Determining optimal ambulance locations is a critical decision in the planning of emergency medical services as it enables emergency vehicles to reach the patients quickly, which can improve patient survival probabilities and reduce morbidities. In this work, we consider the problem of strategic ambulance location considering the impact of variations in travel time and demand over the period of a day. The proposed model considers a continuous survival function-based objective and incorporates station-level service rate, arrival rate, and the busy probability of ambulances. We formulate a mixed-integer non-linear programming model to represent the problem and develop a variable neighbourhood search-based solution approach that uses the solution from a relaxed mixed-integer linear programming model to solve the problem. The proposed approach is demonstrated on test instances developed based on an urban location in India. The travel time variations were estimated from the Uber movement dataset available for the city. Our results indicate that although incorporating temporal variation provides a better estimate of coverage, survival function and ambulance requirements, dividing the planning horizon into a larger number of periods results in diminishing returns.