Climate Reconstruction Using Tree-Ring Data

Instrument records from the past 200 years support a scientific consensus that climate is changing Trenberth et al. (2007). What is less clear is how to interpret these changes in an historical context spanning many hundreds, or thousands, or tens of thousands of years. The statistical question is this: how do we quantify and model historical climate without having direct observations beyond the most recent one or two centuries? The most common solution is to build an index of historical climate change using long-scale proxy observations that are constructed from data such as tree ring measurements, lake sediments, bore-hole observations, coral records and ice cores (see Jones and Mann (2004) for a review as well as Jansen et al. (2007)). Reconstructions based on these sources feature prominently in the 2007 Intergovernmental Panel on Climate Change report (Jansen et al. 2007). There are two steps necessary to reconstruct climate from proxy data. Step one involves filtering out any non-climate features so that only climatic features remain in the proxy. Step two involves using the filtered proxy data, either alone, or combined with other proxy measurements, to reconstruct climate at either the local or global scale. While some studies have not paid enough attention to the first ∗E-mail: [email protected]