Extended Data Fig. 1: Evaluation of the ability of the A-FEN and X-FEN reconstructions to capture extremes in climate targets.
From: Fennoscandian tree-ring anatomy shows a warmer modern than medieval climate
The instrumental temperature data was sorted from coldest to warmest and plotted together with the reconstruction values of the corresponding years. The grey boxes are bound by the 10% coldest and warmest years, and the 10th and 90th percentile of the zscore temperatures, respectively. If an extreme reconstruction value is found within the grey box, the extreme is defined as “captured”. The sum of the captured values divided by the potential sum of values, is calculated and presented as a percentage of extreme value capturing (EVC). In McCarroll, et al.21, a significance testing was implemented, and for 160–170 years of climate data, p < 0.001 is achieved if more than 40% of values are captured. a) A-FEN’s ability to capture JJA temperature extremes. b) X-FEN’s ability to capture JJA temperature extremes. c) and d) show same analysis as a) and b) but using the target MJJA. Both datasets thus display significant amounts of extremes captured, but the A-FEN captures significantly more than the X-FEN for the MJJA target season. The X-FEN captures a higher percentage of cold extremes if the MJJA target season is used but the same percentage of warm extremes regardless of target season. The rationale for using MJJA as the target season for X-FEN is thus less clearcut than for the A-FEN. JJA is the target season used in the publications originally presenting the MXD data2,37 and is thus used in the main text for the other comparisons.
