Stuff I always wondered about astrophysics

[twofish-quant]

That puts you firmly in the camp of Arp, the Burbidges, etc who claim that galaxies can have intrinsic redshift.

Except that it's pretty well established that galaxies do have intrinsic redshift. If you plot galaxies on a Hubble diagram, you get quite a bit of scatter. Are you familar with "fingers of God"? If you plot galaxies on a distance diagram, you get elongated fingers pointing at earth, and those are due to galactic redshift scatter.

Any divergence from the Hubble relationship (either red or blue) needs to be investigated.

We are talking about what? 3000 km/h, which is about z=0.01? That's not going to make much of a difference in your Hubble diagrams.

 

[twofish-quant]

Can you point to where I've made such a claim? Perhaps you are confused?

He is. I proposed a model which I think nicely accounts for the observational results he claims. It might not be right, but by the principle of minimum surprise, you need to explain what's wrong with it first before claiming that what is going on *can't* be explained simply.

 

[twofish-quant]

One other thing. The observationalist that I know tends to be rather suspicous of any statistics that you get from a catalog. The trouble with using statistics from a catalog is that you need to read the catalog very, very carefully in order to make sure what the selection effects are (and in galaxy statistics there are always selection effects).

 

[twofish-quant]

Also keep in mind that I'm a theorist. If an observationalist comes up to me and tells me that it seems to them that the moon is made of green cheese, I'll start calculating the size of the cow that produced it. So if someone tells me that M81 has 11 components all with red shift, I start coming up models for showing how this can happen. If someone says, no they all are blue shifted, I toss that model and come up with a new one.

But the short answer is that even *if* M81 has companions that all are red-shifted, it's not hard to come up with a model that explains that without having to resort to non-standard cosmology.

 

[matt.o]

But the short answer is that even *if* M81 has companions that all are red-shifted, it's not hard to come up with a model that explains that without having to resort to non-standard cosmology.

Yes, the simplest being that M81 does not lie at the dynamical centre of the group.

 

[matt.o]

As for a mechanism for a redshift differential, that is a matter for theorists. We observe and constrain, and that is the role of astronomers, since astronomy is a purely observational science. Trends in the data show that about 1/3 of the small companion galaxies in bridged systems are blueshifted WRT to the large host galaxy, and those differentials are small - well within the range commonly accepted for the peculiar motions of gravitationally-bound companions. The remaining companions are redshifted WRT their hosts, and the differentials can be quite large, despite secondary evidence for interaction.

Let's examine this veracity of this claim. Looking at the data from Table 1 in Jokimaki, Orr & Russell (2008) (downloaded http://www.jorcat.com/tables.htm" ). Of these, 7/21 are blueshifted and 14/21 redshifted.

If those galaxies classified as "bnc" (bridge not connected) are discounted, since the evidence that these are interacting is weak when compared to the "bridge connected" sample, there are 6/16 blueshifted and 10/16 redshifted companions. Is this really a significant difference? Given the small numbers, and the chance that some of these systems could still be chance projections, I would be skeptical.

As expected, a larger number of bnc companions are redshifted (4/14 cf. 1/7 in the blueshifted). This is because it is more likely that a background chance-projection will occur compared to a foreground one, since a larger region is being probed in the background at these low redshifts. This claim is verified by considering those galaxies classified as having "discordant" redshifts in the Jokimaki, Orr & Russell paper (i.e. those with $$\Delta V > 1000\, km\, s^{-1}$$). There are 17 of these systems and 10 of these are classified bnc. The percentage across the whole M51-like sample which were classified as bnc was 16%, compared to ~60% of the discordant redshift objects. So are these really discordant redshifts or just plain old projections?

 

[turbo]

matt, you are right to question statistics on small samples, but if you understand our methodology, you know that our sample size is limited by restricting the set to galaxy associations that fit the M-51 template. Ours is not the first such project, as you can see from our reference list. The sample size (of systems with published redshifts) is also restricted by the dearth of spectroscopy in apparently-interacting galaxy pairs, so that hampers our work. In addition, we consider uncorroborated redshift data from an individual source as tentative, at best.

As for chance projections, there is also a chance that a nearby dwarf galaxy will appear projected over the arm of a more distant spiral and be discordantly blue-shifted WRT it. We found no such examples - and not for lack of trying. Over the course of 2 years, we studied images of many thousands of galaxies looking for interaction.

We address chance projection in our paper, and have done additional work on the statistics of chance projection, placement of companions, etc, in the interim. We will publish that in the future.

 

[twofish-quant]

Also from a theorist point of view, all this is "interesting" but not "we have to totally rethink cosmology interesting." Even if you were to establish conclusively that the interaction is not a statistical artifact (which is *incredibly* difficult), explaining that within the framework of standard cosmology is not terribly difficult. The velocity of these differences are in the hundred to thousand km/s range and there is no shortage of effects that could cause these sorts of effects. The other issue is that even if you were to establish that galaxies have redshift effects that cause things to be off by less than a few thousand km/s, this would cause corrections to distance/velocity plots, but it would not invalidate standard cosmologies.

If you want to work on something that *would* validate standard cosmology, then I think that statistics is just the wrong approach. The problem with statistics is that there are so many possible sources of statistical bias, and that's really hard to convince someone that you have a smoking gun as opposed to statistical bias, and it's messy for theorists because trying to disentangle what's going on when you might have very, very different things going on in each sample, is an impossible task.

If you want theorists to be interested, then rather than look for statistics, it's better if you do high resolution studies of one or two particularly weird objects. If you have an example where there is an interaction between a large main galaxy and a companion with a difference in redshift of z=0.3, then *that* would be interesting since it's hard to come up with an easy explanation. The other thing is publication bias.

Suppose you spend six months studying something and then it turns out that its just statistical bias, you have no consolation prize and nothing publishable. If you have an interaction between a single galaxy and a companion with redshift differences of z=0.3, and you have some very strong reason to think that it's not a coincidence, then after spending six months on it, if you've established that it's some bizarre gas jet effect, then you still have something publishable.

One trick in science is to set up experiments so that even null results are earth-shattering. I remember back in the mid-1990's, when the first results from COBE came in and we were talking to each other saying, "well if we don't see CMB anisotropies soon, then we have a lot of explaining to do." Same, with LIGO and LHC, if those come online and we don't see any gravity waves or the Higgs boson, then this will be something of an earthquake.

The fact is that there are lots and lots of things in the sky that we just don't understand, and if you theorists to be interested in your mystery, you have to provide some reason why your mystery is more likely to completely change people's understanding of the universe more so than the other mysteries out there.

 

[twofish-quant]

Something that would turn heads would be a single object with a huge redshift difference which is obviously not a projection. You can establish non-projection by having a gas bridge with an intermediate red shift. Or high resolution redshifts of the companion object in which you can show that the core has a massive redshift but the edges do not.

Also from a theoretical point of view, all these is much less important than it was in the 1970's because...

1) until the late-1970's, there was to mechanism to account for cosmological quasars. The gas-accretion model basically removed that objection

2) the bar for something Earth'shaking has increased enormously. If all you see is low-redshift objects than rather small intrinsic-redshifts can change greatly impact your data interpretation. Today, we have far, far more data on high redshift objects, which means that if you did establish high intrinsic redshifts in galaxies, it would have less of an impact. Suppose after we *did* establish that faint companions of interacting galaxies had intrinsic redshifts of say 2000-3000 km/s. It would affect calibration, but it wouldn't immediately change the standard cosmology much, because you have a mass of data that is unaffected by this change. If you have 3000 km/s redshifts, this wouldn't affect interpretation of objects with z=0.7 very much.

From a theorist point of view having an interacting galaxy with an apparent intrinsic redshift difference of z=0.1, would be very interesting, not withstanding any impact on cosmology, so as an observational strategy, you get more theorist interest by focusing on particular weird objects than you do by statistical arguments.

 

[turbo]

There is already an impressive grouping that should have been visited in more detail, instead of continually moving the goalposts - NGC 7603.

http://arxiv.org/abs/astro-ph/0203466

In this association, the main galaxy is a Seyfert and it is connected to its companion with 2x the redshift by a luminous bridge. The companion is roughly comparable in brightness to the host, so if it was a chance projection of an object twice as distant it would have to be 4x as bright in absolute magnitude. Embedded in the arm are two emission-line galaxies. One with over 8x the redshift of the Seyfert, and the other with almost 14x the redshift of the Seyfert. This system has been waved off as "chance projection" by the mainstream because there is no way to explain its appearance without invoking intrinsic redshift.

There is no other object visible around NGC 7603 that might have been responsible for the bridge, so the large companion is the most likely culprit despite its redshift. The two emission-line galaxies in the arm are of approximately the same size and brightness, despite the vast differences in their redshifts. If this association had been photographed in detail, with no spectroscopy, it would have been held out as a wonderful grouping of interacting bodies, with plenty of evidence for interaction. When redshift values are published, the entire association becomes "chance projection". To suggest otherwise is impolitic and dangerous for your career in astrophysics and astronomy.

N.A. Sharp made this quite clear in the follow-up paper of 1986.

However conventional the conclusions, there is considerable resistance to work on anomalous systems, and I am grateful to those who supported me, including the time allocation committees of the AAO, the MSSSO, and the KPNO, and to those who encouraged me who may prefer not to be named.

http://adsbit.harvard.edu/cgi-bin/nph-iarticle_query?bibcode=1986ApJ...302..245S

Sharp's paper was published 16 years before the nature of the bright blobs in the arm was known, well before the mesh of coincidences required for the appearance of the system (with conventional explanations) got so far-fetched.

 

[matt.o]

As for chance projections, there is also a chance that a nearby dwarf galaxy will appear projected over the arm of a more distant spiral and be discordantly blue-shifted WRT it. We found no such examples - and not for lack of trying. Over the course of 2 years, we studied images of many thousands of galaxies looking for interaction.

Again, that doesn't surprise me. This is a low redshift sample ($$|z| < 0.02$$) which is further limited by the M51-like criteria which limits the areal region which can be probed. Both of these limitations make the likelihood of finding one chance projection extremely small so I would not expect to find many at all. A further limitation of the sample is the requirement that the companion galaxy must be smaller than the main one, which again increases the likelihood that any chance projection will be a background one, rather than foreground.

In any case, looking at Table 1 again, I see NGC 0646 has a companion blueshifted by ~900km/s, but this difference could be less (~700km/s) depending on which redshift is used from NED. Looking at the Supercosmos sky survey images, it does look like there may be some interaction, but I would say this is a good candidate for being a blueshifted projection. Verification would require better imaging.

 

[matt.o]

There is already an impressive grouping that should have been visited in more detail, instead of continually moving the goalposts - NGC 7603.

http://arxiv.org/abs/astro-ph/0203466

In this association, the main galaxy is a Seyfert and it is connected to its companion with 2x the redshift by a luminous bridge. The companion is roughly comparable in brightness to the host, so if it was a chance projection of an object twice as distant it would have to be 4x as bright in absolute magnitude. Embedded in the arm are two emission-line galaxies. One with over 8x the redshift of the Seyfert, and the other with almost 14x the redshift of the Seyfert. This system has been waved off as "chance projection" by the mainstream because there is no way to explain its appearance without invoking intrinsic redshift.

Do you realize that the magnitude system is on a log scale, so, at a given distance, a magnitude 13 object is much more luminous than a magnitude 15.5 object? Also, there is a galaxy just off to the west with a similar redshift which could be responsible for the tidal features observed.

Edit to add: The absolute magnitude of NGC 7603 at z=0.029 is -22.5, while its NGC 7603b, at z=0.057, has absolute magnitude -21.5. The ratio of the luminosities is 0.36, i.e., NGC 7603 is ~3 times more luminous than its "companion" if the redshifts are cosmological. This is using the r-band magnitudes available from the SDSS http://cas.sdss.org/dr7/en/tools/chart/navi.asp?ra=349.73606&dec=0.24387&opt="

 

[twofish-quant]

turbo-l: There is already an impressive grouping that should have been visited in more detail, instead of continually moving the goalposts - NGC 7603.

True. If those things are interacting then there isn't any easy of explaining *that* particular galaxy without invoking weird physics. What I'd be interested in is looking at the high redshift galaxies to see if there is anything peculiar about them that makes them look different from standard quasars. One thing that would be interesting to see is see if the interacting bridges causes a shadow.

From a theory point of view, one reason for doubting in the 1970's that quasars were cosmological was that there was no good model for getting the required energies. There is now, and the model of quasars as being at cosmological fits so many things that it's going to take more than interacting galaxies to unseat that. The problem is that even if you to show that galaxies do have intrinsic redshifts, you have to also show that they are enough to invalidate distance measurements.

So, yes the goal posts do move over time.

turbo-l: To suggest otherwise is impolitic and dangerous for your career in astrophysics and astronomy.

I don't see why. If you talk to any theorist, you'll find that everyone has some nutty ideas, and you are not a decent theorist unless you have a few totally nutty ideas. Mine is "quantum immortality". To suggest that some redshifts are non-cosmological is hardly going to kill your career as a theorist. Most professionals realize that the evidence for things is nowhere is solid as intro astronomy textbooks suggest, and there are lots of anomalies out there.

But there is a difference from *suggesting* that redshifts are non-cosmological and *insisting* that redshifts are non-cosmological.

 

[twofish-quant]

One other thing (and I'm putting on my explain "how big a cow you need to make a moon of green cheese" hat).

It occurs to me that if galaxies form themselves in cosmological filaments which appears to be the case, then when you see a nearly galaxy, then there is a good chance that you are seeing the near end of a filament in which there are also other galaxies. If you look down an end of a filament of galaxies, you are likely to see a lot of galaxies lining up on a row, and I suspect that for large numbers of coincidences that a inhomogenous universe will give you a lot more galaxies lining up than you'd expect by chance. If the Earth is lined up with a wall of galaxies then you could easily see a dozen galaxies at difference distances right next to each other, whereas the changes of a random coincidence in a homogenous universe is likely to be nil.

In the unlikely event I have some free time, I might try calculating the odds of a chance alignment in a homogenous universe versus that you'd expect in an inhomogenous one, and I suspect that for large numbers of coincidences, that you'd expect a lot more alignments. In a homogenous universe, you'd never have large numbers galaxies at different distances next to each other, but I'm guessing that if the standard cosmology is correct, then this sort of thing ought to be quite common.

(And this has something to curiously this is very much related to my day job on Wall Street. It turns out the improbable events happen much more than by chance, because if two stock go down by 10%, then it's likely that you are in a situation that all of them will go down at the same time. The same thing might be operative here, in that if you see two galaxies at different distances next to each other, you are looking edgewise in a filament, in which you are likely to see twelve galaxies at different distances.)

 

[twofish-quant]

One other thing is that you can rule in and out geometry effects through correlations. Imagine a cone, with the foreground galaxy being a certain fraction of the distance down the cone, as you increase the limit at which the base of the cone registers a galaxy then you ought to see the red/blue asymmetry massively increase as you increase the number of detected objects.

So it occurs to me that one way of seeing whether this is a geometry effect is to look at the magnitudes of the galaxies. If this is chance coincidence then you ought to not only see a red/blue asymmetry but also an asymmetry in the relative brightness of the blueshifted galaxies/core galaxy and the redshifted ones/core galaxy. Also if this is a geometry effect, then as you ought to see the red/blue asymmetry start disappearing as you start limiting your samples to galaxies with a given luminousity ratio.

One other test you can use involves projection angle. I need to think about this some more but if you have one situation in which you have mostly interacting galaxies, then you'd expect to see the interacting galaxies randomly distributed in angle around the main galaxy (maybe). If the galaxies aren't interacting then I think you'd be able to tell the difference be doing statistics of the projection angles.

 

[courteous]

36 -- Why "black hole's temperature gets higher with less 'mass'"?

 

[turbo]

But there is a difference from *suggesting* that redshifts are non-cosmological and *insisting* that redshifts are non-cosmological.

I am not suggesting throwing out the baby with the bath-water. The Hubble flow relationship is on solid ground. There is evidence that in addition to the Hubble relationship and Doppler effects arising from peculiar motion, there is at least one other contributor to the redshifts of celestial bodies. It's not an Earth-shattering concept, but one that encounters a great deal of resistance from some quarters.