Background/Question/Methods Do plant populations from climatic unpredictable ecosystems such as arid and semiarid ecosystems have a higher proportion of dormant and larger seeds than more predictable mesic ecosystems? This question is answered by considering soil seed bank dynamics that are of crucial importance in highly variable and unpredictable environments. They play a determinant role in plant community dynamics, as differential plant germination strategies buffer against inter-annual variability in growing conditions.  Thus, understanding the role of soil seed banks in determining the composition and structure of plant communities under future climatic scenarios becomes an important task in order to understand future global changes. We selected the strong climatic gradient in Israel, that provides an excellent opportunity to test predictions regarding the effects of global climate change on plant community dynamics. This aridity gradient runs from mesic Mediterranean ecosystems in the North to arid desert ecosystems in the South. Climate change scenarios are experimentally tested with manipulations of rainfall amount using rainout shelters and irrigation systems.

Results/Conclusions The results indicated significant differences in seedling densities among the study sites along the aridity gradient. Seedling density increased with increasing rainfall conditions.  Significant differences were also noted when comparing seed density between habitats (open gaps vs. shrub understory). Higher seedling densities were observed at the shrub understory at the arid part of the gradient while the opposite was noted at the mesic part (i.e. higher seedling density at the open gaps between shrubs). This phenomenon indicated shrub facilitation effects at the arid part of the gradient and competitive exclusion at the more humid part of the gradient. Moreover, important differences among years were clearly noted as response of rainfall conditions of the year when seeds were produced. The results collected showed no clear response of the seed bank to the rainfall manipulations. According to our expectations we should have observed a decreased in seed density under droughting conditions while the opposite by increasing rainfall through irrigation. The results were not consistent and in some cases opposite to expected.  The detected short-term resistance on the seed bank density does not necessarily imply resistance to long-term rainfall manipulations.