4 ms·
This doesn't seem newsworthy outside Korea. We see a lot of headlines like this - are they indicative of climate change? There are twelve months in the year an
by dmurray 2y ago
This doesn't seem newsworthy outside Korea. We see a lot of headlines like this - are they indicative of climate change?
There are twelve months in the year and at least four directions to report weather extremes (hottest, coldest, wettest, driest - there's also windiest and snowiest but let's lump them in with the other extremes). That's 50 chances per year for any given area to have its "most X month on record". Reliable weather records go back 100 years, give or take, so any place should experience its "most extreme month on record" approximately every two years.
There are 50 cities as big as 0.75 Seouls [0] or bigger. So worldwide, "megacity has 100-year weather event" comes up every two weeks.
Expand the number of weather axes, relax the definition of a megacity, or realise that many of these places only have 50 years of good record-keeping...pretty soon you can have this headline every two days.
I'm not trying to downplay climate change. I'm just saying you won't find it in these headlines. If you look at the number of places with the most extreme year on record, or model average air or sea temperatures over time, you can find more compelling evidence there.
[0] https://en.m.wikipedia.org/wiki/List_of_largest_cities https://en.m.wikipedia.org/wiki/List_of_largest_cities
- tzs 2y agoNote that if the climate at a place isn't changing we'd expect new records for a given thing to become less frequent over time. For example suppose you had something where it follows a normal distribution with a mean of 100 and a standard deviation of 5. I generated 10 million random numbers with that distribution and noted whenever I got a new record high. The intervals between records were 1, 1, 2, 1, 5, 36, 219, 999, 2558, 21780, 397, 50360, 73288, 200951, 652234, 954240, 547715, 3783984, and 1641056.
- dmurray 2y agoThis is both completely true and completely nonsense in the context. If the numbers are independent, no matter what the distribution, the odds of the nth measurement being a record setter is 1/n. And as above, if you've measured 100 Novembers, the odds of the 100th being the record wettest are 1/100.