A new numerical model for coral distribution

Abstract

ABSTRACT: A critical review of previous numerical work on coral distribution has shown that all previous workers have used models which assume the probability of occurrence of a coral to be spatially uniform over the study area. This is not necessarily true. Here we define a new 'constrained occupancy model' which permits spatial variation in probability. A case study involving extensive data (46 transects) on the distribution and abundance of 8 coral species and 2 octocorals throughout the Caribbean was used to test the new model against the commonly used Poisson model. The Poisson model demonstrated that 4 of these species were highly significantly aggregated in distribution and six were random. The new constrained occupancy model, allowing for variation in probability of occur- rence, demonstrated that, in fact, 1 tended towards aggregation, most were random, and 2 species were significantly uniform in distribution. We believe that accounting for variation in probability of occurrence in sessile organisms is important when analyzing their distribution, and we consider these findings warrant the attention of field ecologists. While no analytic solution to the new model is known, computational procedures are presented, some of which can be used with a Monte Carlo simulation.