| Summary: | Specialization: Forest Biology and Management Degree: Master of Science Abstract: Given that efforts to reduce greenhouse gas emissions appear to fall short, we need to develop adaptation strategies to mitigate anthropogenic climate change impacts, which may include predicting and managing no-analog ecological communities, for example by selecting suitable species assemblages for ecological restoration projects or in reforestation. To develop analytical tools for the future, this study looks at the past to test if no-analog communities can be detected in the North American fossil pollen record, if they were associated with no-analog climate conditions, and if they can be predicted with species distribution models. The working hypothesis is that no-analog communities that are not associated with no-analog climates, but result from differential migration rates will be the most difficult to predict using species distribution modelling (SDM). This study contributes a systematic analysis of 51,427 dated pollen community records from 1,384 sites from the Neotoma Paleoecology Database with climate hindcasts for 21000, 16000, 14000, 11000, 6000 years before present from the CCM1 general circulation model. The analysis relies on a multivariate regression tree approach that delineates pollen communities but constrains the clustering of similar pollen records by using climate variables as partitioning criteria. Modern pollen records are also used to train a species distribution model and create hindcasts up the last glacial maximum to test if past no-analog communities can accurately be predicted. The analysis revealed three pollen communities that were widespread during the last glaciations or during deglaciation but that have no equivalence in today’s pollen record. The most distinct no-analog community was a birch parkland ecosystem with exceptionally high frequencies of Betula, Poaceae, Cyperaceae and forb pollen. This ecosystem also occupied no-analog climate space that was similar to today’s arctic tundra with high ...
|
|---|