Is seed dispersal a bottleneck for passive restoration in semi-arid pampean grasslands? A seed-trap experiment

Abstract Seed dispersal patterns at both species and community levels are widely discussed, but evidence from literature is often limited, especially for natural grasslands. Therefore, we investigated seed dispersal patterns in a natural grassland in Argentina to estimate the reliability of adjacent fields as propagule sources for target species (i.e., native forage species). We installed two different seed trap types (pot and sticky) at distances of 0, 0.1, 2, 4 and 9 m from the vegetation boundaries of an exclosure area, mainly dominated by perennial native grasses. Additionally, we surveyed the aboveground vegetation and estimated the potential seed production for every single species. In total we found 39 plant species in the aboveground vegetation of which 34 were observed to produce seeds and 29 were caught by the seed traps. Over time, species richness, seed density and similarity (Bray-Curtis, Jaccard) to the aboveground vegetation as well as the Shannon diversity index, generally increased. However, these increases were lower at greater distances from the field boundaries, regardless of seed trap type. Overall, pot traps were more effective at catching species and achieved a higher seed density than the sticky traps. Our results highlight the importance of field boundaries as potential sources of seeds for target species and their role in enhancing biodiversity in adjacent grasslands and arable land.