Using Co-Located Lake and Bog Records to Improve Inferences on Late Quaternary Climate and Ecology

Abstract

Records of past climate and ecological dynamics from lakes and bogs provide a long-term perspective that extends the modern observational record. Many different paleoclimatic and paleoecological proxies are used to characterize patterns, processes, and impacts of past climate variability, each with unique strengths and weaknesses. Integrating data across a network of different proxies requires a detailed understanding of each individual proxy. In the eastern United States two common proxies for Holocene hydroclimate conditions are testate amoeba-based water-table depth reconstructions and sedimentary lake-level reconstructions. Testate amoebae tests are preserved in sediments and can be identified to the species level. The community composition of testate amoebae on the surface of an ombrotrophic bog is sensitive to the surface moisture of the bog. Past fluctuations in lake level can be tracked by the changing elevation of the sand-mud boundary along a transect of near-shore sediment cores. Lake-level records have been developed mainly in southern New England and in the Rocky Mountains. Bog water-table depth records have been developed mainly in Maine and the Upper Midwest. For my dissertation, I have developed two pairs of co-located lake-level and bog water-table depth records from Maine and Minnesota. Co-located records allow direct comparison of proxy records that experienced the same past climate conditions. In the first chapter, I present a new testate amoeba record from Caribou Bog and a new Bayesian transfer function model that estimates water-table depth from testate amoeba assemblages. I compare seven different methods for estimating water-table depth from testate amoebae in cross-validation and in reconstruction. In the second chapter, I present a suite of records from Caribou Bog and Giles Pond, both from south-central Maine. I validate and compare the hydrologic reconstructions from Caribou Bog and Giles Pond. Then, I leverage the strengths of each proxy reconstruction to produce an integrated record of past hydrology that captures hydroclimate variability and change from decadal to multi-millennial timescales. Finally, I use the integrated record of past hydroclimate to understand vegetation changes in a new pollen record from Caribou Bog. Holocene vegetation changes represent a complex interaction between large-scale biogeographic patterns, regional and local climate variability and change, and local ecological variability. In the final chapter, I introduce new records from Ely Lake Bog and Knuckey Lake in northeastern Minnesota. I analyze the new records in the context of mid-to-late Holocene changes in the Upper Midwest and mid-latitude North America. I find consistent sub-regional drought events over the past 2000 years and coherent multi-centennial variability in the eastern United States over the past 6,200 years. This overall combination of formal statistical modeling, detailed site-level reconstruction, and single-proxy network analyses represents ways to maximize the utility of proxy records for inferring past environmental variability and change.