Nonlinear mixed effects crown width model for planted L. kaempferi at high altitudes in southern China

• Published in: Trees (Berlin, West) Vol. 39; no. 3; p. 54
• Main Authors: Liu, Jiateng, Tong, Yiwen, Gao, Huilin, Sun, Xiaomei, Chen, Dongsheng
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.06.2025; Springer Nature B.V
• Subjects: Agriculture; Altitude; Biomedical and Life Sciences; China; climate; Climate change; equations; Forestry; heat sums; Height; heteroskedasticity; High altitude; High-altitude environments; Larix kaempferi; Life Sciences; Original Article; Plant Anatomy/Development; Plant Pathology; Plant Physiology; Plant Sciences; Planting; Planting density; prediction; spring; Thinning; timber production; tree growth; Trees; variance; China; Crown width; Thinning intensity; Climate variable; Initial planting density; plantation
• ISSN: 0931-1890; 1432-2285
• DOI: 10.1007/s00468-025-02632-y

Key message Incorporating height to crown base, competition index, climate variable, initial planting density, and thinning intensity into base crown width model significantly improved the performance of the crown width prediction of L. kaempferi at high altitudes in southern China. Crown width (CW) is significantly related to tree growth and the ability to predict crown width with high accuracy is essential. To develop a CW model for planted Japanese Larch ( L. kaempferi ) at high altitudes in southern China, a total of 3950 trees from 78 permanent sample plots with altitudes ranging from 1372 to 1932 m were measured. The best base model including only diameter at the breast height (DBH) was selected by using the evaluation statistics. Tree-level, stand-level, and climate variables were introduced into the basic model using reparameterization approach. Initial planting density and thinning intensity were introduced through dummy variable method. We developed a nonlinear mixed effects crown width (NLME-CW) model. The prediction performance of the NLME model was evaluated using the leave-one-out validation approach. The results show that power equation performed better than other base CW models. Using reparameterization approach, tree level variable of height to the crown height (HCB), competition variable of basal area of the trees larger than the subject tree (BAL) and the climate variable of spring degree-days below 0 ℃ (DD_0_sp) were finally included into the CW model. The developed NLME-CW model including initial planting density and thinning intensity showed good stability and applicability. Power variance equation was used to reduce the heteroscedasticity in the residuals. The six individual trees were randomly selected from each sample plot to estimate random parameters. CW increased with the increasing of DBH and decreasing of HCB, BAL and DD_0_sp. CW decreased with the increasing of planting density and increased with the increasing of thinning intensity. Considering timber production and management costs, we recommend moderate-intensity thinning as the optimal management strategy. The model is widely applicable to L. kaempferi plantations and holds significant practicality for predicting the CW of L. kaempferi in high-altitude areas of southern China.