Environmental filtering of regeneration patterns in a forest dynamic mosaic in south-central Chile

A trade-off between growth in high light and survival under shade is widely regarded as the main driver of secondary succession in humid forests. Outside the humid tropics, however, the microclimates of large openings may select for traits other than those that maximize growth rates. We tested for differential filtering of sapling functional traits by four different microenvironments in a south-temperate forest in south-central Chile. We measured light, temperatures, humidity, soil nutrients and sapling densities in each of four phases of a forest dynamic mosaic: shaded understoreys, tree-fall gaps, Chusquea bamboo thickets, and clearings. We then measured leaf, wood and reproductive traits, as potential predictors of species' regeneration patterns. Clearings were exposed to more than twice as many frost days as any other forest phase, and to much larger vapour pressure deficits in summer. Clearings were compositionally distinct from the other three phases, which were indistinguishable from each other. Saplings of 6 out of 20 common arborescent species were significantly associated with one or other of the four phases. These associations were best predicted by a three-trait model combining seed mass, leaf dry matter content, and % winter leaf retention; though this model explained only 10% of observed variation. No species was significantly associated with bamboo thickets. Community-weighted trait means showed that large seeds were the most distinctive trait of understorey plots, whereas bamboo thicket plots had the highest average wood density, and winter leaf retention was significantly lower in clearings than in any other forest phase. Despite the abundant recruitment of large-leaved (semi)deciduous species in clearings, several evergreens with small leaves (<3.5 cm2) also established primarily or exclusively in clearings. Although this study confirms the strong environmental filtering imposed by the microclimates of clearings in some temperate forests, it reveals more functionally diverse responses to clearing environments than was reported in a parallel study in New Zealand. Smaller sample sizes may partly explain why overall we found less evidence of regeneration niche differentiation and environmental filtering than in the New Zealand study, although attenuation of filtering by vegetative reproduction may also have contributed.