Utah State University
DigitalCommons@USU
All Physics Faculty Publications
5-2001
Large Amplitude Perturbations in Mesospheric
OH Meinel and 87-km Na Lidar Temperatures
Around the Autumnal Equinox
Michael J. Taylor
Utah State University
W. R. Pendleton Jr.
H. L. Liu
C. Y. She
L. C. Gardner
R. G. Roble
See next page for additional authors
Follow this and additional works at: http://digitalcommons.usu.edu/physics_facpub
Part of the Physics Commons
Recommended Citation
Taylor, M.J., W.R. Pendleton, Jr., H-L. Liu, C.Y. She, L.C. Gardner, R.G. Roble, and V. Vasoli, Large amplitude perturbations in
mesospheric OH Meinel and 87-km Na lidar temperatures around the autumnal equinox, Geophys. Res. Lett., 28, 1899, 2001.
This Article is brought to you for free and open access by the Physics at
DigitalCommons@USU. It has been accepted for inclusion in All Physics
Faculty Publications by an authorized administrator of
DigitalCommons@USU. For more information, please contact
[email protected].
Physics
Authors
Michael J. Taylor, W. R. Pendleton Jr., H. L. Liu, C. Y. She, L. C. Gardner, R. G. Roble, and V. Vasoli
This article is available at DigitalCommons@USU: http://digitalcommons.usu.edu/physics_facpub/1232
GEOPHYSICAL RESEARCH LETTERS, VOL. 28, NO. 9, PAGES 1899-1902,MAY 1, 2001
Large amplitude perturbations in mesosphericOH Meinel and
87-km Na lidar temperaturesaround the autumnal equinox
M. J.Taylor
• W.R.Pendleton,
Jr.• H.-L.Liu2,C.Y.She
3 L. C.Gardner
•
R. G. Roble2 andV. Vasoli3
Abstract. Two high-precisionCEDAR instruments,an OH
Mesospheric
Temperature
Mapper(MTM) anda Na Temperature Lidar, have beenusedto investigateseasonalvariability
in the mid-latitudetemperatureat -87 km altitude over the
westernUSA. Here we report the observationof a large
perturbationin mesospherictemperaturethat occursshortly
after the autumnal equinox in close associationwith the
penetrationof planetary-waveenergyfrom the troposphere
intothe mesosphere.This perturbationhasbeenobservedon
three occasionsand exhibitsa departureof up to -25-30 K
from the nominalseasonaltrendduringa disturbedperiodof
- 2 weeks. Such behavior representsa dramatictransient
departurefrom the seasonaltrend expectedon the basisof
currentempiricalmodels. Thesenovelresultscoupledwith a
recentTIME-GCM modelingstudy[Liu et al., 2000] provide
important insight into the role of planetary waves in
mesospheric
variabilityduringthe equinoxperiods.
Introduction
Early investigations
of the thermal structureof the upper
mesosphere
and lower thermosphere
(MLT) region using in
situ probes,radar soundings,and airglow observationswere
basedon relativelyinfrequentsoundings
that providedonly a
coarsepictureof the variability. More recently,Rayleighand
resonancelidar systemshave been developedto probe this
importantatmosphericregion in detail [e.g. Leblanc et al.,
1998]. These studiesutilize only a few nightly (or diurnal)
soundingsper month (smoothedwith an -1-month window
function), for constructingmonthly/seasonal
MLT climatology. Such samplingschemesare unlikely to reveal rapid
changein the MLT dynamicsexpectedaroundthe equinoxes.
Recently,Shepherdet al. [1999] have combineddata from
the Wind ImagingInterferometer(WINDII) aboardthe Upper
AtmosphericResearchSatellite (UARS) with ground-based
airglow observationsfrom three well-separatedsites in the
northernhemisphereto help characterizewhat they term "a
springtimetransitionin atomicoxygen". Their datarevealed
largeperturbations
in the springtimeoxygenairglowemission
ratesthat were planetary-scalein nature,possiblyassociated
In this letter,we reportthe first observations
of an autumnal
perturbationin MLT temperatures
possiblycomplementary
to
the springtimeobservations
of Shepherdet al. [1999]. Our
results indicate that a large departure from the -87-km
climatologicalseasonaltemperaturecan occur shortlyafter
the fall equinox. This perturbationappearsto be relatedto the
penetrationof planetary-waveenergyinto the MLT region
during this transition period. A full descriptionof the
planetary-wavedevelopmentusingthe TIME-GCM model is
givenin an accompanying
paperby Liu et al. [thisissue].
Mesospheric Temperature Measurements
The datawere obtainedusingthe ColoradoStateUniversity
Na TemperatureLidar system and a recently-developed
MesosphericTemperatureMapper (MTM). Details of these
instrumentsare given in She and Lowe [1998] and Pendleton
et al., [2000]. Joint measurements
were made at Ft. Collins,
CO (41øN, 105øW), June 1997 to June 1998. MTM measurementswere continuedfrom the Starfire Optical Range,NM
(35øN, 107øW) from November 1998 to December 2000
alongsidethe University of Illinois Wind-TemperatureLidar
system. The MTM has a 75ø field of view (128 x 128 superpixels), but for this study,only the central5 x 5 superpixels
were used to determinethe zenith temperaturevariability at
the OH emissionheight-87 km (averaged over the layer
width of-8 km). These data were later averagedtogetherto
determinea mean value, and its RMS variability, for each
night of operations(-20 nights/monthcenteredon the new
moon). In comparisonthe CSU lidar systemwas operatedfor
typically4 nights/month
irrespectiveof the moon'sphase.
Seasonal
Observations
A total of 114 nightsof MTM data and 48 nightsof lidar
data were obtained during the 12-month campaign at Ft.
Collins. Figure 1 comparesthe nightly mean OH temperatures for the period June-December,1997 with the lidarderivedtemperaturesfor the 87-km level smoothedover 3.7
km in height. To minimize the effectsof tides and gravity
with the seasonal reversal of the zonal wind field.
Limited
waveson the nightly mean values,only data sets> 4-hr in
dataon the OH and 02 rotationaltemperaturesalsoindicateda
lengthare plotted. A measureof the nocturnalgeophysical
highdegreeof correlationwith the emissionratechanges.
variabilityis given by the verticalbarsas the precisionof the
nightlymeansis <1 K for eachinstrument. An unoptimized
comparison
of the MTM and lidar temperatures
for 12 nights
1Space
Dynamics
Laboratory
& Department
of Physics,
UtahState of overlappingdata during this period (averagedata length
University,Logan,UT.
-7.5 hr) yielded a correlationcoefficient(r) of 0.96 and a
2NCAR
HighAltitude
Observatory,
Boulder,
CO.
meannightlydifferenceof only 0.6 K (excluding3 Nov., see
3Department
ofPhysics,
Colorado
State
University,
Ft.Collins,
CO.
below).
However, variationson a night-by-nightbasis
indicatethatour nightlymeans,referencedto the 87-km lidarCopyright2001 by the AmericanGeophysicalUnion.
Papernumber2000GL012682.
0094-8276/01/2000GL012682505.00
derivedtemperatures,
are compatibleto about+ 5 K. This
result is supportedby the solid curve which shows the
expectedannualvariationin temperatureat the 87-km level
1899
1900
TAYLOR ET AL.: AUTUMNAL
i
i
'
i
i
'
m
i
MESOSPHERIC TEMPERATURE, PERTURBATIONS
i
i
period the total observed temperature change (•30 K)
approached
that expectedfor the entire summer-to-winter
240
[ ß NaLidar
at87-km
-I
annual
variation
(-38 K) (dashedhorizontallines). Although
[ -- She&vonZahn('98).• -]-•T
t
our MTM data do not capture this event completely,
220
measurements
were obtained2 nights prior to the minimum
and for 14 nightsafterwardshelpingto characterizeits shape.
Of particularinterestis the steep, quasi-monotonicrise of
•3.2 K/day (meanvalue) back to normal seasonallevelsthat
occurredover a severalday period following the minimum.
180
(Note: The largest day-to-day change was •16 K.) This
feature representsa major disturbancein the 1997 fall
seasonal
trendthathasnot beenpreviouslyreported.
160
I
I
I
•
' ,
Figure 2b shows the correspondingrelative intensity
1•
2•
240
280
320
3•
changes.The seasonalvariabilityin OH M (6,2) emissionat
Day of Year
Ft. Collins has yet to be fully characterized,and, for
Figure 1. Comparisonof MTM nocturnalmeantemperatures
and reference,datafrom Haute Provence(OHP), France(44øN) is
87-km lidar temperature
for Ft. Collins,CO, fromJun-Dec.1997.
shown[Wiens and Weill, 1973]. The data have beennormalized to unity for the August nocturnalmean since this is
typicallythe minimummonthlylevel of the annualcycle. A
basedon a harmonicanalysisof Na lidar measurementsfor high degreeof correlation(r • 0.88) betweenthe intensityand
the previousyear at Ft. Collins [She and von Zahn, 1998]. temperature
perturbations
is evidentindicatingthat bothwere
Both data sets show the expected-35 K winter-summer characterized
by a major decreasefollowedby a rapid return
difference at-•41 øN and both reveal similar nocturnal trends
to nominalconditions. The magnitudeof the relative inten(not shown) on nights of coincident data reinforcing the sity changeis large (-1.9) and spansthe full climatological
argumentthat under normal conditionsthe OH temperatures rangeof annualvariabilityin an intervalof a few days. In
may be usedas a proxy for atmospherictemperatureat-87
addition,
nocturnal-mean
[OI] •558-nm zenithintensity
data
km [She and Lowe, 1998]. This said, there is a significant obtainedat BearLake Observatory,
Utah (41.6ø N, 111.6ø W)
differencebetweenthe summerand winter datasets. During duringthis event(not shown)also indicatea strongcorrelathe summer months the lidar and MTM data are closely tion (r • 0.94) with the OH intensitydata at Ft. Collins. As
clustered about the 87-km mean and exhibit modest variabithesetwo sites are separatedby about 550 km, and as the
lity; however,duringthe winter monthsboth data setsreveal emissionlayershavea typicalverticalseparationof • 10 km,
substantialshort-term(- a few days)variationsin the average the perturbationmust have been relatively large-scalesince
nocturnaltemperatureand enhancednocturnalactivity.
therewasno apparentphaseshift betweenthesetwo datasets.
This changein mesospherictemperaturevariabilityappears
to occur shortlyafter the fall equinox,when the MTM data
revealthe existenceof a largeperturbation(highlightedby the
OH MTM
.....
dotted box). Unfortunatelythere are no lidar data for this
[ Ft.Collins
1997........ ;i
•--'""•_
[_--••"l
period. However, the magnitude(•23 K) and duration(-2
weeks) of this disturbanceare quite remarkable. A second
unusualfeaturein the late fall data is the exceptionallyhigh
:oo
mean nocturnal temperature(249 K at 87 km) that was
determinedby the lidar on 3 November 1997. The MTM also
190
yielded an enhancedmean temperature(226 K) on this
occasionbut considerablyless than that determinedby the
lidar. This eventis currentlybeingmodeled(S. Melo, private
communication)
andwill not be discussed
furtherhere.
.......
•nn_•a_l
_Ma)_._(O_
h_P_)AJ
200
22ø
I
•(a)iA___
__,nn
180
Autumnal
Perturbation
Characteristics
Figure2a showsan enlargementof Figure 1 centeredon the
temperatureperturbationthat was observed by the MTM
during a 17-day period (UT days 268-284, 1997). For
referencethe dashedlines depictthe _+2 (• band aroundthe
expectedseasonaltemperaturevariation at the 87-km level
(centralcurve) over a 120 day periodas determinedfrom the
Ft. Collins Na lidar measurements
during 1996/97 [She and
von Zahn, 1998]. MTM data setsobtainedprior to and after
the fall equinox,averagedover severaldays(crossedbars),lie
within the confidencelimits establishedby Sheandyon Zahn.
In contrast, at its extreme, the temperatureperturbation
departsmarkedlyfrom the expectedseasonalvalue (- 201 K)
reachingtemperatures
as low as • 178 K, which are more akin
to annualminimum summertimelevels. Indeed,duringthis
---e-MTM
OH(O,2)
iI (b)
? 18 ---o--OHP
OH(6,1)
.g
•
(•
•
'
14
Mean of
'ø/ i _'•'"•"'"'
Intensity
::E
•2
O
1o
200
: •,
'
.
220
240
260
Perturbation
.Annua•
Min._(O_
H_
•_)_
280
300
320
Day of Year
Figure 2. Comparison
of (a) OH temperatureand (b) corresponding
zenithintensityperturbationcharacteristics
for the 1997 fall eventof
Figure 1. The intensity data are plotted against an OH (6,1)
climatologyfromHauteProvence(OHP), France(44øN).
TAYLOR ET AL.' AUTUMNAL
MESOSPHERIC
To investigatethis phenomenonfurther, we have analyzed
two additionaltemperaturedatasetsthat encompass
the same
near-equinoctial
period. Figure3a showsdataextractedfrom
the 1996 Na lidar measurements
at Ft. Collins reportedby She
andvon Zahn [1998]. As expected,the averagedlidar values
prior to and after the fall equinox(large crossedbarsrepresentingsmall-sample
standarddeviations),lie almostcentrally
on the seasonaltrend. However,5 nightsof lidar data during
a 7-day period (UT days 279-285) again show a major
departurefrom this trend. The data are availableat 1 km
heightresolutionand areplottedfor 87 km (solidcircles)and
86 km (crosses)to illustrate height variability of the
temperatureperturbation. The main characteristics
of this
disturbanceare its large magnitude,-36 K, and the steep
recoveryrate. However,in this casethe lidar datasuggestan
"overshoot" in temperature above the seasonal trend
followingthe minimum. Unfortunately,
the lidarwas limited
in its operation,but a subsequent
7-dayaverage(centeredon
UT day335) suggests
a returnto normalseasonal
levels. This
behavioris very similarto the MTM dataof Figure2a.
Figure3b showsthemostrecentMTM dataobtainedduring
the fall 1999 from the StarfireOpticalRange,NM which is
TEMPERATURE
PERTURBATIONS
1901
Table1. Summary
of temperature
results(in K) for Figures2 and3.
Site/Instrument
UT Days
Trmn Truax AT
268-284
178
207
23
3.3
(4 IøN, 105øW)MTM
Ft. Collins, 1996
279--285
(41øN, 105øW)Lidar
186
223
18
4.8
Starfire, 1999
166
199
32
6.6
192
203
9
Ft. Collins, 1997
286-294
MRR
(35øN, 107øW)MTM
Model, 1997
260-285
(42.5øN,105øW)
1..3
(179) (211) (22)
(4.0)
Notes' AT = maximumdeparturefrom seasonalnorm; MRR =
MeanRecoveryRate(K/day); parentheses
= adjusted
modelvalues.
time values. For comparison,averagedMTM temperatures
for datawindowsprior to and after this period (crossedbars)
show fair accord with the TIME-GCM model predictions.
Table 1 summarizes
the main propertiesof thesethreeevents.
Discussion
Thesedata,obtainedover a 4 year interval,stronglysuggest
the developmentof a major perturbationin mesospheric
temperatureand airglow intensityshortlyafter the fall equi-•6ø lower in latitude than Ft. Collins. In this case, we have nox that hasnot previouslybeenreported. The perturbations
are almost certainly related to the large-scale variations
used the NCAR Thermosphere-Ionosphere-MesosphereElectrodynamics
GeneralCirculationModel (TIME-GCM) to observedin the OH M and 02 (0,1) airglow emissionintensidefinethe expected87-km temperaturetrend(solid curve)at ties associatedwith changesin atomic oxygen around the
this site. The accompanying
dottedlinesindicatetemperature spring equinox [Shepherdet al., 1999]. In particular, the
rangefor + 0.5 scaleheights. The rangeof summer-winter geographicallydispersedobservationsof Shepherd et al.
indicated a "pulse-like" enhancementin the OH zenith
variabilityat this latitudeis significantlylessat-•30 K. MTM
data during the period again indicatea large temperature intensity(factor 2-3 times) and temperature(-30 K) over a
perturbationwith a corresponding
intensityvariation(not period of a few dayssaid to be consistentwith the transition
feature. In comparison,our autumshown)of similarmagnitudeand temporalcharacteristics
to of a largeplanetary-scale
thosedescribedearlier. In this case,the temperaturechange nal data,recordedfrom two sitesof similar longitudeover the
exceededthe estimatedsummer-winterannualrange,and the westernUSA, were characterizedby a markeddecreasein OH
temperatureand intensity followed by a steady return to
minimum was -10 K cooler than the model-based summernormal conditions of duration about 2 weeks.
220
ia)!i / @day
33
Na Lidar
These results
are compatiblewith the ideathat the enhancement
(decrease)
is due to an adiabaticcompression(expansion)associated
with strongvertical displacementof the OH layer. Of key
importanceto both of these studiesis the finding that the
perturbationsare transient in nature and geographically
extensivesupportinga planetary-wave-type
perturbation.
Recent quasi-globalstudies of MLT winds [e.g. Smith,
1997] and 87-km temperatures[Drob et al., 2000] provide
Ft.
Collins1996
AnnuallMax.
l-"--'
-'"'•
(41øN, 105øW)
.
--e--
87 km
.
--x-86
km T"•
200
"'
190
Annual
--+--
Min.
Model
x
1 997
--x--Adjusted
Model
210
o MTMData< 4-hr
200
•-,200
'•190
x
if:::190
(•180
180
17O
•
25•
i
200
220
'
26O
•o
•;o
Fall•qubox.
•
200
•o
I
270
I
280
•
,I
•0
300
Day of Year
Day of Year
Figure 4. Comparisonof TIME GCM model results of SPW
Figure 3. Autumnaltemperatureperturbations
observedat (a) Ft.
Collins,CO, 1996 (Na lidar), and(b) Starfire,NM, 1999 (MTM).
transience
effects at 87-km
level with
MTM
data for the 1997
autumnalevent. The adjustedmodelagreeswell with the data.
1902
TAYLOR ET AL.' AUTUMNAL
MESOSPHERIC TEMPERATURE
strong evidence for the importanceof stationaryplanetary
waves (SPW) to the energetics, dynamics, and annual
variabilityof this region. However,very little is known about
changesin the SPW field aroundthe equinoxesresultingfrom
the seasonalreversalof the stratospheric
wind field. Liu et al.
[this issue] have suggestedthat the developmentof SPW
activity in the northern-hemisphere
MLT region during the
fall equinoxtransitionperiodmay havebeenthe sourceof the
temperature
(and intensity)perturbations
documented
here.
To test this hypothesis,the impact of the seasonalSPW
transition on the MLT temperature and wind fields was
studiedusingthe NCAR TIME-GCM model. The 1997 event
(Figure2a) was selectedfor investigationsinceits evolution
wasthe mostfully documented.Realisticforcingof the SPW
at the lowerboundaryof the model(-•30 km) wasprovidedby
temperatureand geopotentialheightvariationsat the 10-mb
level specifiedby the daily National Centers for EnvironmentalPrediction(NCEP) analysis.Full detailsof the model
simulationaregivenin a companionletter[Liu et al., 2001].
Figure 4 shows the MTM mean temperaturesfrom Ft.
Collins for the 17-day period (UT days268-284) compared
PERTURBATIONS
transientduration-2 weeks;and (3) recoveryrate of typically
3-7 K/day. The perturbationhas been observed3 times
duringthe past 4 years,occurringshortlyafter the autumnal
equinox. However, on-goingmodelingstudiesindicatethat
its magnitude,patternand durationat a given locationmay
vary significantlydependingon competingfactors such as
SPW forcing in the lower atmosphereand gravity-wave
forcingof the meanflow in the MLT region.What is clearis
thatwe canexpectto seeincreasingvariabilityaroundthe fall
equinox associatedwith the onset of upper mesospheric
planetarywave activity. The OH temperatureand intensity
changesare highly correlatedduring this time and coherent
over a large geographicarea (>7ø in longitude),consistent
with the growthof planetarywavesand associated
wave transiencein the MLT region.Additionalgeographically
distributed measurements
and data-modelcomparisonsare neededto
unequivocallyestablishthe growth of SPWs in the MLT
region as the primary sourceof the observedperturbations.
Such measurements,performed in conjunction with the
NASA TIMED satellite mission, would be invaluable.
Acknowledgements. The CEDAR MTM was supportedby NSF
with predictions
of 87-kmnocturnalmeantemperatures
based grantsATM-9403474 and ATM-9612810. We are mostgratefulto
on the TIME-GCM results. The observingconditionswere
M.A. White and S.S. Chen(CSU), Major J. Brown (Starfire)and A.
verygoodduringthis periodand 12 of the datasetswere> 4 Luettgen (SDL/USU) for their considerablehelp. The NCAR
hrs duration(range:4.3-9.1 hr; mean-•7.0 hr), the remaining modelingwassupportedin partby NASA grantS-97239-E.
5 data setswere between 1-4 hr in duration. The predicted
perturbation
patternagreeswell in shapewith the MTM data,
butthe observed
magnitudeexceedsthepredictions
by a large
factor. A linear correlationanalysisof thesetwo plotsyields
a coefficientr "' 0.73 with an optimumcorrelation(r - 0.86)
obtained by advancing the model results by one day,
suggestinga high degree of temporal correlation. The
disparityin magnitudeis indicatedby the fractionalRMS
perturbations
of- 1.7% and - 5.1% for the modeland the
data,respectively. This may result,in part,from the model
causingsomedampingof the SPWdueto its relativelycourse
resolution(5ø grid size), and/or to the NCEP analysis
underestimating
the 10-mbarperturbation
amplitudes.Sucha
disparity is not without precedent. Shepherdet al. [1998]
concludedthattheir ground-based
02 temperaturedataagreed
well in shape with the TIME-GCM model but differed in
amplitudefor the tidally-inducedvariationsby a similarfactor
of -•3, with the observations
alsoexceedingpredictions.
To accountfor the disparityin our comparison,the model
was adjustedto have the samemean (192.5 K) and standard
deviation(9.9 K) asthe data,andthen shiftedforwardby one
day for maximumcorrelation.The adjustedmodelresultsbear
a remarkablesimilarity to the MTM data suggestingthat a
rapid phase shift of a quasi-stationaryplanetary wave 1
caused the large mesospheric temperature perturbation
observedaround the 1997 fall equinox [Liu et al., 2001].
However, due to the limited data and data-modelcomparisons,theseinitial resultsare promisingbut not definitive.
In summary, this study provides new evidence for the
developmentof SPW activitynear the mesopause
aroundthe
fall equinox period. The signatureof the transitionperiod
from summer conditions (exhibiting relatively low rms
variability)to winter-timeconditionswhere large, short-term
(- a few days)variationsin the averagenocturnaltemperature
are predominanthas yet to be fully documented. The key
propertiesof this autumnaltransitionperturbationas measuredover the westernUSA are: (1) a significantdeparture
from the climatologicaltrend of magnitude-•25-30 K; (2)
References
Drob, D.P., J. M. Picone, S.D. Eckermann, C.Y. She., J.F.
Kafkalidis, D.A. Ortland, R.J. Niciejewski, and T.L. Killeen,
Mid-latitudetemperatures
at 87-km:Resultsfrommulti-instrument
Fourieranalysis,Geophys.Res.Lett., 27, 2109-2112,2000.
Leblanc, T., I. Stuart McDermid, P. Keckhut, A. Hauchecorne,C.Y.
She and D.A. Krueger, Temperatureclimatologyof the middle
atmospherefrom long-term measurementsat middle and low
latitudes,J.Geophys.Res.,103, D14, 17191-17204, 1998.
Liu, H.L., R.G. Roble, M.J. Taylor, and W.R. Pendleton, Jr.,
Mesospheric
planetarywavesat northernhemisphere
fall equinox,
Geophys.Res.Lett, acceptedFebruary2001 [this issue].
Pendleton, W.R. Jr., M.J. Taylor, and L.C. Gardner, Terdiurnal
oscillationsin OH Meinel rotational temperaturesfor fall
conditionsat northernmid-latitude sites, Geophys.Res. Lett.,
27,1799-1802, 2000.
She. C.Y., and R.P. Lowe, Seasonaltemperaturevariationsin the
mesopauseregion at mid-latitude: Comparison of lidar and
hydroxylrotationaltemperatures
usingWlNDII/UARS OH height
profiles,J. Atmos.Solar-Terr.Phys.,60, 1573-1583,1998.
She, C.Y., and U. von Zahn, Conceptof a two-level mesopause:
Supportthroughnew lidar observations,
J. Geophys.Res., 103,
D5, 5855-5863, 1998.
Shepherd,G.G., R.G. Roble, S.-P. Zhang, C. McLandress,and R.H.
Wiens, Tidal influenceon midlatitudeairglow: comparisonof
satellite and ground-based observations with TIME--GCM
predictions,
J. Geophys.Res.,103, A7, 14741-14751, 1998.
Shepherd,G.G., J. Stegman,P.Espy, C. McLandress,G. Thuillier,
and R.H. Wiens, Springtimetransitionin lower thermospheric
atomicoxygen,J. Geophys.Res.,104, A 1, 213-223, 1999.
Smith,A.K., Stationaryplanetarywavesin uppermesospheric
winds,
J. Atmos. Sci., 54, 2129-2145, 1997.
Wiens, R.H., and G. Weill, Diurnal, annual, and solar cycle
variationsof hydroxyl and sodiumnightglowintensitiesin the
Europe-Africasector,Planet.SpaceSci., 21, 1011-1027, 1973.
M. J. Taylor,W. R. Pendleton,SpaceDynamicsLaboratory,Utah
StateUniversity,
Logan,UT 84322.(e-mail:mtaylor•cc.usu.edu).
H.-L. Liu andR.G. Roble,NCAR High AltitudeObservatory,
PO
Box3000,Boulder,CO (e-mail:liuh•ucar.edu,
roble•ucar.edu).
C.Y. She,Departmentof Physics,ColoradoStateUniversity,Ft.
Collins,CO (e-mail:joeshe•lamar.colostate.edu).
(ReceivedNovember24, 2000; revisedFebruary19, 2001;
acceptedFebruary26, 2001.)