Abstract. We report on a study of three intense ionospheric storms that occurred in September 1989. Using Dst as a reference for storm onset and subsequent main and recovery phases, we analyze the observed worldwide responses of F region heights hmF 2 and densities NmF 2 as a function of universal and local times, latitudinal domains, and storm onset-times; and we compare the characteristics of all three storms. The following points are among the major findings: (1) The negative phase storm was the dominant characteristic, with the greatest intensity occurring in the regions which were in the nighttime hemisphere during the main phase; (2) at middle and low latitudes negative phase characteristics were observed first in the nighttime hemisphere and then corotated with the Earth into the dayside; (3) the most intense negative response occurred in the recovery phase; (4) observations of the negative phase characteristics supported thermospheric upwelling, increased mean molecular mass, and an associated enhancement in dissociative recombination as the principal cause-effect chain; but the observations suggest greater ion-neutral chemistry effects than accounted for in current models; (5) hmF2 was observed to respond quickly to the storm onset (pointing to the importance of electric fields) with enhanced values in all latitudinal and local time domains; (6) positive storm characteristics were among the issues most difficult to reconcile with current descriptions of cause-effect relationships; and (7) the analysis of all storm phases and comparisons with several modeling efforts show that future advances in understanding require a more accurate accounting of the influences of magnetospherically-imposed and dynamo-driven electric fields, plasmaspheric fluxes, and vibrationally excited N 2.

No abstract

No abstract

No abstract

Satellite and ground‐based observations from March 28 to 29, 1992, were combined in the assimilative mapping of ionospheric electrodynamics (AMIE) procedure to derive realistic global distributions of the auroral precipitation and ionospheric convection which were used as inputs to the National Center for Atmospheric Research (NCAR) thermosphere‐ionosphere general circulation model (TIGCM). Comparisons of neutral model winds were made with Fabry‐Perot measurements and meridional winds derived from ionosondes. The peak equatorward winds occurred 1–2 hours later in the model. Gravity waves launched from high‐latitude Joule heating sources reached the equator in about 2 hours and agreed with observed variations in the height of the maximum electron density (hmF2) and in the meridional winds. Joule heating events produced minima in the O/N2 ratio that moved equatorward and usually westward in longitudinal strips which lasted about a day. Changes in the O/N2 ratio and in the peak electron density (NmF2) were strongly correlated so the observed daytime NmF2 values for stations near 50° magnetic latitude were generally reproduced by AMIE‐TIGCM on the second day of the simulation. The AMIE‐TIGCM underestimated the electron density after midnight by up to a factor of 2 in midlatitudes, while the modeled F2 layer was about 35 km lower than the observations at midnight. Shifting the model winds 2 hours earlier at night could double the NmF2 at 0400 LT and increase hmF2 by 20 km. NmF2 could also be increased at night by realistically increasing the TIGCM nighttime downward fluxes of O+ at the upper boundary.