Fragmentation of lateral connectivity and fish population dynamics in large rivers

Lateral Hydrological Connectivity (LHC) has been extensively fragmented in river-floodplain ecosystems of the world. Uncertainties about how LHC affects fishes are great, impeding the design of effective rehabilitation strategies. Existing conceptual frameworks do not provide sufficient mechanistic detail to support the novel decision problems river managers face. We offer a framework of how LHC affects fishes in river-floodplain ecosystems that is, process-based, integrates all life-stages and is spatial; these features, we argue, are required to assess risks and opportunities associated with different LHC rehabilitation strategies. Within river-floodplain segments, LHC affects population processes through five 'effect-classes': effects of floodplain habitat (1) and channel habitat (2); effects of material subsidies from the floodplain to the channel (3) and vice versa (4); and effects of connectivity on lateral dispersal (5). The relative influence of these effect-classes on processes varies among species and life-stages. At the scale of the drainage basin, inter-segment variation in geomorphology generates a longitudinal source-sink structure to habitat quality and quantity, pointing to a need to better understand fish metapopulation dynamics in river-floodplain ecosystems. Given the significant investment in trying to restore river-floodplain ecosystems, we highlight potentially costly and ineffective LHC management decisions. These include certain heavily engineered LHC rehabilitation strategies that do not promote critical population processes at a local scale, and at basin scales implementing strategies that do not facilitate the metapopulation processes that promote species' persistence.