From observation (period 1961–1990), a threshold for an extremely warm summer could be set to 18 °C. Using that same measure of extremity for a different situation, namely the climate scenario for the period 2071–2100, nearly all summers would have to be called extremely hot. This might be true in an absolute way, but it might also be useful to analyze the temperatures relative to the climate of the period in which they occur by using the concept of rarity or relative extremity. In this case, a summer in the period 2071–2100 would probably not be called extremely hot until its mean temperature reaches 24 °C. The red line indicates the mean summer temperature in the summer of 2003. The historical distribution cannot explain this anomaly and it is very plausible that the parameters of the distribution have changed already.
Another discussion is that concerning extremity in the sense of variability. This interpretation aims at characterizing the whole distribution instead of classifying a single event as extreme or not extreme. From a statistical point of view, a distribution is extreme if the observed values vary a lot, that is if the standard deviation is high. When two distributions are compared, the variability is useful since it combines information about the distribution in a single number, the standard deviation. When comparing two distributions with the concept of absolute extremity, statements about changes in the extremity can generally only be made with respect to a specific threshold value. A distribution could be called absolutely more extreme (for high extremes) if the likelihood of excess increases for high threshold values. Within the concept of relative extremity, statements about its changes depend on the choice of the quantile. As a measure for the relative extremity of a distribution one could use the distance of the quantiles to the median. A distribution is relatively more extreme (for high extremes) than another when the distance of the upper quantiles to the median increases. In Fig. 4, three situations of more extreme temperatures are sketched. For the explanation it shall be noted that the focus is on the high extremes.