The hypercube model allows the use of spatially distributed queues in probabilistic location models. It is not unusual that its parameters, considered constants, to be time-dependent in real situations. In this context, this paper presents a nonstationary hypercube model and how to model change of shifts under a preemptive discipline. Moreover, we used an illustrative example to compare the proposed model with a discrete-event simulation model and a stationary approximation in order to verify the accuracy and to find benefits for using the proposed model. Results show that the nonstationary model can handle rush hours without over or underestimate its performance measures even on a low event frequency system. One may use the nonstationary hypercube model as a benchmark for modeling shift-scheduling problems on a spatially distributed situation.