Terrestrial environmental DNA survey substantially improves detection of satellite populations for an invasive insect

Abstract

Abstract Early detection of satellite populations is essential for managing invasive species, but traditional methods often fall short when populations are sparse and widespread. We assessed the effectiveness of terrestrial environmental DNA (eDNA) surveys for detecting spotted lanternfly (Lycorma delicatula) along major highways in New York State, a region at the invasion front in North America. Over two years, we sampled 83 sites covering 1,650 km of high-traffic roadway using a roller-based, host-tree surface eDNA collection method and compared results with concurrent visual surveys and a government-run trap surveillance program (‘circle traps’). We detected spotted lanternfly eDNA at 77% of sampled sites, while visual surveys indicated lanternfly presence at only 5% of these sites. Site occupancy models estimated detection probability at 0.77 per eDNA sample (95% CI: 0.66–0.86), compared with 0.25 per trap (95% CI: 0.21–0.31). Detection probability was lower in 2023 than in 2022 (eDNA: 0.68 vs. 0.89; traps: 0.13 vs. 0.32), likely due to our sampling in locations with lower population densities and distances farther from the invasion core. Extrapolation from hierarchical models indicated that three eDNA samples per occupied site would achieve a 95% cumulative detection probability, whereas circle traps would require 11–23 checks. Our surveys identified previously undocumented spotted lanternfly populations beyond the known range front at the time of the survey, including in high-risk agricultural and forested areas. These results show that terrestrial eDNA surveys can outperform traditional methods in detecting low-abundance invasive insects across large areas. Incorporating eDNA into early detection and rapid response efforts may improve management outcomes and reduce the risk of uncontrolled spread caused by human dispersal.