Assimilating Storm-time Neutral Winds in Ionospheric-Thermospheric State Estimation

Abstract:

During geomagnetic storms at mid-latitudes both electrodynamic disturbances and neutral composition variations contribute to time evolving and localized variations in the plasma content of the ionosphere. While the most plentiful data are typically Global Navigation Satellite System (GNSS) based measurements of total electron content (TEC), assimilation of measurements of the ionospheric-thermospheric state itself, i.e., neutral winds, can improve the fidelity of the result.

Fabry-Perot interferometers (FPIs) measure emissions of thermospheric oxygen, giving line-of-sight wind speeds. During storm-times however, these FPI measurements may also detect a non-thermal oxygen source, yielding a measurement that is not strictly of the thermospheric wind [Makela et al., 2014]. The sign of non-thermal oxygen is in the apparent large 50 to 100 m/s vertical winds. This raises the question: what happens when we try to assimilate direct measurements of the wind but some of those measurements are “contaminated” by a non-thermal source?

We present results of a Kalman Filtered data assimilative experiment ingesting neutral wind measurements made by a 630.0 nm FPI sited in the mid-latitude U.S. during the geomagnetic storm of October 25, 2011. Ionospheric Data Assimilation 4-Dimensional (IDA4D) estimates time-varying plasma densities from GNSS TEC. These densities are ingested without, and with, respectively, FPI neutral wind data into Estimating Model Parameters with Ionospheric Reverse Engineering (EMPIRE). EMPIRE uses background electric potential and neutral wind models, to produce an optimized estimate of both ExB drift and neutral wind based on the data ingested.

We compare the estimated horizontal neutral wind at the FPI measurement locations at about 250 km altitude, first using electron densities without ingesting FPI data. Then plasma densities plus half the FPI data are ingested to estimate neutral winds. These wind estimates are then compared to the FPI data that were not ingested. During periods of strong apparent vertical winds, the estimate of wind disagrees with the FPI measurement, indicating that the data contamination is indeed non-thermospheric. The assimilated estimate agrees quite well with the measurement actually made at that location when FPI data are only thermospheric.