Modeling branch diameter in Larix olgensis using living branches, knot profiles and stem scars: climate–soil associations

Branch diameter (BD) strongly influences crown architecture, knot formation and wood quality, yet BD development is commonly modeled using living branches alone, leaving the diameter attained by dead branches at growth cessation insufficiently represented. Here, we developed a BD model as a function of branch age for Larix olgensis by combining living first-order branch measurements with longitudinal knot profiles and external stem-scar diameters. Data from 60 felled trees in 19 plots were used to estimate age-related BD curves and to evaluate the roles of branch position, tree form, climate and soil factors. Candidate growth functions were compared, extended with environmental covariates and fitted within a nonlinear mixed-effects framework to account for plot- and tree-level variation. BD followed a nonlinear age-related trajectory, with rapid early enlargement followed by gradual stabilization. The Mitscherlich function was selected as the base model, and its performance improved after incorporating branch insertion height, height-diameter ratio, mean warmest-month temperature (MWMT), soil organic carbon and a product of mean annual precipitation and cation exchange capacity (MAP × CEC). The final mixed-effects model increased adjusted R 2 from 0.801 to 0.936 and reduced RMSE from 3.145 to 2.087 mm. Simulations indicated that BD was larger for branches located higher on the stem and under higher MWMT conditions, but smaller in trees with higher height-diameter ratios. The MAP × CEC term described variation in simulated BD across combinations of precipitation and soil conditions. The model provides age-related estimates of branch size that can inform wood-quality assessment and pruning decisions when considered alongside stand age, live-crown retention and management costs.