EXPLORING PATTERNS IN DENDROCHRONOLOGICAL DATA THROUGH CLUSTER ANALYSIS

Resumo

This study employed the dendrogram methodology to analyze time series data obtained from measuring tree growth rings. A total of 64 samples were collected from 21 individual trees. Polynomials were applied to filter the natural growth pattern of the trees and enhance the impact of external factors, such as climate influences. Cluster analysis using Ward's minimum variance and Euclidean squared distance was utilized to group the data based on similarity. Three dendrograms were constructed, consisting of 10, 47, and 64 samples, respectively. The analysis revealed that the samples with the highest correlations, encompassing over 95% of the total samples, formed homogeneous groups. Pearson correlation was also employed to confirm the results obtained from the dendrograms. Consequently, it can be affirmed that the most suitable samples were utilized in constructing the average chronology from the available data.