Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4374563 | Ecological Indicators | 2009 | 10 Pages |
Regulatory agencies need rapid indicators of hydrologic permanence for jurisdictional determinations of headwater streams. The study objective was to assess the utility of larval salamanders and habitat variables for determining stream permanence across a large geographic area. We sampled four core forests (61 sites in IN, KY, and OH) in spring (April–May) and summer (August–September) over a 2-year period. Sites in each forest were selected to cover a gradient of permanence, from perennial to ephemeral. Salamanders were collected by both benthic core sampling and timed visual search on each site visit. Classification and regression tree (CART) models were used to identify indicators of seasonal permanence at core sites that were then tested using data collected from 6 satellite forests (52 sites) located nationwide. Southern two-lined salamanders, Eurycea cirrigera, were numerically dominant and were the only species included in CART models. Salamander diversity declined with distance from the Appalachians and strong longitudinal changes in assemblage composition were evident within streams. Abundance of E. cirrigera was positively correlated with watershed area, whereas dusky salamanders, Desmognathus spp., and spring salamanders, Gyrinophilus porphyriticus, comprised a greater proportion of salamander communities at intermittent sites within their range. Spring and summer CART models incorporated E. cirrigera abundance and measures of channel geomorphology to accurately classify approximately 80% of core sites as either ephemeral, intermittent or perennial. When applied to validation data from national satellite forests, correct classification rates were >85% for intermittent and ephemeral sites, but were only ∼20% for perennial sites. These findings suggest that larval plethodontid salamanders and habitat variables can be valuable predictors of headwater stream hydroperiod, but indicators are largely limited to the regional scale.