A three-dimensional map of an environment with buildings is used to computationally predict locations and times of global navigation satellite system (GNSS) blockages. For example, in urban environments some of the GNSS satellites are occluded by buildings. These blockages can be modeled. A computing system can make a map showing which satellites are or are not visible as a function both of location and time. The map can be used by a mobile GNSS receiver to determine which satellites to use or whether to use a backup system for navigation. The system can determine when a given satellite will enter or leave a GNSS receiver view during a route. The map can be stored in the GNSS receiver (or a host of the GNSS) or can be stored by a network service. This mapping can be used to predict multi-path effects of a satellite transmission at a location.