Evaluating methods for sampling stream salamanders across multiple observers and habitat types

While techniques for sampling pond-breeding amphibians are relatively well-established, comparable methods for stream amphibians are still being developed. Uncertainty about sampling techniques is particularly acute for approaches that involve multiple observers. I evaluated three techniques for sampling stream salamanders with multiple observers—time-constrained surveys, leaf-litter bags, and controlled-cobble added to the stream. I tested for differences among techniques, differences among habitats, and technique by habitat interactions. I also asked whether coefficients of variation across observers and sample dates differed among methods and habitats. For all species and life-stages, time-constrained surveys produced higher counts of stream salamanders and greater richness estimates than did leaf-litter bags or controlled-cobble surveys. However, interactions between techniques and habitat were never significant, indicating that all the methods detected similar patterns in counts among stream habitats. Coefficients of variation in single-species counts tended to be higher for time-constrained surveys, whereas CVs for species richness were significantly lower for time-constrained surveys. Overall, these results suggest that: (1) time-constrained surveys for stream salamanders work surprisingly well even with multiple observers, (2) leaf-litter bags may be quite useful for single-species studies, but are less effective for multi-species sampling, and (3) controlled-cobble added to streams is not particularly effective for sampling salamanders in rocky streams.