Microstegium vimineum is a shade tolerant annual C4 invasive grass in the Eastern US, which has been shown to negatively impact species diversity and succession in hardwood forests. To date, empirical studies have shown that population expansion is limited to <1 m yr−1, which is largely driven by gravity dispersal. However, this likely does not fully account for all mechanisms of population-scale dispersal as we observe greater rates of population expansion. Though water, both riparian and non-riparian water (i.e., ephemeral overland flow), have been speculated mechanisms for M. vimineum dispersal, few studies have empirically tested this hypothesis. We designed an experiment along the slopes of a Southwest Virginia hardwood forest to test the role of non-riparian water on local seed dispersal. We developed a seed marking technique by coating each seed with an ultraviolet (UV) powder that did not affect buoyancy to aid in situ seed recapture. Additionally, a new image analysis protocol was developed to automate seed identification from UV photos. Total seed mobility (summation of individual seed movement within each transect) was positively correlated with precipitation. Over a period of one month with 52.32 mm of precipitation, the maximum dispersal distance of any single recaptured seed was 2.4 m, and the average distance of dispersed seed was 0.21±0.04 m. This is the first quantitative evidence of non-riparian water dispersal in a forest understory, which accounts for an additional pathway of population expansion.