A stochastic eco-hydrological model reveals the impact of active stream dynamics and connectivity on metapopulation

The active portion of river networks varies in time thanks to event-based and seasonal expansion-retraction cycles that mimic the unstradiness of the underlying climatic conditions. These rivers, usually referred to as temporary streams, constitute a major fraction of the global river network. Temporary streams provide a unique contribution to riverine ecosystems, as they host unique habitats that promote biodiversity. Nonetheless, the impacts of network dynamics on ecological processes and ecosystem services are not fully understood. In this contribution, we present a stochastic framework for the coupled simulation of active stream dynamics and the related occupancy of a metapopulation. The framework combines a stochastic model for the generation of synthetic streamflow time series with the hierarchical structuring of river network dynamics, to enable the simulation of the full spatio-temporal dynamics of the active portion of the stream network under a wide range of climatic settings. The hierarchical nature of stream dynamics - which postulates that during wetting nodes are  activated sequentially from the most to the least persistent, and deactivated in reverse order during drying - represents a key feature of the approach, as it enables a clear separation between the spatial and temporal dimensions of the problem. The framework is complemented with a stochastic dynamic metapopulation model that simulates the occupancy of a metapopulation on the simulated stream. Our results show that stream intermittency negatively impacts the average occupancy and the probability of extinction of the focus metapopulation. Likewise, the spatial correlation of flow persistency along the network bears a sizable impact on the mean network connectivity and occupancy. This effect is particularly important in drier climates, where most of the network undergoes sporadic and flashy activations, and species dispersal is regularly inhibited by river fragmentation. This approach offers a robust but parsimonious mathematical framework for the synthetic simulation of stream network dynamics under a broad range of climatic and morphological conditions, providing useful insights on stream expansion and retraction in its ecological significance.