FLASHBACK POST- DECEMBER 23, 2009! At least it would be easier to believe…
Record cold temperatures have arrived to the United Kingdom, Canada (24 consecutive days below -24 °C in a city). Cold Siberian air has also hit Central Europe, France, and Italy. London is colder than Antarctica. Literally. The cold snap is costly.
The temperature in Pilsen and Czechia in general keeps on oscillating around -10 °C, too. The snow around is clean and pretty. The coldest official Czech weather station, Stráž pod Ralskem, has seen -25.1 °C.
Poor people in chilly India solve the situation by burning books; at least 55 people have died. I hope that they have enough copies of An Inconvenient Truth, like in Belgium (I invented the joke before them!). Sorry, the picture above are commies in a warm weather, not poor people in a chilly weather.
Giraffes in a chilly Chinese zoo discovered fire as a way to get warmer.
Cold weather and global warming
A frequent question is whether the repeated record low temperatures imply that global warming either doesn’t exist or it is not serious. Of course that they do. Assume that the temperature in your city is a linearly increasing function of time, “Temp + Slope x Time”, plus fluctuations that are randomly distributed with the standard deviation “SD”.
If you get the opportunity to prove that the linear trend in your city exists at the five-sigma confidence level, i.e. that the net warming since 1900 or so has exceeded five times the natural oscillation “SD”, then it also means that the probability that you get a cold extreme for a certain day will be dropping faster than exponentially: like the Gaussian.
Assuming that the “systematic” global warming accumulated by the linear trend exceeds five times the noise “SD”, which is really necessary for proving that the linear trend in your city is more than noise according to the physics standards, the probability that you measure a new cold weather record should drop roughly one million times (!): check basic articles about the normal distribution and how large fraction of a Gaussian lies below minus five sigma. It is less than 1 part per million.
Such a dramatic decrease of the frequency of record cold temperatures is clearly not happening because the record cold temperatures seem to be as frequent as they were in the past. More precisely, their frequency should be naturally decreasing with the growing temperature records (with time).
If you study the global mean temperature rather than the local temperatures, you obviously increase the signal-to-noise ratio because a large part of the noise gets averaged out (but not all of it). However, at the same moment, the conclusions derived from the global mean temperature are also much less relevant for any particular city in the world.
Every city should actually care about their own temperature, because they can only be affected through it. And for the local temperature, the accumulated warming from the linear trend is much smaller than the unavoidable interannual fluctuations which is why the underlying warming trend is completely irrelevant for every single rational human being in the world.