Atmospheric Density Model Errors and Variations in the Ballistic Coefficient

Atmospheric density model errors are absorbed as variations in the ballistic coefficient or B term by the differential correction process. Unmodeled forces (e.g., geopotential terms) can also lead to variations in B. For satellites with small energy dissipation rates, observability problems also contribute to variations in B. Monte Carlo simulation is used to determine the accuracy of the variation in the ballistic coefficient in absorbing atmospheric density model errors. The standard deviation of the relative change in B from the least squares differential correction covariance matrix is shown to provide an estimate of the accuracy of the variation of B in absorbing atmospheric density model errors. The accuracy of the variation of B depends on the accuracy of the sensor measurements, the differential correction fit span, and the number of independent observations in the fit span. The length of the fit span is most critical for satellites with B term observability problems.