Objective This paper investigates different dominant tree selection strategies and four site quality evaluation methods using natural secondary forests of Castanopsis spp. as the research objects. By analyzing the relationship between site quality evaluation indicators and stand factors, the aim was to determine the optimal dominant tree selection strategy and site quality evaluation method.

Method Based on data from 109 sample plots in southern Jiangxi Province of eastern China, four site quality evaluation methods (site form method, traditional site index method, DBH site index method, and DBH site index method of the thickest tree) were applied. These were combined with four dominant tree selection strategies (no species distinction, only Castanopsis spp. species, Castanopsis spp. species plus the species with the highest basal area proportion, and Castanopsis spp. species plus the top two species by basal area proportion) and two weighting methods (unweighted arithmetic mean and basal area-weighted mean). This resulted in the calculation of six types of dominant height, six types of average DBH of dominant trees, and two types of DBH of the thickest trees, along with their corresponding ages. The Richards equation was used as the base model, and the algebraic difference approach was employed to derive the difference forms corresponding to four site quality evaluation methods. By substituting the reference DBH and reference age, the site quality evaluation indicators for each method were calculated. Rank correlation analysis was used to assess the relationship between site quality evaluation indicators and stand density and stand volume, thereby identifying the optimal dominant tree selection strategy and site quality evaluation method for natural secondary forests of Castanopsis spp. in southern Jiangxi Province.

Result Significant differences were observed among estimates of dominant height, dominant average DBH, and maximum DBH using different methods, all significantly correlated with stand volume but inconsistently correlated with stand density. Among the four site quality evaluation methods, the site index based on dominant height-age relationship demonstrated no significant correlation with stand density (p > 0.05), while exhibiting significant correlations with stand volume (r = 0.145–0.190, p = 0.003–0.026). This pattern aligns with the fundamental requirement for site evaluation indicators: independence from stand density while maintaining correlation with stand volume.

Conclusion In southern Jiangxi Province, it is recommended to adopt a site quality evaluation approach that involves species-specific selection of dominant trees and application of the arithmetic mean to calculate dominant height-age-based site index. This method can more scientifically inform forest management planning and enhance forest productivity.