Forest biomass quantification (Mg ha⁻¹) is essential for ecosystem monitoring, especially in areas under anthropogenic pressure, such as Atlantic Forest fragments. This study aimed to compare remote sensors in biomass mapping and population stock estimation of an Atlantic Forest fragment. Ten 0.1 ha plots were randomly distributed within a 17 ha fragment. Data from Sentinel-1 (S1), Sentinel-2 (S2), digital aerial photogrammetry (DAP), and their fusion were evaluated for the construction of predictive models. Linear models with two predictors were fitted: one for each sensor and another using data fusion, selecting the best predictors among all. The models were applied to estimate stand biomass using a regression estimator. The data fusion model showed the best predictive performance (RMSE = 41%), while the DAP-based model had the highest error (RMSE = 64%). However, the most accurate population estimate was obtained with the S2-based model (SE = 21 Mg ha-1), with a relative efficiency 7% higher compared to the traditional inventory (SE = 22 Mg ha-1). Estimates based on DAP, S1, and fusion were less accurate than those from the field inventory. The selected metrics, such as vegetation indices (S2) and textural metrics (S1), reflected the sensors' sensitivity to canopy structure and foliage abundance. DAP showed limitations, possibly due to its low canopy penetration. It is concluded that although the data fusion between DAP and S2 produced the best model for biomass mapping, S2 alone proved more advantageous for population estimates in forest fragments with limited sampling.
The classification of productive capacity is a fundamental component of forest management, but traditional univariate methods, such as the guide-curve, may oversimplify the complex interactions of factors determining site productivity. Therefore, this study addresses a research gap by assessing the potential of Kohonen Neural Networks (SOM), a multivariate artificial intelligence technique, as an alternative for stratifying commercial eucalyptus stands. The objective was to compare the classification obtained by the guide-curve method with the artificial neural network (ANN) approach, assessing the superiority of the latter. To this end, data from 750 eucalyptus stands in Minas Gerais, Brazil, were classified using both the guide-curve method (Schumacher's model) and an ANN (SOM) with subsequent hierarchical clustering. Validation and comparison were performed using discriminant analysis and a contingency matrix. The results indicated that the ANN provided a more accurate stratification, notably by reclassifying 26% of the stands from the guide-curve's Medium class to the Low class and refining the High-productivity class into a more elite group. The cohesion of the ANN-derived clusters was validated by discriminant analysis, which achieved an overall accuracy of 69.9%. It is concluded that the ANN is superior to the traditional method, providing a more realistic, multidimensional classification with greater sensitivity in detecting low-productivity areas and greater specificity in identifying elite stands. The methodology is thus established as a strategic tool for enhancing planning and optimizing operations in precision silviculture.