Controlling infestations of Parkinsonia aculeata in a riparian zone at the landscape scale

To prioritize weed management activities it is necessary to predict the extent of future infestations of individual weed species. Management effort should focus on populations which are likely to spread and produce new satellite populations. Deciding which aspects of an invading population to manage, new emerging distant satellites or the existing core from which most seed originates has been the subject of much debate. We use extensive survey data to define the core and satellite populations and dispersal function of Parkinsonia aculeata in the semi-arid rangelands of the Northern Territory of Australia. We described populations in terms of three size classes, classified on the basis of plant height (small (< 0.4 m), medium (> 0.4 and < 2.5 m) and large (> 2.5 m)). Data were recorded as a spatial point pattern and a dispersal function was derived from the distribution of the distance of each medium plant to the nearest P. aculeata neighbour that was greater than 2.5 m. All large and medium P. aculeata were classified as members of core or satellite populations. From these classifications, we tested critical assumptions of Moody and Mack's model, namely (i) the satellites form multiple disjunct foci that do not overlap; (ii) the system is homogenous and there are no abiotic or biotic restrictions to the weeds expansion enabling it to expand radially in any direction; (iii) the densities of all foci are uniform; (iv) there is no extinction through stochastic processes; and (v) there are no threshold effects. When the medium size class was considered, the population density and area occupied by the weed increased. Two satellite populations were comprised solely of medium plants, suggesting the number of satellites increased marginally from eight to 10 when data from both size classes were considered. Most individuals were part of the core population. The core population increased in density and we argue this core population is most likely to give rise to new satellites and should be managed in preference to emerging satellite populations. We suggest the invasion is behaving like a source-sink system, where the core population is the source of most new medium plants.