Comparison of methods for estimating forest overstory cover. I. Observer effects

Abstract

Four techniques used to estimate forest overstory cover were tested for differences in estimates derived from different observers. No differences in estimates of overstory cover were found between observers for ocular and gimbal sight techniques. Overstory estimates with the moosehorn were inconsistent between observers over all plots. On a plot by plot comparison, observer effects were less pronounced with the moosehorn (differences on 9 of 23 plots, p < 0.05) than with the convex spherical densiometer (differences on 16 of 23 plots, p < 0.05). No technique showed a consistent observer bias. Only the spherical densiometer showed marked interaction between observer and overstory cover (p < 0.05). Results indicate that the more complex methods requiring evaluation of several grids or dots at the same sample point were less consistent between observers. Application of such techniques will benefit from training to standardize observer readings.