Scalable, reliable underwater communication is still a major challenge limiting the deployment of large-scale cooperative underwater systems, such as swarms of AUVs, which generally rely on frequent updates from surrounding units for control and mission objectives. Communication bandwidth is extremely limited in the underwater domain, hence it is crucial to efficiently manage and distribute this scarce resource between nodes. This paper discusses a symmetric, distributed time-slotted channel access algorithm for large-scale, dynamic ad-hoc networks. In contrast to traditional network algorithms which are mostly designed for sporadic point-to-point communication, this algorithm specifically optimises fast, continuous information distribution, locally and globally, while achieving close to 100% channel utilisation. A theoretical analysis of the performance is presented, along with real-time simulation results for dozens of nodes performing rapid reconfigurations. The impact of swarm dynamics on network performance is analysed in detail.