Local and Landscape-Scale Predictors of Salamander Abundance in New Hampshire Headwater Streams

We investigated the response of Gyrinophilus porphyriticus ( Plethodontidae), the spring salamander, to timber harvest in headwater streams in New Hampshire ( U.S.A.). We conducted cover-controlled, whole-stream surveys for G. porphyriticus in 25 first-order streams in drainages encompassing a range of timber-harvest histories. Streams in two different landscape configurations were chosen, isolated, or paired, based on the presence or absence of a confluent first-order stream-a potential source of immigrants. We analyzed salamander abundance with respect to logging history, presence of the brook trout (  Salvelinus fontinalis), salamander population connectivity, and other measures of physical habitat and aquatic conditions. In stepwise multiple-regression analysis, the best multivariate model of G. porphyriticus abundance included fish presence/absence; effect category, a combined function of years since harvest and substrate embeddedness by fine sediment; and landscape configuration. These three factors explained 74% of the variation in G. porphyriticus abundance. G. porphyriticus was less abundant in streams where fish were present, likely because of predation by S. fontinalis or asymmetric competition between these species. Abundance also declined with increasing substrate embeddedness and/or decreasing years since harvest, indicating an effect of fine sediment inputs and harvest history on this species. Finally, G. porphyriticus was less abundant in isolated streams than in paired streams. This result suggests that landscape-scale population connectivity may buffer this species from the negative effects of local (i.e., stream-scale) habitat perturbation. We hope this work will encourage others to develop multiscale models of land-use effects on stream biota and thereby ultimately increase the range and effectiveness of conservation strategies for these species.