Estimating Geographic Range Shifts of Vespa simillima after Vespa velutina (Hymenoptera: Vespidae) Invasion in South Korea

INTRODUCTION

In South Korea, there are six native species of hornets that include Vespa analis, V. mandarinia, V. simillima, V. crabro, V. ducalis, and V. dybowskii (Choi et al., 2012). Hornets belong to the genus Vespa and the subfamily Vespinae, which has 22 species globally (Perrard et al., 2013). Recently, there has been a growing concern regarding the increase in hornet populations throughout Korea as these species are considered a challenge to the beekeeping industry in Korea. These species, along with the non-native species called yellow-legged hornet (Vespa velutina nigritorax), which was recorded in Korea in 2003 (Kim et al., 2006), have caused damage to beekeepers by attacking beehives. For example, the annual economic loss of hornets to the beekeeping industry in Korea is estimated at 100-170 billion Korean won (Jung, 2012; Jung and Cho, 2015). In Europe, V. velutina has damaged 18 to 50% of honey beehives (Laurino et al., 2019). Therefore, this concern has increased the need to study the distribution of hornets in South Korea (Kwon and Choi, 2020; Zhu et al., 2020) and has led to more studies to identify the threats of hornets and to estimate the effects of the invasion of V. velutina on pollinators and other hornets.

In addition to affecting honeybees and damaging beekeepers, hornets also create disturbances as competing species. For example, Choi et al. (2012) reported that V. velutina became the most abundant species of hornet in South Korea and could occupy habitats of native hornets such as V. simillima. This species has a high ability to occupy different environments beyond the borders of its native niche (Verdasca et al., 2022). Kwon and Choi (2020) compared the aggressive behaviors and body size of V. velutina with five native hornet species to identify the interspecific hierarchies that influence the rate of spread of this species. Their results showed that V. velutina was superior to V. simillima but it was weaker against other hornets. To test this claim, it is necessary to map the distribution of V. simillima before and after the invasion of V. velutina throughout South Korea. Choi et al. (2012) reported that V. velutina had a significant expansion in Korea between 2003 and 2015. Species distribution models (SDMs) are among the tools that can help us to investigate the competitive effects caused by the invasion of V. velutina on V. simillima. In this method, according to the correlation between the climatic variables and their presence points, we can determine the areas that are potentially suitable for species of interest (Rahimi et al., 2021).

Recently, there was a study to develop SDMs to determine high-risk areas according to the presence of hornet species in the world Kim et al. (2021). In this study, they attempted to determine the geographic ranges of V. veluina at a global scale. In their study, the most suitable areas for V. velutina in Korea were the southern areas. According to the above background, the main objective of this study is 1- to determine the geographic range of V. simillima before and after the invasion of V. velutina in South Korea. For this purpose, by using occurrence points of V. simillima in different years, distribution maps of this species are prepared at the country level, and by comparing these maps, distribution changes due to the introduction of V. velutina are determined.

METHODS

To prepare species distribution maps using species distribution models, we need two general types of data, presence data and predictive data (Zurell et al., 2020). To determine the geographic range of V. simillima, we used presence data (300 points) (Fig. 1) of this species between the years 1999 and 2022 freely available at www.gbif.org. We divided occurrence data into two groups before and after 2011. The number of these points was 150 for the year 1999-2011 and 150 for the year 2012-2022.

Fig. 1.

Occurrence data of V. simillima during 1999-2011 (A) and 2012-2022 (B).

Bioclimatic variables were also used as predictors of distribution. These data were downloaded from the WorldClim database (www.worldclim.org). To reduce Multicollinearity among climatic variables, we removed variables that correlated higher than 0.8 (Rahimi et al., 2021). Finally, in this study, layers such as Bio 3, Bio 7, Bio 8, Bio 12, and Bio 17 were used in the modeling process. In the present study, we used the SDM package (Naimi et al., 2016) in R software to model the distribution of the studied species. For this purpose, we used the MaxEnt algorithm to prepare habitat suitability maps based on the presence and climate data. We used 10 runs of subsampling replications to evaluate the model and considered 70% of the presence data as training data and the remaining 30% as a test. Also, to evaluate the effectiveness of the results obtained from the MaxEnt model, we used two statistics of True skill statistic (TSS) and the area under the ROC curve (AUC). The value range of habitat suitability maps is between 0 and 1, where the number 1 represents the maximum suitability. To estimate the changes in the distribution maps, we divided the maps into three classes with low (0-0.3), moderate (0.3-0.6) and high suitability (>0.6).

RESULTS

1. Model assessment

The TSS and AUC values for V. simillima indicated the developed models in this study were reliable (Table 1). The AUC values between 0.7 and 0.9 indicate acceptable results and values greater than 0.9 mean excellent prediction (Valavi et al., 2020). TSS values are between -1 and +1, where 1 indicates complete and correct prediction, while values between 0.4 and 0.6 indicate moderate prediction.

Table 1.

Model validation metrics including TSS and AUC for V.simillima

2. Variable importance

For the period 1999-2011, precipitation seasonality is the most important variable in determining the geographical distribution of V. simillima, and precipitation of driest month has the least effect on the distribution of this species (Table 2). However, for the period 2012-2022, precipitation of driest month is the most important variable that determines the distribution of this species, and precipitation of warmest quarter has the least effect on predicting the distribution of V. simillima (Table 2).

Table 2.

Importance of different climatic variables used in the models

3. Geographic distribution

The distribution of V. simillima between 1999 and 2011 is more limited to the eastern and southern parts of South Korea (Fig. 1). However, between the years 2012 and 2022, the distribution of this species has expanded towards the central and western parts (Fig. 2).

Fig. 2.

Habitat suitability maps of V. simillima during 1999-2011 (A) and 2012-2022 (B). Light colors show high suitability (values close to 1), and dark colors (values close to 0) show low suitability. Therefore, in values close to 1, we have a higher probability and expectation for the presence of V. simillima.

4. Change detection

Each of these classes in the period of 1999-2011 has almost the same area as in the period of 2012-2022. The overlap column also shows the degree of overlap of each of these classes with each other (Table 3). The overlap between the class with high suitability between 2012-2022 with its previous period is only 24% (Table 3). This result shows that in recent years, the suitable areas for V. simillima have had a significant shift to the central and western regions of South Korea (Fig. 2).

Table 3.

Changes of suitable habitats (Ha) by the suitability classes during 1999-2000 and 2012-2022, and their overlapping percentage (%) of V. simillima in Korea

DISCUSSION

Our results showed that between the years 2012 and 2022, the geographical range of V. simillima has significant change compared to the previous years. The difference in the geographical distribution of this species in the present study can be related to two main reasons: (1) the presence of the competing species, i.e. V. velutina, and (2) the uncertainty associated with the presence of data obtained from the GBIF database. This study showed that V. simillima distribution has completely disappeared from the southern parts and this species has shifted to higher latitudes. This phenomenon could be explained by the increase of the habitat of V. velutina. Choi et al. (2012) found that this species occupies most of the southern parts of South Korea. Kim et al. (2021) also showed that the most suitable areas for V. velutina are the southern parts of South Korea. Another hypothesis is that the populations of V. velutina as an invasive species could decrease due to competition with native hornets or the development of defensive mechanisms in European honey bees. Although testing this hypothesis is beyond the scope of this study, the findings of other studies further confirm the hypothesis of the expansion of the population of V. velutina and the decrease of the population of competing species in South Korea (Choi et al., 2012). In addition, the development of defensive mechanisms in the European honey bee as a non-native species is likely to be less efficient than in native honeybee (A. cerena) (Kwon and Choi, 2020; Cappa et al., 2021).

Southern parts of Korea (Busan) have a subtropical climate, while the rest of South Korea has a hot and humid continental climate (Peel et al., 2007), meaning that the subtropical climate can be the favorite climate for V. velutina. Other studies have also shown that climate is an important factor in determining the geographical distribution of this species (Bessa et al., 2016). In Europe, the spread of this non-native species is expanding, which means that it has a high ability to occupy the niches of competing species. For example, Keeling et al. (2017) claimed that, without control, V. velutina could occupy the British mainland rapidly, depending on how the hornet responds to the colder climate in Britain compared to France. Therefore, the difference in the geographical distribution of V. simillima in the present study may be related to the role and presence of V. velutina. However, to further investigate this hypothesis, more studies are needed in this field. Future studies can compare the relative abundance of V. simillima and V. velutina species by trapping in the southern parts of Korea. If V. velutina has occupied V. simillima niches in these areas, we expect V. simillima abundance to show a significant decrease in these areas.

CONCLUSION

Using species distribution models, we tried to estimate the geographic range shift of V. simillima after the invasion of V. velutina in South Korea. These models showed that the southern regions of Korea that are climatically suitable for V. velutina, have been unsuitable for V. simillima between 2012 and 2022. This result was obtained using the GBIF database. These data are valid and hundreds of scientific studies have been published based on this database. However, the southern regions of South Korea might not be sampled enough after 2012, and therefore, fewer occurrences of V. simillima were recorded in these areas. On the other hand, other studies claim that V. velutina is known as an important competitor for V. simillima and can occupy the niches of this species. Our study also confirms these findings and shows that it is possible that in the southern parts of South Korea V. simillima population has decreased due to the presence of V. velutina.

Acknowledgments

This research was funded by by the BSRP through the National Research Foundation of Korea (NRF), Ministry of Education (NRF-2018R1A6A1A03024862), and Rural development Administration Agenda project on climate change monitoring (PJ01480803).

References

• Bessa, A. S., J. Carvalho, A. Gomes and F. Santarem. 2016. Climate and land-use drivers of invasion: predicting the expansion of Vespa velutina nigrithorax into the Iberian Peninsula. Insect Conserv. Divers. 9(1): 27-37. [https://doi.org/10.1111/icad.12140]
• Cappa, F., A. Cini, L. Bortolotti, J. Poidatz and R. Cervo. 2021. Hornets and honey bees: A coevolutionary arms race between ancient adaptations and new invasive threats. Insects 12(11): 1037. [https://doi.org/10.3390/insects12111037]
• Choi, M. B., S. J. Martin and J. W. Lee. 2012. Distribution, spread, and impact of the invasive hornet Vespa velutina in South Korea. J. Asia-Pac. Entomol. 15(3): 473-477. [https://doi.org/10.1016/j.aspen.2011.11.004]
• Jung, C. 2012. Initial stage risk assessment of an invasive hornet, Vespa velutina nigrithorax Buysson (Hymenoptera: Vespidae) in Korea. J. Apic. 27(2): 95-104.
• Jung, C. and S.-k. Cho. 2015. Relationship between honeybee population and honey production in Korea: A historical trend analysis. J. Apic. 30(1): 7-12. [https://doi.org/10.17519/apiculture.2015.04.30.1.7]
• Keeling, M. J., D. N. Franklin, S. Datta, M. A. Brown, G. E. Budge. 2017. Predicting the spread of the Asian hornet (Vespa velutina) following its incursion into Great Britain. Sci. Rep. 7(1): 1-7. [https://doi.org/10.1038/s41598-017-06212-0]
• Kim, J. K., M. Choi and T. Y. Moon. 2006. Occurrence of Vespa velutina Lepeletier from Korea, and a revised key for Korean Vespa species (Hymenoptera: Vespidae). Entomol. Res. 36(2): 112-115. [https://doi.org/10.1111/j.1748-5967.2006.00018.x]
• Kim, S.-H., D. E. Kim, H. Lee, S. Jung and W.-H. Lee. 2021. Ensemble evaluation of the potential risk areas of yellow-legged hornet distribution. Environ. Monit. Assess. 193(9): 1-15. [https://doi.org/10.1007/s10661-021-09406-2]
• Kwon, O. and M. B. Choi. 2020. Interspecific hierarchies from aggressiveness and body size among the invasive alien hornet, Vespa velutina nigrithorax, and five native hornets in South Korea. PloS One 15(7): e0226934. [https://doi.org/10.1371/journal.pone.0226934]
• Laurino, D., S. Lioy, L. Carisio, A. Manino and M. Porporato. 2019. Vespa velutina: An alien driver of honey bee colony losses. Diversity 12(1): 5. [https://doi.org/10.3390/d12010005]
• Naimi, B., M. B. Araujo, M. B. Naimi, B. Naimi and M. Araujo. 2016. Package ‘sdm’.
• Peel, M. C., B. L. Finlayson and T. A. McMahon. 2007. Updated world map of the Köppen-Geiger climate classification. Hydrol. Earth Syst. Sci. 11(5): 1633-1644. [https://doi.org/10.5194/hess-11-1633-2007]
• Perrard, A., K. Pickett, C. Villemant, J.-i. Kojima and J. Carpenter, 2013. Phylogeny of hornets: a total evidence approach (Hymenoptera, Vespidae, Vespinae, Vespa). J. Hymenopt. Res. 32: 1-15. [https://doi.org/10.3897/jhr.32.4685]
• Rahimi, E., S. Barghjelveh and P. Dong. 2021, Estimating potential range shift of some wild bees in response to climate change scenarios in northwestern regions of Iran. J. Ecol. Environ. 45(1): 1-13. [https://doi.org/10.1186/s41610-021-00189-8]
• Valavi, R., J. Elith, J. J. Lahoz-Monfort and G. Guillera-Arroita. 2021. Modelling species presence-only data with random forests. Ecography 44(12): 1731-1742. [https://doi.org/10.1111/ecog.05615]
• Verdasca, M. J., L. Carvalheiro, J. A. Gutierrez, J. P. Granadeiro, Q. Rome, S. J. Puechmaille, R. Rebelo, H. Rebelo. 2022. Contrasting patterns from two invasion fronts suggest a niche shift of an invasive predator of native bees. PeerJ 10: e13269. [https://doi.org/10.7717/peerj.13269]
• Zhu, G., J. Gutierrez Illan, C. Looney and D. W. Crowder. 2020, Assessing the ecological niche and invasion potential of the Asian giant hornet. Proc. Natl. Acad. Sci. 117(40): 24646-24648. [https://doi.org/10.1073/pnas.2011441117]
• Zurell, D., J. Franklin, C. König, P. J. Bouchet, C. F. Dormann, J. Elith, G. Fandos, X. Feng, G. Guillera-Arroita and A. Guisan. 2020. A standard protocol for reporting species distribution models. Ecography 43(9): 1261-1277. [https://doi.org/10.1111/ecog.04960]