Objective The methods used for earlywood and latewood (EW/LW) demarcation significantly affect the accuracy of extracted tree-ring parameters. To determine the optimal demarcation approach, it is essential to systematically evaluate the applicability of various boundary demarcation criteria.

Method Using the visual inspection method (VIS) as a benchmark, this study compared the demarcation results of three dynamic threshold methods (MMmean, MAX80%, and MM2/3+) and the maximum derivative method (D6). Spearman correlation analysis was employed to test the consistency of the parameter series, while Pearson correlation analysis was utilized to examine the relationships between the tree-ring chronologies derived from different methods and climate factors.

Result The density thresholds obtained by the MAX80% and MM2/3+ methods were approximately 640 kg/m³, whereas those from the MMmean and D6 methods were approximately 550 kg/m³. The overall trends of the correlations between the derived chronologies and climate factors were consistent across the different methods, with the D6 method demonstrating the highest agreement with the visual inspection benchmark. Compared to the other methods, the D6-derived earlywood width chronology exhibited more significant positive correlations (P < 0.05) with three temperature indices from the previous November. Furthermore, the earlywood and latewood density chronologies based on the D6 method showed greater sensitivity to the maximum temperature of the previous September.

Conclusion By capturing the inflection point of the tree-ring density curve, the D6 method more accurately characterizes the physiological transition between earlywood and latewood. Earlywood parameters are more susceptible to the choice of demarcation method than latewood parameters. We recommend 550 kg/m³ as the reference density threshold for earlywood and latewood demarcation in Abies nephrolepis.