Modeling diameter structures with the Log- Logistic function in natural forests of Durango, Mexico Modeling diameter structures with the Log- Logistic function in natural forests of Durango, Mexico

Abstract

SUMMARY

Diameter distribution models are useful tools for predicting forest growth and yield and planning sustainable forest management activities. The objective of this study was to analyze the fitting capacity of the Log- Logistic probability density function (PDF) using a percentile-based parameter estimator and to evaluate the accuracy of two alternative modeling approaches for diameter distributions in natural stands in northwestern Durango, Mexico. Six percentile estimators were evaluated and compared with the maximum likelihood method based on the performance of the Kolmogorov-Smirnov (KS), Anderson-Darling (AD), and Cramér-Von Mises (W ² ) statistics. For modeling diameter distributions, the graphical and numerical behavior of the prediction (PPM) and parameter recovery (PRM) methods were assessed with mean bias (SM) and mean absolute error (EMA). The best parameter estimator was the diameter that accumulates the 25th and 79% percentiles, considering the percentage of stands where it was most accurate in terms of KS, AD and W ² , as well as its performance with respect to maximum likelihood. Modeling the number of trees per diameter class with the PRM and PPM approaches resulted in similar accuracy based on the measurement of the square mean diameter, basal area per hectare, height and dominant diameter. This work contributes significantly by providing an easily applicable tool in growth models developed for the natural forests of the Sierra Madre Occidental in Mexico.