Background

Invasive plant species are non-indigenous species that introduced into a new geographic region either deliberately or inadvertently and impacting negatively on agriculture, horticulture, and wild ecosystem (Cullen et al. 2011; Eminniyaz et al. 2013; Ricciardi 2013; Early et al. 2016). They alter the dynamics of plant communities and imperil the stability and functioning of ecosystem by affecting nutrient cycle, increasing soil acidity, competing with indigenous flora, and inhibiting their regeneration (Manchester and Bullock 2000; Ehrenfeld 2003; Hellmann et al. 2008; Bradley et al. 2010; Tiedeken and Stout 2015). The climate change exacerbates the threats and loss in ecosystem through various mechanisms including removal of climatic barriers and enhances to spread invasive species (Hellmann et al. 2008; Bradley et al. 2010). The invasive plant species have a higher tendency to shift their niche more rapidly compared to native species, and they are more likely to adapt in new environment (Hellman et al. 2008; Shrestha and Shrestha 2019). In addition, invasive plant species benefit from atmospheric carbon dioxide (CO2) enrichment and global warming more than native plant species (Liu et al. 2017). Therefore, integrative study of climate change and biological invasion are required for long-term management of invasive plant species.

Globally, approximately 6500 invasive plant species are estimated to occur in terrestrial and aquatic ecosystems, and one sixth of the global land surface is highly vulnerable to invasion (Early et al. 2016). In the Republic of Korea (ROK), 14 species of plant species such as horseweed (Conyza canadensis), German evening primrose (Oenothera biennis), and tall fescue (Festuca arundinacea) are categorized as the invasive plant species and are considered as the ecologically aggressive species (National Institute of Ecology 2019).

Global climate change is considered to be a principle reason for spread of invasive species (Dullinger et al. 2017). Over the last century, the world has experienced a temperature rise of 0.78 °C, and the temperature is expected to rise 2.6–4.8 °C by 2100 (IPCC 2013). In ROK, the rate of climate change has been predicted to be higher than the global climate change rates. The temperature increased 1.8 °C over the last 100 years, and the average temperature is expected to increase by 0.63 °C every 10 years until 2100 and by 5.7 °C by the end of this century (Ministry of Environment 2019). Thus, the potential for the spread of invasive plant species is thought to be high in ROK.

Ecological niche modeling is an empirical tool for simulating the spatial distributions of species, assessing the potential responses of organisms to climate change and resolving species niches based on environmental variables (Guillera-Arroita et al. 2015). Among the various ecological niche models (ENMs), the maximum entropy (MaxEnt) model is a widely used machine-learning technique that has high predictive accuracy while using a small set of data on species presence and environmental variables (Phillips et al. 2006; Koo et al. 2015; Dullinger et al. 2017; Lamsal et al. 2018; Thapa et al. 2018; Shrestha and Shrestha 2019).

Although many studies have addressed the ecology, distribution, and impact of climate change on invasive plant species distribution, most of studies to date have focused on either a handful of species or have been limited to a smaller geographical area of ROK (Cho and Lee 2015; Park and Kim 2015; Lee et al. 2016; Park et al. 2017; Ryu et al. 2017; Nam et al. 2018). Therefore, we predicted suitable habitat across the country for 11 invasive plant species, which are considered ecologically hazardous species in the ROK (Park and Kim 2015). Then, we calculated species richness and species expansion of invasive plant species in different provinces of the ROK to assess the potential for invasion in different provinces of the ROK and to develop a theoretical reference framework for management plans to combat the potential spread of invasive plant species.

Methods

Study area

This study was conducted in all provinces and islands of the ROK (Fig. 1), covering approximately 45.72% of the total landmass of the Korean Peninsula. Geographically, the terrain of the ROK is mostly mountainous in the northern and eastern parts of the country and contains lowland and flat plains in the southern and western regions. The climate of the ROK is categorized into warm-temperate, temperate, and cold-temperate. The southern coast and islands experience a warm-temperate climate, the central and northern parts of the ROK are temperate, and the high mountains have a cold-temperate climate. The southern region is relatively warm and wet, whereas the northern region is cold and continental. The annual precipitation (~ 1200 mm) is high in summer and low in winter, with snow deposition in the mountains. The vegetation of the ROK is categorized as deciduous broadleaf, temperate broadleaf, coniferous, subalpine, and alpine, and the overall biodiversity of the ROK includes 41,483 recorded species including 5308 plants, 1899 vertebrates, and 22,612 invertebrate (Ministry of Environment 2014).

Fig. 1
figure 1

Provinces of Republic of Korea. Each province is indicated by different color and the name of the provinces are given in legend

Full size image

Species presence data

The 11 most problematic invasive plant species in the ROK, based on their rapid expansion and adverse impacts on terrestrial ecosystems (Park and Kim 2015), were selected for distribution modeling (Table 1). Species occurrence data were obtained from field surveys and secondary sources (National Institute of Environmental Research 2013; National Institute of Ecology 2017). At least 10 presence points were used for each invasive plant species to obtain accurate model performance. Random points were chosen using a raster map of the ROK with ArcGIS 10.3 (ESRI, Redlands, CA, USA).

Table 1 List of invasive plant species used in species distribution models
Full size table

Environmental variables

Nineteen bioclimatic variables (Additional file 1: Table S1) were considered to be important for the distribution of invasive plant species. Monthly temperature (minimum and maximum) and precipitation data were obtained from the Korea Meteorological Administration for the estimation of current and future climates of the ROK. The HedGEM3-RA global circulation model was used to predict the future climate change scenarios RCP 4.5 and RCP 8.5 using Package Dismo in R (Robert et al. 2017). The current climate conditions were determined by averaging the data from 1950 to 2000. Similarly, the future climate conditions in 2050 and 2070 were estimated from the period of 2046 to 2055 and 2066 to 2075, respectively, similar to Adhikari et al. (2018). In all climatic data, the spatial resolution was 0.01° (36 s) and approximately 1 km2.

Model development, evaluation, and validation

The MaxEnt Package 1.3.3 for R (https://cran.r-project.org/web/packages/maxent) was used to predict the current and future distributions of invasive plant species in different provinces of the ROK similar to Bezeng et al. (2017), Lamsal et al. (2018), Manzoor et al. (2018), and Thapa et al. (2018). These measurements were achieved from the random split of the data into model calibration and model validation sets at the ratio of 75:25. Accuracy prediction and validation of each model were assessed based on the area under the receiver operator characteristic (ROC) curve (AUC) (Pearsons 2010) and the True Skill Statistic (TSS) (Allouche et al. 2006; Shabani et al. 2018) The AUC, ranging from 0 to 1, acts as a threshold-independent approach to distinguish presence from absence to evaluate model performance (Thuiller et al. 2005). Model performance was categorized as failed (0.5–0.6), poor (0.6–0.7), fair (0.7–0.8), good (0.8–0.9), and excellent (0.9–1), while validating the model results (Swets 1988). The TSS accounts for both specificity and sensitivity and ranges between − 1 and + 1, indicating an agreement no better than random and a perfect agreement between observation and prediction, respectively (Allouche et al. 2006; Lobo et al. 2008; Koo et al. 2015).

Suitable habitat, species richness, and species expansion

Area of suitable habitat for each invasive plant species under the climate change scenarios RCP 4.5 and RCP 8.5 was calculated for the years 2030, 2050, and 2080. The distribution maps of all invasive plant species (Additional file 2) were summed to obtain the current and potential species richness maps. A shapefile of the 17 provinces of the ROK was overlaid on the species richness maps, and average species richness and maximum species richness were extracted using the zonal statistics of the spatial analyst tool in ArcGIS 10.3 (ESRI, Redlands, CA, USA).

Invasive plant species expansion was determined by differentiating between current habitat and future habitat of each invasive plant species and reclassifying them based on the potentially new habitat expansion compared to the current distribution. The habitat expansions of all species were overlaid to produce the potential habitat expansion maps of invasive plant species in different provinces of the ROK. The whole process was performed with R software (https://www.r-project.org).

Results

Variable selection and contributions to models

We performed Pearson’s correlation analysis of the 19 bioclimatic variables (Additional file 1: Table S1) and selected six variables: annual mean temperature (Bio1), mean diurnal temperature range (Bio2), isothermality (Bio3), annual precipitation (Bio12), precipitation of the wettest month (Bio13), and precipitation of the driest month (Bio14). All of these variables are weakly correlated with each other, but they have strong correlations (r < 0.99) with other variables such as Bio4, Bio5, Bio6, and Bio19 (Additional file 1: Table S2). The threshold for weak correlation was r < 0.5. Thus, these six variables were used in the MaxEnt models for studying the distributions of invasive plant species.

The bioclimatic variable with the greatest model contribution differed among species (Table 2). Bio3 had the highest contributions (27.51–41.09%) to the modeling of six species (such as Ageratina altissima and Solanum carolinense), and Bio 14 contributed the highest proportion to the Ambrosia artemisiifolia (38.17%), Hypochaeris radicata (77.02%), and Rumex acetosella (24.41%). Similarly, Bio1 and Bio2 exhibited the highest contributions to the Paspalum distichum var. indutum (32.39%) and Paspalum distichum (30.36%), respectively. In this study, the temperature-related variable isothermality was a dominant driving factor for the ENMs of many invasive plant species.

Table 2 Contribution of bioclimatic variables to models
Full size table

Model evaluation and validation

Altogether, 11 invasive plant species were used in this study (Table 1), and independent ENMs were established for predicting current and future distributions of each species. We used both AUC and TSS values to evaluate model performance (Table 1). The mean AUC value of 11 invasive plant species was 0.808, ranging from 0.734 to 0.976, and the mean TSS value was 0.774, ranging between 0.641 and 0.892. Based on the evaluation, five species showed excellent (e.g., A. altissima), three species showed good (e.g., S. carolinense), and remaining three species showed fair (e.g., A. artemisiifolia) model performance.

Prediction of suitable habitats

The extents of climatically suitable habitats for the 11 invasive plant species were modeled to show the distribution of each species and the calculated area of the potential habitats under current and future climate change scenarios (Table 3). Under current climatic conditions, A. artemisiifolia had the highest suitable area (44,954 km2), while H. radicata had the least area of suitable habitat (3768 km2). Under future climatic conditions, Solidago altissima would have the highest suitable area, and P. distichum would have the least area of suitable habitat by 2030, 2050, and 2080 under the RCP 4.5 (Table 3). We calculated the area of new habitat expansion for each invasive plant species separately, which indicated that expansion of suitable area will be greatest by 2030 for Lactuca serriola, but by 2050 and 2080, P. distichum var. indutum will have greatest suitable area under the RCP 4.5 (Table 4). Although estimated rates and trends of expansion were not consistent, the suitable area for invasive plant species is predicted to expand continuously, except for the suitable area for P. distichum.

Table 3 Suitable habitat for invasive species (km2) under current and future climate conditions
Full size table
Table 4 Habitat expansion of invasive species (km2) under the future climate change scenarios
Full size table

Current species richness of invasive plant species

The potential species richness of invasive plant species under current climatic conditions is presented in Fig. 2 and Table 5. The provinces located in the western (Jeollabuk, Daejeon, Sejong, Chungcheongnam, and Gwangju) and southern (Jeollanam and Jeju) regions showed pronounced species richness, estimated to be average species richness 4.38–5.81 and maximum species richness 9–11. However, the provinces in the northeastern (Gangwon), southeastern (Gyeongsangnam, Ulsan, Busan), eastern (Gyeongsangbuk and Daegu), and central (Chungcheongbuk) regions have relatively low species richness, averaging 1.19–2.05, with maximum species richness predicted to range from 5 to 7.

Fig. 2
figure 2

Potential species richness of invasive plant species under the current climatic condition. The black lines on the map showing 17 provinces of Republic of Korea

Full size image
Table 5 Average and maximum invasive species richness in different provinces of Korea
Full size table

Future species richness of invasive plant species

The potential richness of invasive plant species in different provinces of the ROK was predicted to increase by 2030, 2050, and 2080 under the climate change scenarios RCP 4.5 and RCP 8.5 (Fig. 3 and Table 5). The estimated average and maximum species richness by 2030, 2050, and 2080 are highest in Seoul and lowest in Gangwon under the RCP 4.5.

Fig. 3
figure 3

Potential species richness of invasive plant species under the climate change scenarios RCP 4.5 and RCP 8.5 by the years 2030, 2050, and 2080. The black lines on the map showing 17 provinces of Republic of Korea

Full size image

The predicted rates of increasing average species richness are inconsistent among the provinces. The provinces in the northern (Seoul, Gyeonggi, Incheon), eastern, and southeastern regions are predicted to have relatively high rates of increase in average species richness in the future compared to the provinces in the western and southern regions (Jeollanam), where species richness is estimated to increase 17.8–292% by 2030, 25.75–390.75% by 2050, and 59.15–468.06% by 2080 under the RCP 4.5 compared with the current richness.

Maximum species richness is not predicted to increase in as large as a proportion similar to average species richness in all provinces. The provinces of Seoul, Incheon, Gyeonggi, Chungcheongbuk, Daegu, Gyeongsangbuk, Gangwon, Gyeongsangnam, Busan, and Ulsan have predicted maximum species richness values of 7–9 by 2030, 9–11 by 2050, and 10–11 by 2080 under the RCP 4.5 (Table 5). The other provinces show very similar estimations of maximum species richness.

Invasive plant species expansion

The invasive plant species expansion in different provinces of the ROK is shown in Fig. 4 and Table 6. The average species expansion is predicted to be highest by 2030, 2050, and 2080 in Seoul under RCP 4.5. Under the extreme climate change scenario (RCP 8.5), the maximum species expansion would be highest in Seoul by 2030 (11 species), Ulsan and Busan by 2050 (11 species), and Gangwon, Gyeonggi and Gyeongsangbuk by 2080 (11 species). These results indicate that the provinces in the northern and eastern regions of the ROK will have the highest species richness of invasive plant species in the future.

Fig. 4
figure 4

Potential species expansion of invasive plant species under the climate change scenarios RCP 4.5 and RCP 8.5 by the years 2030, 2050, and 2080. The black lines on the map showing 17 provinces of Republic of Korea

Full size image
Table 6 Average and maximum invasive species expansion in different provinces of Korea
Full size table

Discussion

This study estimated current potential and future predicted habitat of the most troublesome invasive plant species in the ROK (11 species) in different provinces. The performance of the ENMs should not be altered by prevalence (data size), but there should be an integration of both omission and commission errors (Allouche et al. 2006). The AUC value is independent of the prevalence. However, this index is sometimes criticized because it equally weights the commission and omission errors and avoids real prediction (Lobo et al. 2008). Therefore, we applied both AUC and TSS values to evaluate the model performance. The mean value of AUC shows that the model performance was very good, demonstrating good prediction results (Swets 1988). Furthermore, the mean value of TSS reveals that there was a good agreement between observations and predictions of the MaxEnt model (Allouche et al. 2006).

Climate change facilitates the introduction, dispersion, and naturalization of invasive and alien species and also reduces the resilience of aquatic and wild ecosystems to invasive species (Hellmann et al. 2008). Our model predicted that all the invasive plant species will retain their current ecological niches while adding areas of climatically suitable habitats in the ROK in the future, as concluded by Lamsal et al. (2018) and Shrestha and Shrestha (2019). However, the rates and extents of the climatically suitable habitats were not predicted to be consistent among all species. In this study, A. altissima, H. radicata, P. distichum var. indutum, S. altissima, and L. serriola are predicted to show unusual rates of habitat expansion (Additional file 2: Figure S3, Figure S5, Figure S9-S11), estimated at 156.36–659.15% by 2030, 183.75–716.79% by 2050, and 360.23–968.78% by 2080 relative to the current climatically suitable habitat similar to Cho and Lee (2015) and Nam et al. (2018). The change in the suitability of habitats for invasive plant species depends not only on the climatic variables related to precipitation and temperature used in the MaxEnt model but also on different kinds of non-climatic factors, such as land topography, altitude, and latitude, and on each individual plant’s physiological and morphological advancement and ability to cope with an adverse climate (Lamsal et al. 2018). Therefore, the climatically suitable habitat for each species could be different under the same climatic conditions in the future. Climate change may limit the invasiveness of some species, creating opportunities for ecosystem restoration (Manzoor et al. 2018). We predict a decrease in the climatically suitable habitat for P. distichum in the future (Additional file 2: Figure S5), and no more suitable habitat expansion will occur by 2080 under RCP 4.5 (Table 4). In this study, three temperature-related variables (Bio1, Bio2, and Bio3) have found high contribution in its model that suggests increase in temperature is unfavorable to its life cycle. Lee et al. (2016) mentioned that the temperature above 20 °C is harmful for the germination of P. distichum and concluded that the temperature-related variables such as annual mean temperature are critical factors in determining its distribution.

Under the current climate, the average and maximum species richness was estimated to be relatively high in Jeollanam, Jeollabuk, Chungcheongnam, Daejeon, and Gwangju compared to other provinces. Currently, 7–8 invasive plant species, e.g., A. artemisiifolia, P. distichum, P. distichum var. indutum, and R. acetosella are densely distributed in such provinces. However, the other provinces adjoining the sea, e.g., Incheon, Seoul, Gyeongsangnam, Gyeongsangbuk, Ulsan, Busan, and Gangwon, are predicted to be invaded by 5–7 species. In these provinces, invasion is predicted to be concentrated on islands and in coastal areas of the mainland. This result indicates that introduction and dispersion of invasive plant species is relatively high in the ROK from the western and southern coast. Invasion typically occurs during foreign trade and tourism, and in particular, invasive species enter through ports (Hellmann et al. 2008). In the ROK, altogether, 18 international ports of different sizes, e.g., Incheon, Pyeongtaek, Gunsan, Masan, Yeosu, Mokpo, Busan, and Ulsan, exist in various provinces for foreign trade. Invasive species may be introduced to the ROK through these ports either purposely or inadvertently and colonize wild and urban ecosystems of the adjacent provinces (Early et al. 2016), then dispersed across the ROK through various modes of dispersion and human transportation systems, such as road and railway connections, as suggested by Eminniyaz et al. (2013).

Globally, climate change induces habitat expansion of invasive species northward (Petitpierre et al. 2016). In this study, we also predicted range expansion of invasive plant species in the northern and northeastern regions, particularly in Chungcheongbuk, Gyeonggi, Seoul, Incheon, Gyeongsangbuk, and Gangwon and in Jeollabuk, Jeollanam, Chungcheongnam, Gyeongsangnam, Busan, and Ulsan. In Jeju, we estimated habitat expansion of invasive plant species at high elevation near Halla Mountain (1950 m) from the coastal region. These results show that climate change causes expansion of habitat of invasive plant species towards high latitudes (> 36° N) and high elevation (not calculated in scale) in the ROK by removing current climatic barriers and shifting habitats to higher latitudes and elevations (Hellmann et al. 2008). Consistent with our findings, many studies have predicted the expansion of distribution ranges of invasive plants species towards higher latitudes and elevations in response to climate change (Loarie et al. 2008; Lenoir et al. 2008; Bradley et al. 2012; Shrestha and Shrestha 2019).

All the studied invasive plant species were introduced to the ROK either intentionally or unintentionally (Table 1). The S. altissima is native to North America and was introduced to the ROK before 1980 for nectar purpose (Jung et al. 2017). It is distributed across the ROK and is invading the forests. The other plant species, A. artemisiifolia, Ageratina altissima, P. distichum, R. acetosella, and L. serriola were accidently introduced to the ROK, as recorded between 1932 and 1994 (National Institute of Ecology 2019). Although some invasive plant species such as P. distichum could be food resource to wild herbivores, e.g., roe deer (Adhikari et al. 2016), they have poor nutrition value. Most of the invasive plant species originated from the American continents, followed by Eurasia, Europe, and Africa, and may have been introduced to the ROK via trade, tourism, and tidal activity of the sea. The invasive plant species native to tropical countries have much higher critical thermal maxima than do native species, indicating that they can thrive in higher temperatures and could be dominant to native species under climate change conditions. Additionally, invasive species have some important characteristics, such as short reproduction times, high fecundity, strong dispersal ability, and high environmental tolerance, that help them establish well in new environments (Bradley et al. 2010). Along with climate change, several non-climatic components, including anthropogenic activities such as rapid land use and land cover change, road construction, urban development, foreign trade, agriculture, and tourism, promote invasion (Martin et al. 2013).

Although this study provided critical information about potential plant invasions in different provinces of the ROK under current and future climatic conditions, our models were based only on bioclimatic variables, disregarding land use and land cover change, dispersal capacities, biotic interaction (e.g., facilitation and competition), and vectors driving species invasions, as described by Pysek and Richardson (2010) and Martin et al. (2013). This study is a part of ongoing research; we would consider using other variables, such as land use and land cover change, roads, foreign trade data involving dispersion capacities, and biotic interactions, to obtain more accurate prediction in the near future.

Conclusions

In this study, we predicted the current and future potential habitats for invasive plant species for which additional suitable areas are likely to be created in the different provinces due to climate change in the future. The continuous increase in the climatically suitable habitats of invasive plant species has already caused adverse impacts on biodiversity, ecosystem services, food security, and livelihood in the ROK. The results of this study can be used as a precautionary note and might be helpful to develop policy in each province for the cost-effective management of invasive plant species. Additionally, our study will support the prevention and early detection of invasive plant species in their potentially suitable niches. Based on our study, we urge that it is necessary to perform early identification and eradication actions in all provinces, especially focusing on the western, southern, southeastern, and northern regions of ROK, including Jeollanam, Jeollabuk, Chungcheongnam, Daejeon, Gwangju, Gyeonggi, Seoul, and Incheon.