Comparison of various growth functions for predicting long-term stand development associated with different initial spacing in 64-year-old Japanese cedar (Cryptomeria japonica (L.f.) D. Don) plantations

Abstract

Establishing a plantation with varied distances between trees is one way to control stand density. Understanding how stand growth patterns are influenced by initial spacing can help maximize growth and yield. The purpose of this study was to examine the long-term effects of different patterns of initial tree spacing on stand development. A spacing trial with 5 initial distances, namely, treatments I: 1×1 m, II: 2×2 m, III: 3×3 m, IV: 4×4 m and V: 5×5 m, was installed in 1950 on Japanese cedar ( Cryptomeria japonica ) plantations in central Taiwan, managed by the Experimental Forest of National Taiwan University. Each treatment had 3 repeated plots, and all the plots were surveyed from 1955 to 2014, with 8 records for each time series. We simultaneously employed 6 growth functions to predict stand basal area growth for each plot and assessed their predictability using the root mean square error (RMSE) as a criterion. Consequently, a total of 15 plots covering all the treatments, each with 6 RMSEs resulting from various models, were obtained. The repeated measures analysis of variance approach was adopted to compare predictability among models. The Richards growth function stood out from all the other models. As a result, this model was used to analyze stand development following different initial spacing. We found that the predicted curves could effectively exhibit the growth patterns resulting from different initial spacing. Moreover, the parameters help explain some characteristics of stand development, such as the growth potential and the maximum growth rate achieved ( t max ). As a result, a clear trend emerged, showing that the growth potential increased, whereas the tmax decreased with decreasing initial spacing. These results provided valuable information for managing the stand density of this conifer.