Simulating vertical and horizontal multifractal patterns of a longleaf pine savanna

Abstract

Many ecological processes (e.g., individual growth, competition, and mortality) are dictated by existing spatial patterns and lead to the generation of new spatial conditions. Spatial patterns are the result of a spectrum of ecological processes operating at widely different time scales. In this study, a cellular automata model that incorporated autecological information for longleaf pine (LLP) and fire effects was used to simulate one- and two-dimensional spatial canopy/gap properties (i.e., the distribution of crown heights and widths) over a savanna landscape. These were quantified using multifractal analysis and were compared to remotely-sensed data from LLP stands from the Disney Wilderness Preserve located near Kissimmee, Florida. Lidar-derived transect information provided canopy height patterns and aerial photography provided crown width and horizontal distribution patterns. Multifractal spectra and size class distributions were found to be sensitive to spatially interactive parameters (i.e., competition, fire ignition and spread probabilities). Simulations with moderate levels of competition coupled with a relatively high fire frequency (once every 4 years) and a relatively high likelihood of fire spread across five-meter grid cells (dependent on litter fuel loads) were shown to create patterns that closely mimic the quasi three-dimensional remotely-sensed measures of this open canopy system.