Coexisting responses in tree-ring δ13C to high-latitude climate variability under elevated CO2: A critical examination of climatic effects and systematic discrimination rate changes

Abstract

Dendroclimatic methods may quantify a factor or factors affecting δ13C fractionation and thus help to reveal the signals archived in dendroisotope records over the long term. Gathering evidence suggests a repertoire of limitations for photosynthesis to occur, with an increasing possibility that more than one external factor may control the assimilation rate and thus the tree-ring δ13C variations. Here we show that such a situation conceivably describes the tree-ring δ13C data from northern timberline and further illustrate the use of dendroclimatic analyses in separating the coexisting signals in dendroisotope data. While the assimilation rate was primarily controlled by a photon flux, thereby allowing the tree-ring δ13C to provide a proxy for past variations in irradiance, there was some evidence that also a temperature signal is directly present in the tree-ring δ13C data, not merely as a function of an indirect correlate reflecting its interplay between sunshine/cloud cover. Over the period common to all instrumental records (1971–2011), both the sunshine hours and global radiation influenced the δ13C from mid-June to mid-July, whereas the mean maximum temperatures (TMAX) showed an impact on δ13C from mid-July to mid-August. We assume that these climatic associations represent mainly non-stomatal limitations to assimilation rate. Possibly, this response may involve the mesophyll conductance to CO2 transfer from intercellular spaces to chloroplasts, a factor found previously to pose a temperature responsive limitation to photosynthesis. After correction for the Suess effect, the δ13C chronology exhibited a long-term decline attributable to discrimination rate changes under elevated atmospheric CO2 concentration. We make a methodological contribution by comparing the various methods available from literature for estimating the magnitude of this bias in the δ13C chronology. The robustness of the results indicated that this data shortcoming is not critical but can be corrected. The method inter-comparisons yielded very similar results, near the previously suggested change in discrimination of 0.0073‰ per ppmv CO2. We recommend comparative assessments of discrimination rate change to be combined with dendroclimatic analysis.