Plot data for determining minimum sampling scale using moving window method: insights from forest structural and functional indicators

The analysis indicated that an 85 m window (within a 10% deviation rate from the overall plot) was determined as the minimum necessary to accurately capture the spatial variability of all indicators. To further investigate the drivers of minimum sampling scale variation, a random forest model identified the coefficient of variation as a key factor influencing scale sensitivity for basal area, angular scale, neighborhood comparison, and biomass.

This finding underscores the importance of spatial heterogeneity in sampling design. Additionally, partial least squares path modeling (PLS-PM) was used to assess ecological interactions at the minimum scale. Unexpectedly, species diversity showed a negative effect on productivity, potentially due to canopy gaps and asymmetric competition, but positively influenced spatial structure (especially mingling), underscoring its role in enhancing structural complexity.

By integrating minimum sampling scale determination with ecological interaction analysis, this study provides a framework for context-specific sampling strategies that consider multiple ecological variables. The findings contribute to improving forest inventory practices and offer practical guidance for large-scale monitoring and conservation efforts in temperate mixed forests.