PDA

View Full Version : Strange Difference Between Gamma Ray Bursts and Quasars



Fraser
2006-Jul-31, 08:38 PM
Look at a quasar and a gamma ray burst - two of the most luminous objects in the Universe - and you're 4 times more likely to see intervening galaxies in front of the burst. This conclusion was reached by astronomers from UC Santa Cruz, who studied more than 50,000 quasars, and a handful of gamma ray bursts. There shouldn't be a connection between the quasar or burst in the background, and the number of galaxies in the foreground... but there is, and right now that relationship is a complete mystery.

Read the full blog entry (http://www.universetoday.com/2006/07/31/strange-difference-between-gamma-ray-bursts-and-quasars/)

ngeo
2006-Aug-01, 02:53 AM
“The result contradicts our basic concepts of cosmology, and we are struggling to explain it,” said Jason X. Prochaska, associate professor of astronomy and astrophysics at the University of California, Santa Cruz.

I wonder what basic concepts of cosmology are contradicted by this result.

GeorgeK
2006-Aug-01, 04:08 AM
Perhaps the burst is just 4 times further away? The result seems to assume the the bursts are at the same distance as the quasars.

rod57
2006-Aug-01, 09:21 AM
Why couldnt the 15 bursts supposedly behind the 'intervening' galaxies actually be from those galaxies ? Apologies if this is addressed in the original article (which I havent found yet).

hrizzo
2006-Aug-01, 09:49 AM
I've always had an "strange feeling" (unscientific, isn't it?) on quasars. And, as GeorgeK says, the answer could be the on distance.

And that is a solution not examined by de investigators.

Couldn't it be that the quasars redshift is not entirely due to a cosmological reason?. That could put them much nearer, and could explain a lot of things.

trinitree88
2006-Aug-01, 10:49 AM
“The result contradicts our basic concepts of cosmology, and we are struggling to explain it,” said Jason X. Prochaska, associate professor of astronomy and astrophysics at the University of California, Santa Cruz.

I wonder what basic concepts of cosmology are contradicted by this result.

Hey ngeo. How's it going? I'll take a stab at this.
1.The assumption that all of the redshift of a quasar must be intrinsic to it's recessional velocity, is invalid. There could also be (as there was in 1964), gravitational redshift due to a compact object. In the early-mid sixties, gravitational redshift as a cause fell out of favor,because such an object would consume all of it's nearby gas (which emits the hydrogen spectra shifted) fairly quickly. Now that's known not to be true, as a kinematically ejected pulsar, traveling through a gas cloud light-years in extant, can do so for millenia. Occasionally, such a pulsar will run out of such a cloud...and wink out (quasars have in the quasar surveys). Ejection velocity to escape a galaxy runs ~250-300 km/sec....these velocities have been seen in pulsar surveys....up to ~ 2000 km/sec (Shrinivas Kulkarni, Dale Frail).
2. GRB's have been recently theoretically divided into two likely suspects:A. neutron star-neutron star mergers. ...the short bursts
B. Supernovae (Type 1a's and subspecies...Type 2's and subspecies)

A. Neutron star mergers should occur where most of the neutron stars are, physically, and that's in the galactic halo. Most galaxies are thought to contain ~ 10,000-100-000 neutron stars at a range of ~ 400 kiloparsecs from the galactic hub. Only a small subset of the pulsars are kinematically ejected...~10% (Manchester, Kesteven,Cordes, Harrison, Lyne et al ...and "A Survey of Southern Pulsars" circa 1988?) So, when a GRB occurs, it will often be near the periphery of, or behind or infront of a galaxy....either because it's a star in the galaxy going SN type 1 or 2..(long burst)...or a galactic bound pulsar merge (short burst). Not a surprise for them to be galactically associated. (Gamma Ray Bursts a Halo of Neutron Stars at 400 Kiloparsecs? AAPT Meeting, Harvard University, Olney Science Center, 1994, Ben Affleck, Matt Damon, yours truly..1994)

The kinematically ejected pulsars are free to travel intergalactically...even merging out there by lower chances, to create massive beasts (see thread " A New View of Quasars" by Fraser last week. The enormous magnetic field present in them cause a precipitous decline in the half-life of neutrons in the skin. Magnetically decaying neutrons, and neutrino trapping in the interior of them cause polar jets to form releasing the energy due to parity effects. One pole will asymmetrically dominate the symmetry, so they're like birthday party balloons ...instead of air propelling latex across the room...it's degenerate matter propelling these beasts through a neutrino-driven jet. These will generally be found far from their galactic birthsites...in the intergalactic voids, and hence not usually be associated with a galaxy. A ratio of 4/1, not too surprising :dance: Pete

Send my regards to Matt & Ben.

snarkophilus
2006-Aug-01, 11:39 AM
Why couldnt the 15 bursts supposedly behind the 'intervening' galaxies actually be from those galaxies ? Apologies if this is addressed in the original article (which I havent found yet).

My first thought was along this line, though not quite the same. I haven't read the actual paper yet, but here's my initial impression.

Long duration GRBs are typically thought to occur when a high mass star goes supernova. Stars ought to be in rich sections of the universe, specifically in galaxies. More specifically, they are probably more likely to be in galaxies that are parts of clusters. So it would make sense if another galaxy in the cluster might be in the way. However, space is big... and clusters aren't all that dense.

With respect to your idea, high mass stars tend to be closer to the centres of galaxies. So if those fifteen GRBs all originated very close to the centres (or on the other sides) of rich galaxies, then that would make sense. However, "GRB hosts are typically low massive galaxies forming stars at a high rate." (http://www.pha.jhu.edu/~savaglio/ghosts/project.htm) So you wouldn't expect all that much absorption from the galaxy in which the GRB forms. You probably need a bigger one (though I don't know for sure -- it's probably discussed in the paper).

Another possibility is that the absorption lines in the GRB spectra are from gas ejected by the GRBs themselves, rather than from gas in intervening galaxies. But in nearly every case when researchers have taken a closer look in the direction of the GRB, they have in fact found a galaxy at the same position as the gas.

On the other hand, quasars are normally thought to be caused by material falling into really big black holes. That would also suggest a rich section of the universe, except that they tend to be much farther away than GRBs. That means an earlier stage in the universe's development, and perhaps clusters are different at that stage (fewer, more massive galaxies, or fewer galaxies and more mass in the intergalactic medium).

But since quasars are so far away, one would expect the odds of an intervening galaxy to increase. How much, I can't say.

I think it'd be neat to see just how far away all of these intervening galaxies are from the objects they overlap. There's probably a balancing act going on here, with various factors increasing and decreasing the likelihood of seeing an overlap for each of these objects, and a combination of small factors makes it all add up into one surprising result. Or maybe it's just a big statistical fluke.

Oh, and my favourite line from that article:

Those who know more about gravitational lensing than I do tell me it’s unlikely to be the answer.
That is one of my favourite explanations for things I don't know! "Someone smarter than I am says X. I'm inclined to believe it."

Fraser
2006-Aug-01, 02:39 PM
I wonder if intervening galaxies are necessary for us to see GRBs at all. Maybe they're so bright because they require a lensing galaxy for us to see, and they're actually not the brightest explosions in the Universe.

Ozzy
2006-Aug-01, 02:43 PM
Survey: Cosmological cosmic conundrum
SANTA CRUZ, Calif., July 31 (UPI) -- U.S. astronomers say a survey of galaxies observed along the sight lines to quasars and gamma-ray bursts creates a cosmic conundrum -- odd galaxy distribution.

The survey revealed a puzzling inconsistency: Galaxies appear to be four times more common in the direction of gamma-ray bursts than in the direction of quasars.

Quasars are thought to be powered by accretion of material onto supermassive black holes in the centers of distant galaxies. Gamma-ray bursts, the death throes of massive stars, are the most energetic explosions in the universe. But researchers say there's no known reason to expect galaxies in the foreground to have any association with these background light sources.

"The result contradicts our basic concepts of cosmology, and we are struggling to explain it," said Jason Prochaska, associate professor of astronomy and astrophysics at the University of California-Santa Cruz.

Prochaska and graduate student Gabriel Prochter led the survey, which used data from NASA to obtain observations of the transient, bright afterglows of long-duration gamma-ray bursts.

The research -- which might have strange cosmological implications -- has been accepted for publication in Astrophysical Journal Letters.

Copyright 2006 by United Press International. All Rights Reserved

http://www.newsdaily.com/Science/UPI-1-20060731-18442200-bc-us-cosomology.xml


"The result contradicts our basic concepts of cosmology,

How does this information contradict basic cosmological concepts?:think:

Sp1ke
2006-Aug-01, 03:05 PM
Is it the apparent causal link between the number of galaxies and the direction of GRBs? Sounds to me more like an unexplained correlation rather than a cosmological revolution.

ngc3314
2006-Aug-01, 03:38 PM
What's odd about their results is that they use two different kinds of background "searchlights" to examine the statistics of foreground galaxies. They compare a long-standing body of data on absorption lines seen from gas in front of quasars (all at much smaller redshifts than the quasars themselves), especially sets of absorption lines whose properties are most like gas in the outer reaches of galaxies (small velocity spread, strongest lines from metals such as Mg+) with new, similar data on the optical afterglows of gamma-ray bursts. This is useful since GRB afterglows are the only comparably bright probes to quasars for high redshifts, and for some purposes (local environment unaffected by the brilliant UV radiation of a quasar, for example) should give a fairer sample of the Universe. In both cases, the background sources are at such large redshifts that one expects no correlation with the foreground galaxies. However, they report a statistically significant excess of galaxy absorption in front of the GRBs. This makes zero sense given the large redshift differences in both samples. One explanation that springs to mind is that the GRB radiation might be preferentially observed in cases where gravitational lensing by a foreground galaxy makes it brighter - Prochaska argues that this effect is too weak to explain the difference, but I haven't gone through the calculation in detail to see whether there are additional assumptions one has to make (for example, the closer you work to a flux threshold, the more difference even a small sample bias can make).

So, indeed, this results makes no sense in our standard picture, and thus might be trying to tell us something, or might be an unlikely statistical fluke. Wish I could tell which at this point...

Ken G
2006-Aug-01, 04:44 PM
What bothers me is that even if you are going to accept the effect is real and try to build a new cosmology that explains it, how are you going to do it? It's not just that it contradicts our cosmology, it would seem to contradict any cosmology, because gamma-ray bursts come from galaxies and so do quasars, and the redshifts are not all that different right? So you not only have to make distant matter line up with foreground clumps, you have to make the GRB distribution be different from the quasar distribution in a way that is not statistically random. Granted they are different galaxy types, but what cosmology can couple a distant galaxy type to a local foreground distribution? Is it possible that quasars happen more in regions where there is relatively low galaxy density, so they don't lose angular momentum to the dark matter and you get a nice spiraling-in action around the SBH, whereas GRBs are in dwarf ellipticals that are in regions of high galaxy density? If so, then maybe one imagines cosmological "spokes" leading away from us, but that would certainly toss out the cosmological principle. Do you think the steady-state proponents will jump on this as evidence that the quasars and GRBs are not cosmologically distant? Anyway, it seems most likely that this is either an observational anomaly having to do with different instruments and flux levels, etc. or a selection effect, or pure coincidence.

John Mendenhall
2006-Aug-01, 05:10 PM
Non-cosmological redshifts ??!! Let me see, now where have we heard that before? Wasn't there a California astronomer who was banished to Europe for (h)arping on this idea? And weren't high quasar counts around his objects part of his argument?

Mehitabel: "Quiet ---too, too quiet."

ToSeek
2006-Aug-01, 05:25 PM
Related Q&A thread merged in with this one, redirect left.

Don Alexander
2006-Aug-01, 05:56 PM
Okay, maybe that was a bit exaggerated...

First of a, since no one here seems to have read the paper, here are two inks:

Prochter et al.'s paper:

http://arxiv.org/abs/astro-ph/0605075

Frank et al.'s try at an explanation:

http://arxiv.org/abs/astro-ph/0605676

So, let's see if I can answer some stuff:

ngeo asks: "I wonder what basic concepts of cosmology are contradicted by this result."

The cosmological principle: "The universe is homogenic and isotropic." Or in other words: No matter where I look, the universe looks basically the same on large scales, and no matter where I am, the universe looks basically the same on large scales.

Both GRBs and quasars are distributed randomly in the sky, so why should GRB sightlines be different??

GeorgeK asks: "Perhaps the burst is just 4 times further away? The result seems to assume the the bursts are at the same distance as the quasars."

If I recall correctly, the quasar sample was chosen to have the same redhift distribution as the GRB sample.

rod57 asks: "Why couldnt the 15 bursts supposedly behind the 'intervening' galaxies actually be from those galaxies ?"

The redshifts are derived via absorption spectroscopy. The highest redshift system is the GRB host galaxy. The intervening systems are at much smaller redshifts. Here's a nice recent example:

http://arxiv.org/abs/astro-ph/0607483

Concerning non-cosmoligical quasar redshifts: I'm not an expert on quasar spectra, but I'm sure that the evidence for them being cosmological sources is quite oerwhelming.

Concerning trinitree88 comments: Sure, GRBs are associated with galaxies - and short GRBs are thought to be found in halos (mounting evidence for that). But long GRBs are at very small offsets from the cores of their host galaxies, which are almost always compact dwarf galaxies.

But your whole argument is off the topic, as we are talking about intervening galaxy systems here, not the GRB host galaxies. Furthermore, ALL spectra we are talking about are from long (core-collapse) GRBs, not ejected NS mergers.

Snarkophilus: "More specifically, they are probably more likely to be in galaxies that are parts of clusters."

This is actually totally wrong. Long GRB host galaxies have been found to be isolated, while there have been cluster associations for short GRBs, but we're not talking about those. Furthermore, the intervening systems are at totally different redshift (say, GRB at z=1.5 and intervening system at z=0.6 - that's gigaparsecs...), so no cluster association.

"except that [Quasars] tend to be much farther away than GRBs."

Also very wrong. The mean redshift of GRBs is tending toward z=3, while the mean redshift of quasars is closer (I don't have the exact number, but somewhere around 2).

And concerning Jason's quote: That is one of my favourite explanations for things I don't know! "Someone smarter than I am says X. I'm inclined to believe it."
Well, that's science, actually. If someone is an expert in a field and says X, then he's probably right. You can't go around being sceptic about EVERYTHING, and you can't know everything either.

Finally, Fraser himself: "I wonder if intervening galaxies are necessary for us to see GRBs at all."

Well, this has been thought of. But you need quite massive foreground galaxies to create strong lensing, and no such system has been found near a GRB line of sight. Furthermore, many GRB hosts have been imaged in detail by the HST and they show no sign of lensing distortion.

Hope all this helps. The mystery Prochter et al. presented remains.

Alex

VanderL
2006-Aug-01, 09:02 PM
Okay, maybe that was a bit exaggerated...

First of a, since no one here seems to have read the paper, here are two inks:

Prochter et al.'s paper:

http://arxiv.org/abs/astro-ph/0605075

Frank et al.'s try at an explanation:

http://arxiv.org/abs/astro-ph/0605676

So, let's see if I can answer some stuff:

ngeo asks: "I wonder what basic concepts of cosmology are contradicted by this result."

The cosmological principle: "The universe is homogenic and isotropic." Or in other words: No matter where I look, the universe looks basically the same on large scales, and no matter where I am, the universe looks basically the same on large scales.

Both GRBs and quasars are distributed randomly in the sky, so why should GRB sightlines be different??

GeorgeK asks: "Perhaps the burst is just 4 times further away? The result seems to assume the the bursts are at the same distance as the quasars."

If I recall correctly, the quasar sample was chosen to have the same redhift distribution as the GRB sample.

rod57 asks: "Why couldnt the 15 bursts supposedly behind the 'intervening' galaxies actually be from those galaxies ?"

The redshifts are derived via absorption spectroscopy. The highest redshift system is the GRB host galaxy. The intervening systems are at much smaller redshifts. Here's a nice recent example:

http://arxiv.org/abs/astro-ph/0607483

Concerning non-cosmoligical quasar redshifts: I'm not an expert on quasar spectra, but I'm sure that the evidence for them being cosmological sources is quite oerwhelming.

Concerning trinitree88 comments: Sure, GRBs are associated with galaxies - and short GRBs are thought to be found in halos (mounting evidence for that). But long GRBs are at very small offsets from the cores of their host galaxies, which are almost always compact dwarf galaxies.

But your whole argument is off the topic, as we are talking about intervening galaxy systems here, not the GRB host galaxies. Furthermore, ALL spectra we are talking about are from long (core-collapse) GRBs, not ejected NS mergers.

Snarkophilus: "More specifically, they are probably more likely to be in galaxies that are parts of clusters."

This is actually totally wrong. Long GRB host galaxies have been found to be isolated, while there have been cluster associations for short GRBs, but we're not talking about those. Furthermore, the intervening systems are at totally different redshift (say, GRB at z=1.5 and intervening system at z=0.6 - that's gigaparsecs...), so no cluster association.

"except that [Quasars] tend to be much farther away than GRBs."

Also very wrong. The mean redshift of GRBs is tending toward z=3, while the mean redshift of quasars is closer (I don't have the exact number, but somewhere around 2).

And concerning Jason's quote: That is one of my favourite explanations for things I don't know! "Someone smarter than I am says X. I'm inclined to believe it."
Well, that's science, actually. If someone is an expert in a field and says X, then he's probably right. You can't go around being sceptic about EVERYTHING, and you can't know everything either.

Finally, Fraser himself: "I wonder if intervening galaxies are necessary for us to see GRBs at all."

Well, this has been thought of. But you need quite massive foreground galaxies to create strong lensing, and no such system has been found near a GRB line of sight. Furthermore, many GRB hosts have been imaged in detail by the HST and they show no sign of lensing distortion.

Hope all this helps. The mystery Prochter et al. presented remains.

Alex


Thanks Alex for the links and all the (quick) answers. There's one answer in your list that needs more explanation; the notion of non-cosmological redshifts. The finding would be consistent with quasars having an intrinsic component (at least a stronger component than GRB's and their hosts) and could maybe be used to find out how large such a component is (I'm not sure, though). Dismissing the possibility because other evidence seems overwhelmingly to the opposite could prove to be a mistake.

Cheers.

Btw, for people interested in a discussion about intrinsic redshift, see the "Arp et al." thread in the ATM section, that debate is very much alive.

Nereid
2006-Aug-01, 10:53 PM
... can we look at some of the other posts, which Don Alexander didn't cite?

For example, ngc3314's (http://www.bautforum.com/showpost.php?p=797065&postcount=11).

"they report a statistically significant excess of galaxy absorption in front of the GRBs" and "or might be an unlikely statistical fluke". Just how solid is the statistical analysis? Just how much of a fluke would the excess be (if it were a fluke)?

The good news about statistics and GRBs is that, over the next decade, the much larger dataset will allow better testing (and, maybe, squeeze the room for 'a fluke' into oblivion).

"One explanation that springs to mind is that the GRB radiation might be preferentially observed in cases where gravitational lensing by a foreground galaxy makes it brighter - Prochaska argues that this effect is too weak to explain the difference, but I haven't gone through the calculation in detail to see whether there are additional assumptions one has to make (for example, the closer you work to a flux threshold, the more difference even a small sample bias can make)." Selection effects are the bane of an astronomer's life - getting a sufficiently robust handle on all relevant aspects of such effects has taken decades (if not centuries) in some parts of astronomy*. With today's knowledge of GRBs, how confident can we be that this "statistically significant excess" isn't the footprint of one (or more) selection effects?

*For example, "quasar overdensities" can be said to have been laid to rest, in terms of non-cosmological origins of quasar redshifts, only with SDSS (with lensed quasars playing a supporting role) ... some 40 years after a well-known alternative idea was proposed.

Fraser
2006-Aug-02, 02:18 AM
I just interviewed Jason X. Prochaska on this for my podcast. I'll put it live as soon as possible. I'll drop him a note now, and see if he can jump in with a few answers on this thread.

profx
2006-Aug-02, 05:37 AM
At Fraser's suggestion, I thought I'd say a quick hello and see if I could contribute to the thread of discussion. I think it is fair to say that Don Alexander (whom I recently met at a GRB meeting) has well addressed many of the concerns that I saw posted.

I'll only add (for now) that the likelihood of the null hypothesis (namely that the GRB sightlines show the same frequency of intervening galaxies as QSOs) is less than 1 in 10000, which is an unusually strong result for Astronomy. And what is particularly powerful about the experiment is that we have enough QSOs from the Sloan Digital Sky Survey to very nearly make it an apples vs. apples comparison. This doesn't mean I'd be my life that the difference can't be statistical fluke, but it is very unlikely.

Am very curious myself to see where this goes...

Cheers,

X

peteshimmon
2006-Aug-02, 06:02 PM
Mmm...all this is somewhat similar to questions
raised when Radio Astronomers first started
bringing out sharp radio images of Quasars
using long baseline techniques. The images
were exquisitely aligned with the optical
object. But if there was any degree of
ionised matter along the line of sight one
might have thought the radio image would
become dissasociated from the optical image.
Well there just had to be a very clear line
of sight OK! We live and learn:)

VanderL
2006-Aug-02, 07:44 PM
At Fraser's suggestion, I thought I'd say a quick hello and see if I could contribute to the thread of discussion. I think it is fair to say that Don Alexander (whom I recently met at a GRB meeting) has well addressed many of the concerns that I saw posted.

I'll only add (for now) that the likelihood of the null hypothesis (namely that the GRB sightlines show the same frequency of intervening galaxies as QSOs) is less than 1 in 10000, which is an unusually strong result for Astronomy. And what is particularly powerful about the experiment is that we have enough QSOs from the Sloan Digital Sky Survey to very nearly make it an apples vs. apples comparison. This doesn't mean I'd be my life that the difference can't be statistical fluke, but it is very unlikely.

Am very curious myself to see where this goes...

Cheers,

X


Hi prof X, thanks for chipping in. So the results seem to be as solid as they can get, what do you think of the Frank et al. suggestion that different (beam) sizes between quasars and GRB's explains the findings?

Cheers.

VanderL
2006-Aug-03, 08:12 PM
Hi prof X, thanks for chipping in. So the results seem to be as solid as they can get, what do you think of the Frank et al. suggestion that different (beam) sizes between quasars and GRB's explains the findings?

Cheers.

Oops, I missed this in the article:


Frank et al. (2006) have considered an alternative explanation for the observed effect, namely that the difference in sizes between GRBs and QSOs leads to lower equivalent widths in QSO sightlines. We believe, however, that this model is ruled out because one does not observe unsaturated Mg II lines (at high resolution) where the doublet is not in a 2 : 1 ratio (Churchill 1997).

Is there a clear increase in the number of intervening galaxies in front of QSO's with increasing redshift in the SDSS catalogue (or any of the other surveys)? In the article in Figure 2 the lower red curve shows an upwards trend to Z=2, but is this trend also visible for higher redshifts?

Cheers.

turbo-1
2006-Aug-03, 09:24 PM
Is there a clear increase in the number of intervening galaxies in front of QSO's with increasing redshift in the SDSS catalogue (or any of the other surveys)? In the article in Figure 2 the lower red curve shows an upwards trend to Z=2, but is this trend also visible for higher redshifts?

Cheers.I will pitch this in here, since many people who read Fraser's informative breaking-news threads may not gravitate to the ATM's "Arp et al" thread. This link will take you to an on-line video lecture by Michael Strauss of the SDSS team, presented at the Space Telescope Science Institute. One very interesting result that he brings out is that the SDSS team surveyed 161 quasars at z~4-6.5 and not a single one of them was lensed. None. Even if you remove the 19 objects above z~5.7 (essentially single-band z-band detections in the SDSS ugriz bandpasses), you are left with 142 quasars at z~4-5.7 with not a single lensed example, despite the very great column density implied by a mainstream interpretation of their redshifts. This result, combined with profx's result pose some serious questions for mainstream cosmology. The most compelling question to me is if these results are pointing toward the existence of intrinsic redshifts in quasars. Certainly, if quasars are much closer than implied by a Hubble-relationship view of their redshifts, this lack of lensing could be understood, and the 4x larger-than-normal lensed sample of GRBs could be accomodated with no cosmological crisis.

Another important point made by Michael Strauss is that the SDSS team could detect no evolution in the properties of quasars (including absolute and relative metallicities) all the way out to redshift 6.5. As he points out quite gleefully, this result cannot adequately be explained by cosmologists. This is perhaps the best video presentation I have seen in the past year or so, though lectures by Penrose and Kolb are always both very intertaining and informative.

http://www.stsci.edu/institute/center/information/streaming/archive/STScIScienceColloquiaFall2005/

For those of you here who want more presentations, colloquia, courses, etc on video, I highly recommend that you bookmark Serkan Cabi's video link page. A tremendous educational resource.

http://web.mit.edu/people/cabi/Links/physics_seminar_videos.htm

Fraser
2006-Aug-03, 10:08 PM
I'm just working on the transcript and article for my interview, but if you want a sneak peak, you can access it here:

http://www.universetoday.com/audio/UT080306survey.mp3

turbo-1
2006-Aug-04, 01:17 AM
I'm just working on the transcript and article for my interview, but if you want a sneak peak, you can access it here:

http://www.universetoday.com/audio/UT080306survey.mp3That was a great interview, Fraser. You were non-intrusive and let profx do his thing, which he did in an animated and informative way.

Thanks for the preview!

Nereid
2006-Aug-05, 07:03 PM
Before we take this home: "The results suggest that at least one of our fundamental beliefs on absorption line research is flawed" (source (http://arxiv.org/abs/astro-ph/0605075)), perhaps we should go over all the steps which lead to it, and see what sorts of things should be checked in detail.

First, since the GRB afterglow observations were made with telescopes (+spectrascopes+software) different from what was used for the SDSS quasar observations, how much of a difference is there here?

For example, if the various telescopes+ used to make the GRB afterglow observations were used to observe an appropriate sample of SDSS quasars (selected for similarity in redshift and u/b/v/r/i/z magnitude at the time of observation), how many more MgII absorbers along sightlines to the SDSS quasars would have been detected (that meet the Prochter et al. selection criteria)?

(I think the answer is "less than 1", but if we're considering "our fundamental beliefs", isn't it a good idea to check?)

(to be continued)

VanderL
2006-Aug-07, 07:53 PM
Before we take this home: "The results suggest that at least one of our fundamental beliefs on absorption line research is flawed" (source (http://arxiv.org/abs/astro-ph/0605075)), perhaps we should go over all the steps which lead to it, and see what sorts of things should be checked in detail.

First, since the GRB afterglow observations were made with telescopes (+spectrascopes+software) different from what was used for the SDSS quasar observations, how much of a difference is there here?

For example, if the various telescopes+ used to make the GRB afterglow observations were used to observe an appropriate sample of SDSS quasars (selected for similarity in redshift and u/b/v/r/i/z magnitude at the time of observation), how many more MgII absorbers along sightlines to the SDSS quasars would have been detected (that meet the Prochter et al. selection criteria)?

(I think the answer is "less than 1", but if we're considering "our fundamental beliefs", isn't it a good idea to check?)

(to be continued)

And?

Which of our fundamental beliefs are flawed, or are we already trying to forget that this result is against expectations and move to other interesting but less controversial topics?

Cheers.

VanderL
2006-Aug-11, 06:01 AM
Before we take this home: "The results suggest that at least one of our fundamental beliefs on absorption line research is flawed" (source (http://arxiv.org/abs/astro-ph/0605075)), perhaps we should go over all the steps which lead to it, and see what sorts of things should be checked in detail.

First, since the GRB afterglow observations were made with telescopes (+spectrascopes+software) different from what was used for the SDSS quasar observations, how much of a difference is there here?

For example, if the various telescopes+ used to make the GRB afterglow observations were used to observe an appropriate sample of SDSS quasars (selected for similarity in redshift and u/b/v/r/i/z magnitude at the time of observation), how many more MgII absorbers along sightlines to the SDSS quasars would have been detected (that meet the Prochter et al. selection criteria)?

(I think the answer is "less than 1", but if we're considering "our fundamental beliefs", isn't it a good idea to check?)

(to be continued)


To be continued when?

Cheers.

(just a reminder before the item drops off the first page)

Jerry
2006-Aug-11, 06:02 PM
The anti-correlation is weird. Perhaps I can help. Assume both quasars and this type of Gamma ray are intrinsically redshifted. (I would be interesting in knowing exactly how Jason determined GRBs are divorced from intervening galaxies. - i guess I better read.) I will assume instead, they are pole features of a class of galaxy cores and intrinsically redshifted. Like wise, let's assume quasars pop out of a different family or core types, as proposed by arp. That makes these quasar things relatively small, and not visible 'behind' intervening galaxies.

So a galaxy with an 'inny' produces a gamma ray, and a galaxy with on 'outty' produces a quasar;)

Nereid
2006-Aug-11, 07:09 PM
Second, if the analyses were done with SDSS DR2 (say), or DR5, how different would the conclusions be? What if only GRBs in the same several thousand degrees of sky that SDSS covered were slected? Of course, should such analyses be done, and the conclusions differed significantly from those in the Prochter et al. paper, then maybe even more of "our fundamental beliefs" may be flawed!

Third, what room for misidentification (of MgII absorbers in the GRB afterglow spectra) is there? What confirmation (e.g. from other lines) is there of the reported MgII absorbers in those spectra?

Fourth, if we were to make sufficiently deep and wide observations of the regions around the GRBs with MgII absorbers, and (a sample of) SDSS quasars with MgII absorbers, how would the distribution of galaxies along (or near) the sightlines differ between the two? Such a test as this requires, of course, careful planning, lest selection effects make interpretation of results difficult.

Fifth, the GRBs with observed MgII absorbers are a small subset of all (long) GRBs detected, in the relevant time period, and these are a subset of all GRBs, in this same time period, with a) detected afterglows, and b) spectra obtained. The fact that one cannot think of how this kind of selection could possibly affect the number of MgII absorbers along GRB sightlines is rather a non-answer to any question about the effect of this selection on detection (of MgII absorbers).

(to be continued)

VanderL
2006-Aug-12, 09:51 AM
Second, if the analyses were done with SDSS DR2 (say), or DR5, how different would the conclusions be? What if only GRBs in the same several thousand degrees of sky that SDSS covered were slected? Of course, should such analyses be done, and the conclusions differed significantly from those in the Prochter et al. paper, then maybe even more of "our fundamental beliefs" may be flawed!

Third, what room for misidentification (of MgII absorbers in the GRB afterglow spectra) is there? What confirmation (e.g. from other lines) is there of the reported MgII absorbers in those spectra?

Fourth, if we were to make sufficiently deep and wide observations of the regions around the GRBs with MgII absorbers, and (a sample of) SDSS quasars with MgII absorbers, how would the distribution of galaxies along (or near) the sightlines differ between the two? Such a test as this requires, of course, careful planning, lest selection effects make interpretation of results difficult.

Fifth, the GRBs with observed MgII absorbers are a small subset of all (long) GRBs detected, in the relevant time period, and these are a subset of all GRBs, in this same time period, with a) detected afterglows, and b) spectra obtained. The fact that one cannot think of how this kind of selection could possibly affect the number of MgII absorbers along GRB sightlines is rather a non-answer to any question about the effect of this selection on detection (of MgII absorbers).

(to be continued)

Nereid,

Firstly the paper addresses most of your remarks, secondly the results are quite straightforward; the quasar data are based on a very large database, so there's no question about the validity of those data. Leaves only the number of GRB's that should give a statistical meaningful result, which is stated as being a chance of 1 in 10,000. In your opinion, do you consider this significant?

Finally, your way of phrasing the remarks/questions is very confusing, could you be more specific? For example what do you mean with this?


Of course, should such analyses be done, and the conclusions differed significantly from those in the Prochter et al. paper, then maybe even more of "our fundamental beliefs" may be flawed!


Or this?

Fifth, the GRBs with observed MgII absorbers are a small subset of all (long) GRBs detected, in the relevant time period, and these are a subset of all GRBs, in this same time period, with a) detected afterglows, and b) spectra obtained. The fact that one cannot think of how this kind of selection could possibly affect the number of MgII absorbers along GRB sightlines is rather a non-answer to any question about the effect of this selection on detection (of MgII absorbers).

Cheers.

Jerry
2006-Aug-13, 01:27 AM
If quasars are cosmic events in which a significant portion of the redshift is intrinsic as Hrizzo speculated, the answer is simple selection bias. Intrinsic redshifts, if not corrected for, over-estimate the magnitude of quasars. In look-back time, this is interpreted as evolution in the quasar head count (fewer quasars in the past), but if the quasar head-count remains more or less constant, direct comparison with a gamma ray in which all of the redshift is cosmic must present this type of bias. This is consistent with the lack of any measured transverse proximity effect by foreground quasars. This study is entirely consistent with explanations for the Lyman forest in quasar spectra that rely upon a portion of the quasar spectra being caused by intrinsic redshifting.

Nereid
2006-Aug-13, 04:43 AM
[snip]

Finally, your way of phrasing the remarks/questions is very confusing, could you be more specific? For example what do you mean with this?It has been a most frustrating experience trying to post my material.

As with most lengthy posts, I prepared it offline, and simply copy and pasted.

However, BAUT kept returning an odd error.

Several days of sleuthing ensued, with bits and pieces of my material getting successfully posted (and other bits not), and it seems the problem has been identified*, and a patch made (but is it solved? I don't know).

The net is that what got posted isn't what I had intended to post; so apologies for the confusion.

*it seems to be the s-e-l-e-c-t-e-d, which is why, in one post, I wrote 'slected'

Nereid
2006-Aug-13, 05:31 AM
Nereid,

Firstly the paper addresses most of your remarks,I don't think so, but I may be wrong. Let's go through each one, shall we?
secondly the results are quite straightforward; the quasar data are based on a very large database, so there's no question about the validity of those data.Hmm, clearly I've not expressed my points well if this is the 'take-away'.

In a nutshell, a claim about one of our fundamental beliefs (possibly) being flawed should trigger a careful check of all the links in the chains which lead to the results, if only to identify as many of the fundamental beliefs as possible.

For example, the SDSS quasar data is what it is - for an object to be identified as a 'quasar', it had to meet certain criteria; for 'foreground MgII absorption lines' to be identified in the spectrum of an SDSS quasar, certain criteria had to be satisfied.

Has SDSS identified all the quasars in the part of the sky it imaged? Clearly not. Does that matter, for the purposes of Prochter et al.? Yes, and no (for example, they imposed a z = 2 cutoff, in part because, indirectly, the SDSS quasars are thought to have sampled foreground MgII absorption lines with z > 2 too poorly to provide a sound basis for statistical analyses).
Leaves only the number of GRB's that should give a statistical meaningful result, which is stated as being a chance of 1 in 10,000. In your opinion, do you consider this significant?I don't understand this - could you clarify please?
Finally, your way of phrasing the remarks/questions is very confusing, could you be more specific? For example what do you mean with this?
Of course, should such analyses be done, and the conclusions differed significantly from those in the Prochter et al. paper, then maybe even more of "our fundamental beliefs" may be flawed!One fundamental belief that would be shaken would be isotropy - if large subsets of the SDSS data, selected by galactic longitude (say) showed big differences in the incidence of foreground MgII absorption lines, perhaps not all regions of space are the same after all?

Or if the GRBs in the same region of the sky as SDSS observed have a quite different distribution of foreground MgII absorption lines (than the total sample), one that more closely matches the incidence observed in the SDSS quasars?
Or this?
Fifth, the GRBs with observed MgII absorbers are a small subset of all (long) GRBs detected, in the relevant time period, and these are a subset of all GRBs, in this same time period, with a) detected afterglows, and b) spectra obtained. The fact that one cannot think of how this kind of selection could possibly affect the number of MgII absorbers along GRB sightlines is rather a non-answer to any question about the effect of this selection on detection (of MgII absorbers).It's a standard sampling, or selection, question - from the sample used in the analysis, what conclusions can be drawn about the population from which it was taken?

In this case, there is an assumption that the (long) GRBs which had observed (optical) afterglows are no different (statistically) wrt foreground MgII absorbers than the (long) GRBs for which there was no observed afterglow.

Ditto, wrt afterglows for which spectra were obtained vs those for which spectra were not obtained.

And so on.

Can you think of a reason why there should be differences between these different groups (wrt foreground MgII absorbers)? I can think of one or two (but they're pretty far fetched). However, whether you, or I, can think of reasons why the groups could (or should) differ, is irrelevant - if possible, the assumption that there is no significant difference should be tested (if possible).

VanderL
2006-Aug-13, 09:54 AM
If quasars are cosmic events in which a significant portion of the redshift is intrinsic as Hrizzo speculated, the answer is simple selection bias. Intrinsic redshifts, if not corrected for, over-estimate the magnitude of quasars. In look-back time, this is interpreted as evolution in the quasar head count (fewer quasars in the past), but if the quasar head-count remains more or less constant, direct comparison with a gamma ray in which all of the redshift is cosmic must present this type of bias. This is consistent with the lack of any measured transverse proximity effect by foreground quasars. This study is entirely consistent with explanations for the Lyman forest in quasar spectra that rely upon a portion of the quasar spectra being caused by intrinsic redshifting.

I think you're right that it could explain the difference, but is there any reason for a 4-fold differerence? And what about the redshift of GRB's, why wouldn't there be an intrinsic component there as well?

Cheers.

VanderL
2006-Aug-13, 11:00 AM
I don't think so, but I may be wrong. Let's go through each one, shall we?Hmm, clearly I've not expressed my points well if this is the 'take-away'.

In a nutshell, a claim about one of our fundamental beliefs (possibly) being flawed should trigger a careful check of all the links in the chains which lead to the results, if only to identify as many of the fundamental beliefs as possible.

For example, the SDSS quasar data is what it is - for an object to be identified as a 'quasar', it had to meet certain criteria; for 'foreground MgII absorption lines' to be identified in the spectrum of an SDSS quasar, certain criteria had to be satisfied.

Has SDSS identified all the quasars in the part of the sky it imaged? Clearly not. Does that matter, for the purposes of Prochter et al.? Yes, and no (for example, they imposed a z = 2 cutoff, in part because, indirectly, the SDSS quasars are thought to have sampled foreground MgII absorption lines with z > 2 too poorly to provide a sound basis for statistical analyses).I don't understand this - could you clarify please?One fundamental belief that would be shaken would be isotropy - if large subsets of the SDSS data, selected by galactic longitude (say) showed big differences in the incidence of foreground MgII absorption lines, perhaps not all regions of space are the same after all?

Or if the GRBs in the same region of the sky as SDSS observed have a quite different distribution of foreground MgII absorption lines (than the total sample), one that more closely matches the incidence observed in the SDSS quasars?It's a standard sampling, or selection, question - from the sample used in the analysis, what conclusions can be drawn about the population from which it was taken?

In this case, there is an assumption that the (long) GRBs which had observed (optical) afterglows are no different (statistically) wrt foreground MgII absorbers than the (long) GRBs for which there was no observed afterglow.

Ditto, wrt afterglows for which spectra were obtained vs those for which spectra were not obtained.

And so on.

Can you think of a reason why there should be differences between these different groups (wrt foreground MgII absorbers)? I can think of one or two (but they're pretty far fetched). However, whether you, or I, can think of reasons why the groups could (or should) differ, is irrelevant - if possible, the assumption that there is no significant difference should be tested (if possible).

Did you know that besides a difference in Mg II absorbers between QSO's and GRB's, there is also a difference between between BL-Lac's and QSO's (see here (http://xxx.lanl.gov/abs/astro-ph/9709111))?

Questioning the validity of the data is, imo, always useful, but in the paper the analysis is clearly described and Mg II absorbers are easliy identified, so no question about those numbers in GRB's (they even identified some of the actual intervening galaxies). The number of absorbers in front of quasars is calculated based on a huge number of quasars and the numbers were compared:

Prochter et al:

Assuming Poisson statistics, the observed incidence of 14 strong Mg II absorbers is inconsistent with the average value seen towards QSOs at >99.9% significance. We have also assessed the significance of the observation by drawing 10000 sets of quasars from the SDSS-DR4 chosen to have a similar g (z) function as the GRB sightlines. The results of this analysis is presented in Figure 3. We find an average of 3.8 strong Mg II absorbers, that less than 0.1% of the trials have over 10 systems, and that none has 14 absorbers. Therefore, it seems very unlikely that the difference in incidence between the GRB and QSO sightlines is only a statistical fluctuation.

Why don't you actually check the selection bias criteria you mention and address them, instead of listing all the possible errors and hope someone else will.

I think you'll find that all of your concerns were adequately addressed already by the authors, and the data stand. The chance this result is a statistical fluke is calculated by Prochter at al. to be 1 in 10,000. We may start wondering which of the fundamental beliefs are possibly in error, one of which you mentioned already (isotropy), but that would fly in the face of all current observations, not just assumptions.

Cheers.

VanderL
2006-Aug-18, 11:34 PM
Did you know that besides a difference in Mg II absorbers between QSO's and GRB's, there is also a difference between between BL-Lac's and QSO's (see here (http://xxx.lanl.gov/abs/astro-ph/9709111))?

Questioning the validity of the data is, imo, always useful, but in the paper the analysis is clearly described and Mg II absorbers are easliy identified, so no question about those numbers in GRB's (they even identified some of the actual intervening galaxies). The number of absorbers in front of quasars is calculated based on a huge number of quasars and the numbers were compared:

Prochter et al:

Quote:
Assuming Poisson statistics, the observed incidence of 14 strong Mg II absorbers is inconsistent with the average value seen towards QSOs at >99.9% significance. We have also assessed the significance of the observation by drawing 10000 sets of quasars from the SDSS-DR4 chosen to have a similar g (z) function as the GRB sightlines. The results of this analysis is presented in Figure 3. We find an average of 3.8 strong Mg II absorbers, that less than 0.1% of the trials have over 10 systems, and that none has 14 absorbers. Therefore, it seems very unlikely that the difference in incidence between the GRB and QSO sightlines is only a statistical fluctuation.

Why don't you actually check the selection bias criteria you mention and address them, instead of listing all the possible errors and hope someone else will.

I think you'll find that all of your concerns were adequately addressed already by the authors, and the data stand. The chance this result is a statistical fluke is calculated by Prochter at al. to be 1 in 10,000. We may start wondering which of the fundamental beliefs are possibly in error, one of which you mentioned already (isotropy), but that would fly in the face of all current observations, not just assumptions.

Cheers.


Can I conclude from this silence that the data are sound and the results are indeed a major puzzle?

Cheers.

Nereid
2008-Jan-25, 02:02 AM
Reconciling the Metallicity Distributions of Gamma-ray Burst, Damped Lyman-alpha, and Lyman-break Galaxies at z=3 (http://arxiv.org/abs/0801.3273)
We test the hypothesis that the host galaxies of long-duration gamma-ray bursts (GRBs) as well as quasar-selected damped Lyman-alpha (DLA) systems are drawn from the population of UV-selected star-forming, high-z galaxies (generally referred to as Lyman-break galaxies). Specifically, we compare the metallicity distributions of the GRB and DLA populations to simple models where these galaxies are drawn randomly from the distribution of star-forming galaxies according to their star-formation rate and HI cross-section respectively. We find that it is possible to match both observational distributions assuming very simple and constrained relations between luminosity, metallicity and HI sizes. The simple model can be tested by observing the luminosity distribution of GRB host galaxies and by measuring the luminosity and impact parameters of DLA selected galaxies as a function of metallicity. Our results support the expectation that GRB and DLA samples, in contrast to magnitude limited surveys, provide an almost complete census of star-forming galaxies at z=3.
The arXiv Comments is unusual*: "19 pages, 6 figures. Submitted to ApJ. Comments very welcome" (my emphasis)

* To me anyway; maybe I have simply not been paying attention

Jerry
2008-Jan-25, 05:38 AM
Do you think the steady-state proponents will jump on this as evidence that the quasars and GRBs are not cosmologically distant?
Here here!

This type of evidence must exist if the redshifts of quasars and GRBs are biased by an intrinsic component.

Time for a new 'dark spinner' parameter - webs through redshift space that align distance evidents with more local features.