Development of a dynamic stand growth model and optimization of the management of "Pinus pinaster" ait. in Asturias

Abstract

This dissertation comprises three studies that provide tools for improving the management of maritime pine (Pinus pinaster Ait.) stands in Asturias (NW Spain). The studies were based on information derived from two networks of research plots established in stands of this species in the region. In Study I, several age-independent methods were evaluated for site index estimation and height growth prediction, as information about stand age is not always collected in forest inventories. The growth intercept method proved the best for estimating site index, while an age-independent equation that uses climatic variables as predictors behaved the best for height growth prediction. In the absence of age data, these methods can be used to provide the input information required by the dynamic stand growth model developed in Study II. In this model, it was assumed that the state of a stand at any age can be described by dominant height, number of stems per hectare and stand basal area, which can be projected to any other age by using transition functions. Two alternative procedures can be used to estimate total and merchantable volumes from these state variables: a stand volume ratio function or a disaggregation system. The former proved the best method as it is more accurate and computationally more efficient. In addition, comparison of the whole model with those developed for the nearby regions of Galicia and northern Portugal showed that a single model may suffice for the entire NW the Iberian Peninsula. Using the developed model and an optimization algorithm, Study III optimized the stand-level management of the species in Asturias in economic terms, considering the number, timing and intensity of thinning operations, as well as the rotation age as decision variables. The depth-first search (DFS) method was initially used to compare the stand volume ratio function and the disaggregation system: both provided similar results, although the former was computationally more efficient and was therefore selected for further optimizations. The DFS and five direct search optimization methods (one based on one solution vector and four on a population of solution vectors) were then compared using a fixed discount rate. The differential evolution method produced the most consistent results and it was used to evaluate the effect of site quality, stem density and discount rate on optimal management schedules. In general, three heavy thinning operations were considered in the optimal schedule. As site quality and discount rate increased, the optimal timing of cutting occurred earlier, while stem density was not influent.