Ajuste de equações lineares e não lineares para estimativa do volume comercial no estado do Acre, Amazônia Brasileira

Accurate wood volume estimation is essential for planning sustainable forest management in the Amazon. In dense tropical forests, the high structural heterogeneity requires models that account for various forms of variation. In this context, the present study aimed to evaluate twenty volumetric models, both linear and nonlinear, to estimate the commercialvolume of trees in a dense ombrophilous forest under sustainable forest management in the state of Acre, Brazil. A total of 223 trees from 20 species were analyzed, with diameters ranging from 50.4 to 149.9 cm. The performance of the models was evaluated based on statistical indicators such as root mean square error (RMSE), Akaike Information Criterion (AIC), and mean absolute bias. Models that included commercial diameter and commercial height as independent variables showed better performance, particularly the logarithmic forms of Spurr and Schumacher-Hall. The nonlinear Schumacher-Hall model proved to be the most efficient (RMSE = 1.634; AIC = 856.76), with normally distributed residuals, demonstrating statistical consistency. Including commercial heightas a predictor was essential for improving estimation accuracy and is recommended for forests with heterogeneous structure, such as those in the Western Amazon.