The accuracy of predicting the spread of biological invasions is improved if models explicitly incorporate the two main dispersal mechanisms: diffusive spread and jump dispersal. However, quantitative methods for differentiating these two mechanisms in spatial occurrence data are lacking. We present jumpID, an R package using directional analysis of occurrence data to distinguish between jump dispersal and diffusive spread in biological invasions. We applied this method to occurrence data from the spotted lanternfly (Lycorma delicatula) invasion in the US, a pest rapidly expanding its range and impacting the forest and grape industries. Application of jumpID to a dataset of 123,542 occurrence records of spotted lanternfly uncovered 152 dispersal jumps between 2014–2022, with the first jump in 2017, three years after spotted lanternfly’s first find. More than half of the dispersal jumps started satellite invasions the year after their detection. The average jump distance did not change over time, with 89% of jumps shorter than 200 km and just three jumps farther than 300 km. The overall spread rate was 41 ± 24 SD km/year, but reduced to 25 ± 11 SD km/year when considering diffusive spread only. Estimating jump dispersal enhances our understanding of species’ dispersal mechanisms, provides more robust estimates of diffusion rates for spread models, and helps determine the perimeters for containment and control measures. The R package jumpID is openly available to facilitate invasive spread analysis. jumpID equips scientists and managers with a tool to separate the spread of invasive species into diffusion and jump dispersal components, allowing for more precise parameterization of spread models and directly informing management strategies. Application of jumpID to the spotted lanternfly system indicates that management efforts targeting jump dispersal should focus on a 200-km buffer around the invasion boundary which is where 89% of jumps occur.
