With the proliferation of wireless communication and sensor technology, the importance of location-based applications has tremendously increased. In order to support these applications, services that implement algorithms for the efficient processing of spatial queries, such as range and k-nearest neighbor queries, are definitely beneficial. In this paper, we propose definitions of probabilistic range and k-nearest neighbor query semantics that take into account the inaccurate position information acquired from positioning systems. We introduce efficient algorithms for distributed storage of the dynamic information captured by positioning sensors on mobile network nodes, and for processing the aforementioned queries in mobile ad hoc networks. We show by evaluation that the studied algorithms incur small communication costs, and that the query algorithms return accurate results in a timely manner.