Abstract
Following the retreat of the North American glacier at the beginning of the Holocene (around 11,700 years ago), plant colonization took place more rapidly than predicted, compared with migration rates estimated from the current average dispersal capacities of several trees. This lag between observed and estimated migration (Reid's paradox) has been observed in western Quebec, where conventional hypotheses to resolve it cannot be applied. Indeed, no type of proglacial refuge has been identified, and the presence of the large proglacial lake Ojibway between 10,570 and 8,200 years before present (cal. BP) prevented any plant colonization or long-distance dispersal event. Yet, floodplain afforestation was almost straightforward after the drainage of proglacial Lake Ojibway. In this doctoral thesis, the ancient islands of proglacial Lake Ojibway (the paleo-islands) were studied to determine whether they may have acted as colonization outposts and thus accelerated the process of plant colonization. Analysis of lake sediments from two paleo-islands in the Abitibi region confirmed the establishment of the first trees on the paleo-islands around 9500 and 9850 cal. BP, before the drainage of proglacial Lake Ojibway began. A multi-proxy analysis (pollen, macro-remains and sedaDNA) of the sediments shows the progressive establishment of vegetation on the paleo-islands, with a succession of non-arboreal tundra phases, afforestation and diversification before the establishment of the boreal mixedwood forest around 8200 cal BP. Despite common successional phases, the vegetation on the two sites differs, probably in response to different edaphic conditions and fire dynamics. Vegetation dynamics are linked to climate variations. Finally, the comparison of different colonization scenarios and their associated migration rates confirms that the paleo-islands did indeed play a role as outposts of colonization. They may have facilitated and accelerated post-glacial colonization, thus resolving Reid's paradox in western Quebec. The study of paleo-island distances and surfaces suggests that their colonization followed the theory of island biogeography. This thesis contributes to the advancement of our knowledge of postglacial colonization mechanisms, and also discusses the role of paleo-islands in present and future landscapes, and possible perspectives for study.