Comparaison de différentes approches, modèles et tailles d'échantillons pour l'établissement de relations hauteur-diamètre locales

Abstract

This study compares the predictability of 3 approaches, 4 models, and 10 different sample sizes to determine local relationships between total height and diameter at breast height in balsam fir stands less than 50 years old. The results show that a system of height–diameter curves based on the mean diameter and mean height of the sampled trees (approach 3) gives the most precise estimations in comparison with curves resulting from pooling all sampled trees in a single height–diameter model (approach 1) and with those resulting from the application of a height–diameter model for each of the combinations of sample plot–measurement periods (approach 2). Depending on the sample size, the residual variance of the total height or the total volume is reduced 2 to 2.5 times for the individual stems and 2 to 15 times for the mean stem, when using approach 3 instead of approach 1. Approach 3 is more precise than approach 2 for sample sizes that vary between three and five sampled trees per plot. However, this gain in precision becomes negligible when the sample size approaches 10 sampled trees per plot.