Aníbal Pauchard1, Ramiro Bustamante2, Alicia Marticorena3, and Alejandra Jiménez1. (1) Universidad de Concepción, Instituto de Ecología y Biodiversidad (IEB), (2) Universidad de Chile, Instituto de Ecología y Biodiversidad (IEB), (3) Universidad de Concepción
Homogenization of plant communities caused by the invasion on non-native species is a global scale phenomenon. However, homogenization patterns could change depending on the scale of the observations, but little empirical evidence has been collected and analyzed at multiple scales. We used a multiscale approach to understand patterns of homogenization in plant communities in and around protected areas of south-central Chile. We sampled 8 protected areas and their surrounding matrices using a nested plot design. In total, we sampled 144 plots and recorded 445 native species and 131 non-native species. We analyzed changes in plant community similarity using native species assemblages as stage 0 and native plus non-native species assemblages as stage 1 (ANOSIM algorithm with Primer 6.0). Similarity analyses were run at two scales: using plots as sample units (n=144) and using protected areas and adjacent matrices as sample units (n=16). Additional analyses were run at each study site to detect local trends in similarity patterns and to contrast protected areas and adjacent matrices. We found significant differences in community similarity between stage 0 and stage 1 at both scales, stage 1 showed higher similarity between plots (p<0,05) and sites (p<0.05). MDS analyses show that native species assemblages are clearly differentiated for each study area, while non-native species are highly similar even across distant areas. Adjacent matrices show higher percentage of alien species and therefore higher homogenization when compared to protected areas. These results indicate homogenization is occurring even in more pristine protected areas, and that a general trend of homogenization caused by non-native species is consistent across scales. Funded by Fondecyt 1040528 and ICM P05-002.