Indian Journal of Radio & Space Physics
Vol 43, February 2014, pp 83-90
Rayleigh lidar observations of mesospheric inversion layers over Gadanki
(13.5°N, 79.2°E) and their relationship with gravity wave activity
K Ramesh1, S Sridharan2,$,*, S Vijaya Bhaskara Rao1, K Raghunath2 & Y Bhavani Kumar2
1
Department of Physics, Sri Venkateswara University, Tirupati 517 502, India
2
National Atmospheric Research Laboratory, Gadanki 517 112, India
$
E-mail: susridharan@narl.gov.in
Received 20 April 2013; revised 4 July 2013; accepted 23 August 2013
Fifteen nights of Rayleigh lidar temperature profiles over Gadanki (13.5ºN, 79.2ºE) for the period January - February
1999 show the presence of mesospheric inversion layers (MILs) on ten nights indicating their frequent occurrences over low
latitudes. The role of gravity waves in causing the MILs has been investigated using the temperature profiles derived from
the half an hour integrated photon counts. The temperature deviations from the nightly mean temperature are inferred as due
to gravity waves and potential energy per unit mass due to these waves; and their spectra are examined separating MIL
nights and non-MIL nights. The spectra of potential energy per unit mass (Ep) computed from the temperature fluctuations
attain broad band convective instability limit at mesospheric heights on MIL nights, whereas the spectra just tend to attain
the limit on non-MIL nights at longer wavelengths. These results indicate that enhanced gravity wave growth and breaking
due to convective instability may be the potential mechanism for the occurrence of these MILs.
Keywords: Mesospheric inversion layer, Convective instability, Gravity wave, Temperature profile, Brunt-Väisälä
frequency, Potential energy density, Energy content
PACS Nos: 92.60.hc; 92.60.hh
1 Introduction
An inversion of the vertical temperature gradient
from negative to positive observed in the mesosphere
over several kilometers in thickness is called
mesospheric inversion layer (MIL)1. The MILs have
been found to occur more frequently in winter months
at mid latitudes2 and in equinox months3 at low
latitudes. Though, the occurrence characteristics of
MILs have been reported from several sites, it is still
unknown about their causative mechanisms. A few
mechanisms, namely gravity wave breaking, gravity
wave-tidal interactions, chemical heating and
planetary wave critical level interaction have been
suggested for their occurrence1. Of these, the
planetary wave critical level interaction has been
suggested mainly for the lower MIL which occurs in
the altitude region ~65-80 km. Mlynczak & Solomon4
suggested that chemical heating arising from the
quenching of excited species such as OH, which is
maximum at ~90 km during night time, could form
MILs. For the upper mesospheric inversion layers, the
gravity wave tidal interaction has been suggested as
the potential mechanism between 85 and 100 km,
where tidal amplitudes are much larger. Gravity wave
breaking has also been considered as a potential
mechanism for the occurrence of MILs because when
MILs occur, they are accompanied by the regions of
convective or shear instabilities. Convective
instability occurs when the gradient of the total
potential temperature or Brunt-Väisälä frequency
square (N2) becomes negative and the dynamic or
shear instability occurs when the Richardson number
Ri=N2/(du/dz)2 is in between 0 and 0.25. Here, N2, is
Brunt-Väisälä frequency square; du/dz, is the wind
shear. Gravity wave can become unstable at the
height, where the zonal winds begin to decelerate as a
part of the development of a critical layer arising from
the matching of the wave phase speed (c) with the
background zonal wind ( u ), which implies u = c or
u − c = 0 . The decrease in N caused by the higher
temperatures in MIL region reduces the stability,
causing the waves to break and the turbulent heating
arising from the breaking waves provides a feedback
mechanism which can be maintained as a MIL5.
In this paper, the Rayleigh lidar temperature
observations carried out over Gadanki (13.5°N,
79.2°E) during January-February 1999 are used to
find the causative mechanisms for the occurrence of a
84
INDIAN J RADIO & SPACE PHYS, FEBRUARY 2014
few mesospheric inversion layers and their night-tonight variabilities. Gravity waves inferred from the
temperature perturbations and their spectra are
compared for MIL and non-MIL nights and the results
are discussed.
2 Observations and Data analysis
The Rayleigh lidar system at National Atmospheric
Research Laboratory (NARL), Gadanki (13.5°N,
79.2°E) has been operated since March 1998 during
cloud-free nights according to scientists’ requirements.
It employs the second harmonic generation of Nd:YAG
solid state pulsed laser at 532 nm with pulse repetition
frequency of 20Hz and has the energy about 550 mJ
per pulse with pulse width of 7 ns. The Rayleigh lidar
receiver employs a Newtonian type telescope with a
diameter of 750 mm and the back scattered photons
are detected by two photo multiplier tubes of high
(90%) and low (10%) gains referred as R and U
channels and counted sequentially into successive 300
m range bins using Multi Channel Scalar (MCS)
averager for the height range of ~30-85 km. The
strong backscatter from altitudes below 12 km is
blocked by electronic gating, which is synchronized
to trigger the laser. The four minute (250 sec)
averaged photon count profiles corresponding to 5000
laser shots are averaged for 30 minutes and the
temperature and standard errors are determined using
the method given by Hauchecorne & Chanin6.
The Rayleigh lidar technique involves range
resolved detection of molecular backscattered laser
radiation from altitudes above 30 km where the
atmosphere is free from aerosols so that the signal
strength is proportional to molecular number density.
Using the number density taken from an appropriate
model (CIRA-86) for the height of 50 km where the
signal-to-noise ratio is fairly high, the constant of
proportionality is evaluated and thereby, the density
profile (z) is derived. Taking the pressure (P) at the
top of the height range (90 km) from the atmospheric
model (CIRA-86), the pressure profile is computed
using the measured density profile, assuming the
atmosphere to be in hydrostatic equilibrium. Using
the ideal gas equation, the vertical temperature profile
T(z) is calculated using the following expression:
T ( zi ) =
ρ ( zi ) g ( zi ) ∆z
M g ( zi ) ∆z , where
X=
P( zi + ∆z / 2)
R Log (1 + X )
… (1)
The uncertainty in the temperature is calculated by:
δT ( zi )
T ( zi )
=
δLog (1 + X )
Log (1 + X )
=
δX
… (2)
(1 + X ) Log (1 + X )
where, M, is mean molecular weight of air; g, the
acceleration due to gravity; z, the thickness of the
atmospheric layer; R, the universal gas constant; ,
the atmospheric density; and P, the pressure.
3 Results
3.1 Mesospheric thermal structure during January-February 1999
In this paper, the Rayleigh lidar temperature
observations carried out during January-February
1999 are presented. As the system was operated for
several cloud free nights continuously under
equatorial wave campaign, it has provided a valuable
and continuous data set for about 24 nights with
relatively less noise. Figure 1 shows the nightly mean
temperatures during 27 January-19 February 1999
over Gadanki. It can be observed from the figure that
the mesospheric temperatures are relatively larger
during a few nights in the altitude region of ~75-85
km. As the uncertainties in temperature measurements
are larger due to poor signal (less number of photon
counts) at this height region, out of these 24 nights of
observations, only 15 (27-31 January 1999 and 2-3,
5-6, 9-13, 19 February 1999) nights are selected for
the present study.
The fundamental characteristic of an inversion
layer is its amplitude. In the present work, inversion
nights are identified based on the following criteria.
The inversion amplitude can be defined as the
difference between the temperatures at top (~81 km)
and bottom levels (~74 km) of the inversion layer6.
Figure 2(a) shows the inversion amplitude and the
uncertainty values in the inversion region for
27 January-19 February 1999 and the inversion events
Fig. 1 — Contour plot of Rayleigh lidar nightly mean
temperatures (~5 h) during 27 January - 19 February 1999 (24
nights) over Gadanki (13.5°N, 79.2°E)
RAMESH et al.: MILs OVER GADANKI AND THEIR RELATIONSHIP WITH GRAVITY WAVE
are identified as the observed amplitude to be greater
than 20 K, which is reasonably larger than the
maximum uncertainty in temperature (~15K). Thus,
from this figure, the inversion nights are considered to
be 27-29, 31 January 1999, 2-3, 5, 11, 13, 19
February 1999, whereas the non-inversion nights are
30 January, 6, 9, 10, 12 February 1999. The same can
be inferred from Fig. 2(b), which represents the
difference between the inversion amplitude and
uncertainties and the differences are negative
(amplitude < uncertainty) on non-inversion nights and
positive (amplitude > uncertainty) during inversion
nights. Since the data quality is poor due to low
signal-to-noise ratio above ~70 km and operational
duration (<2.5 h) of Rayleigh lidar observations is
less, the remaining nights (30 January, 1, 4, 6-10, 12,
14-18 February 1999) are not considered, even though
the inversion amplitudes are larger or smaller than the
uncertainties. Figure 3(a and b) shows the lidar
vertical temperature profiles for inversion and noninversion events, respectively and they are compared
with MSIS-90 model profiles and each profile is
shifted by 30 K. It is observed clearly from Fig. 3(a)
that on inversion nights, the nightly mean lidar
temperatures are found to be larger (~10-40 K) than
the MSIS-90 model temperatures in the inversion
region, whereas during non-inversion events, the lidar
temperatures are almost matching with the MSIS-90
model temperatures at ~80 km as shown in Fig. 3(b).
It can also be observed from Fig. 3(a and b) that the
maximum temperature is above 220 K for inversion
nights, whereas it is less than 220 K for non-inversion
nights. The thickness of the inversion is equal to the
difference between the altitudes of the top and bottom
levels of the inversion layer and it is found that the
inversion layer thickness is varying from ~ 2.5 km to
Fig. 2 — (a) Mesospheric inversion amplitudes and the
corresponding uncertainties [2 T(z)]; and (b) difference between
inversion amplitude and uncertainties for all the inversion and
non-inversion nights
85
8 km during the inversion nights. It is also observed
that for inversion nights, the environmental lapse rate
( = -dT/dz) is close to the dry adiabatic lapse rate
( d=10 K km-1) at the top level of the inversion but
during non-inversion events, the lapse rate is much
smaller than the dry adiabatic lapse rate (not shown).
3.2 Brunt-Väisälä frequency
The Brunt-Väisälä frequency square is an
important parameter to determine the stability of the
atmosphere. The Brunt-Väisälä frequency squared
values are calculated using the following relation:
N 2 ( z) =
g ( z ) ∂T ( z ) g ( z )
+
T ( z ) ∂z
Cp
… (3)
where, g(z), is the acceleration due to gravity; T(z),
the nightly mean temperature profile; ∂T / ∂z ,
temperature gradient; and Cp, the specific heat at
constant pressure (1004 J kg-1 K-1). The atmosphere is
convectively stable if N2>0 and unstable if N2 0.
Fig. 3 — Altitude profiles of nightly mean temperatures (blue thick
lines) for: (a) mesospheric inversion nights during 27-31 January
and 2, 3, 5, 11, 13, 19 February 1999; and (b) non-mesospheric
inversion nights during 30 January and 6, 9, 10, 12 February 1999
[corresponding MSIS-90 model temperature profiles are shown by
black thin lines; each profile is shifted by 30 K]
86
INDIAN J RADIO & SPACE PHYS, FEBRUARY 2014
Figure 4 shows the Brunt-Väisälä frequency squares
between 30 and 84 km for all the inversion and noninversion events. The first 10 plots (first two rows) are
corresponding to the inversion nights and the next five
plots (third row) are for non-inversion nights. It is
observed from these figures that for the inversion
nights, the N2 values are positive from 30 km to 80 km,
and above that they became zero or negative at the top
of the inversion layer. The N2 values vary from ~5×10-4
(rad s-1)2 at 30 km to ~8.5 - 10×10-4 (rad s-1)2 at 80 km.
On the other hand, for non-inversion nights, the N2
values are positive from 30 km to almost 80-84 km and
the values vary from ~5×10-4 (rad s-1)2 at 30 km to ~6.0
- 9.3×10-4 (rad s-1)2 at ~74-79 km. Thus, it is clear that
the atmosphere is stable for the height regions up to
which the N2 values are positive and the gravity waves
propagate with increasing amplitude in response to the
decreasing density and when N2 becomes negative, the
atmosphere becomes unstable; and the gravity waves
reaching these heights becomes convectively unstable
and breaks. The turbulence generated due to gravity
wave breaking causes the temperature enhancement at
that particular height region, which is maintained as an
inversion layer.
3.3 Vertical wave number spectra of gravity wave potential
energy per unit mass
The average potential energy per unit mass can be
calculated from the following equation:
1
g
Ep =
2 N ( z)
2
2
T ′( z )
T ( z)
o
… (4)
where, g, is the acceleration due to gravity; N(z), the
Brunt-Väisälä frequency; T (z)/To(z), fractional
temperature perturbations7. The power spectral
density (PSD) calculated from the Fast Fourier
Transform (FFT) analysis employed on relative
temperature
perturbations
is
scaled
as
[g/N(z)][T (z)/To(z)], which gives the vertical wave
number potential energy spectral density. Assuming
that the wave induced temperature gradients are
constrained by convective instability and the adiabatic
lapse rates, the vertical wave number (m) and the
potential energy spectral densities are limited to
approximately N2/2m3 for a single mode and N2/10m3
for a broadly distributed spectrum of waves
Fig. 4 — Altitude profiles of Brunt-Väisälä frequency squares for inversion (top and middle rows) and non-inversion (bottom row) nights
[vertical lines indicate the limit of N2=<0]
RAMESH et al.: MILs OVER GADANKI AND THEIR RELATIONSHIP WITH GRAVITY WAVE
interacting only due in superposition. Figure 5 shows
the spectra of relative temperature perturbations as a
function of vertical wave numbers for three different
height regions 30-48.9 km, 48.9-67.8 km and
64.8-83.7 km for the inversion nights. This figure also
shows the broad band convective instability limit for
these height regions calculated from (1/2 )(N2/10m3).
As expected, these spectra reflect that vertical growth
in the spectral magnitude is observed mainly at longer
wavelengths. At higher altitudes (64.8-83.7 km), the
spectrum at longer wavelengths (or shorter wave
number), such as ~10 km, not only attains the
convective instability limit but exceeds this limit
during most of the MIL nights. At lower altitudes
(30.0-48.9 km and 48.9-67.8 km), the spectral
magnitudes are near or below the convective
instability limit indicating that it is yet to attain the
convective instability limit. For example, on 11
February 1999 MIL event, the wave of wavelength
~10 km having spectral power of 5.91×105 m3 s-1 is
not attaining the convective instability limit in the
height region of 30.0–48.9 km but it tends to attain in
the height region of 48.9–67.8 km with spectral power
of 7.29×105 m3 s-1 and in the height region of
87
64.8–83.7 km (superposition height region), the
spectral power of the same wave is 3.69×106 m3 s-1.
This indicates the gravity waves are propagating
upwards with increasing amplitude and attaining the
convective instability limit at a particular height
during the inversion events. Whenever the gravity
wave attains the convective instability limit, the wave
breaks and deposits the energy and momentum carried
by it, which causes an increase in the temperature at
that particular height region. Figure 6 shows the
power spectral densities of temperature perturbations
as a function of vertical wave numbers during nonMIL events. It is clearly observed from this figure that
the gravity waves at higher altitudes (64.8-83.7 km)
do not attain or just tend to attain the convective
instability limit at longer wavelengths (~10 km)
during most of the non-inversion nights. It indicates
that these gravity waves have not attained enough
amplitude to get saturated. As a consequence, the
inversion layers are not present on these noninversion nights.
Figure 7(a and b) shows the averaged spectra of
temperature perturbations for all the inversion nights
and non-inversion nights, respectively. It is clear from
Fig. 5 — Nightly averaged vertical wave number potential energy spectra for inversion nights [broadband convective instability limit
(1/2 )(N2/10m3) lines also shown for corresponding heights]
88
INDIAN J RADIO & SPACE PHYS, FEBRUARY 2014
this figure that at lower altitudes, the gravity wave is
not attaining the convective instability limit on both
inversion and non-inversion events but at higher
altitudes (64.8-83.7 km), the gravity wave of longer
wavelength (~10 km) is having the spectral power
much above the convective instability limit with
maximum value of 2.89×106 m3 s-1 on inversion
events and the wave just tend to attain the convective
Fig. 6 — Nightly averaged vertical wave number potential energy spectra for non-inversion nights [broadband convective instability limit
(1/2 )(N2/10m3) lines also shown for corresponding heights]
Fig. 7 — Total mean vertical wave number potential energy spectra for: (a) inversion and (b) non-inversion nights [broadband convective
instability limit (1/2 )(N2/10m3) lines also shown for corresponding heights]
RAMESH et al.: MILs OVER GADANKI AND THEIR RELATIONSHIP WITH GRAVITY WAVE
instability limit having the averaged spectral power of
1.02×106 m3 s-1 on non-inversion events.
3.4 Energy content spectra
The average energy content spectra calculated by
multiplying the power spectral density (PSD) and the
vertical wave number at the three different altitude
regions 30.0-48.9 km, 48.9-67.8 km and 64.8-83.7 km
for the MIL and non-MIL events are shown in
Fig. 8(a and b), respectively. Here, the PSD is taken
in linear scale and the vertical wave number in log
scale. The area under any portion of this curve is
proportional to the potential energy density contained
in the corresponding wave number band. It is clear
from these figures that the gravity wave of vertical
wavelength of ~10 km (wave number 1×10-4 m-1) has
the lower energies of ~60-70 (ms-1)2 at the lower
height regions (30-67.8 km) but as it propagates
upwards, its amplitude increases and the wave has
higher energy of ~300 (ms-1)2 on inversion events and
~106 (ms-1)2 on non-inversion events in the height
region of 64.8–83.7 km. Thus, the energy content of
the gravity wave is larger in the inversion region
(above 80 km) during MIL events when compared to
that of non-MIL events. But there is not much
difference in energy contents observed on both MIL
and non-MIL events in the lower height region. Thus,
the enhancement in the energy content at higher
altitudes during the MIL events is due to the gravity
wave breaking which did not occur during the case of
non-inversion events due to their smaller energy at
higher altitudes.
4 Discussion and Summary
In this study, a few mesospheric inversion layers
are identified in the Rayleigh lidar temperature
profiles over Gadanki (13.5ºN, 79.2ºE) during
Fig. 8 — Total mean energy content spectra for: (a) inversion and
(b) non-inversion nights
89
January-February 1999 based on the criteria that the
inversion amplitude is greater than 20 K, which is
above the uncertainty limit in the temperature
measurement. The comparison of lidar and MSIS-90
model temperature profiles show that in the height
region of 60-85 km, the lidar temperatures are much
larger than that of model values during inversion
events and almost matching with the model values on
non-inversion nights in the inversion region.
Moreover, the MILs found for ten nights, during
January-February 1999, indicates the frequent
occurrence of MILs over low latitude region Gadanki
and the amplitude of the inversions varies as ~20-48 K
with the thickness of ~2.5-8 km during these MIL
events. The temperature lapse rates reaching the dry
adiabatic lapse rate (10 K km-1) indicates the
existence of the turbulent layers at the inversion
heights. Though several mechanisms have been
proposed for the formation of MILs, the present work
investigates the role of gravity waves in causing the
MILs using Rayleigh lidar temperature observations
over Gadanki.
The stability of the atmosphere, which is essential
for the vertical propagation of the gravity waves, was
estimated from the temperature lapse rates and the
Brunt-Väisälä frequencies for MIL and non-MIL
events. These parameters are further used to obtain
gravity wave potential energy and to verify the
condition for convective instability. Cutler & Collins8
discussed the mesospheric inversion layers above 60 km
with the amplitude of 18 K using Rayleigh lidar
observations over a high latitude station Poker Flat,
Alaska (65ºN, 147ºW). They showed the temperature
gradients at the top of the inversion layer approaching
the dry adiabatic lapse rate which represents the
turbulent layers at that altitude region. The gravity
wave generated in the troposphere propagate
vertically upwards with increasing amplitude as the
atmospheric density decreases in proportion to e(z/2H)
would at some height induce convective or shear
(dynamic) instability. The amplitude growth saturates
above this height (breaking level) by irreversible
extraction of the energy from the wave to produce the
turbulence. Atmospheric gravity waves exist only
when N2 is positive and if N2 becomes negative, the
layer becomes convectively unstable and the gravity
wave amplitudes cannot grow beyond that level. It is
observed that the N2 values are negative above 80 km
indicating the convective instability, which leads to
saturation and breaking of gravity waves during
MIL events.
90
INDIAN J RADIO & SPACE PHYS, FEBRUARY 2014
The temperature deviations from the nightly mean
temperature are inferred as due to gravity waves and
the potential energy per unit mass due to these waves
and their spectra are examined for MIL and non-MIL
events separately. The spectra of potential energy per
unit mass (Ep) calculated from the temperature
fluctuations attained the broadband convective
instability limit at the mesospheric heights during
MIL events, whereas the spectra just tend to attain the
limit for non-MIL events at longer wavelengths.
These results denote the saturation of wave
amplitudes and the wave breaking due to convective
instability are the potential causative mechanisms for
the occurrence of these MILs.
Some main features of the power spectral density
(PSD) have been described here. The PSD of
available potential energy increases for all wave
numbers by roughly one order of magnitude from the
stratosphere to mesosphere over the height regions of
30.0-48.9 km, 48.9-67.8 km and 64.8-83.7 km. But
the observation of the large MIL amplitudes at high
altitudes between ~74 and 81 km is consistent with
longer wavelength gravity waves (~10 km), which are
more stable for a given amplitude than shorter
wavelength waves propagating to higher altitudes. It
is now established by many observational studies that
the amplitude of monochromatic gravity waves
increases with vertical wavelength8. The variation of
PSD with vertical wave number is studied for
different wave numbers for both MIL events and nonMIL events. From the averaged spectral density, it is
clearly shown that on MIL events, the gravity wave of
wavelength ~10 km propagates from the lower
atmosphere and attain the convective instability limit
with larger power spectral density amplitude above
~80 km when compared to lower height regions and
causing the wave breaking and thereby, producing the
mesospheric temperature inversions. But, the wave of
same wavelength (~10 km) during non-MIL nights
just tends to attain the convective instability limit and
it has less PSD when compared to the MIL events.
The averaged energy content, calculated for all the
MIL and non-MIL events, shows that the energy
content on MIL nights is much larger than that on
non-MIL nights in the height region of ~64.8-83.7
km. Thus, it is concluded that the enhanced gravity
wave amplitude and the wave breaking due to
convective instability could be the potential
mechanism for the occurrence of the MILs during
January-February 1999 over Gadanki.
Acknowledgement
The authors would like to acknowledge NASA for
providing the MSIS-E-90 model data through the website
http://omniweb.gsfc.nasa.gov/vitmo/msis_vitmo.html.
References
1 Meriwether J W & Gardner C S, A review of the mesospheric
inversion layer phenomenon, J Geophys Res (USA), 105
(2000) pp 12405-12416.
2 Leblanc T & Hauchecorne A, Recent observations of
mesospheric temperature inversions, J Geophys Res (USA),
102 (1997) pp 19471-19482.
3 Siva Kumar V, Rao P B & Krishnaiah M, Lidar measurements
of stratosphere-mesosphere thermal structure at low latitude:
Comparison with satellite data and models, J Geophys Res
(USA), 108 (2003) 4342, doi: 10.1029/2002JD003029.
4 Mlynczak M G & Solomon S, Middle atmosphere heating by
exothermic chemical reactions involving odd-hydrogen
species, Geophys Res Lett (USA), 18 (1991) pp 37-40.
5 Hauchecorne A, Chanin, M L & Wilson R, Mesospheric
temperature inversion and gravity wave dynamics, Geophys
Res Lett (USA), 14 (1987) pp 935-939.
6 Hauchecorne A & Chanin M L, Density and temperature
profiles obtained by lidar between 35 and 70 km, Geophys
Res Lett (USA), 7 (1980) pp 565-568.
7 Whiteway J A & Carswell A I, Rayleigh Lidar observations
of thermal structure and gravity wave activity in the high
arctic during a stratospheric warming, J Atmos Sci (USA), 51
(1994) pp 3122-3136.
8 Cutler J L, Collins L R, Mizutani K & Itabe T, Rayleigh lidar
observations of mesospheric inversion layers at Poker Flat,
Alaska (65°N, 147°W), Geophys Res Lett (USA), 28 (2001)
pp 1467-1470.