Model Bias Characteristics for Data Assimilation in the Thermosphere

Physics-based models have been used over the past 30 years to improve our understanding of dynamics and driver-response relationships in the upper atmosphere and ionosphere. Aided by data assimilation techniques, several physics-based models of the ionosphere have begun to surpass empirical models in specification accuracy. In principle, these techniques can also be applied to the thermosphere, however, the scarcity of neutral data is a challenging obstacle to overcome. As a first step, we treat the output of the High Accuracy Satellite Drag Model (HASDM) as data for assimilation into the Thermosphere-Ionosphere-Electrodynamics General Circulation Model (TIEGCM). This method has the advantage of simulating excellent data coverage from an empirical model that has been "dynamically calibrated" using ground-based satellite tracking data. Validation of the technique is performed by comparing the corrected TIEGCM output with CHAMP and GRACE total mass density data for both quiet and moderately disturbed periods in April and May, 2004. During this time, the orbital plane separation between CHAMP and GRACE is approximately 90 degrees, yielding data coverage at four distinct local times. The critical issues of thermosphere modeling bias characteristics are addressed and discussed.