0000o

v- DYNAMICS OF THE POLAR MESOPAUSE ANDLOWER THERMOSPHEREREGION AS OBSERVED

O% IN THE NIGHT AIRGLOW EMISSIONS

BY

HANS KRISTIAN MYRA30

NOrE PUBL-C8/1001SN4 tt'1- V. 12

FORSVARETS FORSKNINGSINSTITUTTNORWEGIAN DEFENCE RESEARCH ESTABLISHMENr DTICPUox 25 - N-2007 Kjell, Norway ELEC ED

SSEP 0 21988H

-___ _ _ERi 889 2 C98]7,sbtbution UnWWtd

DYNAMICS OF THE POLAR MESOPAUSE AND LOWERTHERMOSPHERE REGION AS OBSERVED IN THE NIGHTAIRGLOW EMISSIONS

by

HANS KRISTIAN MYRABO

NDRE PUBL-88/1001

ISSN 0800-4412

FORSVARETS FORSKNINGSINSTITU17NORWEGIAN DEFENCE RESEARCH ESTABLISHMENT

P - 'ox 25 - N-2007 Kjeller, Norway

February 1988

To

OH, 02 and Na

without whom

this work would

not have been

possible

2NORWEGIAN DEFENCE RESEARCH ESTABLISHMENT (NORE) UNCLASSIFIEDFORSVARETS FORSKNINGSINSTITUTT (FF1)

SECURITY CLASSIFICATION OF THIS PAGE

POST OFFICE BOX 21 (Whs dotn r"6rd

11-2l JELLER, NORWAY

REPORT DOCUMENTATION PAGE

I) PUBL/REPORT NUMBER 2) SECURITY CLASSIFICATION 3; NUMBER OF

NDREEPUBL-88/1001 UNCLASSIFIED PAGES

Is) JOB REFERENCE 2a) DECLASSIFICATION/DOWNGRAOING SCHEDULE 100

FFIVM/119

41 TITLE

DYNAMICS OF THE POLAR MESOPAUSE AND LOWER THERMOSPHERE REGION ASOBSERVED IN THE NIGHT AIRGLOW EMISSIONS(DYNAMIKK OG TRANSPORTFENOMENER I DEN POLARE MESOPAUSE)

5) NAMES OF AUTHOR(S) IN FULL (iRms., first)

MYRAB0 Hans Kristian

6) DISTRIBUTION STATEMENT

Approved for public release. Distribution unlimited(Offentlig tilgjengelig)

7 INDEXING TERMSIN ENGLISH IN NORWEGIAN:

.1 Nightglow ,1 Nightglow

-I Polar cap bI Polaromrhdet

cI Atmosphere dI Atmosfaren

dl Mesopause d) Mesopausen

. Transport phenomena .) Transport og dynamlkk

THESAURUS REFERENCE: -

8) ABSTRACT (contine On m -;..ide if n ec.ryl

-This work utilizes night airglow emissions to deduce temperatures, dyna-mics, energetics, transport and photochemistry of the polar 80-110 kmatmospheric region. The morphological behaviour of the-polar 80-110 kmregion as seen in the night airglow emissions is best described by quasiregular to regular variations in the temperature and in the intensitiesof the emissions with periods ranging from minutes to a few days.Temperature amplitudes are seen from a few degrees up to ±50 K.Intensity changes up to several hundred percent may occur. Gravity wavesfrom below are generally found to be present In the region, beingresponsible for much of the short period variations. The long periodvariations are seen to be related to circulation changes in the lower

of the 80-110 km region by a ratio approximately twice as large In

9) DATE AUTHORIZED BY POSITION

TIS SIS ONly

26 Feb 88 ErikKlippenber - Director

UNCLASSIFIED

SECURITY CLASSIFICATION OF THIS PAGE

I,- daa. -tr.r

PTO

UNCLASSIFIEDSECURITY C .ASSF(CATION OF THIS PAGE

(wnen data entemd)

ABSTRACT (contiu.d)

amplitude as the heating at the 10 mbar level. The semidlurnal tide isfound to be dominant with a peak to peak amplitude of about 5 K, incontrast to model calculations. Effects from geomagnetic phenomena onthe energetics and dynamics of the region are not seen, and If present,have to be small or rare as compared to the influence from below. Thereis a mesopause temperature maximum at winter solstice. Pronounced dif-ferences in the day to day and seasonal behaviour of the odd oxygenassociated nightglows at the North and South Pole are found. This mayindicate fundamental differences at the two poles in the winter meso-pause region circulation and energetics.

UNCLASSIFIED

SECURITY CLASSIFICATION OF THIS PAGE

(when (Iota enteredl

I-

3

I

PREFACE WITH ACKNOWLEDGEMENTS

The work reported here were initiated by the author during 1092 whil:at the Norwegian Defence Research Establishment (NDRE), Division forPhysics. Most of the work, however, was carried out and completedduring two periods of sabbatical leave in 1982/83 and in 1985 at theGeophysical Institute of the University of Alaska. The author wishesto express his gratitude to colleagues at the Geophysical Institutefor support and encouragement. Particularly appreciated is the closeprofessional contact with Professor C S Deehr, whose advice have beenof great help throughout the preparation of this work. The author isalso indebted to Professor G G Sivjee for the use of spectrophoto-metric data from Longyearbyen and Poker Flats. Special thanks go toMrs Sheila Finch for her expert typing of various publicationmanuscripts. lhe engineers J Baldridridge and D Osborne are acknow-ledged for the great effort put down to keep instruments and computersworking around the clock.

The sabbatical periods at the Geophysical Institute were supported bythe NDRE, by the Royal Norwegian Council for Scientific and IndustrialResearch through fellowship programs, and by the Geophysical Institutethrough grants from the National Science Foundation. Thanks also toMr T Hvlnden, formerly Head of the Division for Physics and to theHead of the Division for Weapons and Equipment, Dr philos B Haugstadand to his Chief Scientist Dr philos P Thoresen for practical supportduring this work.

Kjeller, February 1988

H K Myrabe

Z..

4

CONTENTS

Page

I INDTRODUCTION AND OUTLINE 6

2 SUMMARY OF SCIENTIFIC ACHIEVEMENTS 9

2.1 General morphology 9

2.2 Gravity waves 10

2.3 Stratospheric warming events 12

2.4 Tides 13

2.5 Geomagnetic/auroral effects 13

2.6 Seasonal variations 14

2.7 Differences in winter solstice conditionsat the North/South Pole 16

3 FURTHER RESEARCH 17

References 18

4 PAPERS 23

4.1 Myrabe H K, C S Deehr and G G Sivjee, Large amplitudenightglow OH(8-3) band intensity and rotational tem-perature variations during a 24-hour period at 780 N,J Geophys Res, 88, 9255 (1983) 23

4.2 Myrabe H K, Temperature variations at mesopause levelsduring winter solstice at 780 N, Planet Space Sci, 32,249 (1984) 28

4.3 Myrabe H K, G J Romick, G G Sivjee and C S Deehr, Nightairglow OH(8-3) band rotational temperatures at PokerFlats, Alaska, J Geophys Res, 89, 9153 (1984) 37

4.4 Myrabo H K, C S Oeehr and B Lykekk, Polar cap OH airglowrorational temperatures at the mesopause during a stra-tospheric warming event, Planet Space Sci, 32, 853(1984) 42

5

Page

4.5 Myrabe H K, C S Deehr, R Viereck and K Henriksen, Polarmesopause gravity wave activity in the sodium andhydroxyl night airglow, J Geophys Res, 92, 2527 (1987) 47

4.6 Myrabe H K, Winter season mesopause and lower thermo-sphere temperatures in the northern polar region, PlanetSpace Sci, 34, 1023 (1986) 56

4.7 Myrabe H K and C S Deehr. Mid-winter hydroxyl nightairglow emission intensities in the northern polarregion, Planet Space Sci, 32, 263 (1984) 64

4.8 Myrabe H K K Henrlksen, C S Deehr and G J Romick,02 (blX+g-XE-g) atmospheric band night airglow measure-ments in the northern polar cap region, J Geophys Res,89, 9148 (1984) 74

4.9 Myrabe H K, C S Oeehr, G J Romick and K Henriksen, Mid-winter intensities of the night airglow 02(0-1)atmospheric band emission at high latitudes, J GeophysRes, 91, 1684 (1986) 80

4.10 Myrabe H K, Night airglow 02(0-1) atmospheric bandemission during the northern polar winter, Planet SpaceSci, (in press) (1987) 88

6

DYNAMICS OF THE POLAR MESOPAUSE AND LOWER THERMOSPHERE REG1LN ASOBSERVED IN THE NIGHT AIRGLOW EMISSIONS

SUMMARY

This work utilizes night airglow emissions to deduce tem-peratures, dynamics, energetics, transport and photochemistryof the polar 80-110 km atmospheric region. The morphologicalbehaviour of the polar 80-110 km region as seen in the nightairglow emissions Is best described by quasi regular to regu-lar variations in the temperature and in the intensities ofthe emissions with periods ranging from minutes to a fewdays. Temperature amplitudes are seen from a few degrees upto ±50 K. Intensity changes up to several hundred percent mayoccur. Gravity waves from below are generally found to bepresent in the region, being responsible for much of theshort period variations. The long period variations are seento be related to circulation changes in the lower atmosphere.Stratospheric warmings are generally associated by a coolingof the 80-110 km region by a ratio approximately twice aslarge in amplitude as the heating at the 10 mbar level. Thesemidiurnal tide is found to be dominant with a peak to peakamplitude of about 5 K, in contrast to model calculations.Effects from geomagnetic phenomena on the energetics anddynamics of the region are not seen, and if present, have tobe small or rare as compared to the influence from below.There Is a mesopause temperature maximum at winter solstice.Pronounced differences in the day to day and seasonal beha-viour of the odd oxygen associated nightglows at the Northand South Pole are found. This may indicate fundamental dif-ferences at the two poles in the winter mesopause region cir-culation and energetics.

1 INTRODUCTION AND OUTLINE

The upper mesosphere, mesopause and lower thermosphere region (i.e.

80-110 km) hereafter abbreviated the mesopaise region, has been and

still is one of the least understood and investigated regions of our

atmosphere. (Romick et al, 1986a; Romlck et al, 1986b; Myrabe, 1987).

The main reason for this seems to have been the relative inac-

cessability of this part of the atmosphere to most of the common

atmospheric measuring techniques. Further, the existing observations

have been almost entirely restricted to low and mid latitudes.

7

A basic parameter, such as the temperature, has only been crudely

known in the polar 80-110 km region. As an example of this, prior to

1983, knowledge of temperatures at the polar mesopause height was

mainly based on a few sporadic measurements of the OH night airglow

rotational temperatures, showing temperatures in the range 150-300 K

(Myrabo et al, 1983). Together with a small number of rocketsonde data

from Heiss Island (CIRA 1972), these constituted our observational

information of the atmospheric temperature profile for the 80-110 km

region in the polar cap area. Seasonal variations, amplitudes and pha-

ses of tides, possible effects of gravity waves, connection to stra-

tospheric warmings, auroral effects and the influence of these on the

energetics, small and large scale dynamics and photochemistry had

barely been investigated. Particularly information about the small

scale dynamical phenomena, being dependent on continuous ground based

monitoring, was lacking. This composed a serious limitation in our

understanding of the polar 80-110 km region and therefore also in our

understanding of the polar atmosphere as a whole (Romick et al, 1986b;

Myrabo, 1986; Myrabe et al, 1987).

Consequently, the first goal of this work was to gain basic knowledge

of the polar 80-110 km region, starting with a study of the short and

long term (seasonal) morphological behaviour of the temperatures of

the region, utilizing currently available instrumentation that

actually operated at the latitudes in question.

To resolve the short term dynamical phenomena and to obtain the needed

temporal coverage, ground based operating instrumentation was needed.

The only available gro-id based instrumentation operating within the

polar cap, potentially capable of obtaining temperatures of the 80-110

km region, were the optical spectrometers of the Geophysical Institute

of the University of Alaska. These instruments had operated at

Svalbard (780 N) for some years to study the high latitude and dayside

aurora (Deehr et al, 1980). A spectrometric data base was present.

Inspection of the data base looked promising. Using spectrometers to

measure night airglow features in the polar cap, and to obtain the

neutral temperature from the distribution of the rotational lines

(Kvifte, 1959) of the OH and 02 (0-1) Atm band emissions seemed thus

8

feasible. However, only the winter 80-110 km region could be studied

by this approach, due to the limitation in the observing method, i.e.

nightglow cannot be observed from the ground in the twilight and

daylight atmosphere. In the polar region this excluded the summer

period from equinox to equinox.

The problem with contamination by auroral emissions faced in earlier

high latitude airglow experiments (Meriwether, 1975; Kvifte, 1967) was

partly solved by selecting the vibrational bands, by monitoring in the

same spectra, one or more auroral atomic lines directly excited byprecipitating particles and by using a relatively high spectral reso-

lution. In aodition, the observing site was situated in the polar cap

region where contaminations from the aurora was mainly by atomic lines

and less frequent than in the auroral zone (Sivjee and Deehr, 1980;

Gault et al, 1981).

Also the ability to record short period intensity variations in the

emissions became greatly improved in the 1980s due to new detector

development (i.e. GaAs photomultipliers). Short time period dynamics

could thus readily be studied and unwanted effects by auroral emissionmore easily restricted to fewer spectra and thus to much shorter time

intervals than had been feasible earlier. The acquisition and reduc-

tion techniques were further refined to give reliable temperaturesfrom nightglow spectra down to a few minutes (Myrabe et al, 1987),

which is comparable with the Brunt V iss~lli period of the atmosphere

in the 80-110 km region.

The second objective of this work was to relate the experimental data

to physical phenomena in and outside the B0-110 km region, and if

possible to derive connections and parameters important for the dyna-

mical and photochemical state of the region. The presence of tides,

gravity wave activity, connections to stratwarms etc were thus

searched for, and different experimental setups were designed par-

ticularly to be optimal for studies of each of these phenomena.

The detailed scientific results of the research are given in ten

separate papers included in this report. The general morphological

behaviour of the mesopause region as seen through the night airglow

9

emissions is particularly, but not exclusively, depicted in

papers 4.2, 4.7, 4.8 and 4.9. Evidence and effects of gravity waves

are mainly discussed in papers 4.1 and 4.5. The behaviour of the

polar mesopause region temperature during stratospheric warmings is

reported in paper 4.4. Additional results relevant to stratospheric

warming events are found in papers 4.6 and 4.7. Tidal components in

the temperatures and night airglow intensities are discussed in

papers 4.2, 4.7 and 4.9. In paper 4.7 results from a search for pos-

sible geomagnetic/auroral interactions with the mesopause region are

reported. This is also touched upon in papers 4.9 and 4.10. Seasonal

variations are mainly dealt with in paper 4.6, 4.7 and 4.10.

Paper 4.10 also treats differences in the morphological behaviour ofthe odd oxygen associated night airglow at the North and South Polar

regions. References made in the first three sections to the papers in

section 4 have been underlined.

2 SUMMARY OF SCIENTIFIC ACHIEVEMENTS

As a result of this work one may conclude that the dynamics of the

polar mesopause region and its interaction with other regions of the

atmosphere are found to be more complex than previously believed

(Gdrtner and Memmesheimer, 1984; Offermann, 1985; Forbes, 1982a; Evans

and Nagy, 1981; Holton, 1979). Dynamics is here mainly manifested in

terms of temperature, and may indirectly be related to density,

pressure, wind-velocity etc.

2.1 General morphology

At Spitsbergen (- 780N, geographic latitude) intensities of the night

airglow emissions from different OH bands and from the 02 (0-1)

atmospheric band together with deduced temperatures from the rota-

tional distributions of the emissions at the respective emission

heights, i.e. the mesopause region, were obtained. The measurements of

the night sky covered 5 winter seasons, i.e. from autumn 1979 through

spring 1984. Intensities and temperatures were collected on time sca-

10

les from a few minutes to hours throughout this period. For part of

the winter season 1976 through 1978 mesopaube region temperatures were

also calculated from OH night airglow data at Poker Flats, Alaska

(~ 650 N, geographic latitude).

Both from the Spitsbergen and Poker Flats data it is found that the

morphological behaviour of the polar mesopause region, as manifested

in the deduced temperatures and in the intensities of the different

emissions, is best described in terms of quasi regular to itregular

variations (Myrabe, 1984; Myrabo, 1986). Temperature variations

ranging from a few percent to tens of percent (hundreds in the inten-

sities) and periods lasting from minutes to several days are seen.

Time intervals of several hours void of significant variations are

rare (Myrabo, 1986). This is in contrast to the morphology of the tem-

perature variations at mid and low latitudes (Takahashi and Batista,

1981; Wiens, 1974; Shefov 1969; Takeuchi et al, 1979; Tarrago and

Chanin, 1982). Consequently, the local dynamics and the larger scale

circulation patterns of the polar mesopause region have to be very

different from those at lower latitudes (Myrabo, 1984; Myrabe et al,

1984a; Myrabo, 1986). Temperatures at the mesopause height have been

measured from 298 K to 158 K during the winter season. This is up to

-90 K in deviation from the CIRA 1972 model atmospheric mean.

2.2 Gravity waves

Measurements of one or more night airglow emissions situated at dif-

ferent heights In the atmosphere were obtained during clear weather

situations. Sampling time scales from 30 minutes down to 5 minutes

(the latest being close to the Brunt-VdissV119 frequency at the meso-

pause region) were employed. From the simultaneously obtained emission

intensities and temperatures at the different heights evidence of gra-

vity wave activity could be searched for (see Fritts et al, 1984;

Fritts, 1984; Hatfield et al, 1981; Fredrerick 1979; Noxon, 1978;

Krassovsky, 1972).

It was found that in a large fraction of cases, shcrt period

variations in temperature and intensities displayed similar behaviour,

11

i.e. when strong variations in the temperature occurred, the same was

-er in the intensity, and with approximately equal periods of

variations (Myrabo et al, 1983; Myrabe, 1984). In cases where several

emissions at slightly differert heights were measured, similar

variations, I e same periods, were often seen both in temperature and

intensities at the different heights (Myrabe et al, 1987). Most fre-

quently, in phase variations were observed between intensity and tem-

perature. From analysis of the power spectra of the temperature

variations, an approximately k-5 / 3 dependence on the temperature

variations with wave number was found in a large number of cases,

implying the presence of a quasi saturated or saturated gravity wave

field with breaking gravity waves (Myrabe et al, 1987).

The bulk of the temperature and intensity variations, with periods in

the time domain - 5 minutes to 3-4 hours, may therefore be explained

in terms of gravity waves penetrating or breaking in the 80-110 km

region (Frltts et al, 1984). Both examples of simple cases with

monochromatic waves and more complex cases, with combinations of

several modes including breaking waves were found (Myrabe et al, 1983;

Myrabo et al, 1987). Even a case of standing waves near the Brunt

Vdissalla frequency, probably excited by longer period gravity waves

(Tuan et al, 1979), was observed (Myrabe et al, 1987).

Upper levels of gravity wave induced vertical eddy diffusion coef-

ficient in the range - 106-107 cm2/s were deduced from the intensity

variations of the emissions (Myrabe et al, 1987). This is a pro-

nouncedly larger eddy diffusion coefficient than normally seen at

lower latitudes (Thrane et al, 1985; US Standard Atmosphere, 1976) and

is possibly in disagreement with the predictions by Llndzen (1981)

quoting a minimum in the eddy diffusion coefficient near the mesopause

at high latitudes. The large eddy diffusion coefficient may also be

taken as an additional indication of very strong gravity wave activity

in the polar winter atmosphere. Heights of the sodium emissions and

the OH(6-2) emission were found to be separated by about 1 km, indi-

cating the OH(6-2) emission to be situated approximately at 90 km in

the polar winter atmosphere (Myrabo et al, 1987). This is - 5 km

higher than average quotations for lower latitudes (Taylor et al,

12

1987). It also places the polar winter mesopause at - 90 km (Myrabo

and Deehr, 1984).

Finally the occurence of gravity waves, penetrating or breaking, is

found to be an almost omni-present feature of the polar 80-110 km

region (Myrabe et al, 1986), in contrast to mid and low latitudes

(Tarrago and Chanin, 1962). The sources of the waves are in no cases

found to be associated with geomagnetic or auroral phenomena, thus

implying that the sources of the waves are generally the lower and

middle atmosphere, i.e. the mesosphere, stratosphere or troposphere

(Myrabo et al, 1987; Myrabe, 1984; Myrabo and Deehr, 1984).

2.3 Stratospheric warming events

In the northern hemisphere the stratosphere and lower mesopause is

known to undergo a circulation reversal once or more during mid-winter

before the final circulation reversal takes place in spring (Labitzke,

1977). In connection with the mid-winter reversal there is found to be

a heating of the stratosphere (stratospheric warming) with a corres-

ponding cooling of the mesosphere up to at least - 60 km height

(Labitzke and Barnett, 1985).

To study possible effects of the stratospheric warnings on the polar

mesopause region (i.e. 80-110 km), 12-24 hours averaged mesopause tem-

peratures were compared to temperatures of the stratosphere and

mesosphere (i.e. up to - 60 km height). The stratospheric/mesospheric

temperatures were available from the Nimbus satellite radiometers and

covered the polar cap area (Drummond et al, 1980).

It was found that the polar 80-110 km region cooled considerably (at

least up to the 90-95 km region), during mid-winter stratospheric

warming events (Myrabe et al, 1984a; Myrabe et al, 1986). The ampli-

tude of the cooling at the mesopause was found to be approximately

twice as large as the amplitude of the associated warming in the stra-

tosphere at the 10 mbar level (Myrabe et al, 1984a). Daily mean tem-

perature as low as 170 K were observed at the mesopause during a stra-

tospheric warming at solstice. After the stratospheric warming

vanished, the cooling of the mesopause was followed by a heating above

the "average" temperature by nearly the same amplitude as the cooling

(Myrabo, 1986).

Outside periods of major and minor stratospheric warnings, the long

period (day by day) variations of the temperature at the mesopause

height were found to have similar or equivalent periods of variations

as the day to day variations of the temperature in the mesosphere and

in the stratosphere (tMyrabo, 1984; Myrabo, 1986). This seems to imply

a common origin and that the long tern variations in the temperatures

at the mesopause are somehow connected to the circulation and tempera-

ture changes in the underlying atmosphere.

2.4 Tides

Both emission intensities and temperatures deduced from the emissions

at different heights have been analysed for tidal components (Myrabo,

1984; Myrabo and Deehr, 1984; Myrabo et al, 1986). Fourier analysisand superimposed epoch methods have been employed. The diurnal tide

component at the mesopause, that according to the latest tidal modelsshould dominate in polar region (Forbes, 1982b), is not found to be

present above the noise level, i.e. 1 K. On time scales from a fewdays to about 20 days the semidiurnal tide component is seen to domi-

nate with peak to peak temperature amplitudes in the range 4 to 8 K.The absence of the diurnal component may be explained by strong

interactions of this mode with gravity waves (Beer, 1975). However,

the observed amplitude of the semidlurnal mode is far larger than pre-

dicted by the current tidal models. Even higher semidlurnal amplitudes

are currently reported to exist over short time intervals (Waterscheld

et al, 1986). Tidal models should therefore be revised to match the

observations better.

2.5 Geomagnetic/auroral effects

Including more than one month with almost continuous temperature data

and a good coverage in the intensity data, correlation studies of OH

temperatures and intensity variations with geomagnetic Kp, Ap Index,

14

with sudden storm commencement and with tho interplanetary magnetic

field (Bz) showed no significant correlation (Myrabe and Deehr, 1984).

However, for truncated data set taken from the same period (i e case

studies) both positive, negative or insignificant correlation could be

produced.

The long disputed geomagnetic/auroral influence on the temperature

and dynamics in the 80-110 km region (Krassovsky, 1956; Saito, 1962;

Shefov, 1969; Maeda and Aikin, 1968; Brekke, 1977; Takahashi and

Batista, 1981; Baker et al, 1985) may thus be settled (Myrabe, 1984).

In view of the results above and from corresponding analysis of the

total amount of temperature and intensity data for the other seasons,

it may be concluded that the atmospheric temperature and the dynamics

of the polar mesopause region, at least up to the - 90 km level, are

dominated from below (Myrabe, 1986; Myrabe et al, 1987). Influence

from the aurora, if present, have to be rare or too small to be

detected as compared to the competing interactions from the atmosphere

below. This limits the amplitudes of the perturbations from auroral

sources to be generally one order of magnitude or more smaller than

the perturbations originating In the underlying atmosphere.

2.6 Seasonal variations

Daily averaged temperatures together with 5 or 10 days running avera-

ges were calculated from individual 30 minutes and 1 hour temperatures

in order to study the seasonal variations of the mesopause region tem-

peratures (Myrabo, 1984). Tidal components were removed. Forming 5 and

10 days running averages also removed the day to day variations mainly

connected to the larger scale variations in the circulation (Myrabo,

1984; Myrabe et al, 1984a). In the polar cap area, (I e Spitsbergen,

780N geographic) temperatures for 5 seasons were used, covering - 4

months around winter solstice. Also temperature data obtained from

Poker Flats, 650 Ngeographic, for the seasons 1976 through 1978 have

been analysed with respect to seasonal variations (Myrabe et al,

1984c). Mainly OH emissions have been used in obtaining the tem-

peratures. At 78°N a relatively low temperature in late December

followed by a very warn mesopause in January is found to be consistent

15

for all four winter periods (Myrabo, 1986). A hypothesis is that this

is associated with changes in the transmission of gravity waves to the

upper mesosphere in connection with stratospheric and lower

mesospheric circulation changes.

The average temperature maximum in January was found to be 223 K, data

from five winter seasons being used. This is about 15 K higher than

the CIRA 1972, 700N, January model atmoshpere at 90 km. The tem-

perature variation from November through March is best described by a

standing wave with a period of about 50 days, peaking in early

December and In mid January (Myrabe, 1986). The peak to peak ampli-

tide Is about 20-30 K. The corresponding local minimum around

solstice Is found to be present both during winters with and without

major stratospheric warmings. There is, however, both at the 780N and

at the 650N site an average winter solstice maximum, pointing clearly

to lower temperatures both In the autumn and In the sprng "°-ab

1986; Myrabe et al, 1984c). At the 650N site there Is no clear local

minimum around solstice (Myrabe et al, 1984c).

The Intensities of the 02(0-1) Atm band have been used to deduce oxygen

concentration at the 95 km level (Myrabe et al, 1984b; Myrabe et al,

1986; Myrabe, 1987). No clear minimum In the deduced oxygen con-

centration Is found at winter solstice. The 2-3 months of data around

winter solstice for the two winters of the 02(0-1) band observations

is best described as consisting of strong enhancements In the inten-

sity lasting for days and being superimposed on a constant background

level. This Is contrary to model calculations estimating that a very

clear minimum In the oxygen concentration should occur around winter

solstice (mainly due to the lack of production of odd oxygen through

dissociation of 02 by solar ultraviolet radiation in the polar night)

(Elphlnstone et al, 1984). The lack of a clear minimum indicates that

the origin of the oxygen is outside the terminator. Consequently, a

more effective transport from outside the solar terminator Into the

dark polar cap than previously thought has to be awoked. Calculations

should therefore be reconsidered.

16

2.7 Differences in winter solstice conditions at the north/southpole

Differences in the variations of the emission intensities of the odd

oxygen associated nightglows at the two poles might be used to indi-

cate differences in the dynamics.

To study this, absolute intensities of the 02(0-1) Atm band emission

were obtained from measurements at Spitsbergen (780N) (Myrabo, 1987).

Two winters, I e 1982/83 and 1983/84, of measurements covering

~2h months around winter solstice have been analysed. Contamination by

aurora was removed, securing that the obtained 02(0-1) band inten-sities only reflected the variations in the night airglow component of

the 02(0-1) band.

The 02(0-1) band absolute intensitites from 780N were compared with

night airglow absolute intensities of the 015577 emissions over the

South Pole, reported by Ismail and Cogger (1982), utilizing satellite

data from the ISIS 2 limb scanner. As the 02 Atm band emission and

the 015577 emission co-varies, a comparison reveals any differences in

the odd oxygen variations at the two poles.

Comparing the two sets of data shows striking differences both in the

seasonal and in the day to day variations (Myrabe, 1987).

While the intensities at the South Pole show a relatively smooth

decline towards .a minimum around winter solstice, the emission inten-

sities in the northern polar region reveal$ variations with very

strong emission enhancements lasting for days up to a week or more.

The seasonal trend at the North Pole does not show any clear minimum

in the emission intensities and thus no minimum in the odd oxygen con-

centration.

In the south polar region there are no major stratospheric warmings

during the mid-winter and thus a more consistent circulation pattern

in the stratosphere and in the mesosphere (Shaebend, 1977; Labitzke,

1981). If it Is assumed that the enhanced levels of emissions are con-nected to changes in the circulation and transport within and into the

17

upper mesoshpere and lower thermosphere, the observed differences in

the emission intensities reflect differences in the circulation and

transport in the two hemispheres (Myrabe, 1987). Whether this is

further connected to differences in gravity wave activity caused by

differences in the topography (i.e.land masses) in the two polar

regions or of other origin is open to question.

3 FURTHER RESEARCH

This work has illuminated a polar mesopause region with a dynamical

complexity; involving strong tidal forcing, almost ever present gravity

wave activity, and longer period variations (I e a day or more) in

temperature, density and circulation connected to variations in the

atmospheric regions below. One of the surprises was to find the

atmosphere at mesopause height not particularly affected by geomagne-

tic or auroraly related activity, but with the variations in tem-

perature and density rather controlled by interactions with the atmos-

pheric regions below, all the way down to the stratosphere.

In order to understand the circulation, transport, energetics and che-

mistry of the entire polar winter atmosphere, it is crucial to iden-

tify the transition region in the atmosphere where the atmosphere

changes from mainly being governed by interactions from below, to

regions where interactions caused by geomagnetic activities dominates.

This region is certainly above - 90 km. Monitoring the penetration

and breaking of gravity waves from below and through the 90 km region

and upwards in addition to monitoring geomagnetic activity should

therefore be given priority to identify this transition region. It is

also urgent to get a better measure of the amount of energy deposited

by the almost ever present gravity waves in the polar winter atmos-

phere.

Further, efforts should be made to identify if a coupling between gra-

vity waves and tides exists in the polar mesopause region

(Walterscheid et al, 1986), and eventually be able to estimate the

18

amount of energy transferred from the gravity waves to the tides.

Monitoring the presence of gravity waves over large spatial scales, -

before, during and after stratospheric warmings, to see if gravity

waves might play a part in the stratospheric warming events is another

important item that should be pointed out.

Monitoring of a restricted number of nightglow emissions using ground-

based spectrometric equipment is scarcely likely to answer the

questions outlined above. However, a combination of spectrometric

observations from a number of ground stations with the additional use

of Rayleigh lidars and monochromatic imagers reinforced possibly by

in-situ measurements might provide the information needed. Programs

aimed in this direction seem now to be in progress (Romick et al,

1986a).

References

Baker D F, A F Steed, G A Ware, D Offermann, G Lang and H Lauche,Ground-based atmospheric infrared and visible emission measurements,J Atm Terr Phys, 47, 133 (1985)

Beer T, Atmospheric waves, A Hilger, London (1975)

Brekke A, Auroral effects on neutral dynamics, in Dynamical and Chemi-cal Couping Between the Neutral and Ionized Atmoshpere, edited byGrandal B and Holtet J A, p 313, Reidel, Dordrecht (1977)

CIRA, COSPAR Working Group 4, COSPAR International Reference Atmos-phere, Pergamon Press, Oxford (1972)

Deehr C S, G G Sivjee, A Egeland, K Henriksen, P E Sandholt, R Smith,P Sweeney, C D Duncan and J Gilmer, Ground-based observations of FRegion associated with the magnetospheric cusp, J Geophys Res, 85,2185 (1980)

Drummond F R, F T Houghton, G D Peskett, C D Rodgers, M F Wale, FWhitney and E J Williamson, The stratospheric and mesospheric sounderon Nimbus 7, Phil Trans R Soc Lond , A 296, 219 (1980)

Elphinstone R D, J S Murphree and L L Cogger, Dynamics of the lowerthermosphere consistent with satellite observations of 5577A airglow:II., Atomic oxygen, local, turbulence and global circulation results,Can J Phys, 62, 382 (1984)

19

Evans J V and A F Nagy, Solar-Terrestrial Research for the 1980's,National Academy Press, Washington DC (1981)

Forbes F M, Atmospheric tides; the solar and lunar semidiurnal com-ponents, J Geophys Res, 87, 5241 (1982a)

Forbes F M, Atmospheric Tides I., Model Description and results forsolar diurnal component, J Geophys Res, 87, 5222 (1982b)

Frederick J E, Influence of gravity wave actlvly on lower ther-mospheric photochemistry and composition, Planet Space Sc, 27, 1469(1979)

Fritts D C, Gravity wave saturation In the middle atmosphere: A reviewof theory and observations, Rev Geophys Space Phys 22, 275 (1984)

Fritts D C, M A Geller, 8 B Balsley, M L Chanin, I Hirota, J R Holton,S Kato, R S Llndzen, M R Schoebend, R A Vincent and R F Noodman,Research status and recommendations from the Alaska Workshop on Gra-vity Waves and Turbulence in the middle atmosphere, Fairbanks, Alaska,18-22 July 1983, Bull Am Met Soc, 65, 149 (1984)

Gault W A, R A Koehler, R Link and G G Shepherd, Observations of theoptical spectrum of the dayside magnetospheric cleft aurora, PlanetSpace Sc, 29, 321, (1981)

Gartner V and M Meimesheimer, Computation of the zonally-averaged cir-culation driven by heating due to radiation and turbulence, J Atm TerrPhys, 46, 755 (1984)

Hatfield R, T F Tuan and S M Silverman, On the effects of atmosphericgravity waves on profiles of H, 03 and OH emissions, J Geophys Res,86, 2429 (1981)

Holton F R, An Introduction to dynamical meteorology, Academic Press,Orlando, Fla (1979)

Ismail S and L L Cogger, Temporal variations of polar cap 015577Aairglow, Planet Space Sci, 9, 865 (1982)

Krassovsky V I, Infrared night airglow as a manifestation of the pro-cess of oxygen recombination, in Airglow and Aurora, edited by E BArmstrong and A Dalgarno, p 195, Pergammon Press, New York (1956)

Krassovsky, V I, Infrasonic variations of the OH emissions in theupper atmosphere, Ann Geophys, 28, 739 (1972)

Kvifte G, Nightglow observations at As during the IGY, Geophys Publ,20, 1 (1959)

Kvifte G J, Hydroxyl rotational temperatures and intensities in thenightglow, Planet Space Scl, 15, 1515 (1967)

20

Labitzke K, Stratospheric - mesospheric mid-winter warmings, in Dyna-mical and Chemical Coupling Between Neutral and Ionized Atmosphere,edited by Grandal B and Holter J A, p 17, Reidel, Dordrecht (1977)

Labitzke K, Stratospheric-mesospheric mid-winter disturbances; A sum-mary of observed characteristics, J Geophys Res, 86, 9665 (1981)

Labitzke K and J J Barnett, Stratospheric and mesospheric large-scaleheight and temperature field during the November/December 1980 EnergyBudget Campaign, J Atm Terr Phys, 47, 173 (1985)

Lindzen, R S, Turbulence and stress due to gravity wave and tidalbreakdown, J Geophys Res, 86, 9707 (1981)

Maeda K and Aikin A C, Variations of polar mesopheric oxygen and ozoneduring events, Planet Space Sci, 16, 371 (1968)

Meriwether J W, High latitude airglow observations of correlated shortterm fluctuations in the hydroxyl Melnel 8-3 band intensity and rota-tional temperature, Planet Space Sci, 23, 1211, (1975)

Myrabe, H K, C S Deehr and G G Sivjee, Large-Amplitude NightglowOH(8-3) band Intensity and rotational temperature variations during a24 hour period at 780N, J Geophys Res, 88, 9255, (1983)

Myrabe H K, C S Deehr and B Lybekk, Polar cap OH airglow rotationaltempreatures at the mesopause during a stratospheric warming event,Planet Space Sc, 32, 853 (1984a)

Myrabo H K, K Henriksen, C S Deehr and G J Romick, 02 (bl Z+ g - X3 Z- g)atmospheric band night airglow measurements in the northern polar capregion, J Geophys Res, 89, 9148 (1984b)

Myrabe H K, G J Romick, G G Slvjee and C S Deehr, Night airglowOH(8-3) band rotational temperatures at Poker Flats. Alaska, J GeophysRes, 89, 9153 (1984c)

Myrabo H K, Temperature variations at mesopause levels during wintersolstice at 780N, Planet Space Sc, 32, 249 (1984)

Myrabe H K, and C S Deehr, Mid winter hydroxyl night airglow emissionsintensities In the northern polar region, Planet Space Sci, 32, 263(1984)

Myrabe H K, C S Deehr, G J Romick and K Henriksen, Mid-winter inten-sities of the night airglow 02(0-1) atmospheric band emissions at highlatitudes, J Geophys Res, 91, 1684 (1986)

Myrabo H K, Winter-season mesopause and lower thermosphere tem-peratures in the polar region, Planet Space Sci, 34, 1023 (1986)

Myrabo H K, C S Deehr, R Viereck and K Henrlksen, Polar mesopause gra-vity wave activity in the sodium and hydroxyl night airglow, J GeophysRes, 92, 2527 (1987)

21

Myrabo H K, Night airglow 02(0-1) atmospheric band emission during thenorthern polar winter, Planet Space Sci (in press) (1987)

Noxon, J F, Effects of internal gravity waves upon night airglow tem-peratures, Geophys Res Letters, 5, 25 (1978)

Offermann D, The Energy Budget Campaign 1980; Introductory review, JAtm Terr Phys, 47, 1 (1985)

Romick G J, C S Deehr, M A Bondi, J C Foster, T L Killeen, R W Schunk,C F Sechrist Jr, B A Tinsley and D G Torr, Coupling, Energetics, andDynamics of Atmospheric Regions "CEDAR", Vol 1: Overview, Report tothe Aeronomy Program of the National Science Foundation (1986a)

Romick G J, C S Deehr, M A Bondi, J C Foster, T L Killeen, R W Schunk,C F Sechrist Jr, B A Tinsley and D G Torr, Coupling, Energetics andDynamics of Atmospheric Regions "CEDAR", Vol II, Report to the Aero-nomy Program of the National Science Foundation, p 18 (1986b)

Saito B, Unusual enhancement of night airglow intensity at low latitudes on November 13, 1960, Antarctic Rec, 14, 8 (1962)

Schoebend M R, Stratospheric warmings, observations and theory, RevGeophys, Space Phys, 16, 521 (1977)

Shefov N N, Hydroxyl emissions of the upper atmosphere - I. The beha-viour during solar cycle, seasons and geomagnetic disturbances, PlanetSpace Sci, 17, 797 (1969)

Sivjee G G and C S Deehr, Differences in polar atmospheric opticalemissions between mid-day cusp and night-time auroras, in Explorationof the Polar Upper Atmosphere, edited by C S Deehr and J A Holtet,p 199, Reidel, Hingham, Mass (1980)

Takahashi H and P P Batista, Simultaneous measurements of 0H(9.4),(8.3), (7.2), (6.2) and (5.1) bands in the airglow, J Geophys Res, B6,5632 (1981)

Takeuchi I, K Misawa, Y Kato and I Aoyama, Rotational temperaturesand intensities of OH(6-2) and OH(8-3) bands in the nightglow, J AtmTerr Phys, 41, 387 (1979)

Tarrago A and M L Chanin, Interpretation in terms of gravity waves ofstructures observed at the mesopause level by photogrametry and lidar,Planet Space Sc, 30, 611 (1982)

Taylor M, M A Hapgood and P Rothwell, Observations of gravity wavepropagation in the 01 (557.7 nm), Na (589.2 nm) and the near infraredOH nightglow emissions, Planet Space Sod, 35, 413 (1987)

Thrane E V, 0 Andreassen, T Blix, B Grandal, A Brekke, C R Philbrick,F J Schmidlin, H U Widdel, U Von Zahn and F J LUbken, Neutral air tur-bulence in the upper atmosphere observed during the Energy Budget Cam-paign, J Atm Terr Phys, 47, 243 (1985)

22

Tuan T F, R Hedinger, S Silverman and M Okuda, On gravity wave inducedBrunt-Vdisslla oscillations, J Geophys Res, 84, 393, (1979)

U S Standard Atmosphere, 1976, NOAA-S/T76-1562, Washington 0 C (1976)

Walterscheid R L, G G Sivjee, G Schubert and R M Hamwey, Large-amplitude semidurnal temperature variations in the polar mesopause;evidence of a pseudotide, Nature, 324, 347 (1986)

Wiens R H, Diurnal variations of the (8-3)/(5-0) intensity ratio ofnightglow OH at Agi Udri, Planet Space Sc. 22, 1059 (1974)

-:~.,

41

24

)1L RN At. i0l- GEiiPH SIC \[ RLS[Ali 1 ul \v I),, S ivi1 %"v' Fi flhEl ; v

Large-Amplitude Nightglow OH (8-3) Band Intensityand Rotational Temperature Variations During a 24-Hour Period

at 78,N

H. K. MYRAO..' C. S. DEEHR. AND G G. Sivtrr

Geophvi, ,

l Istitat. L'erett, - I t41aka

We report results from a coninuous 24-hour measurement of the OH .v-31 band mission in thenightilow at -S.4N A mean temperatu t of 2-3 K and a mean band ntensitv of 596 R were obseree.Estreme temperature variauons were tien with amplitudes up to - -0 K from he mna. It is suggested'hat these variations are related to the passage of internal gravity aes If so. the extreme amtplitudes ofthe varations might eply that the OH emoting layer is situated above 90 km ai this latitude in Januaryfihe deduced " vat A I T fn favor the ozone mechanism to be responsible for the OH emtsstonwith the posatbilft of an additional mechanism contributing up to 5'.

I INTRODLCTION These measurements. although providing vatuable data onSince M,!emel [19501 -densitied the OH band system in the temperature and ntenaity with latitude only gave a snapshot

night airflow. a large number of investigators have tred to (and therefore a very limited understandingi of OH morpholoelucidate the physics and chemistr behind the different pat- gy at high latitudes For this purpose continuous measure-tens of behavior of the OH nightglow. ments from the ground and if possible linked with satellite

The study of offemissions alone or together with other measurements from above will provide the best datamghtgJow emissions prosildes important information on tern- Satellite measurements alone have been utilized. Reedperature. density. composition. and behavior of the upper [1976] reported indications of strong polar enhanced OHmesosphere OH emission measuremnents may also be used as ermssions. Further. Walker and Reed [1976] found the effectsan important tool for observing the effects of propagating from an early major stratospheric warmtng in December 196'gravity waves through the atmosphere [Krassovskv, 1972: on the OH emission intensity to be largest between -0 andKrassoskv and Shaqear. 1974; Voxon. 1978 Frederick. 1979: 80'N, Unfortunately. the measurements were broadband opti-Hatfield et al. 1981]. cal measurements. and contarmnation by aurora. as was

AImost all of the observations of OH have previously been pointed out in both these works, cannot be ruled out.made at latitudes less than approximately 70' Measurements The lack of observations at high lautudes is due. in part. toabove latitude 55 are sparse compared to lower latitudes, the problem of contamination by auroral emissions. In the

The only measurements at high ie.. above 70')

latitudes auroral zone this is caused by aurora molecular bands whichseem to be those of Charberlain and Oliver [1953] in Thule. cover almost the entire visible and near infrared [Vallance-the flights by the U.S. Air Force KC-135 jet aircraft in Mach Jones. 1974, Mertwether. 1975].1963 and f964 reported by Vosos [1964]. and the 1968 However. at higher latitudes i.e.. into the polar cap regioni.NASA auroral ,wrborne expedition [Sivyee et al.. 1972]. auroral molecular emissions diminish and the aurora enits

Chamberlain and Oliver's data consited only of a few spec- mainly atomic hnes [Sirfee aid Deehr. 1980] Thus a normal.,ra with relatively coarse spectra and temporal resolution. high-responsivity auroral spectrophotometer operated at aThey found an OH rotational temperature ranging from 300 resolution of 2000 t4-A bandwidth) may be used to measureto 350 K which is believed to be approximately 30-40 K too most of the OH bands and lines clearly resolved from thehigh due to erroneous molecular constants used in the re- auroral emissions. Because of the high geographic latitude ofduction of the data [Kcilte, 1959] the station, observations may be caned out continuously for

Noxon's results, which are based munly on two spectra 24 hours per day for 2 months around each winter solsticeaveraged in time and space over latitudinal ranges 77'-85N The auroral observatory at Longyearbyen may therefore beand 69 -85 N gave temperatures 160 and 185 K. respectively regarded as a suitable site for high latitude OH observatioisHe also reported a significant decrease in intensity and tem- from the ground.

perature with increasing lantudes.The NASA airborne observations performed between Janu- 2. lNSTR NTATON. OBmSVATIONS. AND REDUCTION

ary and March 1968 consisted of a larger number of flights The OH emission data employed in this work were part ofand thus provided more data pertinent to the pictue of the the measurements undertaken dunng the multinational Sval-latitudinal dependence of the temperaure. Latitudes TO'-78'N bard aurora expedition beginning in 1978 at an observatorygave temperatures in the range 190-245 K with a mean close [Deehr et al. 19801 close to Longyearbyen on West Spitsber-to 225 K (Stapee et al. 1972] gen (geographic latitude 78 4-Nt

The I-n and t-m high-throughput Ebert-Fastie spec-'On leave from Norwegan Defense Research Establishment. trophotometers are coupled to a minicomputer and record inNow at National Science Foundation the photon-counting mode The I-m instment. u for

thes measurements reported here. is further described by DickCopyright 1993 npy 'he American Geophysical Lnion et aL [1970] and Sitjee ei al. (1972].

Paper number 1AO690 From three seasons of data. i.e., 1979 1980. 1980 1981. and.144aJ2' 93 003Aa)690SO5 00 1981 1982. we have been able to uttliz limited penods dunng

9255

25

9256 MYRA1a0 er Ai. Ni GWw OH 18-11 BA-ND AT 8 N

to i-tJ3 that t here wan no contamination of Elhe data frioms the N2 11Pe izi ,r2. r2 ,3[ and other aurora) molecular miinsons.

ao on~~i -3 R.ESULTS AND Discussiol,

3: 3.1. Intensities and Rotational Temperatures

--~2 2 Intensities and rotational temperatures an denved from themeasured OH (8-3) band lines between January 6 and 7. 1981,

o 0600-0550 UT are presented in Figure 2. They are given re-spectively as broken and solid lines. The mean 24-hour rota-

, 0 , ,tional temperature is 237 K and the mean intensity of the OH-i00 NGTvu (8-3) band s 596 R. For comparison we may mention that the

Fig. I An example of the spectra used for deducing temperature average temperature and band intensity found by Takahasiand intensity Each spectrum is acquired dunng a [0-in period et al. [1977] at ZYS for the period July 1972 to October 1974suramming 75 individual 8-s scans. In addition to the OH [8-3) lines, was 179 K and 408 R, respectively.also the forbidden oxygen doublet at 7320,7330 A onginating m the No obvious diurnal or semdiurnal trend is visible in theaurora is indicated. 24-hour record presented here.

Previous temperature measurements at latitudes 70'-85'Nclear weather where the mtruments were run in a suitable cover the range 160 to approximately 300 K_ However. such amanner for OH measurements. Of partcular Interest are oh- large variation in temperature as seen here I t '0 K), has notservations from a single 24-hour period of continuous been reported from a single station and a linuted time span.measurement of the OH (8-3) band where the OH intensity Krassovsky and Shagaer [1977] report extremes of -50 Kand temperature were observed to change conaiderably. Un- during the propagation of internal gravity waves through thefortunately, none of the other usable penods in the data col- mesosphere.lection contained a full 24-hour period with as high a time There is no way of ascertaining that the variations seen hereresolution. How often this behavior occurs is therefore not yet are due to the propagation of internal gravity waves. To doknown. that would have required simutaneous determination of tern-

The spectrophotometer was pointing toward zenith and the peratures at least at two altitudes [Nowon. 1978] or measure-a. 7285-A-7350-A region was scanned in 8 s by using the ments taken simultaneously at three places in the sky [Krass-spectrophotometer in the second order with a I-m slit corre- ovskv. 1972; Krassovsky and Shaqaer. 1974].spvnding to a bandwidth of 1.5 A and a resolution of 4900. However. it is difficult to find another mechanism thatEach scan was recorded on magnetic tape. In most cases, an could lead to such extreme variation in the temperature andadequate signal-to-noise ratio was acquired by integrating 75 intensity. An indication that at least part of the variations seenindividual scans over 10 mn. An example of the quality of the here are due to the propagation of gravity waves is the corre.data may be seen in Figure 1. lation between temperature and intensity. A correlation be-

Rotational temperature was calculated by Kvifte's method. tweem temperature and intensity was found by Shaqae [1974]using the intensity ratio of the P,(2) and P,(3) lines [Keife. and Krassovsky and Shagaev [1977] during the propagation of1959]. The band intensity was then based on the measured gravity waves, while under conditions where gravity wavesintensity of the P,(2) line together with the calculated temper- were not likely, poor or not correlation was found [Harrisonature. Absolute intensity calibration was perforsed in the field et al.. 1971; Takahashi er al., 1974; Takeschi ef al.. 1979].by using a standard lamp and a diffusing screen. Figure 3 shows the intensity-temperature plot. The correlation

Probable error in the calculated temperature is estimated to coefficient found is 0.56 using N - 286 points. The best fitbe ± 5 K caused by the ancertainty in defining the P (2) and curve with a slope AlAT - 1.93 RK is indicated.P3) background levels, while the possible error in the abso- It has been shown theoretically by Weinstock [1978] thatlute intensity is approximately 20%. mainly caused by the the relation between temperature and intensity during gravitycalibration uncertainty, wave passage is complex. It might show phase shifts and

The aurora could easily be momtored through the depend upon the scae herght of the respective emnor compo-7320/7330-A Oil lines. Additional care was taken to assure nent fe.g.. H) relative to total scale hetght at the particular

520 6_1.- .-

06 n7 08 09 '2 3a -4 15 6 8 '9 20 21 22 22a " 2

iNIV EQSAL 'i M'HRS

Fig 2. Intensity and routonaj temperatures as derived from mesured OH 8-31 bands during Jalury 6-7. 1981. atLongyearbyen, Svalbard. The intensity and rotaional temperature as obtained each 5 mis from I0 mis of scans 50o,overlapping) are plotted as points. Straight lines are drawn between sueive points.

Ic

26

-u /q

Fil 3 Correlaton btween OH 18-3) rotational temperature ad

qt¢nstty durng8 the 24-hour pe'rnod as obtained employing es h point

5-mln intral) in Figure I diretly The corelation analysis w-

:performed by usi6 g . - 286 poits, vtt a coreelation Cefloet Fig. 5 Power spectra of the ati.ons of OH 8- 31 rotaiionai [e-

- 0.36. perature and intensity A Hanning window was employed.

altitude. The coherent motion of the emitting particles (i.e- ments of different vibrational levels of OH bands durtng De-OH) moving in and out of the field of view may also influence cember 1980 and Januaxy 1981, a positive temperature gradi-the observed relation. According to Weinstock. nonlinear cor- ent was found for the OH emitting region [. .Wao. 1983].rections to the oscillations may be neglected at OH height. i.e. This may be interpreted to mean that the emission region !?85-90 km. situated at or slightly above the mesopause. Owing to the

An attempt to improve the correlation between temperature polar jet, the mesopause is believed to be at a higher altitudeand intensity waves was made by phase shifting the temper- of midwinter above the polar regions than elsewhere [Rodgers.ature with regard to the intensity. The maximum obtained was 1977]. In view of the above, a reasonable interpretation is that0.59. i.e.. not sigmficantly different from that with no phase all or part of the larger amplitude found is due to the highershift. Slight phase shifts between temperature and intensity altitude of OH emitting layer. This contradicts the suggestionand/or changes in the intensity-temperature amplitude ratio of Sijee etr al. (1972), who suggested that the height of theduring the observation period as ught be inferred from OH-emitting layer is independent of latitude. The latter resultsFigure 2 may help explain the relatively low correlation found, are, however, based on a comparison of a very linuted amount

This does not eliminate gravity waves as the source because of temperature data with temperatures taken from a model.correlation coefficients in the same range (i.e. 0.5-0.9, KrasS- Figure 5 shows a Founer analysis of the temperature andouksy and Siagaee [197 ) are typical of gravity wave data. A intensity variations, which exhibit a decreasing energy in thehigher coefficient (i.e. up to 0.8) mght be obtained from our higher frequencies. This is also in agreement with Kransovskydata by selecting only part of the observed period. and Shagaev [1977). who find the same characteristic feature

The relationship between amplitude and periods as shown during gravity wave propagation conditions. Periods in theis Figure 4 als, strongly points to gravity waves as the source, range 3 hours to approximately 20 min are clearly recog-It shows the same linear relationship as found by Krassovsky sizable in the power spectra of both intensity and temperatureand Shagroet [1977] during gravity wave propagation. It dif- shown in Figures 5 and 1. They are all within the domain offes. however, in the striking manner, that for the same period internal gravity waves [Hines, 1960].of oscillation our results show about twice as high an ampli-tude. This may be interpreted to mean that the OH-emttinglayer over Longyearbyen. i.e, at extreme high latitudes, is situ-ated at a slightly highet altitude than at middle to moderately 50high latitudes [Frederick, 1979]. From simultaneous measure-

40

vi

U 30

0'00 0rs 20

so

-8-6-4-2 0 2 4 6 8

R oo o . 'Z . 77-VALUE

Fig, 4. Ampltude of the rotational temperature. AT in degrees Fig. 6. Distribution of (he t value (Al I AT T) for the measure-Kelvin plotted against penod r in minutes for the most well-defined merits on January 6-1 Cases with I > 8 not oduded here were inoscillationsis Figure I the range 0-2 for a tingle interval

27

42 _8 MY0ao FT Ai NIGHTGLOW OH 8-3i 8-) Ar '8 N

All of these indicanons seem to support the hypothesis that operative, since they find q values confined to the range -0 5a large part of the variations in OH temperature and intensity to - 1.5 iThe theoretical vaJue for the process (I is P = 1.reported here have been due to the passing of internal gravity whtle for the perhydroxyl mechanism and process il. " is 3 ' )waves. Weinstock [1978] has reviewed KIrassovsky's q value, intro-

A closer inspection of Figure 2 reveals a tendency of a line ducing a more complex relation between tempermure anj .0

structure of the order of 10-15 min superimposed on the tensity which is dependent upon the scale height of the minorlonger period waves. Fine structure might be seen both in the component fi.e., H or 0) with respect to the total scale height.intensity and in the temperature. Similar fine structure in the Thus phase shifts and amplitude ratio variations may be intro-5577-A and 6300-A night airglow intensity is known to exist duced. However. the main effects of phase shift and variation[Okuda, 1962; Silverman, 1962] and is in some cases believed in the scale height of the minor component during the passageto be asociated with local instability in gravity waves having is most likely to broaden the 17 distribution (i.e.. smear out theamplitudes greater than approximately 20% [Hodges. 19671. relation to .4.). Thus the P values derived here by using Krass-According to Tuan et al. [1979], the ripple frequency should ovsky's relation should be able to distinguish between pro-be close to the buoyancy period, i.e. the Brunt-Vaisil cesses with values as different as I and 3.7penodL, which in our case is of the order of 5-7 tut. Our Figure 6 presents a histogram showing the occurrence fre-integration lime and sampling rate mighL however, modulate quency distribution for the values 'f P caiculated from the 24and obscure the real frequency pattern. i.e, the real oscillations hours of observation -t'rrteu here. The mean I value is 1 08.might be considerably smaller than 10-I5 min. A further in- which imphes the ozone mechanism to be responsible for thevestigation and discussion of this is therefore advisable to OH emassion. Other emission mechanisms with 7 close to Ileave unul data with a much higher sampling rate i.c. the can. however. not be ruled out, but mechanisms with q > 1.5,order of I min) is obtained, such as. for example, the perhydroxyl mechanism, if operativeBabakov [1975], Katicn [1975], and Shaqaev [l97-j report at all seem not to contribute significantly to the OH emis-

the occurrence of gravity waves to be correlated to geomag- sions. An upper level of 5% may be estimated to be the largestnetic activity as measured by the ,-omagnetic K index. ShA- possible contribution.gar [1974] reports southward :.avelling waves, indicating anorigin in polar region, .\rassovsky and Shagaev 1194. 19') 4. SLSO4AitY

report the sourri-s to be connected with active meteorological From continuous 24-hour measurements of the OH (8-31phenomena in the troposphere in connection with jet streams, band nightglow emission at 78.4'N, a mean temperature ofwe-ner fronts, cyclones, anticyclones, etc. But they do not 237 K, and a mean band intensity of 596 R is found. Temper-exclude a dependence on geomagnetic activity. attire variations are very large, showing oscillations with ex-

Svalbard is indeed clos-ly surrounded by both geomagnetic tremes up to t70 K from the mean. The large variation be-and meteorological phenomena. Not etrsgh data have yet tween previous temperature measurements [Chamberlin andbeen accumulated to establish how common this behavior of Oliver. 1953; Noxon, 1964; Sivjee et al. 1972] at latitudesthe OH emission is, and how it connects to the overall mesc- 70'-85N cover the r;gc 160 K to approximately 300 K andspheric temperature. density, composition, wind velocities, etc. was previously thought to be too large when compared to

measurements at lower latitudes. The observation of this vari-3.2. The Observed q Value and the OH Emission .Wechamui anon in a single 24-hour period from a single station lends

The I value was first introduced by Krassovsky [1972] and credence to the validity of all the previous observations. Inis defined as follows: addition, there are indications that these large variations were

due to the passage of internal gravity waves through the= AIfl.ATT = 2.17- I) - A. msosphere. If so, the extreme amplitudes of the variations

where Al/[ and AT/T are relative increments in the intensity imply that the OH-emitting layer at high latitudes is slightly

and temperature 7 is the ratio of the specific heats, and A, is higher in the atmosphere than at middle and low latitudes inthe exponent in the temperature dependent rate coefficient January [Frederick 1979]. The deduced q values uAl IAT 7)(i.e., k - koT - A.4 k and ko being the respective reaction coef- favor the ozone mechanism to be responsible for the OH emis-ficients at temperatures T and To). The theoretical justification sion with the possibility of an additional mechanism contnbu-

of the P7 value and its relation to the emission process in the ting up to 5% of the emission.case of an adiabatic oscillation of the atmosphere is also dealtwith by Krssovsky [1972]. It is an important parameter to d4ckOwledgments. Finanoal support for this research was provid-

ed by National Science Foundation through grants ATM"-.'0837.evaluate and may be used as a means of identifying the emis- ATM77-24838. ATM80-12718. and ATM82-O0114 to the Geophysicalsion process. Kriassovsky [1971] introduced the perhydroxyl Institute of the Uversity of Alaska. One of us M. K. M) is support-process and suggested that other processes may be important ed by a fellowship grant from Royal Norwegian Council for Scientificin addition to the ozone mechanism suggested by Bates and and tndustial Research.

The Editor thanks T F Tuan and another referee for their aust-.'Niolet [1950] and Herrhe-g [1951]. ance in evaluating this paper

O 3 - H - OH9fr : 9) + O0(+ 3.34 eV) (1)

Takahashi and Baiitr, 1-981] find evidence in their data for Babskov. N. A., On the influence of the planetary state of mac-the mechanism netosphere on the parameters of the hydroxyl emission. I:r .4kad.

'taik Tubkm SSR. 5. 67. 1975.HO, + 0 - OHlv 6) i 401+2.4 eV) 12) ates, D R.. and \4 Nicole. The photohermstry of atmospheric

proposed by Virolet [1970] in addition to the ozone mecha. water vapor. J. Geophiys. Res., 35. 301. 1950.Chamberlain. I W. at.J N J Oliver. OH in the airglow at high

rmsm. Takeuchi er al. [1981], however, report strong evidence latitudes. Phys. Re. 90. it1. 1953for the ozone mechanism to be the only emission mechanism Deehr. C S,. G G Sivee, A. Egoland. K. Herinksen. P E Sadhoilt,

28

Mnas ET 4L Ncorrca~ow OH (8-3t BAND yr'1 8N 0

R. Smith. P Sweeney. C Duncan. and I Oilmes. Govia-hasd Reed. E 1. Polar rnancwmentn of nigoigiaw emissions near sf30.ohsersatcns of F .ep.n a-scitd wish the mnagistonpnenc cusp. ophcs Roe Ldwt. 3. 5. 191rJ Gophis Rca. 53. 2185. I9BO0 Rodgers. C D. Morphology of upper atmosphere temperatare. in

Dick. K A. 0 GG Stolen and H, M4 Cr-sawuie. Aircraft aigiow Dsneeal adu Chensicai Caspleso Berweren inr \arai ca Ionizedintensity measurements Variations in OH and O1 15571), Plans .Itesplsrr edited hr, B Grndal and I A Hoe.p 3. D Redel.Space Sc,. 1.8. 6dl. l97S. Hintghanm, Masis. 1977

Frederick. J. E.. Iluis of gravity waces activity on lower ihermiso- Shapaes. 14. Fast sariacons of hydronri right airglow emission. J1spheric phetochemnistry and compositien. Pi-nt Space o.2 4-~s Tree Ph rn.. 3d. 3h7. 39741469, 1979 Sdrverinaan. S& M.. Unusual fluctuations of 5577A 01 airgiow ets-

Hairion. A. W., W F I Erants. and E. I Lleweellyn. Study of the nion intensity on Octoher 28. 29, 1961. .Vacuoe. 195. a81. 1962id-Il1 and ,5-2) hydroxyl baisn in lie night anglers. Ca. J Phvyt. Sides., 0.0.. and C S. Deehe. Difference in polao atmospheric opti-a9. 2509, 1971 cal euninaons heiween mid-day cusps and night-time auroras. in

Hatlield. B. T F Tuan. and S& M. Stisian. On the eemus of Explorato i the Polar U pper Otmoirpkee. edited hy X, Deelir and.2icjucz .a; < -air ,x pio ix '. r14 ntH ON crtoorn. X "oliei p 190 El Resle- "tneham Ma.._ loon1

J Gophre Re,., 18. 2429. 1981 Sielee. G G.. K. 4. Dick, and P D Feldman. Temporal variations inHertzherg. 0. The atmosphere of the planems J. R Aston. Lao. Cant, the night-time, hydroxyl rotational temperature. Planner Space S-t

45 100. 1951 211. 2h61. 1972.H ines. C. 0. Intenal atmosphenec granity waves at ionospheric Takhlashi. H.. and P P Batisia Simultaneous measurements of OH

heights., Ca- J Ph.t.. 38. 1441. 960. 19. 4hL t8. SI, 17. 2L16. 21 and 0. 1i hands in the airglow. J tGoophcoHodges, R. R. Oenesralion of turhulence in the upper atmosphere hy Ret. 86. 5632. 1981

interntal grarty warms, J. Oephs. Res.. 72. 3455. 1967 Takahashi. H.. B. B. Clemesha and V. Sahat. Nightglow OH 8g-liKrassonshy. V L.. The hydronyl emirssion prohleme and pan of its hand intensities and rotational temperatures at 2SYS. Planet Space

solunes. ten. Onophres- 27.2N1,. 1971 Sc,.. 22. 1523, 1974Krossosnhy. V L.. Infrasonifc rarianons of OH esion in the upper Takahashi. H, Y Sahut. B R Clmesha. P P Batinia. and N R

atmosphere. Ane leopnyn.. 28. 739. 1971. Tetutra Diurnal and seasonal carialtios of the OH 18-3I ateglowKrasisnrshy. V I., and M. V Shapaer. Optical method of recording band and ito correlation with Of 5571A. Planet Space Sci. :5. W4.

acoustic or gravity waves in the upper atmosphere J -lisn Ter 1977Phys.. 36, 573, 1974. Takeauchi. 1. K. Misawa. Y Kato. and 1, Aoyama. Rotational temper-

Krasaorshky. V L. and 14. V Shapesr. On the nature of internal alumes and intensities of OH 16-2) and OH (8-3I hands in thegravitational wanes obiserved from hydronyl emissions. Plantr nightglow. S Aiscs. Te, Physn..l 387. 1979Space S., 275. 200. 1977 Taheuchi, IL. K. Mrsawa. V. Kato. and I Aosanna. Seasonat variations

Kiaisin. K. 1L. Intensity oscillauina of the 5577A and 5893A esnti- of the corrlation among nightglow radiations and emission mecha-stns and geoesupiw activity. Aur-a Auqrlow USSR dald. Sot.. n.- of OH nightgllowi -m . J 4-no Tore Ph.o.. 43. 15",2.1328. 1975 1981.

Keite. G, Nightglow obiseraions us Ats during the [CV.. (leophyn. Taan. T F.. B. Hereger. S. M Silserssan. and M4 Ohuda. Os gravityP.Mi., 210, 1. 1959, wave induced Brnint-Vaisalli oscililations J (loophvn Ben.. 814. 593.

fMeo. A. B.. OH minion hands is the specsirnam of the night sky. J979.doieopbys. J.. 1l1. 555. 1950. Vallance Jones. A.. Auror. D Reril Hingham. Mass-. 1974

Meriwether. J. W. High latitude airglow, ohservations of correlated Walker. J. D_. and E. 1. Reed. Behavior of the sodium and hvdrovyl,horitnoin fluctuations in the hydronyl Meanae 8-S heed intensity nighttime eisoons daring a stratosphersc warmhing. J 4tnsos. So..and rotational temperature. Planer. Space Sot. 23. 1211. 1975. .13. 118. 197h.

Myraho. H. K.. Temperaiture variatin at mesonpause, levels at 78-N Wetessoek. J_. Theosry of interaction of prarity wares with 0, 01:)luring wintser wslsic. Planer. Spare Sot.. in press. 1983, airglow.SJ. Genphct. Res- 53. 5175. 1978.

Nicoles .Ozone and lireegn emotions lee dphryn.. 26, 5Sf.1970. C. S. Deehr and H K. Mrabwt. Geophynical Institute Univerity of

Nonon. J, F. The latitude dependcei of OH rotational temperature Alaska. Fairhanks. AK 99'01in the night ateghow. J. Geephyn. Res., 69. 4087, 1964. 0. 0 Siclee National Science Foundation. I80M 0 Si. N w,

Nonon. J F. Effmts of enternal prarity waves apes night airglow Washington. D. C 20550temiperatures (leophyn. Rex Sail. 5,25.1979. Rleersed Felsruarv IS. 1985.

Ohuda. M_ A study of excitation processes insight airglow. Sci. Rep rerised April Ill. 1983.Toh.ha Unte. 15. 9. 1962. accepted May 2. 19831

If

30

A'-~o Spo S,,. Voi 32. No _'.in 249--255. 1984 (X)3263 K 3 OU-UsiPrned u G e:i BOn Pepeov P- L:4

TEMPERATURE VARIATION AT MESOPAUSE LEVELSDURING WINTER SOLSTICE AT 78°N

H. L %IhASO'Geiphysca !nSIEute. Jniverity -f tAl ska.F irbanks. AK 'O 1. 1T

IReceived in fina form 27 June 19831

Asitract-Atmospheric temperatures from the polar mesopause are deduced from spectrophotometricmeasurements of hydroxyl bands and lines in the night aarglow made at 78WN dunng Decetber and January198081 and 1982,83. An overall mean temperature of 220 K is found with a range from 172 to 257 Kin the dailymean values. Several warm periods lasting 3-6 days may be due to heat dissipated by gravity waves. One weekof consistently low temneratures was apparently connected to a stratospheric warming. Both datasets show awarmer mean temperature later in January than for early and mid-December. The polar OH airglow seems topeak at or just above the mesopause. The data also indicate that the mesopause is situated at approx. 90 kmwith an upper temperature gradient of I K km - . indicaung a very shallow mesopause. A superposed epochanalysis of 19 consecutive 24-h periods reveals a semidiurnal variation in the temperature around wintersolstio with an ampttudeof 5 K. No diurnal variation of anmpitude greater than I K is apparent. Average windvelocity deduced from the amplitude of the semidiurnal temperature variation is 9 m s -

I. INTROOUCTION resolution is in the range 15-30 km iRodgers. 1977;Barnett, 1980) leaving 60 km as the approximate upper

Temperature is a basic physical parameter of the height for usable temperature determination.atmosphere. It is involved in most of our understanding While rocket and gun-probe measurements give aof atmospheric processes and behavior, single profile in time and space, satellites may cover the

Common temperature-measuring techniques above entire globe within a 24-h period ISissala. 1975).the troposphere involve radiosonde balloons, grenade Satellites are therefore superb for measuring the largeprobes, rockets, satellites and different methods of scale spatial and temporal variations of temperatureoptical, i.r. and radar remote sensing. caused by the main global transport and heating

Temperatures in the polar mesosphere and lower processes.

thermosphere, i.e. 60-110 ki, are only crudely known Nevertheless, for continuous measurements of thefrom sporadic measurements and only a single, all- dynamical behavior of temperature in local timeseasons 80'N model for the altitude range 25-80 km is (caused by gravity waves, tides, winds, etc.l ground-given in CIRA 1972. Varations assumed to be caused based remote-sensing techniques are the only tool. Inby phenomena such as tides, winds, gravity waves, etc. the mesopatise region (i.e. 80--95 km altitude) oneare not included, ground-based technique is to extract temperatures

,ecently, satellite probes such as those on board the from the measured intensity distribution among theNimbus and TIROS-N NOAA series (Sissala. 1975: rotational lines of particular vibration band of theRodgers. 1977; Schwalb, 1978: Drummond et al.. 1980) hydroxyl nghtglow emission IKvifte, 19591.have provided temperature measurements of the Mesopause region temperatures have been obtainedstratosphere and mesosphere up to the 60-70 km this way since Meinel (1950) identified the OH bandregion. Emissions from the IS pan bands of atmospheric system in the night airglow. Numerous investigators

CO 2 are used to deduce these temperature profiles have reported temperatures, their different variations(Rodgers, 1976:GilleetaL. 1980). Above 80 km CO 2 is and connection to other atmospheric parameters andnot in local thermodynamic equilibrium and thus processes (Kvifte, 1960: Wallace, 1961: Noxon. 1964:

temperatures obtained by radiance inversion from Shefov, 1969: Visconti er al., 1971 : Sivjee et al., 1972:channels including this height could be largely in error Wiens and Weill, 1973 : Takahashi et al-, 1974. 1977.(Drummond et al., 1980). Above 60 kin, vertical 1981; Takeuchi et al., 1979).

Due both to environmental problems and problems

with auroral contamination iMeriwether, 1975) a

On leave from Norwegian Defense Research sparse amount of.OH-derived temperatures exist forEstablishment. N-2(0)7. Kjeller, Norway. high latitudes and only a few occasional measurements

249

31

250 H K MYRA80

are reported for latitudes greater than 10- 1 Myrabo et

al.. 1983). This is also true of other ground basedtechniques. Asa result, the effects of gravity waves. tidesand winds in the polar mesosphere and mesopause _region are not et experimentally investigated. ,

The purpose of this paper is to report and discussresults from recent measure ,' Cii rotational tempera- - -

lures at 78 N

L OBSERVATIONS AND DATA REDUCTION ~/WThe OH emission data employed in this work were

part of measurements undertaken during the 1980 81and 1982,83 campaigns of the Multi-National Fi. 1. AN EXAMPLE OiF THE SPECTRA USED FOR DOrFLCtSG

Svalbard Auroral Expedition fDeehr et al.. 1980) close rTEMPEA rR-

to Longyearbyen on West Spitsbergen 178.4 N, Lat.. Eachspectrumisacquiredduringas0minior s0minipenod

geographic. summing 450 13501 scans. In addition to the OH i8-3 P15E Long. gebranch lines, the forbidden oxygen doublet at -320 30 A

OH emissions are normally predominant in the near- originating in the aurora is indicated.i.r. part of the night sky spectra. Contamination byauroral molecular emission in the auroral zone(Valiance-Jones, 1974: Meriwether, 1975) is normally The quality of the spectra used for temperatureinsignificant at Spitsbergen because auroras are deduction may be seen in Fig. 1. Probable error itn agenerallyatahighaltitudeandemttmainlyatomiclines single calculated temperature is estimated to be -3 K(Sivjee and Deehr. 1980). Thus, a normal. high- caused by the uncertainty of defining the backgroundresponsivity aurora spectrophotometer operated at a levels.resolution around 4 A or less may be used to measure Daily mean temperatures are based on 24-h averagesmost of the OH bands and lines clearly resolved from of the I- and 3-h temperatures. For diumal variationauroral emissions, only I-h integration was used.

At the Longyearbyen Observatory, a I m and a 1 2 m The auroral activity was monitored by the intensityhigh-throughput Eberi-Fastie spectrophotometer are of the 7320,30 A Oil lines. Additional care was taken tocoupled to a trm-computer recording in the photon- assure that there was no contamination from N, I P andcounting mode. The I m instrument, used for the other auroral molecular emissions.measurements reported here, is further described byDick e al. 11970) and Sivjee et al. (1972.

The spectrophotometer was normally operated for24 h a day from December 1982 to January 1983. 3.1. Day-to-dav tariabili v and December-Januarv(During 1980/81 other spectral regions were scanned patternand 24-h operation was not routinely performed.) Temperatures for each day from 5 December to 30Spectra were rejected when Fraunhofer absorption January 1983 as derived from the hourly 13-hi meanslines appeared during the twilight period ti.e. 3-6 h are plotted in Fig. 2. The daily means are indicated byduring mid-day in the last half of January 19831 and filled circles and straight lines are drawn between eachduring full moon with overcast weather. Operation mean. Dashed lines indicate that the temperature forceased only during snow or storm. The sensitivity of the one or more days is missing. Heavier lines indicate theinstrument allowed temperatures on an hourly basis to daily mean temperatures obtained from OH emissionsbe extracted even during overcast sky conditions, during the 1980 81 campaign.

The spectrophotometer was pointed towards the Some of the 1980,81 measurements did not cover azenith and the 7280-7410 A region was scanned in 8 or whole 24-h period. Missing temperatures were added32 s using the spectrophotometer in the second order where possible by interpolation before daily meanswith a 1 mm slit corresponding to a bandwidth of 1.5 A. were calculated. Due to a different observing programEach scan was recorded on magnetic tape. Individual during 1980,81. a number of OH bands other than therotat'onal temperatures were calculated from I- and 8-3 band were utilized to obtain the temperatures3-h integrated scans, by employing Kvifte's method shown.using the intensity ratio of the P1 2), P113). P14) and The most obvious feature in Fig. 2 is the large, week-P,(5) lines )Kvifte, 1959). long cold period arou d January 1. followed by a

32

Temperature vanation at mesopause Ivets

260

-250-

240 /'--.) . -

,230

220 - ' "' '" .

X 7 + . . . . . . .. . . + . . . .

1ODEC 20 DEC 31 DEC -C CAN 2) .AN K ANDATE (UT!

FIG. 2. MEA, OaILY TEMPERArURES FOR DEcEBER 411) JAxt.,RY

warming to a peak in the daily mean temperature of IShashun 'Kina and Yudovtch. 19801. According to257 K in mid-January. The similarity between the Brekke(1977)even ifasignificant part(i.e. 20 %) ofthis1980,81 and 1982,83 temperatures is also striking, .e. a energy is deposited at 80-90 km level, it could nottendency to a higher mean temperature in mid-January increase the temperature by more than some tenths of acompared to early and mid-December. degree. This does not rule out the possibility that

The 2-6 day warming periods previous to the cold gravity waves onginating in the ionosphere inperiod may be explained by heat deposition from connection with geomagnetic storms could causegravity waves dissipating in the upper mesosphere. significant oscillations of the mean temperature in themesopause region Hines, 1965). The energy available mesopause region. We find, however, no clearby dissipating gravity waves seems to be able to provide connection between these 3-7 day warming periodsa heating rate of at least 10 K day - I at 90 km tHines. and geomagnetic activity or substorms.1965: Clark and Morone, 1981) which is sufficient to The I week cold period (<200 K) observed atexplain the warming periods in Fig. 2. Since gravity year's end and the continuous warming from 172 towave activity is believed to be associated with 257 K ti.e. 85 K) in the daily mean temperature seemedtropospheric weather systems such as fronts, cyclones, too large to be connected with gravity waves. Its longjet streams. etc. (Clark and Morone. 1981 Krassovsky duration could suggest that it was connected to largeand Shagaev. 1977) the duration ofthe heating is likely scale transport of cold air from low or mid-latitudes.to beoftheorder ofdays.This fits theobserved data. It Another means of producing such a large change inmay also be noted that from the spectrum of temperature is in connection with a "stratosphericatmospheric winds velocity fluctuations at 86 kn warming- or -stratwarm" Labitzke. 19801. Initialaltitude, as measured from Poker Flat. Alaska 165'N ggLat., 147W gg Long.. Balsley and Carter, 1982), asignificant peak in the power density spectrum appears , _at a period around 3-4 days. Time-averaged OHrotational temperature and mesospheric temperatures 2generally, are closely associated with wind fluctuations t /(Krassovsky and Shefov. 19801. . -

It seems less likely that gravity waves originating in . -

connection with heating caused by ion drag or Joule : - - .

heating during magnetospheric substorms containenough energy to heat the mesopause region tens of - -

degrees for a period of several days. For example, .... . .during the substorm of 15 February 1978 approx. FIG. 3 HOURLY MEAN TEMPRATURES FOR 24-h PERIOD AT1022 erg was released into the ionosphere during 6-7 h Lt,NGYEARRYEN. 78 N.

33

252 H K. MYRABio

warming of the polar stratosphere is expected to temperature between the bands were, in most cases,produce a cooling effect at the mesopause ;Labitzke, within the measurng uncertainty i.e. 4 Ki,1977;Schoeberl, 1978l A stratospheric warmirg is seen Takingall the 1980,81 measuremenisfor Decemberon the 10 mbar level charts around 31 December together, there is a small tendency for the temperatures(Najoukat et al., 19831. Thus, the direct connection obtained from OH 19-41 and OH (9-31 bands to bebetween these two phenomena seen here (i.e. a drop in slightly higher than those from the OH (5-1 and OHtheobservedtemperatureofthemesopausearoundthe (6-2) bands (see Fig. 4). The mean temperaturestratwarm confirms the theory tLabitzke, 1977). difference between the t9-4) and the i5-li band

As one would expect, the large-scale temperature temperatures is found to be 3 K which is statisticall,trend is clearly visible in the data from individual 24-h significant.periods. i.e.thereisnormallyalonger-termtrendtaking Rodgers et al. (1973) reported from their rocketplacewithshorterfluctuationssupermposed lessthan observation at Poker Flat 165N gg Lat.) that theI day). Figure 3 illustrates this by showing a slice of emission height profileofthe lower vibrational bands of"thedownwardtrendinthecoldperiodlateinDecember OH is displaced to lower altitudes. They suggested1982. Figure 3 contains a typical, moderate fluctuation chemical quenching by atomic oxygen to be responsiblepattern, but there are also observations of very large for the observed effect together with an additionaland violent fluctuations in temperature with ampli- excitation process. Simultaneous observations of OHtudes in excess of t-70 K )Myrabo et al., 1983). (9.4), 18,3), (7.2), 16.2) and 5, 1) band intensities bs

Takahashi and Batista 1198 1) may also be interpreted3.2. The height ol'the OH emissions layer, the mesopause as support for the hypothesis that quenching islevel and the temperature gradient responsible for the observed altitude distribution.

Laboratory measurements by Charters et al. (1971) Rodgers' measurements were performed at highshowed that at pressures comparable to those at 80- latitude(65'N gg Lat.)during March while Frederick et100 km height, an increase in the pressure leads to an al. (1978) used data from satellite measurements inincreasein theOH excitation ratem inverse proportion May-June and for low latitudes. Use of these twoto the number of the vibratiunal state. A result of this independent measurements covering different latitudeswouldbethatemissionsfromlowervibrattonallevelsof and seasons to deduce height difference between thethe OH molecule would correspond to a slightly lower OH band emissions as done by Takahashi and Batistaaltitude range in the atmosphere. (19811 is open to question.

This interpretation was suggested by Gattinger Taking the height of the peak v = 9 level emission(1971) and employed by Wiens (1974) to interpret from Frederick er al. (1978) and the v = 5 level fromregular changes of the OH 18-3)/(5-1) band intensity Rodgers et al. (1973) results in a height difference ofratio during the night at Adi Ugn. approx. 1 km. This implies a temperature gradient of

Figure 4 shows the temperatures from the different 3 K km -bands before averaging and interpolation. Some of the From winter observations at Zvenigorod (55-N)scatter may be explained by the different bands not during gravity wave propagation in the mesosphere.covering the same time interval. This was the case for Krassovsky and Shagaev (1977) report a significantlymost of the January measurements. When different larger temperature disturbance in temperaturesbands were acquired simultaneously, the differences in deduced from the v = 9 level then from the v = 4 level.

Average temperature differences between the two levelswere found to be 12 K ranging from 3 to 26 K. Onanother occasion. Krassovsky et al. (1977 report

- , + temperature differences between the v = 9 and v = 5level temperatures of 3-14 K with a mean close to 6 K.

If this is to be interpreted as corresponding totemperatures at the respective emission heights of the

, " different bands, the peak altitude difference as deduced.,,, from Rodgers etal.)(1973) and Frederick et al. 19781(i.e.

1 km) seems too small., .. ,, s+ .+ ~ Using the temperature slope deduced from U.S.

Standard Atmosphere, Supplement 1966 for 60N.FIG. 4. TEMPEpAruis, AS DERIVED FROM 1THE DtFFMENT OH winter, for the 89-98 km level just above the mesopause

SANDSFOR r tE 1980-1981 MASUusttNrs. (i.e. 1.8 K kmi-) and the temperatures derived bySymbols indicate the different bands. Krassovsky et al.. a mean height difference of at least

34

Temperature %ariation at mesopause levels

3 km would be expected between the 9-31 and i5-I) examined for regular vartations with periods less thanbands. Observations during gravity wave propagation or equal to 24 h. onginating from the solar tide ix.e.conditions reported by Krassovsky and Shagaev (197) penodsofthe type 24 m, m being an integer 1.2 3. 1and Shagaev t 19hU) give an average height difference ol Hourly mean temperatures for each of the days *ere4 km. Applied to our results this establishes a positive superimposed resulting in hourly means for the period.temperature slope in the range 0.8-I K km - '. The hourly means are plotted in Fig. 5 and a

From the results by Krassovsky and Shagaev (1977), semidiurnal trend is clearly visible.Krassovsky ei al.. 11977) and our result presented here. A semidiurnal tide curve has been fitted to the datait is seen that the OH emission is situated in an altitude following the method of Petitdidier and Teitelbaumrange with a positive temperature slope. This may be 11977):interpreted to mean that the mesopause is at approx. i"T T) = . cos [I2i A)IZ -Z,) 2 Pit -tr))].80 km with the OH emissions onginating in the 80-85 km region. This interpretation is in sgreement with where A is the vertical wavelength. P is the period. Z,) isthe rocket observation of Rodgers et a .1973) at 65'N the altitude ofOH 18-31 nightglow intensity maximum.in March. is the time of maximum temperature Tat Z0 altitude

However, the very large amplitude temperature and and .4 is the amplitude of the relative variation in theintensity disturbances at 78°N (probably due to the temperature. The best fit gives an amplitude of 5 K andpassage of gravity wavesIreported here and by Myrabo the time of maximum temperature at Of UT. or localet al. (1983) indicate a higher altitude layer. Results Longyearbyen time 02: 00.reported by Shagaev 11980) from Zvenigorood 155-N) Semidiurnal variations in the OH intensity andalso indicate a mesopause above 85 km in winter. A deduced temperature are reported by a number oftemperature slope of 1.8 K km - was reported. authors (Takeuchi er al., 1979: Takahashi et al.- 1974.

The polar winter vortex, which has a peak westerly 1977: Petitdidier and Teitelbaum, 1977) but there seemflow at about 50-60 km altitude (Rodgers, 1977). forces to be no reports based on OH measurements indicatingthe stratopause and mesopause at extreme high a diurnal (i.e. 24-h) temperature trend.latitudes in winter to besituated at higher altitudes than Spizzchino 11969l and Teitelbaum and Blamontelsewhere. A mesopause at about 80 km in early winter 11975) argue that non-linear interaction with gravityat 78°N is therefore unlikely, and a more reasonable waves is more important for the first diurnal mode thaninterpretation is that the mesopause is situated approx. for the semidiurnal. Thus, the effect of averaging overat 90 km and the OH emissions-at least for high and several days ito obtain an adequate signal-to-noiseextremely high latitudes in winter-peak at or just ratiois tocancelthedmrnalmodesinceitwouldbe farabove 90 km. less stable than the semidiurnal mode.

From the temperature slope deduced here (i.e. 0.8- Continuous observations at lower latitudes norm-I K km-') it seems reasonable to point out that the allylastless than 10h, whichemphasizesasemidiurnalextreme high latitude winter atmospheredoes not seem tide over a diurnal one. This selective effec, mayto have a very pronounced mesopause therefore be present in previous data. The data here are

not so limited. A Fourier analysis of penods failed to3.3. Diurnal, semidurnal and short time v'ariations retrieve any diurnal tide component. If it does occur. the

Data from 19 days around winter solstice from 9 amplitude is less than I K.December to 27 December 1982 were selected to be According to Forbes 119821. model calculations

show the diurnal tide to be dominant over thesemidiumal at high latitudes, while according to Beer(1975), the semidiurnal tide dominates over the diurnalat high latitudes. Zonal wind speed data reported bySpizzichino (19701 favor Beer 119751 and the resultsreported here. Possible dominance ofthe diurnal tide at

- . high latitudes on a shorter time scale 1i.e. a few days is,2 however, not ruled out by this result.

Krassovsky and Shefov 11980) have showed that alinear relationship between measured OH tempera-

FIG. 5. HouRLYMEAN TEMPERATURESAVERAGED OVERA 19-day tures and wind data in the radio meteor region exists.'toG 5.ORLYovi wiNTEMR S osTCE FRom 9 DEc M Titro 27 To a first approximation the relation

DECEMs-R 1982.Semidiurnal trend is indicated. AT T i(- W ' v c

35

254 II K. MIRABO

is found to satisfy the temperature-wind velocity is apparent. Wind speeds deduced from the obsersedrelationship, where T is the temperature. AT the OH tide temperature component show a meantemperature increment. i the imaginary unit l ie., - 1 semidiurnal wind of approx. 9 m srepresenting the phase shift, - the ratio of specific heats,v the wind speed and c is the speed of sound. This Acknowledqemenrs -Financial support for this re-relation is also theoretically justified for long period search was provided by National Socience Foundationwaves Hines, 1965). Applying this relation to our through grants ATM80-12719. ATNIS2-00114 andsemidiurnaltideresult. andtakingthespeedofsound, c. ATM82-14642 to the Geophysical Institute of theto be 258 m s - ' i USSA, Supplement 19661. :*. = 1.4, University of Alaska and from the Royal Nor'egianT= 218 K and AT = 5 K. results in a mean Council for S.:ientific and Industral Research throughsemidiurnal wind speed of 9 m s- which seems to be of a fellowship grant.the right order of magnitude (Groves. 19801.

REFERENCES

Balsley, B. B. and Carter D A. 1l9821 The spectrum of4. SUMMARY atmospheric velocity fluctuations at 8 km and 86 km

Geophys. Res. Left. 9. 465Ground-based observations of atmospheric OH Barnett. J J. 119801 Satellite measurements of middle

emission temperature and intensity have been carned atmosphere temperature structure. Phil. Trans R Socoutat78 'Nduring December and January 1980/81 and Land. . 296, 41.

1982.83 (Fig. 2). 3-6 day warm periods were observed Beer, T (1975).Atmospheric Waves A Hilger. LondonBrekke, A. 11977) Auroral effects on neutral dynamics. inboth years in December. In 1982.83 this warm period Dynamical and Chemical Couplinq Between tir Veuiral and

was followed by I week of consistently lower lonized Atmosphere Edited by Grandal. B. and Holtet. Jtemperatures (down to 172 K). Common to both A.). p. 313. D Reidel, Dordrechtseasons is a higher mean temperature level reached Charters, P. E.. Macdonald. R. G. and Polanyt. J C 119711later in January after the cold periods. This is consistent Appl. Opt. 10. 1747

CIRA 11982) COSP.4R Workinq Group 4. COSP4Rwith Quirozs 119691 idea of a higher mesospheric International Reference 4tmosphere Pergamon Press.temperature in January than in early winter. i.e. Oxford.November and December. Clarki. H. and Morone. L T 119811 Mesosphenc heating due

We find evide:';o of a stratospheric warming to convectively excited gravity waves-a case study W.onWeath. Rev. 109.990.

(Najoukat et al.. 1983) near the cold period which Deehr. C S.. Sivjee. G. G.. Egeland. A.. Henrinksen. K..explains the cold mesosphere (Labitzke. 1977). Sandholt. P. E., Smith. R.. Sweeney. P. Duncan. C and

Wesuggest that the warm periods ofthe order of days Gilmer. F 11980) Ground-based observations of F-regionare due to heat dissipated from gravity waves associated wsth the magnetosphenc cusp J Ieophvs Resoriginating in the troposphere. 5. 2185

Dick, K. A..Sivjee,G. G and Crosswhlite. H. M 19'0iAircraftThe overall average temperature for the data from airglow inensity measurements. varations i OH and Of

the 1980,81 and 1982.83 seasons is found to be 220 K. (5577. Planet. Space So. I. 887Assuming that the OH emission at high latitudes in Drummond. F R.. Houghton F T.. Peskett. G D, Rodgers.winter peaks at or just above the 90 km level, this C. D.. Wale. M. F.Whtney. F and Williamson, E.J i i 90i.

The stratospheric and mesosphernc sounder on Nimbus 7temperature is slightly higher than the 80N model Phil. Trans. R. Soc Lond. A 296. 219from CIRA 1972 at 80 km. Forbes, F M 119821 Atmospheric tides: the solar and lunar

The difference in the mean temperature obtained semidiurnal components. J. qeophys. Res. g, 524a1.from OH 19-31 and (5-1) bands, assuming a height Fredenck.F E..Rusch.D W andLiuS.C..ll978lNightglodifferenceof3-4kmiShagaev, 1980)isusedtodeducea emission of OH iX~nl: comparison of theory and

measurements in the 19-31 band. J. qeophvs Res. 83. 2441mean temperatureslpe°f 1K kn ' Thedata further Gattinger. R. L. (1971) Interpretation of airglow in terms ofindicate that OH emissions originate at or just above excitation mechanism in The Radiatinq Atmospherethe polar mesopause (i.e. 88-94 kin). The size of the (Edited by McCormac. b. M.L p 51 . D Reidel. Dordrechtgradient (i.e. I K km - (compared to similar gradients Gille. F. C.. Bailv P L. and Russell. F. M.. 111 0980obtained at lower latitudes (i.e. - 1.8 K km - at 55'N. Temperature and composition measurements from the

I.ri.r. and Itims. experiments on Nimbus 6 and 7. Phd.Shefov. 1980 1.84 K km -

' at 60'N (winter). U.S. Trans. R. Soc. Lond. A296. 205Standard Atmosphere. Supplement 1966) implies that Groves.G.V.I980)Seasoralanddisurnalvariationsofmtddlethe polar mesopause in winter is less pronounced, atmospheric winds. Phil. Trans. R. Soc Lond. .4296. 19

Superimposed hourly means for a 19-day period Hines. C. 0. (19651 Dynamical heating of the upperaround winter solstice reveal a clear semidiurnal trend atmosphere. J. geophvs. Res 70. 1".Krassovsky. V. I. and Shaaev. M V 119777) On the nature ofwith an amplitude of 5 K (Fig. 5). This is connected to hydroxyl airglow Planet Space Sci. 25. 509the solar semidiurnal tide. Noeffect ofsolar diurnal tide Krassovsky. V 1. and Shefov. N N i 19801 Relation between

36

Temperature 3natjon at mesopause resets s5

temperature and circulatiovnnthe mesopause region ueom Sheto, N N f969i flsdr,I enissio ,I Ire .ppe,4 err,n 20. i I1 atmosphere I The hehasior duting soiat scesasr.

Krassossky. V I . Potapsis. B. P. Semenos. N. I .Soholes. V and gleomagnetic disturbances Planet Spa~i- SI 17. -')U , Shagaes. Ms V and Shefoy. N, N. 11971) On the StssalaJ F 19-5rThse'\itl, e.cms uie Goddard Sp-c.equi-tnm nature of hydroxyl airglow. Planet. Space Sct. Flight Center. Greenbelt MD125. 596 Ssjee G G. and De :ir. C S i19S0, Differences in polar

Ksifte.C, G1959) Nightglow observation at Ass dunng the atmospheric optical emissions heiueen miclda, cusp andiLGA . Geophsca Norsegica. Geophtrs. Pubis 20. 1 night-time auroras, in Explforation it (he Puiar E riper

Kvifte. G. (19601 Temperatur-emeasurements from 0H bands 4tinospheie IEdited by Deehr. C, S And Holtet. J Ar1.Planet Spiace. St 5. 153 p. 199. D. Reidel. Dordrecht

Laitzke. K. i 1977) Stratosplienc-mesosphenc mtdwinter Sivjec.G.G.. Dick. K. As and Feldman. P D 1 19721 Tempcrairurmnines. in Dcnamss-aand Chemical Couplinq between the variatton in the night time hsdro'r i rotational temperature.Vetuland 1oni:ed Aimospheet (Edited by Orandal. B, and Planet. Spaics Sii. 20 261Holtet J, A. ,p I' D Reidel. Dordrecht. Spizzchino. A (19691 Etude des interactions entre Ces

Labttzke. K. 119801 Climatology of the stratosphere and differenies composanies du sent dans la haut atmospheremesosphere. Phil Trans. R. Soc Land. 4296.7 4 nn. Gs'oph ys Z5. 93.

isenielA B.ri 1950)lOHemission bands in the spectrum ofthe Spizzichino. A. 0 970) Etude des interactions enire Ie,night sky. .4p J, 111, 555. diffierentes composantes du sent .lans [a haut atmosphere

Menwether. J W 119751 High latitude airglow observations of 4 nn. Geophy s, 1.9

correlated short term fluctuations in the hydrov Memnel Takahashi, H and Batista, P P 1 19811 Simuitaneous8-3 hand intensity and rotational temperature. Planet measurements ofOH 19,4), 18. I i. t- i'Xr.h21and i. hI andsSpace Sci. 23. 1211. in the airglow. J qertphis Res. 86. 5631.

Myraha. H. K.. Deebe. C S. and Sistee. 1G G 119931 Large Takahashi.H..Clemesha.B R.andShai.Y r19

"',otrhtelo.amplitude nightglow OH (8-3) hand intensity and OH 18-31 hand intensities and rotational temperatures atrotational temperaturesvanationdunngA 24-hoar periodat 23 S. Planet. Space Sci 223 23781'4.JI stecph is. Res. tin press). Takahash, H.. Sishat. Y Clemesha. 8 R.. Batista, P B and

Noxon. 1. IF (19641 The latitude dependence of OH rotational Teixeira. N. R.i 119771 Diurnal and seasonal s anations at thetemperature in the night ategiow. J. qeuphvs. Res. 69.4087. OH (8-31 airglow hand and its correlation with 01S A

Petitcdidier. M4. and Teitelbaum. H. (197)1 Lower thermo- Planet. Space Sct. 25.3541sphere emissions and tides. Planet. Space Sci.25. 711, Takeuchi. 1-. Misawa. K.. Kato, Y and koyamna. 1. 19)79

Quiroz. R. S, 11969)The warming of theaupper stratospheretin Rotational temperatures and intensities Jf OH 1,-'I andFebruary 1966 and tie associated structure of the OH (8-31 hands in the nightglow. J asmtov. ten- Phis 41.Mesopasise Moun. Weath. Rev. 9 7,541 387.

Rodgers. C. D. i19761 Rev Genphvs. Space Phvsics 14. 609. Teitelbaum. H. and Blamoni. J E. 119751 Some conseq~uenscesRodgers. C. D 114771 Morphology of upper atmosphere of non-linear effects on tides and gravity waves. J atinos.

temperatures, in Dvis~asical and Chemical Coupliny between terr Phis. 37. 697the Nettral and Ion:ed.4trmosphe (eEdited by Granda. B. Vallance-Jones, A. 11974) 4urnu, D Reidel. Dordrechtand Holtet. J. A-). p,

3. D. Reid. Dordrecht. ViscontiG.,Congeduti. F and Ftocco. G 11971 Flctuation

Rodgers, J. W. Murphy. R. E.. Stair. Jr., A. T.. Ulwtck, J. C.. in the intensty and excitation temperature in the OHBaker. K. D. and Jensen. L. L. 119731 Rocket-borne airglow 18-31iband. in The Radiaiinq.Arrssspherei Edited bvradiometric measurements of OH in the auroral zone. McCormac. B Mri. p 92. D. Reidel. DordrechtJ. yeophvs. Res. 78. 7023. Wallace, L 119611i Seaso)nat sariation and interpretation of the

Schomeherl. 54 R.r t978lSiratosphencwarntngs observations OH rotational temperature in the airglow J, atmos retnAnd theory. Rev. Genphirs. Space Phis. 16. 521. Ph v5. 20.10

Schwalh.A.i 1978lTheTIROS-N NOAAA-GiatellteSertes Wiens, R. H. 19741 Diurnal vsanation of rhe 18.-3 i115-t)i

NO.4. Tech. Me-, NESS 95. intensity ratio of nightglow OH at Adi l..g Plastei SpacesShagues. M. V (19801 Vertical temperature gradients and S-t 22. 10359

dissipation of inel gravity waves near the mesopause Wiens. R. H. and Weill, G~ 119731 Diurnal, annual and solare-, .4e,- 20. 529 cycle varitaions of hydrosyl and sodium nightglow

Shashan *Kina. V M. and hadlosich L. A. t1980) Effects of intensities inihe Europe-Afrtca sector Planem.SipateScr 21.internal gravity waves during the magnetosphenc susorm 1111of Februarv I5. 1 9't9 Geom. Ace,,. 20. 516.

..........

............. ......

38

JO NAL ,OF -EOPHYSICAL RESEARCH, V0L. 19. NO. A10, PAGES 4153-156, 3CTOBR :, 198-

NIGHT AIRGLOW OH (H-3) BAND ROTATIONILL TRPERArJRES AT POKER FL\T, AlASKA

H. K. yrab61

, G. J. Romick, G. G. Sivjee2 and C. S. Deehr

Geophysical institute, University of Alaska, Fairbanks

Abstract. temperatures of the mesopause observations from Heiss Island as part of the

region (85-90 am) have been deduced from OH (8-3) Middle Atmosphere Program (MAP) [Naujoost, et al.,molecular band night airglow emission measure- 1983]. The observations reported in this papermerts made at Poker Flat (659), Alaska. The were obtained from Poker Flat, Alasa (65.11*°,data cover the first 4 months of each of the 147.46*W, geographic coordlnates) ouring the f'rst

years (976, 1977, and L978. Hean monthly tem- 4 months in the years 1976, 1977. and 1978 using aperatures of 229, 224, 213 and 193 K were oh- high-throughput Ebert-Fastie spectrometer to resolvetalned with no significant yearly differences. the rotational structure of the OH (8-3) vibrationalThe mesn temperature for each month is about 15 K bend.higher than the respective monthly temperaturesin the 85 and 90 km, 60 and 70*N, CIR (1972) Observations and Data Reduction

model, and it follows the general decreasing

trend show by the model. A I-s high-throughput Ebert-Fastsi spectrophoto-meter coupled to a minicomputer was used to obtain

introduction the night sky spectre. The instrumentation hasbeen described sore fully in the papers by Sivjee

Temperatures of the upper atmosphere at the et a. (1972] and Roosc 1976]. For the measur e-

menopause level (i.e., 80-90 kin) are Important mens reported here, the instrument was normallyfor studying both the dynamics and the overall pointing toward the zenith, operating over the

circulation pattern of the mesophere and in spectral region 7220 to 7450 A, which covers the OHconjunction with other data can be used to study (8-3) band. The scan time over the band was eitherthe interaction between the aesosphere and the 8, 16. or 32, using a alit width of I ma, corre-thermosphere and stratosphere. Temperature also spoading to a spectral resolution around 2 A. inplays an important part in the local chemistry. the auroral region, auroraily enhanced atomic and

Rocket-,orne probes and ground-based remote molecular emissions can overlap the near-infrared

sensing of the night airglow O rotational line night airgo emissions, making it almost topossi-intensity distribatios have been the min mathode bi t obtain reliable 08 rotational temperaturesused to obtale the 80-90 la altitude teeperaire for long periods of time ]elarcm-J-oes, 197.;]Kvifte, 1967; Shefo, 1971; Sivjse et ai., 1972; Merleethar, 1975; Myrab4, 1984].Lenschow and Petzoldt, :1980; yrabi, 1984]. To avoid contamination by auroral emissions, we

Igh-latitude (i.e., above approxlmately 60) have used relatively high spectral resolution andmesopause region temperatures are very sparse have selected agnetIcally quiet periods mostlycompared t hose at lover latitudes. Rarly during the evening. whin the oval is wll north ofspectral measurements yielding Oi rotation tem- the observing station. Each of the OH spectra havepecatures at high latitudes were those of Chamber- In addition been carefully inspected for any possE-lain and Oliver [1953]; .Mlronov at al. (19581; ble auroral contribution. rhe presence of auroraMcPherson and 7allance-Jones [1960] and Krasov- could easily be detected, as this spectral region

sky et al. [19621. However, most of those spec- contains both the strong 7f21P (5,3) and (6,4) bands

tral measurement were of poor quality for accur- as vell as the 0i (2D- P) lines. Spectra shovingate temperature estiCmaes and some are even knvn any sign of these auroral emissions have been re-t be In error [Kvifte, 1961, 1967; Wallace, jected from the analysis.1909n]. Later and probably more reliable (ground Fros the selected data set, 30-minute and 60-

hmsed) spectral date from which OH temperatures minute Integration periods acre used to obtainare obtained were reported by Kvifte [1967] and the individual 0. rotational temperatures. TheShefov [1969]. Si milar data from airborne observa- OR Pl(2), PL(O), P

1(4), P

1 (5), and PI(b) lines and

tions were described by Moxos [1964] and Slvjee et the method of Rvifte (19591 were used to calcu-ml. (1972]. in sit, temperature measurements from late the temperature. he rotational term valuesrocketmondes were made from Fort Ureely (64'S) and used in the plot vf In (i/o") against P(J)Helss island (81'S) (CIRA, 1972). hc/K were calculated using the data taken from

The inst recesc observations are the O tempera- Dieke and Croswhite (1948], Kylfte (19591,

lures reported from Longycarbye., Spitsbergen Herman and Horobeck [19531, Bass and Garvin78.4'S] ]Myrsbd, 1984 r nd h 1984; (1962], Chamberlain and koesler [1955], and Rosen

4yrab at al.,1983] and the rocketsonde temperature (1970. The ter m values for v 8 are almostidentical to those listed by Coxon and Foster

- eave from Norwegian Defense Research Es- [1982].tablisent, 0-2007, Kjelier, Morway. it -as been suggested that the temperatures

20n leave at AtmospherIc Science Section, Ma- quoted here should be recalculated using thetional Science Foundatlon, Washington, D.C. transition probabilities by Mies (1974] Ic.f.

Meriwether, 19751. Use of Mies's values lowers

Copyright 1984 by the American Geophysical Union. the calculated values by approximately 4%, bring-Ing them closer to the CIPA (1972) model. There

Paper number 4A0766. Is disagreement, however, am in the fIt of ien's0148-O227/84/04A-0766S02.00 calclations to the enperisental data [of.

9153

39

9154 Myr.6 et 41.: Bried Report

S8-3: ively. The January through April lean is 219 K.

- 22j23 P241 a 5 P'J In revtew, the data In Figure 2 shoe no clearP2 1

P)P4) p51 2a(6) tendency for a yearly trend, thus implying that

- 2 3) 4) 3: 0:8)f there is a solar cycie/ geomagnetic activttyA -~- ~ 5,9 c.c. -- hecmesopause temperature,.itprubanlyO [ N2 P has to be of he order vf 5 or es for theZ 15-3) particular period represented hare (i.e., 197h

through 1978). It should, however, ha streased- I! -l chat the amount of data ts too sparse to rule out

to obvious froa Figure 2 that the measuredt , the possibility of such an effect.

OH rotational temperacures are higher than theappropr iateA FL A (19 7 2) model ceaperature Theaverage difference between the 60'N, 90-Ja helge

model teaperature and the ....urej mean (dashed

7250 7300 7350 7400 7450 curve) Is about 15 K during any part of the tia

CLVhltGTs I) period covered. The difference heteeh the C&Afig. p o e e d(1972) 1.9 end 7s'N temperatures for ,the parli-

Fig. o deduce h lr height interval, i.e.* 85-90 ka, and tIme0 rotstional temperatures. The spectrum 15 of the year is oly 1-3 Kend therefore the

obtained by summin li2 lb-s indidual scans over latitudi-l change could little impact on these30 in. In addition to the OR (8-3) line., the results. L' he lifetime of ecited 08 soleoulee

locations of tha 421P band and Ott Lines origi- Is long enough to permit cheraliZatiu L thehating in the aurora are Indicated. The lack of i eni h

any broadening at the base of the P1 (2) and P2 (3) ambient s at 85-90 Lm, the OR rotatoal tee-

iInes indicates the absence of any auroral eel- perature should be representative of the neutralgas temperature of the emitting ragin withi tdegree or so [Erasovkay at al., 1977]. Fromrocket observations, 08 emission is found to peak

darner at ai., i983'. It le questionable, he- between 85 and 90 ha with a typical half width ofever, if the corrections to the lambda-doubling 10 ka (Evans at l., L973; Witi at el., 1979;

para atrs would have any effect on the derived Watanabe at al., 1981. missions from the highertemperatures, especially at the higher vibration- vibrational Levels peak slightly higher in theal quantum numbers in uae here. We have there- atmosphere thas do the emissions from the lowerfore elected to retain the present, traditional vibrational bands [Shagayen, i9801. Thus thereaethod of temperature calculation until we have is no reason to believe that the rotnioal

iavestiga:ted more thoroughly a comparison between temperatures as deduced from the 08 (8-3) bandthe methods, emissions do not represent the neutral gas tsr

An example (Figure 1) ,i a spectrua obtained prcature in the 85-90 ka height region. e ayusing a 30- inute integration time illustrates therefore conclude that the observed neutral gasthe lack of muroral features and the typical temperatures are in agreement 41th the trend butsignal-to-noise ratio for the individual O lines differ somewhat in magnitude copared with theused to calculate the temperature. The similar CIKA (1972) model teaperatures for this particu-shape at the base of the P(2) and P2(3) Lines Lar latitude and time of the year.copered to the Fj(3), P 1

(4), and PI(5) lines A simlar difference was reported by Sivjee Itclearly shows the lack of any auroral contribu- al. (19721 for the 60"-70N region. showingtin. In addition to the OR lines, the wave- mean temperature for the flights close to 225 K

lengths of the main auroral features are indi-cated. 253

The typical uncertainty in each calculated :,7temperature (as given by the standard deviation 240 't.of the regression line (see Kvife, 1959) is a- - .

tinted to be t5 K. From the individual tempera- 2--cures, daily means are calculated. Each daily mean 2 -- - -

covers a minimum observi ngc periLod uf 2 hours,wlth the average being 4 1/2 hours. 210-

Results and Discoabon e200 '2 60 -

Daily nean temperatures deduced e previously 50

described for the years 1976, 1977, and 1978 are

plotted in Figure 2. The temperature values for , 4

tae a n

each year are arked oith d fferent sybols. A roE OF ' aee fit second degree polynomial curve (dashed Fig. 2. Daily mean night airglow O rotational

curve) is drawn through the combined Aet of temperatures at Poker Fist, Alaska, daring theindividual daily means. The CIhA (i972) tempera- first 4 months of the years 197b, 1977, and 1978.

lures for h0 * at both the 85- and 90-k heights A best fit second-degree polynomial curve is

are also plotted In Figure 2 for the appropriate drawn (dashed curve) through the net of Lndivi-time of the year. dual daily means. The CILA (1972), 60', 85- and

The 3-year means for January, February, March, 90-km height temperatures for the respective

and April ore 229, 224, 213, anod 193 K, respect- months ore indicated by the solid carves.

40

4yrabd et a.: Brtef Report ;L55

during Januoary to March 1968, which is at least during the 1.2;.Y., ophy- .10-r,.t12, _ 19591 0 KIn higher than Cthe app roprIate CIRA (1972) Kvifts '2. Temperature en uree... r- 36

mon. temperature.- Further idirraces,%may also beads, P1nt.Sac !t.5 153,19.befound in summery of 0H rotational tepera- Kviftt. C2..T244M* i.,rcy rttIonal temperatu~res

turrev ith latitude by Kvlfte [19671. From the"e ad tatsnsities in the oighcglow, Fluoc.data 6 5 mIdwj nter OH rotation"l temperacure SpacE Sr., 15, 1115, 167.above 240 K say be deduced. The tendenc othe Le.c. w It. . and K. Petzoldi, 3oheniarten der,2

C18A 197) espaue idinter temperatr odel abkr-Plache sovls monatlittle M4itteire r

at high latitudes to be somewhat low is further den zeitraum Oktobor 1970 bis Dezember 1921,

uapoted,:by our ""cent findings froe langyear- Meteorol. 466. Berl~n. 97, 2, 1980.

a e, Spitabe rgen (7 8.4'.1) during the 1982/19613 Ic herCona, D. H., and A. VainAlnn 4 tudy

ueasoo (M1yrahb, 1984;,4yrab4..t al., 1984]. of tbs latitude dependence of 3 Ota oa

hu the avilable wtecr seo apause te~peratures temperatures for Canadian stations. is AtMoa.ded"ced from -on ib racLon-r o ttion epectra at Tart. Phys., 7, 302, 1960.high la1ttudes all show the same tendency; i.e., feriweth r, J 1. 4.High latitude airgiow observe-the measured values are higher that the CIBA tE = of corrslattd short term fluctuations Ift(1972) modm. jbut the general dacrsagtrendo Is hydroxyl Miael S-3 hand intetsity and rota-the same. S Imtannu Crolnd -basd Og roattion- ti onaI t emprature. Pianet.- Space Sdi. , 23..1 temperature measurements from sevetal high- 1271 1975.latitude stations are oeeded to clarify the 4Is F.,* Calculated vibrational transition pro-masopause t emperature behavior at high latitudes. babilities of 08 (X

2,). J. Molec. Spectrosc.,

53, 150, 1974.Acknowledgemnts. Financial support for this Mlrono, A. V., V. S. Frodukina, and NI. N. Shefor,

research wae provided by the National Science Some results of Lavescigatione of night air-Foundation through grants AMO612719, 4TM82-i44h2. glow and aurora, hAn. Geophys, 16, 365, 1958.and AT5M22-00 114 to thS Geophysical lmstitut* of M,,yrabO, gt. K., Temperature variation at 2esopau .the Snivers ity o f Alaska. One of us ( R51.M.) is levele during winter solstice at 78-N, Planet.alao supported by a fellowship grant from the ac * . 2., 249, 1984.Royal ,Norwe gian Council for Scientific and Indu-ya~j~ e nd C. S. Seebr, Mid-winter by-trial Reearch. dromyl ni ght sirgiow emission intesities in

The Eitnor thanks K. J. Llewellyn and another the nor thern pola.r region. Planet. Space Sdi.,referee for their assistance in evaluating thia 32, 263. 1 984.paper. 4yrabo, H. K., C. S. Deehr, and G. G. Siojee,

Large-amplitude aightginw 06L (8-3) band loten-

References sit y and rotational temperature variationsduring a lb-hour period at 78'N, J1. deophys.R2_., a8, 9255-9259. 19 83.

Sas &.3.*ad SD Garvin. oliso the by- Mouohat, B.. K. Fetnoldt. xK. Lesbitnk., and B.d-8oAcylrdclvbaonrtin spectrum Lenschow, Bol. Berl. IdettethArts, 57/83, 1983.between 3900A and il,50o A, 3. Ma1. Spectrogc_ Maon, 3. F..* Th, latitude dependence of OH69, 114, 1962. rotational temperature in the night airglow,

Chamberlain. J. W., and S. J. Sliver, 08 in the J. Geophys. ges., * 9, 4087, 1964.airglow at high latitudes. Phys. Rev., LO9, Romck, G. I., The detection and study of the1118, 1953. vieibie spectrum of the aurora and airgiow,

C hamberlalo. J. W., &ad F. L. Roesler, The 06OHc o.Pht.Ot ntrm n. 1

bands in the infrared airglow, Astronhys. J.., 37, 96121, 541. 1 955. Rosen, 8.,* io fnornarionel Tables of Selected

320000, 4. A., 'nd SC.Foster. Rotational anely- constants. Table 17 spectroscopic data tel.-.i of hydronyl oi~brationrotatinn emission tios to diatomic molecules, Pergamon. New

h a: Molecular con' tats for 08 X2v,6 ok 190

41,Can. J. Phys., LO6, hi 1982. Shagayev, M. V.. Vertical temperature gradientsDslke, 2. 5., and It. g. Crosewhlte, Bumblebee and dissipation of internal gravity waves near

SerCies. Johns Hopkins 00±0., Balti- Che meeopausv, Gaese. Aarnn., Engl. Tranal. 20.

moreo, Md.. 1418. 529, 1980.Eoans. W1. F .,.* 1. 4.Llewellyn, and A. Vallanc,- Shefov, N.4. S'ydroxyl emission of the upper at-

soeAttd itiaio fhdoy ad ophere I, The behavior during solar cycle,of the Aom2 sequence in the nightglow. Can, seasons and geomagnetic disturbances, Planet.J. Py 1, 1288, 1973. Spc c.*17, 797, 1969.

Hesinman, K.'C. and G. A. Hornbeck, Vibrat Ion- Shefr 9. ydroyl emissions of the upperrotation bands of 0H. Astrnphys. J., 118, 214. atmosphere t11, Diurnal variations, Planet.1951 Spars Sdi. 19, 2. 91K- o uuky, 5?. 1., 9. 48. Shefoc, ad V. 1. Tarma, Sioje., . G., . A. Dick, adp .Flmn

Atc of the airglow spectrum 3000-12400 1, Temporal variation in the nighttime hydronylpla. SpceS., 9, 883, 1962. rotational temperature. Planet. space Sri.,

Kresnsky, V. I., a . F. Potapov, A. 1. Semenov. LO. 261, 1972.V.3. nolev I. V hagaev and N. 9. Shefov, Vallance-Jones, A., Auora,. S. Reidel Hingham,

on ts equilibrium.natur o f rotational raper- MAss., W94..Cur: ofdhydroxyl airgiow, Planet. Space Sri.. Wetll"ca, I, Not. on aSlow temperature determL-25. 596., 1977. nations from 0H e pectra. J. Geophys. Res.. * 5,

Kolfte. G., Night airgiow obsetvatlone at Is 921,. 1960.

41

9156 Myrabd ec at.: Brief Report

Wacanabe, T., M. Nakaaura, aad T. Ogswa, Rocket green line in the nightglow, Planet. Space Set.,

easuresents of 0. ,t-ospheric and OH Neiael 27, 341, 1979.bands In the airglow, J. Geopays. Bee., 86, 578,1981. C. S. Deehr, C. J. Roelek, and 0. 0. Si jes,

Werner, .-. * P. osiuae, ad E.-A. Retonch. Mole- Genphy I I.stitut, n vereicy of Alaska,cular properties fros CSCF-SCSP wave func- Fairbanks, AK 9970L.ctone, 1. Accurate dipole aaent functions of H. K. Myrabd, Norwegian Defense ResearchOH, OH-, and Off, J. Che". Phys., 79(2). 905- Establsheant, N-2007, Kjeller, Norway.916, 1983.

Wicc, G., J. Stegsan, B. H. Nolheim, and E. 3. (Received June 28, 1983;L eweilyn, A eeasureaenc of the 02 bjZ + revised Apr11 20, 198h;x Z,,) ocepheric band and the 01 5) accepted May 23, 1 94.)

........ ..

Ale

43

Pnied ,n (00%a Bitn ,.95, Pers.. moo Pre. La

POLAR CAP OH AIRGLOW ROTATIONALTEMPERATURES AT THE MESOPAUSE DURING A

STRATOSPHERIC WARMING EVENT

H. K. MYRABO ad C. S. DEEHR

Geophysical Institote, Univi's:ty of Alaska, Fairbanks, AK 00711 _ S A

B. LYBEXKInsitute of Physics. University of Oslo. Oslo. Norway

(Received infinaljform 15 .Vovember 19831

Absarct-OH 18-3) band rotational temperature was observed at 78.4'N dunng a stratospheric warmingevent. A negative temperature wave of the order of 40 K observed nca dhe mesopause seems to be associatedwith a corresponding stratospheric warming of the order of 20 K. A 1-2-day delay is observed between themaxtmum stratospheric warming and the maximum cooling near the mesopause seen in the OH rotationaltemperature change.

INTRODUCTION the instantaneous mcsopause temperatures at these

The sudden mid-winter stratospheric warming in the latitudes are for the daily or weekly averages is highlyNorthern Hemisphere was first discovered by Scherlag questionable (Myrabo et al.. 1983: Myrabo, 1983. see119521. During such a warming, the circulation pattern also Noxon, 1978 : Weinstock. 1978). This may also beand temperature profile of the stratosphere and lower inferred from the large variations in the temperaturesmesosphere are changed (Finger and Tewels. 1964. observed iCIRA. 1972). More recently, satellites haveBasically the cold polar vortex existing over the provided world coverage of stratospheric and mesos-northern winter pole becomes broken and zonal pheric temperatures up to approximatey 60-70 kmwindflow is weakened. Rocket temperature studies and altitude. However, the satellite temperature data aresatellite radiance observations show that the mesos- not reliable above 60-70 km altitude. Thus. ground-phere cools under these conditions, Changes in OH and based remote sensing of airglow emissions is necessaryNa airglow emissions have also been correlated with to obtain temperatures of the mesopause reeton i i.e..

the stratospheric warmings IHunten and Godson, 85-95 kmi)with a much better temporacontinuity than1967 Rudle and Sullivan. 1972; Shefov. 1973 ; Reed, can be provided by rocket flights.1976: Fishkova, 1978). The cooling in the upper The purpose ofthis paper is to report nightglow OHmesosphere I Labitzke. 1977) lat least in the 70-80 km rotational temperatures in the winter polar cap regionregion), should result in decreased airglow emissions duringastratosphericwarmingevent.Amoreextended(Moreels et al., 1977) contrary to observed results treatment of mesopause temperature behavtour in theIReed, 1976: Walker and Reed, 1976). These observed winter polar cap has been given in an earlier paperenhanced emission rates are therefore probably mainly lMvrabo. 1983).due to a mixing and redistribution of the atmosphericconstituents associated with the breakdown andreversal of the polar circulation pattern. OBSERVED TEMPERATIRES

The development of the rocketsonde led to in situmeasurementsoftemperaturewithaltitudethrough the The OH emission data employed in this work weremesosphere, and typically a few temperatures were taken near Longyearbyen on West Spttsbergenobtained each week from regular observations. Heiss 178.4°N: 15E during the 1982,83 Multi-NationalIsland (81'N: 58°E) has been the main site for Svalbard Auroral Expedition (Deehr et al.. 1980. Theobservations in the polar regions. How representative spectral region 7280-7410,A covering the OH (8-3)

band was scanned employing a I m Ebert-Fastie

-On leave from Norwegian Defense Research Establish- spectrophotometer operating in the photon countingment. N-2007. Kjeller. Norway modeand coupled to an on-linedigital data processing

853

44

154 H K 'ANtss.,t,

systemiSijeeet al. 1972). A band widthof I.5Aclearl ., .. 260resolved the OH 18-31 P lines from auroral emissions. -250

The spectrophotometer was normally run for 24 h a -day during mo,t of the winter solstice period from 5December 1982 to 30 January 1983 Individual scans -230 _

obtained each 8 or 32 s were summed and averaged to -220obtain hourly mean Intensities- The temperatures were 0H(8-3) Rotational-210then read from the houriy means of the OH P, lines I TmDera:ur 200using the method of Ksifte (1959). Probable error in a _-. 100single calculated hourly temperature is estimated to be E

3 K. The occurrence of aurora is indicated on the 9'8high latitude spectra by the appearance of the [Ol] 80 SL C -170"320-30multiplet among the lines oftheOH 8-3) band. 2S- 75 1 • i]~27-1 7 nteari

'Molecular emissions such as the N. I Pos. bands are 0 --only rareh seenfrom LongyearbyeniSileeand Deehr. -19801 and these would tend to increase the deduced s0 .1 \ _30rotational temperature by contributing relatively more -[ 30to the PSfand P,6ilinesMerwether. 19751. Further edetails on the instrumentation, data takinit. and ' 40 -A-50

reduction, and examples of the qualit of the data are S . - ,,'b a

venbyvMvrabo 119831 and Myrabo'etal 119831. -50 28-4 sia r E -- 50

In Fig. I is plotted the daily'mean OH rotational .-----temperatures, together with atmospheric temperatures -. .60 r

-

deduced from the SSU satellite over the northern polar -7 0-.-'.'__region at different pressure levels as given by Naujokat . - Oer ul. 119831 for the 1982 83 winter. iThe temperature is 80 "" V 'J A3 bproportional to the radiance which is given in Fig. t forthe 1 7 and 4 mbar levels. The exact proportionality 20 1 10 201102 0 1 1 1020 1020depends upon the atmospheric composition. so the Nov 52 Dec 82 Jan 83 Feb 83 Mar 83

authors elected to lease the data in units of radiance.) F I OH 1 H RiiyTIOvv. rEsIPES ROMThe temperatures (radiancesi are arranged vs hetight- bio,-v'EARBNEN1S4 N. i A, aND SSL SATELLITE TEPrslPER-30. ll).4and I.7 mbar corresponding to approx. 23. 30. rLRESt ,o A .tViAsES 1,vR THE NoRm POLE AT DiFEE'.T36 and 42 km altitude. The OH (8-31 rotational PR.55 RE LEVELStemperatures are representative for the 90 km region The latter are obtained from Beilage zur Berliner % etterkarne

with an altitude thickness ofabout lOkmWatanabeet 65 1983 tNauokat ,'r a. 19S31

al.. 19811. The first warming event can be seen to occurin the end of December while the second main event lower levels relative to the higher levels ,ould beseen at the 10 mbar level took place near the end of expected. ifthe energy in an upward propagating %axeJanuary Lnfortunatelv. our data do not include the were to remain constant and proportional to it thensecond event, as the density decreases with increasing altitude. the

wave velocity has to increase proportionallyIn addition to differences in fluctuation amplitudes.

the absolute value of the temperature deduced atLongyearbyen may not be representative of the vaiue

From Fig. I tmay beseen that thetemperaturesat all averaged over the polar cap. w hich would correspondaltitudes show irregular fluctuations from day to day. to the satellite data for the stratosphereThe amplitude of the fluctuations is typically twice as The sequence of events reported here is in keepinglarge in the OH data at 90 km as it is in the satellite with a currently ac,:epted theory of stratosphericdata from the 30-40 km region, and even smaller warming by Matsuno i 19'1 in which the beginningfluctuations are normally seen in the 20 km region ii.e.. involves an increase in the amplitude of tropospheric30 mbar. It is also seen that the different heights show planetary waves. This results in increased polessardsignificant deviations and a peak-to-peak correspon- heat transport at lov levels leading to a rising motion atdence cannot always be found, not even from closely- high latitudes and a sinking motion at lo, latitudes.related curves. The difference in amplitude between the The Coriolis force on this rising motion leads to

45

P,,ar ap ()H iuLo. roEatona temperature- -

decreased westerlies at high levels II these are strong 4,knoieuqenesi %Ae would like t thank Dr autoat,enough to become easterlies. a critical leiel is reached Petzoidt. Labtzkeand Lenschow forthe timely disrhu: -n o

for itationarl planetar' ave, The pianetar ale tratospheric and rnesospnerc temperature or ur'e, !or :he- SM2 S3 winter season Financiat support for this research .as

energy thus cannot penetrate thicritical level resulting prouidesdbthe National Science Foundation through grantsin poleward heat transport onil below it This gradient ATMSO-i 219. -T\t82-l.4 and -T'vto2- O. l ii the

in heat transport with altitude lead, to a secondar Geophysical Instituie of the Unlsersirt ol Sassa ) one it as

circulation which includes upward motion in the polar ,HKMi is also supported ny a feiowship grant from Roal

regon accompanied by expansion cooling above the Norwegian Council for Sciences anti Indusirta Roscar;h

critical level nearly balancing the poleard heat

transport below that level. This sequence ofeents has

been observed previously but the observtions re- REFERENCES

portedhereextend to highaltttudesand latitudev witha . IRA. COSPAR AorkingGroup 4 COSPAR [nternaitona.

time history which may b, important in the further Reference Atmosphere. I)2 Pergamon Press L:a (txiord

development ot the theory see Geller. lost) Deehr. C S. Sotee. G G, Egeiano A Henr;ksn. K.

it is clear from Fig. I that the extremeis cold Sandiholit. P E.. Smiuh. R . Sweeene. P Duncan, C andGilmer F i tAtli Ground-based obsersatton, ., F -"ieon

mesopause region over Longyearhyen around I associated with the magneisphenc cusp J apt's.-'

January 1983 may be associated with a correponding 85. t5

warm stratospherc region down to below the Itttbar Finger F G and Tenets. S ,1t,.i The mid--inter 463

level This is in agreement with rocket and satellite stratosperic aarming and circulation nanee .1 apr,

observations at the 00--'G k,.eel during stratospherc ileteor 3 I" Fishkosa. L Ni| i lA95i Intensity tiuctuations il Shr nocturnaiwarmings tLabitzke. 19-2. 19-1 The result presented emission of the upper aimosphere dunng tratosohert-

here shows that the cooling of the lower mesosphere -armings 6iiun 4ere 18 3"

previously associated with stratospheric warrthngs Getler. It A lsmi' Middle atmispphere ynanc andextends throughout the mesosphere up to at least 'sit - composition. in E pl rai.,i iou P a'r L -e' tts',--ner

km iEdited by Deethr C S and Holiet, J A . D Reideialttitude, and also that the central polar cap meso,phere Dordrecht

cools in much the same way as the lower latitude Hunien.D M*I and Godson.\k L ,I'6 I.pperatmosphere

mesosphere observed previously during stratospheric sodium and stratorheric uarminu at hiah ajitude Jwarmings. ar.- So 24. 1O

1-2-day delay between the KiLfe. (a l99) Nightglow onxersations at As cturnt theThere seems to be a teh C G tphis Puht 20. :

mesopause minimum temperature and the correspond- Labti~.e. K., 192 Temperature changes n 'he -nesispnereing stratospheric temperature maximum This may be and sratosphere connetfu with circuaiattn hiancits in

due to the timeit takes thesecondar, circulation system winter J at, S, , 29tO de.elop bove thecritical level or the time required Labiizke. K 'iS", Stratsspherc-mespver,. ma-alve'

to dvelp abve he arminps. .n Disanv . 1a C-111 5nt i, " C urn,for the mesopause to cool radiatively once the source of Ie.tral and l,,nimea tmnperel Edited hy Cr ina B and

heat from the zonal wind has -lowed at the meso- Hotet. J A... p I- D Reidel. Drdrech

pause level t2 days. Rodgers. 19-'). or it myj he Maisuno. T 1,2-i, A ds:iami a moue t -he orii,'snner,

the geometrical effect of the limited coverage of the sudden warming I aini, Si 28 1as-\tenssether.J A 5f ihtttdaris 'sraon'

Loneearbyen measurement compared to the satellite elehrt-tr t!ationsin the ltore.\tne-- - ~~~~~~~~~correlated short -term fl uato mth )d( ' lmlmeasurements integrated over a far larger area 3 band intensiiy and roatisnai temperarare l'ar a3p,-

The initial drop in temperature at the mesopause is Si 23. I21

followed by a heating which extends at least 2N-30 K Moreels. G G . Meete. -\ alilance-Jones A an' ltatt-ner

higher than the average. A corresponding cold R 119-" An orygen-h, drogen atmospnerc m,,e and ;isapplication to the OH omission proelem J .i. i'r.

stratosphere is not evident. This mat be characteristic Phpi 39. 55 O

of the circulation above the critical le el. or it might M trabo H K. Deehr. C S and Sittee. G 6 ', Laroe

imply that the heating is a local effect amplitude nightgiw, OH -3i hand nons ty arnd

Unfortunately. our measurements stop just it the rotational temperature iarions duing a 24-h,,ur period

bengning of the large stratospheric warming at the end at Ao N K o rpni ! Ret 8t ' a25 nAty~ratte. H. K. i Ax i Temperature saiatis'n nd mnstiauseofJanuary 1983. From the data seen in Fig i one ma. levels during winter soitice at 'A N P!anet ri S,: 32.

deduce that at least a 1-2-day dela, between 2a.o

mesopause minimum temperature and stratosphere Naujokat. B. Petzoldr. K. Labitke. K and Les hou. R

maximum may also have been the case for this event. , t991, Beilage zur Berliner AWeterkarte. s- s,

NoxonJ. F i 19'Si Effectsoftnternai gatt wales unon sightFuture plans include an extended obsersing period airglow temperatures G,plr s Re Letr 5. 25

and the establishment ofan additional ohser, ino site it. Reed. E. I i 19TSi Polar enhancements of nightgiow., emissionsAlaska. near 6230 1 Ge,phi, Res Lett 3. 5

46

856 H K MYR^ao et ai

Rodgers. C D t19" Morphology of upper atmosphere Sijee. G G and Deehr C S 119801 Difference, in polartemperatures. in D inamical and Chemical Coupling Between atmospheric optical emissions between mid-da, cusp andthe Neutral and Ioni:ed Atmosphere (Edited by Grandal. B night-time auroras, in Expior ion ot ine Poiar L pperand Holtet. 1.1 D Reidel. Dordrecht. See also p. -4 4imosphere(Edited b_ Deehr.C S and Hoitet.J i.p 199 D

Rundle. H N and Sullivan. H M, 19721 Upper atmospheric Reidel. Dordrechtsodium and stratospheric warmings atmos. Sc 29. 97" Walker. J D and Reed. E T 119161 Belsasior of the sodium

Scherlag R t19521 Die explosionsartigen Straiospharen- and hydroxyl night-time emissions during a stratosphericwarmungen des Spatwinters. 1951-52 Bet Det warming J atmos Si 33. 11SWetierdienst 38. 51 Watenabe. T., Nakamura. M and Ogawa. T 119811 Rocket

Shefov. N N ( 1973, Relation between the hydroxyl emission measurements ofO, atmospheric and OH Meinel bands inof the upper atmosphere and the stratospheric warTnings the a.'rglow J yeophis Res 86. 5768Gerlands Bear Geophis. 82. 111 Wetnstock. J (19781 Theor' of interaction of gravy wavesSivee.G G. Dick. K A and Feldman. P D (1972iTemporal with t0 i'Zlaurglow J qeaophis Res 83. 51'5variations in the night-time hydroxyl rotational tempera-ture Planet Snace Sri 20. 261

-77

q-

"Vm

.. IAI.

-4:

7'.

ji

48

JOURNAL DF .EOPYPIRAL RESEARCH. 'OL. 92, N. 42. PAGES 252'-2534, "ARCH ', '48,

0oLAR MESOPAUSE C RAVITY 4AVE ACTIVITY IN 7E C0Dl!'M AND HYIROXYL NbHNT AERCLCj

4. K. Mvrab6., Z. S. Teehr. and R. -ierck

;eophysical Institute, University of Alaska, Fairbanks

K. Henriksen

Auroral Vbservatory, Troos. Noway

Abstract. The TO (6-) band and the Na I 55-4& line are useful for obtainiog ingornationlines in the night airglow have been observed n gravity aaves, T addition, these eisslonsfrom Spitsbergen '79'N latitude) during a I-day might give valuable information on the listribo-period in the end of November 1983. Regular and tion of minor constituents, odd oxvgen and trans-qaasi-regaiar vaniatians in tenperatace and th part processes, and chemistry of the mesopauseOH and sodiue intensity nec aheerved and are region.here Interpreted in terms of gravity wave theory. The :an parpose of this paper is to reportIn particalar. naves with periods in the range and discuse recent ahservetions durIng the polar

0-120 mm. baring overlying ripple structore, night of gravity wave activity as manifested bowith periods of 5-IN m, ae observed. Fran regular and quasi-regular variatlons in thethe phase difference e.wea the eaves in the Ov neutral nemperanure and the OH (h-2 and Naand a .diac layer a height difference between the night airglow intensities.centroid of the to emiesiana of less than I hecould be dedced. A gravity wave induced nerti- Vbservenions and iota Redactioncal eddy diffusion coefficient in the range n,,6 to 9 . i o cm

2/s s astinated froe PH-an- During a 3-day period with tlear wa they from

hanced intensities. No significant net heating November 20 to November 3O, 1983, emission spec-or caoling nan observed during gravity wave anti- nra of the enith sky, containing the 3H (H-2U Pnip. The gravity wave activity as not associ- brenches, were taken near Longyearbyen on Westated nith geomagnetic activity. Mean SR 26-i) Spitsbergen (I76N latitude, 15E longitude.band and sodium night airglow intensities were geographic). The night airglos sodium doublet1.9 kR and 75 R, respectively. wan also recorded for part of this time (ie.,

from 0200 fT November 27 to Imps lT NovemberIntroduction cR1. Tbe measurements nets part nt b 3/19R4

h ulti-4ational Svalbard Auroral Expedition ODehrThe cesopause and ioer thermosphere region t al., logo]. Two spectrometers. a 1/2 ' and a

(i-e.. 90-100 kef reflect interactions blween the I N Ebert-Fastie, were used to obtain the OR andupper, middle, and lower atmosphere and thus hold sodiu emissions, respectively. The spectro-important keys to the understanding of local and eters are derlbed in detail in the papers byglobal dynamics of the atmosphere [Holton, 198: Siyee et al. 19'21, Roik lqAT, and NvrabNohanakucar, 1996: yrab6 et al.. 19851. et aI. [19611.

Gramity waves play n important role in the The 1/2 N instrument recording the OR (6-2) ppolar mesopause region [Ebel, 8R4: Noton, 183; branches eas set up to scan the spectral regionLindze, I1991 5. Solomon, private comc.nicntions. from 9370 to 95705 in 12 a oith a bandwidth ofIPR8f. Gravity wave activity in the polar winter 00. A sum at 23 indivi doa ecns were used toatmosphere is reported to be mere frequent and obtain a spectrum, corresponding to one spectrumviolent than at ler latitudes (Tarrago and each 5 ,mI An enasple at a typinal 5-in nte-Chanin, 1982: Juramy et l.. 1981; Hirota, 1984; grated spectrum is given in Figure 1. The 1 Myra b6 at i., L9831. Juramy at al .[1981[ instrument wae set up to scan the spectral regionreport a 1000 occurrence freqhency of gravity 5RO nO 5965A in 16 s with a bandwidnh carrespond-waves in sodium density profiles obtained with ing to 4A. A sum of iS scans we used to obtainldar fram Helee Island (81'N) compared to less an integrated speatrue in 2 Am. A typical

than 20% (Tarrago and Chanin. 1982 occurrence at 1/2 min spectrum of the Na D lines with the O1Haust Provene t44'N . Thus gravity eaves may and D

2 lines well resolved Is shown in Flgure 2.

take e nme active part in the circulation, One of the main problems with measuring mightdyneeics, and energy balance of the polar atmo- airglow features in .torral regions is the cnta-sphere than is the case for lower latitudes. eietion by aurora. If the emise.ane are not

The night airglow emissions originating in the excited or otherwise directly influenced by80-100 km region such as the OR eissio.s the 02 uroral particle bachardment, overlying aurarsl(0-I) atmospheric band, the N D lines, and the features may normally be filtered out or accounted

ftr given high enough spectral resolutin fMyrab

et al. , 983. In the polar region., auroral

T eave free Rorwegian Defence Research emissions are mainly those of atomic linen

Establishent, R-S0O7, 1jellet. NorWay. [Siv ee and .eehr, 1980: ault en al. . 1941] dueto the dominating soft electron precipitation

Copyright 1987 by the American Geophysical Jion. spectrum normally present in the cusp and polar

cap area. This takee the night airglow compo-Paper number 5A8316. meets, such La SR emiseione and Na 0 line,. easy

0148-O227/R7/005A-8316S05.00 t isolate. The spectra l region shave in igure

2527

49

252' yrab eal. ?oa v 1 ran.to -ayec mn 1a '0

Seni ssnns nan theref re safel v' :e aen t' terve-

2S 0W ,-2)ra i- th no aurora 0adt4rnuvd onrantna-

tis'ave beens red for ca-rn lti; toe + ota-tional tertperature and the absolute tntensitnen

2la fth H si'hn n the ha 1 line Thea :ensities of the lines have teen alc-ulated (ro

the area under the lines rather th'an onv the

I peak vai.es. OH rottionaL temperatures have

I feen calculated frn" the intesiti lintrihu'tln

ji0 * ofthe P-I.? 3 ? . ed P tne seeF o (igure I . assuming a loItzmann tinrhu tlnn

h e uit plet votioral levels. re.

... i . "'J',i .', Nv'k IA i.1 '. '

-ovrLe.WT. (A) Iv' (Trot) k Trnt

Fig. I. A typical night airg w spectrum caner- here I is thephoton intenatv in 'oto e

Ing the KH (6-2) P branches. The spectrwm oae hre i ' i the intnic-ntator a -

obtined daring a 5-mis integretion of 25 ndiv'- c Te c I )v, i, the electronic-rotvt'orot

daM 12-s scans. The handwidth as 7i. The On partition function for the v' level. Nv' is the

(6-2) P lines and the aarnrai (1) 94461 line are totl voncentration of nolecules in the v' nibra-

indicated. ninl ievel. Trot is the derived r.itionai tem-perature, and Ej', v' (') i .heuneersta eterm value for the vibrationl hand o' v' - he

I also includes the (O) 8446A auroral Ilne as other variables have their usu.I meaing. TheI ur else inEae th (0) is.arna inea

an aarorai reference emission. The (Oi) 9446 instein transition probabilities. A. gloss ho

line is directip encited hy electron bombardment Mies 1q74) have been used together oith energofVailanca-fonee 3973) and thos yan sensitive ereis, fir'(J.' ), by 1vifte :lnsn.

to any auroral particle precipitation in the Absolute calibration of the sectra oas oar-

field of riew. tied out in the field using a stndard source.

in spite of sarly reporte of auroralip en- The tota intensities of the 1H 5-2) band were

hanced sodium enissions [rbom, ig ~: untn,, calculated from the I lIne intensities and the

1955 Hunten, i9 h1. the stiv doublet is n a no tio al temperat ures. The uncertai n n the

believed t: be purely cheicalye1ite, alnl ahsoiute intensities of the OH (h-I' hand and thethrou the pl y emiciiy incted, an Na f .lte is estieated to e 20. The uncertain-

al., 19i. y in rae aheolute Intensities i' ainly due to

calibration uncertainties. The relative uncer-

a- 03

- N-0 a 02 (13 tainty in the emission Intensities due to photon

noise is less than 21. The arerage stndard

Naf . 0 - Na (2

p) . 02 (2 deviation in the temperature is *K.

which emits the Na 2 lines through the transition Results and Discussion

Ma () Na (25) + ho. Using a large data base,

aees et ai, (1975] failed to find anp auroratin Figure 3 gives an over-vies of the sodium D2

enhanced sodium esission. Both the OR and sodium line intensity. The averaged background under

the sodium doublet and the solar depression angle

2ss

a ,

011

0 2 5 ' a 4 22 0 5 e a 1.

eats 5550 ease asenesa, nw v-esWaVNeaM (a Fig. 3. intensity (R) of the sodium D

2 line

Fig. 2. A typical night airglow emissio spc- (middle curve) during tbe continuous 39-hour ob-

true of she eodiam 0(i) and D(2) lines at S81O serving period. The upper curve gives the solar

ced 89 A, respeativeip. I Th spectrmm :as oh- depression ang e, ahile the loer curve shows the

tIned during a 2 i/2-min intagration of 10 average intensity of the barkground on both sides

individual lb-s scans. The bandwidth eas 4f. of the sodium doublet (see Figure 2).

50

Octuc, en ul -Pojor Cay Crariy r-ayes n 'a n gLw2

00 -2602

100

5 OG 5 sw 10 21

3 4 5 0055

TIME (HAS)

rig noen4i.iso of OH (6-21 bnd, Na D2 line, and background during the night otNovember 27. 198 as taken from Loegyearbyen (7'Vo, The calculated 0 !6-> hand ro-

tational temperature is also given (upper curve).

are also plotted f or the nie period whan both ON and the potential temperature. respectiel- andAnd sodium emission measurements were obtained. ao is the unperturbed deity. seuning thathe seen from the figure. massrements were tken the relative variations in emiesion Intenstieson a 24-hour basis for a total of 40 hours, are proportional to the relative va rations inHowever. DH spectra with daytime Fraunhofer density i.e. A0/0 a 3w/u. where and u artabsorption features were not used for obtaining the emission and denait, respective"'v. themintensities and temperatures. Sodium resonant temperature and intensity fluctuation should heecattaring is obviously present after about 0fl30 80e out of phase during the passage of gravityUT (corresponding to approxinately 0845 local waes. However. the bulk ci simultaceous tempera-soIlar time or solar depresion angle i5"). The norm snd OH and sodium intensity variations,lack of data between 10O0 and 2100 UT was due to which arm possibly caused by gravity wanes, showcomputer failure. an in-phase relationship beeen intensity And

temperature 'yrabi eat al.. 1Q3: Shaganmo,3.1 SifuLteneous Teeperature and Intensity Waves 1974; Krassovsky. 197L; Takeuchi and 'lsowa.In the 0 _and Sodium LAyers 1q8l; Takahashi et 51.. 1485].

The sample gimen to Figure seqems :to repre-A closer inspection of Figure 3 reveals wave- sentaner simple case of gravity oars prcpaga-

like riations in the a D2 iine intensity not tton through the 90 to 05-km region as it main>

present in the background. Enhancements are also shows only w singl. A ude wave with an overlyingsuperimposed on the stw iight pert of the sodium ripple structure. The period deduced from theeission. i.e., around 1600 T. November 27. figure is in the range I 1/2-2 hours. lecauseExpnding the period from 0200 to 0900 ULT on the wave is continually present throughout thisNonember 17 and including the OH temperatures and part o the observing period. it probablv pene-intensities for the same period, the wve. are trane the 87 to 05-ks region oihout hreekicg.seen to he present in the sodium emission, the 0h The nertical waveleogth of the I th o o r ive

omission, and the oeutral temperature of the ti- might be estimated to he in the range -15gion (Figore 4). The background intensity is Phibrick en al., 1095; Takabashi e al.. 209.

indicated at the bottom of figure 4. The wave From the negligible phase difference seen be-pattern, present both in OH temperature and Na tasin the O and sodium emission, a ai-uoand 08 intensities. -y well be enPialnad by

gravity waves passing through the emitting regionNonun, IN7; Weinstock, 1978: Herse e at.,

1980; Nytabi t enai Q1;93; Taka hashiel , t-

1035?1 Als the powe r opatrao epeaue5variations versus frequency, presented in section3.d. strongly support this interpretation Smith SNY

en at 19s.-sAccording to linear gravity wave theory, 2

however (Hines, 1960: Holton, 19791, density and

temper.a.tr fi .ct.ation should be 180' outphs uder conditione he. Oo '/9o > p/c 5 ,

.

i.e.*. when density fluctuations due to pressurechanges are amall tompared with those due totemperature changes, then * a 5 5

8 /8 - -O /O. (.) Fig, 0. Intensliy variation of sodium DI and D 2

lineas A. ob taed irom Longyearbyen during theehere c, Is the speed of sound p' and 9' are the morMing hours of November 27. 1083. Background

deviation from the mean values, for the pressure is also shown for comparison.

51

- 0 OM-2) IV4203 ntatoo a ?oar

BEIM IsU.)Z • ,

e12-'e2oe *n 22 ' '2e

0.n Trce of te tsh o-fonsumed the 0 tio 4 :ne ad , -di- rbc et 7%:

ZOO44 T Noeme 29 9e8

- g. if.e re so,.. thei, loemnsitie of th T0he-f brod and the wsodr ertaine. Te th

, -1 _n na be deduced 7' naiumn don. propa'P- right 2r Iun tnl flr .rc-aioa Ccilia-jatlog group ce icity of l0 km/h is --sued. The clone , Sines ]. -~h t o- Phaibrick et al.axiu probable height difference is somewhat [l0851 one ny find -atwaes aith periods itn

less than that found by Takahashi et al. (Q153 the range 5-10 in should he reflected back ..:tcfor low latitudes, where the phase differsnce the lower acosphere fro heights of aproxiately

between the OH and sodium intensity oleariy couid h0-dO kr. respectively. Thus at least the nainhe seen. The bulk of sodium profile easureeents. body of the observed "fie etroctore" aes inboth fron low aed high latitudes [Chanin. 1084: Figure originates at altitudes higher than 02

Juramy ec al.. 1981 report the centroid of the km' which strongly supports that they are eoa-sodiun e-ision at about 0 km height. Assuming t[aed to the nesopaue region and In fact ecitethat this Is the case for our sodium - ueasrements by the longer-period waces. yrn The periods

places the OH-enitting layer at abot 00 km. con- ohserred, one nay estinate oerticai and barizon-firning precious findings of the OH eminsions tai aelngths In the range .O cc and I1-20 ti,

eaking at high altitudes in the polar winter respectively fPhilbrick en i., l 91.,esopause region 9myrab. 19841.

3.3 Evidence for Breaking 3ravity Waves3 2 Possible Gravity Wave Induced Buoyancy2am 5Figures 6 and - show craces of the temperature

and the 39 and sodium intensities durl g theFi gure shows an overlying ripple structure remaining parts of the observing perio. -he

on the I I/l to i-hour waves. "oth 00 the sodium missin g data between i.30 and 2130 UT November 2'

and o8 intensity and on the temperature. The are, as prevously stated. due to coputer failure,aeplicude of the rippie structure in the intensi- while the dta gap in Figure ' between 0000 ondties greatly exceeds the photon noise. htmtpheric i130 UT corresponds to spectra with Fraunhofen

tranceission variations uan also be ruled oat as absorption hands (i.e.. twilight .the source because the background emissions do flring the afternoo. Noneeber 2 there arenot show structures of comparable relative ampli- still quasi-regular wav present in the OHtudes. Figure 5, comparing the intensity vcra- intensity and temleratre. iome correspondencetions of the and O

2 ii_:s, confirms that the with the sodium intensit'" is also seen. -hen In

ripple structures are real, i.e., originate fro the evening and night, large ninultaneous aes

the emissions at the esopcass. integration tine appear both in the O and sodiw , intensities codwas only 2 1/i imates. The hatbavond is also in the temperature. Oscillations In the OPincluded for comparison. Typical periods are intensity of up to 10 of the minimu intensityfound in the range -15 ln, with ampilitdes accompanied by up to 60 K temperatore variation

I0-ill of the tonal in tnsity of the liner. are obserned. .he itecaity of the sodium emls-

limilar variations in the T1 or-s night airglow siuo shows similar behaovio but .ith slightlrhave been reported ecrller (Okuda. IOU ; 'liver-tan, less pronounced variations. After about 350 LTI0h2(. More recently. short-period oscillations. Nove er 28 the wave pactern disappears and isi.e., 5-10 in. have been reported to occur replaced by irregular variations. After thissliultaneousip in both OH lntensity and tepera- time the temperature variations becone smaller,

turn (Takeuchi and Miscec, 10811. However, the but he OH intensity starts to oho periods oflatter ob:1rnation was not clearly superinposed large enhancements (up to 100%) lasting from toon a longer-period eare, and thus night hare a h hoars Some f D. If the gravity aesdifferent origin. Tuan et a. [Iq] have made which were present be f i'si. point start to

an effort to eeplain the ripple oscillation break in the '0-85 km M. on. one would enpect an

onerl~lng the longer-period aes ohierved in the increased edd diffusio -.eInstock, 10'8: '4ein-O0 557Ai data. They find from theoretical uonsi- stock, 1985: thel, S94: VaIterscheid. 10811ferations that the short-period oscillations are The result of an -crewed eddy diffusion inassocted with the buoyany (Brunt-Vaisaial the n0 to OS-km region would be to increase thefreouency of the atmosphere and eocited by the 03 concentration in the 85 to 05-km region

lone period eaves, The sane theoretical consider- Moreels mt cl.. i A' n enhanced 0H emjssionations also apply to the sodium and 0w missions. could result (assuming )3 Is the limiting factor).

52

4. ;t-1, -e, 0 -1 1 -

04'

0 6 4 1 0 12 a '5 20 22 24

Fig. ,,Traces of tb ntensity upper curve) and tmeraturea obta ined from the o

lb-2) band. Measurement period is from 00 UT November 23 to 2130 November 20. lh3.

Further evidence that the gravity wave field is - )Aa 'fsaturated and that the saves teally break duringthis part of the observing period is contained in will give a first approximation. Au is the aver-the power spectrum of the temperature variations, age distanca the 03 or H - olecuies and atoms eedAS indicated in Figure q, on a log/Io

R plot of to nove In the time interval At. Assuming 13 Is

power against frequenry tbe faLioff it tbe porer the limiting factor MreelS at .. 1-'11 andfolloas an

/ 3 lao (N being the frequency). using average oumber densities of 03 versus

This is a strong indicator of saturated gravity height from Noreels & a. tigY!. A is appro-aave fields and breaking waves [Smith et al., . ietely 10-15 km for a '0-lOfT increase In the 031985; euan, lOqN:

0age. io, -. According to density. The 70-100, enhancement in 3I ilntensi-

sore theoretical calculations one should also ties takes place typically over 3-6 bours. Apply-expect a net heating rate from breaking gravity fng these numbers to (10) implies a gravity wavearves of the order of 3-20 K/hr at the mesopause induced vertical eddy diffusion coeffic~ent

0zz

region [bal, iuq4j. Howeer, depending upon in the rooge berween 8 o 1yh and I 10 cm/s.

upproacb and choice of dis.ipatdoe parareter This is a rather large eddy diffusion coefficientthel, iNSA], one night also arrive at a net as compared to average values [Von Zahn and

cooling dalteracheid, iss1i. The rate of 1eig, 10771. ainly quoted around 1l6 c 's. Itheating/cooling is therefore quesfionabie; proba- should be ored that part of the intensity varia-bly bo.th -ircardoc occur. n vie of the clone could be produced by horizntal transport

above indication we find it reasonable to inter- in connection with spatial variation in thepret the variation in intensity and temperature breaking or other atmospheric Irregularities.seen In Figure as effects of breaking gravity Thus our estimate only gives c possible upperaves. Ta places the region of breaking gra- limit of the eddy diffusion coefficie, hoa-ry wavms in the wnter polar niddle atorpheve aver, it is reasonable to believe that eddyslightly higher i.e., - 10 kin) than for the diffusion i

5 highly variable (Weinstock 1q81hO-h5N latitudes jGarcia and Solomon, iqgil, and tat periods aith gravity wave a tivity

The increase in intensity of the 0t esission represent particularly high levels of D -" themight then he used to estiate the upper limit of breaking regive.

enhancemen t of the eddy diffusion coefficient in

connection with the gravity waves (Weinstock. 3.0 OH (6-2) Band and Sodium Night kirglowIN'9. The simplified formula Absolute Intensities

The average intensity of the OH (6-2) band for

PERIOD(MITE) the 3-day obseving period we found to he I .tso 5e kR. This value Is higher hy about 3"-0f than

values nnr.-ly reported for 1w- and ,id-Iatltode

a sttIons [Takabahi and iatista, 1081: IleaStin,to.' IN'S: Takeuthi and Niaawa. 1081. 3 substantlal

0 gn part of this enhancement seems to he related to0o -, the enhanced eddy diffusion In connection withbreaking gravity wayes. Previous results on ON

(3-3, band !ntensity from Longyearbyen have also

atItn-O-

shown average intensities and intensity varatisona ahich ass considerably higher than a id- and

0-a. los-latitude statioms Myrab1 and fleshr, P1841.

Abolute intensities of the OH emissiom foundtn-a tn-a here seem, therefore, to be in good agreement*.EOUEtdah teai with previous findings and cosfir earlier cheer-

Pig. S. htvage power frequency spectmu of the ations of a rather intense OH night air giso in

temperature variation at the meeopaa:e level as the polar cap during winter solstice conditions.obtained from the OH (6-2) roationa* lines The aarags sodium doublet lotensity ic theduring the period lf00 UT Novenbe- .3, tl 130 IT night airgiow derived from the neasurmente a

November 29. 1083. "5 /3

(A - frequency) line fnd to b -, 8. Siniiar intensities have beenio ndicated, reported from high- latitude atations frees et

53

2532 Myrac et aL.: Polar Cap 3ravity waves in Na and OR hirglow

aC., iN'; Henrtksn etl., ,1985 and seer aLeo bten. est Spitebergen ,829 Latitude, L5'

to be typical for Lover-Latitude [Kirhhoff et Longitude, eogrphLc) frem 3205 CT Noveher 2m

ai., 18LbI winter conditions. im- ase both the to LOO LT November 2, 1q3. Spectra covering

rodium number density and sodium night wirglw the P1

linee of the OH band and the '3 and D,

emission normally show Larger rmiatiee variations siees of 9a were obtained with a L/. M and a

than is the case for OH e1issione, it is diffi- tbert spectrophotometer, respoctively. Spectral

cult to say whether the '0 R sodium emission resolution was 7A and 4A for the OH and Na 3.represents an enhanced Level or not. Unfortunate- resolving the P1 and P

5 lines wed the d and S 2

ly, the sodium emission was not teanated during Lines. In the first part of the observing period,

the entire observIng period reported here. regal an wanes sre recorded with perids in the

Whether the sodium emlssion is enhanced by in- range I "l/-2 hours and with overlying ripple

creased eddy diffusion or nor therefore remains structures, periods 5-15 min. Te weves wer:

to be proven by future measurements. present in both the OH and sodium intens ir andin the temperature. strongly indicating passing

3.5 Possible Sources of Graet Wavee in the of gravity waves throagh the emiesion Layers.

Polar Winter esosphere and Lower Thermosphere The tipple structuree coincided in both the 3and

2 lines and appeared in the OH intensity and

Gravity wares are excited In the atmosphere by temperature. There is strong indication that the

a variety of seuhanisms. Penetrative convection, observed buoyant frequency waves (C-15 ein) are

(i.e., when a r!sing current suddenly loses its excited by the larger waves.

buoyancy as it enters a stable region (Towensend, From the donward propegating phase difference

igh8]), frontal acceleration, and oregraphic between the intensity of wanes in the OH andforcing are often qoted as responsible mechanisms. sodium layer, a saxima bight difference between

The most important single source might, however, the centroid of the te emission regions is foundbe wind shears in the lower and middle amosphere to be lees then l km. Assuming an average height

(Ebel. lNis; Einaudi, i9g'N. In polar region. of the Na Layer of 90 is then places the OH laver

speculations of aaroral eoarces launching waves at approximately the eawe height, confirming

and causing temperature and intensity changes of previous findings of the OH emiseion peaking at

the airglow in the 85 to 95-km region have aleo about NO km in the polar sinner atmosphere yvrabt

been put forward Eibel, 1964; laker et i., I S;, ett., L95B.

Takahashi e -aL 1 . The idea that OH eels- From the enbanment of the SN emission, using

sion intensity and temperatures of the 85 to 0, and H profiles from MorseLs eat . [7'1. a

90-k regio shooid be affected by geomagnetic gravity wave induced eddy diffusion coefficient

disturbances was first put forward by Krassovsky In the range nx 06 to 0 n L0

? em2/s was towed.

[19561. Calculations on Joule heating and in- Sean night airglow intensities of the OH (6-S)drag perturbation upon the 90-km region density band and No D lines over the observing period

and temperature during geomagnetic ebbstorms nets 1.9 ir and ,5 , rpectieLy.have, however, given negative resuLts (Irekks. The gravity wave antivity ohserved during thisi97; 4aeda and Aikin, 1968. From earlier period Ioudd not be found to be associated with

observations at Longyearbyen, NyrsbO and leehr geomagnetic activity. Since Krassovsky !19561(19841 showed that negative, positive, and no indicated that upper mesospheric (i.e., 5-:5 km)

correlations could be extracted between OH inten- teeperatarms and night airglon intensities should

slty, temperature, and geomagnetic indices using be modulated by gravity waves launched fromtruncated data sans (i.e,, case studies). A geoagnetic disturbaunces in sauroral regions,

cleat and simple correlation beteen mseopaase theoretinal and esperimental evidence ban appeered,region temperature/ t.perature vrianions, night- both In support of the hypothesis end against it.

glow intensity/intensity variations, and geomag- 40nen caiculetione shnw negligible and unobserveblonetic activity seems therefore doubtful. In view effects (lrekke, 19771 except In very extreme

of the different sour-es of perturbations (gravi- cases. Recently Baker et al. (19.51 reportedty waves from middle and lower atmosphere and positive correlation between &K, and & (OHC tem-

possibly from auroral origin, horizontal and peretures but simulLtaneously also a highly posei-

vertical transport, cheminel reactione) that tire eorrelatlon between a (OH) temperatures and

could act simultaneously, a clear correlanion the variation in 30-mbar temperatures. The

wIth only one of the poesibie sres over a Latter results indicate a connection between 6Kp

longer time interval should perhaps not be en- and the 30-mbar temperature which is even more

petted. The period of observations covered by difficult to explain.the measureents reported here are non a parti-

cularly disturbed geomagnetic period but rather Acknowledgments. Pinascial support for this

quiet. Especially the hours before and during research was provided by National Snienu* Fo-unda-

the morning of November 27 shon no sLgniflcant tine through ATN-83372 to the Geophysical

adroral perturbation capable of penetrenieg Institute of the University of Alaska ed by the

energy down into the ISO-km region. We therefore oyal Norwegian Council for Scient fic and Indus-conclude that t: gravity save activity reported trial Research through a fe llowship grant to one

in this paper has its origin In the lower or of us (NK/t). The obeervations on SvaLbard aremiddle atmosphere and not in the thermosphere in made through a cooperative effort involvingconnection with geomagnetic disturbances, the universities in Fairbeeks, Troms4 end Oslo,

with the help and cooperanion of the Great orwe-Summary giwe Spit-ergem Goal GOmpaer end tha Norwegian

Polar Institute. The authors elso want to thank

TheOH (6-2) bend and the Na D lines in the Dan Osborne and Bob Erinkson foer tchnical assis-

night airgow have been observed from Longyear- tance and programming ssistance.

54

tivrab6 et al.: Polar lap ^ravity We-e in No and 3H0 Airgiow 20

The cdto hinks V. W. J. M. Kirchhoff and aurora. J. Atmos. Tarr. ?Sys.,._ ', ',V.C. Rid for their assistance in evaluating this Juramy, P.. .5 Chenn,' 2.l' i

paper. Toulinov, and Y. P. Ooudoiadcc. LIdac sondingof mesospheri c sodium layer at nigh atltude,

Reference. J2. Atane. , err. Phys. . -3. 209. 1911.

Kirchhoff. V. W1. J. H.. B. R. Cleneeha. and D.Baker, 1. 1., A. P.Sed, G. A. Were, 11. Offer- S Ino h The atnoopheric neatra 1 odiumnen . LAng and H .L e, Ground-b .se.d ler, I, Rect modelling compared to neasure -

at ::phenin inicrerd and visible eission ...nts, J2. G.ophy.. R... L6, 6892, 109.

.:ero.tm nt. 12. A teas. Tern. Ph3., 43. 133. Kirchhoff. V. W. -2 J.. .R l ea n

1985. D. '1. SiLmnich,- Avrg noctra and 'e-rek.A., Autral effects an neutral doen, sn eitoso oimngtlwa 7S

fro Oynamical and chemical coupling hetween 46N. Planet, Space id., 29, 6h5. .981b.

the neta n ote rishne..B rassoveky, V. I., Infrared night airgloc an aCnandal and J. A. Hoitet, D. PRidel. Holland, manifesta ti on of the p roteea of oxvfennm'e"-p. 315, 1977. hioation, In Airglno and Aurora. edited Sy

Chanin. M. L.,* Review of ldor contributions to E. B. Armtrong and A. Dalgarno, p. '103, Per-the description and understanding of middle gamom, Ne. York. 1956.a tmoeSphere, .2. Atmoe. Tant. Phys., . 6, 987, Krasnovehy, V. t., Infrasonic variations of OH1 984. emioion In the upper atmosphere, Ann.

Deehr, C. S, G. V. Sivje, A. Egel1nd. K. Henrik- Veophy.. . 28, '39, 1972.eaP. . Seodhoslt. R. Smitch.*P. wee. ie .' HI Nihtglow, obae - tierio at In doringC. n_ and F. Vilmer. Vroumd-baeed"obeervs the 1. G. Y.. Veophys. Pobi. , 20, 1. iSSO.

tlone Iof F:-r'Simn assnciated with the magneto- LindaeR . Turhulenc. and Otto.. oing tospheri cu. p . Gacphys. Re... 85. 2135. grviy ev and tidal breakdown, .2. Veophvs.

1980. R..86. 77 01.erh Io-m H., Sheervatione of sodium emissions in Lien~eLlyn, E. E_ I .8 Long, and B H. Solhe-Inaurora. .JAtm.. h - 26.1 '9t71, 1964. Tb. quenc.hin g of 080 in the atmosphere, Plane.,

Diean, E..spNE-G e teer: r.eem- Sp.ce Sci. . 26. 52:. 08ling turbulence, Science. 206., 932. 1979. Meeda. K., and A.C. bum,. Voriatlono of polar

Eb. A. CUon tribu tion, of gravity uavee to the neoopheric oons during events, Planet. pc.....co, beat aod torbolent energy budget Sdi., 16. 371, 1968.

o f the upper me soephere ad loner thormoephere. Mia..F. H., Calculated vibrational transition1. At.... Tart.r b.. 46, 727. 1984. probabilities of OH(X

2n1, j..1. spectrOe,

iE~inaF,11 P, iio and G. E:. eona, The.3 150, 1974.role of grav ity vivm. In tropospheric pro- lohenakunar, K., An investigation on the infiounce

_teee Aur Ap. G.ophye.., 117, 627, 1979. of solar cycle ont mesospheric tenperature,

Gage, K. S., E Evidence for a K-/5 '

low internal Planet.pa .S, 33, 795, 19851.range in nsole Oe-imnionel turbulence. MoreeTiVZZ: U. le H. flac-oeand0416tos Sot., - 16 1950, 1 979. 8. L. Gattinger, An oxygen-hydnoe atmohei

Garcia, R. R. and S. Solosm, Tha effect of break- model and its application to th OH emlelninag gravity wr.ethdyacand chemical prbe... AtMOS. Tarr. Phy.., 3q. 5 51., 1 0.

coepeiton f t a he re and loner thermo- Mab6, H. K..' Temperature variarlon at neecycuonephae. ~ pbe.es. 9, 3850, 1935. Ievel during winternoecea S.

Vault. , A4B. AK lr. IF Link, and V. V. Plaet SacSr. 32. 249, 1984.Shepherd, Ssorvaetione of the optical spectrum Nyrab; an P I.S Deehr, lid-winter Sodroc-of the daygide eagoetospheric cleft aurora, yl night Airglaw emission intenhities in the

Plne. pce .Sd , 9, 321. 19t1. oorthern polar region. Planet. Space Sdi. 3Haniken K *V.V.Sivjim. and C, S. Deebr, 63, 1984.

W inter en oceent oatmclithium and Myrkb6. H. K., C. S. Doeebr, and V. V. Sivjee,nodium inheh pelaroupper atmOphoto, J. Large-amplitude olghtglon OH (8-3) band intec-Veophys. Rom. . , 51 5g3. 1980. sity end rotational temperature variations

Here M.. G. . Horeals, and .J. Cieiremsidi. Wares during a 24-boor period at 78*4. J. Vsophvs,

ithe OH emiseion layer: Photogremmatry and Rae.. 88. 925 5, 1983.topography. Appl. Opt..* 19, 355, 1980. 4y rab7, 4. K,. C. S. Deshr. V. J2. Rnvlck, end

Hinee. C. S0 Intrn1 atmospheric gravity wave. K. Heneriksen. Mid-ainter Intensities of theatoosherchihs!a. 2 b' t8 ngtarlw~ (0-1) atmoepheric Send

1441, 1960. eciesiOn at hi gh Latitudes, .2. Voophy,. Ree

Hirot.I., Clmtlg o rvt wref h 91, 1684 , 1986.,middle atophee . ~.Tr n P hyn , Soxon, S. F. I Effec ts of internal gravit' ave46. 767,t1984. upon night &airginer temperatures, Veophys. R.

Holton, .1. B., Am Introduction to Ilyrumic*1 Lett.. 5, 25, 1 978.Meteorology, AcatdeMic. 'Orlando, Fla. . 1979. Okuda. w. , A, study of excitation proceses Im

Holton, F. R. * The inf luence o f grav'ity wav night airgiow. Sri. Rep. Tohoke niol., 14,

br:akiog on the Senaeal crculation o f the. 9 , 1962.middle a phere. 3. Atans. Sci., 40, 2497. Pibriok, C. B., F. J2. Schmidlin, K. V. Gtone-

19nt83. mann., G. Long.. DI. Offervant,,K.,0..Neker,Hut" D. 8., Some photometric observations of D. Krankovasky, and V. von Z&ho, .enI ty naurorel specra. .. dAinO.. Tarr. Phys.., 7, temperature structure oven northern Europe,t14. 1955. J. Arens. Tern. Phye. . 47. 19. 1q85

Hontsn, D9 M. * Excitation of sodium eckeelon in Raes. K. H. G . J. Bomick, and AL. E. Belon. The

55

OS~o Myrah at at.: Pot.iap rciy t - tare i'O and DR Alrglon

intnity of the sodium t ines to the of the Oh '6-2) hand. Ann. 3moptvs., .3- 315.IeAnn. Ceophys.., 31, 311 . 1975. 1981.

spectru oftearr an argi. Pro. tem 'f grat y vasostcurn ha .hoc. Photo. Opt. Eero . g.., 91, 63. at thencesopass tee hr::lb photogrerntr and

1976. tidar. Pltanat. Space. Sit., 30. 6ti, 1082Ohagayen I.V, at a ion of hydranyl Toe.nsend. . A., fact tatoo of ternal anes to

ait airoe saission.' . Asres. Tart, a o ethi-ettatified atmosph:rm aith o-ePhy., 36. 367, 1974. ablm nind shea. 2.j FluidMach..5.15

Shelton. F. D. , C. S. Gardner. and C. F. Sechrist. 1961.ODesty res ponse of th beotph:ri ua .7.F.:. A. Adinger. S. a. Silnernaa. and

eodiu taper to g ravity oar ruhtione. . Ok da .Ongraity oar induced rn-asaleophys. Ass. Lett., ' , 1060. 1980. oscittations, J. Ceophys. Ass., L'. 3n', Igan.

Oltvrnt, S. 4 ., U.nusual fluc tuations of 55"! al Vattance-lone A.. Aurnra. dinghav. Nave.,

uuua ~ ~ ~ 0ro- soutnlovat8n cna -'n D. Reidet. 1073.1961, Oviore. 195, 48i, 1962. Von Zuan. V.. and T. dernig. Inert gas ahundance

Stols. 0 2. andC. S Sahr, :6ff r ans lv as Ii indcatrs ~for thestcretngth of eddy diffue-poa amspeicotia eisin hteea te tutr hopueattitudes. tn Dynamical and

,id-da, ycap a nd ni ght-tie auroras, to Cheminal Coapling Bet-es the 4.autral andtoptortion of the Polar i!per Atenar.nisdAmsphere, edied y.8. radelanedited hy C. S. Seehlarad 2..A. FIaltet, p. .i. u inta t pr., edted A y.idel, d nig a

199. ~ ~ ~ ~ t 9, AU.el Ringitam Mae,195.Mae. 177

hin]ee. i .. 'A . .Ic,'and P.11 5. Fema, Wal'tersohaid. 8. L. , Dynaical cooltng induced hyTemporal variati ons in the nighttime hydreryt dissipating internal gravity vans. CeOphys.rotational teepavaturs,. Pianet. Spoace ti. Res. Lott. . 8. 13235, 1981 .

25 61, 1972.C aetnenoak, J.. Thenry of t~e interaction ofSeith, S. A., S. C.Frints, and T. E. Van Zndt. grolity vane nI'ith' D~37 irgos-, J. ;sophys.

Compari saof :soaph-i. oitd spec tr voitcA a e As. 3. 517, 1 978.grvtyvn- nd1 Radio 3di. .23)6), 1331- -e;tc.JGaiyaestrt: bdd1338, 1985. diffusion in the naddie atnep hece 3 tnTakeash, .. ant:d P.HP.Ba ticta, Sisut taneus Terr Pbys.. 46. 1069. 1nn5.

neareea ef 56( 9-43, 38-3). ('-Z) 6-)avo (5-1) hands in the airgtom w. J.vophym. _____ye-, 86, 5633, t981. C. S. Dsshr, 9. K. tyrahd. and A. Vieri,

Tachashi, 9. , P. P. Batieto. Y. iShael. and Geohysital Institote, Unriverat Of Alaska,R.A Cleha A shcri:.save prop-aaton Pairhanks. AX 99775-5800.

In the vas pas rgon aherad hy the OR K. Henriksen, Auroral lOher torn rT-96,(3-3 ) hand, 4. 0, 32 6 (8 6433' -ad and 01 Nevaay.

5577A nightgilo, emissions . Planet. spatsSal. . 3'1. 1985. (9eceinsd Ooreehe 21. 1987:

akchI.ad. Meaa Shari-perlod o:scilla- revised Ictohet 3, 1986:tions of .isc y ad ro tattianel atepar'tor accepted Octnher 21, 1986.3

-. 47

At

......... :

IL

57

S" iS- S"i %Io iA '0 11 pp 10t i02Y ty 1032 .wii 86 S; 0-iJXPried i Ort BnGam , 46 Penmon oorev Lid

WINTER-SEASON MESOPAUSE AND LOWER THERMOSPHERETEMPERATURES IN THE NORTHERN POLAR REGION

H. K. WnRABO*

Geophysical Institute. University of Alaska. Fairbanks. AK 99754-800. U S A

iRecetred in itnal torm 12 W4at 19861

Abstact-Mesopause lower thermosphere nocturnal temperatures have been deduced fromspectrophotometnc observations of the night airglow OH emissions at Longyearbyen. Svalbard 78 Ni.during four winter seasons 11980-1985). A monthly average temperature maximum of 223 K is found tooccur in January with monthly averages of 206. 212.212 and 198 K respectively for November. December.February and March. A relatively low temperature in late December followed by a very warm mesopause inearly January seems to be consistent for all four seasons and n-ight be associated with changes intransmission of gravity waves to the upper mesosphere in connection with stratosphen and lowermesosphenc circulation changes.

I. INTRODLC'TON method of obtaining temperatures of the 80-100 kmaltitude is to extract temperatures from the measured

Ternperature and temperature vanations are basic rotational intensity distribution of the different nightparameters in the understanding of atmospheric airglow emissions, such as the OH bands and the 0.dynamics, chemistry, circulation and energy flux, and (0-1) Atmospheric bands Myrabo. 1984). Tempera-in the overall interaction between the upper-middle tures obtained from these emissions yield neutral airand lower atmosphere. The 70-100 km region of the temperatures at the emitting height because theatmosphere has been found to hold important keys to emitting molecules are in thermodynamic equilibriumthe understanding of local and global dynamic.s of the wtth the surrounding atmosphere Krassovsky er al..middle atmosphere IHolton. 1983; Frits etal., 1984: 1977). There is no OH emission height profile

- Myrabo et al., 1985). In spite of this. measurements of published for latitudes 75-0 N that could be trulythe temperature in this part of the middle atmosphere representative of the measurements in this paper. Thehave been sparse. This is mainly due to the relative northernmost obtained profile Ia = 68 N: by Witt etinaccessibility of this region of the atmosphere to al., 1979). however, shows an extremely large halfdifferent temperature measuring methods Myrabo. width (15 km) believed not to be typical for the OH1984: Barnett. 19801. layer in polar regions (Myrabo et al.. 19861. Therefore.

In the polar cap, measurements of the mesopause to illustrate the position and shape of the OH layer inand lower thermospheric temperatures and their the atmosphere, the emission height profile by Rogersvariations are even more sparse. Besides rocketsonde e al. 11973) has been plotted in Fig. I together with themeasurements from Heiss Island (81 N). mainly CIRA 1972. 70 N. December, model atmosphericgiving a few snapshots in time and space, the only temperature profile. A half width in the range 5-10 kmmeasurements on a regular basis covenng a longer is typical for the emission, and the peak heighttime period are those reported by Myrabo (1984). Due normally ranges from 52 to 90 km. probably closer .oto considerable short time-scale variations caused by 90 km for winter solstice polar conditions Myrabo ergravity waves, local transport phenomena, and al., 1983; Myrabo. 1984). Also included in Fig. Its thestratosphetir- and lower mesospheric circulation weighting function swith height of the NO.4.4 7. Ch. 27changes (i.e. stratwarms) (Myrabe et al.. 1983 spacecraft radiometer and the CIRA 1972, 70N,Myrabe et al.. 1984). temperatures must be measured December. model atmospheric temperature profile.frequently (covering at least 5-8 h) to give reliable The Ch. 27 radiometer reading is proportional todaily means. Thus, ground-based methods are the temperature and was used to monitor the loweronly means of obtaining this temporal coverage, and mesospheric temperatures and circulation changesresolving dynamical phenomena. over the northern polar cap (Labitzke e al.. 19851. It

In the polar regions, the present ground-based has been shown to anticorrelate with larger

*On leave from the Norwegian Defence Research temperature changes in the 80- 100 km region i MyraboEstablishment. N-2007. Kjeller. Norway et al.. 1984).

1023

58

1024 H K MRAB0

TemOerature K; tempcrature range of 160 to approximatel% 300 K The210 230 250 measurements by Myrabo 11984) yielded a mean

winter solstice r.e December and Januarvi901" OH-Profe - temperature of 220 K for the 1982 83 season The

- purpose of this paper is to extend and generalize these-a -0 results by including mesopause temperatures obtained

- ~from OH observations during the 1983 84 andS 70- 1984 85 winter season at Longyearbyen. 79 N.

y 60,-x 2. ORERv ATIOS AND DAtA REDLCTION

501- - As part of the 1983 84 and 1984 85 Multi-National-- OAAI - Svalbard Auroral Expedition (Deehr et al, 19801,

40 Ch 27 We,grtng spectra of the zenith sky were taken close to-funcdom Longyearbyen on West Spitsbergen (78 N. Lat, 15E.

30,- -. Long.. geographicl. Both winters, measurements

0 0 ocovered a 3-4 month period around solstice. EmissionRelativeamostude spectra were obtained using a 1 2 M Ebert Fastie

FIG. I. EXAMPLE OF A TYPICAL OH EMISSION HEIGHT PROFILE spectrophotometer coupled to a minicomputer and

AS OBTAINED FROM ROCKET MEASLREMENTS (ROGERS t al.. recording in the photon counting mode. The

1973). instrument is further described by Sivjee et al. 119121The weighting function with height of the NO.4A 7. Ch. 27 and Myrabo et al. (1985). The spectral region fromspacecraft radiometer together with the CIRA 1972. 70 N. 8240 to 8740.A was scanned in 12 s using theDecember. model atmospheric temperature idashed line spectrometer in the first order with a 0 5 m'- slit

profile are also indicated, corresponding to a spectral resolution of 7 4. Each

scan was recorded on magnetic tape Indiv'dualIn adidtion to the measurements reported by rotational temperatures were calculated from I and 2-

Myrabo (1984) from Longyearbyen 178 N). using h integrated scans (300 and 600 scans. respectively)night airglow OH emissions, earlier sporadic using the OH 16-2) band An example of a typical I-h

measurements have been reported by Chamberlain integrated spectrum used to obtain the temperature isand Oliver (1953). Noxon 11964) and Sivjee et al. given in Fig. 2. The P, and P. lines of the OH 16-21)11972) at latitudes 70-85 N and covering a band are indicated together with the R and Q

140 " OH 16-2)

R P 12) f3T P14) P5, P(6)

30I_

-20l-

0 8300 8350 8400 8450 8500 8550

Waevelength (A,

FIG. 2. A rYPICAL i-h INTEGRATED SPECTRUM USED TO OBTAIN THE TEMPERATU REIn addition to the different lines and branches of the OH 16-21 hand, th aurora emission from atomic

oxygen at 8446 A is also indicated

59

Mesopause and lower thermosphere temperatures in polar region IQ2_

Day number

315 335 355 010 030 -350 )70

240 -

. 2201,-

2001-

5 15 25 5 15 25 10 20 30 10 20 0Noflmbor ecmombar January February March

FIG. 3 DAILY HEAN FE5mPERA IURE FOR THE WINTER SEASON 1983 84 ASOBTAINFD FROM OH (6-21 BAND NIGHTAIRGLOW EMISSIONS NEAR LONGYEARBYE. 78 N

Straight lines are drawn between the daily means. Temperatures are given in degrLes KdCin KI.

branches. Also shown is the auroral permitted the field using a standard lamp. The averageemission at 8446 A from atomic oxygen. The auroral uncertainty in an individual temperature as denvedemission is excited from particle bombardment ii.e. from a 1- or 2-h integrated spectrum is Z_- 3 K. mainlymainly electrons) IVallence-Jones. 1974 and may be due to uncertainty in defining the backgroundused as an indicator of aurora] activity. Using spectra continuum.with a relatively high resolution as shown in Fig. 2.and containing an auroral reference emission. makes itpossible to eliminate spectra with auroral emission 3, RESLLTSfeatures which may contaminate the OH P, linemeasurements. In the polar cap, auroral emissions Nocturnal mean temperatures were obtaaed byfrom molecular species are less frequent than in the averaging five or more of the 1- or 2-h individualauroral oval, due to a lower energy in the precipitating temperatures. The individual temperatures wereparticles ISivjee and Deehr. 1980). Thus, only spectra normally scattered throughout a 20-24 h penod. andvoid ofauroral contamination ofthe OH P, lines were on the average covered 14 h a night. Although theused to obtain temperatures length of the night possible for OH temperature

The temperatures have been calculated from the determination. i.e. ;_ > 100., lasts only for about 9 hintensity distribution of the P,(2l. PO3, P,(41 and in early March. the correction due to the semidiurnalP, (5) lines assuming a Boltzmann distribution of the tide on the nightly average dung February-March asmultiplet rotational levels, i.e. compared to winter solstice is found to be less than I K

iMyrabo. 1984). The nocturnal averages may thusr'= ,4 , " safel be taken as the average for a 12 h pervod around

2j2J'. +v) -) E. i' socalymidnight over the entire perod of obsero ns(.e November-March). Because no diurnal sanailon

. "" T of temperature amplitude obtained from OH largerwhere I is the photon intensity in photons s - cm . than I K is found at Svalbard (Myrabo. 19841. theQ.(T,,) the electronic-rotational partition function nocturnal averages are equivalent to the dailyfor the v' level, N',. the total concentraton of molecules averages. Thus, no corrections have been made for thein the r' vibrational level. T, the derived rotational semidiurnal or diurnal varation in the temperaturetemperature and E., IV) the upper state term value for iMyrabo. 1984). We therefore use the expression dailythe vibrational band C - C. The other variables have averages for the average temperature obtained withintheir usual meaning. The Einstein transition each 24-h interval. The resulting daily averages for theprobabilities, 4. given by Mies 11974). have been used 1983 84 winter season are given in Fig. 3. A straighttogether with energy levels. E IJ') by Kvifte 11959). lne is drawn between the averages, marked with openAbsolute calibration ofthespectra were carried out in circles. In addition to the day and month, the day

60

026 H K. MYRABO

Day number315 335 355 010 030 050

'110240,-

-100

2201- 10 O ~-90ui' A- Ch 27 - P17a

a OnAA7I-80

- '~; -70

1501- -60

15 25 5 15 25 10 20 30 10 20

November December Januar February

FIG. 4. DAILY MEAN TEMPERATURES FROM THE WYINTER SEASON 1984.85 AS OBTAINED FROM OH 16-21 BAND

NIGHT AIRGLOW EMISSION- NEAR LONGYEARB--E. 7, N

The radiance level (in oWI from Ch. 27 on the NO.-A 7"radiometer integrated over the northern polar region

is also indicated istippled linci.

number is also marked along the top ofthe honzonta deflection of the radiometer, a stratospheric warmingscale of the figure. The temperature is given in degrees occurred near the end of December.

Kelvin (K). In January and the last part of February. Large variations in the temperature may occurthe 1 2 m spectrometer was used to obtain auroral within each daily average. An example ofthis ts shownspectra, resulting in the data gaps seen in the figure. in Fig. 5. depicting the temperature from hour to hourThe temperatures for the 1984,85 winter season are during pan of the very intense warming of the lowershown in Fig. 4. Figure 4 also includes the radiance mesosphere and stratosphere in late December iseelevel from Ch. 27 on the NOAA 7 radiometer isee Fig. Fig. 4). An average trend drawn through the dailyl). as integrated over the norther polar region during means (marked with filled tranglesi is indicated.the first and mid-part of the winter As seen from the Due to the large day-to-day and week-to-week

Day numbe,

359 360 361

220-

S2004 F \

i r -

E 180'- .

170'-

160,

12 16 20 04 08 12 16 20 04 08 12 16 20

25 Decemier 26 De:ember 27 December

Urryrsal ,me h)

FIG 5 ONE-HOUR INTEGRATED TEMPERATURES FROM THE OH (6-2) SAND NtCACr IRGLOW EMISSION D'RINGPART OF THE INTENSE STRATOSPHERIC WARMitG AT THE END OF 1984

The dady means are marked with open circles. An average trend second degree curve is fitted to the dadymeans istippled line).

61

rIesopausc and lower thermosphere temperatures in polar region 102

Day number

240 315 335 355 310 030 050 070

230-

220- , N

1i210, '154- A +\

20o- -

5 15 25 5 5 20 10 20 '0Novern~er December Jauary February Marc,,

FIG, 6. TEN DAYS RUNNING AVERAGE OF THE DAILY MEAN TE.MPERATLRES IMARKED WITH CIRCLES) FOR THE

80/8. 82,83. 83 84 AND 84.85 WINTER SEASON AT 18 NThe monthly means are also plotted icrosses) together with the monthly means for the CIRA 1972. "0 N, Mkin. model atmosphere itnangles). A second degree best tit curve is drawn through the means, both for the

model atmosphere isolid line) and for the actual monthly mean temperatures from these measurements

Istippled curbel.

variations in the temperature, data for both seasons atmosphere (solid (ine) and for the actual monthyhave been added to investigate any seasonal trend and means of these measurements istippled curve.for comparison with model atmospheres The datlytemperatures for the 1980,81 and 1982 83 season asobtained by Myrabe (1984) have also been included.The 1980 81 and 82.83 temperatures were obtained From Figs 3 and 4 it is seen that the day-to-dayusing the OH (8-31 band and slightly different temperatures show large peak-like variations iastingmolecular constants (Myrabo, 19841. A temperature for a few days to a week or two. This was also seen inof 2K was subtracted from the OH (8-3) band the 1980,81 and 82,83 seasons (Myrabo. 19841 Atemperatures, because the OH (8-3) band emits similar pattern was also observed for the 01 5577 Amainly from a level 1-2 km higher in the atmosphere ntght airglow intensity variations from Thule (76 N)than the OH (6-21 band (Charters et al., 1971; IMullen et al.. 1971). As seen in Fig. 4. the relativelyMyrabe, 1984; Hamwey. 19851. The temperature large and consistent cooling of the OH (6-21gradient obtained by Myrabo (1984) was assumed, temperature at the mesopause is accompanied by aFollowing this adjustment of the 80r81 and 82.,83 warmingandcirculationchangeILabtzkeeral, 1985)temperatures, the temperatures were corrected for the of the lower mesosphere and stratosphere This wasdifferences in molecular constant used. Corrections also the case durng the stratospheric warming aroundwere normally in the range 3-6 K. to be subtracted new year in the 82.83 season Mvrabo er al . 1984).from the 80,81 and 82.83 temperatures Thus. on Ismail and Cogger (19821 have also associated largeaverage, the 8081 and 82,83 temperatures were week-to-week changes of the O 557' A night airglowlowered by 5-6 K to make the data sets compatible, from Thule with minor and major stratosphericThe maximum error introduced in the average daily warmngs. Thus. it seems established that thesemeans for the four seasons by employing this apparently irregular large scale temperature changestranstormation procedure to the 80,81 and 82.,83 taking place over days and weeks are mainlytemperatures is estimated to be less than 2 K. connected to effects of the changes in the circulation

Figure 6 shows the 10 day running average of the pattern and temperatures in the lower mesosphere anddailyimeantemperaturesusingthetemperaturesforall stratosphere. They seem also to be present dunngfour seasons. The 10 day running averages are taken at seasons with no major stratespheric warmings5 day intervals. i.e. 50'. overlap. The monthly means From Fig. 4 it may also be seen that the coldare also plotted (crosses) together with the monthly mesopause regions occur before the lower mesospheremeans for :he CIRA. 1972. 70 N, 90 km model and stratosphere start to warm. Simultaneousatmosphere (filled triangles). A second degree best fit unpublished me-opause temperatures from Fairbankscurve is drawn through the means, both for the model (65 Ni during the 1984 85 stratospheric warming

62

028 H K MIRABO

seem to indicate the same pattern. i.e. a cooling of the and in the circulation, temperature or transport in themesopause is seen befoe the heating of the underlying atmosphere A further consequence ot thisstratosphere take place (Viereck et al . 19861. seems to be that. on 'he average. mesopause

Transmission of gravity waves into the mesosphere temperatures in the northern polar cap reach ais controlled by the propagation in and the interaction maximum in January. and not in December aswith the environment. Phenomena as refraction, predicted by the CIRA 1972 model atmospherereflection and critical-level absorption due to The peculianty that the mesopause region seems tovariation in mean zonal wind uil)and the square of the warm up dramatically just after the coolingBrunt-Vaissala frequency .N) with height may cause accompanied by the straiwarms. and somewhatmajor changes in the waves with height iselective compensates for the cold penod. causing an extremelylItertngl. Spatial inhomogeneities in the trans- warm mesopause in January. cannot be explained vetmissavity and or in the gravity-wave sources are However. investigations for the coming years arethought to be capable of exciting large scale gravity planned which may provide an answer to this.waves or planetary waves (Frtts et al.. 1984). Suchlarge scale excited waves may be important in wave-wave interaction in connection with stratospheric .SUVI\ARwarmings (Smith. 1985).

Gravity waves may thus be important both in the Mesopause lower thermosphere temperatures havewarming of the stratosphere and in the cooling of the been deduced from spectrophotometric observationsupper mesosphere. of the night airglow OH emissions close to

If the winter mesosphere and mesopause region is Longyearbyen on West Spitsbergen (78 N, Lat.. 15 E.kept warm largely from energy deposited by breaking Long.. geographic) dunng the 80 81.82 83.83 8

,, andgravity waves onginating in the lower atmosphere. a 84 85 winter seasons. A 1 2 M Ebert Fastie

slight change in the circulation pattern of the spectrophotometer has been used to obtain theunderlying atmosphere would affect the propagation emission spectra with a spectral resolution of 7,A.

of these waves (Dunkerton and Butchart. 1984) and resolving the P- and P,-lnes of the OH 16-2) and 18-the heating rate. Thus. Lmdzen (1981) and Holton 3) bands. A monthly average temperature maximum(1983) have suggested that appearance of easterlies of 223 K is found to occur in January with monthlywill inhibit propagation of quasi-stationary gravity averages of 206. 212. 212 and 198 K. respectively forwaves to the mesosphere. This would allow the November. December. February and March. Themesopause and upper mesosphenc region to undergo January monthly average is about 10 K higher thanradiative cooling. A 10-20 K per day cooling rate the respective CIRA 1972 model atmospherecould occur lChamberlain. 1978: Ebel. 1984). which is temperature at 90 km. 70 N. A relatively lowin the range of actual average cooling seen. It is also temperature in late December followed by a warmconsistent with the observation of the cooling taking mesopause in early January is seen to be consistent forplace before the heating is observed in the lower part of all four seasons. In the 10-day running averagethe atmosphere (Holton. 19831. temperature this appears as. a 20-30 K amplitude

The 10 day running average temperature in Fig 6 temperature wave around years end. This "wave"(average for all four seasons) also shows a peculiar might be associated with circulation changes i.e.drop in the temperature during the end of December stratwarms in the lower mesosphere andfollowed by a very warm period in January. The stratosphere. and physically related to the changes instandard deviation in the I0 days running average is in the occurrence of breaking gravity waves at thethe range 2-5 K, which means that the December mesopause caused by interaction with the general flowJanuary dip is real and has a total amplitude of 20-30 in the underlying atmosphere Large hour-to-hourK. A closer look at the data for each year shows the variations in the temperature are regularly seen.tendency for a meopause cooling to occur around indicating the presence of gravity waves HourlyNew Year each year. even during 1983 84 which was temperatures as large as 283 K and as low as I58 K

not a year with a major statospheric warming have been observed.

iLabitzke et al.. 19851. No auroral effects seem to have Acknowledqenenrs -Financial support for this research wasa seasonal pattern that could account for part of the provided by National Science Foundation through GrantDecember January "wave" in the mesopause and ATM-8313727 to the Geophysical Institute ofthe tniversity

lower thermosphere temperature. Thus. we may of Alaska and by the Royal Norwegian Council for Scientificand Industrial Re.carch through a fellowship grant The

conclude that this "wave" seems to be entirely author also wants to thank Mr Rod Viereck. jim Baldndgeassociated with changes in the gravity wave actlvity and Dan Osborne for programming and technical assistance

63

Mesoipause and lower thermosphere temperatures in polar region 1029

REFERENCES Lindzen, R. S (19811 Turbulence and stress due to gravity

Barnett. J I 1 980i Satellite nmeasurenments of middle wave and tidal breakdown. J qeophvs Res. 86. 9707atnmosphere temperature structure Phil Tranm. R Soc, Mies. F H 119741 Calculated vibrational transition

L~on 2%A 41probabilities of OH (X- , i J nailer Spetros 53. 1%0Lond 296ai.J 41 98~e~otPait Amshrs Mullen. E G., Silverman. S. M and Korf. D) F 119771inntrodtn Wo 17 thery ofvw Pandetay rv At p 32.s Nightglow (5577" nes of O1) in the central polar capAa ndemcto toss tNeiYrk Phsc n lens.p3. Planet. Space Sct 25. 23.Cadbemli. Prs. W YndOlrk. NIi9 Hi h Myraho. H K. 119841 Temperature variation at mesopause

Chaberain I.W ad Oive. N i.1951 O inthe levels duringwinter solstice at 78 N Planet Space Sct 32.airglow at high latitudes. Phi's. Ret . 90. 1118. 249.

Charters. P E.. Macdonald, R. G. and Polanyt. J. C. 11971) M yrabn. H. K..Deehr. C. S and Lyhekk. B. 11984) Polar cap4ppi Optics 10. 1747.

CIRA 1972 11972) COSP.4R working group 4. C0 4 OH airglow rotational temperatures at the rnesopausIntenatonalRefrenc Atospere.Perarno Prss. during a stratospheric warming event. Planet Space Si.

Onterainld. rnc 4mshr. egmnPes 32.853D Oxford. S.Sve.GG. gln.A.H nk Myrabo. H K.. Deehr. C. S , Romick. G I and Hetnksen.

Deeh, CS. ivje. G G. Eglan .He nn K.. K. 11986) Mid-winter intensities of the night uirglow 0.Sandholt. P E.. Soith. R.. Sweeney. P., Duncan. C. and (0-1l atmospheric hand emission at high latitudes. JG3ilmer. F.1 11980) Ground-based observationss of F-region goh'.Rs 1 64associated with the magnetospheric cusp. J. gee phys. Res. ery.Rs91164

85, 2185. Mytabe, H. K., Deehr. C S and Sivjee. G. G. t 19831 Large.Dunkerton. T. J. and Butchart. N. (1984) Propagation ad amplitude nightglow OH (8-31 band intensity andad rotational temperature variations during a 24-h period atselective transnrusaion of internal gravity waves in asudden 78-NJ. geephi's. Res. SIL 9255

warming. J. arises. Sci. 41. 1443. Notion. J. F 119641 The latitude dependence of OHEbel. A. 1 1984) Contibution of gravity waves to the rotinlemetuenthngtaigwJgepy.

momentum, heat and turbulent energy budget of the upper retti. a 69. eatr 4087 ih arlo eohs

mesosphere and lower thermosphere. J. arnsos. ter-, Phys. Res. 69. 408 upy.R7.Sai.. Uik

Fr6t. 727 .Gle.M ABlly .B. hnn . Baker. K. D and Jensen, L. L. 11973) Rocket-homneFrills. D LC. Gelter. M. A.. ato, B.S.. Chin. Mn R.. radiometric measurements of OH in the auroral zone. J

Hirta.1..Holon.3 R. KteS..Linzen R.S.. geephys. Res. 78. 7023Schoeberl. M. .Vincent. R. A. and Woodman. R. F ije GanDeh.CS.198 ifrncsnpor11984) Research status and recommendation from the Sve.G0adDer .S 911Dfeecsi oaAlaska workshop on gravity waves and turbulence in the atmospheric optical emissions betwemn mid-day cusp andmiddle atmosphere. Fairbanks. Alaska. 18-22 July. 1983. nght-time auroras, in Expedition of~ the Polar C pperBall. Am. Wet. Soc. 65. 149 Atmosphere iEdited by Deehr. C. S. and Holtet. J A i.

Hamwsey. R. 11985) Spectroscopy of the night airglow OH p. 199. Reidel, Dordrecht.M. . Tess. eopyscalInsitte.Alaka Siyjee. G. G.. Dick. K. A. and Feldman. P D 1 1972)emissions- .S.Tei.GohsclIsiu.Aak. Temporal variations in the night-rune hydroxyl rotationalHolton, J. R. 11983) The influence of gravity waves breaking tmeaue lntSaeSt2.21

on the general circulation of the middle atmosphere. J. Smth.ertu A. ne K.1 S85 ae tasciec and wae6 ea1oarms. Sci. 40, 2497 tAK.i18 Wvtrninead s-mnfl,

Ismal. . an Coger L.L 1182) empralinteraction caused by the interferee of stationary andIsmail.~~~~~~~~ S.adCge.LL 181Tmoa aitions of traveling waves. J. aries. Set. 42. 529pola ca 01577 airlowPlanr S a SI 3.85 Vallence-Jones. A. i 1974) lAurora. Reidel. Dordrecht.

Krassovsky. V I., Potapos. B. P.. Semenov. A. L.. Sobolev. Viereck. R.. Myrabe. H.- K.. Deehr. C S. and Romick, G IV G. Sagav. . V an Shlin, N N 197) O th (19861 Lower thermosphere and mesopause regionequilibrium nature of rotational temperature of hydroxyl temperatures at Fairbanks 165 N) and Longyenrhyen

airglow. Planet. Space Set. 25. 596. 078N) during the stratospheric warming event around IKvifie. G. 11959) Nightglow observations at As during the an.195ipraatn.

I.G.Y. Geophysica norregica 20, 1. Waurilt utpeartoi

Labilzke. K.. Naujokat. B.. Lensebow. R.. Petzoldl. K.. Ameasuementi..ohet. B, an ' lwln E- 3Z- 1at 9sp9riRajewski. B. and Wohlfart. R. C 11985) The third wint easuemn of the 0.)gre'ln in'~ th tosPhaeiof MAP*-dynaesics. 1984,85; A winter with an extremely bn n h 1

11genln ntengtlwPae

intense and early major warming. Bedl. Berliner WettKarr e Space Sct. 27. 341.67. 1985

65

MID-WINTER HYDROXYL NIGHT AIRGLOW EMISSIONINTENSITIES IN THE NORTHERN POLAR REGION

H. L. MYRA80* W. C S. DE[HRGeophysical Institute, Unversity of.Caska. Fairbanks, AK 99'01. UiS.A

Received infinailjorm 28 June 19831

Abisinic-Groind-based spectrophotometnic measurements of night airgiow OH 18-3) band absolateintensitesithe pbolarusp region 1

78.-Nldunng winter solsice are reported. A mean value of .525 = 4 R is

found for the absolute intensity of the OH (8-3) band. Maxinum and mtininsumn daily mean values were 13Oand 320 R respectively with hourly mean values ranging from ISO to 1020 R Neither a winter solstice iniumor maximum in the intensity is obvious from the data. No consistent correlation was found between theabsolute intensity and geomagnetic and solar activity. A mean transport of 0and 0, into the polar cap regioncorresponding to a mendiottal wind speed of& atceat 20mis-' at 9(3cm height seems necessary to maintain theobserved intensity. A dominant seidiurnal tide componeni is found in the intensity data, both on a 20-dayand a 3-day time scale.

1. iNTRODUCTiO0N Roesler, 1955; Fedorova, 1959; Kvifte. 1961; Noxon.

The night airglowemissions originating from the t x~r 1964; Shefov, 1969a. b. 1970: Hunten. 1971; Sivjee r

'al.. 1972; Kraasovsky, 1972; Wiens and Weill, 1973:-X'rl, transition of the hydroxyl molecule are Vallance-Jones. 1913; Good. 1976; Petitdidier andconfined to the 80-95 km region ofihe atmosphere. The Teitelbaum, 1977:; Kraasovsky et al.. 1 )17 moreels etemnisaions are a consequence of the ozone reaction al.. 1977: Fishkova. 197,8: Shags, v, 1980: TakahashiBates arid Nicolet, 1950: Herzberg, 1951): and Batista 1981).

0, +H - OHNt <_ 91.-O_. 1 By contraati very little is known regarding airglow

Addiiona mehanims ave lso eenpropsed ermi.ssion in the winter polar cap region. Reed 11976)Addiiona mehanims ave lso eenpropsed utlized photometric data at 6230 A 50 A bandwidth)

(Nicolet, 1970: Kraaaovsky, 19721 but have not so far from the Oqo 4 spacecraft and found astrongly variablebeen shown to contribute significantly compared to the enhanced i'.enisitv leve! durn part of the 1967 68ozone mechanism (E vans and Llewellyn, 19'2; itrathg

HarrsonandKenall 193 Lewelyn nd ong witerat ighlatitudes that seems not totallyHarriso Taedh KendaW. 19 1; Takeweshin and Longta explainable by auroral contamination. The enhan,;e-

197: Tkeuhief d.,191: akaash ad Bttaa. ments may be explained by an enhanced OH night198 1: Myrabo et al., 1983). airglow emission and partly'by an increased nightglow

Measurements of the OH emission alone or together continuum from NO, (Gadsden and Marovict' 1973).with other night airglow emissions such asthe Of5577 Walk"r and Reed (1976) showed that the enisancedand 6300A lines and the 020y)~ and C1A) bands may be levels probably were connected with stratosphericused as a means of remotely monitoring the spatial and

temora vaiatonsfth od oyge cocenratonas warming events. In both these works which are based

well as the dynamical behavior of the atmosphere. o htren aa uoa otiuint h

Neutral temperatures at the mrescopause level may be enhancements cannot be ruled out and it seems difficultfoun byanayzin th roatioal ineintensity to attribute the enhancement to a specific emission. i.e.

distribution of the emission bands IKvifte. 19591. OHr aigN esueet n h 2acprgoExtensive studies of the OH ntght airglow have been Ote igomasrensnthp:a cpein

carred out at low and mid-latitudes tn order to clarify arth0157 emsintuycredotttetheexitiio mehaisman beavor fte O ngh tAntarctic station ,.sing photographic equipmenttheeri'.ttinmecanimandehaiorotheH niht (Sandford, 1964). Mu~llen et al. 11977) report ground-

airglow itself, and also~ as a means of remotely studying based measurements of the 01 5577 A emission fromother atmospheric parameters and processes, such as Thule, Greenland. They found the 01 5577 A ntghttemperature, wind, composition. tides, gravity waves. airglow emission in polar Cap region to be quiteetc. (Memnel. 1950; Dufsy, 1951: Chamberlain and different from that ohserved at mid- and low latitudes.

___________________________________ No significant diurnal variation was found in the data:*On leave from Norwegian Defense Research Estab- large amplitude variation, on the time scale minutes to

lishment. N-2007. Kieller. Norway. hours are relattvely commotn and for periods of 4-19

263

66

264 H K \t'sRxBi a - S DEEHR

davs stronglv, enhanced levels are found to exist This time insigniticant at Spitsbergen becau r , .riuapattern is similar to that found in the OH intensitN lower occurrence frequenc of auroras iStringer anoreported here. and may point toward acommon source. Belon. 19h'" and because auroras at these latitudes areThey also show that the intensitn distribution is generally at a higher altitude and emit mainil atomicdifferent from lower latitudes i.e showing a skewed lines Sitjee and Deehr. 19801. Thus, a normal, hiagh-bimodal distribution for polar region compared to a responsiin, auroral spectrophotometer operateo at askewed unimodal at low and mid-latitude) also found resolution of20X)Oi4A bandwidthior more may be usedby Sandford 11964). A correlation between boundary to obtain measurement of the OH bands ano ines

crossing of the interplanetary field and the sign of the clearly resolved from auroral emissions.gradient of the airplow intensity ii.e. increasing or Ait the Longyearhven Observatory. a I m and adecreasing slopei was also found to exist for two of the i 2 m high-throughput Ebert-Faste spectrometer are

observing seasons coupled to a mim-computer recording in the photon-

Ismail and Cogger i1982) extended the studs by counting mode The I m instrument, used lor theMullen et al. 119771 using data from the ISIS-2 satellite, measurements reported here. is further described b,They attributed the enhanced 01 557- A emission Dick et a 119"01 and Sivjee er al. 1 19"21.

periods to increased meridional transport of oxygen The spectrophotometer was normally operated forinto the polar cap and showed that stratospheric 24 h a day from December l98 to January 1983 Forwarming events might result in an enhanced polar cap the absolute intensity measurements reported here.

O 557 A airglow. only penods with clear sk and stable atmospheric

In two recent papers. Myrabo et al. 11983) and transmission were selected. The transparency )I theMyrabo [1983) have reported ground based measure- atmosphere was checked by %isuallv observing aments of OH emissions from Spitsbergen (78.4 N) known sequence of non-variable stars. A *0-20,during winter solstice condition. Large amplitude van- intensity ariation of the stars could routiney beatton of both intensity and temperaturetup to - 70 K detected Within periods where detectable trans-in temperature) is foand in the internal gravity wave parency variation occurred, data were not utilized.period range On a larger time scale, the daily mean Spectra were also rejected when Fraunhofer inestemperatures show a wave-like pattern with deviations occurred during twilight and full moon periods.from the mean, each lasting for about 3-10 days. A The spectrophotometer was pointed towards thesemi-diurnal but no diurnal vanation in the zenithandtheOH 8-3)bandPlinesinthe 290--4111Atemperature is also evident, region were scanned in 8 or 32 , using tme

The purpose of thts paper is to continue the study of spectrophotometer in the second order with a I or2 nm'the OH emission in polar cap region, utilizing the slit corresponding to a bandwidth of 1 5 and I A1982 83 season absolute intensity measurements at respectively Each scan was recorded on magnetic tape.winter solstice. From results reported in the previous Individual rotational temperatures were calculatedpapers and in this paper. it seems obvious that the OH from I -and 3-h integrated scans, by employ ing K, ifte sintensity and temperature pattern in polar cap regions method using the intensity ratio ofthe P,(2l P;i 3 1, P 41are mainly governed by large and small scale transport and P1i5i lines tKvifte. 19591 The total band intensityand wave phenomena. A disturbed OH intensity and was calculated using the well-known formula for thetemperature pattern. i.e. strongly fluctuating with tine variation of the intensity of the lines in a rotational-withperiodsfromminutestodays. seemsmoretobethe subrational hand as a function of t5 a; angularrule than the exception, momentum J iHerzberg, 195MS

2. OBSERVAfIONS AND DATA REDI-ClION 1'= - ' J I'.' e-

The OH emission data employed in this work were where l, is the total i - band intensity. C a constant.part of measurements undertaken during the 19'82 83 S, itJ the line strength. F, tJi the rotational term valuecampaign of the Multi-National Svalbard Auroral in the upper vibrational level i. . the wavelength of theExpedition (Deehr et al.. 19801 close to Longyearbyen line, and h, c. k and T have their usual meaning. Theon West Spitsbergen 178.4 N. Lat., 15E Long. total OH i8-3t band intensity was thus calculated iandgeographic). 'he constant C determined) using the absolute intensity

OH emissions are normally predominant in the near of the Pt3 ;ine together with the temperature Ti.r. part of the night sky spectrum. Contamination by calculated from the P, lines using Kvifte's method.auroral molecular emission as in the auroral zone Summations fCr the R. Q and P branches up to v = 9(Vallance-Jones. 1974: Meriwether. 19751is most of the were performed. The line strength of Honl and London

67

Hydroxyl night airglow emissions tl

11925) and the FJ values as determined from Dieke not cover 24 h, Where less than a 12-h period of dataand Crosswhite (19481 and Kvifte (19591 have been were missing, interpolation was performed using meanused. Absolute intensity calibration was performed in values on each side averaged oser the hours of missingthe field using a standard lamp and a diffusing screen. data. Daily f24-h mean values were then calculated

Probableerror for asinglemeasurement t.e. [-or 3-h from the original and the interpolated data. The OHintegrationi sestimated to be -3 Kin the temperature rotational temperature data for the same period asand less than 10, in the relative intensity. The given by Myrabo (1983) is added to the figure foruncertainty of the absolute scale is about 20'. mainly comparison. In addition, tile daily mean value of thedue to calibration uncertainty. The absolute intensity planetary geomagnetic activity index K, is showngiven s not corrected for atmosphenc extinction, together with direction of the interplanetary magneticCorre ted emission intensity, i.e. at the emission height, field (i.e - - I. Days with geomagnetic storm suddenmay bc found by applying an extinction coefficient of commencements are indicated by a A in the lower part0.2,Myrabo. 19791and a ground albedo for snow at the of the figure.appropriate wavelength of 0.7 IMyrabo et al.. 1982). The dataset consists of absolute intensities from

Example of the quality of the data used is given about 150000 individual spectra recorded during aelsewhere(Myrabo, 19831. Daily mean values are based total time of 402 h covering 20days. Even though thereon 12-24-h averages of the 1- and 3-h values. Auroral are several periods with missing data (due to badactivity was monitored by the intensity of the 7320, 30 A weather) Fig. I shows neither maximum nor minimumOl lines. Aidational care was taken to assure that there around winter solstice. An irregular intensitywas no contamination from N 21 P and other auroral fluctuation around the baseline mean of 425 =40 R.molecular emissioi.s, seems to be a better description of the I month winter

solstice period covered. This is somewhat in3. RESUL.TS AND DISCUSSION disagreement with the results and suggestions reported

by Shefov 11969ai. He reports a winter maximum in the3 1 Winter solstice behavior of the intensity OH emission intensity at latitudes poleward ofapprox.

Absolute intensities of the OH (8-3) band for each 35 igeographic). Both the data from Yakutsk (621N.day as derived from hourly and three-hourly means are geographicland Loparskaya 168-N, geographic) showsplotted in Fig. 1. Thedailymeansareindicatedbyfilled a very strong Intensity maximum around wintersquares and straight lines are drawn between each solstice. Independent measurements by Visconti ei al.mean. Dashed lines indicate that intensities for one or u19

71)at42'NandbyHarrisonetal.U1971(atSl Nalso

more days are missing. Some of the measurements did shows winter intensity maxima in accordance withShefov I 1969a). On the other side, measurements by

Kvifte 11967) from Tromso I 70'N. geographic) reveals a-..280 clear tendency of a declining intensity during the

2 autumn towards the end of November pointing% t ¢.t, lt , •. 20 1. towards a minima at winter solstice.

220 _ Because the main production of 0 is rmiated to solar-. . photodissoctation (Giachardi and Wayne. 1972;

, 10 Simonaitis et al., 19"3, Moreels et al., 1977), by the800 - process

2 G00 60*t* *-O ~0 - hv - 0 -h O O, 12)

saoe . migh: expect the diminshig production of 0 in

200 the :4-h polar night to lead to a decrease in the OH2;? 5~ 2 ~emission. i.e.

10 2 5,12 20112 25,i 'c"i 51i 1 A 0 0 , -o0,TIME (DAY MONTM / 1982-83)

FIG. I DAILY MEAN VALUES OF THE OH 18-3) BAND INTENSIT andARE GIVEN WIT STRAIIGKT LINES DLAWN BETWEEN THE MEANS0Dashed lines indicate data for one or more days is missing. 03 - H - OH - 02 4Planetary geomagnetic disturbance index K, and OH The results reported by Kvifte 119671 and the winterrotational temperature are also given. Days with geomagnetic sstice minimum of the 01 5577 A nght airvowstorm sudden commencements and the direction of oiInterplanetary magnetic field are indicated by A and -. - emission ir polar cap regions found by Mullen ei al.

respectively 1977) also suggest that one rather should expect a

68

266 N, K. MYRAe and C S DEEHR

winter solstice minimum than a maximum at these amountofdata.gtvesexpenmentalevidenceforadirectlatitudes i i.e. > 70'Ni. However, the one season of data effect on the OH emission intennity and temperaturepresented here do not rule out a broad € > I monthi from geomagnetic disturbances. He found a wavelikeminimumormaximum. Thedoninanceofthe wavelike disturbance in the intensity with an amplitude up topattern may also partly mask an eventual trend around 7010 of the OH emission intensity at Zvenigorod 155" Nsolstice travelling equatorward from the auroral zone with an

On a planetary scale the absolute intensity of OH apparentspeedof5-Oms- The amplitudedecreasedemission is found to be at a minimum around 25' as the disturbance moved equatorward.(geographicl, increasing steadily poleward towards Theoretical calculations by Maeda 11968!1 and681N IShefov. 1969a). Absolute intensity measure- Maeda and Aikin (1968) concerning the dissociation ofments above 7 0' latitude have only been recorded from 0 molecules in the case of impact of electrons at theirborne platforms like those by Noxon 119641 and by appropinate altitude, show the effects to be very small

Sivjee et al. 1972) providing data only for a few hours on the 0 and 03 concentration. Dissociation of 0.not sufficient to establish any seasonal trend or mean associated with auroral substorm directly, seemsvalue. The mean winter solstice value of 425 ± 40 R therefore not sufficient to explain Shefov's data. Morefound here for the OH 18-3) band is therefore the first recent discussion (Brekke. 19771 concerning effectsmean value based on a more extensive data set. It is from Joule heating and ion drag heating on mesosphencclose to the yearly mean found by Krassovsky et al. and thermospheric temperatures also concludes that(1956) at 55'N. Compared with more recent the energy present in an auroral substorm seems notmeasurements, it almost duplicates the 2 yr mean of 408 sufficient to alter the mean temperature significantly in± 40 R reported by Takahashi et al. 1977) from Brazil the 90 km region.(23'S) which is very close to the normalizing latitude From comparison between the vanation in the dailyemployed by Shefov ( 1969a) in his work, i.e. the latitude mean OH emission intensity and planetary geomag-of minimum intensity of the OH emission. Thus, using netic disturbance index K,, time of storm suddentLe 2 yr mean Brazilian value to calculate an expected commencement and boundary crossing of thewinter solstice maximum at 68'N employing Shefov's interplanetary magnetic field, no consistent correlation(1969a) latitudinal and seasonal variations, an emission can be seen from Fig. I between OH emission intensityintensity of -950 R is found. Extrapolating this value and the other parameters. Applying 3-h averagesaccording to Shefov's results, i.e. a latitudinal increase instead of daily averages and allowing for time delays.and a winter solstice maximum, gives an even higher still leaves us with an inconsistent correlation. i.e.value at - 78'N by 20-30",, i.e. - 1.2 kR. If we instead positive correlation between one or more peaks may beassume a winter solstice minimum at 80'N, of the same found while a negative or no correlation applies to theorder, i.e. - 30",, an emission intensity of - 530 R is rest of the data. This is very similar to the findings bycalculated for winter solstice condition at 80*N. The Saito (1962) and Weill and Chnstophe-Glaume (1967).latter value is in much better agreement with the actual While Shefov 11969a) argued that these observationsvalue observed, showed negative results because the sites were too far

The data material gathered so far, thus suggest a from the aurora regions and/or data only covered awinter solstice minimum of OH emission rather than a couple of days, this objection cannot be held againstmaximum at extreme high latitudes, i.e. > 70c This is in our data. Svalbard is far more closely surrounded byagreement with the O 5577 A nightglow emission the auroral oval than Zvenigorod.minimum in polar cap regions reported byMllen er al. One may, therefore, conclude that the data so far(1977). However, data for more seasons should be available indicate a very small or no direct correlationgathered before making any firm conclusions on this. between daily mean (or 3-h mean OH emission

intensities and geomagnetic disturbances in polar cap3.2. Relation to solar and geomagneric activity regions. It is evident that this also applies to the OH

It was first suggested by Krassovsky (1956) that OH rotational temperature as given in the upper part ofemission intensity should be affected by geomagnetic Fig. Idisturbances. Shefov (1959) reported an intensificationof the OH emission intensity during magneticallydisturbed periods. Later study by Berg and Shefov 3.3. Relation to transport and wave phenomena

(1962) and Saito (1962) failed to show a definite 3.3.1. Gravty wares. Figure I shows that thecorrelation between OH emission intensity and behavior of intensity and temperature dunnggeomagnetic activity. December reveals a similar pattern It maybe described

In a later work Shefov i 1969a). using an impressive as consisting ofa relatively stable level with three main

69

Hydroxyl night airgiow emissions 26

enhanced periods. This is particularly obvious for the made assumingtemperature. In an earlier paper when only the I themerdionalfluxtoreplacetheloss mailybstemperature pattern was available, Myrabo 1983) equation (1)).argued that enhanced temperatures lasting for a few (2) a loss rate of 4 x 10' [03] cm s - ' at the OH

days most likely were due to gravity wave effects, i.e. ermssion peak 190 kml (Moreels et al.. 1977),direct energy deposition as a consequence of the i3) a dark polar cap 650 km radius at 90 km heightbreakdown of the wavesIHines. 1965) and, or turbulentdiffusion increased by gravity wave action This results in a meridional wind speed of the order of

i Zimmerman and Murphy, 1977) transporting warmer 20 m s '. which may be compared with the 9 m s -'

air down from above. The intensity enhancements semidiurnal tide temperature associated wind speed

simultaneous with temperature enhancements further derived from part of the December temperature data

strengthen this suggestion. Increased turbulent presented in Fig. I iMyrab. 1983). Nisbett and Glenar

diffusion produced by gravity waves may be seen to be a (1977) have analyzed data from a series of high latitude

sufficient mechanism to explain both the enhanced vapor trail experiments and arrived at a meridional

intensity and temperature levels by vertical mixing of transport of 50 m s - ' into the polar cap region in the

oxygen and warmer air from above. Increased altitude range 100-I 10 km. Ismail and Cogger 198"2

turbulent diffusion was successfully applied by Moreels from 01 5577 .- airglow measurements required

etal.) 1977) in an oxygen-hydrogen atmospheric model meridional wind speed into the polar cap of 10-

to simulate mid- and high latitude observed deviations 20 m s - 1 to explain the observed 01 5577 .4 intensity

from a chemical production and loss pattern. Harrison and variations.

" eal.(1971)also pointed out a change in the downward Assuming that a significant fraction of the 03 is

transport or height of the oxygen profile as an transported across the polar cap, an even higher wind

explanation of observed enhancements in the OH speed than 20 m s - would be needed. To explain the

emission intensity, intensity gradients of the enhanced periods and the

Unfortunately, due toconsistent clouds or mist. only levels reached, a mendional wind speed up to approx.

a few days of absolute intensity measurements were 40 m s -' for some hours is needed. This seems to be

available during January. However, both the mean rather high for the particular altitude range under

intensity ater inJanuaryandthebehaviorofthedaily consideration i.e. -90 kmi). but Deehr et al. (19661

means seem to be not significantly different from observed the effects at 67:N. 145-W ofan Li release from

December. 58'N. 95'W within 12 h implying an average speed of60 m s - ' Even though the release was near 160 km

3.3.2. Stratospheric warming. A stratospheric warm- altitude, the Li was at the mesopause by the time of theing is normally associated with a cooling of the observation. Vertical mixing of oxygen from abovemesosphere (Labitzke, 1977. 1980: Schoeberl, 1978). could be one possible source in addition to meridionalConnected with such a warming is also a large-scale transport.redistribution and mixing of the upper atmosphere 3.4. Short term behavior-intensirv and temperature(Mclnturff, 1978; Labitzke, 19811 that is expected to correlationsresult in a large increase in the OH emission intensity corrtations

(Fukyam. 177;Waler nd Red.197; Fshkva. Fluctuations in both OH emission intensity and(Fukuyama, 1977; Walker and Reed, 1976. Fishkova' rotational ',emperaticre in the gravity wa~c period

1978).ThecoldmesopauseduringNewYear 1983does range ii.e. 5 rain or more with amplitud s fromcoincide with astratwarm event )Naujokat efal., 1983), rag, mo oe wt mltdsfocoincdewitastrawarmeentl~ujokaetal. 98, 20 K in temperature and +20' intensity are notbut we can hardly see any large scale enhancement of un20mKon tmr aure 2ntensity arenothe OH emission intensity following this event. This is uctuton apr Ow andrthent all T es

opposite to results reported by Walker and Reed) 1976). fluctuations appear now and then at all latitudes. but itThu, i sems hata sratvar isnotnecssaily is not regular behavior (Shefov. 1969a: Wiens and

Thus, it teems that a stratwarm is not necessarily Weill, 1973: Takahashi et al.. 1974; Misawa andneeded to result in a large enhancement of OH emission Takeuchi, 1978 .intensities in the polar regions. From OH emission data obtained during the

3.3.3. Meridional transport. As previously men- 198081 observed season at Longyearbyen. it becametioned in Section 3.1, there is no production of odd clear that very large amplitude regular and irregularoxygen compounds by solar photodissociation in the oscillations in both intensity and temperature alsopolar cap region during mid-winter. To keep the OH could occur in the polar cap region. A special event,emission intensity at t1 , observed 400 R level for the believed to be due to the passage ofgravity waves. with(8-3)band requires a certain meridional transport from amplitudes up o - 70 K in temperature within a fewthe terminator. An approximate calculation was hours was reported by Myrabo et al. 119831.

70

H K MsRAse and C S DEEHR

OH data obtained in 1982 83 further confirm that observations at mid- and low latitudes. assuming h asmedium-to-large amplitude oscillations in both a typical observing night. Choosing separaie S-htemperature and intensity are not exceptional for OH periods corresponding to 10-15 nights we are able toemissions in the polar cap region. produce both positive. slightly negative and insignifi-

.As found from observations at F2 region heights cant correlation between temperature and intensityiThome, 1968: Shashun'Kina and Yudovich. 19801 Consideration only of the data between 25 Decembersubstorms and geomagnetic disturbances are likely to t300 UT and 27 December 0700 U T., altogether 39generate internal gravity waves. Internal gravity waves, points, results in a correlation coefficient of 0.81 Onlaunched from geomagnetic disturbances and travel- the other hand, the data from 26 December 1600 U Tling downward into the mesopause region, though not to 27 December 2200 U T., consisting of 30 points.containing enough energy to alter the mean gives a negative correlation. i.e. -0.09 The negativetemperature significantly, may cause OH temperature correlation is not significant. Figure 3 shows anand intensity oscillations around the mean values, intensity-temperature plot in the case of a correlationObservations by Shagaev 11914) and Krassovsky a ul. coefficient of 0.81 The above analyses were also11977) showed that at least some of the gravity wave performed using 30-m integrated spectra, thusevents recorded at mid- and high latitudes near the providing twice the number of points. i.e. 78 and 60mesopause probably onginate in the auroral zone respectively. The result was not significantly different.and that they were connected with geomagnetic For limited time intervals, typically less than 12 h. adisturbances. higher correlation coefficient could sometimes be

Figure 2 shows a 75-h (i.e. -3 daysi continuous obtained by shifting the intensity and temperaturerecordoftheOH intensityandtemperatureasobtained curves relative to each other. Whether this indicates afrom 1-h integrations of emission spectra. This seems to real phase shift between intensity and temperature, asrepresent a relatively normal behavior, and as can be reported by some authors iNoxon, 1978. Takahashi erseen, considerable variations in both temperature and a.. 1979) is difficult tojustify because applying the sameintensity take place. time shift, in other cases results in the opposite effect. I e.

The correlation between intensity and temperature lowers the correlation coefficient.for the OH emission has been the subject of many Mechanisms, such as horizontal transport, verticalpapers and widely different and seemingly opposing transport, turbulent mixing diffusion, tides and gravttyresults have been reported, i.e. both positive, negative waves act differently on the temperature and theand no correlation (Takahashi et al., 1974. 1977 intensity. We believe that the rather confusing patternsMisawa and Takeuchi, 1978, Visconti et al., 1971. observedinbothintensityandtemperatureresultinginHarrison et al., 1971, Shefov, 1970: Takeuchi et al., positive, negative and insignificant correlation, and1981). sometimes all three possibilities within a 24-h period.

The 400 h of data, presented in Fig. 2 cover an only reflect the dynamical and chemical complexity ofequivalent time period of about 50 nights of the OH-emitting region. From data so far collected,

70, 270

am - 260

S00 250

400 2.0A < 'u e 230

V7, _"v 25

300 , , J , 230

0i 00 04 08 12 16 20 00 04 08 12 i6 2 0 54 02 5 6 SQ 26-- 25 DECEMBER ----- 20 DECEMBR - -- -- 27 OECRMBER

TIME (HOUR, DAY)

Fic. 2. 1-h AVRAGEsOF TTe OH 18-31 0AND 1NrENvSTY AND ROTATIONAL T M PERA TL RE ASORTAiNED EAcm HOURDURING s 3-OY PERIOD IN 1982.

71

Hydroxvl night airglow emissions :.

Boo- oo

goo a oo7000

0

6000-

a 00 , 5000-z , 5,

3 200 4000-

3 0 00- a

200 210 220 230 240 250 .2 [L

TEMPERATURE () oooFIG. 3 OH ig-3) BAND INTENSITY-TEMPERATUE PLOT USING 00_

THE %EAsuRsEtENrs ErwEEN 25 DECEMBER 1300 U T ro 27 ,-DECEMBER 0700 U T 1982 SEEF G. 024 12 8 4 3

The best-lil linear relationship is indicated by the least square PERIOD (HOURS)regression analysts line. The slope is I1 2 R K. FIG 4

it seems that transport and wave phenomena are alsoconfirmsourpreviousresultusingthetemperaturedominant factors in the overall behavior of the OH data Myrabe. 1983).

emitting layer in polar regions.

4. SUMMARY

3.5. Tidal ciraponents Ground-based observations of atmospheric OHIn a previous paper Myrabo 11983) reported the emission intensity and temperature have been carned

presence of a semidiurnal tide in the December 1982 out at 78'N during December and January 1982 83.temperature data by applying a superimposed epoch The results from the analysis of the temperature data isanalysis of 19 consecutive days. A diurnal tide com- reported elsewhere (Myrabo. 19831ponent was not evident. A mean value of

425 40 R is found for the absolute

Using Fourier analysis of the intensity data coverng intensity of the OH 18-31 band. Maximum andthe same time interval, a dominant power peak around minimum daily mean values were 770 and 320 R12 h is found with far less power around 24 h. According respectively while the hourly mean vai,es ranged fromto a model calculation by Forbes (19821 one should 180 to 1020 R.expect the diurnal tide mode to be dominant at high The daily main values of the intensity show largelatitudes and polar regions. Spizzichino 1 1969) and amplitude variations qualitatively very similar to thoseTeitelbaum and Blamont 119751 have argued that found in the O1 5577 A airglow emission in polar caprandom interaction with gravity waves has more regions IMijalen et aL, 1977). There is, however, noimportant affect on the first diurnal mode than on the obvious minimum or maximum seen within the Isecond. Thus. the effect ofaveraging over a period of 19 month-long period covered around solstice as theredays could be to cancel out the diurnal mode. To check was for the 015577 A airglow emission) Mitlen et al.this we have applied Fourier analysis to the three 1977). A broad minimum or maximum cannot be ruledconsecutive days of intensity data shown in Fig. 2. As out, however the relatively low mean value of 425 _ 40 Rseen from the result presented in Fig. 4. the semidiurnal rather indicates a minimum than a maximum The lacktide mode is again the dominant mode. The energy seen of production of 0 by photodissociation of O in thearound 4-5 h and around 3 h are on the limit of what is polar cap region dunng the polar night also suggest asignificant, and may partly be due to side lobes minimum. The flux of 0 and O necessary to keep theappearing from the Fourier analysis and noise in the OH emission at the observed level indicates aintensity data meridional transport corresponding to a wind speed of

The finding of the semidiumal tide mode also to at least 20 m s at 90 km height. However. verticaldominate on a time scale as short as 3 days may be transport induced by gravity waves and mixing wouldinterpreted to be due to a very high occurrence lower the necessary wind speed considerablyfrequency of gravity waves effectively interacting with The daily mean OH emission intensity andthediurnal tidemode but not with thesemidiurnal. The temperature s,,ow no evidence of any consistentdominance of the semidiurnal tide in the intensity data correlation with geomagnetic or solar activit,

72

H, K MsYRea and C. S DEEHR

A dominant semidiurnal tide component in the OH Variations during iratosphenc warming evnts. J atmis,intensity is found and itrongly confirms the previous iemr Phvs. 39. 317finding of the same variation in the temperature data Gadsden. M. and Marovich, E. 1973) The nighiglsa

continuum. J .armos. trt Ph vs. 35. 1601.Myrabo, 1983). Even on atime scale ofthree days, the Giachardi. G.J. and Wayne, Rt. PW 1972) Proc. R Soc. L.nd.

semidiurnal component is dominant. The power A330, 331spectra appearing from Fourier analysis barely reveals Good. Rt E. (1976) Deterniinaiion of atomic oxvgen densit vany peak around 24 h, i.e. the diurnal component. from rocket borne measurenment of hydroxyl tirglowv.

From the OH data gathered so far, it seems that Planet. Space Sci. 24. 389.Harrison.A. W and Kendal.J .W (1973) Airglow hvdrosid

transport and wave phenomena are the dominant intensity measurements 0.6-2.3 gin. Planet Space Sri 21.mechanisms defining the behavior of both OH intensity 1731and temperature. Further airglow observations are Harrison. A. W.. Evans. W F J. and Llewellyn. E. J 1971)necessary to understand the complex menopause Study of the (4-1) and 15-2) h ' droxyl bands in the night

airglow. Can. J. Ph vs. 49 2509.region dynamics and photochemiastry in the polar lferberg, G, (1950) Spectra q/ Diatomic MWolecles~ Vanregions. Nostrand. New York.

Acknowledgements- Financial support for this research was Herzberg, G3. ) 19511 The atmosphere of planets. J R. Astron.provided by National Science Foundation through grants Sinc. Ca. 119500. nsa etn o h peATM80-1271 8 and ATM82-001 l4toGeophysicai Instituteof Hns .0 161Dnmclhaigo h pethe Isivesiy of Alaska. One of us (HKM) was supported by a atmosphere. J. geophiys. Res. 70. 1 77fellowshipgrant from Royal Norwegian Council for Scientific HonIl. H. and London. F. 11925) Z. Phys. 33. 803

Hunten. D. M. (1971) Airglow-irodaction and review. in Theand Industrial Research. Radiating Atnmosphere IEdited by McCormac, B. M.I. p, 3.

REFERENCES Reidel. Dordrecht.tsmail. S. and Cogger. L. L (19821 Temporal variations of

Bates. D. R. and Nicolet. M. (1950) The photochemistry of polar cap 015577 A airglow. Planer Space Sci. 301.865atmospheric water vapor. J. geophys Res. 55. 301. Krassovsky. V. 1. 11956) Infrared night airglow as a

Berg. M. A. and Shefov, N. N.0l62)Emsussionsof thehydroxy! manifestation of the process of oxygen recombination inbands and ofthe(0-2) .n8645 Aatmosphetc band ofoxygen Airglow and Aurora (Edited by .u."trong. F Rt andin the nightglow. Planet. Space Sri. 9, 1. Dalgarnio A.i p, 193. Pergamon Press. Oxford.

Brekke. A. (19771 Auroral effects on neutral dynamics in Krassovsky, V. 1. I t972) Infrasonic variations ofOH emissionDynamical and Chemical Coupling Between the .esaral and in ihe upper atmosphere. A. Geoph vs. 28. 739Ionized Atmosphere. (Edited by Grand". B. and Holtet. Krassovsky, V. tI- Potapov. B. P.. Semenos. A. t., Shagaev. M.J. A.). p. 313. Readel, Dordrecht. V., Shefos. N. N.. Soboev. V. G. and Toroschelidze. T 1.

Chiamberlanj W. and Roesler.F. L. I1955)TheOH bands in (19771 Internal gravity waves near the menoipause and thethe infrared airglow. Ap J 121, 54 1. hydroxyl emission. .4. Geophys. 33. 347

Deebe, C. S_. Romick. G. J. and Belon. A. E. 11966) An Kvifte. G. (1959) Night glow observations at As during theobservation of artificial lithium twilight emission. J.A.rmos. I.GY. Geophys. ,Vornegica 20. 1Sri. 23. 362. K vifte G. 1196 1) Temperature measurements from 031H hands.

Deehr. C. S., Sivjee, G. G., Egeland. A,. Henriksen, K_ Planet. Space Scs. 5. 15 3.Sandholt. P E,. Smith. ft. Sweeney. P.. Dunicant. C. and Kvifte, G. 1. 01967) Hydroxyl rotational temperatures andGilmer. F. ( 1980) Ground-based observations of F-region intensities in the nightglow. Planer Space Scz. I5. 1515associated with the magnetospheric cusp. J1. geophvs. Res, Labitzke. K. 119771 Strstospheric-mesosphenc mid-winter95, 2185. warnings, in Dynamical and Chemiical Coupling Betwaeen the

Dick. K. A, Sivjee G. G. and Crosswhite. H. M. (1970) Aircraft NVeutral and Ionized Atmosphere (Edited by Grandal. B. andairglow intensity measurements; variations in OH and 01 Holiet. J1. A.), p. 17. Reidel. Dordrecht.(5577). Planer. Space Sci. I1. 887. Labitzke. K. (1980) Climatology of the stratosphere and

Dieke. G. H. and Crosswbite, H. M. (1948) Bumblebee Series. mesosphere. Phil. Trans. R. Soc. Lond. A296. 7John Hopkins Univ. Rep. No. 87. Labitzke. K. (19811 Siratospheric-esesospheric mid-winter

Dufay, 3.11951) Bandes demission des molecules OH et 0, disturbances: A summary of observed characteristics. J.damt le spectre duciel nocturne. entre 9000et I 1000 A. Ann. geophvs. Res. 86.9665.Geophys. 7. 1. Llewellyn. E. J. and Long, B. H. t 19781 The OH Metnel hands

Evans, W, F. 3. and Llewellyn, E.J. (197Z( Excitation rates of in the asrglow -the radiative lifetime. Can. .1 Phvs. 56. 58I1the vibrational levels of hydroxyl and nightglow intensities. Maeda, K. (1968) The auroral 0,-dissociation and thePlanet. Space Sdi. 20. 624. infrared OH entission. A. Geoph vs. 24. 111

Fedorovs, N. 1. (1959) Hydroxyl emission in the apper Maeda, K. and Aikin. A. C. (19681 Variations of polaratmosphere (translated title). Izv. Ak-ad. .Vauk. SSSR. Ser. mesospheric onygen and ozoneduring events. Planer.SpareGeofiz. 6,936. Sri. 16.,371.

Fishkova, L. M.(19781 Intensity fluctuations of the nocturnal Mclnturff. R. M. (19781 Stratospheric warmings- Synoptic.emission of the upper atmosphere duning stratospheric dynamic and general-circulat.Ton aspects. NAS4A Ref- Pabl.warnings. Geomagn. .4e,on. 1& 375. 1017.

Forbes. F. M. (1982) Atmospheric tides: The solar and lunar MetnelA. B. I1950)OH emission hands int the spectrum of thesemidiurnal components. J. qeoph vs. Res. 87, 524t. night sky. Ap. J. tI1, 555.

Fukityama. K. (1977) Airglow variations and dynamics in Meriwether.J. W.0975)(High laijtudearglowvobsrvationsofthe lower thermosphere and upper mesosphere - II. correlated short term fluctuations in [he hydronsd Metnel 8-

73

H-Iydroxyl might airglovv emissions

3 band intensity and rotational temperature. Planer. Space itmosphere -11 Effects of a nht atmosphere PlanetSei. 23. 1211 Spare Sri. 17, 1629.

Misawst. K. and Takeuchi. 1, 119781 Correlations among 0, Shefov. N N. 119701 Hydroxyl emission of the upper1111) atmospheric band, OH (8-33 band and [01) 5577.A atmosphere -Ill. Diurnal variations. Planet. Space Sc; 19.line among P~j2). P1t31 and P,141 lines ofOH 38-31 band. J. 129.aitnos. ter. Phvs. 40. 421. Sunionattis. R.. BraslavskY, S., Heikien. I and Nicoiet. M.

MoreelsG.. Megte G, Vailance-Jones, A. ansdGattinger, R. j 19733 Chsem Plys. Lett 19.601,1977) An oxygen-hydrogen atmospheric model and its Sivjee. 0. 13. and Deehr C. S. (1980) Differences in polar

application to the OH emission problem. J. aonos. ten atmospheric optical emissions betwieen mid-day cusp andPltys 39. 551 night-time auroras. in Exploration of the Polar Upper

Mullen, E. 13.. Silverman, S. M. and Korf. D. 1. 1977 AtimosphereEditedby Deehr. C. S.and Holtet, J A.i; 199Nightglow 1557.7 sni of 011 in the central polar cap. Planer. Reidel. Dordrecht.Space Sci. 25. 23. Sivee. G. G.. Dck.K.A. and FeldtmanP D. (1972)Temporal

Myrabe. H. K. 11979) Atmospheric extinction during clear varitions in the sight-lime hydroxyl rotational tempera-winter nights in the Skiboth Valley, Northern Norway. tare. Planet. Space Sci 20. 261.1974-1977, Report F-340. Norwegian Defense Research Spimzchino. A. 11969) Etude des interactions entre 1esEstablishment. differentes composantes du vent dans Is haul atmosphere.

Myrabo. H. K. 319831 Temperature variation at menopause A4. Geopsys. 25.93.levels, during winter solstice at 78'N, Planet. Space Sri., in Stringer, W. J. and Belon. A. E. 9167) The statistical auroral

press.zone during IQSY and its relationship to magnetic activity.NMyrabe. H. K._ Deehr, C. S. and Siviee. G3. 131983) Large J. genphtvs. Res. 72. 245.

amplitude nightglow OH (8-3) hand intensity and Takahtashi. H. and Batista. P P 119811 Simultaneousrotational temperature variations during a 24-h period at measurement of OH 19.4.18. 3),(7. 21,16,2) and (5,.1) bands78'N. J. geophys. Res, in press. mn the airgiow. J. geophYs. Res. 86. 5632.

Myrabe, H. K.. Lillesceter, 0. and H0tMYr. T. 31982) Portable Takahashi, H_. Batista- P P.. Clemesha, B. R.. Suionich. D. M.field spectrometer for reflectance measurements 340- and Sahat. Y. 01979) Correlation between OH. NaD, and 012500 nim. Appl. Opt. 21., 2855. 5577 A emissions in the airglow. Planet. Space Sri. 27.801.

Naujokat. B_. PetzoldI. K_. Labitzke. K. and Lenschow. Rt. Takahashi. H.. Clemeshia. B. R. and Sahat. Y.( 19743 Nightglowf1983) Beilage cur Berliner Wettenkaste. 57 81 OH 38-3) band intensities and rotational temperatare at

Nicolet. M.1 970)Ozoneand hydrogen reactions. .Geophys. 23'S. Planet. Space Sri. 22. 1323.26C 53 1. Takahashi, H.. Satat. Y.. Clemesha. B. R.. Batista- P P. anid

Nisbett. F. S. and Glenar. D. A. 31977) Thernospheric Teixetra, N. R. 1977) Dturnal and seasonal variation of OHmeridional winds and atomic oxygen depletion at high 18-3) airglow band and its correlation with 0f 5577 Alatitudes. J. qeopltys. Res. gz.4685. Planer. Space Sri. 25,.541.

Noxon. J. F.3(1964) The latitude dependence of OH rotational Takeuchi I., Misawa. K.. Kato. Y. and Aoyaina. 1. 319813temperature in the night airglow. J1. geopltys. Res. 69. Seasonal variations of the correlations among nightglow4087. radiations and emission mechanism of OH4 nightglow

Petitdidier. M. and Teitelbaum 14 31977 Lower thermo- emission. J. aimos. re-r Ptvs. 43. 157.sphere emissions and tides. Planet. Space Sri. 25. 711. Teitelbaum. H. and Blamont, J, E. 119

75) Some conseq~uences

Reed. E. 1. 119763 Polar enhancement of nightglow emissions of non-linear effects on tides and gravity waves. J. atnios.near 6230 A. Geophys. Res. Letr. 3. 5. tenr Plrys. 37. 697

Saito, B. 319623 Unusual enhancement of night airglow Thome G..319683 Long-period waves generated in the polarintensity at low latitudes on November 13. 1960. Antarctic ionosphere during the onset of magnetic storms. J3. qeopliys.Rec. 14.8. Res. 73. 6319

Sandford, B. P 31964) Aurora and airgiow intensity variations Vallance-Jones A. 11973).The infrared spectrum of thewith time and magnetic activity at southern high latitudes. airgiow. Space Sri. Rev, 15. 355.j.taros. terr. Phty.2 26749. Vallance-Jones. A. 1 9743 Aurora. Reidlel. Dordrecht

Schoeberl. M. R. 31978) Stratospheric warnings; Obser- Visconti G3.. Congeduti. F. and F iocco, G3.1 197 1) Fluctuationvattons and theory. Rev. Geopltys. Space Phys. 16. 52 1. in the intensity and excitation temperature in the OH

Shagaev. M. V. 319743 Relation of rapid iariations of the airglow (8-3) band. in ThseRadiating Atmtosphere (Edited byrotational temperature of atmospheric hydroxyl to MAcCormac. B. \. p. 82. Reidel. Dordrecht.geomagnetic activity. Geonssqn. Aeron. 14. 649. Walker.J. D. and Reed. E. 1. (19761 Behavior of the sodium and

Shagaev. .M. V. 119803 Vertical temperature gradients and hydrosyl night-time emissions during a stratosphericdissipation of internal gravity waves near the mesopause. warming. J. .Amps. Sri. 33. 118.Geonssqn. Aeron. 20 529. Weill, 13. and Christophe-Gluame. J. 11967) Ciel nocturne et

Shashun'Kina. V. M. and Yudovich. L. A. 31980) Effects of aurormsde basse latitudeen petodedorange magnetque..internal gravity waves during the magnetospheric substorm Geophys. 23.317.of February I5. 1979. Geonssgn. Aeron. 20. 516. Wies R. H. and Weill,. 119733 Diurnal, annual and solar

Shefov. N. N. 31959) Intensities of some twilight and night cycle variations of hydrosyl and sodium nightglo%airglow emissions: spectral. electrophotometrical and intensities tn the Europe-Africa sector. Planet. SpaceSri. 21.radar rmsearches of aurorae and airgiow. U.S.S.R. Acad. Sri. 1011.1. 25. Zimmerman. S. P anid.Murphy. E. A. I 1977)Stratospheric and

Shefos. N. N. 0 169a) Hydroxyl emission of the upper mesospheric turbulence in Dyniamical and Chemicalatmosphere- 1. The behavior during a solar cycle, seasons Coupling Between the Neutral and Ionized .Atmsosphsereand geomagnetic disturbance. Planer. Space Sri. 17, 797. (Edited by Grandal. b and Holtet. .1. Al). p. 17. Reidel.

Shefov. N. N. 3 1969bi Hydroxyl emission of the upper Dordrecht.

£~&~ ~k'

JOURNAL O)F GEOPHYSICAL RESEARCH,

75

0, (b1:* - X3Z-) ATMOSPHERIC BAND NIGHT AIRGLOW

AEASU MENTS N THE NORTHERN POLAR CAP REGION

H. K. Myrab i

Geophysical Institute, University of Alaska, Fairbanks

K. Henriksen

Auroral Observatory, University of Troms6, Norway

C. S. Deehr and G. J. Romick

Geophysical Institute, University of Alaska, Fairbanks

Abstract. The 02 atmospheric (0-1) band at emission intensity was estimated usiig the (0-1)8645 has been observed in the airglow during band and a Franck-Condon factor. Fraser et al.the 1982/1983 winter solstice from the ground at (1954] predicted a ratio of 21 for the intensityLongyearbyen, West Spitsbergen (78.4*N latitude, ratio between the (0-0) and (0-i) bands, Ignoring5E longitude, geographic). The average (0-i) any possible variation In the electronic transi-

band intensity for the continuous 18-hour period of tion moment. Laboratory measurements by Noxonthe observations is found to be 445 R ± 40 R. The [1961] gave a value of 20 t 4 while Nichollsmean temperature deduced from the 02 (0-i) atmos- [1965] reported a factor of 20. A measuredpherc band rotational structure Is found to be intensity ratio for the (0-0) to (0-1) bands wa.254 K ± 3 K compared to a mean of 249 K ± 2 K for found to be 17 t 2 by Wallace and Hunten 119681the 0H (8-3) band rotational temperature. This using dayglow measurements. A value close to 20indicates a shallow mesopause with an upper temper- has thus been commonly accepted [Greer et al.,ature gradient of - I K/km. Comparison with pre- 1981].vious observations shows little or no latitude Observations of 02 atmospheric band airglowdependence, although there is considerable scatter cover a wide range of latitudes, but both thein the data Indicating that the 02 airglow is ground-based and rocket measurements have beenhighly variable in time of the order of hours or made at less than 60* except one recent rocketless. experiment by Witt et al. (1979] at 68°N. It

Introduction has been suggested [Deans at al., 1976; Feldmann,1978] that the atomic oxygen concentration and

The 02 (b11 - X

3Z ) atmospheric band system thus the night airglow atmospheric band emission

was discovered in thegnight airglow by Meinel would decrease toward high latitudes. This has119501 who observed the (0-i) band emission at 9645 been contradicted by Witt et al. [1979], whoA. Morphology, diurnal and seasonal intensity found a relatively high 0Z (0-0) band nightvariations were extensively studied by Berthier airglow emission at 68*N.(19561 from Haute-Provence (43.9*N, geographic) A low value of the odd oxygen concentration atusing photographic techniques. Early absolute in- 95-100 km height should result in a very lowtensity measurements were reported by Barbier 5577 A green line night airglow and a low 02(1956] and Dufay (19581. Mean values of 2000 R and atmospheric band emission. A minimum (100 R) for1500 R were given for the (0-I) band. Later the 5577 A green line night airglow duringground-based measurements have showed significantly winter solstice in the polar cap region is re-lower values, giving mean intensities in the range ported by Mullen et al. (1977] and by Ismail and400-500 R for the (0-I) band [Berg and Shefov, Cogger [1982]. 02 atmospheric band night airglow1962; Broadfoot and Kendall, 1968; Wallace and observations have not been reported from the polarHunten, 1968; Shefov, 1971]. cap region. It is the purpose of this paper to

The (0-0) band is not observed from the ground report recently obtained spectrophotometric nightdue to absorption by 02 in the lower atmosphere. airglow measurements of the (0-1) band at 78.a'NObservations of the volume emission rate as a during winter solstice.function of altitude have been made from rockets[Packer, 1961; Tarasova, 1963; Witt er al., 1979; Observations and Data ReductionU-ta-sb- er -1 , 19?]. The altitudes of themaximum emission from these measurements of the The spectra used in this work were part of(0-0) band center about 94 ± 3 km except for the measurements made during the 1982/1983 campaignresults reported by Tarasova (1963] who found a of the Multi-National Svalbard Auroral Expeditionprofile maximum at 80 km. [Deehr et al., 1980] close to Longyearbyen on

Before the rocket measurements, the (0-0) band West Spitsbergen (78.4 N latitude, lbYE longitude.geographic).

rOn leave from Norwegian Defense Research Night airglow OH emission features are normal-

Establishment, Kjeller. ly predominant In the near-infrared spectralregion in the polar cap (Gault et al., 1981;

Copyright 1984 by American Geophysical Union. Myrabi et al., 1983]. The aurora in the polarcap is usually at a high altitude, and emits mainly

Paper number 4AO770. atomic lines [Sivjee and Deehr, 1980]. Thus, a0148-0227/84/004A-0

7 70$02.00 high-responsivity auroral spectro-photometer may

9148

MyrabO et al.: Brief Report

76

intensities is approximately 20%, mainly due tocalibration uncertainty.

OS-2) The yncidencc of aurora was monitored by the, o5 )p9 M5)p) intensity of the 7320/30 ! 0 II lines, the %

NlS) lines and the N2 1 P (2,1) band. The measurements

oMB- reported were obtained during a period devoid of.aurora except around 0700 UIT in the morning of

1, VIA a, January 12. The relative contribution from auror-ally excited 02 (0-1) emission could thus beestimated together with intensities of other auror-al lines and bands. It was observed that an

0o5(7-3) increase of the 02 (0-i) band of 20% to 33%corresponded to a simultaneous Increase of the7320/30 A lines by more than a factor of 5.Thus, the observed intensity of the 7320/30[0 11] emission served as a monitor of auroral

activity and an auroral contribution to the 02(0-I) band of less than 1% could easily be de-

000 6200 5400 eeo aso tected. The intensity measurements of the 02WAVELEMOrT (a) (0-i) band around 0700 UT have been excluded from

Fig. I. An airglow spectrum including the spec- the data in order to insure that the Intensitytral region from 7830 to 8800 A obtained during and temperature data used here contain no aurorala 12-min integration of 12-s scans at Longyearbyen. components.A pressure-broadened sodium line scattered fromLongyearbyen during overcast and very hazy weather is Results and DiscussionIndicated. The 1/2-m spectrophotometer with abandwidth of 7 A was used to record the spectrum Absolute Intensity of 0O Atmosphericwhich Is uncorrected for instrumental response. Bands (0,0) and (0,1)

Absolute intensity measurements of the 02atmospheric (0,I) hand taken during a continuous

be used to measure the night airglow features 18-hour period in January 1983 at 78* are shownclearly resolved from auroral emissions. Figure in Figure 3. The mean intensity during theI shows an example of a spectrum which includes observational period is 445 R ± 40 R. The datathe 7800 to 8800 A region. The bandwidth was have not been corrected for atmospheric extinc-approximately 7 A and the OH and 02 emissions tion. The OH (8-3) band intensity is given inare indicated. Figure 3 for comparison. Both the 02 and OH

At the Longyearbyen observatory, I-i and 1/2-m emission intensities tend to increase during thehigh-throughput Ebert-Fastie spectrophotometers observing period. The mean OH (8-3) band emis-operate in a photon-counting mode and are coupled sion intensity during this 18-hour observingto an on-line digital data processing system. period was 518 R. This may be compared with aThe coaligned, steerable instruments are described mean of 425 R found for the whole 1982/1983 obaerv-by Dick et al. [1970] and SivJee et al. [1972].

During an 18-hour period of clear weather(1900 UT, January 11, to 1300 UT, January 12,1983), the spectrophotometers were pointing N;A

toward the zenith and the spectral regions 7280- 2

7410 A and 8580-8830 A covering the OH (8-3) 25 ' 3

band and the 02 (0-1) atmospheric band were N2

0P.

recorded. The OH (8-3) and 02 (0-i) atmosphericband were scanned in 32 and 12 a, respectively. < 2"

with the I-m and 1/2-m Ebert-Fastie spectrometer __.....

Instruments at a spectral resolution of 1.5 and2.5 A_,, respectively. Each scan was recorded on N 2c 20

tape and spectra were obtained each 30 min for OH Z1 0-.6-2aod each hour for 02 by summing individual scans. Iw-vxmples of the OH (8-3) band spectrum are given by '0

4yr.b6 [1984]. Figure 2 shows an e;:ample of a

i-hour integration of the 0 2 (0-1) atmospheric ,bands at a resolution of 2.5 A. 5 -

Absolute intensity calibration was performed

in the field using a standard lamp and a diffus-log screen. The OH rotational temperature has 0been calculated using the relative intensities P1 860C 7070 8800

(2), P1 (3), P1 (4) and P1 (;) lines and Kvifte's WAVEENGTH A

method [Kvifte, 1959]. The 02 (0-1) atmospheric Fig. 2. An airglow emission spectrum typical ofbend temperature was estimated by comparing with the 02 (0-1) atmospheric band spectra used in thesynthetic spectra (Henriksen et al., unpublished intensity and temperature calculations. Themanuscript 1983). spectrum was )btained from a 1/2-m Ebert-Fastie

Probable error in the rotational temperatures spectrometer with a bandwidth of 2.5 A by summingcalculated from relative intensity ratios is individual 12-s scans for one hour. The location

estimated to be 1 10 K. The error in the absolute of auroral lines and bands is indicated.

9150 Myrabt et al.: Brief Report

77ing season [Myrab6 and Deehr, 19841. Thus, the a __

wean value for this 18-hour period was approxi-mately 25% higher than the seasonal mean.

A search for other ground-based observations,unfortunately, yields very few recent measurements IDgiving absolute intensities. Measurements byMisawa and Takeuchi [19781, Misawa et al. [1980)and Takeuchi et al. [1981] are all in arbitrary 0

units. Ground-based absolute measurements avail- 0able for comparison after 1960 are by Berg and 0 o 30 40 11 6 It so

Shefov [1962], Broadfoot and Kendall [19681 and LATITUOE (De.ss

Shefov [1971]. Absolute emission intensities of Fig. 4. 02 (0-0) atmospheric band intensity ver-the (0-1) band were also estimated by Wallace and sus latitude. G signifies the intensity calcu-Hunten [1968] and by Evans and Llewellyn [1970] lated from ground-based observations of the (0-I)using Broadfoot and Kendall's spectrum. band using a Franck-Condon factor of 20 and R

From the above measurements, we have calcu- signifies rocket-borne observations of the (0-0)

lated the expected (0-0) band emission intensi- band.ties by applying a Franck-Condon factor of 20.The resulting ground-based calculated (0-0) bandintensities are plotted versus latitude in Figure (See Figure 3, or Noxon [1978], and Weinstock4 together with available absolute measurements [1978]). A mean value of 4.1 kR with a standardof the (0-0) band emission intensity as obtained deviation of 2.7 is found for the rocket measure-from rocket flights. The values calculated from ments while the ground-based estimates have aground-based observations are marked with a G, mean value of 8.4 with a standard deviation of 1.1.while the rocket measurements are mArked with an It should be mentioned that absolute measurementsR. The rocket measurements are those by Packer from the ground of the (0-1) band prior to 1960[1961], Megill et al. [1970), Witt et al. [1979] have been excluded from these data as it is feltand Deans et al. [19761. that they may contain a systematic calibration

Although there seems to be little or no varia- error (Barbier [1956), 2000 R; Dufay 11958), 1500 R)tion with latitude, the (0-0) intensities calcu- compared to later results (see Figures 4 and 5).lated from the time-averaged ground-based obser- The value of 2000 R by Barbier (1956[ and 1500 R byvations of the (0-1) band show a small scatter Dufay [1958] would only compound the discrepancy(20%) at about twice the value of the rocket between the rocket and ground-based observations.observations, although the latter are much more A rocket experiment to observe both the (0-0) and

scattered. Correction for extinction and re- (0-1) bands has been carried out (F. R. Harris,smittance of self-absorbed (0-0) band emission is personal communication, 1983) and preliminarynot included in the ground-based calculations The results using the generally accepted Franck-Condonextinction should be about 10% for the appropri- factor of 17 to 22 indicate that the differencesate wavelength region [Allen, 1963], while the between the ground-based and rocket data may becorrection due to reemittance is found to be reconciled and much higher intensities can occurabout 10% [Wallace and Hunten, 1968]. These on a time scale less than that of the time-averagedeffects should, therefore, cancel out. ground-based data.

The scatter in the rocket data may be ex-plained as a result of the short time interval of 0? (0-I) Atmospheric Band and OH (8-3)each measurement of the highly variable airglow Band Rotational Temperaturescompared to the time-averaged ground-based data. Temperatures deduced from the 02 (0-1) atmo-

spheric band rotational structure and the OH00_ (8-3) band PI lines, are shown in Figure 5.

These are 1-hour and 30-min averages for the18-hour period in the same manner as for the in-tensity data. The overall temperature differencebetween the 02 and OH band temperatures is seen to

600 be small and mean values of 254 K t 3 K and249 K t 2 K are found tor the respective emissions.

O(8-3) The two temperatures may also be seen to follow asimilar trend, which is reasonable since they

Z1.500 originate in the same atmospheric height domain;zthe OH intensity maximum is usually found 3-8 km

lower than the 02 emission [Moreels et al.,1977; itea. 197 9; Wataoabe et al., 19811.

400 OO Although considerable variation in both the 02

and the OH temperature takes place during the ob-serving period, nn wavelike pattern that could be

20 22 24 02 04 06 08 10 12 attributed to gravity wave effects as found byUNIVERSAL TIME (HRS) Noxon [1978] is seen in this particular data

Fig. 3. The hourly and half-hourly mean zenith collection. The variations in both the 02 and OH

intensities, at Longyearbyen of the 02 (0-1) temperatures iy rather be interpreted as being

atmospheric band and OH (8-3) intensities plotted due to transport and mixing phenomena which seem

as a function of UT for January 11/12, 1983. to govern the dynamical behavior of the polar

Straight lines are drawn between each mean and winter atmosphere at mesopause height )Myrab6,

missing data are indicated by a dashed line. 1984; Myrab6 and Deehr, 1984].

Myrab et al.: Brief Report Qj5i

78

280 Acknowledgments. Financial support for this

02 Temoersure - research was provided by National Science Founda-27C Ition through grants ATM80-127i9, ATM82-1462,

-01 Temeraturee ATM82-00114 and ATM-8313727 to the Geophysical

260, t-" - Institute of the University of Alaska and bythe Royal Norwegian Council for Scientific and

;j -.. 0 Industrial Research through a fellowship grant toone of us (H.K.M.). The observations on Svalbard

,240 f J are made through a cooperative effort involvingthe Universities in Fairbanks, Troms6 and Oslo,

"3 P,oble Erro 02 with the help and cooperation of the Great230 Norwegian Spitsbergen Coal Company and the Nor-

i s00 ErroO " I wegiao Polar Institute.

220 2 24 02 04 0 08 10 12 The Editor thanks the two referees for their

UNIVERSAL TIME (HRS) assistance in evaluating this paper.

Fig. 5. The hourly and half-hourly mean tempera- Referencestures from the R and P branches of the 02 atmo-spheric (0-1) bands and the OH (83) band P1 lines Allen, C. W., Astrophysical Quantities, p. 122,observed in the zenith airglow at Longyearbyen on Athlone, London, 1963.January 11/12, 1983. Straight lines are drawn Barbier, D., The airglow, in Vistas in Astron-between each half-hourly and hourly value, respec- only edited by A. Beer, p. 929, Pergamon,tively. London, 1956.

Berg, M. A., and N. N. Shefov, Emission of the hy-

droxyl bauds and of the (0,1) 18645A atmo-Assuming a reasonable difference in the peak spheric band of oxygen in the nightglow,

height of the 02 atmospheric emissions and the OH Planet. Space Sci., 9, 167, 1962.(83) band of 5-6 km, the mean temperature differ- Berthier, P., Etude spectrophotometrique de laence AT of 5 K gives a positive temperature gradi- luminescence nocturne des bandes des moleculesent of I K/km for the 90 to 95-km region. This OH at 07 atmospherique, Ann. Geophys., 12,confirms previous findings of a shallow polar 113, 1956.mesopause with a relatively small positive tempera- Broadfoot, A. L., and K. R. Kendall, The airglowture gradient in the 90 to 95-km region [Myrabi, spectrum, 3100-10,0000, J. Geophys. Res..1984] for winter solstice conditions. 73, 426, 1968.

Deans, A. 3., G. G. Shepherd, and W. F. J. Evans, ASummary rocket measurement of the 02 (b

1' X

3:-)

atmospheric band nightglow altitude diftibution,The (0-1) atmospheric bands in the night Geophys. Rea. Lett., 3, 441, 1976.

airglow have been observed from Longyearbyen Deehr, C. S., G. G. SivJee, A. Egeland, K.(78*N) during winter solstice conditions using Henriksen, P. E. Sandholt, R. Smith, P. Sweeney,spectrophotometric equipment. The observations C. Duncan, and F. Gilmer, Ground-based observa-reported here were confined to an 18-hour contin- tions of F region associated with the magneto-uous clear-weather observing period from 1900 UT, spheric cusp, J. Ceophys. Res., 85, 2185, 1980.January 11, to 1300 UT, January 12, 1983. The raw Dick, K. ., G. G. Sivjee, and H. M. Crosawhite,data were recorded with a spectral resolution of Aircraft airglow intensity measurements: Vari-2.5 A, which resolved the rotational structure ations in OH and 01 (5577), Planet. Spaceof the P branch of the 02 (0-1) atmospheric band. Sci., 18, 887, 1970.

A mean 02 (0-1) atmospheric band intensity of Dufay, M., Sur les intensities des bandes445 ± 40 R, ranging from 350 R to 5.3 R, is found. d'emission du ciel nocturne dens leThe mean temperature deduced from the 02 (0-1) proche infrarouge, C.R. Hebd. Seances Aced.atmospheric band rotational structure is 254 K ± 3 K Sci. Paris, 246, 2281, 1958.compared to a mean of 249 K : 2 K for simultaneous Evans, W. F. J. and E. 3. Llewellyn, Molecular oxy-measured O (8-3) band rotational temperature. gen emissions in the airglow, Ann. Geophys.,The mean temperature difference between the 02 26, 167-178, 1970.(0-1) atmospheric band and the OH (8-3) band Feldmann, P., Auroral excitati-i of optical emis-yields a positive temperature gradient for the sions of atomic and molecular oxygen, J..90 to 95-km region of I K/km, assuming a reasonable Geophys. Res., 83, 2511, 1978.height difference between the two emissions of 5-6 Fraser, P. A., W. R. Jarmain, and R. W. Nicholls,km. Vibrational transition probabilities of dia-

All available time-averaged ground-based ob- tomic molecules: Collected results, i, N2,

servations of the (0,1) band intensity were used CN, C2, 02, TiC, Astrophys. J., 119, 286,to calculate the intensity of the (0,0) band. 1954.These data were plotted together with available Gault, W. A., R. A, Koehler, R. Link, and G. G.rocket observations of the (0,0) band as a func- Shepherd, Observations of the optical spectrumtion of latitude. No latitude dependence was of the dayside magnetospheric cleft aurora,observed, and although the rocket data were half Planet. Space Sci., 29, 321, 1981.the intensity of the ground data on the average, Greer, R. G. H., E. J. Llewellyn, B. H. Solheim,the scatter was considerably larger. This indi- and G. G. Shepherd, Observations of thecates that a further correction to the ground- optical spectrum of the dayside magneto-based or the rocket data may be necessary and spheric cleft aurora, Planet. Space Sci.,that the 02 airglow is highly variable on a time 29, 321, 1981.scale of hours. Henriksen, K., G. G. Sivjee, C. S. Deehr and H.

9152 1Myrabi et al. i Brief iteport

79K. 4 rab6~ Ground-based observations Of 02 upon night airgiow temperatores. Geophys.(bit - KIZ*) atmospheTic bands in high Res. Lett., 5, 25, L978.latifude au~oras, J. Geophys. Res., has been Packer, D. 4., A ltitude of the light airgiow rad-accepted, in press, 1984. Lations, Ann. Geophys., 17. 67, 1961.

Ismail, S., and L.. U. Cogger, Temporal variations Shefov, 14. 4., Hydroxyl emission of the upper at-of polar cap 01 5577A airgiow, Planet. Space mospntere. UV, Correlation with the molecularSci., 30, 865, 1982. oxygen emission, Planet. Space Sri., 19, 795,

Kvifte, G., Nightglow observations at As during 1971.the 1.G.Y., Geophys. Pubi. , 20, 1, 1959. Sivjee, G. G., and C. S. Deehr, Differences In po-

Megill, L. R., A. H. Despain, D. J. Baker, and K. lar atmospheric optical emissions betweenD. Baker, Oxygen atmospheric and Infrared at- mid-day cusp and night-time auroras, Inmospheric bands in the aurora, J. Geophys. Exploration of the Polar Upper Atmosphere,Res., 75, 4475, 1970. edited by C. S. Deehr and J. X. Roltet, Hingham,

Meinel, A. B., 02 emission bands in th, Infrared Mass., 1980, p. 199, D. Reidel.spectrum of the night sky, Astrophys. J., 112, Sivijee, G. G., K. A. Dick, and P. D. Feldman, Tem-1464, 1950. porai variations in the aighttime hydroxyl

N1isawa, K., and 1. Takeuchi, Correlation among 02 rotational temperature, Planet. Spate Sri..0-i) atmospheric band, 08(8-3) band and 101) 20, 261, 1972.5577A line and among ?1(2), P1(3) and P1(3) Takeuchi, I., K. Misawe, Y. Kato, and 1. Aoyama.and P1 (4) lines of 0H (8-3) band, J. Atmos. Seasonal variations of the correiation amongTerr. Phys., 40, 421, 1978. nightglow radiations and emission mechanisms

Misaws, K., 1. Takeuchi, Y. Kato, and I. Aoyama. of OH nightglow emission, J. Atmos. Terr.Correlation between 02 (0-1) atmospheric band Phys., 43, 157, 1981.and the NaD lines, J. Atmos. Terr. Phys., 42, Tarasova, T. 4., Night-sky emission-line latensi-179, 1980. ty distribution with respect to height, Space

Moreels, G., G. Megie, A. Valiance-Jones, and R. Res., 3, 162, 1963.L. Gatttnger, An oxygen-hydrogen atmospheric Wal-lace, T,., and D. M. Huntes, Dayglow of themodel and its application to the OH problem. oxygen A band. 3. Geophys. R~es. , 73, 4813,3. Atmos. Terr. Phys., 39, 551, 1977. 1968.

Mullen, E. G., S. .. Silverman, and D. P. Korff, Watanabe, T., M. Nakamura, and T. Ogawa, ROCretNightglow (557.7 nm of 01) in the central measurements of 0-,, atmospheric and OR Meinelpolar cap, Planet. Space Sri., 25, 23, 1977. bends in the airglow, J. Geophys. Res., 86,

Myrabi, H. K., Temperature variation at 2esopause 5768, 1981.levels during winter solstice at 78*N., Weinstock, J., Theyry of interaction of gravityPyr t - S acS.. 32, 249-255, 1984. waves with 3 U airglow, J3. Geophys.

.Irb21 ! n .S. IDeehr, flid-winter hy- R..s., 83, 5135, 1978.droxyl night airgiow emission Intensities in Witt,_G C.*. Stegman, B. R. Soiheim, an E. J.the northern polar region, Planet. Space Sri., t.iewellyn, A measureme nt of the 01( +

32, 63-21, 984.K -) atmospheric band and t he01()

My'ab6, H. K., C. S. Deehr, and G. G. Sivjee, gr een line in the nightglow, Planet. SpaceLarge-amplitude nightglow OH (8-3) band inten- Sri., 27, 341, 1979.sity and rotational temperature variationsduring a 24-hour period at 78 .N, J. Geophys. C. S. Deehr and G. J. Romlic, GeophysicalRes., 88, 9255, 1983. Institute, University of Alaska, Fairbanks, AK 99701.

Nicholls, R. W., Franck-Condon factors to high K. Henriksen, Auroral Observatory. Univ~rsttyvibrational quantum numbers, V, 02 band systems, of Troma6, 9000 Troms6, Norway.J3. Res. Natl. Bur. Stand.,* Sect. A, 69, 369, H. K. Myrabi, Norwegian Defense Resear'r

1965. Establishment, N-2007, Kjeller, Norway.Noxor, J3. P., Observation of the (bit -

a1 t, ) transition in 02, Can. J. Phfs., 39, (Received June 23, 1983;11,1961. revised May 23, 1984;

Noyon, J. F., Effect of internal gravity waves accepted >fay 24, 1984.)

-4-9~~w~ W r J acte iii KW XNerksmn, Kid-wintert Istti at tUs -figt, atrpl 02(-1~pw~ band

state athihlttta. stOPhys US. 1664 (lOS)

81

thAI Nk ;c ;lpq4yS: 'A-' ctdIARli' c . . , AFh-- F R

iD-mTER NTENST ElS 77 0 E qGHT AIRGLOW 2 A - A'hSERC qAhI EMIESUjN A- 11,29 ;T 7 :FS

9. ,vrabe

-.wegan lefence Research Estanlishment, Keller, Norwao

S. Deen 2. Romick

eersity or Alaska, FairbanksK. ienr:csec

urocral 'bservat-or n1,.ersltv or 7roms, Norway

Abstract. Absolute lntensities or the 22 ('-l) fore give an :naependent easre of odc gn

atmospheric band nognt airgiow emission have cartations in cne 90-1,: km region. "hIch procan-been observed Prom near Longyeary en on dent iy reflects toe oheniuai and dynamical nenaviorSpitsbergen 78.4'S latitude. b'0 longitude, of the or- Invest part of the thermospbere

geographic) during a 2-month period around winier Measurements of he (0 (-i) at.osphertc band

solstice li982-i983). Intensities ranging Iron eission may also contribue to the understanning:10 8 to 1590 R wih a sean of 577 9 n 61 6 are of uncertainties in the chemical reations in-

observed. There is no ulear maximum or sinimum volved In the atonIc and enIerular oxygen emissionaround solstice. A senidiarnal tide osponent n the night airglow Wraight, a9 Rates. 978;

givtg rise to a 25-30: intensity variation of reer eta., 1988 Isnail snd losger, 19:..the 22 (0-i) atmospheric emission is present in The purpose of aths paper Is to report a tthe data. The maxiem- and minimum are foand to olacuss recent observations or the - -e

occur at about 0400 and 1000 local time. restec- asospherlc hand night airgIow emission. Thesetively. O diurnal tide component larger Ihan observatlons were made near longyearhyen, Aaihardt3-u% is present in the data. The day-to-day at 79'i daring 38 ays arond oe 1982 winterand short time variations both show a quasi- solstice.egular wane pattern which San be associated

cith gracity waves. 2. Observacions and Data Reduction

I. Introduction During the :982/Z1983 Multi-qattonai IvalbaraAuroral Expedit io., MNSA ) Deehr e t ol., .990'.

NIght airglow observations provide a powerful emisscon spectra of the zenith sky. ncainhgand relatively simple scans of studying the the -30 1-1) atmusyteric band emission, were

dynamics and chemistry of the upper atmosphere. taken near Longyearhyen on cent Spitsbergeniround-based observations are important for study- 8..'N latitude. YR longitude, gengrapnic,.Ing snaller-scale dynamical phenomena hbecause From the end of November 1982 throughout January

rocket measurements only gine a snapshot in time 1983, a 12 m Ebert-Fastie spectrvyhotometer wasand satellite ceasoresens are asually Integrated used to obtain the epectra. Soe inportantonee a large area at widely spaced time intervals, parameters for the instruent is given In Toble

Around-based -nignt airglow observations In the I. For friber reference, a slmular 1: n_ oto-polar regions ar e e r compared to lower sent is nscribed mre d uly in the papers bylatitades. This is aiso the vase for rocket ioeee et al. (19721 and Ramlrk [926], Themeasorements. tn nose recent papers by Myrab6 spectral region from 9340 to 9870 was scanned

et al. [i983, 19 84a], Myrabt [19841 and Myrahd during 12 s with a bandwidth of 7 i, ettept for

and Deehr (1984, ground-based measrementI of a brief lb-hour period when the 8 586-930 g regionLte OH night airglow have been reported. OF! was recorded with a 1.5 A bandwidth. The purposeemissions originate in the 95-90 ha region (i.e., of the high-resolution recordings was lainv lv

hear the sesopause). Measurements of the )17 get an accarate rotational temperature lot the2577 A Line in inn lower thermosphere (i.e., -107 95-ks lene for comparison and chec vcth 31-

tm1

have been reported by Sandford (1964], Muilen obtaIned temeratures. A brief repot on -hlc

et al. (1977) and :.mail and 7og8r [19921. iN um been given by yrabe ec, . -n).tor e

high-laticude, ground- ased observations of the resolution data, spectra were obtainednight airgLow )1 (0-i) atmospheric hand near -95 averaging 150 individual scans yielding one spec_

kh altitude (Deans et at. .6; Aiti et at., trun every 30 sin.

1979] tune been nublinhed. exceyr for a brief Figure t gives an enample of two [-mIn spec-report by Myrab6 et al. [13984b]. tra, one during quiet geomagnetic condltions snd

The Ov 20-i) atmospheric band Is no rmally one during auroral activity with relatioely

Pound to foilow the intensity variations of the strong molecular emissios. The di.ference

1ol 5577 A line (Dufav 1959, and sh.old there- between the spectra is toe presence or aurora.emissio in the bottom panel. last p-ontneh

among these are the 01 9--n S lne. the 0<.

Copyright I-6 by the Amertcan -eophysical Union. band, the slightly enhanced 02 (7-i) acmosneric

hand and the srongest lihes of the 01 r)s'sPaper number A8652. buyP) nitlplet.

1,.s8-722!, 6/7bA-6b2h530 he of toe main problems wi th meass-vg night

82

"yrahe er at.: Midwinter intensities it ,ignt kirglw_

TABLZ 1. Zhar,scteriatics of Specirsoeter uscertasntocs...nc .0..:; ncertao=:c :._poton noise an reading off eacha1 d :zteo::'vis than 3.,

Component Derail

3. Results and DiacusstonType of spectrometer 1/2 s Ebert-FaatieField of view 6 6' 3.1. Short Time Variationo

Dark count rate 1-3 count/s The typical short time pattern Or one trnenc-Emtrancelexit slit height 65 ox ty variations of the 0. (0-I) hand emission isSpectral dispersion somewhat similar to that found for the MH enis-

(current experiment) 14 S/ox Sian in the polar region [Myrab6 et 01., .983;Spectral sensitvity Myrab and Deenr, !9,oif. intenstv uar:aions(with 7 5 resolution of of I or sore within a few nours areand current experiment) -5 R/count at 86OO common. The intensity versus time plots often

Detector type Hamamasu R943-O Ga show wavelike cyclic changes which shift fre-di. tube quently only after a few cycles or are inter-rupted by peaks or step changes. it is rather

usousual to see longer periods (8-12 hoors) w'ithlow amplitude, smooth variations as is often thecase at lower latitues [rthier. 195.; M'sac.

airglow features in auroral regions is the con- and Takeuchi. 19-8].tamination by aurora. if the emssions are not Figure 3 shows a typical example of the inten-excited by auroral particle bonbardment, overly- sity warlacion with time during a 19-hour period

Ing auroral features nay normally be filtered Out from December 25-.6, 1982. It seems reasonabieor accounted for if a high enough spectral reso- to believe that these variations are associatecltion is used. If the airgIow emission also with gravity waven. Normaly, lt"ie or noOCcurs in ibm aurora. as is the case for the [01] correlation is observed between he Intensito5577 A lIne and the 0, (0-1) atmospheric band,there is no direct way of knowing if the photonsoriginate In the aurora or airglow. A simulta-..e..u .msitoreg of pur . o.... far .es' such (RELATIVE INTENSITY1as the 3914 or 4278 A N ISIS hand, could beused to select tine pe r.od devnd. of asirtal swemission features. This method was used by P0 m T M71Pfuilen et .. 119771 and Ismil and Cogger 11982)to isolate the night airglow component of the1011 5577 k line. Toe problem of esatablishing a 360/14:00threshold intensity for auroral incidence isalleviated by using a spectrophoometer with asufficient resolution to record the night airglowand the aurorai reference emissions is the sanespectra. At very high latitudes, atomic lines .1dominate the auroral spectrum for most of thetime lfivjee and Deehr. 1980; Gault et aI. 1981;Myrab6 and Deehr, 1984]. This makes the nightairglow component of the 02 (0-1) atmosphericband much easier to isolate and far less sensi-tive to auroral emission than the [01] 5577 4line.0 ''' s'

The spectral region shown in Figure I includes esa., -s.*the S1 8446 auroral line as an auroral referenceemission in addition to the 02 (f-I) band. An 1|auroral component in the 0, atm hand of 2-3Z In

i ndicated bp the 01 .3446 line rining high above 3013t he 08 Sines, by a factor of 2 o, .0nre. For36/40caculating the absolute intensities of the 0

2(0-I) atmospheric band, we have thus only select-ed spectra with no auroral emissions and only

during clear and stable weather conditions.

Weater conditions were checked visually byviewing standard stellar sequences. The result-ing daily distribution of the spectra is depictedIn Figure 2.

Absolute calibration of the spectra werecarried out in the field using a standard source. 0The absolute intn.ities..werecaIculated by usIng 40the area nder the band emission and represent

only tbe night airglow component of the 02 (0-I0 WAVELENGTH (A)atmospheric band. The uncertainty in the band Fig. 1. Examples of spectra averaged for 30 minintensity measured from each 3(-mn spectrum Is during quiet conditions (top) and during oderateestimated to be 25%. mainly due to calibration auroral conditions (bottom).

83

1686 Syrane taI e .: e Mwcer lntesIctien cc Nignt lirg Ia

<NUMBER OF O ABA A or correspondence tnan is actua.ly cbservedshuld be expected if e intensity -art aicosin the two night airglow emissions large r

" caused by the same gravity waves traveling tom

below throughout the menopause region o at -eastabuse 95 k.. The OH temperature and D2 ()-I]

i p~~~~~ro,Imaely thes a:e eodad le40 ; 1 oal a~~~ a ~w L~ ,= ,.a

7JPeriod$. 1.i 1. normally not seen, which s.p-

20 ports the view that the wavelike variations of

the two airglow emissions night have two differ-~ent rigins,

If gravoty wanes were to breac cn the 95-45neight reg-on, they soid .isipate sost fI their

01 09 15 21 energy here and cause a very complicated and Its-UNIVERSAL TIME (HOURS) tarbed density and temperacure pattern m Holton,

Fig. 7. Himtogram showing the distribution in UT 1979). Gravity waves due t.o auroral activity

of the spectra used is oaicuiaring the absolute above the 100-km region travel downward and could

inrensities of the 02 (0-1) atmospheric bad. affect the ,n 0-I) emissio more strongly thanthe OH emission [Frederick, 1979]. 'c this basisa hypothesis may be put forward chat variations

eariations of the 02 (0-1) band and those of the in H emissions are affected by gravity eatesOH emisnions which are thought to be ahout 5-10 from below, i.e., from lower atmospheric andkm lower in the atmosphere (i.e., at the meso- tropospheric phenomena, wnile variation in the )2pause), even if a time shift is introduced (see (0-1) atmospheric band .re affected by gravityalso Myrab6 et al. (1984b]). An eample of two waves originating from above, i.e., from auroral-subsequent spectra, including both the 02 (0-1)

bend and the OH (6-2) P branches, is shown inFigure 4. A 70-80% decrease is seen in the 02

(0-1) band Intensity, while no change is seen in RELATIVE INTENSITYthe OH intensity. Simultaneous observations ofthe OH (6-2) intensity and temperature and the 02 / ne-z.(0-1) band intensity are given in Figure 5,

showing the lack of correlatios between the two P IV 3 P4I) PHI Pil 6P1l P171emissions.

According to Weinstock ]1978],remperature andintensity variations of the 02 (b 0) airgiow /150emission need not alwaye t urelate sring theIpassage of gravI ty waves. Observatios by Noo(1978] show eamplea of this. The intensityvariation may be both phase-shifted and have amuch smailer amplitude as compared to lthe temper-

aturesf ficrtua tions. A complete corelation&Abetween OH temperature or intensity and 0

2(b Z+) intensity should therefore not be en- i

petad during passage of gravity waves through r

both emission layers. However, a higher degree 7 y '7

1000-6ii~Y ~350/15:30

250. 64 6 600 67 87"0

13 15 17 19 21 23 01 03 WAVELENGTH (A)UNIVERSAL TIME (HOURS) Fig. 4. Two cansecsrive 30-mis spectra oF the

Pig. 0. Typicai esapie of the intensity Carla- 8340-8740 A region showing an almost I00 de-tion of the 0 [0-i) atmospheric band am obtained crease in the 0

2 (0-1) band, while the OH emis-

during a 14-hour period from December 25-26, 1982. sion is left unchanged.

84

MyrabO e al.. o1dwonter IntensIe. af Night Airgloo "

R I E following tnhe method at Petitdldier ahd Teotlodom9 o. The best fit curve gives maxlou and

oIniouc of the emission intensity at ho) .T OhdA e2200 2T. Universal time and local solar time

p differs ool by about 20-30 ml, which places the/ posicion of the maxima and mlnima as expressed i.

* / local solar ime (ILST) at approxicmately J400 and075T

, , A ee. in Figure 6, there is n00 ign of ainal variation. test this further, Fourier

o 5 0'-1 Am' analpsis of the data dam. ndertuken. ho di rol%, component larger than n3-Z. appeared ih the

Fourier spectra. Mllen v al. M197l 1oune nodiurnal variation in the Oi 55,7 0 nIghtgIow at

Thule (77*N) Larger than tSI.Myrand 1984) and Myrabe and Decor (1984;

1 14 ' ' 20 22 '2 02 . oa found no diurnal variatlon "n the OH rotationa.temperature and t toe Of lntensity largen thn

VERBAL TIME (HOURS) e t and t5%, respmctinely. The absence of a

Fig. b. The 15-mn average intensitles of the diurnal variation in the 02 (0-I) atmosphericOH (6-2) band and the 02 (0-1) acm band obtained on band intensity is therefore not surprising. ADecember 17, 1982, from Longyearbyen. negligible diurnal tide effect in the 80-Ill k

region at extreme high latitudes seems thereforeto be established. This is contrary to the model

ly related phenomena. Seen from a dynamical calculations of Forbes (19821 but agrees withpoiot of view, chis wo.ld divide the polar upper Beer [19751 and Spizzichino (1970]. According toatmosphere in two height regions, i.e., below Teitelbaum and Hlamont 11975] and Spizzichino85-90 a and above 90-95 hm, dominated from below )1969], the first diurnal mode effectively inter-and above respectively, by breaking gravity waves ac-. with gravity waves. Averaging over severalwit a turbulent-free mesopause region in between. days cancels the diurnal code. The short timeThis picture aiso fits with rocket measurements behavior of the intensity variations also (seefrom And6ys (b9*N) of the turbulence in the section 3.1) indicates strong gravity wave actl-80-100 kms region as recorded during the 4rA.FWI vity which, in connection with the above hypothe-csmpaign 1983/1984 (T. A. Hlix, private comunl- sis, should explain the absence of the first

cation, 984). diurnai tide node.Mullen etal. (1977] also failed to find any

3.2. Diurnal and Semldiurnal Variations semidi r.al tide component at the 100-ha level.However.,recent OH observations from Spitsbergen

The large, irregular intensity variations with [Myrab, l9Hi; Myrah and Deehr, 1984] haveperiods less than 24 hours mask any diurnal or revealed a dominant semidiurnal tide componentsemi diurnal variations in any single 24-hour both in the rotational temperature and emissionperiod. in order to bring out chese daily varis- Intensity data. The semidiurnal variations intioss we have used a superimposed epoch method, the 02 (01l atmospheric hacd intensity found

employing all of the absolute intensity data for here confirms the presence of a dominnt semi-the 20-day period between December 6 and December diurnal tide at extreme high latitudes, at least6, 1982. Figure 6 gives the result, showing a in the 85-95 km region. However, the amplitudesemidiurmal intensity variation of the 02 (0-I) of the tidal component is slightly weaker than

atmospheric band with an amplitude variation of normally found for the 02 (0-i) hand and theapproximately ±15% from the average. a sem- [01] 5577 A at mid and high latitudes (Brentondiurnal tide curve has been fitted to the data and Silverman, 1970; Petlitdidier and Teitelbau-,

!I R LognGERIemn l'I 1162111tll

S800-

100- " ml

go N 09 IL 0. IS 6-I1UNIVERSAL TIME (HOURS) I Issas

Fig- 6. Hourly man intensities superposed for ,sUmcV

each 4-hour period in UT for the 20 days from as - As - cDecember 6 to December lb, 1982. The semldlurcal "an scumstrend Is indicated by a solid line with the Fig. 7. Sin-hour average 02 (0-1) band absoluteminima and maxima marked by short arrows. Longer intensities at 78N during the 1982/ 1983 CLNSAarrows mark local geomagnetic and solar noon. Expedition. Winter solstice is indicated.

85

1068 Myrabe et 01. -idwiter intensities of Nfh: Airglow

19771. The time for the maxima and minima fits nounced minimum in the emission of the -'very well with the calculations of Petididier atmospheric band during winter solstice. Sut onand teitelbaum [1977], which predicts a maximum the other hand, aorthern [0I] 5577 A hemisphereintensity of the green line at 100-ka height night airglow data from Thule (77'N) during threeto occur around 0130 lo-sI time for winter condt- 1.- w1 .. ...... .. , ,n--, -tiens at 45'h. A shift of the emitting layer of no clear minimum around winter solstice. Indeedh-7 k, i.e., around 93-94 km which is a reason- the results by Ismail and Cogger from the north-able height fur the 07 (0-1) hand (Deans et al., e hemisphere are not conclusive. Thus the

1976], leads to the maximum occurring 2 to 2 1/2 result from Ismail and Cogger showing the clearhours later [Pet didier and Teitelbaum, 1977]. minimum in the southern hemisphere where there isThis in very close to the maximum seen in Figure no midwinter sudden stratospheric warming, indi-6. The aiculato s of Petitdidier and caes ht the lack of seasonal effect In theTelteaum also agree very well wIth experimental northern hemisphere may be due to the effecz Il[01] 5577 £ data from mid-latitudes [Brenton and the stratospheric warming.Sllverman, 1970]. The day-to-day dynamical hehavior in or 0'

(1-) atmospheric data shows the same type or3.3. Seasonal and Day-to-Day Variation pattern as seen in the (I] 5577 A night airglow

from the northern polar region [Mullen et al..Six-hour mean intensities have been obtained 1977] during winter solstice, i.e., large Irrego-

by averaging 30-in spectra. A minimum of three Iar variations. This type of behavior therefore30-mn spectra were used to calculate each 6-hour seema to be common both for the OR, [011 5571 Imean. Normally, 6 to 12 spectra were used. line and the 02 (0-1) atmopheric band night air-Figure 7 shows the resulting intensity versus glow in the northern polar region. It is differ-tim plots. The pattern in the day-to-day vara- ent from the normal morphological patter of theclone is similar to that found for the (001 night airglow at lower latitudes and is therefore5577 A line at Thule by Mullen et al. [19771 and reasonable to connect to the dynamical situationthat by ismail and Cogger [19821 from the ISIS 2 of the polar atmosphere in the 80-119 km regiondata. It is characterized by irregular peeks (Myrab6 et ai. , 1983,1984b; Myrabd and Deehr, 1984.with intensity variations up to a factor of 5 or We think chat these Irregular variations aremore in a few days. The largest 6-hour mean partly connected to temporal and spatial varia-found in the 02 (0-1) atmospheric band is 1125 X tions in the occurrence of gravity. waves andand the lowest value is 165 R with a seasonal their interactions with the zonal winds in theaverage of 570 R ± 60 R. The seasonal average stratosphere and the mesosphere. Model calcula-in slightly larger than that reported for a tions by S. Soloman (private communication, 1984)shorter period by Myrabd et al. (1984bi. using a chemical-dynamical model including gravl-

In the data from Thule of the (01], 557 5 ty wave parameterization (Lindzen, 1981; Holton,line, Mullen et al. (1977] found a correlation 1983] have shows how the (OI] 5577 A night air-between interplanetary magnetic field polarity, glow spring maximum at high latitude reported byi.e., "away" or "toward" sectors corresponding to Cogger et al. (1981) can be reproduced in the3

x positive or negative and By negative or posi- model as a direct result of the seasonal varI-

t iv, and increasing and decreasing, respectively, tion in the propagation and breaking of gravity[1] 5577 5 nightglow trend. Such a correlation waves and the associated diffusion of odd oxygen.Is not clear in this data set. as Seh as in- Thus the maximum in the oi] intensity is pro-creasing and decreasing trend is found for posi- duced as a result of the final warming of thetine interplanetary magnetic field polarity. It stratosphere and corresponding cooling of themight be pointed out that day 344-345 (December mesopause region. Similarly, one would expect10-11) and January 1-I0 were disturbed periods smaller-scale and local gravity wave blocking byand that both these periods show pronounced conal winds to result in shorter periods coolingmaxima in the 02 (0-1) intensity, although the of the mesopause and possibly enhanced nightltter may be partly associated with a strato- airglow levels in the altitude regions affected.spheric warming event. Such enhanced levels of the night airglow

From the ISIS 2 data, Ismail and Cogger [19821 resulting in gravity wave interaction and su ensee an enhanced level of polar [01) 5577 A night- stratospheric warmings or minor warmings couldglow emission during stratospheric warming events. slso explain the lack of a clear minima duringUnfortunately, the time intermal covered by our solstice in the northern hemisphere night airglowdata set only contains one stratospheric warming as compared to southern hemisphere where suddenevent around year's end, Is the satellite Jat warnings during midwinter do not occur [labitzke,for the polar area, there is a warming tendency 1977].on January 6-8, 1983, that only reaches the 4-mbar level [flujokat et al., 1983; Myrab6 et al.,1984b). Unfortunately, due to severe stores and 3.4. 0, atm [0-1] Intensitiesfull moon conditions, we have a large gap in ourdata just in this period. However, a very steep The seasonal mean value of 570 R ± 60 R foundincrease in the 07 (0-1) atmospheric band isten- here for the 02 (0-1) atmospheric band alsosity is seen between January 2 and January 10 appears rather large as seen from a photochemicalwhich may be associated with the stratospheric point of view; i.e., the polar cap is for most ofevent of January 6-8. the winter solstice period without solar irradi-

Ismail and Cogger [1982] found a pronounce d ante at the 95-km level. This should cause

minimum in the (01 5577 8 line In the soathern little or no phuoa*cliaion of 02 into escitedhemisphere during winter solstice. A similar states (Moreels et a., 1977]. The presence of atendency should be expected for the 02 (0-1) strong 02 (O-1) atmospheric emission thereforeatmospheric hand; however, we do not find s pro- requires other sources for exci ing the 02 mole-

86

Myrah et al.: Itidwinter Uens, ies )f N:fht A irglow ,

cules. Vertical and horizontal transport of add sent larger cnan t3--1 is present is toe data.oxygen is such a candidate. The day-to-day and snort tine aria clnns hoth scow

One say also compare the observed ': ;Y-1[ a quasi-irregular wave and peaticke pattern.

dennity assuming the ChapIan reaction to he between oariotion" in the OH and '" ;3-I) dmo-

responsible for the emission (Wallace and Hunten, spheric band emisoions. be helece therefore

1968; Deans et al., 19761, i.e., that the short time vrIalaons seen In the 0'(1-1) hand are connected with turbulence and

0 , 0 + M - 02 (bIZ+) + m (I) breaking gravity waves directed from above and

associated with auroral phenomen. The OH varla-Different reaction rates have been used. It tions seen to be influesced by turbulence con-

is aell knn [Wallace and Hunten, 19681 that the nected with breaking gravity gaves from the lower

laboratory obtained reaction rate by Young and atmosphere and traposphere. Further chservations

Slack [19651 Is not sufficient to account for the of OH and 2 (0-1) band intensities and tempera-observed emission rates. Aasoming the reaction tures ore seeded to verify this hypothesis. One

rate coefficient given by Barth [19611 or almost should also consider the posslbllity of jasing theS equialent by Campbell and Trush (19671 (taking night airglow [oil 557' i line at about 100 Octhe neutral temperature to be 120 K a 95 ,m); as an additional indicator of speed and phase .,fone may derive a approximate expression for the gravity waves.

overage oxygen atom nu er density over theemission heIght: Acknowledgements. Financial support for this

research was provided by the hationai ScIence[0(2 - (I f/k[M] a) 121 Foundation through grants ATh-H31I727, ATMH2-

14642 and A182-Olu, to the Geophysical InstituteFor the (0-I) band I is the observed emission of the University of Alasks a nd by the Royalrate of the (0-I) band In photon" cm

- 3 s5, f is Norwegian Council for Scientific and Industrial

the ratio between the total atmospheric bands Research through a feilowship grant to one of -s

emission and the (0-1) band. taken to be 20 [Dean (H.K.M.). The observations on Svalbard are madeat a., 19761; [M] is the number density of '

2 at through a cooperative effort involving the

95 km, taken to be 4 x 1013 cm-3

(CIBA, 1972); a universrties io Fairbanks, Troma, and Oslo, withis the half width of the (0-1) atmospheric band in the help and cooperation of the Great Norsegtancentimeters, taken to be 15 km oesa et al., Spitsbergen CoaI Company and the Norwegian Polar19761; and A is the reaction rate coefficient Institute. The authors aso want to thank j.from Barth [19611, k . 9 x 10

- 3 3 cm

0 s-1

. Using Baldridge for programing assistance is connec-equation (2) and the above values, one arrives at tion with the analysis of the data. The tditoran average oxygen atom number density of -1.5 x thanks P. Harris and another referee for theit1011 C.

- 3 for the emission intensity of 570 R. assistance in evaluating this paper.

For the emission rate 1590 R the corresponding 0number density is 4.2 . l0l1

c m- 2. Both these References

numbers are well within observed 0 number densi-ties for the high-latitude midwinter 90-100 k. Barth, C. A., Nitrogen and oxygen atomic reoc-

region (Thoma 1980; Mullen et al., 19771. In tions in the chemosphere, in Chemicalthe above estimation, quenching is not taken into Reactions in the Lower and Upper Atmosphere,account, but it is reasonable that even if quench- pp. 303-326, Wiley-Interscience, New York,

ing is taken Into account, the number density 1961.will stay well below [012 cm-3, which is a large Bates, D. H., Forbidden oxygen and nitrogen linesbut not unreasonably large number density of 101 at in the nlightglow, Planet. Space Sci., U5, 897,90-100 km for a shorter period of time [Mullen et 1978.

al., 19771. Beer, T., Atmospheric Waves, A. Hilger, London,Berthier, P., Etude spectrophotometrlque de a

4. Summary luminiscence nocturne des bandes des molecules

The 02 (0-1) atmospheric band in the night OH et 02 atmospherique, Ann. Gecphns. *__2,airglow has been observed from near Longyearbyen, 113, 1956.

Svaibard (7H N) during a 2-month period around renton, F. 0., and S. M. Silverman, A study of

winter solstice using spectrophotometric equip- diurnal variations of the 5577 A [O[j airglowment. The raw data were recorded with a spectral emission at selected ICY stations, Planet.resolution of 7 A which clearly resolved the R Space S c i., 13, 641, 1970.and P branches but not the rotational structure Campbell, 1. 4., and S. A. Crush, The association

within the branches. Absolute intensities have of oxygen atoms and their combination withbeen calculated from 30-tin averages of the nitrogen atom , Pros. R. Soc. London, Ser.emission spectra. and a mean 02 (0-1) atmospheric A, 296, 222. 1967.bend intensity of 570 R ± 60 R, ranging from Coger, L. L., R. I. Elphinstone, and J. S.110 R to 1590 R, is found. No clear intensity Murphy, Temporol and latitudinal 5577 A air-maximum or minimum around solstice is observed, glow variations, Can. J. Phys. , 59 1 1296, 1981.but comparison with other data indicates that Deans, A. J., G. G. Shepherd and W... Evans,this may be due tn intensity increases associated A rocket measremenrt of theY 0 (bwith mid winter stratospheric warnings. A semi- x I atmospheric band nIghtglow altitude

diurnal tide cmponent giving rise to a 25-30. disributon, Geophys. Res. Lett.. 3, 441,intenslty variation of the 02 (0-1) atmospheric 1976.band emission is present in the data. oMaxima and Deehr, C. S.. G. G. Sivjee, A. fgeland, K.mina are found to occur at about 0400 and 1000 Henriksen, P. E. Sandholt, R. Smith, P.local time, respectively. No diurnal tide compo- Sweeney, C. Duncan, and F. Gilmer, Ground-

87

1690 MoranO et a,: Midwiner IntenItes a: Night Airglou

based observations of F region associated B.oat , ., K. ?ezoicet, KC. Uahitc, and K.with the agnetospheric cusp, J. G.ophys. -enschow, Sol. Bert. leiterarte, :781, 982.

Res., 85, -185, 1980. Noxon, j. F., Effect of internal gravity waves

Dufay. M. , Etude photoelectrique du spectre du upon night airglow temperatures, Gaophvs. Sea.

ci nocturne dana ie proche infra-range, Lect. , 5. 25, 1978.

-nn. .on1o, 134, 1959. Fetitdidier, M., and H. Teitelbaum, Lower thermo-

Forbes, F. S., Atmospheric holes: The solar and sphere emissions and tides, Pianet. Sace

lunar semidiurnal components, . eophys. Sci., 25, 711, 1977.

e__.. 87, 5241, 1982. Rhmick, G. J., The detection and study of the

Frederick, J. E., Influence of gravity waves visible spectrum of the aurora and airglow,

activity on power thermospheric photochemistry Proc. Soc. Photo. Opt. Instrum. Eng., 91,

and comparison, Planet. Space. Sci. 7, 63, 1976.1469, 1979. Sandford, 3. P., Aurora and airglow iotennltv

luei', f. A., R. A. Koehler, R. Link and . v. carlations with time and magnetic ctlvitv

Shepherd. Observations of the optical spectrum at southern high latitudes, J. AtmoS. Terr.

of the dayside magnetospheric cleft aurora, Phys., 26, 749, 1964.

Planet. Space Sci., 29, 321, 1981. Sivjee, G. G.. and C. S. Deehr, Differences In

Greet. R.G.H,, E. J. Lleweilyn., B. H. Soihelo, polar stmospheric optical emiss~ons between

and G. Witt, The exciratlon of 0, (b'Z ) mid-day cusp and nighttime auroras,

in the nightglow, Planet. Space ScS., 9, 383, Exploration of the Polar Upper Atmosphere,

1981. edited by C.1. Deehr amd J. A. Holiter. p. 199,

Holton. F. R., An Introduction to Dynaic D. Reidel, Hingham Mass., .980.

Meteorology, Academic, Orlando, Fl. 1979. Siljee, . 0., K. A. Dick, and P. 7. Peidnan,

Holton, P. R.. The influence of gravity wave Temporal variation in the nighttime hydroxyl

breaking on the general circuletion of the rotational temperature, Planet. Space Sci.,

mlddie atmophare, J. Amos. Sci.. 40, 2497, 20, 261, 1972.

1983. Spizzichio, A., Etude des interactions entre len

Ismail, S, and L. L. Cogger, Temporal variations differentes composantes du vent dans La havt

of polar cap 01 5577A airglow, Planet. Space atmosphere, Ann. Oeophps., 23, 93 1969.

Sci., LO, 865. 1982. Spiecichino, A., -tude des interactions entre les

Labitzke, K.. Stratopheric-measopheric aid-winter differentes du vent dame La haut atmosphere,

warmings, Dynamical and Chemical Coupling don. Geophy ., 2, 9, 1970.

Between the Neutral and Ionized Atmosphere, Teitelbaum, H., and J. E. Blaont, Some conse-

edited by B. Grandal and J. A. Holter, p. 17, quencies of non-linear effects on tides and

D. Ridel, Hingham, Mass., 1977. gravity waves. 2. Atmos. Tart. Phys. 3/, 697,Lldzen, R. S., Turbulence and stress owing to 1975.

gravity wave and tidal breakdown, J. Gesphys. Thomas, L., The composition of the mesoaphere andBe_., 86, 9707, 1981. lower thermoephere, Philos. Crans. R. Soc.

Misawa, K., and I. Takeuchi, Correlation among 02 London, Set. A, 296, 243, 1980.

(0-15 atmospheric band, 08 (8-3) band and Sallace, L., and D. N. Hunten, Dayglow of the

(01] 5577A line and among P1 (2), Pl3) and onygen A band, J. Geophys. Res., 73, 4813,

Pj(3) and P1 (4) lines of H (8-3) band. J. 1968.

At... Tert. Phys., *4, 421, 1978. Weinstock, J., Theory of interaction of gravity

Morsel, Megie, A. Nailance-Jonas, and waves with Z, Or) airglow, J. Geophys. Res.,R.L. Gattinger, An oxygen-hydrogen atmospheric _, 5175, 9uodel and its application to the 08 problee, ~wt7 G., J. Stegman, B. S. Solthein, and E. J.2. AtmOs. Tert. Phys., 39, 51, 1977. L

9ewellyn, A measurement of the 7.( '.' -3 I - X Z- )

atmospheric batdad e I(S

Mullen, 5. 0-, 5. N. Silverman, and D. F. Korff, g 1) atos hec nd and the 01 S) e

Nightglow (557.7 m of 01) in the central grein line in th nightglow, Planet. Space

polar cap. Planet. Space Sci., 25, 23, 1977. Soi., 27, 341, 1979.Myrabi, H. K., Temperature variations at mesopause Wraight. P. C., Association of atomic oxygen an.

levels during winter solstice at 78-N, airglow excitation mechanics, Planet. Space

Planet. Space Sc., 32, 249, 1984. Sci., 30. 251, 1982.

Myrab, H. K., and C. S. Deehr. Mid-winter hy- Young, R A., and G. Black, Meaauremen~s of the

dronyl night airgo w emission intensities in rae coefficient of 0.(a16) + 0 (al) - 11

the northern polar region, Planet. Space Scl., (b E) -0 2 (KX ). J. che- Phys, .2, 1,701

32, 263, 1984. 1965.

Myrab6, H. K., C. S. Deehr, and C. G. Sivjee,Large-apliltude nightglow OH (8-3) band inten- C. S. Deehr and C. J. Romick, Geophysical

city and rotational temperature variations Institute. University of Alaska, Paihanks,

during a 24-hour period at 78°N, J. Ceophyn. AK 99775.

Res., 88, 9255, 1983. K. Henriksen, Auroral Observatory, University

Myrab:6, H. K-, C. S. Deehr, and B. Lybekk, Polar of Trome6, N-9000, Troms , Norway.

cap 08 airglow rotational temperatures at H. K. Myrabh, Norwegian Defence Research

the mesopause during a stratospheric warming Establishment, N-2007, Kjeller, Norway.

event, Planet. Space Sci., 32, 853, 1984a.

Myrab , H .K. Henrksen C. S. Deehr, and

G.J. Komick, 02 (b1 'E3) atmospheric

band night airglow measurements in the northern (Received May 3, 1985;

polar cap region, J. Geophys. Res., 89, 9148, revised August 5, 1985;

1984b. accepted August 29. 1985.)

,x. 4" 1 I

-- 7

89

Night airglow 02 (0-1) atmospheric band emission

during the northe-n polar winter

H K Myrabe

Norwegian Defence Research Establishment

P 0 Box 25, N-2007 Kjeller, Norway

ABSTRACT

The 02 (0-1) atm band night airglow has been observed from theground near Longyearbyen (78 N) during the 1983/84 winter solstice

period. The 2h months of observations show no minimum in the emission

around winter solstice, but rather large variations with enhancementslasting for days. An average atomic oxygen concentration at the 95 km

level of 1.5 x 101 cm-3 is deduced from the average emission inten-

sity of 405R. The absence of a clear minimum in the oxygen con-

centration during the northern polar winter solstice period, as com-

pared to the southern polar region, is believed to reflect differences

in the circulation and transport in the upper mesosphere and lower

thermosphere.

1 INTRODUCTION

The height interval 80-120 km is an important region of the upper

atmosphere. It contains the transition region between the ionosphere

and the lower neutral atmosphere, i e, from conditions of diffusive

separation to a well-mixed mesosphere. Gravity waves from below

(generated in the troposphere) and from above (in connection with

auroral activity) are believed to break in this region and deposit a

considerable amount of energy, causing enhanced eddy diffusion flux

90

and affecting the chemistry (Ebel, 1985; Myrabo et al, 1986). At high

latitudes the 80-120 km region is known to undergo large temperature

changes (cooling) and disturbances in the circulation associated with

stratospheric warming events (Labitzke et al, 1985; Matsuno, 1971;

Myrabo et al, 1984). The atmospheric atomic oxygen concentration

peaks at this altitude (Murphree et al, 1984).

Night airglow emissions originating from atomic and molecular oxy-

gen have been extensively used to monitor the oxygen concentration and

to study transport processes and chemistry of the upper atmosphere

(Dufay, 1959; Silverman, 1970; Petitdidier and Teitelbaum, 1977;

Cogger et al, 1981). However, most of this work has been concentrated

at mid- and low-latitudes. Recently Elphinstone et al (1984) utilized

a large data base of ISIS 2 01 5577A limb scans to construct a global

circulation model of the 80-120 km region consistent with the airglow

data. Only latitudes less than 400 were included.

At high latitudes both the 01 5577A and the 02 b' Ig emissions

might be used to monitor the peak oxygen concentration, transport and

dynamics of the 80-120 km region (Murphree et al, 1984; Deans et al,

1976, Myrabe et al, 1986). Night airglow measurements of these

emissions are sparse in the polar regions. Except for the obser-

vations by Myrabo et al, (1986). they are based on photometric data

and thus may have a higher risk of auroral contamination (both

emissions also occur in the aurora). Ground based photometric obser-

vations of the 01 5577A line from Thule (76 N) have been reported byMillen et al (1977). Results from Antarctica, utilizing ISIS 2 01

5577A limb scans are given by Ismail and Cogger (1982). A clear

winter solstice minimum is seen in the Antarctica data, while the

Thule observations are characterized by large amplitude irregular

variations lasting for periods of days to weeks (Millen et al, 1977).

Recently Myrabo et al (1986) reported similar irreguiar behaviour of

the 02 (0-1) atm band (i e, 02 (b1 Z* g - X1 V g)) emission during

winter solstice at 78°N. A possible seasonal trend around winter

solstice could not be clarified due to a rather short (38 days)

observing period. New 02 (0-1) atm band night airglow emission datafrom Spitsbergen (780N) covering a 2 month per-iod around winter

solstice, is consistent with the 01 5577A Thule observations, clearly

91

showing that the northern polar region has no winter minimum in the

oxygen concentration.

2 OBSERVATIONS

The 02 (0-1) atm band emission were observed from near Longyear-

byen on West Spitsbergen (78.4 N; 15 E) during the 1983/84 Multi-

National Svalbard Auroral Expedition (Deehr et al, 1980). The

spectral region from 8235 to 8685A was scanned in 12 seconds employing

a h M Ebert-Fastie spectrophotometer pointing towards the zenith. A

band-width of 7A clearly resolved the R and P branches of the 02 (0-1)

atm band and the P1 and P2 lines of the OH (6-2) band. Individual

scans (300) were sunmed and averaged to obtain hourly mean spectra.

Only spectra during clear weather periods and with no auroral features

were used. The aurora was monitored by the 01 8446A line present in

the spectrum (Myrabo et al, 1986). An example of a typical spectrum

used for calcuating the 02 (0-1) atm band intensity is presented in

Figure 1. The OH (6-2) P-lines, the 01 8446A auroral line and the OH

(7-3) R b'anches are indicated in addition to the 02 (0-i) atm band.

The absolute intensities were obtained by calibrating the spectropho-

tometer in the field, using a standard lamp and a diffusion screen

(Hamwey, 1985). For the 02 (0-1) ati- bard the area under the R and ,

branches represent the absolute intensity (see Figure). The uncer-

tainty in the absolute intensity is estimated to be 20%, mainly due to

uncertainty in the calibration. The relative uncertainty is mainly

due to photon noise and to the uncertainty in the background subtrac-

tion. This is less than 5% for a single one hour mean. From the

hourly values of the intensities, daily averages were formed. At

least 3 hourly values have been used to obtain a daily average.

A ± 12-15% bias due to the semi-diurnal variation has been removed,

assuming the 3mplitude aiwd phase found by Myrabo et al (1986).

Because the daily averages were obtained from 3-10 hourly values, the

average correction was less than 5%. Possible seasonal variations of

the amplitude and phase of the tides should ,herefore not effect the

result significantly. Probably the largest uncertainty in the daily

values are caused by short time variations of the intensity (Myrabe et

92

al, 1986). An example of this is given in Figure 2 showing 20 minute

averages of intensities during the evening of 25 December.

30 P(2) P(3) P PM(P(3) I

01 1 I I 0184461

25- I-.--)

20 - 02(0-1) Atm.- tO A R P

10 U5

8250 8300 8350 8400 8450 8500 8550 8600 8650WAVELENGTH (A)

Figure I Example of I hour of av~eraged scans used to obtain the ab-solute intensities for the 02 (0-1) atm band. The OH (6-2)P-lines, the auroral 01 8446A and the R branches of the OH(7-3) band are also indicated.

The daily means of the 02 (0-1) atm band intensities as obtained

during the observing period from 25 November 1983 to 6 Februa'y 1984

is presented in Figure 3. The result from the 1982/83, 38-days of

observations, is given on the same plot (dashed line) for comparison,

and as an extension of the data set. Winter solstice is marked with

an arrow. Figure 4 compares the 02 (0-1) band 864.5 nm night airglow

data from Spitsbergen to reflect differences in the dynamics of the

two regions. The dashed line indicates the seasonal trend in the

Antarctica data. Both the 0I 5577A line and the 02 (0-1) atm band

intensities are given with similar relative scales, i e, one unit

corresponds to the same relative variation.

93

I I L I

300

200

I-

zLI,.

z 100

0 I I18 20 22 24

UNIVERSAL TIME (Fi-S)

Figure 2 Variation of the 02 (0-1) atm band intensities in the eveningof 25 December. Intensities are obtained from 20 minutes ofaveraged individual scans. Observed values are marked.

1000 /

< 750

ESoo\ 40 250 fv! 1

25 30 5 10 15 20 25 30 5 10 15 20 25 30 5

1 DECEMBER JANUARY

Figure 3 The daily mean intensities of the 02 (0-1) atm band for the1983/84 season. The result from the 82i83 season is given onthe same plot (dashed line) for comparison.

94

NOVEMBER DECEMBER JANUARY FEBRUARY MARCH

1 10 20 1 10 20 1 10 20 1 10 20 1 10

300 750

C,, *, .500z 200 \ -

z N -

100-1001 * . .- * 250

- C 1 I 0

MAY JUNE JULY AUGUST

Figure 4 The daily averages of the 02 (0-1) atm band intensities as

obtained from Longyearbyen (780N) during the 1983/84 season.The intensities of the 01 5577A line from Antarctica obser-vations (ISIS 2 data) during winter solstice conditions arealso plotted (filled triangles). The dashed line indicatesthe seasonal trend for the Antarctica data. Antarctica datais from Ismail and Cogger (1982)

3 DISCUSSION

3.1 The winter solstice pattern and the differences between the twopoles

The 01 5577A line and the 02 (0-1) atm band emissions originate at

about the same altitude (Greer et al, 1981) and are known to covary

(Dufay, 1959). As pointed out in an earlier paper by Myrabo et al

(1986) the 01 5577A line and the 02 (0-1) atm band emission should

show the same seasonal variation and reflect the atomic oxygen con-

centration In the 85-105 km region.

From Figure 3 it can be seen that both the 1982/83 and 83/84

winter season at 780N show irregular peaks of enhanced 02 (0-1) atm

band intensity levels, lasting for a few days to a week or more. This

95

is very similar to the O 5577A night airglow observations from Thule

(76'N) reported by MUllen et al (1977). It is also seen from Figure 3

that there is no minima around winter solstice in the 02 (0-1) atm

band emission at 780 N. When the 02 (0-1) atm band data from the

1983/84 season are plotted together with the Antarctica data from

Ismail and Cogger (1982) (Figure 4), the differences are easier to

see, i e, the 02 (0-1) atm band from the northern hemisphere show much

larger day-to-day variations than do the O 5577A emission from

Antarctica.

Each of the filled triangles in Figure 4, 1 e, the Antarctic data,

corresponds to averaging of several limb scans during d single pass.

From the temporal coverage in the Antarctic data as compared to the

ground based observations a relatively larger scatter should be

expected In the Antarctic data than in the ground based data. From

Figure 4 the opposite is seen. If the day-to-day variation in both

the 01 5577A line and 02 10-1) atm band emissions were mainly con-

nected to auroral activity, one would expect both the northern and

southern polar regions to show a simlar type of behaviour. Ismail and

Cogger (1982) have compared the enhanced O 5577A line emission from

Thule with 30 mbar temperatures over northern Canada and find a relati

vely good correlation if a 0 and 14 day delay is allowed for. Myrabo

et al (1984) have shown that the mesopause (i e, -90 km) temperatures

are affected by disturbances in the circulation pattern at stra-

tospheric levels (i e, stratwarms). It Is therefore reasonable to

believe that the main part of the variations are connected to cir-

culation changes in the stratosphere and lower mesosphere, affecting

the lower thermospheric oxygen concentration and temperature through

vertical mixing and horizontal transport. The differenes between the

two poles may then be explained as differences in mid-winter cir-

culation disturbances, i e, the northern winter stratosphere is known

to have large disturbances while the southern winter stratosphere has

only small scale disturbances (Schoeberl, 1978).

Some of the enhancements may indirectly be connected to auroral

activity producing odd oxygen followed by a downward transport of the

oxygen from the auroral source. The period 8-12 January 1983 shows a

2-3 times enhancement of the 02 (0-1) atm band which is closely con-

nected to a period of high auroral average activity around 8-10

96

January. In order to enhance the 0 atom concentration by a raio e,

i e, -2.7, during a 24 hour period at the 95 km level, assuming down-

ward diffusion from 110 km, an eddy diffusion coefficient,

1) kI = Ah2 /TD = 2.6 x 101 cm2 /sec

is needed, Ah being the hight difference and TD the time over which

the diffusion take place. Thus assuming that the mechanism for exci-

tation of the 02 b'E state of 02 is a two step process (Torr et al,

1985), i e,

2) 0 , 0 + M 02** + M

followed by

3) 02** + 02 02 (b1X) + 02

Follo.4ing the photochemical scheme and the model dtmosphere given

in the paper by Torr et al (1985), an eddy diffusion coefficient of

the order uf 6 x 101 cm2 /sec is needed to account for a 2.5 times

enhancement nf the intensity over 24 hours. The effect of eventually

neglecting the quenching by oxygen and even assuming a direct Chapman

mechanism for the excitation of the 02 b'I state is negligible on the

deduced oxygen profile below about 97-98 km. This is clearly seen

from Mc Deans et al (1976) (Deans et al assumed a direct Chapman exci-

tation mechanism and neglected the quenching by oxygen).

The value 6 x 107 cm2 /sec is a rather large eddy diffusion coef-

ficient compared to average values, mainly quoted around 10' cm2 /sec

(Von Zahn and Herwig, 1977). It is not unreasonably large, however,

since the eddy diffusion coefficient may be highly variable (Wein-

stock, 1985; Thrane et al, 1985). We may therefore conclude that

parts of the enhancement could be associated with auroraly produced

oxygen. However, a further separation of the sources have to await

measurements of several emissions at different heights, including

probably both OH, Na, 02 (0-1) and 01 5577A emissions (Myrabo et al,

1987).

97

3.2 Absolute intensities of the 02 (0-1) atm band emission and anestimate of the oxygen concentration and its variation

The mean intensity of the 02 (0-1) atm band for the 1983/84 season

was found to be 405R. Daily averages ranged from 180 to 780 R with

lowest and highest hourly value of 160 and 860 R, respectively. This

is slightly lower than the 570 R 1982183 seasonal average, but still a

relatively high 02 (0-1) atm band night airg~ow intensity as compared

to lower latitudes (Packer, 1961; Deans et al, 1976; Megill et al,

1970).

Assuming a two step proot.bs ; ( 2) and (3)) to be responsible for

the emission and using the reaction rates and coefficients given by

Torr et al, (1985) one might, to a first approximation, deduce an

atmospheric oxygen concentration of 1.5 x 1011 cm-3 at 95 km.

The highest and lowest measured intensity corresponds to an oxygen

concentration of -2 x 1011 cm-3 and -1 x 1010 cm- , respectively. The

neutral temperatures are taken to be 220 K for these estimates

(Myrabo, 1984). Thus, the oxygen concentration may vary by more than

1000%. Some of this variation might, however, be related to tem-

perature variation as the excitation of 02 (b1 X" g) might be tem-

perature dependent (Wraight, 1982). The relatively high oxygen con-

centration found in the northern polar region during winter solsticemust be either transported to the 90-100 km region from outside orproduced locally by auroral events and downward diffusion of oxygen.

Knowing the differences in the behaviour of the night airglow 02 (0-1)

atm band and the 01 5577A line in the two polar caps, disturbances in

the stratosphere and lower mesosphere causing circulation and eddy

diffusion changes in the mesopause and lower thermosphere region could

be responsible for the variation In the supply of atomic oxygen.

Whether this supply of oxygen comes from aurorally produced oxygen or

from photodlssociation of 02 In the sunlit lower latitude atmosphere

or both, is not yet known. However, experiments is prepared to

clarify these questions.

98

Acknowledgement

Financial support for this research was provided by National

Science Foundation through grant ATM-8313727 to the Geophysical

Institute of the University of Alaska and by Royal Norwegian Council

for Scientific and Industrial Research through a fellowship grant. The

author also wants to thank Mr B Erickson, Mr D Osborne and Ms D L

Wilkinson for programming and technical assistance.

References

Cogger, L. L., R. D. Elphinstone and J. S. Murphree, 1981.

Temporal and latitudinal 5577A alrglow variations, Can. J. Phys.,

59, 1296.Deans, A. J., G. G. Shepherd and W. F. J. Evans, 1976. A rocket

measurement of the 02 (bl X* g - X3 7-3) atmospheric band night-

glow altitude distribution, Geophys. Res. Lett., 8, 441.

Deehr, C. S., G. G. Sivjee, A. Egeland, K. Henriksen, P. E. Sandholt,

R. Smith, P. Sweeney, C. Duncan, and F. Gilmer, 1980.

Groundbased aobservations of F-region associated with the magne-

tospheric cusp, J. Geophys. Res, 85, 2185.

Dufay, M., 1959. Etude photoelectrique du spectre du ciel nocturne

dans le proche infra-rouge, Ann. Gephys., 15, 134.

Ebel, A., 1985. Contribution of gravity waves to the momentum, heat

and turbulent energy budget of the upper mesosphere and lower

thermosphere, J. Atm. Terr. Phys., 46, 727.

Elphinstone, R. D., J. S. Murphree and L. L. Cogger, 1984. Dynamics

of the lower thermosphere consistent with satellite observations

of 5577A airglow: I. Atomic oxygen, local, turbulence, and glo-

bal circulation results, Can. J. Phys., 62, 382.

Greer, R. G. H., H. J. Llewellyn, B. H. Solhelm and G. Witt, 1981.

The exitation of 02 (bl V g) In the nightglow, Plant. Space Sci.,

29, 383.Hamway, R., 1985. Spectroscopy of the night alrglow OH emissions, M.

S. Thesis, Geophysical Institute, Alaska.

Ismail, S. and L. L. Cogger, 1982. Temporal variation of polar cap

01 5577A airglow, Planet. Space Sc., 9, 865.

99

Labitzke, K., B. Naujokat. R. Lenschow, K. Petzoldt, B. Rajewski and

R. C. Wohlfart, 1985. The third winter of MAP-dynamics, 1984/85:

A wiiter with an extremely intense and early major warming,

beilage zur Berliner Wetterkarte 67/85.

Matsuno, T., 1971. A dynamiLal smodel of the stratospheric sudden

warming, J. Atmos. Sci., 28, 1479.

Mc Dade, I. C., E. J. Llewellyn and R. R. Harris, 1985. Atomic oxygen

concentration In the lower auroral thermosphere, Adv. Space Res.,

5, 229.Megill, L. R., A. M. Despain, D. J. Baker and K. D. Baker, 1970.

Oxygen atmospheric and infrared atmospheric bands in aurora, J.

Geophys, Res., 75, 4775.

Moreels, G., G. Megie, A. Vallance-Jones and R. L. Gattinger, 1977.

An oxygen-hydrogen atmospheric model and its application to the OH

problem, J. Atm, Terr. Phys., 39, 551.

MUllen, E. G., S. M. Silverman and 0. F. Korff, 1977. Nightglow

(555.7 nm of 01) in the central polar cap, Planet. Space Sci., 25,

23.

Murphree, J. S., R. D. Elphinstone and L. L. Cogger, 1984. Dynamics

of lower thermosphere consistent with satellite observations of

5577A airglow: I. Method of analysis, Can. J. Phys., 62, 370.

Myrabo, H. K., C. S. Deehr, B. Lybekk, 1984. Polar cap OH airglow

rotational temperatures at the mesopause during a stratospheric

warming event, Planet. Space Scl., 32, 853.

Myrabe, H. K., 1984. Temperature variations at mesopause levels

during winter solstice at 78°N, Planet. Space Sci., 32, 249.

Myrabe, H. K., C. S. Deenr, G. J. Romick and K. Henriksen, 1986.

Midwinter intensities of the night airglow 02 (0-1) atmospheric

band emission at high latitudes, J. Geophys. Res., 91, 1684.

Myrabo, H. K., C. S. Deehr, R. Viereck and K. Henriksen, 1987. Polar

mesopause region dynamics deduced from night airglow emissions, J.

Geophys. Res., (in press).

Packer, D. M., 1961. Altitudes of the night airglow radiations, Ann.

Geophys., 17, 67.

Petitdidier, M. and H. Teitelbaum, 1977. Lower thermosphere emissions

and tides, Planet. Space Sci., 25, 711.

100

Shoebend, M. R., 1977. Stratospheric warmings, observations and

theory, Rev. Geophys. Space Phys., 16, 521.

Silverman, S. M., 1970. Night airglow phenomenology, Space Sci. Rev.,

11, 331.

Thrane, E. V., 0. Ardreassen, T. Blix, B. Grandal, A. Brekke, C. R.

Philbrick, F. J. Schmidlin, H. U. Widdel, U. Van Zahn and F. J.

LUbken, 1985. Neutral air turbulence in the upper atmosphere

observed during the Energy Budget Campaign, J. Atm. Terr. PhyS.,

47, 243.

Torr, M. R., D. R. Torr and R. R. Laher, 1985. The 02 atmospheric 0-0

band and related emissions at night from Spacelab 1, J. Geophys.

Res., 90, 8525.

von Zahn, V. and T. Herwig, 1977. Inert gas abundances as indicators

for the strength of eddy diffusion at turbopause altitudes, in

Dynamical and Chemical Couping between the Neutral and Ionized

Atmosphere, editors B. Grandal and J. A. Holtet, p. 49, Dortrecht,

D. Reidel.

Weinstock, J., 1985. Gravity wave saturation and eddy diffusion in

the middle atmosphere, J. Atm. Terr. Phys., 46, 1069.

Wraight, P. C., 1982. Association of atomic oxygen and airglow exci-

tation mechanism, Planet. Space Sci., 30, 251.

DAT