Extended Data Fig. 2: Individual tree age is a better metric than mean stand age for examining age-dependent drought responses.
From: Younger trees in the upper canopy are more sensitive but also more resilient to drought
Comparison of Quercus alba drought responses at 18 sites between composite chronology with mean stand age (a) and individual series with individual tree age approach (b) using the same dataset as in Au et al.16. Hence, the same data can lead to very different results due to diluting effect of aggregating data at the coarser stand level. Here, we advocate for analysing the raw data on individual tree-ring time series rather than analysing the aggregated the information at the stand level for age-dependent drought responses. Shaded ribbon in panel a indicates the 95% confidence interval for prediction from a linear model. The age cohort classification in panel b follows the cutoff age for Quercus alba listed in the Supplementary Table 12. The numbers at the top of panel b represent the p-values of pairwise differences in percentage of growth reduction between age cohorts that were identified by Tukey honest significant differences. The numbers at the bottom of panel b represent number of tree individuals for the youngest, intermediate and oldest age cohort of Quercus alba, respectively. Boxes show the interquartile range (IQR) while upper and lower whiskers are defined as the third quartile (Q3) plus 1.5×IQR and the first quartile (Q1) minus 1.5×IQR, respectively. Values that are less than Q1–1.5×IQR or greater than Q3+1.5×IQR are plotted as closed circles. The bold lines and open squares in the boxplot represent the median and the mean values, respectively.
