Abstract
The nutria, (Myocastor coypus), is a semiaquatic rodent native to the subtropical and temperate regions of South America. The species was introduced to South Korea for meat and fur production purposes and a wild population has become established. The species subsequently invaded aquatic ecosystems and destroyed aquatic vegetation and cultivated crops. Thus, it is essential to understand their current distribution and future range expansion for effective control and eradication strategies to reduce the risk of colonization into new regions. In this study, we used niche modeling procedure to identify potentially suitable habitats for M. coypus under current and future predicted climate change using the maximum entropy algorithm. We found that the main habitat area of M. coypus is expected to expand under a warming climate from ~ 4069 km2 in the southern and southeastern regions of South Korea, to the northern border of the country, with estimated ranges of 21,744 km2, 55,859 km2, and 64,937 km2 by 2030, 2050, and 2070, respectively. The findings of the present study assist in identifying the future distribution and potential dispersion routes of M. coypus in South Korea, which is important for informing the government regarding essential management actions plans at regional and local scales.
Similar content being viewed by others
Introduction
Myocastor coypus is a large, semiaquatic rodent indigenous to the subtropical and temperate regions of South America1 and is currently included on the International Union for Conservation of Nature (IUCN) list of the 100 worst invasive global species2. M. coypus was introduced to North America, Europe, Asia, and Africa for meat and fur production3,4,5; however, because of the decreasing demand and the consequential reduced market price, individuals were either released or escaped from farmhouses, and wild populations have since established in natural environments5,6. Subsequently, M. coypus have invaded aquatic ecosystems and caused large economic losses by destroying cultivated crops, aquatic vegetation, and tree species because of their aggressive foraging habits, in addition to burrowing-related damage to river banks, dykes, and irrigation canals5,6,7. Additionally, M. coypus causes various zoonotic diseases such as leptospirosis, trichinosis, and toxoplasmosis6,8,9.
In South Korea, M. coypus were first introduced from France in 1985 for breeding purposes; however, immature breeding techniques and poor maintenance resulted in mortality of all the imported individuals. M. coypus were again imported from Bulgaria in 1987, and breeding was successful with production peaking in 20013,7,10. Since approximately 1999, M. coypus have been escaping from their breeding environments and have settled as wild populations in the Nakdong River basin and across water bodies in the southeastern regions of Chungcheongbuk and Jeju Provinces3,7,10. By 2012, their presence was recorded in 13 administrative districts (Ads) in South Korea7. M. coypus pose several threats to aquatic ecosystems and agriculture in South Korea, and destruction of aquatic plant habitat, consumption of endangered aquatic plants, and damaging cultivated crops adjacent to their habitats have been well documented; however, there has been no record of serious threats to irrigation canals or other physical structures in South Korea3,7,10.
Controlling M. coypus populations and inhibiting their range expansion are difficult because of their high fecundity, with ~ 2–3 litters per year and litter sizes of 1–12 individuals11. For instance; lethal control via systematic trapping and hunting has been successful at the local level and has been adopted in several countries in the European Union (EU), United States (US), and South Korea6,7,12. The South Korean government planned to capture M. coypus in response to public requests to protect agricultural crops and aquatic ecosystems, and a pilot project was conducted in 2013 which successfully captured 3349 individuals7. Subsequently, the Ministry of Environment established a 5-year Nutria Eradication Project in 2014, which was operated by a common cooperative system involving 18 organizations including the National Institute of Ecology, local governments, and local environmental agencies. The project facilitates public participation by providing incentives to the public when captured M. coypus are shown to local governments. The project had total operating funds of $9.5 million USD7. In total, 27,487 M. coypus individuals were captured between 2014 and 20187. Accordingly, the distribution and habitat tracks (i.e., species occurrence within a 0.01 km2 area) of M. coypus within South Korea were investigated from 2014–2018 to provide baseline data for informing eradication projects.
The threats posed by M. coypus can be exacerbated by changing climates and the correlated expansion of suitable habitat, which enhances reproduction and survival success13,14. Extreme cold winter temperatures have been suggested to limit M. coypus distribution to areas with ≤ 80 annual freezing days per year15; thus, increasing global temperatures owing to climate change could favor M. coypus habitat expansion14. In South Korea, the mean temperature has increased by 1.8 °C over the last 100 years, and is predicted to increase by an additional 0.63 °C every decade until 2100, with the increase reaching > 4.7 °C by 210016 and potentially creating a highly favorable environment for the spread of invasive species in South Korea.
Successful management of M. coypus requires an adequate understanding of their ecology and distribution13. In particular, assessing M. coypus distribution under current and projected climate and land-cover change scenarios can help prioritize management and monitoring regions for early detection, in addition to informing the preventative measures adopted by governments and conservationists. Therefore, it is essential to develop reliable methods for predicting future M. coypus invasion so that inspection or management policies can be adjusted to decrease the risk of establishment, propagation, and expansion into new areas.
Species distribution models (SDMs) are an empirical tool used in ecological studies to estimate the potential spatial distributions of species in response to climate change, and determine the ecological niches of species based on records of species occurrences and environmental variables17. Among the various available SDM algorithms, maximum entropy (MaxEnt) is a machine-learning technique with high predictive performance based on a small subset of species presence data18. Although various studies have examined M. coypus in South Korea, to date most have focused on the introduction, ecology, distribution, diet, and adverse impacts of M. coypus on aquatic ecosystems and agriculture3,7,10,19,20,21. To the best of our knowledge, a detailed investigation of their distribution considering bioclimatic factors, land cover, altitude, and distance from water has not previously been conducted. Accordingly, the main objectives of the present study were to: (1) Conduct an extensive survey of M. coypus across South Korea; (2) analyze the associations between bioclimatic, land cover, geographic, and water proximity variables and current M. coypus distributions; and (3) model the potential M. coypus habitat under current and future climate and land-cover change scenarios in South Korea. Based on this study, high risk locations for M. coypus invasion were identified. The results can assist with understanding the distribution and potential future dispersion of M. coypus in South Korea, and support the development of a theoretical framework for management policies and strategies to control their potential spread.
Results
Species occurrence survey
A species occurrence survey was conducted to support the “M. coypus Eradication Project” established by the Ministry of Environment, South Korea, from March 2014 to December 2018. Based on the field survey and habitat track study, the highest number of M. coypus was recorded in 2014 (3,805 counts) and successively declined thereafter; i.e., to 1,451 in 2015, 502 in 2016, 235 in 2017, and 187 in 2018 (Fig. 1). In 2014, M. coypus were detected in 19 local administrative districts (LADs) across six ADs: Chungcheongbuk (2 LADs), Jeju (2 LADs), and local ADs of Gyeongsangbuk, Daegue, Gyeongsangnam, and Busan (15 LADs). The main habitat was in the midstream and downstream regions of the Nakdong River basin, between Busan and Daegu. The numbers of habitat track were detected in 15 LADs between 2015 and 2017, and in 14 LADs in 2018. These results showed an estimated 54% reduction in M. coypus habitat tracks between 2016 and 2017, demonstrating progress in eradication following implementation of the M. coypus eradication project.
Model performance and variable importance
Pearson’s correlation test was performed, and based on the weak correlations (r < 0.70), five bioclimatic variables (e.g., annual mean temperature [Bio 1], maximum temperature of the warmest month [Bio 5], mean temperature of the wettest quarter [Bio 8], precipitation seasonality [Bio 15], and precipitation of the wettest quarter [Bio 16]) and three environmental variables (i.e., altitude, distance from water, and land-cover change; Table S1) were selected and used for MaxEnt modeling. A jackknife test was performed to confirm the contribution of the included variables in the model (Fig. S1) and the results showed that the temperature-related variables Bio 5, Bio 8, and Bio 1 were the strongest bioclimatic predictors of the M. coypus distribution, explaining 29.23%, 21.05%, and 15.01% of the variance, respectively (Table 1). Similarly, both altitude and distance from water played relatively important roles, explaining 14.3% and 8.25% of the observed variance, respectively. Other variables had markedly lower values, thus appearing to play more minor roles in the M. coypus distribution model. The area under the curve (AUC) values of the (training 0.968 and test 0.954) indicated a high model performance for the prediction of the M. coypus distribution. Similarly, the true skill statistic (TSS) values (0.910) showed strong agreement between the observations and modeled predictions.
Predictions of current and future distributions of M. coypus
The extent of climatically suitable M. coypus habitats was modeled to show the distribution and the potential habitat area across all South Korean ADs under current and future climate change scenarios RCP 4.5, and RCP 8.522 in 2030, 2050, and 2070 (Table 2). The current potential M. coypus habitat was identified in 12 provinces (Fig. 2A), with a total area of ~ 4,069 km2, equivalent to 4.26% of the total country land area. The seven provinces concentrated along the southern and southeastern regions of the country, i.e., Gyeongsangnam (2,211 km2), Jeollanam (506 km2), Gyeongsangbuk (498 km2), Busan (188 km2), Daegu (213 km2), Chungcheongbuk (160 km2), and Jeollabuk (142 km2), were predicted to be highly suitable for M. coypus habitation (Table 2). Suitability maps are independent of thresholds and commonly represent the probability of focal species to occur in a given geographical area. Therefore, we predicted the probability distribution maps of M. coypus and presented the current suitable habitats in Fig. 3A. Under the current climate, both binary distribution maps and suitability maps showed consistent predictions for M. Coypus habitation. We calculated the mean and standard deviation of suitability values in each AD, indicating that the maximum mean suitability value under the current climate was in Busan (0.23, standard deviation = 0.40) (Table 3).
Under climate change projections, the M. coypus habitat range was predicted to expand northward along the western and eastern coasts of the Korean Peninsula, spreading to all provinces by 2050 (Figs. 2B,C, 3B,C). The future suitable habitat areas were estimated to be 11,335 km2, 35,986 km2, and 50,199 km2 by 2030, 2050, and 2070, respectively, under RCP 4.5, and 21,744 km2, 55,859 km2, and 64,937 km2, respectively, under RCP 8.5 (Fig. 4A). Similarly, the mean suitability between the RCP 4.5 and RCP 8.5 for the years 2030, 2050, and 2070 was highest under RCP 8.5, suggesting that a warming climate will favor M. coypus habitation (Fig. 4B). Although the rates and trends of habitat expansion were inconsistent, suitable habitats were expected to increase across all provinces (except the cities of Ulsan and Busan) by 2030 under RCP 4.5. The areas with the highest suitability were predicted to be in Jeollanam, Gyeongsangnam, Gyeongsangbuk, Chungcheongnam, and Gyeonggi by 2030, 2050, and 2070 under both examined climate change scenarios (Table 2). However, the mean suitability value was estimated to be the highest in Gwangju City where it was predicted to range from P = 0.70–1.00 between 2030 and 2070 (Table 3). The ADs located on the west coast (i.e., Jeollabuk, Chungcheongnam, and Gyeonggi) maintained higher rates of habitat expansion compared to those in the central region and along the east coast. Gangwon Province (northeastern region) and Jeju (southern region) had comparatively small areas of suitable habitat. In addition, high mountain regions (e.g., the Baekdu‒daegan mountain range and Hall Mountain of Jeju [elevation, 1950 m]) also exhibited low habitat suitability (Fig. 3).
Vulnerability estimations of different administrative divisions
We performed vulnerability estimation of different ADs based on the mean habitat suitability values of M. coypus in different provinces, which revealed that all the ADs were considered to have low vulnerability. However, in the future, vulnerability will be extended from the southern region to central and northern regions and high vulnerability with be exhibited in two ADs (e.g., Daejeon and Gwangju) by 2030, eight ADs (e.g., Seoul, Incheon, and Daejeon) by 2050, and 11 ADs (e.g., Seoul, Incheon, and Gyeonggi) by 2070 (Table 4).
Discussion
A 5-year field survey was performed to investigate the M. coypus distribution in South Korea, including identifying their presence in regions of the Nakdong River basin, Han River, and Jeju Island. The majority of habitat tracks were observed in the midstream and downstream regions of the Nakdong River basin, thus identifying this basin as the area that was the most populated by M. coypus. The midstream and downstream reaches of the Nakdong River are characterized by relatively slow water speeds upstream of several tributaries and wetlands10, warm winter temperatures, and abundant food sources, creating appropriate habitats for M. coypus growth and establishment3. Although it was observed in this study that the M. coypus habitat distribution was reduced to 14 LADs along the Nakdong River in 2018, their presence has continued in this area despite several management efforts, including the M. coypus Eradication Project7. However, there were no recorded sightings in either Jeju or Chungcheongbuk Province after 2014, implying significant population declines following the implementation of the eradication project. Globally, temperature is the most importnat factor for biological invasions14. In particular, M. coypus are warm-adapted species that are vulnerable to colder weather; therefore, they tend to propagate rapidly in areas with warm winters23. M. coypus populations can decline owing to adverse effects in natural environments with temperatures ranging between 0–5 °C14. In extreme cold, they fail to reproduce and often show high mortality (≤ 71%) when sheet ice forms on water bodies1,24,25. The Nakdong River, located in the southeastern region of South Korea, maintains relatively warm temperatures year-round, and has notably warm winters compared to the rest of the country; whereas, the other major rivers, such as the Han and Geum Rivers, are located in cold-temperature and temperate regions, respectively, with recorded winter temperatures < − 26.2 °C (KMA 2021, unpublished data). Although Hilts et al.14 suggested that M. coypus can persist where there are ≤ 80 annual freezing days, their presence was notably undocumented along the Han and Geum Rivers. The extreme cold winter water temperatures could explain the absence of M. coypus in these locations; however, these areas could potentially become suitable habitats under increasing temperatures in future3,26.
As previously stated, the Nakdong River is currently the most suitable habitat for M. coypus in South Korea. Increasing temperatures facilitate invasive species growth, range expansion, and naturalization27, while simultaneously reducing the resilience of ecosystems to exotic species14. The model used here predicted that M. coypus will retain their current distribution and that additional habitat expansion will occur in the future, which is supported by the findings of Hilts et al.14 and Pereira et al.28. Climatic factors generally have a dominant influence on species distribution at broad spatial scales29. The findings in the present study show that temperature-related variables are important determinants of climate-based M. coypus habitats (similar to the findings of Schertler et al.13. The results also predict habitat expansion from the southeastern region (particularly Gyeongsangnam, Jeollanam, Gyeongsangbuk, Busan City, and Daegu City) toward the central and northern regions of the country, which, in addition to Jeollabuk, Chungcheongnam, Chungcheongbuk, Gyeonggi, and Seoul City, were predicted to become highly vulnerable in future decades. These results support previous findings that show that increasing temperatures will expand the invasion risk northwards30, in addition to earlier projections indicating habitat expansion of invasive species with increasing temperatures13,14,31. Future expansion of the suitable M. coypus habitat is likely to occur gradually from south to north through the water channel connections of the five major rivers in South Korea. Northward expansion is expected to occur mainly from the west coast. The terrain of South Korea is characteristically low along the west coast and high along the east coast31. Consequently, most major rivers flow westward into the Yellow Sea and are relatively long, maintaining smooth slopes and wide basins. In contrast, the rivers flowing into the East Sea are relatively short and steeply sloped. The geographic terrain and river systems along the west coast, combined with increasing future temperatures, could favor the expansion of M. coypus habitat, supporting the future establishment of M. coypus populations. The derived model used in this study did not predict M. coypus habitats in high mountains or on sharp elevational gradients, such as the Baekdu‒daegan mountain range (701 km), Mt. Halla on Jeju Island (1,950 m), the northern part of Gyeongsangbuk Province, and most of Gangwon Province (as similarly shown by Sheffels23 and Pereira et al.28). These areas could thus serve as a geographical barrier, blocking any future dispersion. In addition, high-elevational areas experience extremely cold weather in winter, maintaining snowpack for long time periods, thus creating inhospitable conditions for M. coypus15.
In addition to bioclimatic variables related to local-scale features, habitat covariates, such as a high proportion of freshwater and forested shrub wetlands (i.e., marshlands or swamps) close to other wetlands, appeared likely to support increasing M. coypus populations14. Freshwater swamps are relatively productive systems that enhance high floral diversity, including densely layered understories of herbaceous plants, shrubs, young trees, and overstory trees32, possibly generating foraging habitat and thermal protection during extremely cold winters. M. coypus tend to prefer various aquatic plants (e.g., Paspalum disthichum, Panicum tricholaenoides, Lemna minor, Spirodela polyrrhiza, Carex spp., and Schoenoplectus spp.), depending on their habitat and the season33,34. In South Korea, Hong et al.19 reported M. coypus diets consisting of similar aquatic plants in the families of Poaceae, Cyperaceae, and Salviniacease, with some terrestrial plants belonging to the families of Chenopodiaceae, Asteraceae, and Amaranthaceae. Freshwater swamps, particularly those connected to groundwater, freshwater rivers, and lakes, could be appropriate M. coypus habitats considering the preference of this species for foraging near the edges of slow-speed water bodies, potentially maintaining connectivity between rivers and wetlands, facilitating dispersion and range expansion10,21,35. M. coypus can travel ≤ 5–6 km overnight36, and Aliev37 reported M. coypus dispersals of ≤ 120 km in eastern Europe within 2 years. Similarly, Hong et al.3 reported that the dispersal distance of M. coypus increased by 42.6 km between 2007 and 2008 in South Korea. Human activities (e.g., landscape fragmentation) are additional important factors affecting M. coypus occurrence7. Similarly, human disturbances (e.g., urban population density and distance to human settlements) are thought to be additional important factors affecting M. coypus distribution7,38. These results highlight the importance of local environmental factors in addition to bioclimatic variables for M. coypus dispersal and range expansion.
M. coypus can cause large-scale damage to aquatic ecosystems, with significant effects on crops and riverine vegetation39, potentially resulting in shortages of food resources for other wild herbivores inhabiting areas adjacent to M. coypus habitat (e.g., roe deer)40. M. coypus can weaken dams and irrigation systems by burrowing on riverbanks, thereby creating direct economic losses, as well as indirect losses connected to floods. Accordingly, the Ministry of Environment in South Korea conducted the M. coypus Eradication Project from 2014–2018, which was considered to have successfully reduced the range of M. coypus habitats in Jeju and Chungcheongbuk Provinces; however, major habitats persisted in the Nakdong River.
Based on the results of the present study, future suitable M. coypus habitats are likely to expand from the southeastern region to the northern international border. Currently, all the ADs are under the low vulnerable; however, by 2070, 11 ADs including Gwangju, Jeollabuk, Busan, Daegu, and Seoul, will be considered to have high vulnerability. The continuous M. coypus habitat expansion has already had adverse impacts on aquatic ecosystems and agriculture within South Korea, and the SDM-based prediction of M. coypus distribution provides a valuable tool for targeting areas with the highest risk of invasion. The results can assist with understanding M. coypus distribution and dispersion in South Korea, in addition to informing government and conservation agencies regarding management action plans and policies aimed at controlling or eradicating M. coypus at regional and local scales.
Methods
Study area
The study area included the mainland and all the islands of South Korea, with a total land mass of 100,033 km2 (Fig. 5). The country is divided into 17 ADs (provinces and metro cities, referred to here as cities), which were utilized to assist in delineating M. coypus distribution localities and informing provincial governments when adopting future management actions. Geographically, the terrain of South Korea is mostly mountainous, dominating the northern and eastern parts of the country, whereas lowlands and flat plains occupy the southern and western regions. The Baekdu-daegan mountain range spans ~ 701 km, starting from Hyanro-bong in Goseong County of Gangwon Province and ending at Cheonwang-bong on Mt. Jiri in Jeollanam Province. The climate of South Korea is readily divisible into warm-temperate (southern coast and islands), temperate (northern and central), and cold-temperate (high mountains)41. Mean winter temperatures range between − 6 and 3 °C, whereas mean summer temperatures range from 23–26 °C16. Annual precipitation is between 1200–1500 mm, with a relatively higher rate of precipitation in the southern regions (including Jeju Island) compared to the central and northern regions16. The forests can be roughly categorized as deciduous broadleaf, temperate broadleaf, coniferous, subalpine, and alpine, with overall total species diversity of 41,483, including 5308 plant species, 1899 vertebrate species, and 22,612 invertebrate species (including 15,651 insect species)42.
M. coypus survey and species occurrence points
A pilot survey was conducted to investigate the spatial distribution of M. coypus in South Korea based on earlier reports from the South Korean Ministry of Environment43 and previous research10. The survey revealed the presence of M. coypus across all areas of the Nakdong River basin and its drainage system, in the surrounding lakes, in the downstream area of the Namhan River, and in the Jeju Island wetlands. Detailed surveys were then performed in the Nakdong River basin and its tributaries (127° 29ʹ–129° 19ʹ E and 35° 03ʹ–37° 13ʹ N), the Namhan River and Chungjuho artificial lake located downstream (127° 50ʹ–128° 20ʹ E and 36° 50ʹ–37° 05ʹ N), and across all the Jeju Island wetlands (126° 08ʹ–126° 58ʹ E and 33° 06ʹ–34° 00ʹN) (Fig. 5). Field surveys were performed from March 2014 to December 2018.
During the surveys, all surveyors traveled on foot across river basins with an average walking speed of < 2 km·h−1 to monitor for footprints, living dens, feces, or make direct observations of individuals. Camera traps and M. coypus catching traps were installed to further confirm habitat and signs of presence. Geographic coordinates were recorded for all observations using a handheld geographical positioning system (GPS V; Garmin Inc., KS, USA). Habitat tracks were determined using the coordinate information within an area of 0.01 km2 in ArcGIS 10.3 (Esri, CA, USA)7. GPS coordinate information for some representative photographs from the M. coypus survey can be seen in Fig. S2. The species occurrence points for each year are shown in Fig. 1.
Environmental variables
Nineteen bioclimatic variables44, in addition to distance from water, altitude, and land-cover change (Table S2), were considered important for predicting the M. coypus distribution in South Korea. Climatic data, including precipitation and the monthly minimum or maximum temperatures, were obtained from the Korea Meteorological Administration (KMA; https://www.kma.go.kr/eng) to estimate the current climate and future climate change scenarios of South Korea. Two greenhouse gas emission scenarios were selected for analysis, i.e., RCP 4.5 and RCP 8.5, for 2030, 2050, and 2070. The moderate and maximum emission scenarios (RCP 4.5 and RCP 8.5, respectively) correlate to predicted global mean surface temperature increases of 1.4–1.8 °C and 2.0–3.7 °C by 2100, respectively. Among the numerous available global circulation models, the HadGEM3-RA is a regional atmospheric model produced by the Met Office Hadley Centre (https://www.metoffice.gov.uk/), based on the atmospheric component of the current Earth System Model45. The KMA used HadGEM3-RA45 to create a national climate change scenario for South Korea, revealing that it tended to model small-scale features more accurately than other global circulation models, such as HADGEM2-AO, which contains more complicated topographies, lengthy and uneven coastlines, and thousands of islands along the Korean Peninsula. Accordingly, the HadGEM3-RA global circulation model was used here to determine the impacts of the RCP 4.5 and RCP 8.5 scenarios, according to the R Package Dismo v. 1.346. Similarly to Jeon et al.47, the current climate was determined by averaging climatic data recorded between 1950 and 2000, and future climates for 2030, 2050, and 2070 were estimated from predictions for 2026‒2035, 2046‒2055, and 2066‒2075, respectively. We used a threshold of 0.7 for Pearson’s correlation coefficient48 for selecting bioclimatic and environmental (as described in Shin et al. 49 and Adhikari et al.50). Seven landcover categories (i.e., urban, cropland, forest, grassland, wetland, barren, and water) were used in the model developed by Song et al.51, based on shared socioeconomic pathways using a scenario generator in which the future population and predicted urban area were adopted for a transition matrix containing the land-cover change trends of each class. The current and future land-cover change scenarios (shared socioeconomic pathways 1) developed by Song et al.51 were obtained from the Korea Environment Institute MOTIVE (www.motive.kei.re.kr). Changes in the altitudinal range were quantified in the M. coypus distribution according to a digital elevation model. The variable distance from water included the distance from all kinds of rivers, lakes, and ponds, which were determined using the Euclidean distance option of the Spatial Analyst tool in ArcGIS Desktop 10.8. All the bioclimatic and environmental variables used in this study have a spatial resolution 0.01° (36 s), or ~ 1 km2, as explained by Fick and Hijmans52.
Model development
SDM is a method used for predicting potential distribution of species throughout global space and time by using a correlation between the species’ geographic occurrence and its surrounding environment53. SDMs typically estimate the ecological niche by statistically relating environmental variables directly to species occurrence. Among the various SDM algorithms, the MaxEnt is a widely used machine learning technique for habitat suitability, which has high predictive performance using only presence data18. In this study, we performed MaxEnt modeling using “Biomod2” Package v.3.5.1, selecting single model MAXENT.Phillips.254to predict the current and future M. coypus distribution in South Korea. MaxEnt exhibits a high predictive performance using presence-only occurrence data18,55 and is often the best option for invasive species because presence–absence survey data are usually not available, as was the case in the present study. As the MaxEnt model required background point data (e.g., pseudo-absence), we determined the background points of the study area using ArcGIS Desktop 10.8 (as suggested by Barbet-Massin et al.56). Multiple species presence points within the same ~ 1 km2 grid were eliminated, leaving only a single unique point per grid by adopting a spatially rarefy occurrence data tool in the ArcGIS SDM toolbox v.2.457, to prevent overfitting and incorrect inflation of model performance owing to spatial autocorrelation58. The total species occurrence points for M. coypus were reduced to 2535 from 6180, and those points were used in the MaxEnt modeling. The five bioclimatic variables, distance from water, altitude, land-cover change, and the species occurrence points were used as the inputs for MaxEnt.
During modeling, the species occurrence data were randomly split into two parts, i.e., a training set and a test set, at a ratio of 3:1(as described in Adhikari et al.59) . The training data set was used for model fitting, which estimated the parameters of the model (model calibration). The test data set was used for model assessment, which estimated the performance of the estimated model (model validation). The other model options were run with default settings and 100 iterations, and cross-validation was maintained in the replicate runs to guarantee model accuracy60.
Model evaluation and validation
The goodness-of-fit of the model was assessed, and a validation was performed with the AUC values of the receiver operating characteristic curves61 and the TSS62. The training and test data sets were used to compute the AUC values. The AUC value measures the model capacity in a model that discriminates the observed occurrences from the background data using the training and test datasets63. AUC is independent of dataset size (prevalence) but is limited because it weights both commission and omission errors equally, avoids the actual probability values, and is dependent on geographical extent64. In particular, when expanding the geographical extent beyond the present range, the AUC values increase; thus, TSS was used as an alternative criterion for model performance validation. TSS corresponds to the modeled sum of sensitivity (omission error), which is the proportion of observed presences correctly predicted, and specificity (commission error), which is the proportion of observed absences correctly predicted61. AUC values vary from 0‒1, whereby higher values suggest model superiority. Here, model performance was classified as poor (0.6–0.7), fair (0.7–0.8), good (0.8–0.9), or excellent (0.91–1), as suggested by Swets65. TSS is threshold-dependent and determines model performance by examining classification accuracy after selecting a threshold value. TSS values range between − 1 and + 1, where low values indicate an agreement that is no better than random chance, and high values represent the perfect alignment of observations and predictions65. In addition, a jackknife test was performed to evaluate the importance of each variable with respect to model performance18,66. The database and a brief conceptual flowchart of the methodology are presented in Fig. 6.
Prediction of potential habitat and new habitat expansion and estimation of vulnerability
The MaxEnt model yielded both binary distribution maps and suitability maps of M. coypus under current and future climate change scenarios RCP 4.5 and RCP 8.5 for 2030, 2050, and 2070. The binary distribution maps were used to determine the suitable habitat areas across the 17 ADs of South Korea, and suitability maps were used to determine the mean and standard deviation of suitability values for each province. To calculate the distribution area, mean suitability values, and standard deviation of suitability values for each AD, the shape file representing 17 provinces of South Korea was overlain on the species distribution maps separately, and an extraction of the multiple values to points was performed using the zonal statistics of the spatial analyst tool in ArcGIS Desktop 10.831,67. The vulnerability of each AD was determined and classified into three categories (low, moderate, or high) based on the mean suitability values from 0‒0.33, 0.34‒0.66, and 0.67‒1.00, respectively. We used a linear scale to classify the vulnerability and the method of Adhikari et al.59 with minor modifications. M. coypus habitat expansion was determined by differentiating between the current and future habitats using R package ‘Raster’ v3.568 as described by Jeon et al.47.
Ethics approval
Permission to collect species presence data are obtained from the Ministry of Environment, Republic of Korea.
Data availability
All data generated or analyzed during this study are included in this article.
References
Kim, I. R. et al. Genetic diversity and population structure of nutria (Myocastor coypus) in South Korea. Animals 9, 1164. https://doi.org/10.3390/ani9121164 (2019).
GISD. Of the World's Worst Invasive Alien Species. Global Invasive Species Database. http://www.iucngisd.org/gisd/100_worst.php. 100, (2021).
Hong, S., Do, Y., Kim, J. Y., Kim, D. & Joo, G. Distribution, spread and habitat preferences of nutria (Myocastor coypus) invading the lower Nakdong River, South Korea. Biol. Invas. 17, 1485–1496. https://doi.org/10.1007/s10530-014-0809-8 (2015).
Ojeda, R., Bidau, C. & Emmons, L. Myocastor coypus (errata version published in 2017). The IUCN Red List Threat. Species (2016): e.T14085A121734257.
Tsiamis, K. et al. Baseline Distribution of Invasive Alien Species of Union Concern (Publications Office of the European Union, 2017).
Carter, J. & Leonard, B. P. A review of the literature on the worldwide distribution, spread of, and efforts to eradicate the coypu (Myocastor coypus). Wildl. Soc. Bull. 30, 162–175 (2002).
Kim, Y. C. et al. Distribution and management of nutria (Myocastor coypus) populations in South Korea. Sustainability 11, 4169. https://doi.org/10.3390/su11154169 (2019).
Park, J. H. et al. The first case of Capillaria hepatica infection in a nutria (Myocastor coypus) in Korea. Korean J. Parasitol. 52, 527–529. https://doi.org/10.3347/kjp.2014.52.5.527 (2014).
Fratini, F., Turchi, B. E., Ebani, V. V. & Bertelloni, F. The presence of Leptospira in coypus (Myocastor coypus) and rats (Rattus norvegicus) living in a protected wetland in Tuscany (Italy). Vet. Arh. 85, 407–414 (2015).
Lee, D. H., Kil, J. H. & Kim, D. E. The study on the distribution and inhabiting status of nutria (Myocastor coypus) in Korea. Korean J. Environ. Ecol. 27, 316–326 (2013).
Guichón, M. L., Doncaster, C. P. & Cassini, M. H. Population structure of coypus (Myocastor coypus) in their region of origin and comparison with introduced populations. J. Zool. 261, 265–272. https://doi.org/10.1017/S0952836903004187 (2003).
Bertolino, S., Perrone, A. & Gola, L. Effectiveness of coypu control in small Italian Wetland areas. Wildl. Soc. Bull. 33, 714–720. https://doi.org/10.2193/0091-7648(2005)33[714:EOCCIS]2.0.CO;2 (2005).
Schertler, A. et al. The potential current distribution of the coypu (Myocastor coypus) in Europe and climate change induced shifts in the near future. NeoBiota 58, 129–160. https://doi.org/10.3897/neobiota.58.33118 (2020).
Hilts, D. J., Belitz, M. W., Gehring, T. M., Pangle, K. L. & Uzarski, D. G. Climate change and nutria range expansion in the Eastern United States. J. Wild. Manaag. 83, 591–598. https://doi.org/10.1002/jwmg.21629’ (2019).
Jarnevich, C. et al. Evaluating simplistic methods to understand current distributions and forecast distribution changes under climate change scenarios: An example with coypu (Myocastor coypus). NeoBiota 32, 107–125. https://doi.org/10.3897/neobiota.32.8884 (2017).
Korean Metrological Administration, (2020). Korean Climate Change Assessment Report 2020.
Guillera-Arroita, G. et al. Is my species distribution model fit for purpose? Matching data and models to applications. Glob. Ecol. Biogeogr. 24, 276–292. https://doi.org/10.1111/geb.12268 (2015).
Phillips, S. J., Anderson, R. P. & Schapire, R. E. Maximum entropy modeling of species geographic distributions. Ecol. Modell. 190, 231–259. https://doi.org/10.1016/j.ecolmodel.2005.03.026 (2006).
Hong, S., Cowan, P., Do, Y. & Gim, J. S. Seasonal feeding habits of coypu (Myocastor coypus) in South Korea. Hystrix 27, 123–128 (2016).
Kim, H. S., Kong, J. Y., Kim, J. H., Yeon, S. C. & Hong, I. H. A Case of Fascioliasis in A Wild Nutria, Myocastor coypus Republic of Korea. Korean J. Parasitol. 56, 375–378. https://doi.org/10.3347/kjp.2018.56.4.375 (2018).
Do, Y., Kim, J. Y., Im, R. Y. & Kim, S. B. Spatial distribution and social characteristics for wetlands in Gyeongsangnam-do Province. Korean J. Limnol. 45, 252–260 (2012).
IPCC. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. (2013).
Sheffels, T. R. Status of Nutria (Myocastor coypus) Populations in the Pacific Northwest and Development of Associated Control and Management Strategies, with an Emphasis on Metropolitan Habitats, PhD Thesis (Portland State Univ., 2013).
Doncaster, C. P. & MlCOL, T. Annual cycle of a coypu (Myocastor coypus) population: Male and female strategies. J. Zool. 217, 227–240. https://doi.org/10.1111/j.1469-7998.1989.tb02484.x (1989).
Reggiani, G., Boitani, L. & Stefano, R. Population dynamics and regulation in the coypu Myocastor coypus in Central Italy. Ecography 18, 138–146. https://doi.org/10.1111/j.1600-0587.1995.tb00334.x (1995).
Cha, Y., Cho, K. H., Lee, H., Kang, T. & Kim, J. H. The relative importance of water temperature and residence time in predicting cyanobacteria abundance in regulated rivers. Water Res. 124, 11–19. https://doi.org/10.1016/j.watres.2017.07.040 (2017).
Hellmann, J. J., Byers, J. E., Bierwagen, B. G. & Dukes, J. S. Five potential consequences of climate change for invasive species. Conserv. Biol. 22, 534–543. https://doi.org/10.1111/j.1523-1739.2008.00951.x (2008).
Pereira, A. D. et al. Modeling the geographic distribution of Myocastor coypus (Mammalia, Rodentia) in Brazil: Establishing priority areas for monitoring and an alert about the risk of invasion. Stud. Neotrop. Fauna Environ. 55, 139–148. https://doi.org/10.1080/01650521.2019.1707419 (2020).
Pearson, R. G. & Dawson, T. P. Predicting the impacts of climate change on the distribution of species: Are bioclimate envelope models useful?. Glob. Ecol. Biogeogr. 12, 361–371. https://doi.org/10.1046/j.1466-822X.2003.00042.x (2003).
Rogers, C. E. & McCarty, J. P. Climate change and ecosystems of the mid-atlantic region. Clim. Res. 14, 235–244. https://doi.org/10.3354/cr014235 (2000).
Adhikari, P. et al. Potential impact of climate change on plant invasion in the Republic of Korea. J. Ecol. Environ. 43, 36. https://doi.org/10.1186/s41610-019-0134-3 (2019).
Welsch, D. J., Smart, D. L., Boyer, J. N. & Minkin, P. Forested Wetlands: Functions, Benefits and the Use of Best Management Practices (US Dept of the Interior Fish and Wildlife Service, 2021).
Borgnia, M., Galante, M. L. & Cassini, M. H. Diet of the coypu (nutria, Myocastor coypus) in agro-systems of Argentinean pampas. J. Wildl. Manag. 64, 354–361. https://doi.org/10.2307/3803233 (2000).
Colares, I. G., Oliveira, R. N. V., Liveira, R. M. & Colares, E. P. Feeding habits of coypu (Myocastor coypus Molina 1978) in the wetlands of the Southern region of Brazil. An. Acad. Bras. Cienc. 82, 671–678. https://doi.org/10.1590/s0001-37652010000300015 (2010).
Corriale, M. J., Arias, S. M., Bó, R. F. & Porini, G. Habitat-use patterns of the coypu (Myocastor coypus) in an urban wetland of its original distribution. Acta Theriol. 51, 295–302. https://doi.org/10.1007/BF03192681 (2006).
Linscombe, G., Kinler, N. & Wright, V. Nutria population density and vegetative changes in brackish marsh in coastal Louisiana. In Worldwide Furbearer Conference Proceedings (eds Chapman, J. A. & Pursley, D.) 129–141 (Worlwide Furbearer Conference Inc, 1981).
Aliev, F. Contribution to the study of nutria migrations (Myocastor coypus). Saugetierkd. Mitt. 16, 301–303 (1968).
Farashi, A. & Najafabadi, M. S. A model to predict dispersion of the alien nutria, Myocastor coypus Molina, 1782 (Rodentia) Northern Iran. Acta Zool. Bulg. 69, 65–70 (2017).
Vilà, M. et al. How well do we understand the impacts of alien species on ecosystem services? A pan-European, cross-taxa assessment. Front. Ecol. Environ. 8, 135–144. https://doi.org/10.1890/080083 (2010).
Adhikari, P. et al. Seasonal and altitudinal variation in roe deer (Capreolus pygargus tianschanicus) diet on Jeju Island, South Korea. J. Asia Pac. Biodivers. 9, 422–428. https://doi.org/10.1016/j.japb.2016.09.001 (2016).
Koo, K. A., Kong, W. S., Nibbelink, N. P., Hopkinson, C. S. & Lee, J. H. Potential effects of climate change on the distribution of cold-tolerant evergreen broadleaved woody plants in the Korean Peninsula. PLoS ONE 10, e0134043. https://doi.org/10.1371/journal.pone.0134043 (2015).
National Institute of Biological Research. Korean Red List of Threatened Species 2nd edn. (Ministry of Environement of Korea, 2014).
Kil, J. et al. Monitoring of Invasive Alien Species Designated by the Wildlife Protection Act (VII) (Natl Inst. of Environmental Research, 2013).
Busby, J. R. In Bioclim, a Bioclimatic Analysis and Prediction System in Nature Conservation: Cost Effective Biological Surveys and Data Analysis (eds Margules, C. R. & Austin, M. P.) 64–68 (CSIRO, 1991).
Lee I. H., Park S. H., Kang, H. S. & Cho C. H. Regional climate projections using the HadGEM3-RA in Proceedings of the 3rd International Conference on Earth System Modelling; Hamburg, Germany. 17–21 September 2012. (2012).
Robert, J. H., Phillips, S., Leathwick, J. & Elith, J. Package ‘dismo’ version 1.3. , https://cran.rproject.org/web/packages/dismo.pdf (2021).
Jeon, J. Y., Adhikari, P. & Seo, C. Impact of climate change on potential dispersal of Paeonia obovata (Paeoniaceae), a critically endangered medicinal plant of South Korea. Ecol. Environ. Conserv. 26, S145–S155 (2020).
Dormann, C. F. et al. Collinearity: A review of methods to deal with it and a simulation study evaluating their performance. Ecography 36, 27–46. https://doi.org/10.1111/j.1600-0587.2012.07348.x (2013).
Shin, M. S., Seo, C., Lee, M. & Kim, J. Y. Prediction of potential species richness of plants adaptable to climate change in the Korean Peninsula. J. Environ. Impact Assess. 27, 562–581 (2018).
Adhikari, P. et al. Northward range expansion of southern butterflies according to climate change in South Korea. KSCCR 11, 643–656. https://doi.org/10.15531/KSCCR.2020.11.6.643 (2020).
Song, C. et al. Estimation of future land cover considering shared socioeconomic pathways using scenario generators. KSCCR 9, 223–234. https://doi.org/10.15531/KSCCR.2018.9.3.223 (2018).
Fick, S. E. & Hijmans, R. J. WorldClim 2: New 1-km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37, 4302–4315. https://doi.org/10.1002/joc.5086 (2017).
Dukes, J. S. & Mooney, H. A. Does global change increase the success of biological invaders?. Trends Ecol. Evol. 14, 135–139. https://doi.org/10.1016/s0169-5347(98)01554-7 (1999).
Thuiller, W., Georges, D., Gueguen, M., Engler, R. & Breiner, F. Package ‘biomod2’: Ensemble Platform for Species Distribution Modeling, version 3.5.1. https://cran.r-project.org/web/packages/biomod2/biomod2.pdf (2021).
Elith, J. et al. Novel methods improve prediction of species’ distributions from occurrence data. Ecography 29, 129–151. https://doi.org/10.1111/j.2006.0906-7590.04596.x (2006).
Barbet-Massin, M., Jiguet, F., Albert, C. H. & Thuiller, W. Selecting pseudo-absences for species distribution models: how, where and how many?. Methods Ecol. Evol. 3, 327–338. https://doi.org/10.1111/j.2041-210X.2011.00172.x (2012).
Brown, J. L. SDM toolbox: A python-based GIS toolkit for landscape genetic, biogeographic and species distribution model analyses. Methods Ecol. Evol. 5, 694–700. https://doi.org/10.1111/2041-210X.12200 (2014).
Veloz, S. D. Spatially autocorrelated sampling falsely inflates measures of accuracy for presence–only niche models. J. Biogeogr. 36, 2290–2299. https://doi.org/10.1111/j.1365-2699.2009.02174.x (2009).
Adhikari, P., Lee, Y. H., Park, Y.-S. & Hong, S. H. Assessment of the spatial invasion risk of intentionally introduced alien plant species (IIAPS) under environmental change in South Korea. Biology 10, 1169 (2021).
Hong, S. H., Lee, Y. H., Lee, G., Lee, D. H. & Adhikari, P. Predicting impacts of climate change on northward range expansion of invasive weeds in South Korea. Plants 10, 1604. https://doi.org/10.3390/plants10081604 (2021).
Pearsons, R. G. Species distribution modeling for conservation educators and practitioners. Lessons Conserv. 3, 54–58 (2010).
Allouche, O., Tsoar, A. & Kadmon, R. Assessing the accuracy of species distribution models: Prevalence, kappa and the true skill statistic (TSS). J. Appl. Ecol. 43, 1223–1232. https://doi.org/10.1111/j.1365-2664.2006.01214.x (2006).
Thuiller, W., Lavorel, S. & Araújo, M. B. Niche properties and geographical extent as predictors of species sensitivity to climate change. Glob. Ecol. Biogeogr. 14, 347–357. https://doi.org/10.1111/j.1466-822X.2005.00162.x (2005).
Lobo, J. M., Jiménez-Valverde, A. & Real, R. AUC: A misleading measure of the performance of predictive distribution models. Global. Ecol. Biogeography. 17, 145–151. https://doi.org/10.1111/j.1466-8238.2007.00358.x (2008).
Swets, J. A. Measuring the accuracy of diagnostic systems. Science 240, 1285–1293 (1988).
Baldwin, R. Use of maximum entropy modeling in wildlife research. Entropy 11, 854–866. https://doi.org/10.3390/e11040854 (2009).
Adhikari, P. et al. Potential impact of climate change on the species richness of subalpine plant species in the mountain national parks of South Korea. J. Ecol. Environ. 42, 36. https://doi.org/10.1186/s41610-018-0095-y (2018).
Hijmans, R. J. et al. Package ‘raster’ v 3.5: geographical data analysis and modeling. https://cran.r-project.org/web/packages/raster/raster.pdf, (2021).
Acknowledgements
This research was supported by a grant from the Korea Environment Industry & Technology Institute [2020002990009] funded by the Ministry of Environment in Republic of Korea.
Author information
Authors and Affiliations
Contributions
The study was designed by P.A. and D.H.L. Data collection, analysis, and manuscript writing were performed by P.A. B.J.K. and H.S.H. supported the research design and D.H.L. participated in the survey, review of the manuscript, and coordinated the overall research. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Adhikari, P., Kim, BJ., Hong, SH. et al. Climate change induced habitat expansion of nutria (Myocastor coypus) in South Korea. Sci Rep 12, 3300 (2022). https://doi.org/10.1038/s41598-022-07347-5
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-022-07347-5
This article is cited by
-
Predicting future climate change impacts on the potential distribution of the black howler monkey (Alouatta pigra): an endangered arboreal primate
Environmental Monitoring and Assessment (2024)
-
Distribution of sika deer (Cervus nippon) and the bioclimatic impact on their habitats in South Korea
Scientific Reports (2023)
-
Global spatial distribution of Chromolaena odorata habitat under climate change: random forest modeling of one of the 100 worst invasive alien species
Scientific Reports (2023)
-
Climatic niche dynamics in the invasive nutria, Myocastor coypus: global assessment under climate change
Biological Invasions (2023)
-
Risk assessment for the native anurans from an alien invasive species, American bullfrogs (Lithobates catesbeianus), in South Korea
Scientific Reports (2022)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.