Designing fecal pellet surveys for snowshoe hares

Index methods can be valuable for monitoring forest-dwelling vertebrates over broad spatial or temporal scales. Fecal pellet counts are often used as an index of density or habitat use of snowshoe hares, Lepus americanus, but previous surveys have used different plot types and sample sizes, leading to problems comparing results from different studies and questions about the inferential power of each study. In this paper, we use field data and simulations to examine how the precision, bias, and efficiency of four commonly used plot types vary with plot type, pellet density, and sample size. Although no one plot type was consistently superior, we recommend thin rectangles (5.08 cm × 305 cm (2 in. × 10 ft), 0.155 m²) or 1 m² circles over 0.155 m² circles or 10 cm × 10 m (1 m²) rectangles. We recommend that researchers explicitly address the power of their survey design to detect different pellet densities, because much larger sample sizes are needed at low pellet densities than at high pellet densities to obtain similar precision. Small sample sizes are also much more likely to be biased, which could lead to incorrect inferences about management of snowshoe hare populations. Both uncleared and cleared plots performed well and will have value in different research contexts.