Figure 12

Scatter plots for the predicted 2-m temperature anomalies (aT2) or EHF against circulation indices: BL, AL, AZH, and NAO (from left to right). In each panel, the box and whisker plots show the lower quartile (Q1), median (Q2), upper quartile (Q3), and the outliers are determined by the inter-quartile range (Q3–Q1) for plotting. For each box and whisker plot, sample was drawn from the daily forecast values and stratified according to 4 quartile-based bins, namely the values within the ranges of minimum–Q1, Q1–Q2, Q2–Q3, and Q3–maximum, of aT2 or EHF for a selected period. The vertical box and whisker plots are for aT2 or EHF, and the horizontal box and whisker plots are for the corresponding circulation indices. (a,b) For 2-m temperature anomalies (aT2) and EHF, respectively, against the 4 circulation indices for SC-based-binning. (c,d) For WE-based-binning. Only the first and fourth quartile-based bins are shown in (a,b,d) for brevity, and denoted by blue and red colors, respectively. All 4 bins are shown in (c), with green and orange denote the second and third quartile-based bins. The solid and hollow boxes denote the forecast values for lead-week 1 and 4, respectively. The statistical analysis was conducted for the key period of 12 July–8 August, 2018.