Selecting a sampling method to aid in vegetation management decisions in loblolly pine plantations

Abstract

Objective methods to evaluate hardwood competition in young loblolly pine (Pinustaeda L.) plantations are not widely used in the southeastern United States. Ability of common sampling rules to accurately estimate hardwood rootstock attributes at low sampling intensities and across varying rootstock spatial distributions is unknown. Fixed area plot, polyareal plot, and Batcheler's triple-point distance sampling estimators of hardwood rootstock density, total height, and crown area were evaluated using computer simulation. Rootstock age, density, and spatial pattern combinations defined 24 artificial populations representative of hardwood rootstocks found in young loblolly pine plantations. Relative bias, relative precision, and within-and between-sample variance of the estimators were used as evaluation criteria. Eighteen sampling estimators with sampling intensities of approximately 1% were evaluated. Mean relative bias ranged from 1.2 to 7.0%, −1.5 to 1.0%, and −3.8 to 3.4% for fixed area, polyareal, and distance estimators, respectively. Spatial pattern affected evaluation criteria. Relative precision was highest in uniform populations and lowest in clumped populations. Variance was lowest in uniform populations and highest in clumped populations. Batcheler's distance estimator performed poorly. The 0.01-acre (1 acre = 0.405 ha) circular plot, vertical line, and horizontal point estimators were deemed "best" for density, total height, and crown area estimation, respectively. Vertical line sampling estimated all three attributes well and is recommended for field testing in young conifer plantations.