This paper introduces the problem of modeling urban transportation systems in a database where certain aspects of the data are probabilistic in nature. The transportation network is composed of multiple modes (e.g., automobile, bus, train, pedestrian) that the user can alternate between. A trip -- a path between an origin and destination subject to some constraints -- is the central concept. How these trips and the network can be represented as both a graph and relational model, as well as the requirements for querying are the main contributions of this paper. A set of operators are defined to work over these transportation concepts and they are integrated within a SQL-like syntax to express queries over the uncertain transportation network. Additionally, the paper shows how this model can be integrated within other moving objects and spatio-temporal data models, and how these graph-based queries can be processed.