Milena Holmgren1, Juan L. Celis-Diez2, Heleen Woerle1, and Juan J. Armesto2. (1) University of Wageningen, (2) Fundación Senda Darwin, Pontificia Universidad Católica & Center for Advanced Studies in Ecology and Biodiversity (CASEB)
Background/Question/Methods Mediterranean-type ecosystems are highly fragmented global biodiversity hotspots. Regeneration of these semiarid ecosystems has proved difficult but can be strongly facilitated by plant-nurse interactions. Here we test if the interaction of microclimatic and herbivore conditions under existing vegetation patches define a patch-size threshold for tree seedling recruitment and if that is dependent on overall climate conditions and plant functional types. We used a combination of correlational observations and a two-year field experiment in two evergreen shrublands (350 and 500 mm annual precipitation) of central Chile. We planted experimental seedlings of a rather drought-tolerant (Quillaja saponaria) and a drought intolerant (Cryptocarya alba) species at open and shady conditions under different shrub patch sizes (1, 5, 10-15, >30 m in diameter) and at increasing distances from the canopy edge (-5,-0.5, +0.5, +2, +5, +15 m). Half of the seedlings were protected against mammal herbivores. We relate seedling midday stomatal conductance, photosynthetic capacity, growth and survival with the environmental conditions found along those gradients, and compare our experimental results with the natural patterns of seedling recruitment found in these shrublands. Results/Conclusions We found stronger effects of herbivores and canopy conditions (patch size and distance from the canopy edge) at the drier shrubland. There, herbivores reached higher densities and caused higher mortality at open places particularly close to small shrub patches. When protected from herbivores, seedling survival increased under shade. Interestingly though, survival peaked at 2-m distance from the edge decreasing further into the middle and large shrub patches. Under moister conditions (site or year), canopy size and distance effects became less important. Stomatal conductance and photosynthetic capacity closely resembled these patterns indicating that seedling survival was strongly linked to physiological performance under different water and light availabilities. Natural regeneration was found only at the moister site, mostly at shady places close to the canopy edge of small and medium shrub patches. Our results demonstrate a switch in the net effect of facilitative interactions that depends on overall moisture conditions. This switch produces a humped-shaped response curve with a peak in facilitative effects towards intermediate moisture conditions. Our results have important implications for restoration ecology of these semiarid ecosystems. We suggest the combination of herbivore exclusion and the use of nurse plants to facilitate restoration of drier shrublands. Larger chances of success should be close to the canopy edge of existing shrubs. Under moister conditions (years or sites), herbivore protection may be enough.