A generalised hybrid process-empirical model for predicting plantation forest growth

A generic model of plantation growth was developed for Australia's National Carbon Accounting System to allow spatial estimation of carbon stocks over time. Unlike the primary goal of most forest growth models, which is to predict log volume at harvest age, the international guidelines for carbon accounting require estimation of current annual increments of total (above and belowground) biomass. In contrast to most commercial forestry systems that are concerned with rotations of many years, capturing the effects of annual climate variability is important, a feature that would otherwise be largely ameliorated over a forest rotation. While yield tables can provide the basis from which empirically based models can predict an 'averaged' performance over time, a process-based model can capture the effects of variability over short time periods. To utilise the valuable empirical data contained in yield tables, while also capturing the effects of process drivers, a hybrid model has been developed that integrates: center dot a spatially and temporally explicit site class index based on a process model, center dot a simple growth equation modified by the spatial and temporal site index, center dot known empirical constraints on growth (as an average) sourced from yield tables, center dot management effects. Management effects may either increase overall site productivity, and hence biomass accumulation, or accelerate the rate of approach toward site carrying capacity. The effects of management are important and need to be captured in the model. Some 5000 forest management regimes, representing different species, regions, site qualities and variants in management in Australia have been described for application within the model. A national program of identifying both forest areas and forest age classes, using a 30-year archive of Landsat satellite data has been undertaken to provide inputs to the model. Crown Copyright (c) 2006 Published by Elsevier B.V. All rights reserved.