Timing and spatial patterning of seed dispersal and redistribution in a South American warm desert

Abstract

We measured newly-produced seeds entering the soil (Potential Seed Bank) to assess the timing and spatial patterning of Phase I dispersal in the central Monte desert, Argentina. Rates of forb- (6.13 mg m-2 d-1) and shrub-seed input (48.9 mg m-2 d-1) were maximum in early summer. The rate of grass-seed input, instead, was similar in early and late summer (7 to 8 mg m-2 d-1). About 90% of forb- and shrub-seed mass entered the habitat through protected (i.e., under canopy) areas, whereas 70% of grass-seed mass did so through exposed areas. Adult plant location and the uneven impact of wind on shrub, forb and grass seeds may explain such patterns. We also compared the Potential Seed Bank with the soil seed bank in the following spring (Realized Seed Bank). Seeds that form transient banks in other ecosystems (e.g., shrub seeds of the genus Larrea, or perennial grass seeds like those of Pappophorum and Trichloris) prevailed in the Potential Seed Bank. Some annual forb seeds, instead, appeared to form a more persistent seed bank, and prevailed in the Realized Seed Bank (e.g., Chenopodium). Horizontal redistribution did not affect the spatial patterning of forb and shrub seeds, but produced a more homogeneous distribution of grass seeds in the habitat. The impact of wind could explain the redistribution pattern of grass seeds. Finally, we found almost 80% of total seeds in the top 2 cm of soil. The smallest grass and forb seeds (Sporobolus and Descurainia) as well as some medium-sized and large forb seeds (e.g., Glandularia, Sphaeralcea, Phacelia) were able to reach deeper soil layers in the central Monte desert.