Climatic suitability of South Africa for giant reed, Arundo donax (Poaceae) and a candidate biological control agent, the rhizome- and stem-feeding scale insect, Rhizaspidiotus donacis (Hemiptera: Diaspididae)
Sasha-Ann PILLAY1, Angela Bownes1, 2, Terence Olckers1
1School of Life Sciences, University of KwaZulu-Natal, Scottsville 3209, South Africa
2Agricultural Research Council-Plant Protection Institute (ARC-PPRI), Private Bag X6006, Hilton 3245, South Africa
Arundo donax (L.) (Poaceae) is a widespread and highly invasive reed in South Africa. Following progress in the USA, the scale insect, Rhizaspidiotus donacis Leonardi (Hemiptera: Diaspididae) is being considered as a candidate biological control agent in South Africa. An important component of pre-release evaluations in weed biological control programmes is to evaluate the climatic niches of candidate agents and the target weed to identify areas that are climatically suitable for their persistence in areas of introduction. This study aims to determine whether A. donax has reached its optimum invasion distribution in South Africa and whether South Africa is climatically suitable to support R. donacis populations. The climate-modelling programme CLIMEX was used to model climatic similarities between native range locations of R. donacis and South Africa, and the climatic suitability of South Africa for R. donacis and A. donax, using model parameters based on both native and introduced range (i.e. USA) distribution data. The results indicated that most regions of South Africa are climatically suitable for R. donacis establishment or are climatically similar to locations in the USA where the scale insect has established following deliberate release. Predictions from the combined use of native and introduced range parameters indicate that A. donax is likely to have reached its fundamental distribution range in South Africa. Furthermore, the improved model predictions obtained through the use of both sets of parameters, demonstrate the potentially limited predictive value of using native range parameters alone when modelling distributions of invasive alien plants.