Abstract
Incoherent scatter (IS) radars are invaluable instruments for
ionospheric physics, since they observe altitude profiles of electron
density ($N_e$), electron temperature ($T_e$), ion temperature
($T_i$) and line-of-sight plasma velocity ($V_i$) from ground.
However, the temperatures can be fitted to the observed IS spectra only
when the ion composition is known, and resolutions of the fitted plasma
parameters are often insufficient for auroral electron precipitation,
which requires high resolutions in both range and time. The problem of
unknown ion composition has been addressed by means of the full-profile
analysis, which assumes that the plasma parameter profiles are smooth in
altitude, or follow some predefined shape. In a similar manner, one
could assume smooth time variations, but this option has not been used
in IS analysis. We propose a plasma parameter fit technique based on
Bayesian filtering, which we have implemented as an additional Bayesian
Filtering Module (BAFIM) in the GUISDAP analysis package. BAFIM allows
us to control gradients in both time and range directions for each
plasma parameter separately. With BAFIM we can fit F$_1$ region ion
composition together with $N_e, T_e, T_i$ and $V_i$, and we have
reached 4~s/900~m time/range steps in
four-parameter fits of $N_e, T_e, T_i$ and $V_i$ in E region
observations of auroral electron precipitation.