Simulating neutral biodiversity in large-scale networks

The neutral model of biodiversity describes a Markov process of death and replacement of neutral individuals constituting the species occupying an ecosystem. It predicts a number macroscopic ecosystem observables, including local and global abundances, and can serve as advanced null-hypothesis for more complex models. We present a massively parallel approach to simulating the neutral model which both enables simulations on large networks, to, e.g., capture fine details of real river networks, and efficiently produces averages results to enable comparisons with empirical data or other models at high statistical significance or accurately compute correlations or responses.