Abstract:
This research project involves the use of digital terrain modelling and point sample data analysis to build a predictive surface of site quality across a heterogeneous landscape of plantation forestry. Statistical analysis of existing site quality data, guided by a deterministic conceptual framework of the environmental controls on tree growth, enabled the development of a robust data model. This data model was implemented over the study site using map algebra functions in a Geographic Information System; that is, a virtual surface of site quality was generated from terrain variables extracted from the digital elevation model. This virtual surface used elevation, slope and a measure of solar radiation input and was able to explain 40% of the variance in the actual site index. This is a similar degree of explanation to that observed in the literature where similar methodologies have been employed. The generated virtual surface of site quality was tested against control data and used to assess the variation of site quality measurements within individual pods, which are the management units of commercial forestry operations. The within-pod variation in site quality was related to the variation in the virtual site quality surface. Therefore, it follows that it is possible to explain the variance in site quality by terrain characteristics. This is an important outcome because it suggests that site quality for forest plantations can be adequately modelled from terrain characteristics alone. This means that site quality may be modelled more objectively and without the need for detailed soil or other environmental data, which is expensive to collect and difficult to model because of the complexity of natural systems. The techniques and methodology used here are analysed, and future applications for this type of research are discussed, such as the modelling of forestry sustainability.
