Modeling and Prediction of Thermosphere Density and Aerodynamic Drag

Abstract:

Forecasting and modeling atmospheric drag is the main challenge when predicting current and future locations of objects in Low Earth Orbit (LEO). Atmospheric drag is proportional to the mass density and composition of the neutrals in Earth’s thermosphere, which is highly variable, both spatially and temporally, mainly due to fluctuating solar UV/EUV emissions and notably geomagnetic activity. Various errors contribute to the total modeling and forecasting error in atmospheric drag:

1. inadequacies of the solar and geomagnetic proxies used, on short time scales in particular,
2. too low temporal resolution of the solar and particularly geomagnetic proxies,
3. sparseness/errors/inconsistencies in the compiled density data used in the model fit,
4. errors in the predictions of the solar and geomagnetic proxies,
5. simple and coarse modeling algorithm in case of semi-empirical models,
6. errors in the satellite model (e.g., shape, mass, and notably aerodynamic coefficient).

Items 1-5 pertain to errors in the prediction of density, i.e., the thermosphere model, but point 6 actually affects the model too.

In the framework of the H2020 project SWAMI funded by the European Commission (EC), which started in January 2018, a new whole atmosphere model (0-1500 km) is under development. The model will be constructed by blending two existing models, the Drag Temperature Model (DTM) and the Unified Model (UM). The CNES thermosphere specification model DTM2013, which was developed in a previous EC project (ATMOP), is being improved by assimilating more density data to drive down remaining biases as a function of solar activity and seasons mainly. The intermediate model DTM2018, which is still based on the Kp index, will be presented. The final DTM model will be constructed in 2019 using high cadence geomagnetic indices, so-called Hp indices.

A short review of the DTM model and the assimilated data will be given, and DTM2013 and DTM2018 performance is evaluated by comparisons with data. Examples of the effects of the listed items 1-5 will be discussed.