A Large Survey of Electron Acceleration Events

The probability of electron acceleration events as a function of MLAT and MLT

Presented bilingually. Click to toggle: Scientific Presentation Plain English.

Here statistical abstracts of the electron acceleration events which cause the discrete aurora are presented. Some instantaneous views of the oval from satellite images can be found in Story 2

Precipitation data from all DMSP satellites (DMSP F6 through DMSP F10) over a 9 year interval were surveyed for electron acceleration events. Only instances where the acceleration is 3-4 times larger than the thermal energy were selected. Such intense events can be identified by their "monoenergetic" spectra (which really means a sharp drop off above the spectral flux peak):

An example of the type of spectra selected for statistical study herein

The following spectrogram has white vertical lines underneath the x-axis indicating where the program identified electron acceleration events. Both large scale and small scale events exist.

In all more than 1.5x10**8 individual spectra were examined for signs of such substantial field-aligned acceleration. With no separation by IMF or any other parameter the results are as shown below:

The chance of electron acceleration events above 0.25 ergs/cm2/s with no selection by IMF.

As expected, a strong IMF dependence exists. For southward IMF, many more electron acceleration events are observed:

The same, except for southward IMF

For northward IMF, the oft debated "1400 hot spot" stands out clearly, although it is actually centered around 1500 MLT:

The same, for northward IMF.

A weaker morning warm spot also exists. These two spots correspond to places in the Ijima-Potemra pattern where upward field-aligned currents maximize; or alternately where velocity shears in the Heppner-Maynard pattern are largest. However for southward IMF these spots are not distinct.

The distributions above all apply to all events above 0.25 ergs/cm**2 s. Very intense events -- those above 1.0 erg/cm**2 s -- are statistically a nightside phenomena as shown below:

The probability of observing electron acceleration events above 1.0 ergs/cm2/s when the IMF is southward. Some data gaps (insufficent crossings for statistical significance) occur between 00 and 01 MLT.

Notice that a difference exists between the occurence patterns of all acceleration events, and those above 1.0 ergs/cm2/s. The latter are more concentrated near midnight, but with an asymmetry still favoring post-midnight. At higher energy flux thresholds the concentration towards the nightside is still greater.

The previous figures demonstrate that aurora are quite variable depending on local time, interplanetary magnetic field conditions, and intensity. One surprising consistency appeared in our study. The latitudinal widths (north/south extent) of aurora have an exponential scale length distribution. This is true independent of local time or interplanetary magnetic field.

The plotted fits are <w>**-1exp(-x/<w>) where < w > is the mean width of the arcs. Note that there are no free parameters in this fit, yet the observed distributions fit this simple functional form (closely related to Poisson statistics) surprisingly well.

Borovosky [1993] considered 22 possible mechanisms for producing the scale sizes of discrete aurora. The two that produce mean widths closest to our values of <w> = 28-35 km are

The ionospheric conductivity feedback mechanism; and
Velocity shears in the central plasma sheet.

Either one of these possibilities might account for why the most intense events are observed on the nightside.

It is mathematically straightforward to construct an exponential distribution if each individual spectra is considered to be an independent trial, uncorrelated with its neighbors. Most researchers would argue that events such as the spectrogram presented above rule out such an interpretation.

Back to main page
Available Data Sets

Please send comments or questions to:

Dr. P. Newell