A SPATIAL-ANALYSIS OF THE ENVIRONMENTAL RELATIONS OF RAIN-FOREST STRUCTURAL TYPES

The environmental relations of rainforest structural and physiognomic characteristics (in the Wet Tropics of Queensland, Australia) were quantified using the Monomax algorithms. The probabilities of occurrence for these characteristics were predicted in relation to indices of soil parent material and long-term mean monthly climate estimated at the point of survey for a set of 256 sites. A relatively small number of indices proved the best predictors (minimum temperature of coldest month; maximum temperature of hottest month; precipitation seasonality; precipitation of driest quarter; nutrient supply index). Spatial predictions of potential distributions were generated by coupling the predictive functions to gridded estimates of the indices used as independent variables. Landsat MSS data were used to restrict the analysis to forested land. Numerical classification and ordination techniques were also employed to gain further insight into the spatial relationships. The assumptions of monotonicity required by the Monomax algorithms drew upon hypotheses of plant-environmental relations. The results supported the notion that as long as moisture is not limiting, the likelihood of both larger leaf size and complex structure increases with increasing values of temperature and nutrient supply. The potential distributions of structural types were suggested as useful biological units for ecological comparison and conservation evaluation.