There is evidence of an accelerating rate of non-native species introductions beyond their natural distribution ranges (Roy et al., 2024). When non-native species remain confined to relatively small ranges, they may increase biotic differentiation between regions (Olden and Poff, 2003); when these species establish self-sustaining populations over extensive areas, they can promote biotic homogenization (McKinney and Lockwood, 1999; Olden et al., 2004). Taxonomic homogenization, defined as increased similarity in species composition between communities or regions (Olden and Rooney, 2006), is the most extensively studied in relation to non-native plants. Often, the Jaccard Similarity Index was used to quantify these effects (e.g., McKinney, 2004; Wani et al., 2023) in various ecosystems and spatial scales, yielding contrasting outcomes (e.g., McKinney, 2004; Qian and Ricklefs, 2006; Dar and Reshi, 2015; Qian and Qian, 2022; Keck et al., 2025). Understanding this process at regional scales is essential for assessing ongoing reshaping of biogeographic patterns, but it has been relatively underexplored (e.g., Wani et al., 2023).

Whether non-native species contribute to floristic homogenization or differentiation among regions may depend on the degree of climatic, geographic, and political distances between them. Geographically closer regions tend to share more similar non-native floras, and human-mediated dispersal can weaken the natural decay of floristic similarity with distance (Yang et al., 2021; Gulzar et al., 2024). Biotic homogenization driven by non-native species spread tends to increase between areas with similar climatic conditions, even when geographically distant (Capinha et al., 2015; Daru et al., 2021; Yang et al., 2021). Moreover, floristic homogenization has been found to be stronger between regions with shared or historical political ties, suggesting that administrative relationships, trade intensity, and policy harmonization contribute to the similarity of non-native species (Yang et al., 2021). Therefore, considering geographic, climatic and political distances between ecoregions might help understanding the biogeographic consequences of non-native species spread.

The influence of non-native species on floristic homogenization and differentiation has received limited attention in biodiversity hotspots, where the impacts may be especially important (e.g., Wani et al., 2023). One such region is the Andes mountain range in South America, a global biodiversity hotspot that spans over 8,000 km across seven countries. Its tropical portion was home of major pre-Columbian civilizations and later became the site of some of the earliest and most intensive European colonization (Grau et al., 2024), and currently faces different changes including climate, land-use, and biotic changes such as biological invasions (Tovar et al., 2022; Fuentes-Lillo et al., 2023).

Numerous studies have analyzed the ecology and biogeography of Andean non-native plants (e.g., Pauchard and Alaback, 2004; Fuentes-Lillo and Pauchard, 2019; Fuentes-Lillo et al., 2023; Fernandez et al., 2025) but few of them have focused on their local effects on biotic homogenization or differentiation (e.g., Castro and Jaksic, 2008; Hurtado-M et al., 2021). A regional-scale evaluation of homogenization or differentiation encompassing the full Andean range is lacking. Thus, our study aims to provide the first regional assessment of whether non-native plant introductions are associated with floristic homogenization or differentiation among Andean ecoregions.

We use the global biodiversity data (GBIF) and the Global Register of Introduced and Invasive Species for the Andes region (GRIIS) I) to assess whether non-native plant introductions contribute to floristic homogenization or differentiation and II) to compare the relationship between similarity of non-native and native plants and three potential contributing factors: (a) geographic distance, (b) climatic distance and (c) political dis-similarity (i.e. number of countries shared between pairs of Andean ecoregions).

We expect plant similarity between Andean ecoregions to decline with geographic distance, but this decline will be less pronounced for non-native species than for native ones due to human-mediated dispersal. We also expect the decay in similarity associated to ecoregion's climatic differences will be stronger for native species, as their distributions are more tightly constrained by long-term adaptation to local environments while many non-native species tend to be generalists or related to human activities that may bypass natural climatic barriers. Political boundaries influence environmental governance and non-native plant introductions (Montti et al., 2024). Ecoregions that span many countries are subject to distinct cultural legacies, national priorities, legal frameworks, and conservation policies, likely affecting introduction, establishment and spread of non-native plant species. Thus, we expect non-native floras will be more similar between ecoregions that share the similar political context (i.e., they are included in the same countries); similarity of native species, however, is expected to be largely independent of political boundaries.

Our study provides novel insights into the role of non-native plant introductions in shaping floristic patterns across a large heterogeneous and transnational region, as well as a global biodiversity hotspot. Contrary to the previous findings of a strong process of floristic homogenization related to non-native plants across regions of one or a few countries (e.g. Qian and Qian, 2022; Wani et al., 2023) we found similar levels of both floristic homogenization and differentiation processes occurring between Andean ecoregions as a result of non-native plant introductions. Approximately one-third exhibited positive HJ values (i.e., homogenization), another third showed negative values (i.e., differentiation), and the rest showed no net change, implying that the introduction of non-native species can drive contrasting biogeographic patterns depending on regional environmental and social conditions (Guo et al., 2018).

The observed pattern in mean homogenization index (HJ) values reveals regional differences in how non-native plant introductions are influencing floristic composition across the Andes. Ecoregions in the Southern Andes—such as the Central Andean dry puna, Chilean matorral, High Monte, and Southern Andean steppe—showed slightly positive mean HJ values, suggesting a subtle tendency toward floristic homogenization. Southern Andes have experienced long-standing human settlement and development, particularly in regions like the Chilean matorral, which may have facilitated the repeated introduction and establishment of the same non-native plants (Pauchard and Alaback, 2004; Fuentes-Lillo and Pauchard, 2019; Fuentes-Lillo et al., 2023). Regarding the ecoregions with negative mean HJ values—such as the Atacama Desert, Cordillera de Mérida páramo, Cordillera Central páramo, and Marañón dry forests— these span the three major Andean regions (Northern, Central, and Southern Andes), suggesting that floristic differentiation is not strictly a latitudinal pattern but rather the result of local ecological constraints, such as extreme aridity (Atacama Desert) or high-elevation climatic specialization (páramos). However, given that the index values remain close to zero, these results should be interpreted with caution, as they could be influenced by biases associated to uneven sampling intensity or availability of non-native plants records across Andean ecoregions.

Consistently with classic distance-decay expectations due to dispersal limitation and spatial turnover (Nekola and White, 1999), geographic distance was negatively associated with similarity in both native and non-native floras). However, contrary to our prediction, the slope of the relationship was significantly steeper for non-native than for native species (Fig. 4a), highlighting a distinct spatial structure in the distribution of non-native plant species. For non-native plant species, we found that the effects of geographic distance and political dissimilarity were not independent, but interacted significantly. Specifically, the negative effect of political dissimilarity on non-native plant similarity was stronger among geographically proximate ecoregions, suggesting that neighboring ecoregions governed by different countries may result in different policies, cultural practices, or introduction pathways that reduce floristic similarity. For example, differences in import regulations, biosecurity measures, and horticultural or agricultural preferences between countries likely contribute to this pattern. In contrast, at large geographic distances, this pattern reversed: greater political dissimilarity was associated with increased non-native floristic similarity. This reversal may reflect the influence of globalized economic and cultural processes that operate beyond national policies. Widely traded ornamental and crop species are often exchanged through international markets and follow similar introduction pathways worldwide. Consequently, distant regions with distinct political systems may still accumulate comparable sets of non-native plants, as large-scale socioeconomic forces—such as international trade and the circulation of agricultural and horticultural species (e.g., Montti et al., 2024)—may override local political differences and promote convergence in introduced floras.

While climatic distance accounted for a greater proportion of the variance in native plant species similarity compared to non-native plant species, the similarity in the slopes of these relationships suggests that climatic distance between regions exerts a comparable influence on both groups. In the case of native plant species, their long-term evolutionary and biogeographic history has been shaped by climate, leading to stronger spatial structuring (Ricklefs, 2004). Non-native species, although generally more ecologically flexible, are still subject climatic filtering after introduction, which limits their successful establishment to ecoregions with more similar climatic conditions (Thuiller et al., 2005; Bradley et al., 2012; Capinha et al., 2015). Overall, our results are in line with those found in previous global studies that showed how regions with similar climatic conditions shared more non-native plant species (Yang et al., 2021; Capinha et al., 2023).

Our study has certain limitations. First, our analyses with Jaccard index are based on presence–absence data at a single point in time, which does not capture temporal dynamics or abundance differences-both of which are important to fully understand the trajectory and impact of invasions (Olden and Poff, 2003; Wani et al., 2023). Second, sampling effort and data completeness are uneven across the Andes, particularly for non-native species (Tovar et al., 2022), which might induce biases in spatial patterns detected (Pino et al., 2009). Third, our analysis did not account for specific socio-ecological characteristics of the ecoregions. Ecoregion-specific ecological, biogeographic, and anthropogenic variables such as elevation range, land use, habitat heterogeneity, disturbances, connectivity, trade volume, and human population density may influence the establishment and spread of non-native species, as well as the degree of floristic homogenization or differentiation.

Despite these limitations, our study is the first to assess floristic homogenization and differentiation driven by non-native plants across the Andes, a heterogeneous mountain system spanning seven countries. Our results show that with the introduction of non-native plants some Andean ecoregions increase their floristic similarity while others increase their differentiation, and that political borders and geographic distance play a significant role controlling these trends. Future research should consider the functional traits of non-native plant species such as growth form that contribute to homogenization (Dar and Reshi, 2015), as well as incorporate ecoregion-specific environmental, land use and socio-economic variables. Such efforts will be useful for anticipating the long-term consequences of biological invasions and for informing coordinated conservation and management strategies across this biodiverse region.

Datasets of plant species downloaded from the GBIF and GRIIS databases for the Andes region are included as Supporting Information Table S1. Climatic characteristics of the Andean ecoregions, obtained from the WorldClim BIO Variables V1 dataset, are included as Supporting Information Table S2. Data and code for the Jaccard index, homogenization index, climatic distance, and regression analysis are available on GitHub: https://github.com/YohanaJ/HJ_Andes/tree/main.