Slashdot Mirror
Flu Epidemics Coincide with Sunspots
Croatian Sensation writes "According to this article flu epidemics are four times more likely when sunspot activity is high. They've analyzed records going back to 1729 and indeed, the correlation is statistically significant. Neat." The research seems a little... thin.
Yeah, what he said.
BTW, what's "holocost"? Is that the admission price for a turn in the Holodeck?
There is a well-defined mathematical procedure for finding out if a correlation like this is statistically significant or not. To do this, assume the things are not related and calculate the probability that these results would be obtained anyway. Say the probability was 0.001, then the result is statistically significant at the 99.9% level.
This is meaningful.
The research seems a little... thin.
2% chance of it being random? That's not exactly thin. Though of course every slashdot story has to end with someone casting doubt on the story subject, since everyone seems deathly afraid to be thought of as gullible or naive...
--
If sunspots do influence weather by heating the Earth in uneven ways then that might create certain patterns of pressure fronts. IIRC, sunspots are locations of hotter sun activity so presumably the Earth would get more high pressure zones? But if that were true how would this affect the spread of influenza? Isn't influenza spread during cold seasons and wouldn't this require low pressure zones as opposed to high ones?
Another thing to consider is that influenza seems to originate with birds in China, Australia and some other places that I can not recall at the moment. So somehow the way weather is affected by sunspots causes the migration of these disease infected birds to spread to more populous areas thus infecting more people?
"sweet dreams are made of this..."
Is it possible that the occurence of a large number of sunspots increases the mutation rate of the flu virus, therefore causing an outbreak of the disease as no one's immune to it yet?
I adblock all animated gifs.
Blessed be the prime numbered slashdotters
OK,
- B
--
http://www.bradheintz.com/
- updated
The problem is, this ignores selective reporting. Why are we looking at this particular point of data? Because our attention is already drawn to the statistical clumping around sunspots. We assume there is going to be something found there, because of such things as tree-growth rings and whatnot. If the clump had to do with people's heights varying every ten years, our clump-seeking brains would have been drawn the fact to our attention. Our compluation of the significance level tacitly excludes many other factors that DO NOT clump.
The human brain filters vast quantities of data, seeking things that appear unusual, and only then does it send out a conscious signal: "Wow! Look at that!" The wider we case our pattern-seeking net, the more likely it is to catch a clump.
People used to think that a woman's cycle corrsponded with the phases of the moon, because they were roughly 28 days. However, they are NOT exactly 28 days, and a Gibbon has a much shorter period while a mountain gorilla has a longer one.
People look for and expect to find patterns (even such things as the shape of the pyramids in Giza having astronomical relevence in their proportions. We expect to find patterns, and we find them, but that doesn't mean it's significant. It might just be a statistical clump.
Fuzzy Knights: New RPG Strips Tuesday and Friday!:
http://www.fuzzyknights.com
No, no, no... what *I* want to see is evidence that all those millions of people being sick somehow induces sunspots...
Imagine that... "Cover your mouth when you sneeze, Johnny, or you may cause another sunspot."
TheNewWazoo
Some folks here that don't think there could possibly be a connection, that this study is just an example of the abuse of statistics. But I'd like to bring up a point.
I can't think of a possible physical mechanism that directly links sunspots and influenza. But how about an intermediate step? Sunspots influence weather. And weather... influences the transmission of influenza. Influenza is a seasonal disease, after all.
A slight change in the climate wouldn't by itself trigger an epidemic, and it probably wouldn't stop one -- epidemics/pandemics are probably triggered more by antigenic drifts/shifts. But it might ever so slightly change the odds that a new strain will have the chance to gain a foothold, maybe enough to be statistically detectable.
And who knows? Maybe there's even a linkage with the antigenic changes as well. Suppose that, like in humans, these weather changes are influences the transmission of animal strains of influenza in wildlife and farm animals too. It's thought that the mixing of strains from different species (Mainly Human/Avian/Porcine) in a host susceptible to more than one variety (Like pigs) is what drives antigenic shift -- which gives us epidemics. Hmmm...
They're talking about both parallel and concentric layers of gas: the concentric ones are the outer convective layer, the inner radiative layer, and the thin shear layer between them, known as the tachocline. The convective and radiative zones rotate at different speeds (not "opposite directions!"), while the tachocline changes speed periodically; the speeds of the layers above and below the tachocline also change periodically, but in opposite directions (the changes in speed are in opposite directions, meaning one speeds up while the other slows, not opposite rotations) -- which implies that the tachocline is oscillating.
While the radiative zone rotates essentially as a solid body (despite the fact that it's actually a highly-compressed plasma), the convective outer zone doesn't. In fact, the polar regions of the convective zone have a one-year oscillation coupled to the tachocline, while the equatorial regions have a 1.3-year oscillation. These, I think, are the "parallel layers" from the article.
What's entirely unexpected about this is the period: everyone thought it would be connected to the 11-year sunspot cycle, but instead there are two separate periods, 1.0 and 1.3 years, neither of which has any obvious relationship with the sunspot-cycle period. Once again, we find that the simple models aren't a great match for reality -- and science is nowhere near the end of its search for understanding of the universe. (Which is a good thing!)
As for flu epidemics, I am not educated.
--
My ID is in the 200k's.