Tropical forests may have many species, indeterminate ages, and a wide range of growth habits and stem sizes and thus require special modeling techniques. But technique contributes only part of model quality, and much depends on the calibration data and access to the model. Whole stand models have limited utility in these forests, as it is hard to describe the forest adequately with few stand-level variables. Stand table projection and matrix models may be useful where summarized data are available and computer resources are limited, but the many classes required detract from the method. Tree list models offer greater flexibility, enable projections under a wide range of conditions, and provide diverse information. All growth equations should ensure reliable predictions over all tree sizes, sites, and stand conditions. Mortality may be modeled with logistic functions fitted to individual tree data. Two-stage recruitment models are a practical way to predict regeneration where there are many species. Several existing models could be calibrated for tropical rainforests if suitable data were available. Sustainable use of rainforests may depend on maintaining nutrient cycles and ecosystem linkages, and new data and innovative models will be necessary to fully appraise these aspects.
The publisher's version of this article is available at http://www.ingentaconnect.com/content/saf/fs