Search

The ability to colonize new habitat is essential for wild populations affected by disturbance or other forms of habitat change. For aquatic insects in small streams, overland flight is an important strategy for dispersal when barriers to in-stream migration exist and when populations are isolated in upland habitats. Two Ozark-endemic water beetles (Heterosternuta sulphuria and Heterosternuta phoebeae) have shown little overlap in distributions, with the former frequently occurring in small upland watersheds and the latter occurring in aquatic habitats farther downstream in larger watersheds. Because H. sulphuria has been associated with perennial aquatic habitats, we hypothesized that H. sulphuria individuals could exhibit low capacity for flight, thereby affecting population distributions over time. Laboratory flight observations showed that zero individuals of H. sulphuria flew (n = 67), whereas 17 of 76 individuals of H. phoebeae were observed to fly. Stream habitat drying experiments provided further evidence of the weak capacity for flight and overland migration of H. sulphuria, with low probabilities of survivorship in microhabitats exposed to drying. Weak flight capacity and apparent intolerance to habitat drying have important implications for the evolutionary history and conservation of H. sulphuria in small Ozark streams exposed to variable flow regimes and stream margins vulnerable to disturbances.

Stoneflies are a globally threatened aquatic insect order. In Arkansas, a diverse group of winter stonefly (Capniidae: Allocapnia) have not been surveyed since the 1980s, likely because species-level identification requires the rarely-collected adult form. Allocapnia mohri, a regional endemic, was previously commonly found in mountainous, intermittent streams from the Ouachita Mountains ecoregion north to the Ozark Highlands, but no species distributional models including land use or climate variables exist to our knowledge. We collected adults from 71 stream reaches from the historic Arkansas range from November to April 2020 and 2021. We modeled distributions using random forest (RF) models populated with landscape, climate, and both data to determine which were most predictive of species presence. Correlations between landscape or climate variables and presence were examined using multiple logistic regression. The landscape RF models performed better than the climate or landscape + climate RF models. A. mohri presence sites tended to have a greater elevation, a lower mean July temperature, and a greater percentage of very slow infiltration soils in the watershed, compared to absence sites. A. mohri was absent at the Ouachita Mountains sites and may be experiencing a range contraction or migration northward.