Canadian Journal of Forest Research · 1992 · 13 citations · 0 references
ComputingEarth SciencesData AnalyticsPaleographyTime Series AnalysisAutocorrelation AnalysisQuaternary ChronologyTrend AnalysisAutoregressive Moving Average ModelsCross-correlation AnalysisInput Data AnalysisSignificance CriteriaSignificance LevelStatistical Significance TestingStatistical Significance ThresholdsTree-ring Records
CORREL is a FORTRAN program that employs cross correlation to (i) determine potential cross-dating (matching) positions for "floating" (undated) ring series; (ii) detect missing or false rings; and (iii) estimate the statistical significance of potential dating positions. To work properly, CORREL input data must be detrended and modeled using the autoregressive moving average procedure. To guard against spurious dating, the output's best date should be checked for dating consistency. The significance level of the best date is obtained by adjusting its single-dating-trial significance for multiplicity (repeated dating trials). Ideally, COREL should be used with the detrending tree-ring programs ARSTAN or INDEX, and with the data quality-control program COFECHA.