A Dataset of Potential Distributions and Species Richness of Ecosystem-Disrupting Invasive Alien Animal Species in South Korea under Shared Socioeconomic Pathways Climate Scenarios

AInvasive alien species are major drivers of biodiversity change, yet spatially explicit datasets that integrate their distributions with future climate scenarios remain scarce. In South Korea, several ecosystem-disrupting animal species have rapidly expanded across multiple ecosystems, highlighting the need for standardized, national-scale spatial data products to support management and research. Here we present a gridded dataset of current and future potential distributions and species richness for seven ecosystem-disrupting animal species designated by the Ministry of Environment in South Korea: <i>Pochazia shantungensis, Lycorma delicatula, Vespa velutina, Pseudemys concinna, Metcalfa pruinosa, Trachemys scripta</i>, and <i>Lithobates catesbeianus</i>. Species occurrence records were compiled from the National Ecosystem Survey (1997-2021), national park monitoring data (2003-2022), and published literature, then filtered and merged to unique 1 km<sup>2</sup> grid cells. Environmental predictors include six bioclimatic variables (BIO1, BIO2, BIO3, BIO12, BIO13, BIO14) and three climate-trend indices derived from 1 km<sup>2</sup> ensemble climate data provided by the Korea Meteorological Administration under SSP2-4.5 and SSP5-8.5 scenarios for the 2030s, 2050s, 2070s and 2090s. For each species, we fitted Random Forest models using presence and pseudo- absence data, and evaluated performance using receiver operating characteristic (ROC), true skill statistic (TSS), and Cohen’s kappa. Across all species, mean ROC was 0.752 (range, 0.664-0.845), with corresponding TSS and kappa values indicating robust discrimination. Thresholded suitability maps were then summed across species to derive 1 km<sup>2</sup> species-richness surfaces (0-7 species) for the baseline and each future time slice. These richness layers were further aggregated to mean values at the sigun- gu administrative level. The resulting dataset consists of species-level suitability maps, grid- and administrative-level species richness, and metadata on model performance and variable importance. It provides reusable spatial information that can be integrated with land-use, socio-economic, or conservation-planning layers to support invasive-species risk assessment, prioritization of management areas, and climate- change impact studies.