Exorcising the `segment length curse': summer temperature reconstruction since AD 1640 using non-detrended stable carbon isotope ratios from pine trees in northern Finland

Stable carbon isotope ratios from the latewood cellulose of 12 trees from two sites in northern Finland are used to construct an isotope chronology covering AD 1640 to 2002. By measuring isotopic ratios of every sample independently it is possible to identify and remove the juvenile portion of each δ ¹³C series, correct the individual series for anthropogenic changes in atmospheric carbon dioxide isotopic ratios and concentrations, and to quantify changes in signal strength through time. Most importantly, it is possible to demonstrate that there are no long-term trends in the carbon isotope series that are related to tree age. This means that it is not necessary to detrend the series and so they have the potential to retain climate information at all temporal frequencies. The correlation between the non-detrended carbon isotope series and July/August mean temperature is high (r=0.72) and comparison with meteorological records suggests that the dominant control over tree ring δ¹³C at these high latitude, moist sites is photosynthetic rate rather than stomatal conductance. Summer temperature reconstructions based on three different calibrations are presented, with verification based on a mixture of jacknife and split period designs, providing robust and near identical results. Reconstructed late summer temperatures in the early 1900s are very low but the years centred around AD 1660 and 1760 appear to have experienced warmer summers than the late twentieth century, thus our late summer reconstruction does not show a recent warming trend. Our results are in agreement with other palaeoclimate reconstructions for northern Fennoscandia, which show late twentieth-century warming occurring predominantly in the winter. Our results suggest that, where replication and common signal strength are sufficiently high, stable carbon isotope dendroclimatology may provide high resolution proxy time series that also record climate information at lower temporal frequencies, thus avoiding the `segment length curse' that can apply to palaeoclimate reconstructions based on other tree ring parameters such as ring widths and density.