Quasar anomalies

  1. Chronos

    Chronos 10,225
    Science Advisor
    Gold Member

    I am very disturbed by the lack of time dilation in the variability of quasar v redshift studies.

    http://www.arxiv.org/abs/astro-ph/0105073
    Title: Time Dilation and Quasar Variability
    Authors: M.R.S. Hawkins
    The timescale of quasar variability is widely expected to show the effects of time dilation. In this paper we analyse the Fourier power spectra of a large sample of quasar light curves to look for such an effect. We find that the timescale of quasar variation does not increase with redshift as required by time dilation. Possible explanations of this result all conflict with widely held consensus in the scientific community.

    I think this is a very serious challenge to modern theory.
     
    Last edited: Mar 11, 2006
  2. jcsd
  3. Garth

    Garth 3,535
    Science Advisor
    Gold Member

    All quiet on the Western Front!

    It is indeed a disturbing paper, thank you Chronos, Fred Hoyle would have loved it!
    Reactions?

    Garth
     
  4. Chronos

    Chronos 10,225
    Science Advisor
    Gold Member

    The the lensing explanation is a nice try:

    http://www.arxiv.org/abs/astro-ph/0306434
    Title: Can Microlensing Explain the Long-Term Optical Variability of Quasars?
    Authors: Erik Zackrisson, Nils Bergvall, Thomas Marquart, Phillip Helbig
    . . . Although controversial, the scenario of microlensing as the dominant mechanism for the long-term optical variability of quasars does provide a natural explanation for both the statistical symmetry, achromaticity and lack of cosmological time dilation in quasar light curves.

    Unfortunately, it does not appear to hold water:

    Structure Function Analysis of Long Term Quasar Variability
    http://www.arxiv.org/abs/astro-ph/0411348
    Authors: W. H. de Vries (1,2), R. H. Becker (1,2), R. L. White (3), C. Loomis
    . . . We find the following: (1) the outbursts have an asymmetric light-curve profile, with a fast-rise, slow-decline shape; this argues against a scenario in which micro-lensing events along the line-of-sight to the quasars are dominating the long-term variations in quasars.

    The authors understate the case in asserting this study 'argues against' microlensing. There is overwhelming evidence quasars are not local [i.e., within this galaxy], but, this is a serious blow to 'cosmology as usual', IMO. P.S, I'm not ready to buy any 'tired light' stock.

    I searched long and hard trying to find a fatal flaw, or sensible explanation to the issue raised; and came up empty. What I find shocking is how little attention this has been given in the literature. I'm experiencing that same feeling I had as a child when mom, after intense grilling, admitted there was no Santa Claus.
     
    Last edited: Mar 12, 2006
  5. Nereid

    Nereid 4,014
    Staff Emeritus
    Science Advisor
    Gold Member

    well worth studying further!

    Some suggestions:
    • Hawkins' analysis. It's highly unlikely that it contains gross errors (it wouldn't have passed peer-review), but maybe there are some subtle ones?
    • An alternative analysis of the data. He employed one particular approach to determining a 'time dilation' footprint in the data - perhaps the same data, analysed a different way, might give a different result?
    • Selection effects. This is the bane of astronomers' life, and some of these effects are subtle indeed. Perhaps a cold, hard look at the inputs - as compared with what we now know about quasars, and what we know we don't know - might turn up patterns that weaken the stated conclusion? For example:
    • a) redshift-apparent magnitude. In the standard view, the higher the redshift, the greater the distance. Among other things this also means that two quasars with different redshifts but the same apparent magnitude will have quite different absolute magnitudes. In a (apparent) magnitude-limited selection (that was Hawkins' input catalogue, yes?), this introduces biases in the sample.
    • b) colour. The input catalogue wasn't (AFAICS) characterised for quasar colour; there could be all kinds of selection effects relating to quasar colour (more later, if I have time).
    • c) Variability. The quasars to be studied were not selected because of their (then known) variability. However, the extent to which the variability (or lack of it) played a part in them being detected as quasars in the first place should definitely be looked into.
    • The nature of quasars. Unlike Cepheids or SNe, the details of quasars' variability is essentially unknown ('accretion disks' might be responsible for most of such variability, but just how do they give rise to the observed frequency spectrum?) Without some handle on this, the intrinsic causes and nature of quasar variability, I don't see how alternative explanations for an apparent absence of time dilation could be ruled out - intrinsically brighter quasars are variable in ways that are different from those not so bright? evolution effects? observed variability is due to the integration of several, quasi-independent causes?
    I don't see much chance of a resolution via theory any time soon (there's just too much work to do, involving a lot of heavy-duty physics), so maybe an observational project could yield results (confirmation, or not) in the short term?

    For example, studies of the observed variability of quasars, in several time regimes, in spectra, in different wavebands (gamma, X-ray, UV, NIR, FIR, ...), ... to tease out any systematic effects and put constraints on them.
     
    Last edited: Mar 12, 2006
  6. wolram

    wolram 3,784
    Gold Member

    http://arxiv.org/abs/astro-ph/0106220

    Fourier Analysis of Gamma-Ray Burst Light Curves: Searching for Direct Signature of Cosmological Time Dilation
    Authors: Heon-Young Chang (Korea Institute For Advanced Study)
    Comments: 12 pages, 2 figures, 2 tables, accepted to ApJL

    We study the power density spectrum (PDS) of light curves of the observed gamma-ray bursts (GRBs) to search for a direct signature for cosmological time dilation in the PDS statistics with the GRBs whose redshifts $z$'s are known. The anticorrelation of a timescale measure and a brightness measure is indirect evidence of its effect. On the other hand, we directly demonstrate that a time dilation effect can be seen in GRB light curves. We find that timescales tend to be shorter in bursts with small redshift, as expected from cosmological time-dilation effects, and we also find that there may be non-cosmological effects constituting to this correlation. We discuss its implication on interpretations of the PDS analysis results. We put forward another caution to this kind of analysis when we statistically exercise with GRBs whose $z$ is unknown.
    An older paper.
     
  7. SpaceTiger

    SpaceTiger 2,977
    Staff Emeritus
    Science Advisor
    Gold Member

    I wouldn't give the peer review process that much credit. My roommate recently reviewed a paper that was absolutely loaded with errors and poor reasoning -- and it was nearly identical to several that had already been published. Unfortunately, many peer reviewers are just lazy.

    That said, Hawkins doesn't appear to be a crackpot, so I wouldn't dismiss the paper outright. It will be interesting to see how this is resolved.
     
  8. Garth

    Garth 3,535
    Science Advisor
    Gold Member

    Mike Hawkins is a doctorate member of staff of the Royal Observatory, Edinburgh and certainly not a 'crackpot'.

    I agree with Chronos
    The most disturbing thing about this paper is the lack of interest. I would have thought that it would have stirred up a hornet's nest, either of refutations or a major effort of reconciliation with the standard model. After four years there has been more of less nothing published. The 'silent treatment' perhaps?

    Those were very good points Nereid, especially the question of selection and colour dependent variability effects. From that OP paper of Hawkins:
    Quite a statistically significant sample?

    We note the raw details of observed cosmological time dilation so far: Time dilation is seen in distant S/N decay profiles, in slow GRBs but not (apparently) in quasar variability. Could it be (if slow GRBs are distant hyper-novae) that in the first two the engine is basically 'normal' matter undergoing nuclear conflagration, whereas with a quasar the engine is degenerate mass, i.e. a black hole?

    In SCC non-degenerate matter and degenerate matter behave differently, because the scalar field is coupled to non-degenerate matter but decoupled from relativistic matter, therefore cosmological time dilation is to be expected in distant S/Ns but not BHs. I just thought I'll add that as a suggestion!

    Garth
     
    Last edited: Mar 13, 2006
  9. Nereid

    Nereid 4,014
    Staff Emeritus
    Science Advisor
    Gold Member

    Thinking about this some more, I see less reason to be concerned at the apparent absence of a time dilation signal in the data.

    First and foremost, in the mainstream view of quasars, we have quasar evolution - pure luminosity evolution, perhaps. If high-z quasars are different from low-z ones, then, a priori, expecting their (rest frame) variability to be unchanged is rather foolish, isn't it?

    It may be possible to take some different views though Hawkins' data (as "a Fourier power spectrum was calculated for each [quasar's] light curve"), to look for an evolutionary effect, but the data as presented in the paper is too coarsely binned to do such a study.

    Another way to test this may be to look at the power spectra of local AGNs (Seyfert nuclei, BL Lacs, ...).

    Second, slicing and dicing the dataset in other ways would have been helpful - in terms of how the quasars were detected, for example, or weighted for 'completeness' (by comparison with data from SDSS, say), or X-ray or radio brightness.

    Third, some 'field controls' would have increased confidence in the outputs (e.g. power spectra of variable stars on the plates) - I wonder why this wasn't done?

    Fourth, why not make some artificial quasars? Modelling (monochromatic) variability, in a computer, should be a piece of cake - the outputs from the artificial quasars, as discrete 'plate magnitudes' (complete with errors and 'cloudy nights'), could then be fed into Hawkins' analysis pipeline. By tweaking the 'quasar properties' (evolutionary history - luminosity, variability, colour, ...), it should be fairly straight-forward to find out what sorts of things could mimic Hawkins' ~600 power spectra.

    Finally, I think a different kind of analysis (approach) would be welcome.
     
    Last edited: Mar 12, 2006
  10. Chronos

    Chronos 10,225
    Science Advisor
    Gold Member

    My impression was Hawkins bent over backwards trying to find a way to explain this unexpected result. His publication history suggests he is not afraid to go out on a limb, but they tend to be thick.

    Nereid, my objection to relying upon quasar evolution is the extraordinary fine tuning it appears to require. The coarse binning actually strengthens the result, IMO. This smoothes out the oft seen outrageous probabilities suggested by over-binning [a typical crackpot tactic]. It is also evident there are two broad color classes of quasers. Assuming different mechanisms are involved, it appears improbable both would seemlessly cancel out the expected time dilation effect [please don't sentence me to TD :smile:]

    I don't entirely get the variable star thing. I suppose that would be a way of testing for selection effects, but given the large data set, I doubt any systematic effects of that nature are in play.

    I like the computer modelling suggestion, Nereid. Sounds like good thesis material. This is surely something that demands further study.

    I am also concerned about how this same logic could be applied to the apparent time dilation of supernova and GRB light curves. This inconsistency is what I find most troubling.

    Nonetheless, Nereid, your observations are soundly based and logically consistent. And that is exactly what I expected to find from the cosmology community. But, their silence is deafening.
     
    Last edited: Mar 13, 2006
  11. Garth

    Garth 3,535
    Science Advisor
    Gold Member

    Absolutely, Nereid your ideas appear to be worthy of a good hearing, if nobody else has done so, why not write a paper? :approve:

    Garth
     
  12. Nereid

    Nereid 4,014
    Staff Emeritus
    Science Advisor
    Gold Member

    All the more curious then that he seems not to have done several of the (to me) more obvious 'sanity checks' (or, if you prefer, quality controls).
    I think the operative word is "appears" - without doing some simulation, I don't see how anyone could say just how 'extraordinary' any 'fine tuning' would need to be (more later, not necessarily in this post).
    Not at all - we tend to get tunnel vision, when it comes to 'colour'. 'Red' and 'blue' actually differ very little, when the full range of EM that quasars are well known to emit over, in significant quantities, is considered.
    You could think about it as just a 'being careful' step, or you could say that without this, the conclusions rest on weak foundations.

    Imagine there are ~600 variables in the same field(s) where the quasars are to be found. Imagine their colour and (apparent) magnitude distributions are similar to those of the quasars. Imagine Hawkins produced 600 Fourier power spectra, of these variables. Imagine that a computer program blindly, and randomly, applied 'time dilation' to subsets of those spectra, in a distribution which matched the z's of the quasars.

    Now feed those two sets of 600 power spectra into Hawkins' "bin and analyse" pipeline ... what would the results be?

    At the very least, we would have (hopefully) internal consistency checks on the approach (and tools) which Hawkins used.
    How about we lay the groundwork, here in PF? (more later).
    Different bathtub of fish ... quasars are a zoo, of animals which range from shrews to elephants; SNe and GRBs are private collection of different species of mice.
     
  13. Nereid

    Nereid 4,014
    Staff Emeritus
    Science Advisor
    Gold Member

    Just quickly - another systematic effect might be evident in the fainter (apparent mag) quasars.

    The errors in estimating magnitude would have been - presumably - greater for fainter images; to what extent was this incorporated in the analysis? The paper seems to say estimates of magnitude were all given the same (per plate?) errors.

    Also, for quasars near the (faint) limit, how many 'disappeared' over the history of the data collection (i.e. dimmed below detectability)? If the mode was 0.6 mag, "with a tail extending to 2 mag", this is not a trivial consideration.
     
  14. Dare one ask why anyone hasn't repeated the observations - to at least prove them wrong?
     
  15. Nereid

    Nereid 4,014
    Staff Emeritus
    Science Advisor
    Gold Member

    Welcome to Physics Forums, ratfink!

    From the Hawkins paper:
    IOW, it isn't easy to 'repeat' them.

    It is certainly worth checking various observatory (plate) records, to see what sort of comparable consistent, (historical) datasets there might be.

    Any suggestions?
     
  16. Thanks,
    It is just that I read this quote

    So do the plates lie? If not, are they part of the public domain? If so, they could be checked to see if 'non time dilation' is what the results show and that it is not a 'mistake'. Would other scientists looking at the same data set come to the same conclusions - or different ones? and if so why haven't they done it?
     
  17. SpaceTiger

    SpaceTiger 2,977
    Staff Emeritus
    Science Advisor
    Gold Member

    Different researchers have reached this same conclusion -- that there is no apparent time dilation in the quasar light curves. I don't think many people argue the results, it's mostly a question of interpretation. In my opinion, the most likely explanation is some kind of evolution (see Nereid's posts). I don't think it would take a great deal of fine-tuning to reproduce these results because the error bars are still large (the high- and low-redshift subsamples appear to only be a few sigma apart). If we reduce the size of the error bars by another factor of a few and still see no significant difference in the subsamples, then I think that would be cause for concern.

    Keep in mind that the author, in the introduction to his paper, makes note of previous studies that had already confirmed the existence of cosmological time dilation in other sources (GRBs, SNe), so I suspect we're not seeing much interest in this topic because the mainstream has already dismissed this effect as evolution (or selection effects). Although this kind of evolution is interesting in of itself, it's considerably less boat-rocking than a challenge to mainstream cosmology.
     
  18. Thanks,
    But I thought that in science to 'prove' something then one had to look at the same sort of thing from all sorts of angles and show that it is consistent. If one has time dilation in SIa's but not in quasars then isn't 'confirm' the wrong word to use for time dilation?
     
  19. SpaceTiger

    SpaceTiger 2,977
    Staff Emeritus
    Science Advisor
    Gold Member

    I'm not sure what you mean. Cosmological time dilation has been confirmed to exist in SNe and (possibly) GRBs, but not in quasars. Nothing has been proven 100%, but the effect has been observed in those objects to whatever significance the observations can be trusted. It's possible there is some exotic explanation for the effect (like Garth's theory), but I'm just talking about observations, not theory.
     
  20. Thanks,
    I was just wondering why supernova Ia results are said to be 'proof' and consequently quasar results must be 'flawed' as far as time dilation is concerned. Why not the other way around? i.e. Quasar 'non time dilation' show the Universe is not expanding and supernovae Ia results are an 'exotic effect?'
    BTW what about
     
  21. selfAdjoint

    selfAdjoint 8,147
    Staff Emeritus
    Gold Member

    Go back and read Nereid's post on this; if it's too compressed for you I'm sure she'd be glad to expand it. Then pay attention to Space Tiger's remarks about error bars. From just these two posts we find: (1)There is a plausible explanation for the plates, and (2)Astronomical data is inherently noisy. And so the case is still open. Time dilation is supported by a huge amount of independent evidence, and this contradictory evidence is perhaps not all it seems to be.

    In science you never "prove" anything, you just have more or less support. I repeat, dilation has an enormous amount of supporting evidence.
     
Know someone interested in this topic? Share this thead via email, Google+, Twitter, or Facebook

Have something to add?