Premium
How uncertainty in the neutral wind limits the accuracy of ionospheric modeling and forecasting
Author(s) -
David Michael,
Sojka J. J.,
Schunk R. W.
Publication year - 2016
Publication title -
journal of geophysical research: space physics
Language(s) - English
Resource type - Journals
eISSN - 2169-9402
pISSN - 2169-9380
DOI - 10.1002/2015ja021544
Subject(s) - ionosphere , earth's magnetic field , solar wind , environmental science , tec , atmospheric sciences , meteorology , latitude , physics , total electron content , geophysics , computational physics , plasma , magnetic field , astronomy , quantum mechanics
One of the most important input fields for an ionospheric model is the horizontal neutral wind. The primary mechanism by which the neutral wind affects ionospheric densities is the inducement of an upward or downward ion drift along the magnetic field lines; this affects the rate at which ions are lost through recombination. The magnitude of this effect depends upon the dip angle of the magnetic field; for this reason, the impact of the neutral wind is somewhat less in polar regions than at mid‐latitudes. It is unfortunate that observations of the neutral wind are relatively scarce, as compared for example with observations of the Earth's electric field or auroral precipitation, and that the existing climatological models of the neutral wind are thus sharply limited in theirresolution. The observational data base of thermospheric winds is not sufficient to adequately constrain a three‐dimensional model across a variety of conditions such as solar cycle, season, geomagnetic activity, and so on. Using the physics‐based Time Dependent Ionospheric Model (TDIM) of Utah State University, we look for a quantitative answer to this question: How severe is the limitation imposed on ionospheric models by an uncertain specification of the neutral wind? We find that ionospheric modeling depends upon a detailed specification of the neutral wind to the extent that, if a climatologically averaged wind model is being used as a driver, this will lead to unavoidable uncertainties of 20‐30% in the modeled F‐region densities or Total Electron Content (TEC).