SATELLITE drag coefficients are the primary source of un-certainty in atmospheric drag prediction, as well as in the deduction of atmospheric densities from satellite drag. The drag coefficient is heavily dependent on the interaction of the atmospheric gas with the satellite surface. The energy-accommodation coefficient is a key factor in this interaction and can be thought of as the average fraction of energy lost by molecules impinging on the surface. This parameter has previously been determined empirically for satellites orbiting in the Earth’s thermosphere, and tabulated values exist as a function of altitude for solar-maximum and solarminimum conditions [1–3]. At the time of writing, the accommodation data represent the state of the art in computing drag coefficients that respond correctly to the changing environment at a range of altitudes in the upper atmosphere. When applying the existing information, however, an altitude fit is performed to a set of values that may not correspond to the solar conditions in question. The purpose of this paper is to describe the first energy-accommodation model consistent with satellite-accommodation data that can be applied at all levels of solar activity while taking into account some of the basic physics of gas–surface interactions. Rather than compiling a series of ad hoc fits along the altitude component at a variety of atmospheric conditions, the formulation described herein has a theoretical underpinning that makes it generally applicable while fully leveraging the available data. It enables the computation of between all solar-maximum and solar-minimum conditions, as well as extrapolation to extreme space weather behavior, such as the deep minimum of solar cycle 23.

Early laboratory measurements of gas–surface accommodation have been fraught with uncertainties caused mainly by the conditions and cleanliness of the surface [4]. Furthermore, other researchers have noted that conditions on the satellite surface are not represented adequately in the laboratory [5]. One of the reasons for this discrepancy is that atomic oxygen in the thermosphere adsorbs onto satellite surfaces. Instead of interacting with a metallic lattice or the glass surface of a solar cell, incoming molecules collide with the lighter adsorbate species, resulting in a more inelastic collision. Atomic oxygen adsorption has been confirmed by observation from pressure gages [6] and is the primary reason for the altitude dependence of the drag coefficient [3, 7]. The pressure gauge measurements suggest that time scales for adsorption and desorption and, therefore, changes in energy accommodation could be on the order of seconds. This response time has not been measured in the context of drag coefficient changes, so it is not possible to verify its applicability at the present time. The determination of characteristic time scales for changes in satellite accommodation is beyond the scope of this paper and left as a subject of future work.