Seeing the soil through the trees: the utility of stem shape and taper in the butt swell for predicting soil depth in Australian Pinus plantations

Abstract

"Precision forestry" is an emerging concept that is currently shaping the management of commercial forests worldwide. An increasing proportion of the world's wood is grown in plantation forests, so more site-specific or precise management of these forests is important for global wood supply, as well as for the profitability and sustainability of individual plantation estates. Managing plantations at such a fine scale requires a much more detailed understanding of spatial variation in soil resources than has been the case in the past. Current approaches to soil mapping are constrained by the logistic limitations and considerable expense of soil sampling. For this reason, finer-scale soil mapping across the large areas that characterise plantation estates is not feasible. Relationships between soil properties and attributes of tree growth are well established for major plantation species. The relationship between soil properties and stem shape and taper in the base or butt swell section of the tree stem has not previously been established or quantified, but physiological models of stem development suggest an association between this section of the stem and the quantity of soil resources. This thesis investigated the nature and utility of stem shape and taper in the butt swell section for predicting forest soil depth.

Fieldwork to address this topic was conducted at a series of case study plantation sites in southern and eastern Australia. A regression model for predicting soil depth from tree shape measurements in the basal 2 m of the stem was developed and improved progressively as data was acquired from each case study site. Stem shape and taper in the butt-swell section of the stem was found to be strongly and predictably related to soil depth across all sample sites, which encompassed a broad range of soil types and depths. At its current stage of development, the model requires calibration for the location at which it is applied from a relatively small number of soil depth measurements. The quality of the map generated by the model was found to be equivalent to that produced using the conventional method of soil mapping; and better than the map generated by a model based on the well-established tree height.

The results of this thesis are encouraging for fine-scale soil mapping in plantation forests. They suggest, firstly, that variation in the shape and taper of the butt swell section of the tree stem can be used to predict soil properties - certainly soil depth, and possibly, with further development, other soil properties relevant to tree growth; and secondly, that there is considerable promise for a simple tree-based approach to mapping fine-scale spatial variation in at least soil depth, and perhaps other soil properties. This, in turn, offers the prospect of a feasible and low-cost means of generating the information necessary to support finer-scale plantation management, such as that envisaged by precision forestry. -- provided by Candidate.