Performance of sampling methods to estimate log characteristics for wildlife

Abstract

Accurate estimation of the characteristics of log resources, or coarse woody debris (CWD), is critical to effective management of wildlife and other forest resources. Despite the importance of logs as wildlife habitat, methods for sampling logs have traditionally focused on silvicultural and fire applications. These applications have emphasized estimates of log volume and weight, but wildlife applications require estimates of percent cover, density, size, and length. Given the diverse number of log variables of interest to wildlife managers, understanding the strengths and weaknesses of common methods of log sampling is essential. Consequently, we evaluated the empirical performance of line-intersect (LIM) and strip-plot (SPM) methods for sampling logs [(≥15 cm large-end diameter (LED); ≥1 m long)] in forests of Oregon and Montana. In unharvested stands, the precision and efficiency of LIM and SPM were similar for all variables except log density, for which SPM performed better. In harvested stands, precision and efficiency of LIM and SPM varied widely for several estimates. Strip-plot method was more precise and efficient than LIM as an estimator of density and volume for logs ≥15 cm LED. Moreover, SPM required less time to sample large logs (≥25 cm LED). Both methods yielded unbiased estimates except for total log length, which LIM underestimated in harvested stands. Importantly, SPM performed better than LIM as a predictor of wildlife use. Our results provide the first empirical validation of these methods, and offer useful context for design of log inventories for research and management.