The height-diameter relationship is one of the foundational equations of forest science. Measuring the diameter at breast height, or DBH, is quick and cheap; climbing or using instruments to measure total tree height is laborious and expensive. Because height is essential for estimating timber volume, biomass and carbon stocks, foresters rely on models that predict height from diameter. The research team, led by Yan Lu and Huiping Zeng of Southwest Forestry University together with colleagues at Ningxia University, measured 910 Chinese fir trees across nine 20-meter-by-20-meter plots, spanning young stands of eight years, middle-aged stands of seventeen years and mature stands of twenty-five years.

Nine candidate growth functions were fitted to the data, ranging from simple linear and power equations to the Korf, Michaelis-Menten, exponential and mixed models. The 910 trees were split by stratified random sampling into a calibration set of 728 trees and an internal validation set of 182 trees, with performance judged on the coefficient of determination, root mean square error, mean absolute error and the Akaike information criterion. The Quadratic model emerged as the overall winner, achieving the highest R-squared value of 0.6429 in validation, with root mean square errors between 2.39 and 2.43 meters and mean absolute errors between roughly 1.90 and 1.92 meters. The Mixed and Exponential models followed closely behind.

The authors are refreshingly honest about what those numbers mean. An R-squared near 0.5 during calibration indicates that diameter alone explains only about half of the variation in tree height, a level of noise that reflects microsite variability, competition and management history in these dense plantations. For individual trees, prediction uncertainty of around 2.4 meters is considerable. Yet for stand-level applications, estimating mean height, approximating timber volume and conducting large-scale inventories, the Quadratic model provides a useful, economical alternative to direct height measurement. Notably, the best-performing models diverge from earlier Chinese fir studies in Fujian and southeastern China that favored Richards-type or nonlinear mixed-effects models, a difference the researchers attribute to the distinctive soils, climate and growth conditions of the southern Jiangxi mountains.

Beneath the uniform fir canopy, the team documented a surprisingly rich hidden world. Across the nine plots, the understory contained 74 plant species belonging to 64 genera and 43 families, split evenly between 37 shrub species and 37 herbaceous species. The spreading fern Dicranopteris pedata proved to be the most ubiquitous companion, appearing across all stand ages, while dominant species shifted with forest age, from Myrsine africana and Smilax glabra in young stands to Rubus corchorifolius and Gardenia jasminoides in middle-aged ones.

Diversity indices told a nuanced story. Species richness declined gradually with stand age, with young forests holding the most species and mature forests the fewest, a pattern consistent with previous findings in pine and rubber plantations. Species diversity, measured by the Simpson and Shannon-Wiener indices, peaked numerically in middle-aged stands, while evenness was highest in mature stands. The shrub layer consistently showed higher diversity and evenness values than the herb layer. Floristic analysis revealed a strong tropical signature: the 43 families fell into seven distribution types and the 64 genera into thirteen, with tropical elements accounting for nearly 58 percent of genera, and clustering analysis showed that plants from similar or adjacent biogeographic regions tended to group together, hinting at deep phylogeographic roots in this subtropical flora.

The most provocative part of the study is its use of Mantel correlation tests to link tree growth with understory diversity. The tests showed that tree height and DBH were mostly insignificantly and negatively correlated with understory diversity indices, though both were positively associated with the Simpson evenness index. The researchers interpret this cautiously: the growth of dominant fir individuals may primarily affect how evenly resources are allocated among understory plants rather than the total number of species present, with a few shade-tolerant species thriving under larger trees while light-loving herbs decline. Among the diversity indices themselves, significant differentiated association patterns emerged, including a significant positive link between the Shannon-Wiener and Pielou evenness indices and a significant negative link between species richness and evenness.

Crucially, the team refuses to overclaim. With only nine plots, three per age class, and topographic factors partially confounded with stand age, the statistical power was limited and causal inference impossible. The authors explicitly frame all results, from the model rankings to the correlation analyses, as exploratory and descriptive, intended for hypothesis generation rather than definitive conclusions. This kind of methodological transparency is increasingly valued in ecology, where small-sample studies often generate headlines that later fail to replicate.

The practical implications remain substantial. As China pushes forward with carbon-neutral forestry initiatives, accurate growth models for Chinese fir are essential for monitoring timber production, estimating yields and planning ecological services. The study suggests that the Quadratic model, with its stable and interpretable parameters, is the most suitable local tool, while cautioning against extrapolation beyond the observed diameter range of 3.3 to 29.1 centimeters. Future work, the authors argue, should expand sampling across broader environmental gradients and incorporate site-specific covariates such as soil properties, light availability and management history to sharpen both the models and the ecological inferences they support.

For now, the study stands as a model of careful, self-aware field science: a modest dataset, honestly analyzed, that maps the growth equations of one of China’s most planted trees while revealing a floristically rich and historically layered understory beneath its shadows. It reminds foresters and ecologists alike that even in intensively managed monocultures, biodiversity persists, and that the relationship between the giants above and the small plants below is subtler than any single correlation coefficient can capture.

Subject of Research: Height-diameter modeling and understory vegetation diversity of Cunninghamia lanceolata plantations in southern Jiangxi, China

Article Title: Height-diameter models and understory vegetation diversity analysis of Cunninghamia lanceolata plantations in the mountainous areas of Southern Jiangxi, China

Article References: Lu, Y., Zeng, H., Li, X., Long, Q., Shen, Q., & Dong, Q. (2026). Height-diameter models and understory vegetation diversity analysis of Cunninghamia lanceolata plantations in the mountainous areas of Southern Jiangxi, China. Discover Forests, 2(1), Article 66. https://doi.org/10.1007/s44415-026-00127-3

Image Credits: AI Generated

DOI: 10.1007/s44415-026-00127-3

Keywords: Cunninghamia lanceolata, Chinese fir, height-diameter model, plantation forestry, understory vegetation, species diversity, Mantel test, southern Jiangxi, forest ecology, Quadratic model, biodiversity, subtropical forest