Subspace identification algorithms have proven efficient for performing output-only identification of the eigenstructure of a linear MIMO system subject to uncontrolled, unmeasured, and nonstationary excitation. Such a problem arises in mechanical engineering, for modal analysis of vibrating structures and machines. A common practice there is to collect data from varying sensor locations, using both fixed and moving sensors, in order to mimic the availability of a larger set of sensors. The purpose of this paper is to investigate how subspace-based identification can be adapted to handle such a situation, to prove its consistency under nonstationary excitation, and to report on a real application example.