Effect of the altitudinal variation of the gravitational acceleration on the thermosphere simulation

[1] For simplicity, a constant gravitational acceleration (~g) is assumed in many general circulation models (GCMs). To estimate the influence of the altitudinal variation of the ~g on the thermosphere simulation, two runs have been made under the solar maximum condition using the non-hydrostatic Global Ionosphere Thermosphere Model (GITM), including one with a constant ~g (8.7 m/s) and the other with an altitude-dependent ~g. During geomagnetic quiet time, globe averagely, the constant ~g case overestimates the neutral density by 30% and underestimates the temperature by 10% around 120 km altitude compared with the altitude-dependent gravitation case. A post-processing then has been done to the constant ~g case by shifting the atmosphere vertically according to the altitude-dependent~g. After this shifting, the global average density difference decreases to 10%, and the temperature difference also changes correspondingly to 2% at 120 km. The~g specification directly causes the vertical shift of the atmosphere through changing the scale height. Meanwhile, it changes the temperature profile, which feeds back to the altitude profile of the neutral density. In order to separate these two effects, three simple tests with one-dimensional semi-realistic atmosphere have been conducted, and the primary results are that the vertical shifting and the temperature variation caused by the ~g specification contribute 20% and 25% of the density difference, respectively. This study gives a reference to other GCMs about the validity of the constant gravitational acceleration assumption.