ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-18-6721-2018Observation of Kelvin–Helmholtz instabilities and gravity waves in the summer mesopause above Andenes in Northern NorwayShort-period GWStoberGunterstober@iap-kborn.dehttps://orcid.org/0000-0002-7909-6345SommerSvenjaSchultCarstenLatteckRalphhttps://orcid.org/0000-0002-0001-7473ChauJorge L.Leibniz Institute of Atmospheric Physics at the University Rostock, Schlossstr. 6, 18225 Kühlungsborn, Germanynow at: Fraunhofer Institute for High Frequency Physics and Radar Techniques, Fraunhoferstr. 20, 53343 Wachtberg, GermanyGunter Stober (stober@iap-kborn.de)14May20181896721673213December20172January201812April201825April2018This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/This article is available from https://acp.copernicus.org/articles/18/6721/2018/acp-18-6721-2018.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/18/6721/2018/acp-18-6721-2018.pdf
We present observations obtained with the Middle Atmosphere Alomar Radar
System (MAARSY) to investigate short-period wave-like features using polar
mesospheric summer echoes (PMSEs) as a tracer for the neutral dynamics. We
conducted a multibeam experiment including 67 different beam directions
during a 9-day campaign in June 2013. We identified two Kelvin–Helmholtz
instability (KHI) events from the signal morphology of PMSE. The MAARSY
observations are complemented by collocated meteor radar wind data to
determine the mesoscale gravity wave activity and the vertical structure of
the wind field above the PMSE. The KHIs occurred in a strong shear flow with
Richardson numbers Ri< 0.25. In addition, we observed
15 wave-like events in our MAARSY multibeam observations applying a
sophisticated decomposition of the radial velocity measurements using volume
velocity processing. We retrieved the horizontal wavelength, intrinsic
frequency, propagation direction, and phase speed from the horizontally
resolved wind variability for 15 events. These events showed horizontal
wavelengths between 20 and 40 km, vertical wavelengths between 5 and 10 km,
and rather high intrinsic phase speeds between 45 and 85 m s-1 with
intrinsic periods of 5–10 min.
Introduction
The middle atmosphere is a highly variable atmospheric region driven by a
variety of waves. In particular, the dynamics of the mesosphere–lower
thermosphere (MLT) region is characterized by waves spanning various temporal
and spatial scales, e.g., planetary waves, tides, and gravity waves (GWs).
Our current knowledge about the energy dissipation of breaking mesoscale GWs
at the MLT is limited due to the lack of continuous high-resolution (spatial
and temporal) temperature and wind measurements at these altitudes.
Optical observations, such as lidar or airglow imagers, depend on the weather
conditions (cloud-free conditions) and are often restricted to nighttime
measurements only. Airglow imagers
e.g., and
the mesospheric temperature mapper have the ability to
resolve the horizontal structure in the field of view as well as to obtain
information about the temporal evolution of mesoscale GWs or wave-like
structures, often called ripples . These ripples are
excited when mesoscale GWs break and dissipate their energy and momentum. The
nature of these ripple structures is associated with either convective or
dynamical instabilities. Dynamical instabilities often evolve into Kelvin–Helmholtz instabilities (KHIs),
whereas the convective instability has its origin in superadiabatic
temperature gradients . Airglow observations as well as
models suggest that KHI can generate secondary
instabilities of a convective nature. Although optical measurements provide
valuable information on such small-scale dynamics, they are limited to
nighttime and cloud-free conditions. In addition, airglow observations are
lacking precise altitude information.
Although observations of KHIs have been reported previously using
very high-frequency (VHF) radars , we try to identify such events using polar
mesospheric summer echoes (PMSEs) as a tracer. There are several studies of
KHIs from radar observations in the troposphere (e.g.,
;
) or in the equatorial mesosphere
e.g.,. Optical observations of
noctilucent clouds (NLCs) e.g.,, which are
closely related to PMSE, show that KHIs occur rather frequently in the summer
mesopause at polar latitudes and, hence, might also be seen using PMSE as a
tracer.
Although the PMSE occurrence rate reaches almost 100 % during the summer
months at Andenes with MAARSY , the morphology itself is
rather variable. In particular, the layering within PMSE is significant
and has to be taken into account when PMSE is used as a tracer
for the MLT dynamics. The PMSE layer is affected by tides and GWs, leading to
characteristic altitude variations with a minimum occurrence of better
strength in the afternoon between 14:00 and 18:00 UTC.
In this paper we present measurements under daylight conditions with a high
spatial and temporal resolution using a multibeam radar experiment. The
observations were conducted with the Middle Atmosphere Alomar Radar System
(MAARSY) in Northern Norway (69.30∘ N, 16.04∘ E) during summer
2013. PMSEs are a common phenomenon at this latitude and are suitable tracers
of neutral dynamics e.g.,.
KHIs are investigated from the signal morphology of the PMSE as well as from
the obtained Doppler measurements. Our Doppler velocity measurements permit
us to determine the amplitude of the instability and to estimate the
characteristic scales during a strong shear flow. The PMSE wind observations
are complemented by meteor radar winds in order to access the mean winds
above and below the PMSE layer and to estimate the mesoscale stability
computing the Richardson number (Ri) taking
NRLMSISE-00 as
background temperature . In the second part of the paper
we investigate 15 events with wave-like features using the imaging
capabilities of the radar system to obtain horizontally resolved radial
velocity images, which are analyzed with respect to the propagation
direction, horizontal wavelength, and phase speed .
The paper is structured as follows. A short summary of the technical details
of MAARSY as well as the experiments are presented in Sect. 2, also including
a brief description of the Andenes meteor radar. The wind analysis is
outlined in Sect. 3. Section 4 contains a description of the analysis of two
GW-induced KHI events seen in the morphology of a PMSE,
and we review the GW analysis from horizontally resolved radial velocities
and present the obtained GW properties (observed and intrinsic phase speed,
observed and intrinsic period, horizontal and vertical wavelength). Our
results are discussed and related to other observations in Sect. 5. Finally,
we summarize and conclude our results in Sect. 6.
Experimental setup
MAARSY is located at the Northern Norwegian island of Andøya
(69.30∘ N, 16.04∘ E). The system operates within the VHF band at
53.5 MHz. The radar employs an active phased array consisting of 433 individual antennas. Each antenna is connected to its own transceiver module,
which is freely adjustable in phase, power, and frequency (within the assigned
2 MHz bandwidth around the carrier frequency). MAARSY has a peak power of
866 kW and a beam width of 3.6∘. The beam is freely steerable within
off-zenith angles up to 35∘ without generating grating lobes. A more
detailed description of the radar is given in and an
overview of wind field analysis using multibeam experiments can be found in
.
In summer 2013, MAARSY conducted several multibeam experiments to provide
systematic scans of the horizontal structure of PMSE using 67 unique beam
pointing directions. The experiments were optimized to ensure a horizontal
coverage of 80 km in diameter while keeping a fast enough sampling speed to obtain
reliable Doppler measurements. A complete scan of the observation volume
consisted of four experiments with 17 beams each. Each experiment did contain
the vertical beam and 16 oblique beams. Figure shows the beam
positions for the complete sequence as a projection above the North Norwegian
shoreline (black lines). The red circles denote the diameter of the
illuminated area assuming a 3.6∘ beam width.
The total time resolution between successive scans with the multibeam
experiments was 3 min 35 s. The shortest GW period that could exist at the
PMSE altitude is given by the Brunt–Väisälä period, which is
approximately 4 min at the summer polar mesopause. Given that the vertical
beam is sampled at a higher temporal resolution (it is included in each
experiment), it is possible to resolve even higher periodicities of
approximately 1 min. So far, the spatial and temporal resolution of the multibeam observations
is sufficient to resolve short-period wave-like features. The images derived
from the multibeam experiments permit us to directly access the intrinsic GW
properties similar to airglow observations.
Projection of MAARSY beam positions for the multibeam experiments
during summer 2013. The red circles show the illuminated radar beam area
assuming a beam width of 3.6∘ at 84 km altitude. The black lines are
the shoreline around the North Norwegian island
Andøya.
Data analysis
A summary of the experiment configuration is given in Table .
We analyzed the recorded raw data with regard to the signal-to-noise ratio
(SNR), the radial velocity, and the spectral width using four incoherent
integrations. Further, we obtain the statistical uncertainties from a
truncated Gaussian fit to the spectra. The fitting routine is based upon the
concept presented in , , and . This spectral
Gaussian fitting takes the effects of the rectangular window and the temporal
sampling into account.
In Fig. in the upper two panels we show contour plots of
the radial velocities as well as their associated statistical uncertainties.
We removed potential meteors to avoid a contamination of the measurements.
Meteors are removed by checking all data points with an SNR >-7 dB,
if
there are other adjacent points with a SNR larger than our noise floor
(approximately SNR <-7.5 dB). If there is only one point with an enhanced SNR
and all surrounding ones are smaller or comparable to the noise floor
±0.2 dB, we consider this measurement to be contaminated by a meteor.
Further, we suppressed a potential side lobe contamination at the edges of
the PMSE layer by using the interferometry and assumed consistency in the
vertical profile of the radial velocities. If there is a jump in the vertical
profile of more than 6 m s-1 from the core region towards the edges between
adjacent pixels, we removed these measurements.
The lower panels in Fig. show a histogram of statistical
uncertainties of the radial velocity and a SNR vs. radial statistical
uncertainty scatterplot. The histogram peaks at a statistical uncertainty
of 0.17 m s-1 and has a median of about 0.56 m s-1. We truncated the contour plot
color scale at 5 m s-1 as the histogram shows almost no radial velocity
uncertainties larger than 5 m s-1. The SNR vs. statistical uncertainty radial
velocity scatterplot further visualizes the L shape, which means that large
errors are often associated with low SNR measurements, as is expected.
Stack plot of the radial velocity measurements with MAARSY. The
lower panels show the color-coded statistical uncertainties of the radial
velocity measurement.
MAARSY has a multichannel receiver system, which is used for coherent radar
imaging (CRI) (e.g., ;
; ). When PMSEs have
relative steep gradients at the edges of the layer, the CRI technique is
useful to correct for beam filling effects leading to differences between the
nominal beam pointing direction and the strongest returned signal
. This is of particular relevance for the oblique beams
with off-zenith angles of more than 10∘ as there could be a deviation
of several degrees from the nominal beam pointing direction causing
substantial errors in the derived horizontal wind velocities and altitude
errors of up to 2 km (,
).
Further, we use wind observations from a collocated meteor radar. The system
operates at 32.55 MHz and has a peak power of 30 kW. The radar employs
a crossed dipole antenna for transmission and five crossed dipole antennas
for reception . The radar detects meteors within a
diameter of 600 km. During the summer months we observe between 15 000 and
20 000 meteors per day. The winds are processed with 30 min temporal and
1 km altitude resolution.
Winds presented in this study are computed using a full error propagation of
the statistical uncertainties from the radial velocity measurements and are
based on a new retrieval technique described in
. These wind estimates have been compared and
validated in and . Here we
describe the key features of the developed wind retrieval algorithm. The
starting point of the retrieval is the so-called all-sky fit for both data
sets, which can be considered as a more general DBS (Doppler-beam swinging)
analysis . The advantage of this approach is that we can
use an arbitrary number of measurements (at least three) at different
positions. We additionally implemented a regularization in time and altitude
to retrieve a reliable wind estimate using at least four meteors. The winds are
obtained solving the radial wind equation iteratively to ensure a proper
error propagation due to the statistical uncertainty of the radial wind
velocity measurement and the pointing directions in azimuth and zenith.
Typically we need five iterations until we achieve convergence. Typical errors
in our obtained winds are of the order of less than 1 m s-1 for MAARSY and 1–10 m s-1 for the meteor radar. The largest uncertainties occur at the upper and
lower boundaries of the observed altitudes.
The multibeam experiments are also appropriate for applying more sophisticated
wind analysis methods such as the velocity azimuth display (VAD)
or the volume velocity processing (VVP)
. If radial velocities are not available for all beam
directions due to the patchy PMSE structure, it turns out that the VVP is more
suitable and robust . In addition, the benefit of the VVP
technique is to access higher-order kinematic terms such as horizontal
divergence, stretching, and shearing deformation in the wind field. As we show
in the second part of the paper, VVP allows the decomposition of the wind field into
mean winds, mesoscale distortions (e.g., GW with horizontal wavelengths
larger than the observation volume), and ripples or wave-like features.
ResultsKelvin–Helmholtz instabilities
In June 2013, MAARSY was used for a multibeam experiment campaign. During
this period, the PMSE strength was rather variable regarding the duration and
its horizontal and vertical extension. On 21 June, we observed an
interesting PMSE structure with several thin layers showing signs in the
morphology, which seem to evolve into KHIs. Figure shows the
SNR, the radial velocity, and the spectral width of the vertical beam. There
are times when the morphology of the PMSE layer is forced by strong upward
and downward motions, which appear in the radial velocity as well. We
identify two possible KHI events around 00:30 and around 15:45 UTC. The
events are indicated by the morphology of the PMSE showing a strong wave-like
up and down modulation followed by a more smeared structure indicating a
so-called cat eye pattern. Further, the radial velocity measurements show
large amplitude changes within minutes from ±10 m s-1.
An advantage of the radar measurements is the Doppler information from which
we obtain the 3-D winds independent of the cloud conditions and during
daylight. In particular, the stability of the flow can be investigated from
the vertical wind shear when the KHIs occur. Meteor radar (MR) data are used to extend the
altitude coverage and to complement the MAARSY winds. Figure
shows both observations. Both horizontal wind components obtained from MAARSY
are shown in Fig. 4a, b. Figure 4c, d present the meteor radar zonal and
meridional winds. The zonal and meridional winds are dominated by tides,
which typically have amplitudes of 15–25 m s-1 (semidiurnal) in the
summer mesosphere and the altitude of PMSE . Further,
the zonal wind reverses at approximately 88 km, separating the westerly
mesospheric jet in the lower part from the easterly thermospheric jet above.
The wind reversal and the strong semidiurnal tide generate strong wind shears
in the flow. Differences in the wind magnitude between both observations are
mainly attributed to the different sampling volumes (80 km diameter for
MAARSY and 400 km diameter in the case of the MR) and the temporal
resolution (approx. 4 min for MAARSY and 30 min for the MR winds). The
mesoscale wind patterns are in reasonable agreement between both systems.
Measured SNR from the vertical beam for 21 June 2013. Radial
velocity determined from spectral analysis using a truncated Gaussian fit.
Computed spectral width for the vertical beam.
Color-coded zonal and meridional winds for 21 June 2013.
Panels (a) and (b) show the observations from MAARSY. Panels (c) and (d) show the data from the collocated
meteor radar.
It is known that KHIs evolve in dynamical unstable flows due to strong shears
with Ri=N2/S2<0.25, where
N is the Brunt–Väisälä frequency and S describes the
horizontal wind shear. The Brunt–Väisälä frequency is computed
from
N=gTdTdz+gcp.
Under the assumption of a mesospheric temperature close to the mesopause of
T=130 K, a temperature gradient of dTdz=0, a
gravitational acceleration of g=9.64 m s-2, and a specific heat at
constant pressure of cp=1009 J kg-1 K-1, we obtain a
Brunt–Väisälä period of approximately 4 min at the altitude of
the PMSE. We also estimated the Brunt–Väisälä period above and
below the PMSE layer using a NRLMSISE-00 profile .
The obtained Ri numbers are shown in Fig. assuming a NRLMSISE-00
background temperature profile shifted to match 130 K at the mesopause. The
upper panel shows the Ri obtained from MAARSY and the lower panel
shows
the MR-derived results. In particular, the MAARSY data show that there are
often low Ri numbers within the PMSE layer, which is expected considering that
turbulence is an important factor in the formation of PMSE .
Similarly, Ri numbers are obtained from the MR, although the coarser vertical
resolution as well as the vertical averaging in the wind analysis have to be
taken into account to estimate the Ri. Our wind measurements confirm that
KHI occurred during times showing a strong vertical wind shear in the
horizontal wind speeds that could have generated a sufficiently small
Ri< 0.25, supporting the notion that the instabilities are of a dynamical
origin. However, as we use an empirical temperature background profile, the
low Ri values indicate in first place a strong vertical wind shear instead
of absolute measurements of Ri.
Richardson number estimated from the
vertical wind shear and Brunt–Väisälä frequency. The Richardson
numbers were calculated from the vertical wind shear and the
Brunt–Väisälä frequency assuming a NRLMSISE-00 background
temperature profile shifted by 10 K to lower
temperatures.
The wave characteristics are derived applying a Stokes analysis
. In a first step, we computed the wavelet
spectra for all three components .
Figure shows the resulting spectra for the zonal and
meridional wind component after the subtraction of the mean wind and the
tidal components. Figure 6a, b indicate two wave bursts with periods T<30 min, which coincide with the occurrence times of a strong wave-like
modulation of the morphology of the SNR (see Fig. ). For the
first KHI event, we observed a mean period of 11.5±1.5 min in all three
wind components (zonal, meridional, and vertical) between 00:00 and 00:50 UTC. The
second KHI occurred between 15:00 and 16:00 UTC and had a mean observed period of
20.3±1.0 min in all three wind components. However, the wavelet spectra
shown in Fig. also indicates some spread in the observed
periods, suggesting some dispersion between the events.
Wavelet spectra of the zonal and meridional wind using the MAARSY
wind measurements after the diurnal, semidiurnal, and terdiurnal tide was
removed.
Zoom in on the SNR (a, b), radial velocity (c, d),
and spectral width (e, f) for the two Kelvin–Helmholtz instabilities
observed on 21 June 2013.
The mean horizontal wavelength of the first group of KHIs was determined to
be λh=10.7±5 km and for the second event we
estimated a wavelength of λh=12.3±5.3 km. These
values are obtained assuming that the KHIs are advected by the mean winds
through the radar beam. The vertical extension of the KH billows is estimated
from our range time insensitivity (RTI) to be λz=3±0.5 km. conducted
direct numerical simulations (DNSs) to characterize the KHI evolution at the
MLT and investigated the evolution of KHI in the presence of the mean shear
flows and GW-induced shear flows for small Ri= 0.05–0.20. The
DNSs are compared to actual NLC observations . This
leads to a depth-to-wavelength ratio of 0.3–0.4, suggesting a small initial
Ri.
(a–f) Zonal and meridional amplitude wavelet spectra for
different altitudes. The time series were filtered to remove the mean wind as
well as the tidal components. The white lines indicate the periods of the
diurnal, semidiurnal, and terdiurnal tide.
In Fig. , we zoom in on the SNR, the vertical velocity, and the
spectral width for both KHIs. The SNR indicates a train of ripples (first KHI
event) and a single wave-like event for the second KHI passing through the
vertical beam. Such structures are rather common in airglow images
. Depending on the temporal evolution of the
KHI, we observe strong vertical motions as visualized in Fig. c
and d. Considering PMSE as an inert tracer, the layer follows the upward and
downward motion of the propagating billows. After the passage of the KHIs,
the layer appears to be smoother and vertically smeared compared to the more
confined structure that existed before, which is likely related to the
turbulence generated by the KHIs. From our spectral width measurement in
Fig. e and f, we obtained that the vertical motion of the
KHIs is accompanied by an increased spectral width, which is associated with
an increased turbulence generation.
We further identified the presence of some GWs that become unstable
and likely generated the ripple structures. In Fig. we
present the zonal and meridional wavelet spectra for three altitudes at 83, 90, and 95 km. The wavelet spectra show the MR wind after removing the
tides and mean flow. The pictures point out that there are some components
with GW-like periods with significant amplitudes at 83 km altitude, which
more or less disappear at 90 km and then grow again. This is in particular
obvious for the zonal wind component and to a smaller degree for the
meridional wind.
Gravity wave statistics using horizontally resolved radial velocities
In the previous section, we presented results of ripples and/or billows causing
modulations in rather thin PMSE layers. However, there are times when PMSE
covers a much larger vertical and horizontal volume. Sometimes the complete
scanning area shown in Fig. was filled with PMSE and
provided a sufficiently strong backscatter signal to obtain reliable radial
velocity measurements for each beam direction. This permits the construction
of
radial velocity maps of the horizontal wind variability caused by ripples or
GWs propagating through the layer.
Horizontally resolved radial velocity maps using PMSE as a passive tracer for
neutral dynamics were already introduced by . Here, we
apply this method to enhance our statistics, investigating several days of
our multibeam observations from 21 to 30 June 2013. Considering our previous
experience retrieving monochromatic GW properties from horizontally resolved
radial velocity images, we modified the experiment to ensure that our
sampling time (time for a complete scan) is faster than the
Brunt–Väisälä period for the summer mesopause, which is around
4 min. Further, we improved our analysis to fit directly for the horizontal
wavelength, propagation direction, and phase speed.
Before we can extract ripple or GW features from our radial velocity
measurements, we need to remove the mean wind and large-scale distortions or
contributions from waves with scales larger than our observation volume.
Therefore, we fit for the wind field using the VVP approach
. The basic idea is to drop the
assumption that the wind field has to be homogenous within the observation
volume. expressed the wind field by a Taylor series,
u=u0+∂u∂x(x-x0)+∂u∂y(y-y0)v=v0+∂v∂x(x-x0)+∂v∂y(y-y0).
Here, u0 and v0 express the mean zonal and meridional wind in the
observed volume, respectively, and ∂u/∂x, ∂u/∂y and ∂v/∂x, ∂v/∂y express a zonal
and meridional wind gradient in the x and y directions. For simplicity, we
assume that the radar is located at x0=0 and y0=0. Although the first-order approach outlined here does not account for all the variability within
the observation volume, it provides a good approximation of the mesoscale
situation. The first-order zonal and meridional gradient terms in the x and
y directions can be associated with waves and inhomogeneities larger than the
observation volume.
Sequence of nine successive radial velocity residual images measured on
30 June 2013.
We use the first-order wind approximation given above to retrieve smaller
structures within our field of view. We decompose each radial velocity
measurement by subtracting the VVP solution to obtain a radial velocity
residual vrres for each beam,
vrres=vrobs-vrVVP.
Here, vrobs is the individually observed radial velocity for each beam.
In fact, the radial velocity residuum now includes the wind variability
smaller than the observation volume.
A sequence of nine successive radial velocity images containing all different
beam directions is shown in Fig. . The measurements were
taken on 30 June 2013 and are representative for the type of features that
can be seen in these images. Some frames show rather coherent and wave-like
features; other images seem to be more dominated by random structures, which
may be caused by the superposition of several waves or ripples moving through
our field of view. Whenever we found a coherent wave-like structure in these
images (by looking at the images) that lasted at least three successive
frames, we tried to fit the wave features, i.e., horizontal wavelength, phase
speed, and propagation direction.
(a) Histogram of the observed GW periods.
(b) Histogram of the determined phase speeds using the 2-D fit.
(c) Obtained horizontal wavelengths. (d) Histogram of
derived intrinsic GW periods. (e) Statistics of the computed
intrinsic phase speeds. (f) Histogram of estimated vertical wavelengths assuming linear theory.
Following , a GW is described in the linear
theory by
(u,v,w)=(u′,v′,w′)⋅ei(kx+ly+mz-ωt)+z2H.
Here, u′, v′, and w′ are the zonal, meridional, and vertical amplitude of
the GW, respectively; k and l are the zonal and meridional horizontal wave
numbers; m denotes the vertical wave number; ω is the Eulerian GW
frequency; and H is the scale height. As we just observe the
horizontal structure of the wave, we modify Eq. () by introducing
a phase φ=mz-ωt. Further, we can neglect the term for the
amplitude growth with altitude, z2H, as we only have information
about the wave at a fixed altitude. Thus, we can rewrite Eq. (),
(u,v,w)=(u′,v′,w′)⋅ei(kx+ly+φ).
Assuming only a slow change of the intrinsic frequency and vertical
wavelength over successive frames, we infer the vertical wave number m and
the frequency ω using the time derivative of the phase φ,
dφdt=-ω.
The advantage of the outlined procedure is that we directly obtain the
intrinsic wave or ripple characteristics. The intrinsic wave frequency
ω^ is straightforwardly computed as we know the horizontal
wavelength, the propagation direction, and the mean mesoscale wind components.
The intrinsic frequency and phase speed c^ are given by the Doppler
relation;
ω^=ω-k→⋅u→,c^=c-u.
In total, we were able to identify 15 ripple events within the 9-day
campaign. For that, we searched all 2-D radial velocity images for coherent
wave-like features that lasted at least three successive frames. The obtained
parameters of these ripples are presented in Fig. with
regard to the (intrinsic) period, the (intrinsic) phase speed, and the horizontal and
vertical wavelength. The duration of the GWs or ripples within the scanning
volume varied between 10 and 50 min. Most of the events lasted approximately
20–25 min, viz. for more than four frames. We did not obtain intrinsic
periods shorter than the Brunt–Väisälä period, but they are
already rather close to this limit. It is remarkable that the computed periods
are much larger with 20–90 min, as most of these waves or ripples move
against the background mesoscale flow. This behavior also appears in the
intrinsic and observed phase speeds. Intrinsic phase speeds had values
between 50 and 90 m s-1, whereas the observed phase speeds have values between
2 and 23 m s-1. At phase speeds faster than 60 m s-1 the wave-like structures would
travel more than the diameter of the radar beam, and hence the positive and
negative phase fronts would cancel each other out. Due to the size of the
scanning volume and the subtraction of the mesoscale variability, the
horizontal wavelengths represent the characteristic scale of our scanning
volume between 20 and 40 km. The obtained vertical wavelengths are rather
short,
with values ranging from 5 to 10 km.
The polar diagram in Fig. shows the distribution of the
propagation direction for all 15 events. For simplicity, we just indicated
the mean wind direction with a red arrow. However, individual ripples moved
at a certain angle to the prevailing winds. These angles between the
prevailing wind and the propagation direction of the wave or ripple showed
values between 90 and 180∘.
Discussion
Analyzing the wind structure from MAARSY as well as the MR wind observations,
we showed that the observed modulations in the morphology of PMSE are likely
caused by dynamical instability. This is supported by the computed Richardson
number Ri< 0.25, which is related to a strong shear flow. Further, we are
able to show that there are mesoscale GWs present during the observation of
the KHI. These mesoscale GWs show significant amplitudes at the height of the
KHI, a much smaller amplitude 5 km above, and an increased amplitude at 95 km
altitude. According to , upward-propagating waves do not
dissipate their whole energy at once. They dissipate some energy, decreasing
the GW amplitude, and start growing again above, which is well represented in
the MR data.
Polar diagram of GWs and ripple propagation direction. The red arrow
denotes the mean wind direction.
Other observations above Andenes suggest that KHIs seem to occur frequently
at the polar MLT. and analyzed
two KHI events from a NLC camera network and inferred the background wind
situation by tracing the ice clouds over a much larger field of view than the
billows. They retrieved a horizontal wavelength of 20–30 km for the
generating of GWs. investigated different events and
derived horizontal scales of 5–10 km for the KHI, which are in remarkable
agreement to our measurements of 7–12 km. Due to the weather conditions
(clouds in the troposphere) there are no data available to perform a direct
comparison with our data. Comparable horizonal and vertical wavelengths are
also reported from midlatitude GRIPS (Ground-based Infrared P-branch
Spectrometer) measurements .
Comparing our results to previous optical observations from a NLC camera at
Trondheim monitoring the MLT above Andenes suggests slightly higher observed
phase speeds and much shorter observed wave periods .
The horizontal wavelengths are comparable and had values between 20 and 40 km.
However, these measurements were taken between summer 2007 and summer 2011
and do not cover the period presented in this paper.
Our observations are also consistent with what was reported previously from
airglow observations, although these measurements were not taken at the same
location. showed that breaking mesoscale GWs form ripple
structures that evolve into KHIs. derived some statistics of
ripples and investigated whether they were of dynamical or convective origin.
They inferred the atmospheric stability from lidar and MR winds. The
described features are similar to what we observed with MAARSY.
Comparing our statistical results with the model data from
provides further indication that we have observed
dynamically unstable KHI. Their model resolved convectively generated
mesoscale GWs that propagate up to the mesosphere. Once the GWs
reach the mesosphere they become dynamically instable and form the ripple
features. They obtained a typical billow scale of 8–15 km that occupies a
region with 30–50 km in diameter. The model shows that such events last about
25–40 min. Considering our observations, we likely observed similar
scales and durations, which reassure us that most of the observed events are
dynamically instable KHI. However, we cannot absolutely rule out the
possibility that some of the wave-like features were of convective origin.
Such secondary instabilities can occur after KHIs .
Further, we have to point out that in the model from
the source of the mesoscale GWs is convective clouds. This is likely not the
case at Andenes. We assume that tropospheric jet instabilities are the more
likely source of mesoscale GWs at polar latitudes. A detailed discussion about
the wave sources is beyond the scope of this paper and requires dedicated
model runs.
Conclusions
In this study we present unique observations of ripples and KHI in the wind
field during full daylight conditions at polar latitudes above Andenes. The
observations were conducted with MAARSY during summer 2013 using PMSE as a
tracer for neutral dynamics. The wind analysis was complemented with data
from a collocated MR in order to infer and validate the mesoscale GW
activity.
We were able to identify two KHIs from the morphology of the PMSE layer and
estimated the characteristic scale of these billows to be of the order of
8–12 km. Our measurements indicate an increased spectral width accompanied
with an increased turbulence while the billows occurred in the vertical beam.
In addition, we inferred from the MR observations the presence of mesoscale
GWs that dissipated a part of their energy between 83 and 90 km altitude and
an amplitude growth above, which is at least in qualitative agreement with
. More recent modeling results from
and
with comprehensive models also show qualitative agreements with our
observations. These results suggest that GW wave packets may undergo a
significant acceleration in their phase speeds when propagating upward,
reaching different background flows and evolving into different forms of
instabilities. Other mesospheric observations already indicated that secondary wave
generation due to wave ducting can lead to waves propagating with rather high
phase speeds .
Further, we demonstrated that multibeam experiments are suitable to directly
obtain ripple properties such as horizontal wavelength, intrinsic frequency, and
propagation direction. The observed values are in reasonable agreement with
model simulations of breaking mesoscale GWs generated from convective
tropospheric clouds and are also consistent with airglow
observations .
The data are available upon request to stober@iap-kborn.de.
The authors declare that they have no conflict of interest.
This article is part of the special issue “Sources,
propagation, dissipation and impact of gravity waves (ACP/AMT inter-journal
SI)”. It is not associated with a conference.
Acknowledgements
We thank the IAP technical staff for keeping MAARSY operational. The helpful
discussions about gravity waves with Peter Hoffmann are acknowledged. This
work was partially supported by the WATILA project (SAW-2015-IAP-1). Svenja Sommer was funded by ILWAO (SAW-2012-IAP-4). MAARSY was built under grant
01LP0802A of the German Federal Ministry of Education and Research. Some of
the contributing researchers are supported by the German research grant
MATMELT (SAW-2014-IAP-1). Carsten Schult was supported by grant STO 1053/1-1
(AHEAD) of the Deutsche Forschungsgemeinschaft (DFG).
Edited by: Markus Rapp
Reviewed by: two anonymous referees
ReferencesBaumgarten, G. and Fritts, D. C.: Quantifying Kelvin-Helmholtz instability
dynamics observed in noctilucent clouds: 1. Methods and observations,
J. Geophys. Res.-Atmos., 119, 9324–9337,
10.1002/2014JD021832, 2014.
Browning, K. and Wexler, R.: The Determination of Kinematic Properties of a
Wind field Using Doppler Radar, J. App. Meteorol., 7, 105–113, 1968.Chau, J. L. and Kudeki, E.: First E- and D-region incoherent scatter spectra
observed over Jicamarca, Ann. Geophys., 24, 1295–1303,
10.5194/angeo-24-1295-2006, 2006.Chilson, P. B., Yu, T.-Y., Palmer, R. D., and Kirkwood, S.: Aspect
sensitivity measurements of polar mesosphere summer echoes using coherent
radar imaging, Ann. Geophys., 20, 213–223,
10.5194/angeo-20-213-2002, 2002.Demissie, T. D., Espy, P. J., Kleinknecht, N. H., Hatlen, M., Kaifler, N.,
and Baumgarten, G.: Characteristics and sources of gravity waves observed in
noctilucent cloud over Norway, Atmos. Chem. Phys., 14, 12133–12142,
10.5194/acp-14-12133-2014, 2014.Fritts, D. and Alexander, M. J.: Gravity wave dynamics and effects in the
middle atmosphere, Rev. Geophys., 41, 1–64,
10.1029/2001RG000106, 2003.Fritts, D. C., Baumgarten, G., Wan, K., Werne, J., and Lund, T.: Quantifying
Kelvin-Helmholtz instability dynamics observed in noctilucent clouds: 2.
Modeling and interpretation of observations, J. Geophys.
Res.-Atmos., 119, 9359–9375, 10.1002/2014JD021833, 2014.Fritts, D. C., Laughman, B., Lund, T. S., and Snively, J. B.: Self acceleration
and instability of gravity wave packets: 1. Effects of temporal localization,
J. Geophys. Res.-Atmos., 120, 8783–8803,
10.1002/2015JD023363, 2015.
Fukao, S., Luce, H., Mega, T., and Yamamoto, M. K.: Extensive studies of
large-amplitude Kelvin-Helmholtz billows in the lower atmosphere with VHF
middle and upper atmosphere radar, Q. J. Roy.
Meteor. Soc., 137, 1019–1041, 2011.Hecht, J. H., Fricke-Begemann, C., Walterscheid, R. L., and Höffner, J.:
Observations of the breakdown of an atmospheric gravity wave near the cold
summer mesopause at 54N, Geophys. Res. Lett., 27, 879–882,
10.1029/1999GL010792, 2000.Hecht, J. H., Liu, A. Z., Walterscheid, R. L., and Rudy, R. J.: Maui Mesosphere
and Lower Thermosphere (Maui MALT) observations of the evolution of
Kelvin-Helmholtz billows formed near 86 km altitude, J. Geophys.
Res.-Atmos., 110, D09S10, 10.1029/2003JD003908, 2005.Hecht, J. H., Liu, A. Z., Walterscheid, R. L., Franke, S. J., Rudy, R. J.,
Taylor, M. J., and Pautet, P.-D.: Characteristics of short-period wavelike
features near 87 km altitude from airglow and lidar observations over Maui,
J. Geophys. Res.-Atmos., 112, D16101,
10.1029/2006JD008148, 2007.Hickey, M. P., Taylor, M. J., Gardner, C. S., and Gibbons, C. R.: Full wave
modeling of small scale gravity waves using Airborne Lidar and Observations
of the Hawaiian Airglow (ALOHA 93) O(1 S) images and coincident Na
wind/temperature lidar measurements, J. Geophys. Res.-Atmos., 103, 6439–6453, 10.1029/97JD03373, 1998.Hocking, W., Fuller, B., and Vandepeer, B.: Real-time determination of
meteor-related parameters utilizing modern digital technology, J.
Atmos. Sol.-Terr. Phy., 63, 155–169,
10.1016/S1364-6826(00)00138-3, 2001.Horinouchi, T., Nakamura, T., and Kosaka, J.-I.: Convectively generated
mesoscale gravity waves simulated throughout the middle atmosphere,
Geophys. Res. Lett., 29, 3-1–3-4, 10.1029/2002GL016069, 2002.Jones, J., Webster, A. R., and Hocking, W. K.: An improved interferometer
design for use with meteor radars, Radio Sci., 33, 55–65,
10.1029/97RS03050,
1998.Klostermeyer, J. and Rüster, R.: Radar observation and model computation of
a jet stream-generated Kelvin-Helmholtz instability, J. Geophys.
Res.-Oceans, 85, 2841–2846, 10.1029/JC085iC05p02841, 1980.Klostermeyer, J. and Rüster, R.: Further study of a jet stream-generated
Kelvin-Helmholtz instability, J. Geophys. Res.-Oceans, 86,
6631–6637, 10.1029/JC086iC07p06631, 1981.Kudeki, E., Bhattacharyya, S., and Woodman, R. F.: A new approach in incoherent
scatter F region E × B drift measurements at Jicamarca,
J. Geophys. Res.-Space, 104, 28145–28162,
10.1029/1998JA900110, 1999.Latteck, R. and Bremer, J.: Long-term changes of polar mesosphere summer echoes
at 69∘ N, J. Geophys. Res.-Atmos., 118,
10441–10448, 10.1002/jgrd.50787, 2012.Latteck, R., Singer, W., Rapp, M., Vandepeer, B., Renkwitz, T., Zecha, M., and
Stober, G.: MAARSY: The new MST radar on Andøya – System
description and first results, Radio Sci., 47, RS1006, 10.1029/2011RS004775,
2012.
Lehmacher, G., Guo, L., Kudeki, E., Reyes, P., Akgiray, A., and Chau, J.:
High-resolution observations of mesospheric layers with the Jicamarca VHF
radar, Adv. Space Res., 40, 734–743, 2007.Lehmacher, G. A., Kudeki, E., Akgiray, A., Guo, L., Reyes, P., and Chau, J.:
Radar cross sections for mesospheric echoes at Jicamarca, Ann. Geophys., 27,
2675–2684, 10.5194/angeo-27-2675-2009, 2009.
Lindzen, R. S.: Supersaturation of Vertically Propagating Internal Gravity
Waves, J. Atmos. Sci., 45, 705–711, 1988.Lue, H. Y., Kuo, F. S., Fukao, S., and Nakamura, T.: Studies of gravity wave
propagation in the mesosphere observed by MU radar, Ann. Geophys., 31,
845–858, 10.5194/angeo-31-845-2013, 2013.McCormack, J., Hoppel, K., Kuhl, D., de Wit, R., Stober, G., Espy, P., Baker,
N., Brown, P., Fritts, D., Jacobi, C., Janches, D., Mitchell, N., Ruston, B.,
Swadley, S., Viner, K., Whitcomb, T., and Hibbins, R.: Comparison of
mesospheric winds from a high-altitude meteorological analysis system and
meteor radar observations during the boreal winters of 2009–2010 and
2012–2013, J. Atmos. Sol.-Terr. Phy., 154, 132–166, 10.1016/j.jastp.2016.12.007, 2017.Miles, J. W. and Howard, L. N.: Note on a heterogeneous shear flow, J.
Fluid Mech., 20, 331–336, 10.1017/S0022112064001252, 1964.Picone, J. M., Hedin, A. E., Drob, D. P., and Aikin, A. C.: NRLMSISE-00
empirical model of the atmosphere: Statistical comparisons and scientific
issues, J. Geophys. Res.-Space, 107, 1468, 10.1029/2002JA009430, 2002.Pokhotelov, D., Becker, E., Stober, G., and Chau, J. L.: Seasonal variability
of atmospheric tides in the mesosphere and lower thermosphere: meteor radar
data and simulations, Ann. Geophys. Discuss.,
10.5194/angeo-2018-17, in review, 2018.Rapp, M. and Lübken, F.-J.: Polar mesosphere summer echoes (PMSE): Review
of observations and current understanding, Atmos. Chem. Phys., 4, 2601–2633,
10.5194/acp-4-2601-2004, 2004.Reid, I. M., Ruester, R., and Schmidt, G.: VHF radar observations of
cat's-eye-like structures at mesospheric heights, Nature, 327, 43–45,
10.1038/327043a0, 1987.Sheth, R., Kudeki, E., Lehmacher, G., Sarango, M., Woodman, R., Chau, J.,
Guo, L., and Reyes, P.: A high-resolution study of mesospheric fine structure
with the Jicamarca MST radar, Ann. Geophys., 24, 1281–1293,
10.5194/angeo-24-1281-2006, 2006.Snively, J. B. and Pasko, V. P.: Breaking of thunderstorm generated gravity
waves as a source of short period ducted waves at mesopause altitudes,
Geophys. Res. Lett., 30, 2254, 10.1029/2003GL018436, 2003.
Sommer, S., Stober, G., Chau, J. L., and Latteck, R.: Geometric
considerations of polar mesospheric summer echoes in tilted beams using
coherent radar imaging, Adv. Radio Sci., 12, 197–203,
10.5194/ars-12-197-2014, 2014.Sommer, S., Stober, G., and Chau, J. L.: On the angular dependence and
scattering model of polar mesospheric summer echoes at VHF, J.
Geophys. Res.-Atmos., 121, 278–288, 10.1002/2015JD023518, 2016.Stober, G., Sommer, S., Rapp, M., and Latteck, R.: Investigation of gravity
waves using horizontally resolved radial velocity measurements, Atmos. Meas.
Tech., 6, 2893–2905, 10.5194/amt-6-2893-2013, 2013.Stober, G., Matthias, V., Jacobi, C., Wilhelm, S., Höffner, J., and Chau,
J. L.: Exceptionally strong summer-like zonal wind reversal in the upper
mesosphere during winter 2015/16, Ann. Geophys., 35, 711–720,
10.5194/angeo-35-711-2017, 2017.Stober, G., Chau, J. L., Vierinen, J., Jacobi, C., and Wilhelm, S.:
Retrieving horizontally resolved wind fields using multi-static meteor radar
observations, Atmos. Meas. Tech. Discuss.,
10.5194/amt-2018-93, in review, 2018.Suzuki, S., Shiokawa, K., Otsuka, Y., Kawamura, S., and Murayama, Y.: Evidence
of gravity wave ducting in the mesopause region from airglow network
observations, Geophys. Res. Lett., 40, 601–605,
10.1029/2012GL054605, 2013.
Taylor, M. J., Pendleton Jr, W. R., Pautet, P.-D., Zhao, Y., Olsen, C., Babu,
H. K. S., Medeiros, A. F., and Takahashi, H.: Recent progress in mesospheric
gravity wave studies using nigthglow imaging system, Revista Brasileira de
Geofisica, 25, 49–58, 2007.
Torrence, C. and Compo, G. P.: A Practical Guide to Wavelet Analysis, B.
Am. Meteorol. Soc., 79, 61–78, 1998.
Vincent, R. A. and Fritts, D. C.: A climatology of gravity wave motions in the
mesopause region at Adelaide, Australia, J. Atmos. Sci., 44, 748–760,
1987.
Waldteufel, P. and Corbin, H.: On the Analysis of Single-Doppler Radar Data, J.
Appl. Meteorol., 18, 532–542, 1979.Werne, J. and Fritts, D.: Anisotropy in a stratified shear layer, Phys. Chem.
Earth Pt. B, 26, 263–268, 10.1016/S1464-1909(01)00004-1, 2001.Werne, J. and Fritts, D. C.: Stratified shear turbulence: Evolution and
statistics, Geophys. Res. Lett., 26, 439–442,
10.1029/1999GL900022, 1999.Wilhelm, S., Stober, G., and Chau, J. L.: A comparison of 11-year mesospheric
and lower thermospheric winds determined by meteor and MF radar at
69∘ N, Ann. Geophys., 35, 893–906,
10.5194/angeo-35-893-2017, 2017.Woodman, R. F.: Coherent radar imaging: Signal processing and statistical
properties, Radio Sci., 32, 2373–2391, 10.1029/97RS02017, 1997.Wüst, S., Offenwanger, T., Schmidt, C., Bittner, M., Jacobi, C., Stober,
G., Yee, J.-H., Mlynczak, M. G., and Russell III, J. M.: Derivation of
horizontal and vertical wavelengths using a scanning OH(3-1) airglow
spectrometer, Atmos. Meas. Tech. Discuss.,
10.5194/amt-2017-350, in review, 2017.