 |
| Titel |
Vlasov simulations of magnetic field-aligned potential drops |
| VerfasserIn |
Herbert Gunell, Johan De Keyser, Emmanuel Gamby |
| Konferenz |
EGU General Assembly 2011
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 13 (2011) |
| Datensatznummer |
250054345
|
|
|
|
|
|
| Zusammenfassung |
| In the auroral zone, electrostatic fields that are parallel to Earth’s magnetic field are known to
exist and to contribute to the acceleration of auroral electrons. Parallel electric fields
can be supported by the magnetic mirror field, creating potential drops extending
over great distances (Alfvén and Fälthammar, Cosmical Electrodynamics, 2nd ed.,
1963).
The current-voltage characteristics of the auroral current circuit was studied by Knight
(Planet. and Space Sci., vol. 21, 741-750, 1973) using a stationary kinetic model. Later, fluid
and hybrid models have been used in the study of potential drops and of Alfvén waves and
their relation to the formation of potential drops (e.g., Vedin and Rönnmark, JGR, vol. 111,
12201, 2006).
Recent observations have shown that field-aligned potential drops often are
concentrated in electric double layers (e.g. Ergun, et al., Phys. Plasmas, vol. 9, 3685-3694,
2002). Vlasov simulations of the part of the flux tube where most of the auroral
acceleration takes place have recently been performed (Main, et al., PRL, vol. 97, 185001,
2006).
We present results from Vlasov simulations using a model that is one-dimensional in
configuration space and two-dimensional in velocity space. We verify the model by
comparison with a double layer experiment in the laboratory, and the model is then applied to
the auroral field lines.
A flux tube is modelled from the equator to the ionosphere. The spatial grid is
non-uniform, finer close to the ionoshere and coarser near the equator, in order to improve
computational efficiency. We can run the simulation on a coarser spatial grid and
with a longer time step by introducing a relative dielectric constant εr such that
ε = ε0εr, because λD ~- εr and Ïp ~ 1--εr (Rönnmark and Hamrin, JGR, vol. 105,
25333–25344, 2000). We start with a large εr value, filling the simulation region with
plasma from the ends. We then conduct a series of simulation runs, successively
decreasing εr toward realistic values. Thus, we arrive at a solution, without relying
on assumptions about the distribution function in the interior of the simulation
region. |
|
|
|
|
|
|