The thermospheric neutral density response to the 7–9 September 2017 storms is investigated based on the Swarm satellite observations and the thermosphere-ionosphere-electrodynamic general circulation model (TIEGCM) simulation. The Swarm data depicted a prominent interhemispheric asymmetry (IHA) in the afternoon sector during the second storm, a feature that was yet explained. Driven by realistic high-latitude electric potential and electron precipitation patterns, the TIEGCM is able to reproduce the observed storm-time neutral density response. The TIEGCM simulation reveals that the differences in the traveling atmospheric disturbances (TADs) is largely responsible for the observed IHA in the neutral mass density response at low and middle latitudes, whereas the difference in mean molecular mass between the two hemispheres may contribute to the IHA in neutral density at higher latitudes. The IHAs in TADs and mean molecular mass are attributed to the IHA in Joule heating dissipation on the night and dawn sides.

Thermospheric temperature and density recovery during the 5 April 2010 geomagnetic storm has been investigated in this study. Neutral density recovery as revealed by Thermosphere-Ionosphere-Electrodynamics General Circulation Model (TIEGCM) simulations was slower than observations from GOCE, CHAMP, and GRACE satellites, suggesting that the cooling processes may not be fully represented in the model. The NO radiative cooling rate in TIEGCM was also compared with TIMED/SABER measurements along satellite orbits during this storm period. It was found that the model overestimated the NO cooling rate at low latitudes and underestimated it at high latitudes. The effects of particle precipitation on NO number density and NO cooling rate at high latitudes were examined in detail. Model experiments showed that while NO number density and NO cooling rate do change with different specifications of the characteristic energy of auroral precipitating electrons, neutral temperature and density recovery remain more or less the same. The reaction rates of key NO chemistry were tested as well, and the NO number density between 110 and 150 km was found to be very sensitive to the reaction rate of N(²D) + O₂ → NO + O. A temperature-dependent reaction rate for this reaction proposed by Duff et al. (2003) brought the TIEGCM NO cooling rate at high latitudes closer to the SABER observations. With the temperature-dependent reaction rate, the neutral density recovery time became quite close to the observations in the high-latitude Southern Hemisphere. But model-data discrepancies still exist at low latitudes and in the Northern Hemisphere, which calls for further investigation.