You are here:
THE USE OF CELLULAR AUTOMATA MODELING APPROACHES TO UNDERSTAND POTENTIAL IMPACTS OF GM PLANTS ON PLANT COMMUNITIES
Citation:
Colasanti, R., R. Hunt, AND L S. Watrud. THE USE OF CELLULAR AUTOMATA MODELING APPROACHES TO UNDERSTAND POTENTIAL IMPACTS OF GM PLANTS ON PLANT COMMUNITIES. Presented at New Kind of Science, Boston, MA, April 22-25, 2004.
Description:
The development of models is of interest to ecologists, regulators and developers, since it may assist theoretical understanding, decision making in experimental design, product development and risk assessment. A successful modeling methodology for investigating such characteristics must include both a plant's spatial and functional behaviour within a plant community as a function of its physiology. However, can the full dynamics of a plant community emerge from the physiological properties of its constituent species? Over the last twenty years successful simulations of spatial processes of vegetation have been conducted by a number of authors (Barkham & Hance 1982; Crawley & May 1987; Green 1989; Inghe 1989; Colasanti & Grime 1993) using a cellular automata (CA) method of modeling where a plant's distribution within a two-dimensional environmental grid is determined by rules relating to phenomena such as seed dispersal, clonal expansion and interactions with adjacent plants. We believe that these same methodologies can be utilized in order to predict the potential impacts of genetically modified (GM) plants. To do so we must understand how the engineered traits are expressed in an ecological context. It would be a daunting task to experimentally evaluate the full multiplicity of potential pair-wise interactions between GM plants and native plants under a broad variety of actual environmental conditions. We propose therefore to model interactions between GM plants and the natural environment by describing the plants and the effect of the GM trait in terms of plant functional types (Grime 1979). This approach takes the external factors which limit the amount of plant material present in any habitat and classifies them into two categories: (1) stress, defined with regard to the availability of nutrients and (2) disturbance, which refers to the destruction of plant material. The ecological characteristics of plants both natural and genetically modified, can be described based on functional type i.e. as determined by their quantifiable physiological relationships to stress and disturbance. By ascribing the large number of plant ecological characteristics to a smaller number of functional types the problem becomes tractable. The presentation will describe a CA individual-based model constructed from a plant functional type rule base, that does indeed reproduce all of the main population attributes and vegetation processes that are observed in field conditions. Competition (both natural and genetically modified) in relation to environmental factors (both natural and man-made) has been successfully simulated, and in considerable detail. We will also describe how the model will be validated using data from field and mesocosm studies bing carried out by researchers from the Western Ecology Division.