Spatial analysis to identify invasion colonization strategies and management priorities in riparian ecosystems

Abstract Exotic species have different strategies to invade a new area. Invasive species limited by establishment have a wide dispersion range and high seedling mortality rate. Riparian forests are harsh environments for the establishment of plants. In this sense, we tested the hypothesis that the invasive process of G. triacanthos is limited by establishment in a riparian forest of the Uruguay River and that it begins in the fluvial bank at the transition zone between forest and grassland and moves towards the shore border. We expected that: (1) the distribution pattern of adults of G. triacanthos would be decoupled from the distribution pattern of seedlings and saplings (pre-reproductive stages); (2) the population would have a cluster distribution pattern related to the occurrence of punctual events that favoured the development to adult stages; and (3) the abundance of seedlings, saplings and adults would be higher in the transition zone (between forest and grassland) than in the shore border of the fluvial bank. The spatial pattern of G. triacanthos along the Uruguay River was analyzed using spatial point pattern analysis: autocorrelation indexes, hot spots and wavelet analysis. The effect of the fluvial bank position on the abundance of G. triacanthos was evaluated using non-metric multidimensional scaling (NMDS). This methodology allowed us to detect a primary invasion area, an invasion focus and the dispersal area. This pattern corresponded to an infiltration invasion pattern limited by establishment. It did not seem to be affected by position in the fluvial bank. This information could be relevant to establish correct decision-making in the control of an invasive process and provide an applicable methodology for any other invasion process, especially in riparian ecosystems.