Statistical analysis of series of dendrochronological indices

We analyze series of tree-ring indices to understand whether they can be indicators of long-term climate changes. We calculated the synchronous cross-correlation coefficients (CCs) for all possible pairs composed of nine series, of which only one CC (for a pair of the longest series) proved to be sufficiently high (0.72). We revealed a time interval for which this CC is equal to 0.88. We failed to locate sufficiently larger time intervals with a higher CC for other pairs of series. Similar calculations for a series using a low-frequency filter (with allowance for periods of 100 or more years) led to a noticeable decrease in almost all CCs. This confirms the statements circulating in the literature that tree-ring chronologies include climate signals with characteristic times of several decades and that the secular and supersecular oscillations are significantly distorted as a result of standardization and cross-linking. The results presented in this paper show that the complex of series being considered cannot be used as a climate-change indicator without an additional analysis. Here, we describe a possible procedure for such analysis. As a result, we found that only two time series at some time subinterval can provide useful information about climate changes.