González-Moreno P.; Pino J.; Carreras D. et al. 2013. “Quantifying the landscape influence on plant invasions in Mediterranean coastal habitats”. Landscape Ecology.

Abstract

Landscape pattern might be an important determinant of non-native plant invasions because it encompasses components influencing the availability of non-native plant propagules and disturbance regimes. We aimed at exploring the relative role of patch and landscape characteristics, compared to those of habitat type and regional human influence on non-native plant species richness. For this purpose, we identified all non-native plant species in 295 patches of four coastal habitat types across three administrative regions in NE Spain differing in the degree of human influence. For each patch, we calculated several variables reflecting habitat patch geometry (size and shape), landscape composition (distribution of land-cover categories) and landscape configuration (arrangement of patches). The last two groups of variables were calculated at five different spatial extents. Landscape composition was by far the most important group of variables associated with non-native species richness. Natural areas close to diverse and urban landscapes had a high number of non-native species while surrounding agricultural areas could buffer this effect. Regional human influence was also strongly associated with non-native species richness while habitat type was the least important factor. Differences in sensitivity of landscape variables across spatial extents proved relevant, with 100 m being the most influential extent for most variables. These results suggest that landscape characteristics should be considered for performing explicit spatial risk analyses of plant invasions. Consequently, the management of invaded habitats should focus not only at the stand scale but also at the highly influential neighbouring landscape. Prior to incorporate landscape characteristics into management decisions, sensitivity analyses should be taken into account to avoid inconsistent variables.