Study area

The study sites were selected as part of the “Monitoring of Invasive Alien Species” project conducted by the NIE. To investigate the distribution patterns of S. angulatus in riparian areas, four sites characterized by riparian habitats were selected for analysis. The four study sites included one location in the Han River basin (Sa1) and three locations in the Nakdong River basin (Sa2-Sa4) (Fig. 1, Table 1).

The monthly average temperature (°C) and monthly precipitation (mm) recorded from 2019 to 2024 at the nearest automatic weather observation stations to each study site are shown in Fig. 2 (Korea Meteorological Administration, 2025).

Survey and analysis method

The four selected study sites were designated as fixed monitoring plots, and field surveys were conducted annually over a period of approximately 5 years, from 2020 to 2024. Field surveys were conducted between July and September, the period when S. angulatus reaches its peak growth.

To understand the competitive interactions and spread patterns of S. angulatus within the surrounding vegetation at each site, a current vegetation map was prepared. The same survey area was maintained as much as possible at each fixed site. However, due to the characteristics of the river environment and anthropogenic disturbances (e.g., river construction), the surveyed area varied slightly between years. To accurately assess the relative dominance of S. angulatus communities, the ratio of the S. angulatus community area to the total surveyed area was calculated and analyzed.

In addition, vegetation surveys were conducted in parallel to examine the distribution characteristics of S. angulatus. The vegetation survey was conducted using the quadrat method, recording the coverage of species occurring in each vegetation layer. At each study site, ten 1×1 m quadrats were randomly established within the entire area of the S. angulatus community for sampling. The coverage of each species was recorded using a nine-grade scale based on the modified Braun-Blanquet cover-abundance scale (Westhoff & van der Maarel, 1978). Based on the collected data, the importance value (IV) for each species was calculated as the mean of its relative coverage and relative frequency, in order to compare the relative dominance of the target species within the community. Using the IV, the Shannon-Wiener diversity index (Shannon, 1948) was calculated to evaluate annual changes in community diversity.

Statistical analysis

Annual variations in the Shannon-Wiener diversity index (H′) and IV of S. angulatus communities were tested for each study site from 2020 to 2024. Since the data were repeatedly measured at the same sites, a non-parametric repeated-measures test (Friedman test; Friedman, 1937) was conducted. When significant differences were found, pairwise Wilcoxon signed-rank tests were performed as post hoc analyses to identify specific differences between years (Wilcoxon, 1945). To correct for Type I errors arising from multiple comparisons, the Holm method was applied (Holm, 1979).

To examine changes in vegetation community structure, a non-metric multidimensional scaling (NMDS) analysis was performed (Kruskal, 1964). The NMDS was based on Bray–Curtis dissimilarity (Bray & Curtis, 1957) and reduced to two dimensions, and the analysis was conducted using the metaMDS function of the vegan package (Oksanen et al., 2025). The stress value was presented as an indicator to assess the goodness of fit of the community structure. To examine differences among communities, permutational analysis of variance (PERMANOVA) was performed using the adonis2 function with 999 permutations (Anderson, 2001), and the homogeneity of multivariate dispersion among groups was tested using the betadisper function (Anderson et al., 2006; Warton et al., 2012).

All statistical analyses were performed using R version 4.4.1 (The R Foundation, Vienna, Austria; R Core Team, 2024), with a significance level set at P<0.05.

Temporal and spatial variation in Sicyos angulatus communities

Over the 5-year period, the proportion of the S. angulatus distribution area relative to the total vegetation area at each site exhibited a dynamic pattern, characterized by repeated increases and decreases (Table 2). In particular, site Sa1 showed the greatest fluctuation, with the distribution area sharply decreasing to 11.5% in 2021 and then increasing to 65.3% in 2023. Similarly, site Sa2 exhibited a sharp rebound, increasing from 5.3% in 2021 to 49.3% in 2022, indicating a high degree of community recovery resilience. At site Sa3, the distribution area decreased to 29.6% in 2021 and then increased again to 40.8% in 2022, showing a moderate level of fluctuation. In contrast, site Sa4 maintained a relatively stable proportion of approximately 15-20% throughout the 5-year period. These dynamic changes in S. angulatus communities are associated with the periodic disturbances in river ecosystems, and are likely to be strongly influenced by summer rainfall patterns. S. angulatus prefers humid, nutrient-rich riverine environments and possesses traits that favor seed dispersal by flooding and rapid recolonization of open niches created by the removal of competing vegetation (Uchida et al., 2012). Therefore, heavy rainfall events are considered decisive environmental factors that optimize the growth conditions of S. angulatus, enhance its early dominance, and consequently expand its community coverage.

The species diversity within S. angulatus communities showed a decreasing trend over time, whereas the IV of S. angulatus exhibited a continuous increase (Fig. 3). When the annual average changes were analyzed by combining all study sites, statistically significant differences were observed overall (Friedman test, P<0.05) (Fig. 4). The species diversity index within the communities declined from 3.1 in 2020 to a minimum of 2.0 in 2023, indicating a progressive simplification of the community structure. In contrast, the IV of S. angulatus increased sharply from 5.6% in 2020 to 32.9% in 2023, confirming its intensified dominance. These results are consistent with previous studies reporting that the intensity of invasion negatively affects the species diversity of native vegetation (Valone & Weyers, 2019). As the dominance intensity of invasive species increases, the reduction in native plant diversity within the community becomes more pronounced (Valone & Weyers, 2019). In particular, S. angulatus effectively excludes competing plants by maximizing light interception and spatial occupation through its dense climbing growth form (Önen et al., 2015; Smeda & Weller, 2001b). However, the post hoc pairwise Wilcoxon test did not reveal any significant differences between specific year pairs. This is likely because the number of study sites was limited and the magnitude of annual variation was relatively small, making it difficult to achieve statistical significance in pairwise comparisons.

The NMDS analysis conducted to examine changes in the plant community structure within S. angulatus stands yielded a stress value of 0.212, indicating an acceptable level for interpreting the community structure. The community ordination showed partial separation both by year and by site (Fig. 5). In the annual analysis, the 2020 community was relatively distinct from those of other years, while the 2023 community also exhibited greater variability, indicating pronounced structural changes in specific years. Sites Sa1 and Sa4 showed relatively higher temporal variability in their distribution patterns compared to other sites, and particularly, Sa4 tended to form communities spatially separated from the others. These annual separations and increased variability in community structure are closely associated with the previously observed decline in species diversity and the intensified dominance of S. angulatus (Figs. 3, 4). In other words, the continuous spread and dominance of S. angulatus are interpreted to have simplified riparian plant communities over time and driven progressive homogenization of vegetation structure across sites (Valone & Weyers, 2019). The phenomenon whereby increased dominance of invasive species reduces spatial community diversity has been documented in several studies, and in frequently disturbed riparian environments, invasive species tend to rapidly reestablish after disturbance, thereby accelerating temporal and spatial homogenization of vegetation (Anderson et al., 2006; Warton et al., 2012). The relatively high interannual variability observed at Sa1 and Sa4 suggests that these sites experienced greater environmental disturbance or heterogeneity in the invasion process compared to the other sites. Therefore, the community separation patterns and high variability identified in the NMDS results are interpreted as reflecting transitional stages in which hydrological disturbances and the spread of invasive species jointly alter vegetation stability and species composition at each site.

PERMANOVA results revealed that both year (R²=0.048, P=0.001) and site (R²=0.075, P=0.001) had significant effects on the structure of plant communities. However, the test for homogeneity of multivariate dispersion indicated slight violations of the homogeneity assumption for both year and site groups (P<0.05). Thus, the significant PERMANOVA results likely reflect not only differences in species composition among groups but also the combined influence of community variability observed in the NMDS analysis.

Suggestions for management strategies

Based on the findings of this study, S. angulatus progressively reduced species diversity within its communities over the 5-year observation period, while its IV increased sharply, indicating a progressive intensification of dominance within the community (Figs. 3, 4). These changes contributed to the simplification and spatial homogenization of vegetation structure over time, consistent with previous findings that higher invasion intensity leads to a pronounced decline in native plant diversity (Smeda & Weller, 2001a; Valone & Weyers, 2019). Furthermore, the year-specific community separation and high variability observed at sites Sa1 and Sa4 in the NMDS analysis suggest transitional stages of reduced vegetation stability, driven by the combined effects of disturbance intensity and invasive species expansion (Anderson et al., 2006; Warton et al., 2012). Taken together, the expansion of S. angulatus represents not merely an increase in alien species abundance, but an ecological threat capable of reducing the structural resilience of riparian plant communities. In particular, riparian ecosystems are highly vulnerable to disturbance and exhibit high reinvasion rates after removal (Richardson et al., 2007). Therefore, instead of relying solely on physical removal, an integrated management approach combining removal, restoration, and long-term monitoring is essential for sustainable control.

In this context, management efforts should be strategically designed to reflect the temporal and spatial dynamics revealed in this study. Rapid removal immediately after major rainfall or flooding events is critical, as these disturbances promote seed dispersal and rapid recolonization. Because germination in Korea mainly occurs from April to May and can continue until mid-September (Kang, 2014; Kurokawa, 2009; Messersmith et al., 1999; Smeda & Weller, 2001a), two to three follow-up removals throughout the growing season are recommended to prevent seed set and minimize seed-bank replenishment. Priority should be given to sites that exhibited high interannual variability or rebound patterns, such as Sa1 and Sa2, whereas relatively stable sites such as Sa4 can be monitored at a baseline level. To prevent reinvasion after removal, active revegetation with native riparian species possessing dense growth and strong rooting ability is essential to occupy open niches and stabilize the soil (NIE, 2021; Richardson et al., 2007). In addition, management should aim for “zero seed set” by removing vines before flowering and fruiting, considering that each plant can produce thousands of seeds (Esbenshade et al., 2001; Kurokawa, 2009; Smeda & Weller, 2001a). Regular monitoring of IV and H′ during the growing season (July-September) will help detect early reinvasion, and action thresholds (e.g., IV ≥30% or area coverage ≥30-40%) may be adopted to trigger intensified control (Valone & Weyers, 2019). Furthermore, coordinated management between upstream and downstream reaches should be implemented to suppress propagule flow, particularly after floods, since seed and plant fragments can easily disperse along water currents (Richardson et al., 2007; Uchida et al., 2012).