What do trees remember? Disentangling climate, biological, and ecological signals from tree-ring time series

Abstract

Dendrochronology, the study of tree rings, uses variation in the annual growth patterns of trees to make inferences about the past. Most commonly used as records of past climate, tree ring analysis has built our understanding of how the climate has varied during the Common Era. Tree rings are also physical records of tree growth, and are often used in ecosystem science to understand how climate and environment influence the health of forests. In order to use tree rings to infer past changes in environment and ecosystems, researchers must content with the complexity of exogenous (climatic and environmental) and endogenous (biological) signals contained within annual tree rings. A range of statistical and data processing techniques are used to filter tree-ring time series to extract specific signals and discard unwanted sources of variance, but these techniques are imperfect and may result in biased estimates of the relationship between climate and tree growth through time. In my dissertation, I evaluate how assumptions made about the sources of variance contained within tree rings, as well as choices made during data development and analysis, influence the outcome of tree-ring studies. I explore these themes through three separate case studies, described in four chapters. In my first chapter, I compared different data processing methods commonly used by dendrochronologists to determine their effects on final chronologies. In my second and third chapters, I investigate the relationship between drought and radial growth patterns, and explore whether drought should be thought of following a classic ‘disturbance’ model, or as an ongoing climatic phenomenon. In my final chapter, I analyzed whether evidence of anthropogenic warming is visible in tree-ring series from Minnesota – a region where trees are not typically considered sensitive to variations in climate.