Two methods for predicting locations of rare darter habitat

Rare species are a substantial component of biodiversity, but difficult to locate and assess; their optimal habitats are poorly known. New York’s fish fauna includes >170 species and almost 1/3 are considered rare. Conserving these uncommon species is a major objective of many ecological management plans, including those of New York State. Darters represent some of New York’s rarest fish fauna. Observed fish data from 1978-2002 showed Varigated Darter (Etheostoma variatum) present in 27 streams, Longhead Darter (Percina macrocephala) in 18 streams, Bluebreast Darter (Etheostoma camurum) in 2 streams, and Spotted Darter (Etheostoma maculatum) in one stream. We applied two empirical modelling methods (species-habitat and assemblage-habitat neural networks) to predict occurrence of these four rare darters, based on stream habitat conditions in New York’s Allegheny drainage. Models performed well (R² ≥ 0.9) and where both species-specific and assemblage model results agreed suitable darter habitat was predicted to occur. Predictions for all Allegheny streams were mapped. These predicted distributions indicate potential habitats most likely to support the selected species and may assist with prioritization of habitats to be examined and possibly managed for biodiversity conservation. Additional independent data are needed to validate these models.