Reliability and precision of pellet-group counts for estimating landscape-level deer density

Abstract

This study provides hitherto unavailable methodology for reliably and precisely estimating deer density within forested landscapes, enabling quantitative rather than qualitative deer management. Reliability and precision of the deer pellet-group technique were evaluated in 1 small and 2 large forested landscapes. Density estimates, adjusted to reflect deer harvest and overwinter mortality, were compared with a drive count on the small landscape and with aerial counts using forward-looking infrared videography (FLIR) on the large landscapes. Estimates by 2 expert and 2 novice counters (range = 17.6 to 18.6 deer/km 2 ) on the small landscape were not different from each other and three of the four were not different from the drive count (17.4 deer/km 2 ). FLIR density estimates were approximately 30% lower than pellet-group estimates on the large landscapes (P < 0.04), an expected result. Precision on the small landscape was high; 95% confidence intervals for individual counters were <7.5% of mean estimates of density, and coefficients of variability were <10%. Precision on the larger landscapes was acceptable: 95% confidence intervals were 18.4 to 30.4% of mean estimates and coefficients of variability were <25%. The pellet-group technique produces reliable and precise estimates of deer density, is inexpensive, requires little training to implement, and is best suited to northern hardwood forests where snow and cold result in minimal deterioration of pellet groups. Unless corrected for hunter harvest and overwinter mortality, pellet-group counts represent average overwinter density and overestimate spring density.