Validating generalised diameter-height models for several species and heterogeneous crop structures.

Abstract

An examination of the suitability of generalised height-diameter models for growth modelling and augmenting inventory measurements was undertaken. A large database of repeated measurements taken from crop structure experiments since 1963 to the present in Ireland was used. We used a distance independent individual tree height-diameter model to investigate whether inclusion of competition variables can be used to predict variations in height across a wide range of species and silvicultural management regimes. To this end, we stratified the heterogeneous dataset post-hoc into a variety of constituent species, management and silvicultural strata. In addition, we attempted to control for site-specific effects and serial correlation by using a mixed-effects framework in an effort to identify site specific heightdiameter variables not explained by the model. The generalised model typically performed well for each species and silvicultural treatment. The most noticeable impact of treatment was observed in plots of Sitka spruce (Picea sitchensis (Bong.) Carr.) with differing spacing. The magnitude of inter-plot variability as modelled by a random effect related to the height asymptote varied between species, possibly as a result of inter-species differences in tolerance to variability in environmental growing conditions. Following validation against external data, we show that these generalised models could be used when, in the case of growth modelling for example, it is sometimes necessary to derive individual tree heights from individual tree diameters, perhaps in standard inventory plots where tree height is not measured on every instance that DBH is measured.