PS 4-56 - Modelling the effect of spatial factors on the short- and long-scale ecological dynamics of an endemic Florida ecosystem.

Abstract: Ceratiola ericoides (Ericaceae), known commonly as Florida rosemary or Florida sand-heath, is an endangered, long-lived shrub endemic to south-eastern North America. This species inhabits dry scrub and dune habitats, reproduces through dioecious flowering and anemochoric seed dispersal, and has senescent allelopathy that suppresses the recruitment of allospecific plants. In scrub systems, it becomes a dominant species (often co-dominant with oaks (Quercus spp.)), eventually forming a spatial system characterised by large gaps between individuals with inhibited recruitment of allospecific plants. Despite this, some plant species are specialised for recruiting in these systems. One of these species, Hypericum cumulicola (Hypericaceae) is an herbaceous plant endemic to the scrub habitats of the central Florida peninsula. We are developing a model to examine the intra- and inter-species dynamics of C. ericoides in the Florida sand-heath scrub system, as well as the role that distance among plants plays in determining the strength of interactions in this ecosystem. By employing a mix of agent-based models (ABMs) with multi-species occupancy models (MSOMs) and Bayesian statistical modelling, we examine the strength of interactive effects across distance gradients for individual plants. We also incorporate the influence of abiotic factors, such as long-scale fire and precipitation regimes, on recruitment and population parameters to assess the ecological dynamics of the Florida sand-heath scrub system. Here, we present some initial findings from this modelling approach, as well as discuss future directions for the further development of the model for this as well as other ecosystems. Our hope is that this modelling strategy can be employed for other ecological contexts in which spatial relationships between subjects are a (or the) defining feature of the ecology of the system. Components could be incorporated into the model to account for dynamic spatial relationships between agents, to account for motile research subjects as well, opening the possibilities future modellers and researchers.