Rapid recovery of thermosphere density during the October 2003 geomagnetic storms

Thermosphere densities from the CHAMP and GRACE satellites are utilized, for the first time, to study the recovery of the thermosphere as a function of latitude during the October 2003 storms. Our results show that the relaxation times, defined by the e-folding time of the poststorm recovery of thermosphere density, are about 6 and 8 h for two recovery phases of the October 2003 superstorms, respectively. Geomagnetic activity index Kp or ring current index Dst is incapable of describing the rapid recovery of thermosphere density. Moreover, a weak altitudinal dependence of the relaxation times was observed between the CHAMP and GRACE altitudes at middle and high latitudes, but no coherent latitudinal dependence was found. The MSISE00 and TIEGCM neutral densities are compared with the observations to assess their capability in predicting the thermosphere response during the recovery phase of extremely severe storms. Neither the MSISE00 nor the TIEGCM reproduced the rapid recovery of thermosphere densities seen in the CHAMP and GRACE, although the TIEGCM captured most of the salient features observed by CHAMP and GRACE when AMIE convection and precipitation patterns were used to specify the high-latitude drivers. The relaxation times of the MSISE00 and TIEGCM nighttime densities at 390 km are generally longer than those from the CHAMP observations by about 4 h, and even longer in the geographic latitudinal range of 25°N - 50°N. The TIEGCM recovery times of thermosphere density are shorter on the dayside than on the nightside, whereas the MSISE00 densities show substantially longer relaxation times at low latitudes. Thus, not only are the relaxation times of the MSISE00 and TIEGCM densities longer than observed in the CHAMP and GRACE data, but they also show much larger day-night differences. No clear explanation can be found to fully understand the causes for the slower recovery of the thermosphere density simulated by the TIEGCM.