kalen
Your Daddy
- Joined
- Mar 11, 2004
- Messages
- 933
-
Security incident: ISF was recently accessed by intruders. Please change your password, and change it anywhere else you used it. Read more
Merged Group measures radioactive decay rates changing with seasons - blames sun
Is this the same discovery dot com that's affiliated with the Discovery Channel? The same people that pitch all sorts of wooish nonsense as valid science? The same channel that features a show called "Ghost Lab" that solicits viewers for their own "ghostly" experiences? The show that features three over-acting yokels "... as they travel the country using some of the world's most sophisticated equipment to gather powerful supernatural research"?
That Discovery channel?
Pffft! Sheer balderdash! BALDERDASH, I say!
That Discovery channel?
Pffft! Sheer balderdash! BALDERDASH, I say!
Last edited:
Madalch
The Jester
- Joined
- Nov 17, 2006
- Messages
- 9,763
Andy_Ross
Penultimate Amazing
- Joined
- Jun 2, 2010
- Messages
- 86,664
a good thread on this is already well underway on the Bad Astronomy & Universe Today Forums The-strange-case-of-solar-flares-and-radioactive-elements
Skeptic Ginger
Nasty Woman
- Joined
- Feb 14, 2005
- Messages
- 96,955
I'm going to wait for the guy's data to be replicated. He is an older scientist (could have an out of date education) he doesn't have a lot of published papers (he does have a book), and he did have a UFO hypothesis that wasn't very mainstream. Not that any of this discredits him as a scientist. He does have a teaching position and proper credentials.
I could see solar flares speeding up radiometric decay. But I also wonder why it hasn't been detected before. An awful lot of people have verified the decay rate of an awful lot of elements. One would think discrepant results would have come up and been checked and rechecked.
I could see solar flares speeding up radiometric decay. But I also wonder why it hasn't been detected before. An awful lot of people have verified the decay rate of an awful lot of elements. One would think discrepant results would have come up and been checked and rechecked.
The Fischbach & Jenkins result has been talked about a bit in nuclear physics circles, so I'm reasonably well qualified to comment on it. First, a couple of general comments:
There are a large number of experiments (looking for dark matter, extrasolar planets, other exotica, etc.) which involve looking for a seasonal change in some detector. The detector has more counts in summer than winter, or higher-energy counts in spring than in fall, or the distant star's redshift seems to vary on a 7-day timescale, or whatever. The main lesson of 30 years' experience with these experiments is it's maddeningly difficult. One very rarely understands all of the environmental effects on one's detector---does it care a little about humidity? That varies seasonally, but of course one can try to measure or control for it. Does it care about the quality of your mains power? That varies seasonally (In some places, there's a seasonal switch from hydro to coal, with detectable effects on the power.) Does it care who's babysitting the experiment? (The personnel change seasonally---undergrads in the summer, for example.) How about the experiment two doors down the hall with the stray magnetic field? How about the jackhammers which are outside all summer? Etc.
It's similarly difficult to find day-night effects and 7-day effects. The former encounter all sorts of systematics (temperature, power, humans, etc.) and the latter encounter weekends. This is seriously a big deal: take any raw data from any (say) neutrino experiment, do a Fourier transform on it, and you'll see a big peak at 24 hours and 7 days. There are more people poking at the experiment 9-5 weekdays, and fewer at other times, and this always sneaks into the detector performance somehow. (Day-night differences are actually very interesting in neutrino experiments, but you can see that the good data on these topics is supported by a maddening level of calibration. You almost have to treat it like two detectors, the "day detector" and the "night detector", which you subject to totally separate calibrations.)
Fischbach and Jenkins report correlations for 32Si (a low-energy beta emitter) and 226Ra (a fairly typical alpha emitter). Rick Norman at LBL looked back at several long datasets of (see here) including 22Na, 44Ti, 108Ag, 121Sn, 133Ba, and 241Am, and saw nothing. Peter Cooper at Fermilab looked at the Pioneer Pu244 source and saw no sun-dependent variation whatsoever. http://arxiv.org/abs/0809.4248---certainly not the particular one proposed by Fischbach.
It's also worth noting that Fischbach has published something sort of like this before. his 1986 paper Reanalysis of the Eotvos experiment was *also* a re-analysis of someone elses' data, and *also* claimed to see evidence for a new fundamental force in the details of the data. (This prompted lots of reanalysis at the time; modern experiments fairly quickly ruled out the effect Fischbach reported.)
Finally, let me just comment on the actual data Fischbach is looking at. This data is reported in Astroparticle Physics (and the preprint version here. I think the content is the same, but I notice that the preprint has bitmapped figures. If you're interested in the details and have access to the Astropart. Phys., it is worth zooming in on the higher-quality graphs.)
First, some of the data is shown, not as raw data, but as a 5 point moving average. Moving averages introduce "fake" correlations when there are none in the raw data, see http://bactra.org/weblog/666.html for discussion. So ignore Figs. 1 and 4. (The actual analysis is done on raw data---I think---so this seems to be just a display/presentation issue.)
Where they do show raw data, it reveals the problem. The raw data has systematic errors in it. You don't have to take my word for it---you can see it. The count rate is drifting all over the place, both high and low---for example, look at the end of the trace in Fig. 2. (In addition, of course, it usually drifts in the directions suggested by Fischbach's hypothesis; the waves which are obvious to the eye are the same ones that pop out in his statistics.) Why is it able to drift so far? Remember that the claim, or the assumption, is that this is a detector which is only detecting this radionuclide and which is immune to minor environmental effects. Well, sorry, no it's not. The hump at the end, and the variation in the size of the peaks, are all perfectly clear evidence that this detector's count rate depends on something other than radioactivity and the Sun-Earth distance.
So the author wants to say "This detector is picking up the exponential decay of the nuclide, and also a season-dependent term"; he infers that the season dependent term is evidence for new particle physics. But that's not what the data show. The data show the detector picking up at least three things: 1) the exponential, AND 2) a season-dependent term, AND 3) some other unexplained medium-to-long-term drifts. But this doesn't suggest new physics at all; if there are medium-to-long-term unexplained detector drifts, you have no reason to believe that one of those drifts isn't a seasonal one. In spite of perfectly clear evidence that their detector has systematic drifts, Fischbach allows himself to assume that this particular annual effect is not such a drift. This assumption is unjustified.
Likewise in Fig. 3---this data shows that there's something other than just statistics+seasons altering these count rates; that's simply not what statistical-error-dominated data looks like.
There are a large number of experiments (looking for dark matter, extrasolar planets, other exotica, etc.) which involve looking for a seasonal change in some detector. The detector has more counts in summer than winter, or higher-energy counts in spring than in fall, or the distant star's redshift seems to vary on a 7-day timescale, or whatever. The main lesson of 30 years' experience with these experiments is it's maddeningly difficult. One very rarely understands all of the environmental effects on one's detector---does it care a little about humidity? That varies seasonally, but of course one can try to measure or control for it. Does it care about the quality of your mains power? That varies seasonally (In some places, there's a seasonal switch from hydro to coal, with detectable effects on the power.) Does it care who's babysitting the experiment? (The personnel change seasonally---undergrads in the summer, for example.) How about the experiment two doors down the hall with the stray magnetic field? How about the jackhammers which are outside all summer? Etc.
It's similarly difficult to find day-night effects and 7-day effects. The former encounter all sorts of systematics (temperature, power, humans, etc.) and the latter encounter weekends. This is seriously a big deal: take any raw data from any (say) neutrino experiment, do a Fourier transform on it, and you'll see a big peak at 24 hours and 7 days. There are more people poking at the experiment 9-5 weekdays, and fewer at other times, and this always sneaks into the detector performance somehow. (Day-night differences are actually very interesting in neutrino experiments, but you can see that the good data on these topics is supported by a maddening level of calibration. You almost have to treat it like two detectors, the "day detector" and the "night detector", which you subject to totally separate calibrations.)
Fischbach and Jenkins report correlations for 32Si (a low-energy beta emitter) and 226Ra (a fairly typical alpha emitter). Rick Norman at LBL looked back at several long datasets of (see here) including 22Na, 44Ti, 108Ag, 121Sn, 133Ba, and 241Am, and saw nothing. Peter Cooper at Fermilab looked at the Pioneer Pu244 source and saw no sun-dependent variation whatsoever. http://arxiv.org/abs/0809.4248---certainly not the particular one proposed by Fischbach.
It's also worth noting that Fischbach has published something sort of like this before. his 1986 paper Reanalysis of the Eotvos experiment was *also* a re-analysis of someone elses' data, and *also* claimed to see evidence for a new fundamental force in the details of the data. (This prompted lots of reanalysis at the time; modern experiments fairly quickly ruled out the effect Fischbach reported.)
Finally, let me just comment on the actual data Fischbach is looking at. This data is reported in Astroparticle Physics (and the preprint version here. I think the content is the same, but I notice that the preprint has bitmapped figures. If you're interested in the details and have access to the Astropart. Phys., it is worth zooming in on the higher-quality graphs.)
First, some of the data is shown, not as raw data, but as a 5 point moving average. Moving averages introduce "fake" correlations when there are none in the raw data, see http://bactra.org/weblog/666.html for discussion. So ignore Figs. 1 and 4. (The actual analysis is done on raw data---I think---so this seems to be just a display/presentation issue.)
Where they do show raw data, it reveals the problem. The raw data has systematic errors in it. You don't have to take my word for it---you can see it. The count rate is drifting all over the place, both high and low---for example, look at the end of the trace in Fig. 2. (In addition, of course, it usually drifts in the directions suggested by Fischbach's hypothesis; the waves which are obvious to the eye are the same ones that pop out in his statistics.) Why is it able to drift so far? Remember that the claim, or the assumption, is that this is a detector which is only detecting this radionuclide and which is immune to minor environmental effects. Well, sorry, no it's not. The hump at the end, and the variation in the size of the peaks, are all perfectly clear evidence that this detector's count rate depends on something other than radioactivity and the Sun-Earth distance.
So the author wants to say "This detector is picking up the exponential decay of the nuclide, and also a season-dependent term"; he infers that the season dependent term is evidence for new particle physics. But that's not what the data show. The data show the detector picking up at least three things: 1) the exponential, AND 2) a season-dependent term, AND 3) some other unexplained medium-to-long-term drifts. But this doesn't suggest new physics at all; if there are medium-to-long-term unexplained detector drifts, you have no reason to believe that one of those drifts isn't a seasonal one. In spite of perfectly clear evidence that their detector has systematic drifts, Fischbach allows himself to assume that this particular annual effect is not such a drift. This assumption is unjustified.
Likewise in Fig. 3---this data shows that there's something other than just statistics+seasons altering these count rates; that's simply not what statistical-error-dominated data looks like.
Not in any obvious way. Even if there is a small drift in decay rate, that's not going to add correlations to the data.
What I find really questionable about this is that it's only being noticed now. Artificial neutrino sources can produce something like 10 billion times more neutrinos than solar neutrinos. It seems odd that no one would have discovered this. If it works, it would mean that this is a remarkably sensitive way to detect neutrinos.
Then why not use pseudo-random number generator for all applications ?
I thought that pseudo-random were not quite good enough for certain cryptographic applications.
Gord_in_Toronto
Penultimate Amazing
- Joined
- Jul 22, 2006
- Messages
- 26,921
Thanks Ben. I'll cease panicking now. I was worried that the Fundamentalist Christians might be right that the Universe was only created 6,000 years ago and the current radio isotopic measurements showing otherwise had been affected by a huge Neutrino flux in the past.
Last edited:
noreligion
Unregistered
- Joined
- May 19, 2010
- Messages
- 999
Thanks Ben. I'll cease panicking now. I was worried that the Fundamentalist Christian might be right that the Universe was only created 6,000 years ago and the current radio isotopic measurements showing otherwise had been affected by a huge Neutrino flux in the past.
You can rest assured that they will try to claim this is some kind of evidence for a young Earth but you can also rest assured that they are wrong as usual. GPS has VISUALLYconfirmed continental drift over the past few years at a rate of approximately 10 cm per year. If this were a young Earth one can stand on the coast of New Jersey and see Europe since it would be less than a mile away if the young Earth crew were right. Anyone claiming young Earth is a history and science denier.
Gord_in_Toronto
Penultimate Amazing
- Joined
- Jul 22, 2006
- Messages
- 26,921
Oh. Sure. And can you disprove that Earth's mantel was not a lot softer 6000 years ago and the continents did not rush around as fast as high speed trains rather than a piddling 4" per year? Can you? CAN YOU?
dasmiller
Just the right amount of cowbell
Can you elaborate on that? I'm not sure how GPS could "visually" confirm something like that.
noreligion
Unregistered
- Joined
- May 19, 2010
- Messages
- 999
Are you seriously saying that we can't see the continental drift on GPS readings? Glad I can safely place you in the young Earth camp.
dasmiller
Just the right amount of cowbell
I'm seriously saying that GPS doesn't do much in the visible spectrum. When you said "GPS has VISUALLYconfirmed," what did you mean by "VISUALLY?"
That's an odd use of the word "safely."
noreligion
Unregistered
- Joined
- May 19, 2010
- Messages
- 999
You do not know what the word visually is? Let me ask you something, when a military strike is made and a target is destroyed according to radar is visual confirmation from a living breathing human necessary? Odd that you think GPS made the confirmation itself without human intervention. Now tell me something if a human operator sees satellite pictures and live satellite feeds of the continents 10 cm further apart each year which confirms GPS readings and confirms GPS displays, that isn't to be considered GPS confirming anything? Ok, be pedantic if you want.
dasmiller
Just the right amount of cowbell
I'd expect nothing else.
Okay, I typed a whole line-by-line response but we're pretty seriously off-topic here, so I deleted it. I'm satisfied that the word "VISUALLY" wasn't significant (in any sense) despite the caps, so I don't see much point in pursuing the question.
Olowkow
Philosopher
- Joined
- Oct 29, 2007
- Messages
- 8,230
Oh. Sure. And can you disprove that Earth's mantel was not a lot softer 6000 years ago and the continents did not rush around as fast as high speed trains rather than a piddling 4" per year? Can you? CAN YOU?
MUD, the continental "cdrift proponentists'" nightmare.