Near-Earth satellites travel through a rarefied atmospheric
medium. Atmospheric density varies with altitude and is highly dependent
on solar heating and geomagnetic activity. The relative
elemental constituents (N
, N, O, O
, H, He, Ar) also vary
with height and geographical location further complicating accurate
modeling of drag forces.
Atmospheric drag models commonly in use for these calculations include the DTM model [ Barlier et al., 1987] and the MSIS model [ Hedin, 1987]. The state of the current atmospheric density models for satellite drag modeling is reviewed by Marcos et al., [1993] where the influence of surface properties was also evaluated. These models are based on either in situ atmospheric spectrometer measurements (MSIS) or satellite orbit dynamics (DTM). Although these models are extensively utilized, they suffer from incomplete global coverage, long time constants requiring a great deal of averaging, and aliasing from other unmodeled non-conservative forces for models estimated from satellite tracking data. These models are also undersampled at geodetic satellite altitudes (>750 km) and during times of high solar and/or geomagnetic activity where they produce density profiles based largely on extrapolations. Therefore, in order to achieve the accuracy needed for precision orbit determination, it is a common practice to solve for several drag scaling parameters to better model the observed satellite motion.
Until recently, the combination of satellite dynamical and in situ measurements had not been attempted; however, on-going model development is focused on producing combination solutions [ Cunningham et al., 1994]. This effort attempts to improve the MSIS model at higher altitudes through the inclusion of dynamically reduced geodetic satellite data. Nuth [1991] utilized the SLR tracking on Starlette and Ajisai in an attempt to improve the density modeling at geodetic satellite altitudes. A horizontal wind model [ Hedin et al., 1991] along with accurate spacecraft attitude and non-conservative force models are being developed to decouple the drag density signal from other forces. Satellite data combined with atmospheric wind data could then be used to increase the accuracy of wind field models.
There remain several weaknesses in state-of-the-art drag modeling. Currently, no atmospheric wind effects are being considered within these models nor are they being applied externally in orbit drag computations. Even from what little is known about the mean wind fields at satellite altitudes, the assumption that the atmosphere rotates with the Earth is clearly invalid, especially towards the poles [ Berger and Barlier, 1991]. It is also a common practice to compute only the along track drag acceleration based on the computed projected area in the velocity direction. Out of plane drag forces are thereby neglected. Modeling these effects and/or estimating drag coefficients in the off-velocity directions can be used to further enhance modeling. These approaches are discussed in Ries et al., [1993a].