Fine-scale prioritization of fish persistence in the Daly River, NT: Condition and conservation re-revisited

Systematic conservation planning in aquatic systems gained momentum over the last 5 years, when scientists, managers and NGOs started embracing new methods to adapt to the peculiarities of rivers namely longitudinal and lateral connectivity. In this study, we are using modeled distributions of 41 native fish species extrapolated to 5723 stream segments and their corresponding subcatchments in the Daly River, NT. Distribution models were built from a range of GIS-derived predictor variables including hydrology, catchment topography, geology, climate and present-day tree cover using a multi-response artificial neural network model. We used the conservation planning software MARXAN to derive best solutions and irreplaceability maps for representation targets ranging from 60-120 habitat kilometers per species. Longitudinal connectivity was accounted for by putting exponentially decrasing penalties on solutions that do not include upstream segments. We then compared conservation plans that ignored condition (which is most often the case), with conservation plans that included river disturbance. The river disturbance index (RDI) was used as an additional penalty on subcatchments with a high level of anthropogenic activity. Conservation plans that avoided river disturbance changed significantly demonstrating that including condition measures is important avoiding costly conservation mistakes. However, 100% selection frequency of some lowland regions shows that even highly disturbed rivers cannot be excluded from conservation and restoration planning activities if they contain endemic species.