Predator-avoidance behavior in anuran tadpoles: a new bioassay for characterization of water-soluble cues

Most surveys of large wood in streams are conducted by counting and measuring every piece of large wood within a reach, a technique that is effective but time-consuming. In this study we evaluated an alternative method that takes less time and can be employed in studies in which an estimate of total large wood volume along a stream reach is the primary metric of interest. In first- through third-order streams we estimated in-stream large wood volume and large wood frequency, comparing large wood census estimates to those from a modified a line-intercept technique that has been commonly used in terrestrial forest surveys. Estimates of large wood volume from line transects located in the geographic center of the stream (parallel to stream axis and equidistant from bankfull margins) were highly correlated with those from the wood census (P < 0.001, r ² = 0.88, Pearson’s r = 0.935), but produced slightly greater estimates of large wood volume (regression slope = 1.28, SE = 0.16). Line-intercept estimates of large wood frequency (number per 100 m of stream) were significantly correlated to the wood census counts, but the line-intercept method underestimated frequency by about 50% (P = 0.016). Differences in the estimated large wood volume between line-intercept and wood census surveys were associated with variability in the diameter of the large wood, but unrelated to stream bankfull width, for the range of stream sizes evaluated in this study (≈ 2 to 11 m). Our results suggest that in small constrained streams, line-intercept surveys are an effective method for estimating in-stream large wood volume and that these estimates better approximate results from whole-stream census techniques where the diameter of in-stream wood is relatively consistent. Handling editor: K. Martens