Seascorr: A MATLAB program for identifying the seasonal climate signal in an annual tree-ring time series

A common research task in dendroclimatology is identification of the monthly or seasonal climate signal in an annual time series of indices of ring width. A MATLAB function, seascorr, is introduced as a general statistical tool for identifying the signal. Monthly time series of primary and secondary climate variables are input to the function along with a tree-ring time series and specifications for seasonal groupings. The relationship of the tree-ring series with the seasonalized primary climate variable is summarized by simple correlations. The relationship with the secondary climate variable is summarized by partial correlations, controlling for the influence of the primary climate variable. Confidence intervals on sample correlations and partial correlations are estimated with the help of Monte Carlo simulation of the tree-ring series by exact simulation, which preserves the spectral properties of the observed series. Results are summarized in graphical and statistical output. The function is illustrated with examples from Tunisia and Russia.