Note: I’m going to leave this as a sticky “top post” for a day or so. new stories appear below.
Nigel Calder asks us to republish this post for maximum exposure. He writes:
Today the Royal Astronomical Society in London publishes (online) Henrik Svensmark’s latest paper entitled “Evidence of nearby supernovae affecting life on Earth”. After years of effort Svensmark shows how the variable frequency of stellar explosions not far from our planet has ruled over the changing fortunes of living things throughout the past half billion years. Appearing in Monthly Notices of the Royal Astronomical Society, it’s a giant of a paper, with 22 figures, 30 equations and about 15,000 words. See the RAS press release at http://www.ras.org.uk/news-and-press/219-news-2012/2117-did-exploding-stars-help-life-on-earth-to-thrive
By taking me back to when I reported the victory of the pioneers of plate tectonics in their battle against the most eminent geophysicists of the day, it makes me feel 40 years younger. Shredding the textbooks, Tuzo Wilson, Dan McKenzie and Jason Morgan merrily explained earthquakes, volcanoes, mountain-building, and even the varying depth of the ocean, simply by the drift of fragments of the lithosphere in various directions around the globe.
In Svensmark’s new paper an equally concise theory, that cosmic rays from exploded stars cool the world by increasing the cloud cover, leads to amazing explanations, not least for why evolution sometimes was rampant and sometimes faltered. In both senses of the word, this is a stellar revision of the story of life.
Here are the main results:
- The long-term diversity of life in the sea depends on the sea-level set by plate tectonics and the local supernova rate set by the astrophysics, and on virtually nothing else.
- The long-term primary productivity of life in the sea – the net growth of photosynthetic microbes – depends on the supernova rate, and on virtually nothing else.
- Exceptionally close supernovae account for short-lived falls in sea-level during the past 500 million years, long-known to geophysicists but never convincingly explained..
- As the geological and astronomical records converge, the match between climate and supernova rates gets better and better, with high rates bringing icy times.
Presented with due caution as well as with consideration for the feelings of experts in several fields of research, a story unfolds in which everything meshes like well-made clockwork. Anyone who wishes to pooh-pooh any piece of it by saying “correlation is not necessarily causality” should offer some other mega-theory that says why several mutually supportive coincidences arise between events in our galactic neighbourhood and living conditions on the Earth.
An amusing point is that Svensmark stands the currently popular carbon dioxide story on its head. Some geoscientists want to blame the drastic alternations of hot and icy conditions during the past 500 million years on increases and decreases in carbon dioxide, which they explain in intricate ways. For Svensmark, the changes driven by the stars govern the amount of carbon dioxide in the air. Climate and life control CO2, not the other way around.
By implication, supernovae also determine the amount of oxygen available for animals like you and me to breathe. So the inherently simple cosmic-ray/cloud hypothesis now has far-reaching consequences, which I’ve tried to sum up in this diagram.
By way of explanation
The text of “Evidence of nearby supernovae affecting life on Earth” is available via ftp://ftp2.space.dtu.dk/pub/Svensmark/MNRAS_Svensmark2012.pdf The paper is highly technical, as befits a professional journal, so to non-expert eyes even the illustrations may be a little puzzling. So I’ve enlisted the aid of Liz Calder to explain the way one of the most striking graphs, Svensmark’s Figure 20, was put together. That graph shows how, over the past 440 million years, the changing rates of supernova explosions relatively close to the Earth have strongly influenced the biodiversity of marine invertebrate animals, from trilobites of ancient times to lobsters of today. Svensmark’s published caption ends: “Evidently marine biodiversity is largely explained by a combination of sea-level and astrophysical activity.” To follow his argument you need to see how Figure 20 draws on information in Figure 19. That tells of the total diversity of the sea creatures in the fossil record, fluctuating between times of rapid evolution and times of recession.
The count is by genera, which are groups of similar animals. Here it’s shown freehand by Liz in Sketch A. Sketch B is from another part of Figure 19, telling how the long-term global sea-level changed during the same period. The broad correspondence isn’t surprising because a high sea-level floods continental margins and gives the marine invertebrates more extensive and varied habitats. But it obviously isn’t the whole story. For a start, there’s a conspicuous spike in diversity about 270 million years ago that contradicts the declining sea-level. Svensmark knew that there was a strong peak in the supernova rate around that time. So he looked to see what would happen to the wiggles over the whole 440 million years if he “normalized” the biodiversity to remove the influence of sea-level. That simple operation is shown in Sketch C, where the 270-million-year spike becomes broader and taller. Sketch D shows Svensmark’s reckoning of the changing rates of nearby supernovae during the same period. Let me stress that these are all freehand sketches to explain the operations, not to convey the data. In the published paper, the graphs as in C and D are drawn precisely and superimposed for comparison.
There are many fascinating particulars that I might use to illustrate the significance of Svensmark’s findings. To choose the Gorgon’s story that follows is not entirely arbitrary, because this brings in another of those top results, about supernovae and bio-productivity.
The great dying at the end of the Permian
Out of breath, poor gorgon? Gasping for some supernovae? Named after scary creatures of Greek myth, the Gorgonopsia of the Late Permian Period included this fossil species Sauroctonus progressus, 3 metres long. Like many of its therapsid cousins, near relatives of our own ancestors, it died out during the Permo-Triassic Event. Source: http://en.wikipedia.org/wiki/Gorgonopsia
Luckiest among our ancestors was a mammal-like reptile, or therapsid, that scraped through the Permo-Triassic Event, the worst catastrophe in the history of animal life. The climax was 251 million years ago at the end of the Permian Period. Nearly all animal species in the sea went extinct, along with most on land. The event ended the era of “old life”, the Palaeozoic, and ushered in the Mesozoic Era, when our ancestors would become small mammals trying to keep clear of the dinosaurs. So what put to death our previously flourishing Gorgon-faced cousins of the Late Permian? According to Henrik Svensmark, the Galaxy let the reptiles down.
Forget old suggestions (by myself included) that the impact of a comet or asteroid was to blame, like the one that did for the dinosaurs at the end of the Mesozoic. The greatest dying was less sudden than that. Similarly the impressive evidence for an eruption 250 million years ago – a flood basalt event that smothered Siberia with noxious volcanic rocks covering an area half the size of Australia – tells of only a belated regional coup de grâce. It’s more to the point that oxygen was in short supply – geologists speak of a “superanoxic ocean”. And there was far more carbon dioxide in the air than there is now.
“Well there you go,” some people will say. “We told you CO2 is bad for you.” That, of course, overlooks the fact that the notorious gas keeps us alive. The recenty increased CO2 shares with the plant breeders the credit for feeding the growing human population. Plants and photosynthetic microbes covet CO2 to grow. So in the late Permian its high concentration was a symptom of a big shortfall in life’s productivity, due to few supernovae, ice-free conditions, and a lack of weather to circulate the nutrients. And as photosynthesis is also badly needed to turn H2O into O2, the doomed animals were left gasping for oxygen, with little more than half of what we’re lucky to breathe today.
When Svensmark comments briefly on the Permo-Triassic Event in his new paper, “Evidence of nearby supernovae affecting life on Earth,” he does so in the context of the finding that high rates of nearby supernovae promote life’s productivity by chilling the planet, and so improving the circulation of nutrients needed by the photosynthetic organisms.
Here’s a sketch (above) from Figure 22 in the paper, simplified to make it easier to read. Heavy carbon, 13C, is an indicator of how much photosynthesis was going on. Plumb in the middle is a downward pointing green dagger that marks the Permo-Triassic Event. And in the local supernova rate (black curve) Svensmark notes that the Late Permian saw the largest fall in the local supernova rate seen in the past 500 million years. This was when the Solar System had left the hyperactive Norma Arm of the Milky Way Galaxy behind it and entered the quiet space beyond. “Fatal consequences would ensue for marine life,” Svensmark writes, “if a rapid warming led to nutrient exhaustion … occurring too quickly for species to adapt.”
One size doesn’t fit all, and a fuller story of Late Permian biodiversity becomes subtler and even more persuasive. About 6 million years before the culminating mass extinction of 251 million years ago, a lesser one occurred at the end of the Guadalupian stage. This earlier extinction was linked with a brief resurgence in the supernova rate and a global cooling that interrupted the mid-Permian warming. In contrast with the end of the Permian, bio-productivity was high. The chief victims of this die-off were warm-water creatures including gigantic bivalves and rugose corals.
Why it’s tagged as “astrobiology”
So what, you may wonder, is the most life-enhancing supernova rate? Without wanting to sound like Voltaire’s Dr Pangloss, it’s probably not very far from the average rate for the past few hundred million years, nor very different from what we have now. Biodiversity and bio-productivity are both generous at present.
Svensmark has commented (not in the paper itself) on a closely related question – where’s the best place to live in the Galaxy?
“Too many supernovae can threaten life with extinction. Although they came before the time range of the present paper, very severe episodes called Snowball Earth have been blamed on bursts of rapid star formation. I’ve tagged the paper as ‘Astrobiology’ because we may be very lucky in our location in the Galaxy. Other regions may be inhospitable for advanced forms of life because of too many supernovae or too few.”
Astronomers searching for life elsewhere speak of a Goldilocks Zone in planetary systems. A planet fit for life should be neither too near to nor too far from the parent star. We’re there in the Solar System, sure enough. We may also be in a similar Goldilocks Zone of the Milky Way, and other galaxies with too many or too few supernovae may be unfit for life. Add to that the huge planetary collision that created the Earth’s disproportionately large Moon and provided the orbital stability and active geology on which life relies, and you may suspect that, astronomically at least, Dr Pangloss was right — “Everything is for the best in the best of all possible worlds.”
Don’t fret about the diehards
If this blog has sometimes seemed too cocky about the Svensmark hypothesis, it’s because I’ve known what was in the pipeline, from theories, observations and experiments, long before publication. Since 1996 the hypothesis has brought new successes year by year and has resisted umpteen attempts to falsify it.
New additions at the level of microphysics include a previously unknown reaction of sulphuric acid, as in a recent preprint. On a vastly different scale, Svensmark’s present supernova paper gives us better knowledge of the shape of the Milky Way Galaxy.
A mark of a good hypothesis is that it looks better and better as time passes. With the triumph of plate tectonics, diehard opponents were left redfaced and blustering. In 1960 you’d not get a job in an American geology department if you believed in continental drift, but by 1970 you’d not get the job if you didn’t. That’s what a paradigm shift means in practice and it will happen sometime soon with cosmic rays in climate physics.
Plate tectonics was never much of a political issue, except in the Communist bloc. There, the immobility of continents was doctrinally imposed by the Soviet Academy of Sciences. An analagous diehard doctrine in climate physics went global two decades ago, when the Intergovernmental Panel on Climate Change was conceived to insist that natural causes of climate change are minor compared with human impacts.
Don’t fret about the diehards. The glory of empirical science is this: no matter how many years, decades, or sometimes centuries it may take, in the end the story will come out right.
===============================================================
For those who would doubt our cosmic connections, Svenmark’s work and Calder’s article reminds me to remind you of this well known quote:
The nitrogen in our DNA, the calcium in our teeth, the iron in our blood, the carbon in our apple pies were made in the interiors of collapsing stars. We are made of starstuff. – Carl Sagan
kbmod
Please contact me about this as I don’t remember posting some of the stuff you have cut.
thanks
Adrian
Adrian Kerton says:
April 25, 2012 at 3:19 am
[snip . . OT, and covered extensively already . . kbmod]
Excellent work, will take some time to digest all of it though..
tallbloke says:
April 26, 2012 at 12:37 am
Hi Leif. How much does all that partygoing and bumping and grinding contribute to the cosmological background temperature of a few degrees above absolute zero?
The cosmic microwave radiation is made of photons [that has been stretched some 1100 times since the last scattering when the Universe was 372000 years old and became transparent] and not protons or cosmic rays, so the grinding has no effect on this.
Leif Svalgaard says:
April 26, 2012 at 5:59 am
tallbloke says:
April 26, 2012 at 12:37 am
Hi Leif. How much does all that partygoing and bumping and grinding contribute to the cosmological background temperature of a few degrees above absolute zero?
The cosmic microwave radiation is made of photons [that has been stretched some 1100 times since the last scattering when the Universe was 372000 years old and became transparent] and not protons or cosmic rays, so the grinding has no effect on this.
OK, so you’re saying the heat generated by all the myriad interactions you make an analogy to with the idea of crossing a crowded room at a cocktail party is so completely negligible it doesn’t affect the observed CMBR temperature? Doesn’t sound like such a wild party anymore. How come all this multitude of interactions which make the near lightspeed CGR particles arrive “millions of years” late to the CLOUD party on Earth don’t create measurable heat which masks the CMBR?
tallbloke says:
April 26, 2012 at 6:52 am
How come all this multitude of interactions which make the near lightspeed CGR particles arrive “millions of years” late to the CLOUD party on Earth don’t create measurable heat which masks the CMBR?
Because the density of cosmic rays is REALLY low and the scattering is by gyration in magnetic fields which does not cause energy loss or heating. No need to put the fact of millions of years in quotes.
tallbloke says:
April 26, 2012 at 6:52 am
How come all this multitude of interactions which make the near lightspeed CGR particles arrive “millions of years” late to the CLOUD party on Earth don’t create measurable heat which masks the CMBR?
Another way of looking at it is to note that there are 413 photons per cubic centimeter of the CMB, but less than 0.000,000,001 cosmic rays per cc.
Leif Svalgaard says:
April 26, 2012 at 7:18 am
by gyration in magnetic fields which does not cause energy loss or heating.
I should have said: ‘does not cause significant energy loss’. In fact, the scattering often helps to accelerate the cosmic rays.
Leif Svalgaard says:
April 26, 2012 at 7:41 am
In fact, the scattering often helps to accelerate the cosmic rays.
To elaborate on this: after bouncing around for 10 millions years cosmic rays can escape the Galaxy and can then heat the intergalactic medium to 10,000 degrees by not bouncing around anymore, but that medium is so dilute that its temperature does not matter. It is a bit analogous to the thermosphere of the Earth. We don’t get any heat from it in spite of its temperature being 1000+ degrees, or from the solar wind’s 100,000+ degrees. Bottom line: it all comes down to a matter of density.
I get nervous when people use data that contains internal contradictions.
might be something spectacular like 0.001, and we might conclude that while the correspondence in the curves is provocative, it is far from statistically meaningful as yet. With the error bars as they are given, 
is visibly pretty low, especially for that portion of the curve that is the primary “feature”.
is more or less irrelevant — all points receive roughly equal flux from sheer symmetry.
square kilometers and a depth the order of a light year. That’s a lot of volume — we won’t exactly receive a shotgun blast of GCRs at many times the baseline from it. I don’t have a good feel for how big an excess we’ll get from Betelgeuse — certainly a big chunk of contemporaneous gamma rays and lots of neutrinos, but the actual charged particle flux might still be fairly small compared to the background.
Me too, Willis. Of course, if one assigns a more uniform error bar to all the data points, assuming for example that the true measure of our uncertainty increases systematically with time backwards, then the entire figure might well live in the error bars,
And as I pointed out, there are alternative hypotheses for the variability that are supported at least as well as this one — that is to say, not yet terribly well.
Just sayin’, it might be wise not to rush out there and give Svensmark his Nobel Prize just yet. There has already been one absurd Nobel granted over climate science, one that it is very likely will go down in history as the biggest single mistake ever made by a the King of Sweden, a kind of cosmic IGNobel Prize of the sort that is usually only satire. Personally I think we don’t need another, and that we are decades away from being able to even THINK about validating the hypothesis. I get a headache even thinking about how one might proceed to reliably establish “supernova rates” over a billion year time frame into the past, and a bigger headache trying to solve the associated PDEs in my head that would indicate how e.g. GCR flux on Earth might vary.
Just a hint — the baseline assumption — even over a billion years — is that supernovae are Poissonian. That is, if one chops up something like a galaxy into cells, with each cell scaled to contain large numbers of stars, and looks at the Hertzsprung-Russell diagram:
http://casswww.ucsd.edu/archive/public/tutorial/HR.html
that gives, approximately, the expected distribution of stellar sizes, types, and ages of main sequence stars, one can actually guestimate fairly accurately how many stars one expects to be on the right part of the stellar evolution cycle to go supernova
http://casswww.ucsd.edu/archive/public/tutorial/SN.html#sn
Although small galactic volumes are somewhat inhomogeneous and often contain regions of older stars, regions of star birth (such as the aforementioned Pliades, which are not “exploding” BTW but rather igniting and which have nothing to do with Svensmark’s hypothesis so sorry to those quoting Job, or Orion, ditto), and yes, regions with more older stars. Those regions are certainly capable of producing some modulation of GCR rates due to the higher/lower rates of supernovae in the volumes in the immediate galactic neighborhood of those volumes.
However, then one has to subject the model to some rather brutal statistical analysis. The further away you are from the point of interest, the more stars there are in the time-lagged spherical shells that contribute to the total SN-linked GCR rates. The more stars there are, the smaller the statistical fluctuations around the mean. To put it another way, some fraction of the GCRs that fall upon the Earth come not from “local” exploding stars but from stars that are very far away indeed and that exploded long long ago. The entire volume of space out to just short of the big-bang radius contributes to this background flux, and given hundreds of billions of galaxies with hundreds of billions of stars, there are supernova happening all the time in this enormous contributory volume. Statistical fluctuations away from the mean scale pretty brutally over averages like this, and because the entire active volume is contributing
Spatial deviations from this are due to broken symmetry. Our local galaxy has a spatially inhomogeneous distribution of GCRs it produces in the plane of the galaxy, for example, at least “near” the galaxy itself. It is possible that we get some spatial modulation from the contributions from the Magellenic clouds and Andromeda in our immediate galactic neighborhood. However, the further you are away the less this matters. By the time you are order of 10 or 20 galactic radii away, little asymmetry should survive aside from — perhaps — some remnant asymmetry due to more probable axial rotation of the exploding stars in the galactic plane.
So the real question is — and I do not know how to answer it — what is the temporal inhomogeneity possible given this gross structure. To what extent are supernova not Poissonian events in an old galaxy (like our) consisting of mostly second or third generation stars, high in metals, ripe for life. Our galaxy is full of regions of star birth, but a bit sparse (as far as I know) in local clusters of stars that are all likely to die at or close to the same time. Betelguese has been mentioned — if it exploded it would have visible and easily measurable consequences on our local environment, but only for a relatively short time!. The supernova itself would “last” a matter of days and remain bright for up to a year; massive particles driven by the supernova would be temporally smeared out but also diminished by distance and by the same smearing process — I would guess that even this nearby a supernova would produce a bump in local GCR rates that lasted at most a few years. To put it another way, I doubt that we’re still getting a lot of excess GCRs from the explosion that produced the Crab nebula visible 1000 years ago. Or perhaps, we haven’t really gotten them yet — if they were produces 6000+ years ago, and they travel only 5-10% the speed of light, we might not get that burst of heavier particle radiation for tens of thousands of years!
6500 LY away is almost in our lap, but that bolus of particles will be spread out in a spherical shell with an area of
In the visible Universe, supernovae occur at roughly the rate of 1/second. That means that you are quite possibly getting hit every second or every few seconds with at least one quantum of energy that originated in a supernova — a gamma ray, a neutrino, a muon produced by a massive particle collision in the upper atmosphere. It is very difficult indeed to see how this baseline rate could secularly vary on a billion year timescale. It is very difficult to see how local fluctuations could secularly vary on a billion year timescale. The one is locked down by the laws of large numbers, barring a hypothesis that modulates synchronous galaxy formation rates in some signficant way that isn’t smeared out temporally at very large distances. The other is too local — Beteleuse might well bump our local radiation rates enough to affect climate if the various hypotheses connecting GCRs and climate are correct, but not over a timescale long enough to (probably) affect long term climate, that is be responsible for hundred million year-scale climate swings.
The only way I can imagine Svensmark’s result holding up is if there is a true all-length scale temporal fluctuation in supernova rates across the entire Universe. The problem is that there is nothing special about where the Earth is — all points are in the middle of a (practically, possibly) infinite volume. One cannot sanely hypothesize a synchronous event in a spherical shell (say) 400-500 million light years away from the Earth that produced a meaningful surplus of supernovae compared to rates at interior points or exterior points. Every time I try to mentally take rates of 1/second on a very large volume, smear the 31 million such events per year out over decades and vast contributing volumes, and then try to guestimate the poissonian fluctuations they end up coming out appallingly small, and smaller and smaller the further out you go. You get a meaningful modulation of the flux only from nearby events, but there aren’t enough of them to establish a hundred million year trend. By the time you can get enough to modulate a hundred million year trend, the modulation is literally poissonian noise and incapable of producing a significant hundred million year signal.
These are all just musings, of course — to really put teeth into them I’d have to do the computations (more likely simulations) and demonstrate that no reasonable distribution of supernovae in agreement with our time-lagged direct observations of their spatiotemporal frequency is capable of producing a significant modulation of GCR rates in turn capable of modulating the climate given that GCR rates modulate the climate at all — a hypothesis that is not yet proven, although there is some evidence to support it.
It does, however, indicate the kind of systematic doubt this paper and hypothesis should and will be subjected to. What the hell was “special” about the Universe N million years ago in a spherical shell of stars N/25 million LY in radius and N/10 LY thick (where N is order of 1000) that might have caused all of the stars in the Universe — not just stars in this geocentric shell — to have a statistically significant increase in the supernova rate compared to the present or compared to an interval (say) 2N or N/2 million years ago?
I’m guessing the correct answer is “nothing”. It is certainly the default answer. So before granting Nobel Prizes here and there, perhaps we should check the actual arithmetic.
rgb
rgbatduke says:
April 26, 2012 at 9:07 am
The entire volume of space out to just short of the big-bang radius contributes to this background flux, and given hundreds of billions of galaxies with hundreds of billions of stars, there are supernova happening all the time in this enormous contributory volume.
The cosmic rays we observe are almost all from our own Galaxy, not from the hundred of billions of other galaxies. This is because the Galactic Cosmic Rays [correct name] are trapped by the magnetic field of each galaxy. After 10 millions years they begin to leak out into intergalactic space and their lose their energy eventually [except the ultrahigh ones that have much more energy to lose from].
tallbloke says:
April 26, 2012 at 4:15 am
Yes, and since I’m banned from your site, I can’t answer … sadly, that’s altogether too typical, tallbloke. You still don’t seem to get it. Your site lost all credibility when you started banning people because you didn’t like our scientific ideas. You won’t get it back by attacking me there.
And as if that weren’t enough, perhaps as a result of now having a generally compliant readership who don’t ask the hard questions, your echo chamber now publishes astonishing things. As an example, consider the claims of a man who believes the earth’s temperature is due to “charge”, and that if you simply multiply the charge on a proton by the difference in size between the proton and the earth, you get the “charge” on the earth … I’m not kidding, that’s the claim. And if you want a real laugh, you should try to follow the man’s math, it’s absolutely priceless.
Tallbloke, after my experiences with Nikolov and Zeller, and with Jelbring, and with folks like you and Lucy Skywalker, and after my reading of some of the madness you are publishing like the paper from Miles Mathis the proton charge man, I wouldn’t comment at your site if you un-banned me and then paid me. My hipboots would not be deep enough, and I don’t have a pair of chest waders.
w.
rgbatduke says:
April 26, 2012 at 9:07 am
“It is very difficult to see how local fluctuations could secularly vary on a billion year timescale.”
Only if you believe that it makes no difference whether the solar system is in or between galactic arms. The fact that the solar system moves in and out of the galactic spiral arms indicates to me that what appears to be spiral arms are in fact density (Alfven?) waves, and that as we traverse the spiral arms the local density of galaxies and hence the density of interstellar plasmas (the coslmic rays) will fluctuate. Especially if, as Leif says, the plasma is localized by galactic magnetic fields.
pochas says:
April 26, 2012 at 12:55 pm
rgbatduke says:
April 26, 2012 at 9:07 am
“It is very difficult to see how local fluctuations could secularly vary on a billion year timescale.”
Only if you believe that it makes no difference whether the solar system is in or between galactic arms.
some confusion here. rgb’s argument was that if the cosmic rays came from all over the Universe, then there would be no variations, but as they don’t, the argument falls away.
rgbatduke: The only way I can imagine Svensmark’s result holding up is if there is a true all-length scale temporal fluctuation in supernova rates across the entire Universe. The problem is that there is nothing special about where the Earth is — all points are in the middle of a (practically, possibly) infinite volume. One cannot sanely hypothesize a synchronous event in a spherical shell (say) 400-500 million light years away from the Earth that produced a meaningful surplus of supernovae compared to rates at interior points or exterior points.
It’s difficult to apply probabilistic models to events that have already happened, such as the probability that life might have originated on Earth by chance, or the probability that the MiIky Way Galaxy would have exactly this spiral shape at this lag time after the big bang. . Statistically, no part of the universe is special. However, the actual distribution of stuff is not uniform at all spatial scales; thus, Svensmarks question could be phrased: What is the conditional probability that a sequence with a particular pattern would happen on Earth by chance, given that another sequence of events had already occurred many years before in other parts of the galaxy; is the observed Earth record remarkably different from what would occur by chance?
you quote Willis: I get nervous when people use data that contains internal contradictions.
I would be extremely suspicious of a data set that did not contain internal contradictions. So many real data sets have them, that I would suspect fraud.
Willis Eschenbach and rgbatduke: I get nervous when people use data that contains internal contradictions.
I would be extremely suspicious of a data set that did not contain internal contradictions. So many data sets have them that I would suspect extreme amounts of data massaging — “homogenization”, so to speak.
Nigel Calder is the author of the high praise found in this thread’s original post, not Anthony Watts.
Matthew R Marler says:
April 26, 2012 at 2:41 pm
Say what? Most datasets don’t contain internal contradictions, at least the ones I use. The raw datasets often contain errors and missing data, that is indeed common.
But the errors are generally removed during early quality control, and as a result, usually we’re not looking at error bars that not only don’t overlap, but are miles apart, as in this example.
Having said that, a more accurate statement would be
Bear in mind that the dataset under discussion is actually a collection of a variety of different proxies of CO2. Svensmark has done what Mann and others have been correctly criticized for doing—grabbing a bunch of proxies and never even considering if all of them are valid.
Svensmark’s mistake is that if you are going to use proxies, you have to have ex-ante criteria for proxy selection.
My thanks to you as always,
w.
Willis Eschenbach: Svensmark’s mistake is that if you are going to use proxies, you have to have ex-ante criteria for proxy selection.
Now you are objecting to his selection of proxies? He selected proxies capable of testing his hypothesis.
from the paper: In this paper the aim is to use the least model-dependent approach to the course of events in the past 500 Myr, by deriving the star formation rates and supernova rates directly from open star clusters in the solar neighbourhood, and using the SN rate as a proxy for the GCR flux to the Solar System.
[Willis says:]
Bear in mind that the dataset under discussion is actually a collection of a variety of different proxies of CO2.
Are you saying that he does not in fact have a useful proxy for sea level or diversity of marine invertebrates?
Matthew R Marler says:
April 26, 2012 at 3:46 pm
Yeah, but all proxies are not created equal …
Heck, I haven’t even considered his evidence for those, other than my comment about the thickening of the edges of continental plates … no, I’m saying that if you have three CO2 proxies complete with error bars for a given time period, and the error bars don’t overlap, and you use all three proxies without comment, that you are engaged in guesswork rather than science.
w.
oops, the “bear in mind” is a quote from Willis, not from the paper as it appears.
[Fixed. -w.]
It struck me that whereas Willis has been writing, in effect, “Don’t believe this paper”, I have been writing, in effect “Don’t ignore this paper.”
Willis Eschenbach: I’m saying that if you have three CO2 proxies complete with error bars for a given time period, and the error bars don’t overlap, and you use all three proxies without comment, that you are engaged in guesswork rather than science.
I would call it “scientific guesswork.”
Leif Svalgaard says:
April 26, 2012 at 7:31 am
tallbloke says:
April 26, 2012 at 6:52 am
How come all this multitude of interactions which make the near lightspeed CGR particles arrive “millions of years” late to the CLOUD party on Earth don’t create measurable heat which masks the CMBR?
Another way of looking at it is to note that there are 413 photons per cubic centimeter of the CMB, but less than 0.000,000,001 cosmic rays per cc.
Thanks Leif, this and your other replies have helped me get a handle on the way the standard model quantifies the energies involved. Since the interstellar medium is full of magnetic fields a long way from the matter generating the GCR’s, and the star generating the fields presumably a comparatively strong organised magnetic field such as the heliosphere will scatter them more. So does the heliosphere contribute a lot to the ‘coming from all directions’ nature of the GCR flux?
tallbloke says:
April 27, 2012 at 12:45 am
So does the heliosphere contribute a lot to the ‘coming from all directions’ nature of the GCR flux
Some, but not a lot, as the modulation of cosmic rays is only about 10% [depending on energy].
‘And another drop of water fell to the floor of the cave and splattered on the rocks..’
Science takes time. Relax. Don’t be impatient. If anyone tells you the sky is falling, or not, you’ll still have enough time to live several lifetimes and be able to take all your thousands of kids to the Zoo more times than you can count. This drop of water is most important because we all got a little wet, and in this cave, things don’t dry much at all.
The cosmic rays we observe are almost all from our own Galaxy, not from the hundred of billions of other galaxies. This is because the Galactic Cosmic Rays [correct name] are trapped by the magnetic field of each galaxy. After 10 millions years they begin to leak out into intergalactic space and their lose their energy eventually [except the ultrahigh ones that have much more energy to lose from].
Ah, I can see that I’m going to really have to get my mind wrapped around the turbulent dynamo concept. I would have assumed by default that the average magnetic field of a galaxy was negligible, but apparently there is a very large scale process related (perhaps) to the smaller scale process that goes on in stars and planets that produces not only magnetic fields, but fields with intriguing structure and symmetry in galaxies. Clearly I have a lot to learn here.
However, that still leaves my original “problem”, just now it is expressed on a smaller scale. GCR rates thus modulate within the galactic plane, fine, but is that modulation correlated with SN rates within the galaxy on the right time scales? Also, what precisely is the granularity of the modulation? If I understand the structure of the fields produced from the online article I found, they are highly filamentary and local even within the ISM of the galaxy(s) themselves, and if charged particles are helically following the field lines one would expect the modulation of the GCR rate to be predominantly related to the local magnetic field strength, not a global modulation of supernova rates in the galaxy. Indeed, it makes the Svensmark hypothesis less tenable, does it not, because the rate of supernovae in the visible Unverse is on the order of one per second, making the rate of supernovae in any given galaxy far, far smaller. There is still a serious problem in statistics and estimates of plausible signal to noise here.
As a matter of curiosity, does the local galactic magnetic field temporally vary on a directly measurable scale in the vicinity of the Sun (measurably over the baseline of instrumentation available to measure it)? Is there secular variation of GCR rates from non-solar sources over that same time frame? I think that I recall from Svensmark’s first paper that he was more inclined to correlate climate cycles with the bobbing of the sun up and down in the galactic plane, moving it in and out of domains of greater or lesser GCR rates, but if the GCRs are associated with localized filamentary structures in the galactic magnetic field, might they not modulate on a much shorter (and much more random) time scale?
rgb