Molecular motors drive the directed transport of presynaptic vesicles along the narrow axons of nerve cells. Stationary clusters of such vesicles are a prominent feature of axonal transport, but little is known about their physiological and functional relevance. Here, we develop a simulation model describing key features of axonal cargo transport with a view to addressing this question, benchmarking the model against our experiments in the touch neurons of C. elegans. Our simulations provide for multiple microtubule tracks and varied cargo motion states while also incorporating cargo-cargo interactions. Our model also incorporates obstacles to vesicle transport in the form of microtubule ends, stalled vesicles, and stationary mitochondria. We devise computational methodologies to simulate both axonal bleaching and axotomy, showing that our results reproduce the properties of both moving as well as stationary cargo in vivo. Increasing vesicle numbers leads to larger and more long-lived stationary clusters of vesicular cargo. Vesicle clusters are dynamically stable, explaining why they are ubiquitously seen. Modulating the rates of cargo motion-state switching allows cluster lifetimes and flux to be tuned both in simulations and experiments. We demonstrate, both in simulations and in an experimental system, that suppressing reversals leads to larger stationary vesicle clusters being formed while also reducing flux. Our simulation results support the view that the physiological significance of clusters is located in their role as dynamic reservoirs of cargo vesicles, capable of being released or sequestered on demand.