside of the F layer.
Using ray trace studies, 27 out of 33 of these
waves were identified as secondary GWs from
Tropical Storm Noel

2000 km to closest convective overshoot
30
Evidence in support of the existence of
thermospheric body forces 1) EISCAT case study
downward and upward propagating GWs originated
from z200 km at 6UT. Occured during and after
rapid deceleration of the geomagnetic meridional
winds from -80 m/s to zero in 1-1.5 hrs at 5 UT
(Shibata and Schlegel, 1993). 2) PFISR case
study SEward propagating GWs observed
simultaneously with large SEward accelerations in
the extracted neutral thermospheric winds at
z180-200 km. GWs likely excited by mountain
wave breaking (Vadas and Nicolls, 2009)
(Vadas and Nicolls, 2009?)
31
Conclusions

• Deep convection excites primary GWs with ?H5 to
  300 km and cHlt200 m/s.
• These primary waves are filtered by winds in the
  lower atmosphere, and by winds and viscosity in
  the thermosphere. Those waves which survive to
  the thermosphere tend to propagate against the
  neutral winds. This increases the intrinsic
  frequency, which increases the dissipation
  altitude.
• Hotter thermospheric temperatures imply deeper
  penetration into the thermosphere.
• Those primary GWs which survive to z200 km have
  ?Hgt100 km, ?zgt50 km, and cHgt100 m/s. In the
  thermosphere, ?z (of the surviving waves)
  increases approximately exponentially with
  altitude.
• When small scale primary GWs from convection
  dissipate and break, they create horizontal body
  forces in the thermosphere at z120 to 200 km.
  This acceleration excites secondary GWs
  with??H100 to 3000 km and cH100 to 600 m/s.
• GWs excited in the thermosphere can propagate to
  much higher altitudes than GWs excited in the
  lower atmosphere because H is larger, thereby
  allowing for larger wave frequencies for a given
  ?H.
• IMPORTANT Waves in the thermosphere and F region
  from deep convection can be either primary waves
  (directly excited) or secondary waves (excited
  from thermospheric body forces). Both have
  scales of ?Hgt100 km and cHgt100 m/s.
• Analysis of new TIDDBIT ionospheric sounder data
  at Wallops Island (using ray tracing) has allowed
  for the identification of secondary GWs from
  Tropical Storm Noel. These secondary waves
  maximize at ?H100-300 km, cH100-300 m/s, and
  ?15-40 min. None of the waves were identified
  as primary waves.

32

Gravity waves generated from deep convective
plumes which overshoot the tropopause
3 May, 1999, near Oklahoma City, Oklahoma
Gravity Waves
Note the wave-like circular ripples that move out
from the overshooting convective plumes Gravity
Waves move upwards and away from the
thunderstorm, carrying energy and momentum with
them
33

Evidence in support of the existence of
secondary GWs in observational data 1) DE2
satellite measured O long-wavelength
perturbations of 2 at z300 km near the equator
during geomagnetically quiet times (Hedin and
Mayr, 1987) This agrees well with model O
perturbations 1.5 for single convective plume.
Typically, thunderstorms consist of many to tens
of plumes, which would slightly increase O
perturbations (although this increase is limited
by wave breaking in the thermosphere (e.g., Yigit
et al, 2009)). 2) CHAMP satellite measured RMS
density perturbations of 4-6 at z400 km near
the equator during geomagnetically quiet times
during solar minimum This agrees well with model
density perturbations as large as 3.6 for a
single convective plume. 3) TIDBIT observed
mostly NW/Nward propagating GWs at Wallops
Island. 27 out of 33 of these waves were found
to be secondary GWs from Tropical Storm Noel via
ray tracing (Vadas and Crowley, 2009 in
preparation). See Vadas/Crowley poster for more
information.
34

Evidence in support of the existence of
secondary GWs in observational data 1) DE2
satellite measured O long-wavelength
perturbations of 2 at z300 km near the equator
during geomagnetically quiet times (Hedin and
Mayr, 1987) This agrees well with model O
perturbations 1.5 for single convective plume.
Typically, thunderstorms consist of many to tens
of plumes, which would icrease O perturbations
z300 km, long wavelength waves
Perturbation amplitudes of long wavelengths GWs
(400-4000 km) are virtually independent of the Ap
index below 60o magnetic latitudeTHEREFORE, NOT
FROM AURORAL HEATING
Hedin and Mayr, JGR,1987
35
Supporting evidence of secondary GWs, cont
2) CHAMP satellite measured RMS density
perturbations of 4-6 at z400 km near the
equator during geomagnetically quiet times
during solar minimum
This agrees well with model density
perturbations as large as 3.6 for a single
convective plume. Larger average amplitudes are
expected for thunderstorms, since the body force
amplitudes are proportionately larger.
(Bruinsma and Forbes, 2008)
36
Supporting evidence, cont 3) HF Dopper in Japan.
Waves with cHlt300 m/s propagated opposite to
wind. Waves with cHgt300 m/s were
omnidirectional. Because of the omnidirectional
property of the fast waves, . This agrees very
well with our model results, which show that
secondary Gws propagate in all directions away
from a body force, except that perpendicular to
the body force direction.
Many high-speed Gws could not originated from
auroral heating
Azimuth is clockise from north
NOTE, WAVES WITH cHgt300 m/s CANNOT PROPAGATE IN
THE LOWER ATMOSPHERE!
(Shibata, 1986)
37
Supporting evidence, cont 4) TIDBIT data
measured directions of propagation for TIDs near
Wallops Is., USA. Most of the waves originated
southeastward of Wallops Is., near Tropical Storm
Noel, even though the Kp index was high (Crowley
et al, in preparation). Because of their
origin, most could not have originated from
auroral heating. This agrees very well with our
model results, which show that secondary Gws
propagate in all directions away from a body
force, except that perpendicular to the body
force direction.
38
Evidence of thermospheric body forces 1) EISCAT
case study downward and upward propagating GWs
originate from z200 km at 6UT. Occured during
and after rapid deceleration of the geomagnetic
meridional winds from -80 m/s to zero in 1-1.5
hrs at 5UT (Shibata and Schlegel, 1993). 2)
PFISR case study PFISR observed Seward
propagating GWs. Extracted neutral winds showed
rapid thermospheric accelerations twice in
several beams at z180-200 km at same time. GWs
likely secondary GWs excited by mountain wave
breaking (Vadas and Nicolls, 2009)
Spatial inhomogeneities in the extracted neutral
windsagrees with theory. Time scale for
acceleration and dissipation also agree with
theory.
z180 and 190 km
SEward thermospheric accelerations at the same
times in nearby beams of 0.1-0.2
m/s2 (Amplitudes of secondary GWs from wave
breaking are only a few of the primary breaking
GW. This may be the reason the acceleration is
5-10 times smaller than for convectively-generated
GWs) GWs disappeared in beams farthest to the
SE
(Vadas and Nicolls, JASTP, 2009?)
39
Evidence in support of the existence of
thermosheric body forces in observational
data 1) EISCAT case study downward and upward
propagating GWs from z200 km at 6UT. ?z larger
(smaller) above (below) the force. Occured during
and after rapid deceleration of the geomagnetic
meridional winds from -80 m/s to zero in 1-1.5
hrs at 5UT. Agrees well with model result A
body force is created at altitudes of z150-220
km, which is decelerated by molecular viscosity
within a few hours. Upward and downward
secondary GWs are excited from thermospheric body
forces. ?z is larger (smaller) above (below) the
force because of the increasing temperatures.
(Shibata and Schlegel, 1993)
40
Supporting evidence of body forces, cont 2)
PFISR case study high-frequency, SEward, upward
propagating GWs. Extracted neutral winds showed
spatial inhomogeneities.
z180 and 190 km
This agrees well with model GW dissipation
causes spatial inhomogeneities in the background
winds. The time scales, 40 min, are also
consistent.
SEward thermospheric accelerations at the same
times in nearby beams of 0.1-0.2 m/s2 GWs
disappeared in beams farthest to the SE
(Vadas and Nicolls, JASTP, 2009?)
41
These accelerations are likely caused by the
dissipation of secondary gravity waves excited
from mountain wave breaking NW of Poker Flat by
400 km
Mountain waves created by strong northward winds
over high Alaskan mountains.
(Vadas and Nicolls, JASTP, 2009?)
42
Additionally, large mean horizontal neutral
wind perturbations of 100-400 m/s are created in
the body force region
(Vadas and Liu, 2009)?
43
Gravity waves excited from a modeled convective
plume
compressible
Boussinesq

• secondary gravity waves are excited

(Vadas, in preparation)?
44
Comparison between OH airglow data and ray trace
convective model data
Modeled GWs excited from the convective plume
reproduce the data very well, thereby validating
the scales and amplitudes of the convective plume
model
45
White noise GW momentum fluxes due to wind,
temperature, and dissipative filtering
Showing GW momentum fluxes scaled by the
background density ???u'w')/?i?u'w')i (0.5 is
bold)?
GWs are from the lower atmosphere, and are
propagating eastward into a westward wind of 200
m/s.
46
?????????????????????????????????? ???????????????
????????????????????????????????????????? ??????
??????????????????????