Objective This paper verifies the feasibility of site quality evaluation method based on potential annual increment of forest and optimize this method. The potential annual increment method was used to estimate the potential productivity of basal area and stand volume, providing a scientific basis for site quality evaluation and silvicultural management of Larix olgensis plantations.

Method The dominant height growth model of L. olgensis plantations was developed using generalized algebraic difference method based on the permanent sample plot data from the national forest inventory and forest management unit-level inventory in Jilin Province of northeastern China. Site grades were ranked according to the site index. The growth models of basal area and volume, incorporating site grades as dummy variables, were constructed, and the optimal stand density index was searched using golden section method. The corresponding potential productivity was calculated and verified by self-thinning line. Realized productivity was calculated from the actual stand age and stand density index. The productivity improvement gap was comprehensively evaluated by combining the difference between realized and potential productivity.

Result (1) The models of dominant height, basal area and volume growth of L. olgensis plantations showed good performance. The adjusted coefficients of determination values were 0.877, 0.985 and 0.966. The root mean square error values were 1.037 m, 0.886 m²/(ha·year) and 9.756 m³/(ha·year), while the relative root mean square error values were 7.353%, 6.588% and 11.890%. The average site index values for the five site grades were 19.8, 17.8, 15.7, 13.5 and 11.6 m, respectively. (2) For the same stand age, the optimal stand density index differed for basal area and volume potential productivity. At the base age of 30 years, the potential productivity for stand basal area ranged from 0.861 to 0.941 m²/(ha·year), while the potential productivity for volume ranged from 5.319 to 8.775 m³/(ha·year). The productivity gap between realized and potential productivity was larger in young and middle-aged stands compared with near-mature, mature and overmature stands, indicating a greater potential for productivity improvement. (3) During stand growth, the optimal stand density corresponding to potential productivity was always less than or equal to the maximum density determined by self-thinning process.

Conclusion The potential productivity calculated after constraining the initial stand density is more realistic, further validating the feasibility of site quality evaluation method based on potential mean annual increment. During the young and middle-aged stages, the realized productivity of L. olgensis differs significantly from its potential productivity, with the difference increasing as site quality declines. The results provide the optimal stand density corresponding to potential productivity, offering a basis for tending operations and management of L. olgensis plantations.