Predictive models of resin production in Pinus pseudostrobus Lindl., in Michoacán State, Mexico

Abstract

Pinus pseudostrobus is a conifer widely used in the state of Michoacán for the extraction of resin; however, current resin extraction methods are empirical and without knowledge of the potential production. Based on mensuration variables and using the ordinary least squares (OLS) method and mixed effects models (MEM), a predictive model for estimating resin production was evaluated. The normal diameter, crown diameter and total height of 215 resin trees were measured; in addition, the resin production per face (2 186 faces) was quantified in an altitudinal range of 2 226 to 2 785 m. After debugging the database and constructing the combined variable (d2tH), a logarithmic linear model was adjusted under two statistical approaches ―OLS and MEM―, using the R® software. By including the altitude covariate in the MEM as a grouping variable, there is an average statistical gain of 17 % with respect to the OLS approach. No issues of non-compliance with the normality regression or homoscedasticity assumptions were observed. A model with global parameters is proposed to estimate the average resin yield for the region and three variants with random parameters, where the altitude of 2 500 m has the highest production. The estimated productivity and its relationship with the altitudinal ranges can be a guideline for the establishment of forest plantations for resin extraction purposes or the development of forest management plans for the species in the Indigenous Community of Nuevo San Juan Parangaricutiro, Michoacán.