Measuring Forest Site Quality Using the Parameters of a Dimensionally Compatible Height Growth Function

Abstract

Abstract This article presents a new approach to modeling forest site quality which centers on interpreting as measures of site quality the parameters of a stand height growth model. It is argued that site index, which relies on a single height at a particular age, is not capable of reflecting important differences in the shapes of height growth trajectories. In the proposed approach, the parameters of the height growth function and the function itself completely describe the projected dominant height growth curve. In order to make the approach applicable in a wide variety of modeling situations, a method of allowing the number of site quality parameters to vary is proposed. The most flexible model configuration allows the estimation of all parameters separately for each stand and requires remeasurement or stem analysis data that span a relatively long period of time. Less flexible model configurations involve the simultaneous estimation of parameters that are different for each stand (stand-level parameters) with parameters that are the same for all stands (global parameters). Stand-level parameters measure different aspects of site quality that are unique for each stand and global parameters measure different aspects of site quality that are unique for each stand. These less flexible model configurations may be useful when modeling objectives warrant combining stand-level and global parameters or in situations where the mixture of longitudinal and cross-sectional data precludes fitting the most flexible model form. The approach works best with a growth function that offers maximum flexibility with a minimum of parameters and when the parameters of the growth function have intuitive biological interpretations. A new growth function with these properties is presented in this paper. The new growth function is dimensionally compatible and has two biologically meaningful parameters: a rate parameter and a maximum size parameter. The integrated form of the growth function also includes initial conditions which can be specified with data or as an additional parameter to be estimated. Two applications of the approach to modeling dominant height growth of loblolly pine (Pinus taeda L.) are described: one where sufficient longitudinal data allowed fitting several model forms and another where data limitations required the use of less flexible models. For. Sci. 38(2):409-429.