A Procedure for Choosing Tree-Stem Volume Equations Previously Fitted in a Forest.

Tree-stem volumes constitute critical measurements for forest management. We propose a general two-criterion procedure for choosing the most reliable and robust volume equation(s), among a set of them previously fitted in a forest. Also, the proposal can be used as an additional criterion when fitting new equations to a single sample. Through simulated volumes, we considered how well a model estimates the volume generated by itself (the more accurate estimation, the better) and measured a model's variability of the mean errors when it estimates volumes rebuilt by other models (the lower variability, the better). The procedure is exemplified with volume equations of Pinus patula in the Forestry Management Unit 1302 in Hidalgo, Mexico. In the example, we found that a linearized form of the combined-variable equation results reliable and robust. This selection coincided with the one made with data of a recent 150 trees sample (validation), where the rebuilt-estimated mean volume differs 0.034 m3 from the observed mean volume; it represents around US$2.6. Therefore, it is shown that simulation becomes an effective tool to identify equations that do not reproduce the volume of trees well in the current forest and helps to choose reliable and robust equations. [ABSTRACT FROM AUTHOR]