Assessing forest productivity at local scales across a native eucalypt forest using a process model, 3PG-SPATIAL

To assess the appropriateness of a simplified process model to predict forest growth over a diverse area of forest, we used a geographic information system (GIS) implementation of the 3-PG model (3PG-SPATIAL) with spatial layers of climate and soils to predict forest growth and compare it with existing plot-based data. The study was undertaken across 50,000 ha of eucalypt forest within Bago-Maragle State Forest, Australia with stands ranging in age from 16 to 87 years. At the 16 validation plots, the accuracy of the 3PG-SPATIAL predictions varied depending on the forest variable modelled. There was a strong relationship between predicted and measured standing volume (adjusted R-2 = 0.86, p < 0.001, S.E. = 82 m(3)/ha, n = 22). The relationships between stocking and mean diameter at breast height (131314) were less (stocking: adjusted R-2 = 0.86, p < 0.001, n = 22; and mean DBH: adjusted R-2 = 0.59, p < 0.001, n = 24). Although the independent validation set is small, the results indicate that 3PG-SPATIAL described the spatial variability in native forest growth in the region using relatively simple input data and parameter values based on current physiological knowledge. There is evidence to suggest that it is now possible to provide forest managers with operationally relevant information over large areas using a simplified process model within a GIS environment. (C) 2001 Elsevier Science B.V. All rights reserved.