Discrete Aurora Are Suppressed in Sunlight

Presented bilingually. Click to toggle: Scientific Presentation Plain English

(Newell et al., Nature, 1996 )

Out of the 22 theories for discrete aurora reviewed by Borovosky [1993], only one, that involving the ionospheric conductivity feedback mechanism, seemed to give latitudinal scale sizes that are at all appropriate (most predict tens or even hundreds of km, but the most frequent size we observe is a single spectrum, about 6 km in north/south extent). It occurred to us that an instability based on ionospheric conductivity feedback would be unlikely to operate if the precipitation were not the dominant source of conductivity.

Therefore we investigated whether sunlight had any effect.

The Data Base:

152 million individual electron spectra taken over a 9 year interval from 5 different DMSP series satellites were examined. (Dave Hardy of Phillips Lab is P.I.). These spectra were examined for signs of field-aligned acceleration, namely "monoenergetic" peaks [Evans, 1968]. Below is a sample spectrum flagged:

A monoenergetic spectrum of the type surveyed

Such spectra are observed only when the accelerating potential is at least 2-3 times the thermal energy of the electrons. Therefore ours is a study of electron acceleration events rather than direct auroral observations. It is not possible that intense localized fluxes of electrons into the ionosphere would fail to produce intense localized emissions of light (discrete aurora), and in any event the observational coupling of the two is now well established.

The results:

The first results of our survey [Newell et al., JGR 1996] showed that electron acceleration events occur at all local times:

Probability of observing electron acceleration events as a function of MLT and MLAT. All events above 0.25 ergs/cm2 s are included.

Although there is a broad minimum around noon (the gap is actually centered about 11 MLT), electron acceleration events can be observed at all local times, and almost as often on the dayside as the nightside. But the situation changes radically when INTENSE electron acceleration events are considered:

Same as previous except only for events above 5.0 ergs/cm2 s. Notice that the great majority of such events occur in the dusk-midnight sector.

The monoenergetic peak value gives the accelerating potential. It turns out that the dayside accelerating potentials are rarely above 1 keV, and generally only several hundred eV. However in the dusk to midnight sector, many keV or even a few tens of keV accelerating potential can be found. It is this difference which causes the difference in the relative occurrence of intense energy fluxes.

To directly address the role of sunlight, we binned the acceleration events according to solar zenith angle. A solar zenith angle greater than 90 degrees means the observation is in darkness, while less than 90 degrees means the observation is in sunlight. Compare the two following figures, which respectively give the chance of finding intense electron acceleration events in sunlight and in darkness:

Probability of observing e- acceleration events above 5.0 ergs/cm2 s in darkness.

Same as previous figure, except in sunlit conditions.

Discussion: Why are Intense Events Concentrated in the Dusk-Midnight Sector?

We know that auroral arcs are associated with field-aligned currents, and that the acceleration occurs in the range 1-3 RE. But currents are stronger on the dayside than the nightside. It is interesting to consider the quiet time picture to really bring this out. Here is the probability of observing electron acceleration events above 0.25 ergs/cm2 s for northward IMF:

Acceleration events for Bz>0 (>0.25 ergs/cm2 s)

The peak around 1500 MLT corresponds to the point of the largest currents out of the ionosphere according to Iijima and Potemra [1978], and also to the largest convection shears in statistical patterns [Heppner and Maynard, 1987]. A lesser shear occurs in the morning region. But what accounts for the nightside maxima? And with all the dynamics of the dayside, why do so few little electron acceleration events reach large values?

As already mentioned, we believe it is because the dusk-midnight region is peculiarly sensitive to the conductivity feedback mechanism. To make this more explicit, we calculated the ionospheric conductivity from its two main contributors, namely the diffuse aurora and sunlight. For diffuse aurora we used our own particle observations (with the electron acceleration events discarded) and the equations of Robinson et al. [1987]. For sunlight we used Rasmussen et al. [1988]. Superposed are the statistical currents from Iijima and Potemra [1978] (without any adjustments -- the coordinate systems are almost identical):

Magnetosphere-ionosphere coupling requires large scale currents out of the ionosphere (closed by currents in the ionosphere). Such currents require ionospheric conductivity. Because hot keV electron curvature and gradient drift towards dawn, there is an asymmetry about midnight for this population (lower energy electrons do not contribute to ionospheric conductivity because they do not reach low altitudes where collisions allow currents to flow). The figure clearly shows that the only region where currents are required but the conductivity does not exist to support them is the region where intense electron acceleration events are routinely seen.

Acknowledgements: This work was supported by the National Science Foundation grant ATM-9531489 (Aeronomy). Dave Hardy of the Air Force Research Laboratory is the effective P.I. on the DMSP SSJ/4 particle detectors. We thank R. Lepping, J. King, and the Goddard NSSDC for the IMP-8 data.

Return to Auroral Particles and Imagery Page

Return to Main Stories Page

Examine Available Data Sets

Send science questions/comments to Dr. P. Newell
Send WWW questions/comments to webmaster