Abstract Tree biomass equations are important yet difficult, time-intensive, and expensive to develop. However, the calibration of previously developed, species-specific models could be a viable alternative, particularly for highly diverse and protected forests like the Atlantic Forest of Brazil. Consequently, the primary research goal of this study was to conduct a comprehensive evaluation of the potential to calibrate an existing individual tree aboveground biomass model for a new species and/or site by using linear mixed-effects. Specific research objectives were to determine the optimal approach for effective calibration by allowing sample selection method, sample size, and range of tree sizes sampled to vary. In particular, a certain set of species was used as a primary dataset to fit both generalized and species-specific biomass models, that were then calibrated for a secondary dataset at a different site and location. Both similar and divergent species at the secondary site were used to calibrate and evaluate the previous models. Our results suggested that species-level calibration was efficient for the majority of the species or individuals examined that can greatly improve the performance at much lower sample sizes required to develop a new equation, especially for the larger trees in the stand. In general, one to three randomly selected trees were sufficient to effectively calibrate a biomass model for a new species. We expect the combination of model calibration for abundant species associated with the use of the previous developed generalized model for less abundant species can drastically reduce the need for destructive sampling and improve predictions, which is important for highly threatened forests like the Atlantic Forest in Brazil. Overall, the results highlight the potential of model calibration to significantly improve both biomass and carbon estimates in species-rich forests like those in the tropics.

中文翻译：

---

校准单个树木生物量模型以对比巴西原生大西洋森林中不均匀年龄地点的热带物种：替代方法、样本大小和样本选择方法的直接比较

摘要 树木生物量方程很重要，但开发起来很困难、耗时且昂贵。然而，校准先前开发的特定物种模型可能是一个可行的替代方案，特别是对于高度多样化和受保护的森林，如巴西的大西洋森林。因此，本研究的主要研究目标是通过使用线性混合效应对为新物种和/或地点校准现有单个树木地上生物量模型的潜力进行全面评估。具体的研究目标是通过允许样本选择方法、样本大小和采样的树木大小范围变化来确定有效校准的最佳方法。特别是，某些物种被用作主要数据集来拟合广义和物种特定的生物量模型，然后针对不同地点和位置的辅助数据集进行校准。次要站点的相似和不同物种都用于校准和评估以前的模型。我们的结果表明，物种水平校准对大多数被检查的物种或个体是有效的，可以在开发新方程所需的样本量低得多的情况下大大提高性能，尤其是对于林分中较大的树木。一般来说，一到三棵随机选择的树木足以有效地校准新物种的生物量模型。我们期望将丰富物种的模型校准与之前开发的针对较少物种的广义模型的使用相结合，可以大大减少对破坏性采样的需求并改善预测，这对于高度受威胁的森林（如巴西的大西洋森林）非常重要。总体而言，结果强调了模型校准在显着改善热带地区等物种丰富的森林中生物量和碳估计的潜力。