COMPORTAMIENTO DE DOS MÉTODOS EN LA PREDICCIÓN Y REPRESENTACIÓN GRÁFICA DEL ÍNDICE DE SITIO

The productivity of forested lands .is influenced, mainly, by the site or stand quality that is determined through the highest wood harvest that a forest produces at a time. Frequently it has been assessed through indirect methods such as the Site lndex (SI) that comes from the adaptation of the height-age variables growth curves of normal yield tables. Growth models are crucial for forest management planning. The definition of mathematical functions that show tree growth correctly needs the support of several methods and models in arder to find those that fairly represent the group of data in each particular case. In arder to generate polimorphic site index curves for Pinus pseudostrobus in the Hidalgo-Zinapecuaro region in the state of Michoacan, the parameter prediction and algebraic difference methods were used as well as the Schumacher and Chapman-Richards models. Height-age data were provided by 51 felled dominant trees. Both methods and models were compared on the basis of their statistical fitness and accuracy to represent the data. Results show that the models formulated with the algebraic difference method had an excellent statistical fitness but were defficient in representing the data. In contrast, the models developed with the parameter prediction method were statistically sound and acccurate for data representation.