Monochromatic AGW in Airglow

1
Studies Of Monochromatic AGWs From Airglow In The
Mesosphere
Rezaul Haque, Lee Rumsey, and Gary R.
Swenson Department of Electrical Computer
Engineering, University of Illinois at
Urbana-Champaign, Urbana, IL 61801.
2
Abstract
Dynamical perturbations redistribute mesospheric
minor species which influence chemical rate
processes and resultant emissions in the 80-110
km altitude region. Acoustic Gravity Waves (AGWs)
which have large vertical wavelengths effect the
emission layers and studies of layer brightness
provide insight into wave phenomena. Simultaneous
measurements of winds extend the ability to
determine the intrinsic wave characteristics from
which wave energy and momentum flux can be
calculated. The data from 5 nights data taken in
Albuquerque, NM are described. The data is
presented in a spatial plot with a superposition
of theoretical vertical wavelengths imposed by
the dispersion relationship. Values of measured
momentum fluxes and implications of wave ducting
will be presented.
3
Instrumentation A picture of the imager used to
acquire images is presented in the figure to the
left. It combines high-quality, large-aperture
optics to acquire quality image resolution using
low-noise CCDs. It uses a Loral 1024 x 1024 CCD
binned to 512 ? 512 pixels and incorporates a
large-format Mamiya all-sky lens with telecentric
relay optics similar to that mentioned by Mende
et al. The 15-?m CCD pixels give an effective
angular resolution of 0.366o/pixel (or 0.5-km
horizontal surface resolution at 87-km altitude).
With the high spatial and angular resolution of
this imager, airglow waves near the 87-km
altitude region were identified over a radial
distance greater than 400 km from the ground
station. The broadband imaging method includes
the starfield in the exposure, which is used to
calibrate the angular resolution and aspect
information. The wavelength and OH vibrational
bands detected are described by Swenson and
Mende. Photon and readout noises are negligible,
with the major sources of error being background
continuum noise and stars.
4
All-sky Imager Setup
Optical System
5
Processing of images for observed phase velocity
(Co)
A time-difference (TD) image is created from two
subsequent raw images taken at a time separation
of 2 min. This was done to remove the dc
background from the raw images. Because the TD
process allows removal of the entire dc level,
the moving wave structures had a higher contrast
than unprocessed images.
6
Acoustic Gravity Wave Parameter Estimation
Algorithm
Start
Is the minimum MSE below a preset threshold?
Load two consecutive images
N
No wave found Exit
Y
Calculate the Mean-Squared Error (MSE) between
the first image and shifted copies of the second
image (Shift step size8 pixels)
Further minimize the MSE using finer
steps (Shift step size1 pixel)
Identify the shift that minimizes the MSE
Using the shift found above, estimate
wave bearing and velocity
Record info Finished.
7
Fixed values of Vertical Wavelength (?z) plotted
as a function of Intrinsic Phase Speed (Ci ) and
Horizontal Wavelength (?h). The dispersion
relationship for AGWs is used to characterize the
variables. The top (heavy) line is the limit
imposed by ?z8 and the bottom (heavy) curve is
the limit imposed by the thickness of the airglow
layer (?z 10 km).
Same as above, but includes all data for 5 nights
from Albuquerque, NM Swenson et. al., 1999
8
Acoustic Gravity Wave Spectral Analysis

• Problem Given a sequence of OH airglow images,
  can we determine the horizontal bearing and
  wavelength of acoustic gravity waves?
• Approach Compute the Unambiguous 2-D
  Horizontal Wave Number Spectrum
• Calculate the 3-D power spectral density for
  airglow perturbation p
• Integrate FOH on (-?,0) to obtain horizontal
  (k,l) spectra w/o conjugate peaks
• Remaining spectral peaks indicate the direction
  of propagation
• Example

9
Image Processing Challenges
Our specific imaging situation presents some
unique image processing problems. Challenges 1.
The CCD imager samples the airglow layer on a
non-uniform grid 2. The all-sky lens introduces
radially-symmetric intensity distortion 3.
Spatial and temporal data are limited in
extent/duration 4. Stars and other objects
pollute the images Solutions 1. Resample the
data on a uniform Cartesian grid using sinc
interpolation 2. Flat-field the images by
estimating the degree of distortion 3. Use
spatial and temporal windowing, prewhitening, and
recoloring 4. Employ a despeckling filter to
remove unwanted spatial features
10
2-D Unambiguous Spectral Analysis Code
Detrend each interpolated image
Start
Load a sequence of airglow images
Prewhiten spatial and temporal data
Remove radial intensity distortion from each
image (Flat-fielding)
Compute the 3-D (k, l, ?) spectrum
Remove Stars and Normalize each image
Recolor the spectrum
Compute the unambiguous 2-D (k, l) spectrum
Interpolate each image and re-sample on a uniform
Cartesian sampling grid
Plot spectrum Finished.
11
Synthetic Image Test
We generated a 20-image sequence of synthetic
waves to test our code. Note the distortion
introduced by the all-sky camera
view. Performance is similar in the presence of
noise.
Synthetic All-sky Image 3 airglow
perturbation (2) 14 km waves, (10)
stars Quadratic Intensity distortion
Flat-fielded Image Note that radial intensity
distortion is attenuated
Interpolated Image 128x128 Cartesian sampling
grid Stars have been removed
12
Unambiguous 2-D Spectral Plots
Synthetic Image Set
Evening of 03 Feb 95
Spectral peaks indicate the propagation
direction for each wave
Ticks indicate waves detected by
Haques parameter estimation algorithm
13
Future Goals
Optimize flat-fielding routine for speed and
accuracy Explore fast spectral estimation
algorithms, such as ESPRIT Further verify our
code with synthetic image sets Develop a flexible
user interface for our code
References
Coble, M., Computing two-dimensional unambiguous
horizontal wave number spectra from OH airglow
images. Ph.D. Dissertation, ECE Department,
University of Illinois at Urbana-Champaign,
1997. Gulati, K., Computing two-dimensional wave
number spectra and gravity wave momentum fluxes
from OH airglow images. M.S. Thesis, ECE
Department, University of Illinois at
Urbana-Champaign, 1996. Mende, S., Eather, R.,
and Aamodt, E. Instrument for the monochromatic
observation of all sky auroral images. Applied
Optics 16, pp. 1691-1700, 1977. Swenson, G.,
Haque, R., and Alexander, J., Dispersion imposed
limits on acoustic gravity waves in the
mesosphere observations from OH airglow.
Geophysical Research Letters (submitted). Swenson,
G., and Mende, S., OH emission and gravity
waves (including a breaking wave) in all-sky
imagery from Bear Lake, UT. Geophysical
Research Letters 21, pp. 2239-2242, 1994.