Windborne Dispersal of Arthropod Vectors as a Pathway for Lumpy Skin Disease Virus Introduction into South Korea: A Multi-Species, Source-Attributed Modelling Study

Lumpy skin disease (LSD) is a rapidly expanding transboundary vector-borne disease threatening livestock systems across Asia, with increasing risk of introduction into previously unaffected regions such as South Korea. This study developed an integrated, spatially explicit framework to assess the plausibility, timing and source-region structure of LSD virus (LSDV) introduction via windborne dispersal of infected arthropod vectors. Maximum Entropy models based on climatic, environmental, and topographic variables were used to determine habitat suitability for four major vector families (Culex spp., Musca domestica, Stomoxys calcitrans, and Culicoides spp.). All four were run via the transport simulation; S. calcitrans had the best experimental transmission efficiency, whereas Culicoides spp. provides the strongest evidence for passive long-distance transport. Livestock density and vector suitability were used to calculate the risk of LSD incidence in the adjacent source regions. A fourth-order Runge–Kutta trajectory model driven by ERA5 reanalysis winds at 850 hPa (about 1458 m) was used to predict windborne dissemination. This model included degradation of mechanically transmitted virus as well as survival limitations based on temperature, humidity, and wind. The simulated trajectories arrived in South Korea within the retention period of mechanically transported LSDV, with a median transit time of 18 h (interquartile range 10.0–30.3 h), 61.4% arriving within 24 h, and 87.1% within 48 h. North Korea accounted for 76.5% of weighted arrivals, with a median transit time of 11 h. Transit durations varied significantly by source location. For all four vector species, the predicted infectious arrival mass peaked in September and decreased thirteen-fold by November. Infected farms were found in regions of greater anticipated exposure than non-infected controls when compared to the October 2023 Korean epidemic (58.1% vs. 34.3% in the two highest risk groups; median exposure 0.214 versus 0.045; p < 0.001), although randomly relocating the dispersal surface reproduced comparable agreement in 23.6% of permutations, indicating that outbreak locations provide only limited validation of a pathway-specific introduction model. Random permutation of the dispersion surface produced a comparable result in 23.6% of permutations, indicating limited robustness of the observed spatial relationship to spatial randomization. With pairwise geographical and temporal distances significantly associated (Mantel r = +0.136, p = 0.013), spatiotemporal analysis revealed that the 74 reported outbreaks originated from a maximum of 25 introduction events over a 13-day period, consistent with significant farm-to-farm dissemination after introduction. The seasonal window, source-region structure, and transit timeframes of windborne LSDV introduction into South Korea are described in this paper. Reported epidemic sites only partially validate a pathway-specific introduction model since they reflect both the point of introduction and subsequent local transmission. This evaluation is pathway-specific and does not evaluate overall incursion risk; windborne transport is one of numerous possible introduction paths to consider.