A rapid method for the production of robust millennial length stable isotope tree ring series for climate reconstruction

We developed a millennial length tree ring delta C-13 chronology from Pinus sylvestris at a site known as Laanila, in northern Finnish Lapland. In order to measure this series rapidly and with adequate replication, we adapted a multiple year pooling system. Using a combination of offset cut 5-year blocks, plus a novel method of joining tree cohorts together without the loss of a reliable absolute mean, allowed for the preservation of low frequency information in the series. Our 'Join-Point' method addressed systematic error in delta C-13 between trees and allowed the development of a robust millennial length chronology with a considerably lower number of measurements when compared to annual sampling. Our pooled chronology retains a replication at each year of n = 5 such that it is produced from just similar to 1000 mass spectrometry measurements. The Join-Point verification of the absolute delta C-13 value, at periods where cohorts of different aged trees meet, requires the analysis of a much larger number of trees at 8 Join-Points. We term this methodology offset-pool plus Join-Point, and describe the characteristics of the series it produces. Use of this method required an assessment of age-related trends in delta C-13 at the site and an investigation of potential differences between the living and sub-fossil wood material used, prior to the implementation of the method. With the assumption that delta C-13 series produce robust climate reconstructions, the emphasis has shifted to producing well-replicated series for palaeoclimate purposes. With this aim the analytical precision of a single measurement is considerably less important than the confidence attached to the mean value for each time step. The most efficient way to increase the precision of the estimate of mean delta C-13 is by increasing replication. Here we have demonstrated a novel method for doing this without the analysis of an unrealistic number of measurements and without the offset in delta C-13 between trees causing error to accrue back in time. (C) 2011 Published by Elsevier B.V.