Morphology of evening sector aurorae in λ557.7-nm Doppler
J. M. Holmes1, M. Conde2, C. Deehr, and D. Lummerzheim
Geophysical Institute, University of Alaska Fairbanks
An all-sky scanning Fabry-Pérot spectrometer was used to observe temperatures of auroral
OI(557.7-nm) emissions over Poker Flat. The sudden temporal and spatial changes in Doppler
temperatures that were observed are likely due to the emission height changing as a response to
variations in the characteristic energy of the precipitating electron population. Three cases were
analyzed: (1) A Doppler temperature drop (~200K) over the entire sky occurred immediately
after an auroral brightening; the temperature remained lower after the auroral intensity had
resumed its quiescent levels. (2) A local increase of Doppler temperature, colocated with a
weak auroral arc, occurred 25 minutes before a westward propagating substorm onset. When
the auroral luminosity suddenly increased the Doppler temperature had a sharp decrease. (3)
The region inside a loop like auroral arc showed elevated Doppler temperature relative to that
of the arc itself. Auroral fading just prior to onset was accompanied by increased temperatures.
Beginning with the original interferometric
measurements of the auroral λ557.7-nm line by Babcock
, there has been considerable interest in the
measurement of auroral and airglow emissions and their
applications to upper atmospheric physics. Determination
of temperatures from Doppler-broadened emissions,
including both atomic [Armstrong, 1958; Wark, 1959;
Nilson and Shepherd, 1960] and molecular species
[Vegard, 1932; Lytle and Hunten, 1960], contributed to the
early quantification of thermospheric and mesospheric
temperatures. A discussion of the applications of optical
Doppler measurements for this era is given by Hunten
More recently, interferometric techniques have been
used with an emphasis on determining neutral winds from
observed Doppler shifts [Hernandez and Killeen, 1988].
Investigations of thermospheric dynamics have been made
using these techniques ranging from small-scale studies of
specific phenomena [Rees, 1984b] to global-scale
observations [Killeen et al., 1988].
With the advent of the ground-based all-sky Fabry-Pérot
spectrometer (ASI-FPS) [Rees and Greenaway, 1983; Rees
et al., 1984a; Biondi et al., 1995; Nakajima et al., 1995;
Ishii et al., 1997; Conde and Smith, 1995, 1997], it became
possible to record Doppler spectra in many locations across
the sky, as opposed to previous narrow-field Fabry-Pérot
spectrometers which routinely observed only the zenith
plus the four cardinal azimuths. This novel technique that
provides modest spatial resolution is now used to explore
the relationship between auroral intensity and Doppler
The auroral λ557.7-nm [O I] emission is a result of the
metastable transition between the excited 1D and 1S states
of atomic oxygen, the bulk of which is produced in the
height range of 100-150 km. The state has a lifetime of
0.74s, which provides adequate time for thermalization in
this altitude range. Thus, it can be tacitly assumed that the
emitting population is in thermal equilibrium with the
neutral atmosphere. Since the thermosphere has a positive
temperature gradient, higher energy electron precipitation,
which penetrates farther into the lower thermosphere,
produces more intense aurora with lower temperatures as
measured by all-sky instruments such as the Scanning
Doppler Imager [Rees, 1989].
This inverse relationship between Doppler temperature
and auroral intensity has long been known [Størmer, 1955],
and is known to persist in spite of increased localized Joule
heating [Ishii et al., 2001]. Narrow-field interferometric
measurements of auroral Doppler temperatures, when
compared with intensities, show an identical relationship,
although deviations from simple inverse-proportionality are
observed [Hilliard and Shepherd, 1966]. Owing to the
known relations between observed temperatures and
intensities, optical temperature data, in the absence of
auxiliary height information, are of limited use in
estimating the rate of auroral energy deposition and Joule
heating [Turgeon and Shepherd, 1962].
The Poker Flat SDI makes measurements of the λ557.7-
nm [O I] emission using an imaging Fabry-Pérot
capacitance-stabilized étalon of 20-mm air-spaced gap and
105-mm working aperture. The all-sky optics of the
instrument have a 140º field of view. The imager is unique
in that unlike other imaging FPS instruments, the SDI
acquires spectra by repeatedly scanning the Fabry-Pérot
étalon plate separation through one order of interference,
producing spectra as a function of etalon gap (as opposed
1Now at the University Centre in Svalbard, Longyearbyen,
2Now at Latrobe University, Melbourne, Australia
HOLMES ET AL.: AURORAL MORPHOLOGY IN DOPPLER TEMPERATURES
Figure 1. Example SDI λ557.7-nm zone map with
fitted spectra (white). Exposure time: 1 minute.
Background shading depicts zone boundaries.
to the individual fringe shape) [Conde et al., 2001a]. This
allows the acquisition of images and spectra which are not
distorted by temporal and spatial brightness fluctuations in
the aurora during one scan [Conde and Smith, 1997].
To produce high signal-to-noise spectra with integration
times on the order of minutes, the SDI analysis software
divides images into arbitrary “zones”, the overall signal in
each zone yielding an individual spectrum. Figure 1 shows
the zones and their recorded spectra for a one-minute
integration. By fitting each spectrum with a suitable model
profile (in this case, a Gaussian emission function
convolved with a measured instrument function), the
Doppler broadening, Doppler shift and emission and
background brightnesses are determined. Doppler shifts
provide line-of-sight winds, while Doppler broadening,
when calibrated, returns the temperature of the emitting
For the cases analyzed, the SDI is configured to group
signal into 67 zones, thereby providing modest spatial
resolution without resorting to longer integration times
unsuitable to investigate the desired temporal auroral
structure. In order to ensure consistent uncertainties and
signal levels, the integration time is automatically adjusted
during observations according to signal brightness, with
exposure times varying from about 1 minute for bright and
7-8 minutes for very weak aurora.
Uncertainties in calculated temperatures vary depending
on auroral intensities and also the location of the zone
relative to the zenith, but are typically between 5 and 10
degrees Kelvin under bright auroral conditions. Figure 2(a,
b, c) are histograms of temperature uncertainty and their
variation with zone number (zone numbers increase with
zenith angle). Also, Figure 2(d) shows the decrease of
temperature uncertainty with auroral intensity. A
description of the treatment and determination of
uncertainties can be found in Conde [2001b]; a detailed
discussion of instrumental particulars can be found in
Conde and Smith .
Three cases were chosen to illustrate different types of
dynamics of the λ557.7-nm optical temperature relative to
the motions of the visible aurora. All cases display
evening sector discrete aurora and/or substorm activity.
Although measurement of temperature derived from the
spectral width of auroral emissions can be used in part for
identifying Joule and particle heating, care must be taken
since it is well known that the height of emission (and
therefore the temperature in the ideal thermosphere) varies
primarily by changes in the energy spectrum of the particle
precipitation [Sica et al., 1996].
3.1 Case 1: December 2, 2002
The first case was observed on December 2, 2002.
Figure 3 shows the SDI data in an image sequence of
temperature maps. For reference, the Geophysical Institute
white light all-sky camera is shown above the maps.
Since the previous brightening event at 0844 UT (not
shown), a gradual increase in measured temperature
occurred, starting near zenith then nearly filling the entire
field of view of the instrument at about 0930 UT. This is
depicted in the first two images in Figure 3, at 0944 and
0947 UT. It is interesting to note that there was little or no
change in the measured brightness during this period of
The westward-propagating brightening event at 0949
UT displays an immediate temperature drop of roughly
200K within the intensified region. Note that after the
intensity enhancement had passed (either dimming
overhead or advecting out of the instrument’s field of
view), the lower temperature present during the event
persisted for many tens of minutes thereafter.
Figure 2. Temperature uncertainty histograms calculated
for 22 March 2003: all zones (a), zenith zone (b), edge
zone (c). A scatter plot showing temperature uncertainty
versus intensity is shown in (d).
HOLMES ET AL.: AURORAL MORPHOLOGY IN DOPPLER TEMPERATURES
3.2 Case 2: March 22, 2003
The second case was observed on March 22, 2002. The
sequence in Figure 4 shows a band of discrete arcs near
zenith accompanied by a relatively high temperature
(shown in yellow and red). Preceding the beginning of the
sequence, a thin arc of both increasing temperature and
brightness was observed slightly equatorward of the
brightening arc shown in the first frame in the figure.
Immediately after the emission intensity began to increase
at approximately 0852 UT, a temperature drop occurred
only in the vicinity of the brightening features, i.e. from
zenith eastward toward the edge of the temperature map
(0856 – 0858 UT). This case illustrates large gradients and
discrete features in the temperature data evolving in only
3.3 Case 3: March 24, 2003
The final and most interesting case occurred on March
24, 2002. Figure 5 shows the time evolution of a
westward-traveling arc. The SDI measured low to
moderate temperatures in the vicinity of the discrete loop-
like band, which is expected based on previous
measurements However, inside the band, a much warmer
region developed (0852-0857 UT) shortly before the band
broke up at 0859UT, filling almost the entire field of view
Figure 6 depicts the elevation of neutral temperature in
the zenith zone during the above time period. The ratio of
the λ630.0-nm [O I] emission to the λ427.8-nm (N2+)
emission, as measured by the Meridian Scanning
Photometer (MSP) at Poker Flat, was used to make an
estimate of the characteristic energy of deposited auroral
electrons in the magnetic zenith using a technique
Figure 4. A time series depicting an auroral brightening event and associated changes in the λ557.7-nm
optical temperature for 22 March 2003 from 0850 to 0909UT.
Figure 3. An image sequence depicting an auroral brightening event and associated changes in the λ557.7-
nm optical temperature for 2 December 2002 from 0944 to 0958 UT. The upper row shows the Poker Flat
All-sky Camera and the lower row shows the temperature maps produced by the Scanning Doppler Imager
Figure 5. A time series showing the WTS event on 24 March 2003 from 0848 to 0900UT. Note that as the
loop-like feature moves westward, the region of elevated optical temperature inside it appears to follow the
HOLMES ET AL.: AURORAL MORPHOLOGY IN DOPPLER TEMPERATURES
described in Lummerzheim, et al. . The figure
indicates a decreased characteristic energy, which follows
if it is assumed that the increase in Doppler temperature
resulted, at least partly, from an increase in the emission
There exist commonalities between all three cases
discussed in this study. First, when considering discrete
arcs, the combined SDI and ASC data clearly show the
commonly accepted inverse relationship between auroral
intensity and temperature.
Next, as mentioned earlier, the data frequently exhibit
large temperature variations over very short spatial and
temporal scales. Several median temperature time series,
taken over the entire field of view, have shown changes in
temperature of hundreds of degrees Kelvin in a matter of
only several minutes. For the first case of December 2,
2002, shortly after 0800 UT (not shown), a near all-sky
brightening event occurred and the measured temperature
from the SDI dropped nearly 200K. Since the temperature
decreased, it is evident that auroral heating or possibly
other in situ thermospheric dynamics are not primarily
responsible for the bulk of the temperature change; the
drop occurred from a sudden lowering of the altitude of
auroral energy deposition, which corresponds to an
increase in characteristic energy. However, neutral heating
is almost certainly taking place, since it draws more than
50% of the power associated with auroral particle
precipitation [Rees et. al., 1983]. The presence of heating
in the emitting region causes the estimated emission height
to be overestimated (when simply converting from
temperature using a model such as MSIS); the actual height
of emission could in fact be lower depending on the
amount of heating.
It is also noteworthy that for the case of 0848-0900UT
on 24 March 2003, the portion of the loop-like feature
closest to the zenith decreased in intensity from 15 kR at
0854 UT to 3 kR immediately before the onset westward-
traveling surge. The λ427.8-nm channel of the MSP
revealed a similar reduction while the λ630.0-nm channel
remained approximately constant. This sudden increase in
the red/blue ratio and corresponding decrease in
characteristic energy is indicative of “auroral fading” as
described by Pellinen and Heikkila . Although the
softening of the auroral precipitation was not accompanied
by a significant increase in number flux, and the
equatorward hydrogen arc detected by the MSP λ486.1-nm
(Hβ) channel did not move significantly during this
“fading” period, the region of enhanced λ557.7-nm
Doppler temperature, located just equatorward of this arc,
increased in temperature by at least 100K in only several
minutes prior to surge onset.
The case studies presented here are only interpreted
qualitatively; they are used to illustrate some of the
phenomenology of the neutral thermosphere Doppler
temperature in the auroral energy deposition region and its
relation to optical auroral emissions. In addition, the
Doppler temperatures recorded using this imaging
instrument show variations with respect to auroral activity
that are consistent with older non-imaging instruments.
Although temperature data from the Scanning Doppler
Imager alone cannot be used to deduce the sources of the
dynamics observed, it is nonetheless an indispensable
instrument for the observation of spatially resolved
phenomena in the vicinity of auroral arcs and the
characterization of neutral atmospheric properties in the
context of auroral effects on the thermosphere.
Acknowledgments. This work was supported by a joint NASA
and NSF TIMED-CEDAR grant, number NAG5-10069. We
would also like to thank Brian Lawson for the operation of the
Poker Flat All-sky Camera and Meridian Scanning Photometer.
Armstrong, E. B. The temperature in the atmospheric region
emitting the nightglow OI 5577 line and in regions above faint
auroral arcs, J. Atmos. Terr. Phys., 13, 205-216, 1959.
Biondi, M. A., D. P. Sipler, M. E. Zipf, and J. L. Baumgardner,
All-sky Doppler interferometer for thermospheric dynamics
studies, Appl. Opt., 34, 1646-1654, 1995.
Conde, M. and R. Smith. Mapping thermospheric winds in the
auroral zone, Geophys. Res. Lett., 22, 3019-3022, 1995.
Conde, M. and R. Smith. Phase compensation of a separation
scanned, all-sky imaging Fabry-Pérot spectrometer for auroral
studies, Appl. Opt., 36, 5441-5450, 1997.
Conde, M., J. D. Craven, T. Immel, E. Hoch, H. Stenbaek-
Nielsen, T. Hallinan, R. W. Smith, J. Olson, and Wei Sun,
Assimilated observations of thermospheric winds, the aurora,
and ionospheric currents over Alaska, J. Geophys. Res., 106,
Conde, M, Analysis of Fabry-Pérot spectra of lidar backscatter
echoes, ANARE Reports, edited by Morris, R. J. et al., 146, 91-
Hernandez, G. and T. L. Kileen, Optical measurements of winds
and kinetic temperatures in the upper atmosphere, Adv. Space.
Res., 8(5), 149-213, 1988.
Hilliard, R., L. and G. G. Shepherd, Upper atmospheric
temperatures from Doppler line widths –IV, A detailed study
using the OI 5577 Å auroral and nightglow emission, Planet.
Space Sci., 14, 383-406.
Hunten, D. M., Temperatures deduced from aurora and airglow
spectra, Ann. Géophys., 17, 249, 1961.
Figure 6. Time series of zenith temperature (solid line)
and derived characteristic energy (dashed line)
0832 and 0916 UT on 24 March, 2003. Dotted lines
represent MSIS-90 heights.
HOLMES ET AL.: AURORAL MORPHOLOGY IN DOPPLER TEMPERATURES
Ishii, M., M. Conde, R. W. Smith., M. Krynicki, E. Sagawa, and
S. Watari, Vertical wind observations with two Fabry-Pérot
interferometers at Poker Flat, Alaska, J. Geophys. Res., 106,
Ishii, M., S. Okano, E. Sagawa, S. Watari, H. Mori, I. Iwamoto,
and Y. Myrayama, Development of Fabry-Pérot
interferometers for airglow observations, Proc. NIPR Symp.
Upper Atmos. Phys., 10, 97-108, 1997.
Killeen, T. L., J. D. Craven, L. A. Frank, J. J, Ponthieu, N. W.
Spencer, R. A. Heelis, L. H. Brace, R. G. Roble, P. B. Hays,
and G. R. Carnigan, On the relationship between dynamics of
the polar thermosphere and morphology of the aurora: Global-
scale observations from Dynamics Explorers 1 and 2, J.
Geophys. Res., 93, 2675-2692, 1988
Lummerzheim, D., M. H. Rees, and G. J. Romick, The application
of spectroscopic studies of the aurora to thermospheric neutral
composition, Planet. Space Sci., 38, 67-68, 1990.
Lummerzheim, D., and J. Lilensten, Electron transport and energy
degradation in the ionosphere: evaluation of the numerical
solution, comparison with laboratory experiments and auroral
observations, Ann. Geophys., 12, 1039-1051, 1994.
Lytle, E. A. and D. M. Hunten, Dawn enhancement of Auroral N2+
emission, Can. J. Phys., 38, 477, 1960.
Nakajima, H., S. Okano, H. Fukunishi, and T. Ono, Observations
of thermospheric wind velocities and temperatures by use of a
Fabry-Pérot Doppler imaging system at Syowa Station,
Antarctica, Appl. Opt., 34, 8382-8395, 1995.
Nilson, J. A. and G. G. Shepherd, Upper atmospheric temperatures
from Doppler line widths –I. Some preliminary measurements
on 0I 5577 Å in aurora, Planet. Space Sci., 5, 299-306, 1961.
Pellinen, R. J., and W. J. Heikkila, Observations of auroral
fading before breakup, J. Geophys. Res., 83, 4207, 1978.
Rees, D., and A. H. Greenaway, Doppler imaging system, an
optical device for measuring vector winds, 1, General
principles, Appl. Opt., 22, 1078-1083, 1983.
Rees, D., A. H. Greenaway, R. Gordon, I. McWhirter, P. J.
Charleton, and Å. Steen, The Doppler imaging system, initial
observations of the auroral thermosphere, Planet. Space Sci.,
32, 273-285, 1984a.
Rees, D., R. W. Smith, P. J. Charleton, F. G. McCormac, N.
Lloyd, and Å. Steen, The generation of vertical winds and
gravity waves at auroral latitudes, I, Observations of vertical
winds, Planet. Space Sci., 38, 667-684, 1984b.
Rees, M. H., B. A. Emery, R. G. Roble, and K. Stamnes, Neutral
and ion gas heating by auroral electron precipitation, J.
Geophys. Res., 88, 6289-6300, 1983.
Rees, M. H., Physics and chemistry of the upper atmosphere,
Cambridge, New York, 1989.
Størmer, C., The polar aurora, Clrendon Press, Oxford, 1955.
Turgeon, E. C. and G. G. Shepherd, Upper atmospheric
temperatures from Doppler line widths –II, Measurements on
the OI 5577 and OI 6300 Å lines in aurora, Planet. Space Sci.,
Vegard, L., Results of investigations of the auroral spectrum
during the years 1921-1936, Geophys. Publ, 9(11), 1-71, 1932.
Wark, D. Q., Doppler widths of the atomic oxygen lines in the
airglow, Astrophys. J., 131, 491-501, 1960.