Comparison of non-linear models for determining the growth rhythm of stem height in Populus deltoides seedlings

摘要

Seven different non-linear models were constructed for stem heights of 200 Populus deltoides progeny in a full-sibs family in Nanjing, China. Based on the determination coefficient (R~2), the best R~2 estimate was obtained by the logistic model, followed by the Gompertz and Von Bertalanffy (>0.990). Residual Mean Square (RMS) and Mean Absolute Deviation (MAD) analyses revealed the logistic model produced the lowest RMS (29.95) and MAD (13.17). Meanwhile the parameter A of the logistic model was notably close to the ending observation value. Considering all statistical analyses, the logistic model presented the best fit in describing the growth pattern in P. deltoides seedlings, and then the Gompertz and Von Bertalanffy models. Additionally, we established an ideal statistical model to track the changes in growth rates of a Populus pedigree. This study provides essential information and biological interpretation for understanding the dynamic growth rates for Populus spp.