Balancing Accuracy and Efficiency Optimal Plot Design for Regeneration Sampling in Amazonian Secondary Forests

Abstract

Background: Optimal sampling designs are crucial for accurate ecological and forestry assessments, particularly for regeneration studies in Amazonian secondary forests, which play an important role in biodiversity conservation and carbon sequestration. This study evaluated different sampling plot configurations for estimating regeneration diversity and structural attributes in a 30-year-old secondary forest in Belém, Brazil. Within a one-hectare permanent plot (100 × 100 m), all trees with a diameter at breast height (DBH) ≤ 10 cm were measured, identified, and geolocated, totaling 3,003 individuals. Trees were classified into two diameter classes: DBH < 5 cm and 5 cm ≤ DBH ≤ 10 cm. Resampling simulations using the bootstrap method subdivided the one-hectare plot into four sampling plot sizes (4 m², 25 m², 50 m², and 100 m²) with rectangular and square shapes. Simulations tested sample sizes ranging from four to (N − 1) units, with 1,000 iterations per configuration.

Results: Accuracy and precision for diversity metrics (species richness and Shannon–Weaver index) and structural attributes (tree density, stem density, and basal area) were evaluated using Mean Absolute Error (MAE) and Relative Sampling Error (RSE). Results indicated that 4 m² sampling plots were the most suitable for estimating diversity metrics across both diameter classes, regardless of plot shape. For structural variables, square 4 m² plots performed best for trees with DBH < 5 cm, whereas rectangular 50 m² plots were optimal for trees with 5 cm ≤ DBH ≤ 10 cm. The influence of plot shape varied depending on the variable analyzed and the sampling plot size.

Conclusion: Overall, sampling plots of 4 m² and 50 m² are recommended for efficient regeneration sampling in Amazonian secondary forests, as they provide better accuracy and precision for diversity and structural estimates across different diameter classes.