Diaz, J., Garcia, R., Lopez, C., Linares, C., Tobias, A. And Prieto, L. 2005. Mortality impact of extreme winter temperatures. International Journal of Biometeorology 49: 179-183.
What was done
The authors examined "the effect of extreme winter temperature on mortality in Madrid for people older than 65," using data from "1,815 winter days over the period 1986-1997, during which time a total of 133,000 deaths occurred."
What was learned
First of all, Diaz et al. report that daily Tmax was more closely correlated with mortality than was daily Tmin, because, as they describe it, "very low Tmin occur mostly during stagnation episodes, characterized by very cold nights and sunny days, with a typical temperature range of between 15°C and 20°C," while "most of the days with very low Tmax occur under cloudy conditions, with very limited temperature ranges of around 5°C," so that "human exposure to low temperatures during these days is longer than that occurring during the stagnation days associated with a very low Tmin." In addition, they note that "Tmin is usually recorded around 7 a.m., when very little human activity occurs outdoors, while Tmax is usually recorded at around 4 p.m." Hence, they conclude that "both the time of occurrence and the associated synoptic situation should explain the higher correlation exhibited by Tmax."
Second, when total mortality was considered, the maximum impact occurred about 8 days after a low temperature extreme, while for deaths due to circulatory diseases the lag was 7-14 days, and when respiratory causes were considered, two mortality peaks were observed, one at 4-5 days and one at 11 days after the low temperature event.
Third, a plot of daily mortality from all causes at a lag of 8 days after the occurrence of an extreme Tmax "pointed to a change of slope at a Tmax of 6°C," so that "an unusually cold day (UCD) can be defined as one in which Tmax is equal [to] or lower than 6°C." Consequently, as Tmax dropped below this value, "the impact on mortality also increased significantly." Diaz et al. note, in this regard, that the average winter mortality associated with UCDs (measured 8 days after the UCD occurred) "was 106.2 deaths/day, while for the other days it was 96.1 deaths/day." In addition, they found that the impact of UCDs increased as the winter progressed, with the first UCD of the season producing an average of 102 deaths/day at a lag of 8 days and the sixth UCD producing an average of 123 deaths/day at a lag of 8 days. This latter observation suggests, in their words, that "acclimatisation does not occur, with every cold spell enhancing the pathologies produced in previous spells." Hence, whereas Diaz et al. report that "the impact of heat waves is reduced as they occur during a certain season, suggesting an acclimatisation to heat," just the opposite occurs in the case of recurring cold, which becomes ever more deadly with each new occurrence.
What it means
(1) Cold kills, (2) extreme cold kills even better, and (3) recurring extreme cold kills best of all, even in a climate as moderate as that of Madrid.