Towards thermospheric density estimation from SLR observations of LEO satellites: a case study with ANDE-Pollux satellite.

The present contribution investigates the possibility to obtain thermospheric neutral density estimates using satellite laser ranging (SLR) observations of low Earth orbiters (LEOs). This approach is based on the analysis of the satellite atmospheric drag, driven by the fact that the drag force is the largest non-gravitational perturbation acting on LEOs. Due to the uncertainty of current thermospheric models, it is the main error source in the LEO orbit determination process. Moreover, the drag is physically related to the thermospheric density distribution, the interaction of the satellite surface with the surrounding thermosphere and thermospheric winds. For this investigation, a spherical satellite called "Atmospheric Neutral Density Experiment-Pollux" (ANDE-P) developed by the Naval Research Laboratory (USA) is adopted as a case study. The satellite flew at the very low altitude of about 350 km. The most important perturbing acceleration at this altitude, the atmospheric drag, is easier to model for a spherical satellite like ANDE-P than for a satellite of complex geometry. A precise orbit determination of ANDE-P was performed with the DGFI Orbit and Geodetic parameter estimation Software (DOGS) over a period of 49 days (16 August 2009 until 3 October 2009) using SLR observations to this satellite and different thermospheric models. In total, we tested four thermospheric models, namely CIRA86, NRLMSISE00, JB2008 and DTM2013. Correspondingly, scale factors of these reference models are estimated with a 6-h resolution. The results confirm that the estimation of force model parameters from SLR measurements of the ANDE-P satellite is sensitive to differences in the density distributions provided by different models. As a consequence, information on the discrepancies between the various models and the true density can be derived from SLR measurements. Moreover, it is found that SLR observations to LEO satellites at very low altitudes are capable to estimate corrections to (scale factors of) the integrated thermospheric density if all other perturbing accelerations are modelled with sufficient accuracy. We derived time series of estimated scale factors of the thermospheric densities provided by the models and obtained the following mean values during the processed period of time at the ANDE-P altitude: 0.65±0.26 for CIRA86, 0.65±0.25 for NRLMSISE00, 0.79±0.24 for DTM2013, and 0.89±0.27 for JB2008. This suggests that all models overestimate the true thermospheric density along the ANDE-P trajectory during the processed period to a certain extent. The thermospheric densities need to be scaled downwards to fit ANDE-P SLR observations with JB2008 requiring the least amount of scaling. [ABSTRACT FROM AUTHOR]