The successful design of an Optical Deep Space Network (ODSN) greatly depends on the selection of optimal telescope sites. At the highest system level, there are two main factors to consider in the design of a global optical communications network for deep space applications: telescope size (i.e., aperture size) and the distance between stations. The size of the individual telescope aperture needs to be selected based on mission needs (e.g., maximization of received photons per hit). At the same time, because of weather effects and Earth rotation, a number of telescopes have to be placed within certain distances around the Earth in order to achieve global coverage. The distance between the adjacent telescopes is driven by other secondary factors, which are basically derived requirements from: 1) outage tolerance; 2) continuity in data stream; 3) operational cost; and 4) minimal requirements on the spacecraft payload design. To perform properly, ground stations must he placed on highaltitude peaks (for better visibility and high atmospheric transmission) around the Earth. However, the scarcity of peaks, along with geopolitical issues, may cause difficulties in the selection of the telescope sites in a global network. In an optical deep space link, the characterization of the atmospheric channel requires great attention. In fact, cloud opacity is the first evident impairment to the successful closure of a space-to-ground (and vice versa) optical link. Likewise, aerosol distribution in the atmosphere can significantly increase the optical thickness of the atmosphere with a detrimental attenuation of the laser signal. Moreover, an optical communicatiodtracking network must operate during daytime, and in this case, an increase of background sky radiance can dramatically affect the receiver performance by increasing system noise.In this paper, therefore, we present an analysis of site selection for an optical deep space network as performed by the ODSN study team at JPL. Given a set of mission requirements, we illustrate how the high-level requirements, along with the properties of the atmospheric channel, can be used to determine the site selection and the architecture of an ODSN. Moreover, we characterize candidate sites for a global optical network and their possible suitability fur global architectures such as the linear dispersed optical subnet (LDOS) and cluster optical subnet network (COS).
