Large, prolonged flooding and pool persistence promote floodplain fish diversity in a threatened river.

Abstract

While it is widely recognized that reduced river-floodplain connectivity has contributed to the decline of biodiversity in floodplain rivers, surprisingly few studies have quantified the relationship between connectivity, pool persistence, and fish assemblage structure to the level required to generate measurable targets for management. The task is further complicated by the inherent complexity of accurately describing fish assemblages. We maximized our capacity to describe unbiased hydrology-fish relationships by sampling fish assemblages in floodplain pools with a variety of connection histories (60 sampling events), and by using a hierarchical multispecies occupancy model that accounts for changes in sampling design and species detection. Our study was conducted in a tropical wet-dry river threatened by water resource development and elevated temperatures associated with climate change, the Fitzroy River (Western Australia). Our results revealed that wet season (river-floodplain connectivity) and dry season (pool persistence) components of the hydrological cycle influenced fish occurrence in floodplain pools. Pools that were connected to the river by short distances were substantially more species rich than distal pools. This effect was strong at distances <2000 m but negligible at distances greater than 3000 m. Species richness in floodplain pools increased when wet season connection to the river lasted more than 25 days, and when river stage height exceeded 6 m. Prolonged connection to the river (up to 90 days) during overbank flooding (river stage height >11 m) maximized fish species richness in floodplain pools. Dry season components of the hydrological cycle also influenced fish assemblage structure, with pools that persisted during the preceding dry season twice as species rich as those that dried. Our model revealed that sampling gear influenced species detectability, indicating that accounting for variable detection is critical when assessing fish assemblage structure. Given that large flood events are less likely to be impacted by water take, we recommend that managers seeking to maintain floodplain fish diversity ensure that water resource development does not negatively impact pool persistence during the dry season.