Spatial pattern analysis and modeling of Heterotheca subaxillaris and Lespedeza cuneata in a South Carolina old-field

Abstract

Investigating and modeling spatial patterns of co-occurring plant species can reveal potential biological interactions. Occurrence of Heterotheca subaxillaris, an annual, potentially depends on the neighborhood density of a nitrogen fixing perennial, Lespedeza cuneata. After all individuals of both species were mapped and initial calculation of Cross-K functions revealed a positive spatial relationship between the species at all lag distances, the spatial positions of Heterotheca were modeled as a Poisson process whose intensity was dependent on a neighborhood density index (NDI) of Lespedeza. Both a juvenile and an adult data set allow for comparison of patterns at different life stages of Heterotheca, and various forms of the NDI were compared to determine the best characterization of the neighborhood with respect to Lespedeza. After parameterization using maximum likelihood estimation, model fit was assessed by a Cramer- Von Mises test statistic. Quadratic models performed better than linear models. Even the best fitting models adequately fit the data for only certain lag distances, suggesting that there may be complex biological mechanisms operating in this system.