Research on connected and autonomous vehicles (CAVs) is moving towards first deployments around the world. For complete vehicle autonomy, on top of sensors there is a need for an effective communication system. Due to the critical safety, transmission requirements for these communications are stringent. In an urban environment, with high density of vehicles, standardized dedicated short range communications (DSRC) solely does not perform well. Avoiding costs for new DSRC infrastructure, heterogeneous networks integrating long term evolution (4G-LTE mobile network) and DSRC have shown promising results. With the ever increasing number of vehicles, an optimal integration is required in order to balance the capacity load on the two networks. This paper proposes a systematic approach to designing multitier heterogeneous adaptive vehicular (MHAV) networks. With extensive system level simulations modeling Glasgow city center, incorporating proposed algorithms, scaling of the network along with load balancing between LTE and DSRC have been investigated in this paper. With the design criteria proposed for MHAV, results show that under realistic conditions the probability of end-to-end communication delay to be less than 50 ms is above 90% for a density of 250 vehicles/km 2 .