# Potential Distribution of Linepithema humile (Hymenoptera: Formicidae) in South Korea: An Ensemble Species Distribution Modeling Approach

## TL;DR

This study predicts the potential spread of Argentine ants in South Korea using models trained on data from their native and invasive ranges, identifying high-risk coastal areas.

## Contribution

The study demonstrates how ensemble species distribution models can predict invasive species spread in new regions with limited local data.

## Key findings

- The predicted distribution includes southern and southwestern coastal areas of South Korea.
- Model projections were consistent with the initial detection site in Busan.
- High-risk areas beyond Busan were identified for early detection and management.

## Abstract

Linepithema humile
is a well‐known invasive alien species that was first detected in 2019 at the cargo handling area of Busan Station in South Korea and has since established a population near the discovery site. Despite sufficient time for potential spread, no additional occurrence records have been reported outside the initial detection site. To predict its potential distribution across South Korea, we constructed species distribution models (SDMs) using occurrence data from its native range in South America and invasive range in North America. Modeling was conducted with the BIOMOD2 platform using five algorithms (ANN, GBM, MARS, MAXENT, and RF) and six environmental variables (Bio03, Bio04, Bio13, Bio16, Bio18, and the mean annual cumulative soil temperature degree‐days above 15.9°C). Model performance was evaluated with Kappa, ROC, and TSS, and only models meeting thresholds (Kappa ≥ 0.7, ROC ≥ 0.7, and TSS ≥ 0.5) were used for ensemble modeling via EMmean, EMwmean, and EMca methods. Projection accuracy was assessed using chi‐square tests based on occurrence data not used in model training. The predicted potential distribution included southern and southwestern coastal areas, which was consistent with the record in Busan. This study demonstrates the utility of SDMs trained on occurrence data from outside Korea in predicting the potential distribution of
L. humile
with limited domestic records and highlights high‐risk areas beyond Busan. Such approaches may support early detection and management strategies in the initial stages of biological invasion.

The Argentine ant (
Linepithema humile
) was first discovered in Korea in Busan in 2019. Due to a lack of occurrence data, its distribution in Korea remains difficult to predict. This study constructed a single and ensemble model based on BIOMOD2 using occurrence data and environmental variables from South and North America and projected it to Korea. The predictions confirmed a high probability of distribution along the southern coast and Jeju, which can be used to develop countermeasures against the spread or further invasion of Argentine ants.