Exploring biomass modeling in the Amazon Forest: assessing the effect of sample size of model calibration datasets

Tree-level models are commonly used to estimate the mean and total forest aboveground biomass (AGB). However, tree biomass predictions can vary significantly between models due to limited understanding of how calibration datasets influence their performance. The aim was to recommend a sample size for a dataset to calibrate tree biomass models that estimate mean AGB per unit area with the best possible precision and accuracy. The methodology consisted of (i) simulating an Amazonian forest using the Monte Carlo Method (MCM) and m out of n Bootstrap, (ii) estimating the average forest biomass using models calibrated with datasets of varying sample sizes, (iii) analyzing the precision and accuracy of the estimate of the mean AGB in Mg ha-1 per unit area using the models calibrated with different sample sizes, and (iv) relating characteristics of the calibration datasets with the errors produced by the models. Our findings indicate that increasing sample size improves the precision of biomass estimates, with notable gains at 200 trees. The study highlights that larger samples better represent tree diversity and support reliable biomass modeling, which is essential for effective forest management. We conclude that a minimum calibration sample size of 200 trees is sufficient to optimize biomass predictions while balancing practical sampling constraints. These results can aid in developing forest management strategies and improving carbon credit calculations in tropical ecosystems.