Seasonal dependence of gravity wave occurrence frequency and
propagation direction using the OH imager at Platteville, CO
Tao Li, B. P. Williams, and C. Y. She
(Department of Physics, Colorado State University)
L. Kieffaber, and A. Peterson
(Whitworth College)
The modefied Whitworth College OH imager with automatically running feature was deployed in Platteville,CO (40.2N, 104.7W) since August 2001. So far we have most clear nights data except some gaps due to computer problems. This paper will present seasonal dependence of gravity wave occurrence frequency and propagation direction based on the 13 month data from September 2001 to September 2002. From our observation, we found that in summer the gravity wave occrrence frequency is very high, around 100% compare to other seasons, around 85%. On the other hand the seasonal dependence of gravity wave propagation direction shows a strong northward (poleward) preference in summer and a weaker southward (equatorward) preference in winter. The westward travelling waves were very rare, but eastward travelling wave were often in Fall. The possible reasons for the east/west asymmetry and north/south anisotropy were also discussed in this paper.
Keywords: OH imager; gravity wave; seasonal dependence; occurrence frequency; propagation direction
Gravity waves play an important role for the transportation of momentum and energy from the lower atmosphere to the middle atmosphere, and thus significently influence on the temperature structure and global-scale circulation in this region. So observation of gravity wave properties has become a great interest of many middle atmospheric scientists.
When the gravity waves propagate through the airglow layers, the perturbations of density are introduced, and variations on the visible and near-infrared emission intensity for certain species can be dectected from ground. All sky images of the OH (~875 km), OI (~965km), O2 (~945 km), and sometimes Na (~905km) airglow layers have been made by groundbased video and CCD imagers for the past several decades. (Taylor, Hecht, Swenson, etc.) These images can measure gravity wave horizontal wavelength, observed phase velocity (speed and direction), observed period (5 min to 1-2hrs, typically), and intensity amplitude (ΔI/I0 typically 1-4%). Meanwhile these imagers are inexpensive and not instrumently complicated, and also can be setup to run automatically. But to deduce the intrinsic wave period and phase speed, they need to be co-located with a more expensive instrument to measure the background wind, such as a radar, lidar, or interferometer.
We brought the Whitworth College OH imager to Colorado to take advantage of and complement the wind measurements made by the Colorado State University sodium resonance lidar and the Platteville radar cluster operated by the University of Colorado and collaborators. The imager also forms a link in the Rocky Mountain chain of OH imagers coordinated by M. Taylor and G. Swenson for TIMED science. This paper reports the initial results of the imager on the seasonal variation of gravity wave occurrence frequency and propagation direction.
2. Instrument
The Whitworth College OH imager has been taking broadband OH images at a 2 min cadence on most nights since it was installed in a trailer at Platteville,CO (40.2N, 104.7W) in August 2001. The imager uses a Nike f/2.8 fisheye lens to give 180 degrees field-of-view. It has a 25mm diameter broadband OH filter (680nm-980nm) with a cut-on feature at 680 nm. The CCD camera is an Apogee with a Kodak 1024x1024 CCD chip binned to 512x512 with a 120sec exposure time and 12 bit digitization. The camera sensitivity provides the cut-off before the filter does. This imager was built at Whitworth College and used for undergraduate education projects for many years. In the summer of 2001, it was brought to Colorado and modified to run automatically every day after sunset and before sunrise. We have images for most clear nights from September 2001 to September 2002, with only a few 2 week or shorter gaps due to computer problems.
3. Data Processing
All images taken by the Whitworth College OH imager were saved in FITS format every 125 sec with 120 sec exposure time and 5 sec data transfer time to computer. Since the Platteville site has no high speed ethernet access, we had to travel to the trailer every 2 weeks to burn CDs of the raw data for processing.
Fig.1. An example
of typical images that we classified as hight, low and no gravity
wave acitivity. in the upleft side shows a very strong gravity wave propagating
northward, but very weak gravity acitivity was shown in the downleft side.
Usually we observed wave like picture shown in the upright side. When gravity
wave is extremely weak, we counted them as no wave acitivity as shown in the
downright
side.
The first step we used to process raw image data
is that we rotated and shifted the images to place the North Star at
the proper altitude and azimuth. Then, in order to look at short
period waves, we subtracted sucsessive 2min images to form difference
images. This subtracted out the nonvarying background light and much
of the star background. The differencing acts effectively as a
low-pass filter, with maximum response at 4min periods, falling off
as the period increases. The difference images for each night were
scaled to 40 count limits,
compared to a typical value of 1000 counts in the raw image. This
gives us a range of 4% with
waves visible down to less than 1% perturbation. The raw and
difference images were made into an MPEG file for each night. Further
analysis was done by visually reviewing the MPEG files. We first
determined the hours in each month with clear, moonless skies, which
ranged from a low of 14.5 hours to a high of 85 hours. Then, we
counted the number of hours in each month with visible gravity waves
in the difference images and divided these hours by the total number
of clear, mooonless hours to determine the monthly gravity wave
occurrence frequncy. At the same time, we classified the intensity of
these occurred gravity waves as high, media, low, and no wave
acitivity. Figure 1 shows an example of typical images with high,
media, low and no gravity wave activity. The observed directions of
wave propagation was divided into 45 degrees bins (N, NE, E, etc.)
and the dominant direction of propagation for each month was
determined afterwards. We counted the larger scale waves that extend
over much of the field of view that are often referred to as band
structures [Taylor,?]. There were often 2 or more band structures
present, so we used the propagation direction of the strongest wave.
We also noted that in winter smaller scale 'ripples' ocurred more
frequently than that in summer [Taylor,?], but we did not include
them in the statistics.
4. Results
Month-by-month details of the results are listed in Table 1. Gravity waves were present in most of the images with the mean monthly occurance frequency over 90% (Figure 2) and multiple waves often present at the same time (Figure 1). The occrrence frequency of gravity waves was ~100% from April to August, dropping to lower values in
Tab.1. Monthly detailed result from September 2001 to September 2002.
1.Month |
Hours of Mostly clear sky |
Hours of GW presence |
Number of nights used |
GW occurrence frequency |
Predominant Propagation Direction of GW front |
|
September 2001 |
40 |
33 |
7 |
82.5% |
North (71%) |
|
October |
53.5 |
45.5 |
11 |
84% |
Southeast (47%) |
|
November |
68.5 |
65.5 |
12 |
95% |
East (46%) |
|
December |
29 |
22 |
6 |
75% |
North (40%) |
|
January 2002 |
77 |
72.5 |
16 |
94% |
South (42%) |
|
February |
85 |
77.5 |
17 |
91% |
South(32%) |
|
March |
14.5 |
12.5 |
4 |
86% |
East (68%) |
|
April |
20 |
20 |
7 |
100% |
Northeast (97.5%) |
|
May |
17.5 |
17.5 |
7 |
100% |
Northeast (54%) North (46%) |
|
June |
75 |
75 |
18 |
100% |
North (55%) |
|
July |
76 |
76 |
21 |
100% |
North (53%) |
|
August |
54.5 |
54 |
15 |
99.00% |
North (45%) |
|
September |
42.5 |
37.5 |
7 |
88.00% |
North (49%) |
|
Total |
653 |
608.5 |
148 |
93.00% |
North |
Fig.2. Monthly occurance frequency of gravity wave from September
2001 to September 2002. From April to August 2002, the gravity
wave appeared almost in every frame of OH image.
fall and winter with the lowest value of 75% in December. In winter, the occurrence of small scale "ripples" increased. This is in contrast with the low fall and winter occurrence frequency observed at Peach Mountain Observatory in Michigan (Wu and Killeen, 1996), with no waves observed in the winter. They reviewed the raw images, which were sensitive only to the largest waves with amplitudes greater than 5%. We did see significantly more large amplitude waves in summer than in winter consistent with the Wu and Kileen (1996) results. The predominant propagation direction of GW fronts also showed seasonal dependence (Table 1 and Figure 3). Northward propagation dominated in the summer and early fall from June to September. October showed a preference for southeast propagation, with east largest in November. December was largely bimodal north and south and southward was preferred in January. February showed a variety of wave directions, with a slight preference for southward. In March, we only saw eastward (68%) and northward (32%) and April was 98% northeast. May had waves travelling both northeast and north, providing a smooth transition to the summer northward domination.
Fig.3. Seasonal dependence of dominant gravity wave propagation direction was shown in
4 graphs above representing 4 different seasons. In summer, northward propagation
direction preference clearly shows that the gravity wave source located in the south of
Platteville. But in wind a weaker equatorward preference was shown.
To summarize the seasonal results, we saw a strong northward (poleward) preference in summer and a weaker southward (equatorward) preference in winter. Spring was predominantly northeast and fall was mixed with a slight preference for east. Westward travelling waves were rare and were the dominant direction in none of the months. The percentage of waves in the NW, W, and SW directions ranged from 0% for spring to 30% for autumn. A strong majority of the wave phase fronts propagated north, south, or into the eastward hemisphere.
5. Discussion and Conclusions:
Filtering by winds in the lower and middle atmosphere can account for the east/west asymmetry, (Taylor, 1993) but location of the gravity wave sources in the troposphere is likely responsible for the north/south shift from summer to winter. Walterscheid et al. (1999) gave similar results for the southern hemisphere with poleward propagation in the summer and equatorward in the winter. They expalin this north-south anisotropy as an effect of wave ducting by the mesopause thermal structure of waves generated by the strong tropical deep convection to the north or storms to the south in winter. They model the effect of the mean winds on the wave duct to account for the absence of east-west propagation.
We studied the seasonal dependence of gravity wave occurrence frequency and found that the average frequecny was over 90% based on the 13 month OH images taken in Plantteville, CO with almost 100% occurrence frequency in summer and ~85% in other seasons. The predominant propagation direction of gravity waves also show the strong seasonal dependence with a northward (poleward) preference in summer and a weaker southward (equatorward) preference in winter. No westward preference was predominant in any month.
The result presented in this paper is still initial analysis for 1 year OH image data. Out future work will concentrate on looking at the temperature and wind profiles from the nearby CSU Sodium lidar during the wave presence to determine wave thermal ducting.
References:
Hecht, J.H., Walterscheid, R.L., Hickey, M.P., Franke, S.J., 2001. Climatology and modeling of quasi- monochromatic gravity waves observed over Urbana, Illinois. Journal of Geophysical Research 106, 5181-5195.
Nakamura, T., Higashikawa, A., Tsuda, T., Matsushita, Y., 1999. Seasonal variations of gravity wave structures on OH airglow with CCD imager at Shigaraki. Earth Planets Space 51, 897-906.
Nakamura, T., Tsuda, T., Maekawa, R., Tsutsumi, M., Shiokawa, K., Ogawa, T., 2001. Seasonal variation of gravity waves with various temporal and horizontal scales in the MLT region observed with radar and airglow imaging. Advances in Space Research 27 (10), 1737-1742.
Smith, S.M., Mendillo, M., Baumgardner, J., Clark, R.R., 2000. Mesospheric gravity wave imaging at a subauroral
site: First results from Millstone Hill. Journal of Geophysical Research 105 (A12), 27119-27130.
Taylor, M.J., Hapwood, M.A., 1988. Identification of a thunderstorm as a source of short period gravity waves in the upper atmosphere nightglow emission. Planetary and Space Science 35, 975-985.
Taylor, M.J., Ryan, E.H., Tuan, T.F., Edwards, R., 1993. Evidence of preferential directions for gravity wave propagation due to wind filtering in the middle atmosphere. Journal of Geophysical Research 98, 6047-6057.
Walterscheid, R.L., Hecht, J.H., Vincent, R.A., Reid, I.M., Woithe, J., Hickey, M.P., 1999. Analysis and interpretation of airglow and radar observations of quasi-monochromatic gravity waves in the upper mesosphere
and lower thermosphere over Adelaide, Australia (35S, 138E). Journal of Atmospheric and Solar-
Terrestrial Physics 61, 461- 478.
Wu, Q., Killeen, T.L., 1996. Seasonal dependence of mesospheric gravity waves (<100km) at Peach Mountain Observatory, Michigan. Geophysical Research Letters 23 (17), 2211-2214.