How do alien plants invade : an analysis of the historical spread and potential distribution of the invasive alien plant Chromolaena odorata in and around Hluhluwe Game Reserve.

Thesis (M.Sc.)-University of KwaZulu-Natal, 2009.

Alien plant invasions and human-induced land transformations are the two most important factors contributing to loss of biodiversity, and Chromolaena odorata (C. odorata) is one of the world’s worst invader plants. It impacts negatively on biodiversity conservation, forestry and agriculture due to its rapid rate of spread, fast growth and hence an ability to transform habitats. It was introduced into KwaZulu-Natal, South Africa in the mid-1940s, and has spread north and southwards along the coastline and adjacent interior. In 1961 it was observed in Hluhluwe Game Reserve, and has invaded and transformed large parts. In this thesis I use C. odorata distribution maps from 1978 to 2001 to investigate invasion pathways, and whether these are related to species or habitat characteristics. Invasion in the twelve vegetation types identified in the Reserve showed two patterns. In some vegetation types invasion expanded linearly, while in others there were three phases of invasion, driven by interactions between habitat suitability, habitat fragmentation and propagule pressure. Expansion of existing invasions (phalanx or wave invasion) through short distance seed dispersal was the primary mode of spread, rather than the formation of new infestations (guerrilla or diffuse invasion). Known distribution of dense stands was used to develop a spectral signature and run a supervised classification of satellite images. I utilized the dense stands and other GIS data sets of climate, altitude, solar radiation, distance from roads, distance from rivers and normalized differential vegetation index from 1973 and 2001 to determine a probability model, and assess the impact of the invasion on biodiversity conservation. The supervised classification underestimated the extent of the invasion, as C. odorata is a cryptic understory species in some habitats. Supplementation by using multiple regression to develop a probability model, and combining these into one map provided a realistic result. I used this map and GIS data sets of vegetation types and protected areas to assess the impact of the invasion on biodiversity conservation. This showed that C. odorata was preferentially impacting sensitive areas, and the conservation of threatened vegetation types is not guaranteed by including them in protected areas.