| The main objective of the present dissertation is to investigate past and recent forest-atmosphere interactions across biomes. A review of the literature and analyses of available databases show that in extratropical zones forest expansion was the dominant response to climate warming during the early to mid-Holocene (12,000-4,000 YBP). Tropical forests started to expand after this period, possibly in response to wetter conditions. Recent changes in the atmosphere are expected to promote forest expansion worldwide, but the effect of this process on global climate remains unknown. As a result of extensive palynological studies, forest expansion in temperate and boreal regions is well understood, but in the tropics alternative techniques are needed to elucidate mechanisms of expansion. Using soil organic matter isotope composition I show that after millennia of stability forest borders and patches have continuously expanded into grasslands in southern Brazil. Soil-plant feedbacks appear to be involved in this process, as colonizing trees increase soil fertility and microbial activity, promoting the establishment of other tree species. Nevertheless, expansion in this region has been orders of magnitude slower than historically observed in temperate and boreal regions, which could possibly be explained by competition, dispersal limitations and disturbances. Dendrochronological and isotopic analyses reveal that the colonizing tree Araucaria angustifolia has been affected by recent atmospheric changes. Systemic increases in water use efficiency occurred in response to rising atmospheric CO2, but warming-induced stress has caused tree growth decline. Similar results are observed for two deciduous (Quercus rubra and Acer rubrum) and two conifer (Picea mariana and Pinus resinosa) species in temperate and boreal forests of Canada. Despite increases in water use efficiency over the past century, tree growth decline has been the prevalent response. Basal are increment has been negatively correlated with temperature and positively correlated with precipitation, suggesting warming-induced water stress. Relationships between water use efficiency and basal area increment through time can be used to anticipate the effects of atmospheric changes on forests worldwide. Dendrochronological and isotopic measurements can be used with this purpose, but future studies should account for intrinsic autocorrelations between water use efficiency and atmospheric CO2 levels. |
