Living on the edge: Predicting invertebrate richness and rarity in disturbance‐prone aquatic–terrestrial ecosystems

Abstract

Abstract Temporal fluctuations in cause the spatial extent of wet and dry habitats to vary in aquatic–terrestrial riverine ecosystems, complicating their biomonitoring. As such, biomonitoring efforts may fail to characterize the species that inhabit such habitats, hampering assessments of their biodiversity and implementation of evidence‐informed management strategies. Relationships between the dynamic characteristics of aquatic–terrestrial habitats and their communities are well known. Thus, habitat characteristics may enable estimation of faunal assemblage characteristics such as taxonomic richness, regardless of in‐channel water levels. We investigated whether indicators summarizing habitat survey data can predict two metrics representing terrestrial invertebrate assemblages (e.g. taxa richness) in two aquatic–terrestrial habitats: exposed riverine sediments and dry temporary streams. We also compared the performance of unimetric and multimetric habitat indicators in making predictions. In exposed riverine sediments, >88% of predictions were correlated with observed taxa richness and an index of conservation status. Values predicted by exposed riverine sediment samples were correlated with those observed in temporary stream channels with comparable riparian (i.e. largely agricultural) land use, but not those observed in channels with contrasting (i.e. more urban) land use. Unimetric habitat indicators performed similarly to more complex multimetric indicators, with each explaining ≤6% of the variability in taxa richness and the index of conservation status. The different spatial scales at which invertebrates respond to habitat conditions and at which indicators record habitat conditions, and a more comprehensive training dataset that incorporates a full range of habitat conditions (i.e. land use), may improve future predictions. We demonstrate that invertebrate assemblage characteristics can be predicted regardless of in‐channel water levels. Agreement between exposed riverine sediment predictions and temporary stream observations suggests that these predictions are transferable among a range of aquatic–terrestrial habitat types, and could thus be widely applied to aid conservation of riverine biodiversity in dynamic aquatic–terrestrial ecosystems.

Keywords

• disturbance
• exposed riverine sediment
• habitat survey
• inundation
• prediction
• rarity