|
Assimilation and Modeling of Thermospheric Neutral Density Using Observations from LEO Satellites
|
Ghazaleh Ghazavi   , Saeed Farzaneh * , Mohammad Ali Sharifi |
| University of Tehran |
|
|
Abstract: (307 Views) |
Global and multi-level estimation of thermospheric neutral density (TND) is essential for upper atmosphere studies and for accurately calculating drag forces acting on satellites in Low Earth Orbit (LEO), with direct applications in precise orbit determination. Empirical models are often employed to forecast variations in TND for orbit prediction experiments. However, their performance can be enhanced by addressing the inherent limitations in model structure and the sparse spatiotemporal sampling of input data. In this study, a calibration algorithm is developed to assimilate freely available TND measurements derived from CHAMP and GRACE satellite accelerometer data into the empirical model NRLMSISE-2.0. A primary focus is placed on thermospheric parameters—specifically neutral density—which, while estimable via physical or empirical models, often lack the required accuracy for space applications. Improving these empirical models remains a significant challenge in thermospheric science. Satellite motion is highly sensitive to atmospheric drag acceleration, which is a function of thermospheric density. To enhance the accuracy of TND estimations, satellite-derived data can be assimilated into empirical models, allowing for calibration of key parameters and improved predictive capability during data-sparse intervals. Consequently, this study aims to enhance empirical thermospheric models by calibrating their parameters using various geodetic observations, particularly neutral density data from LEO satellites. The methodology involves extracting key model parameters and calibrating them using accelerometer-derived mass densities along the orbital tracks of CHAMP and GRACE. Model correction and enhancement are performed using spherical harmonic expansion in geographic latitude and local solar time. The improved empirical model is then evaluated for its impact on orbit determination. Results show that the enhanced model induces observable changes in satellite orbits and improves model performance based on three evaluation indices: ARD, R, and LD. Validation with 10% of the dataset yields RMSE-based improvements of 40.23% and 45.80% for CHAMP and GRACE, respectively.
|
|
| Keywords: Thermospheric Neutral Density, Empirical Model, Calibration, Geodetic bservations |
|
|
|
Type of Study: Research |
Subject:
Geodesy
Received: 2025/04/24 | Accepted: 2025/12/13 | ePublished ahead of print: 2026/01/31
|
|
|
|
|
|
|
| Send email to the article author |
|
|
