Adjustment for proxy number and coherence in a large-scale temperature reconstruction

Abstract

Proxy records may display fluctuations in climate variability that are artifacts of changing replication and interseries correlation of constituent time-series and also from methodological considerations. These biases obscure the understanding of past climatic variability, including estimation of extremes, differentiation between natural and anthropogenic forcing, and climate model validation. Herein, we evaluate as a case-study, the Esper et al. (2002) extra-tropical millennial-length temperature reconstruction that shows increasing variability back in time. We provide adjustments considering biases at both the site and hemispheric scales. The variance adjusted record shows greatest differences before 1200 when sample replication is quite low. A reduced amplitude of peak warmth during Medieval Times by about 0.4°C (0.2°C) at annual (40-year) timescales slightly re-draws the longer-term evolution of past temperatures. Many other regional and large-scale reconstructions appear to contain variance-related biases. Copyright 2007 by the American Geophysical Union.