CHAMP and GOCE thermospheric wind characterization with improved gas-surface interactions modelling

The CHAMP and GOCE satellites provided high-resolution thermosphere data between 2000 and 2013, improving our knowledge of atmosphere dynamics in the thermosphere-ionosphere region. However, the currently available data sets contain inconsistencies with each other and with external data sets and models, arising to a large extent from errors in the modelling of aerodynamic forces. Improved processing of the wind data for the two satellites would benefit the further development and validation of thermosphere models and improve current understanding of atmospheric dynamics and long-term trends. The first step to remove inconsistencies has been the development of high-fidelity models of the satellite surface geometry. Next, an improved characterization of the collisions between atmospheric particles and satellite surfaces is necessary. In this article, the effect of varying the energy accommodation coefficient, which is a key parameter for describing gas-surface interactions (GSI) is investigated. For past versions of the thermosphere density and wind data from these satellites a value of the energy accommodation coefficient of α_E=0.93 was selected. The satellite accelerometer measurements, from which the thermospheric data are derived, have now been reprocessed using high-fidelity geometries and a wide range of α_E values. Lowering the α_E value used in the processing leads to an increase in the lift over drag ratio for those satellite panels that are inclined to the flow. This changes the direction of the modelled acceleration, and therefore the interpretation of the measured acceleration in terms of wind. The wrong choice of α_E therefore leads to the introduction of satellite attitude-dependent wind errors. For the CHAMP and GOCE satellites, we have found that values of the energy accommodation coefficient significantly lower than 0.93 (0.85 for CHAMP and 0.82 for GOCE) result in increased consistency of the wind data. A comparison between the two missions and an overview of the influence on the results of filtering for solar activity and seasonal and diurnal variations is presented.