Modelling the spread of an invasive woody taxon: Rhododendron ponticum L.

Abstract

Simulation of the present-day distribution and abundance of rialRhododendron ponticum L. at the Glen Etive study site in the Western Highlands of Scotland was achieved using a simple deterministic model (MIGRATE). The model utilises the demographic and dispersal parameters characteristic to a species and a knowledge of the environmental history of the area through which it spreads to simulate patterns of spread. Biotic parameter values were derived from simple field measures and from data in the literature. "Habitat maps" were constructed on the basis of observations made in the field as to the likely relationships of Rhododendron to biotic and abiotic features of the habitat. Habitat features and their attributes were digitised and recorded in an ARC/INFO Geographical Information System (GIS). The simulation of changes in habitat through time was attempted using different habitat maps composed of cells containing unique values for relative carrying capacities, which were representative of the state of the habitat at a certain time. These habitat maps could only influence the dynamics of spread at the intervals between generations. Implementation of habitat changes was dependent on the cohort structure of the model which limited the resolution and exact order of changes that could be taken into account. Model simulations were tested for accuracy against the present-day distribution and abundance of the invading population as mapped in the field, and as seen in aerial photographs from 1946."Null" simulations showed that environmental factors were important determinants of the migration rate. Having achieved accurate simulation of a past and present distribution at a fine spatial scale from two initial foci of introduction in 1910, predictions were made as to the likely pattern of future spread. Predictions for the future were then made considering the effects of control regimes. The importance of the implications of the pattern of spread to migration research and to conservationists, considering the ecological impacts of Rhododendron observed at the study site are discussed in relation to previous findings. More specifically the importance of the long-distance dispersal function to the invasion process is highlighted, and it is suggested that evolution should favour strategies resulting in long-distance dispersal. The reason for large seed crops is discussed in this light. This project represents an integration of field techniques, biotic data available from the literature, a deterministic model, a GIS and aerial photography.