Abstract
Forest plantations have increased in Mexico and statistical tools are required to evaluate and predict their productivity, to plan their use. The objective was to determine a generalized total height-normal diameter model for Pinus leiophylla forest plantations in Patamban, Michoacán, Mexico. Trees were measured in 33 temporary quadrangular random sites in plantations with initial densities of 2,500 trees ha-1 and survival greater than 75%. 779 total height-normal diameter data pairs were used; with values for height of 1.86-19.60 m and 1.5-36 cm for normal diameter. The fit of seven models was checked through their evaluation by nonlinear least squares with the nls function in R®. For the selection of the model and its readjustment, the parameters of significance, coefficient of determination, root mean square of the error, Akaike criterion, Bayesian and bias were compared. The resulting generalized total height-normal diameter structure was refitted with mixed effects between and within sites. The Wang and Tang model presented the best indicators of statistical fit. The mixed effects models showed an improvement in the explanation of sample variability (coefficient of determination: 2.4%), fit statistics (24.5% on average) and in the precision of the estimates in the root mean square of the error and bias greater than 14%. The generalized model was efficient to represent said sample variability in the P. leiophylla forest plantations of the Patamban Indigenous Community, Michoacán, Mexico.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call