Technical Abstract: Ephemeral gully erosion is now recognized as a significant, if not dominant source of sediment from agricultural lands worldwide. Ephemeral gullies are typically plowed in and tilled across annually or more frequently, thus restoring the original swale and allowing erosion processes to become reactivated. Mechanized tillage redistributes soil from convex areas to swales in amounts that may exceed soil losses due to water erosion. This ‘conveyer belt’ repeatedly re-supplies concentrated flow zones with erodible material, potentially exacerbating the long-term impacts of ephemeral gully erosion on losses of soil material and crop productivity, as topsoil thickness is reduced not only in the location of the gullies themselves, but across entire fields. Previous studies have extended the basic theoretical framework of the Ephemeral Gully Erosion Model (EGEM) by refining existing components, incorporating additional components, and adapting the model to operate within the USDA Annualized Agricultural Non-Point Source model (AnnAGNPS). Models of ephemeral gully erosion such as EGEM, and now AnnAGNPS, limit the depth of an ephemeral gully channel to the tillage depth or depth to a less-erodible layer (e.g. fragipan). Once evacuated to this depth (through incision and headcut development and migration), channel widening by sidewall erosion dominates and erosion decreases. While the perceived magnitude of ephemeral gully erosion may be masked after gullies are repaired, the action of plowing in these channels reduces topsoil thickness and crop productivity over a much wider area than the channel itself. This study demonstrated that filling ephemeral gullies on an annual basis during tillage operations may be more destructive than realized. These results provide land managers an additional incentive for adopting soil conservation practices such as no-till.