Do voids in the Universe influence the orientation of galaxies?

[Astronuc]

Shedding Light on the Cosmic Skeleton
http://www.eso.org/public/outreach/press-rel/pr-2009/pr-41-09.html

Astronomers have tracked down a gigantic, previously unknown assembly of galaxies located almost seven billion light-years away from us. The discovery, made possible by combining two of the most powerful ground-based telescopes in the world, is the first observation of such a prominent galaxy structure in the distant Universe, providing further insight into the cosmic web and how it formed.

“Matter is not distributed uniformly in the Universe,” says Masayuki Tanaka from ESO, who led the new study. “In our cosmic vicinity, stars form in galaxies and galaxies usually form groups and clusters of galaxies. The most widely accepted cosmological theories predict that matter also clumps on a larger scale in the so-called ‘cosmic web’, in which galaxies, embedded in filaments stretching between voids, create a gigantic wispy structure.”

These filaments are millions of light years long and constitute the skeleton of the Universe: galaxies gather around them, and immense galaxy clusters form at their intersections, lurking like giant spiders waiting for more matter to digest. Scientists are struggling to determine how they swirl into existence. Although massive filamentary structures have been often observed at relatively small distances from us, solid proof of their existence in the more distant Universe has been lacking until now.
. . . .

Does this imply a continuous network, web or skeleton of the universe? What are the cosmological implications of this discovery?

Galactic Clusters - http://www.eso.org/gallery/v/ESOPIA/GalaxyClusters

 

[marcus]

"The most widely accepted cosmological theories predict that matter also clumps on a larger scale in the so-called ‘cosmic web’, in which galaxies, embedded in filaments stretching between voids, create a gigantic wispy structure...

These filaments are millions of light years long and constitute the skeleton of the Universe: galaxies gather around them, and immense galaxy clusters form at their intersections, ..."

...
Does this imply a continuous network, web or skeleton of the universe? What are the cosmological implications of this discovery?
...

It certainly is a very interesting discovery, and by no means a new one!
We were seeing pictures of the wispy cobwebby structure already 10 years ago. And the formation of such structures is a central part of the accepted theory of structure formation in the early universe.

Here is a 2001 press release about the observation of the wispy cobwebby structure, which was earlier predicted by models. This wouldn't be the earliest appearance of the idea, just a sample:
http://www.eso.org/public/outreach/press-rel/pr-2001/pr-11-01.html

One of the important implications has to do with dark matter. People do computer models of the formation of structure, where matter (most of it dark) is allowed to condense from near-uniformity (by ordinary gravity) and they find the simulations give pictures that look pretty much like what we observe!

The dark matter---because so much of it, it dominates---is what condenses into strands---and denser regions where strands intersect. Then the ordinary matter (only about 1/10 as much) is attracted to these concentrations of dark matter and collects along these strands and especially at intersections---and it is what we see.

Structure formation computer modeling (and it's striking agreement with observation) is one of a several interlocking kinds of evidence supporting the assumption of dark matter.

One of the places one can see these computer simulations of early universe structure formation is in the popular TED lecture by George Smoot. Google "TED Smoot". It is only 15-20 minutes and it's well worth watching the whole thing.

=========================
A further question is what can we expect to learn from this newly reported 2009 observation of a particular wisp at redshift z = 0.55?
Presumably other things like this have been detected and they lend themselves to technical refinements of detail.
Here is the Tanaka et al. abstract:
http://www.aanda.org/index.php?option=article&access=doi&doi=10.1051/0004-6361/200912929
They say "The observed structure is among the richest ever observed in the distant Universe. They will be an ideal site for quantifying environmental variations in the galaxy properties and effects of large-scale structure on galaxy evolution."
Here is the Tanaka et al. preprint:
http://arxiv.org/abs/0909.3163
The spectroscopically confirmed huge cosmic structure at z=0.55

 

[mikeph]

Each dot is a galaxy, the distance to the right and left are distances from Earth above and below the galactic plane (the region we can see). I think this is one of the best cosmology pictures that exists, you can see this "web" very well... amazing, really.

 

[marcus]

Nice picture Mikey!

Just a BTW comment, I was happy to see that the ESO press release used the present-day distance---the distance used in working astro and cosmology. They did not use "light travel time" interpreted as a distance. We should encourage this among ourselves.

If you plug z = 0.55 into wright's calculator you get that the distance NOW to the structure is 6.7 billion lightyears.
That would be the radar distance if you could freeze expansion today and send a signal to it.
And ESO said "nearly 7 billion light years." Clearly they had that 6.7 in mind.

On the other hand the calculator will also tell you that the light travel time was 5.4 billion years. Light travel time is not related in any simple fixed way to actual distance, because the expansion rate has varied considerably over time. Since they said 7 billion lightyears, they could not have had the travel time (of 5.4 billion years) in mind.

There is a website that caters to middle-schoolers (and younger teens in general) that used to use light travel time as an expression for distance, this contributed to confusion---but I think they may have cleared that up somewhat recently. For astronomers and most of the rest of us, distance is normally actual distance (with expansion frozen at some definite time so that it can be well-defined.)

 

[Chronos]

This, IMO, is misleading marcus. Submarines use sound travel time to compute the future distance and position of other ships. If this approach is flawed, how do they hit their targets? Sound travel time is used to compute the 'now' distance and predict the future distance. Light travel time is no different.

 

[Sorry!]

Hey Mikey thanks for posting that image. As well I'd have to agree with Chronos... as long as it is made clear which distance method is used it won't make a difference... will it? (aside from the obvious difference in the numerical value)

@OP
Astronuc as you can see from the many posts here the web structure has been observed for quite sometime. A lot of simulations you can find on youtube show this web structure in 3D...




When I first saw this simulation it blew my mind away.

 

[Wallace]

On the other hand the calculator will also tell you that the light travel time was 5.4 billion years. Light travel time is not related in any simple fixed way to actual distance, because the expansion rate has varied considerably over time. Since they said 7 billion lightyears, they could not have had the travel time (of 5.4 billion years) in mind.

There is a website that caters to middle-schoolers (and younger teens in general) that used to use light travel time as an expression for distance, this contributed to confusion---but I think they may have cleared that up somewhat recently. For astronomers and most of the rest of us, distance is normally actual distance (with expansion frozen at some definite time so that it can be well-defined.)

(emphasis mine)

I'm sure ESO could have used any distance measure they wanted in the press release, all that matters is that it is in the Billions and a big number. But you're right, it looks they are using the transverse Proper distance as defined by FRW co-ordinates for this.

But please don't use phrases like 'the actual distance' to describe this quantity. It is no more special, unique or correct than any other distance, and to say that it is is thoroughly misleading. It's perfectly acceptable, and probably true, to say that this might be the simplest, easiest to understand and most usefull distance to use at a pop sci level, but stop there!

Let me make this clear, cosmologists do not use the transverse proper distance as defined by FRW co-ordinates when talking to each other! You would only use measureable distances for this, such as angular diameter distance, luminosity distance or redshift. What ESO uses to convert these into some distance for a press release should not be interpreted as saying anything about scientific discourse. For starters, you have to assume a set of cosmological parameters (I know you understand all this) which makes this distance ambigous and prone to revision in a way that measurable distances are not.

I've read probably hundreds of cosmology papers and attended at least a dozen or so cosmology conferences. I've still yet to see a distance be communicated by the tranverse proper distance in co-moving co-ordinates, which would be surprising if indeed this was the correct 'actual distance'...

(note to be clear: 'Proper distance' is a technical term, it doesn't imply a unique or correct description. In fact the definition of this distance requires arbitrary gauge choices to be defined).

 

[Astronuc]

@OP
Astronuc as you can see from the many posts here the web structure has been observed for quite sometime. A lot of simulations you can find on youtube show this web structure in 3D...

It was my impression that the web structure has been studied for sometime, although it's not something I follow closely. I believe SpaceTiger did his PhD on a certain large structure or set of large structures.

I had read a Yahoo article from Space.com, so I went to ESO's site for the actual story. I thought it might be of interest here, and that more knowledgeable folks here could provide more insight into the significance of the 'new' discovery and how it fits with the 'known' web/skeleton.

 

[turbo]

There have been suggestions that the filamentous structures could imply a fractal distribution of matter, as well. Self-similarity might not play well with the observed near-by Fingers of God effect, nor with the Kaiser effect, which is observed on more distant systems. Both effects might point to an inconsistency or unexplained contribution to redshift that is not normally assumed by the effects expected from the Hubble distance/redshift relation, nor the contributions explainable by the Doppler effects resulting from the peculiar motions of the galaxies.

 

[Sorry!]

There have been suggestions that the filamentous structures could imply a fractal distribution of matter, as well. Self-similarity might not play well with the observed near-by Fingers of God effect, nor with the Kaiser effect, which is observed on more distant systems. Both effects might point to an inconsistency or unexplained contribution to redshift that is not normally assumed by the effects expected from the Hubble distance/redshift relation, nor the contributions explainable by the Doppler effects resulting from the peculiar motions of the galaxies.

Woah turbo I never knew you knew stuff about cosmology .

@Astro. Sorry I thought you meant that the new information was suggesting our universe is of what structure.

 

[Wallace]

There have been suggestions that the filamentous structures could imply a fractal distribution of matter, as well. Self-similarity might not play well with the observed near-by Fingers of God effect, nor with the Kaiser effect, which is observed on more distant systems. Both effects might point to an inconsistency or unexplained contribution to redshift that is not normally assumed by the effects expected from the Hubble distance/redshift relation, nor the contributions explainable by the Doppler effects resulting from the peculiar motions of the galaxies.

On the contrary, the web like structure has long been a prediction from simulations assuming the standard model. The continued observational evidence for its existence is further evidence in favour of the standard model then, and not evidence for anomalies or inconsistencies as you suggest.

 

[turbo]

On the contrary, the web like structure has long been a prediction from simulations assuming the standard model. The continued observational evidence for its existence is further evidence in favour of the standard model then, and not evidence for anomalies or inconsistencies as you suggest.

Filamentous structure was discovered through surveys, as was the existence of "walls" of galaxies and voids. Once such structures are observed, cosmologists employ computer simulations to see if the existence of the structures can be accommodated within current models, including some hierarchical models of matter formation. Plausible accommodation should not be construed as "prediction".

The apparent distortion of clusters is another matter entirely. When we map nearby clusters of galaxies, we find that galaxies near the center of the clusters are preferentially blueshifted WRT to the other cluster members. If we map the galaxies' distances using their redshifts, this gives us maps in which the cluster assumes the appearance of a wedge, with the central galaxies at the cusp, and pointed directly at us (the observers). This was not expected. It is assumed that in a virialized cluster, smaller members held in the gravitational sway of the larger members would have a wide range of peculiar motions, including some with substantial motion toward us or away from us (blueshifted and redshifted, respectively). This is not observed. Instead, we get the Fingers of God effect in which the redshift distributions of the cluster members make them appear to be distributed preferentially with respect to the observer. Central members closer to us (using redshift=distance model) and outlying members farther from us. The opposite effect (Kaiser effect) is observe in distant clusters, with the cluster maps assuming a flattened shape.

 

[Wallace]

Filamentous structure was discovered through surveys, as was the existence of "walls" of galaxies and voids. Once such structures are observed, cosmologists employ computer simulations to see if the existence of the structures can be accommodated within current models, including some hierarchical models of matter formation. Plausible accommodation should not be construed as "prediction".

Sorry Turbo, but this is simply incorrect. N-body simulations were predicting the cosmic web a decade before we were able to see it is things like the 2DFGRS and SDSS surveys. It's a basic prediction of not just LCDM but more broadly any FRW like model with an effective tilt in the power spectrum of fluctuations in the range that we observe. The earliest simulations in the 1980's (e.g. Efstathiou and collaborators) showed this kind of structures, long before we had large galaxy surveys.

The apparent distortion of clusters is another matter entirely. When we map nearby clusters of galaxies, we find that galaxies near the center of the clusters are preferentially blueshifted WRT to the other cluster members. If we map the galaxies' distances using their redshifts, this gives us maps in which the cluster assumes the appearance of a wedge, with the central galaxies at the cusp, and pointed directly at us (the observers). This was not expected. It is assumed that in a virialized cluster, smaller members held in the gravitational sway of the larger members would have a wide range of peculiar motions, including some with substantial motion toward us or away from us (blueshifted and redshifted, respectively). This is not observed. Instead, we get the Fingers of God effect in which the redshift distributions of the cluster members make them appear to be distributed preferentially with respect to the observer. Central members closer to us (using redshift=distance model) and outlying members farther from us. The opposite effect (Kaiser effect) is observe in distant clusters, with the cluster maps assuming a flattened shape.

This is also not at all true. The 'Fingers of God' simply refers to line of sight redshift space distortions, which are observed in surveys in much the same way as you get from simulations (or even just expectations from perturbation theor). We have even used these red-shift space distortions to learn some reasonably accurate things about dark energy. See for instance the Nature paper by Guzzo et al. There is no discrepancy between theory and observations along the lines of what you are suggesting. If you have any references to the contrary please share them, but I could give you dozens of references looking at redshift space distortions and their value to cosmology, and in no study I've seen has anything like this been shown to be a significant issue.

 

[twofish-quant]

On the contrary, the web like structure has long been a prediction from simulations assuming the standard model. The continued observational evidence for its existence is further evidence in favour of the standard model then, and not evidence for anomalies or inconsistencies as you suggest.

The interesting thing about science is what constitutes the "standard model" changes over time. In the 1980's, there were two competing models for the cosmology. Cold dark matter and hot dark matter. If the dark matter is cold, then you should see very complicated structures like what we do see. If the dark matter is hot, then you shouldn't because the thermal movement of the dark matter would wipe out these structures. (Imaging throwing an ice crystal into boiling water.)

Physics cage fighting. Two competing theories. Look at the observations. One theory dies. The other gets the "standard model" crown.

 

[Wallace]

Depending on the temperature you assume for the hot dark matter, you'd still see filamentary like structures forming (unless it was really hot). Really, it is quite a generic prediction of non-linear gravitational structure formation in an expanding Universe, it doesn't actually tell you that much about the physics of say dark energy or dark matter etc since you get this kind of structure in a very wide parameter range.

 

[twofish-quant]

Also it's not just filament structure, but the *type* of filament structure. You run statistics on the the N-body simulations, get a number. Run statistics on observations, get another number. See if the numbers match.

It's really important to use statistics to do these sorts of comparisons since "gee the pictures look the same/different" doesn't work very well.

Also IIRC the statistics that you see are pretty clearly non-fractal, as the power spectrum doesn't show any self-similarity.

 

[twofish-quant]

Depending on the temperature you assume for the hot dark matter, you'd still see filamentary like structures forming (unless it was really hot).

Yes but the structures start looking very different. They start "melting".

it doesn't actually tell you that much about the physics of say dark energy or dark matter etc since you get this kind of structure in a very wide parameter range.

The fact that you get structures doesn't tell you very much. The detailed statistical properties of the structures look like tells you a great deal. One of the really important things about the standard model is that it doesn't merely tell you that you get structures, it tells you in pretty large detail what those structures are going to look like for a given set of input parameters.

One probably with popular descriptions is that they try to avoid math, which means that the miss the point that the standard model can predict in a lot of detail not only that there are structures but what those structures look like.

 

[Wallace]

Exactly right, we shouldn't overstate the importance of this whole filamentary structure thing, it comes up a lot in press releases (because it makes pretty pictures) but it's a very broad qualitative description of the density field. As you say, detailed statistical properties such as the power spectrum, bi-spectrum, halo mass function etc are the things that are studied in simulations and compared to observations.

And yes you're right, we don't see a fractal bevhaviour or self similiarity in the distribution of material in the Universe. Turbo erronously suggested that the filamentary structures seen have some bearing on the Universe having a fractal distribution of matter (see post #9) which is spurious. You could have a web-like non-fractal distribution or a non-web like fractal one, the two are unrelated.

 

[twofish-quant]

Exactly right, we shouldn't overstate the importance of this whole filamentary structure thing, it comes up a lot in press releases (because it makes pretty pictures) but it's a very broad qualitative description of the density field.

One thing that someone could do (and someone really should do) is to set up some sort of web application in which people can punch in a power spectrum, and see what it looks like. At that point, you can do things like, "this is what the universe looks like", "this is what the universe would like like if it were five billion years old", "this is what the universe would look like if there were no dark matter."

On problem with "eyeballing" a picture, is that you don't know whether things look the same or different because they really are different, or if it's because of the way you drew the picture. The other problem is that if you think that something looks similar, and someone else says that it looks different, then there's no real way of saying who is right.

As you say, details statistical properties such as the power spectrum, bi-spectrum, halo mass function etc are the things that are studied in simulations and compared to observations.

And I should point out that all of this stuff is really useful in finance. If you plot out stock prices, all you get are wiggly lines. There is a lot of statistics involved in telling whether one wiggly line behaves the same or different than another wiggly line.

And yes you're right, we don't see a fractal bevhaviour or self similiarity in the distribution of material in the Universe.

One thing that I've found is that people use "fractal" in popular language to mean something that it doesn't mean in mathematics. A "fractal" is a specific type of geometric structure. The distribution of matter in the universe is not "fractal" at all.

Also once you understand what a "fractal" looks like, it's pretty clear that the universe is not one.

 

[turbo]

On the contrary, the web like structure has long been a prediction from simulations assuming the standard model. The continued observational evidence for its existence is further evidence in favour of the standard model then, and not evidence for anomalies or inconsistencies as you suggest.

Funny! When Geller and Huchra discovered the "Great Wall" in 1989, it came as quite a surprise. Somehow the existence of thin, wall-like structures wasn't "predicted" by the standard model. Until filaments were observed I doubt that they were predicted either. At the time of Geller and Huchra's survey, it was commonly believed that superclusters were the largest structures, and that they were distributed quite uniformly. That was only 20 years ago.

 

[turbo]

And yes you're right, we don't see a fractal bevhaviour or self similiarity in the distribution of material in the Universe. Turbo erronously suggested that the filamentary structures seen have some bearing on the Universe having a fractal distribution of matter (see post #9) which is spurious. You could have a web-like non-fractal distribution or a non-web like fractal one, the two are unrelated.

Huh! I "erroneously suggested" that there appears to be a fractal-type distribution of matter in the universe? There is a large, well referenced body of work on this subject. It doesn't take a lot of research to pull up papers on this subject.

Concordance cosmologists are pretty sold on an isotropic and homogeneous universe on very large scales, but that doesn't mean they are correct. Mapping of galaxy clusters and filaments is pretty much limited by selection-effects, since distant clusters are very difficult to get spectroscopy on, and their fainter members get undetectable pretty fast, leaving only the brighter outliers. Flattening of any fractal pattern at great distances could be a sign of homogeneity, or it could be a statistical artifact of selection effects in faint observations. We need more data.

 

[twofish-quant]

Funny! When Geller and Huchra discovered the "Great Wall" in 1989, it came as quite a surprise. Somehow the existence of thin, wall-like structures wasn't "predicted" by the standard model.

That's because what people defined as the standard model in 1989 isn't what it people defined it in 2009. In 1989, there were a number of viable cosmological models, some of which predicted large scale structures, some didn't.

The one's that didn't became non-standard.

At the time of Geller and Huchra's survey, it was commonly believed that superclusters were the largest structures, and that they were distributed quite uniformly.

And if that were the case, then what we now consider the standard model of cosmology just wouldn't work, and what we'd be calling the standard model would be some variant of hot dark matter.

That was only 20 years ago.

20 years is an eternity in cosmology. Stuff that is a year old is ancient history.

 

[twofish-quant]

Huh! I "erroneously suggested" that there appears to be a fractal-type distribution of matter in the universe? There is a large, well referenced body of work on this subject.

Pull up three review papers that argue in favor of a fractal distribution. Every power spectrum that I've seen has been very non-fractal. Also you need to define "fractal-type". A fractal has a very well defined mathematical definition.

If you are using the term "fractal-type" you need to define what you mean. If you mean a distribution with power at all scales, you are using terms in very non-standard ways.

It's pretty obvious just looking at the distributions that they aren't fractals. The thing about fractals is they are scale invariant. If you magnify the a Julia set, it looks a lot like another Julia set. If you take one of the three-3 surveys of the universe, magnify it and then superimpose it on the original picture, it looks very different.

Mapping of galaxy clusters and filaments is pretty much limited by selection-effects, since distant clusters are very difficult to get spectroscopy on, and their fainter members get undetectable pretty fast, leaving only the brighter outliers. Flattening of any fractal pattern at great distances could be a sign of homogeneity, or it could be a statistical artifact of selection effects in faint observations. We need more data.

Selection effects doesn't help you here.

I think we have enough data to rule out fractal distributions. The thing about fractals is that they are self-similar. Take a fractal distribution of galaxies. Now put in some sort of selection effect. What you end up with is still a fractal. It's a different looking fractal, but it's still a fractal.

Also non-fractal doesn't mean homogeneity.

 

[twofish-quant]

turbo-l: Mapping of galaxy clusters and filaments is pretty much limited by selection-effects, since distant clusters are very difficult to get spectroscopy on.

But you can get rid of these sorts of effects by not looking at distance but instead looking at angular correlations.

 

[turbo]

That's because what people defined as the standard model in 1989 isn't what it people defined it in 2009. In 1989, there were a number of viable cosmological models, some of which predicted large scale structures, some didn't.

My point is that cherry-picking some models that happen to fit observations after the fact is not equivalent to "prediction", since there was no consensus on viability of the models prior to the observations. With enough freely adjustable parameters, you can make almost any model fit observations, though that is a shaky way to conduct science.

 

[Sorry!]

My point is that cherry-picking some models that happen to fit observations after the fact is not equivalent to "prediction", since there was no consensus on viability of the models prior to the observations. With enough freely adjustable parameters, you can make almost any model fit observations, though that is a shaky way to conduct science.

I don't think this is what happened overall.

Before we made these observations there were scientific theories which predicted a large asortment of things... as we get more precise observations these theories become 'weeded-out'. The ones that are completely wrong are dismissed mostly... the ones that had predictions most like our observations were continued... this elimination process is always occurring. The scientists don't just make an observations and throw in a constant here arbitrarily. (Sometimes they do but they accept that it's just for now.) There are obviously adjustments to theories based on the observations we do make...

If you're suggesting that science must make 100% accurate predictions or else its not science then I don't even know if science exists.

 

[turbo]

If you're suggesting that science must make 100% accurate predictions or else its not science then I don't even know if science exists.

I don't suggest any such thing. Theoretical cosmology relies on astronomical (and other) observations for confirmation, and since astronomy is an observational science, the most potent tool available to cosmologists is predictability. Make some predictions that can be verified by observation, and test the model as data comes in. Retrodictions do not test models, but only serve to separate the cats from the dogs. Show me a model that predicted the discovery of a cosmic void a billion light-years across, not one that contained a plausible fit after the fact.

 

[twofish-quant]

My point is that cherry-picking some models that happen to fit observations after the fact is not equivalent to "prediction", since there was no consensus on viability of the models prior to the observations. With enough freely adjustable parameters, you can make almost any model fit observations, though that is a shaky way to conduct science.

That's precisely how science works. You have model A predict A. Model B predict B. You see A. Well it looks like B is wrong.

Of course there is going to be no consensus on the viability of models before observations. You don't have anything to base the consensus on.

 

[twofish-quant]

I don't suggest any such thing. Theoretical cosmology relies on astronomical (and other) observations for confirmation, and since astronomy is an observational science, the most potent tool available to cosmologists is predictability.

Hot dark matter predicts no structure. Cold dark matter predicts structure. We see structure. Looks like hot dark matter is wrong.

Show me a model that predicted the discovery of a cosmic void a billion light-years across, not one that contained a plausible fit after the fact.

CDM was first published in 1984. You didn't see observations of voids until 1989.

 

[turbo]

That's precisely how science works. You have model A predict A. Model B predict B. You see A. Well it looks like B is wrong.

It's not that simple, as you know. There can be a plethora of models, so either-or is a vast over-simplification. Also any number of models can remain viable, at least marginally, if they can be patched to conform to observations.

Of course there is going to be no consensus on the viability of models before observations. You don't have anything to base the consensus on.

That's the point with cosmology, isn't it. There are lots of possibilities, and observations have to constrain them and weed out the dogs. It is disingenuous for cosmologists to weigh in after the fact with a surviving model and claim that cosmology is predictive, when in fact it is not.

 

[turbo]

CDM was first published in 1984. You didn't see observations of voids until 1989.

Please show me how CDM predicted voids a billion light-years across, much less the 3.5 Gly void turned up by the 6dF Galaxy Survey (Very current science as surveys go.) Fitting models to observations retroactively is SOP for cosmologists, and it is not a sign that the field is well-developed, much less mature. I won't go as far as Michael Disney in deriding the field, but I agree with him that there are some serious weaknesses regarding cosmology as a hard science.

I do have a dog in this fight, since my collaborators and I are studying and publishing about redshift distributions in interacting galaxies. It is enlightening to find out how strongly theory and politics can trump observations, and even inhibit observations to the point where decent statistical analyses are difficult to perform due to a paucity of observations. It would shock you to find out how many relatively bright galaxies have no published redshift data simply because they are part of an interacting system. Arp's banishment has had a chilling effect for decades.

 

[twofish-quant]

Please show me how CDM predicted voids a billion light-years across, much less the 3.5 Gly void turned up by the 6dF Galaxy Survey (Very current science as surveys go.)

CDM is a framework. You can use cmbfast to set the parameters to get you those voids. You might complain that you can adjust the parameters to get realistic looking power spectrum, but I don't know of any other framework that gets you even that far.

Fitting models to observations retroactively is SOP for cosmologists, and it is not a sign that the field is well-developed, much less mature.

1) I thought science was all about fitting models to observations retroactively.

2) Well cosmology *isn't* a particularly well-developed, mature field. It's not like say classical mechanics or electromagnetism where we think we pretty much understand everything, or even particle physics were it's been at least a decade since anyone has observed anything that doesn't fit in our models.

The fact that we are invoking inflation, dark matter, and dark energy to get models that approach observations is a clear sign that we really don't know what's going on. And this is a problem because?

I do have a dog in this fight, since my collaborators and I are studying and publishing about redshift distributions in interacting galaxies. It is enlightening to find out how strongly theory and politics can trump observations, and even inhibit observations to the point where decent statistical analyses are difficult to perform due to a paucity of observations.

You seem to be inconsistent here. You were just complaining that people are modifying their theories to fit observations, and now you are talking about how theory trumps observations.

I don't understand this talk of lack of statistical analysis. Most of the evidence in favor of CDM involves coorelation functions and power spectrum, and I don't see a lack of observations. I'd be interested in hearing what objections you have to the statistics that is used for galactic observations, but I don't think that you can plausibly argue that that don't exist or that there is a lack of data. I'd like to see you try...

It would shock you to find out how many relatively bright galaxies have no published redshift data simply because they are part of an interacting system. Arp's banishment has had a chilling effect for decades.

The trouble is that if you take small numbers of bright galaxies, you just don't end up with enough data to get decent coorelation functions and power spectrum, and then you end up wondering about selection effects. The trend over the last decade or two has been to take massive number of measurements so that any selection effects are overwhelmed by sheer numbers. If you don't have a human being in the loop deciding which galaxies to measure and which one's not to, then it's much, much easier to model statistical bias and get good statistics. Once you put a human being in the loop to decide what to measure and what not to measure, it becomes more difficult.

But the idea that we don't have enough statistics and observations to do cosmology is one that I find very, very odd.

The other thing is that if your argument is that we ought to be interested in interacting galactic systems because the redshifts suggest that something weird is happening there, that that suggests to me that when doing large scale cosmological surveys that one should try to *exclude* those galaxies, since whatever is causes weird stuff in those system is going to bias the thing that people are interested in studying.

Also, if you do have reason to believe that "weird stuff" is happening in an interacting cluster, then at that point I think that it's a bad idea to use statistics because since you don't know what the "weird stuff" is or if it is the same "weird stuff" and you'll end up mixing apples and oranges and not figuring out what is going on.

 

[Astronuc]

Perhaps it would be worthwhile to look at James (Jim) Peebles text, Large-Scale Structure of the Universe, Princeton University Press, 1980 to see where the understanding (models) was then (ca 1970's) and where we are now.

https://www.amazon.com/gp/product/0691082405/?tag=pfamazon01-20

 

[turbo]

Perhaps it would be worthwhile to look at James (Jim) Peebles text, Large-Scale Structure of the Universe, Princeton University Press, 1980 to see where the understanding (models) was then (ca 1970's) and where we are now.

https://www.amazon.com/gp/product/0691082405/?tag=pfamazon01-20

Thanks for that link. The first chapter lays out the history of the study of large-scale structure up to the 1970's, which is handy. It seems that cosmologists were expecting homogeneity at some large scale, but kept discovering clumping at larger and larger scales as time went by. Things haven't changed much, since we keep finding structures.

6dF turned up a void 3.5 billion light-years across, so there is currently evidence of large-scale structure at least that large, and there may be much larger structures that we can observe, given improvements in instrumentation, including better optics and greater detector sensitivity. With voids framed by filaments and walls of galaxy clusters, the distribution of matter in space takes on a foamy appearance. A challenge for modern cosmology could arise from the discovery of structures on larger and larger scales, because the formation of structure through gravitational accretion takes time, and the Big Bang theory contains a self-imposed limit on the time available in which structure can form.

 

[twofish-quant]

A challenge for modern cosmology could arise from the discovery of structures on larger and larger scales, because the formation of structure through gravitational accretion takes time, and the Big Bang theory contains a self-imposed limit on the time available in which structure can form.

True, and one reason that we have "warm fuzzies" about the big bang is that there is this observed massive cutoff in power spectra at very large angles. One reason why we think that "dark matter" exists is that if it didn't, then then cutoff for structure happens way, way below where we observationally observe it. If you don't have dark matter, the universe expands quickly, and you don't have time for structure to form. The more dark matter there is the universe, the slower the universe expands and the more time there is for structures to form.

From the location of the cutoff, you can calculate where to set the parameters for the LCDM model, and you then get "warm fuzzy feelings" because where you set those parameters from structure happens to be where you set them up from completely different calculations (such as big bang nucleosynthesis). Basically by looking at the distribution of galaxies, you can figure out how much dark matter there is in the universe, and if that number matches the other estimates of dark matter, then you think you have some clue about what's going on. If you have different estimates of dark matter that give wildly different numbers, then there is a good chance that your basic model is wrong.

Also, one big advance is that we now have cheap computers in which we can use with LCDM to make very, very detailed predictions about what the universe looks like.

One big problem with popular descriptions about what is going on is that without numbers these sorts of arguments are impossible. Yes there is a cutoff in the big bang. If you talk about power spectra, you can figure out where that cutoff is, and if it is consistent with observations.

 

[twofish-quant]

6dF turned up a void 3.5 billion light-years across, so there is currently evidence of large-scale structure at least that large, and there may be much larger structures that we can observe, given improvements in instrumentation, including better optics and greater detector sensitivity.

It's fairly unlikely. You can do angular correlations, and if you do that, you'll find that there is a huge dropoff in the power spectrum once you exceed a certain angle. One problem with doing distance correlations is that you don't know whether or not you are seeing a real correlation or some galaxy evolution or observational selection effect.

With voids framed by filaments and walls of galaxy clusters, the distribution of matter in space takes on a foamy appearance.

Which is pretty inconsistent with anything that has been suggested that is wildly different from the big bang. Foams aren't fractals. If the distribution of the galaxies were determined by something like plasma physics, you'd get a cascade at all scales in which you have something that looks nothing like a foam.

A challenge for modern cosmology could arise from the discovery of structures on larger and larger scales, because the formation of structure through gravitational accretion takes time, and the Big Bang theory contains a self-imposed limit on the time available in which structure can form.

Yes, but remember that all this was written in the 1970's. Things have progressed since then.

So people in the 1980's and 1990's were specifically looking for structures that could challenge the big bang. The stuff that they did find required some model tweaking, killed a few scenarios, but ended up not challenging the BB in a major way. Now it is true that you have to assume three "tooth fairies" to get everything to work (inflation, dark matter, and dark energy), but that's a lot better than any alternative theory. Once you assume these three tooth fairies, you get massively detailed predictions about how the galaxies are distributed, abundances of elements, and the characteristics of the cosmic microwave background. The fact that people outside the field aren't aware of this, is more a failing of science journalism than anything else.

The problem is that you just can't wave your hands and say "plasma physics" and get people to take you seriously, because we have enough data that you really have a huge job trying to explain how something really weird explains the data better than any of the theories that you have.

One big problem that you have if you start invoking plasma and magnetic fields is that what tends to happen when you have plasma and magnetic fields is that structures start breaking up. You get what is known as a Kormogarov cascade in which big eddies start breaking up and forming smaller eddies, and you end up with a hell of a time trouble to explain how you get big voids from that. If you take things that we know are due to plasma and magnetic fields (which you can do in the laboratory), you come up with a specific power spectrum that looks totally different from what we see in the galaxies.

 

[turbo]

Yes, but remember that all this was written in the 1970's. Things have progressed since then.

That observation regarding the time-budget for formation of structure was my own, not from Peebles, and was made in light of some fairly recent (6dF) observation.

So people in the 1980's and 1990's were specifically looking for structures that could challenge the big bang. The stuff that they did find required some model tweaking, killed a few scenarios, but ended up not challenging the BB in a major way. Now it is true that you have to assume three "tooth fairies" to get everything to work (inflation, dark matter, and dark energy), but that's a lot better than any alternative theory. Once you assume these three tooth fairies, you get massively detailed predictions about how the galaxies are distributed, abundances of elements, and the characteristics of the cosmic microwave background. The fact that people outside the field aren't aware of this, is more a failing of science journalism than anything else.

If we need to invoke DM, we have to do so in ways that allow it to "fix" observations in various scenarios, including the too-flat rotation curves of spiral galaxies, and the too-large binding energy of clusters. So we not only have to believe in that tooth fairy, but also believe that it is distributed "just so" to do its magic. It's not "clean".

The problem is that you just can't wave your hands and say "plasma physics" and get people to take you seriously, because we have enough data that you really have a huge job trying to explain how something really weird explains the data better than any of the theories that you have.

One big problem that you have if you start invoking plasma and magnetic fields is that what tends to happen when you have plasma and magnetic fields is that structures start breaking up. You get what is known as a Kormogarov cascade in which big eddies start breaking up and forming smaller eddies, and you end up with a hell of a time trouble to explain how you get big voids from that. If you take things that we know are due to plasma and magnetic fields (which you can do in the laboratory), you come up with a specific power spectrum that looks totally different from what we see in the galaxies.

I think you're confusing me with someone else. I never invoked any plasma cosmology. Anywhere, any time. Nor do I think that is a productive line of research.

 

[twofish-quant]

Here is a link that talks about the power spectrum

http://www.astro.caltech.edu/~george/ay21/eaa/eaa-powspec.pdf

The reason people believed in the 1970's that there wouldn't be any large scale structure was that the general belief was that the universe consistent only of baryons. If you assume that the universe consists only of baryons then what happens is that the baryons collide and clump pretty quickly, so you don't get any large structures.

If you assume CDM the you get a closer match to what you actually see. It's not an exact match, but it's close enough so that you can argue that your basic picture is right and the differences are due to details. Now if galaxies formation was the *only* reason to believe that CDM exists, then it would smell fishy, but there are other reasons to think that it's there. So it kind of makes sense.

 

[twofish-quant]

If we need to invoke DM, we have to do so in ways that allow it to "fix" observations in various scenarios, including the too-flat rotation curves of spiral galaxies, and the too-large binding energy of clusters. So we not only have to believe in that tooth fairy, but also believe that it is distributed "just so" to do its magic. It's not "clean".

Reality is messy. If you have any better ideas, let's hear them. Whenever someone comes up with a "clean model" it usually means that they don't have much useful data come in.

The fact that dark matter is a pretty large tooth fairy is why people are interested in modified gravitation models. The good thing about dark matter is that you just have to involve one tooth fairy with really one parameter, and it fixes a lot of things. LCDM has basically nine or ten parameters, and you get an incredible amount of stuff out if you have those parameters.

I think you're confusing me with someone else. I never invoked any plasma cosmology. Anywhere, any time. Nor do I think that is a productive line of research.

Curiously I think that plasma cosmology *is* a productive line of theoretical research. The good thing about plasmas is that 1) we know they exist and 2) there are all sorts of weird effects associated with plasmas, so I wouldn't be that surprised at all of there was some weird effect that explains a lot. I wouldn't write my Ph.D. dissertation around plasma cosmology, but to use a financial analogy, weird stuff like that is like a lottery ticket that might make you a millionaire if you are lucky.

When I say "here are the problems with plasma cosmology" it's not to end discussion, it's to further it.

 

[turbo]

Reality is messy. If you have any better ideas, let's hear them. Whenever someone comes up with a "clean model" it usually means that they don't have much useful data come in.

The fact that dark matter is a pretty large tooth fairy is why people are interested in modified gravitation models. The good thing about dark matter is that you just have to involve one tooth fairy with really one parameter, and it fixes a lot of things. LCDM has basically nine or ten parameters, and you get an incredible amount of stuff out if you have those parameters.

Modified gravity would be "clean" in the sense that no additional entities (tooth fairies) would be needed. Consider that G is not a constant, but is variable, increasing with the average matter-density of the space in which G is to be calculated between interacting masses. Scaled appropriately, variable G could eliminate the "missing mass" problems observed on galactic and cluster scales. We already have lots of estimates of mass densities of galaxies and clusters, and those are used to compute an apparent deficit in the mass required to produce the observed gravitational effects, which is then used as a quantification of the DM required. What would we get if we instead scaled up G in every instance until the observed gravitational effects are reproduced and the mass-deficit is erased? Variable G is probably not going to become popular anytime soon, since there is so much momentum in LCDM, but it is simple and clean with no additional entities or parameters.

 

[turbo]

I followed your link to the power-spectrum paper and from there to another on structure formation.

Galaxy formation stutters into action around z ∼ 5. Only a tiny fraction of the stars present today would have formed prior to that time. By z ∼ 3, the epoch when galaxies isolated by the ‘Lyman-break’ technique1 are observed, galaxy formation has started in earnest, even though only 10 per cent of the final population of stars has emerged. The midway point is not reached until about a redshift of 1–1.5, when the universe was approximately half of its present age.

This model is already ruled out by SDSS observations of quasars from z~5.7-6.5. Since luminosity falls off as a square of the distance, if these high-z quasars are at the distances implied by their redshifts, they would have to be the most luminous objects in the universe, powered by accreting BHs of billions of solar masses, feasting on host galaxies of perhaps a trillion solar masses or more. How did such beasts form only a few hundred million years after recombination? Another puzzle is that all of these quasars feature Solar or super-Solar metallicities with no observed evolution in either absolute nor relative metallicities. Moreover, despite the huge column densities to such distant objects not a single one was found to be lensed. These observations are explained in detail in Michael Strauss' presentation to the Space Telescope Science Institute on 11/02/2005. You will enjoy the talk.

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

 

[twofish-quant]

Consider that G is not a constant, but is variable, increasing with the average matter-density of the space in which G is to be calculated between interacting masses. Scaled appropriately, variable G could eliminate the "missing mass" problems observed on galactic and cluster scales.

The trouble with that is that that means that you have increased gravity in the early universe which means that all of the things that seem to match get blown to hell. BBN nucleosynthesis breaks. Structure formation breaks. You end up with too little deterium, you end up with galaxies that are too clumpy, etc. etc.

What you really want (and you get that with some modified Newtonian and f(R) models) is gravity that increases when you have large distances. The thing about f(R) models is that they make some concrete predictions about structure formation which last I've heard, don't match.

What would we get if we instead scaled up G in every instance until the observed gravitational effects are reproduced and the mass-deficit is erased?

Because in order to match observations you end up having to have a different gravity law for every single galaxy and cluster you are trying to measure.

Variable G is probably not going to become popular anytime soon, since there is so much momentum in LCDM, but it is simple and clean with no additional entities or parameters.

It's not social momentum, but fitting things to observations. Once you try to fit modified gravity to observation it becomes you end up with something that's more complex and with a lot more parameters than LCDM. People have tried. People have tried very hard.

If someone wrote a paper that says we assume this one crazy thing about G and everything works, then that would turn heads (which is what happened with inflation). There has been a lot of effort in getting this to work, and no one has been able to come up with anything. The "problem" (and I use that term ironically since it's a good thing) is that we have just too much damn data.

 

[twofish-quant]

Since luminosity falls off as a square of the distance, if these high-z quasars are at the distances implied by their redshifts, they would have to be the most luminous objects in the universe, powered by accreting BHs of billions of solar masses, feasting on host galaxies of perhaps a trillion solar masses or more. How did such beasts form only a few hundred million years after recombination?

Since CDM implies top-down formation of galaxies, I don't see the problem.

Another puzzle is that all of these quasars feature Solar or super-Solar metallicities with no observed evolution in either absolute nor relative metallicities. Moreover, despite the huge column densities to such distant objects not a single one was found to be lensed.

We are getting into the details of galaxy formation which is a process that no one quite understands, so yes, there are huge numbers of things that make no sense right now. And this means?

 

[turbo]

Because in order to match observations you end up having to have a different gravity law for every single galaxy and cluster you are trying to measure.

Same law, but with a variable G instead of a constant. We already quantify the mass-deficit to adjust the projected DM to fit observations. Leave DM out of it and scale G until the observed gravitational effects are fit. Do this for galaxies, loose groups, clusters, etc, rinse and repeat. If the relationship doesn't hold for clusters of similar masses and velocity dispersions, then it's back to the drawing board, but I'd like to see the consistency of the fit, and whether it can be used to predict the behavior of structures that have not previously been modeled.

 

[turbo]

Since CDM implies top-down formation of galaxies, I don't see the problem.

the author that you quoted on power-spectra had linked to the paper that I quoted, citing bottom-up formation, with galaxy formation only getting underway by z~5 or so. The apparent existence of trillion-solar-mass galaxies at z~6.5 seems a bit problematic, doesn't it? Anyway, if you have the bandwidth and the time, I highly recommend Strauss' presentation. He is the science spokesperson for SDSS and his presentation is clear and entertaining.

Edit: Here is the link to that 2006 paper citing bottom-up hierarchical matter formation.

http://www.astro.caltech.edu/~george/ay21/eaa/eaa-strucformsims.pdf

 

[Chronos]

My point is that cherry-picking some models that happen to fit observations after the fact is not equivalent to "prediction", since there was no consensus on viability of the models prior to the observations. With enough freely adjustable parameters, you can make almost any model fit observations, though that is a shaky way to conduct science.

That sounds suspiciously similar to your tenants. Do you have a ground breaking revelation in mind, or are satisfied by sniping at mainstream interpretations?

 

[twofish-quant]

Leave DM out of it and scale G until the observed gravitational effects are fit.

Been tried. Look up "modified Newtonian dynamics" (MOND), there are problem several dozens (if not hundreds) of papers that are trying to do that. The problem is that what you end up with looks (at this point) to have more tooth fairy than LCDM.

One problem is trouble is that you can't get a galaxy rotation curve for even one galaxy to work right without assuming a custom gravitational law. Now if you try multiple galaxies and clusters then it turns out that you have to put in a custom power law for each galaxy and cluster and what's worse, no one has been able to come up with a rule that says what law of gravity works with what galaxy and cluster.

Maybe someone will, but it's pretty clear that it's not as simple as "just change G." No one has been able to get anything simple to work.

If the relationship doesn't hold for clusters of similar mases and velocity dispersions, then it's back to the drawing board, but I'd like to see the consistency of the fit, and whether it can be used to predict the behavior of structures that have not previously been modeled.

Do a literature search for MOND. It's pretty clear that anything really simple won't work. Now you can make any model work by making it sufficiently complicated, but a complicated theory has no predictive power. Right now (and things could change in the next few years), LCDM ends up being far, far simplier and more predictive than any modified gravity model that anyone has come up with.

One problem is that MOND is right now at the level of just trying to come up with a coherent gravity rule. As far as I know, no one has even thought about the implications of MOND to CMB anisotropy and galactic structure formation.

 

[twofish-quant]

the author that you quoted on power-spectra had linked to the paper that I quoted, citing bottom-up formation, with galaxy formation only getting underway by z~5 or so. The apparent existence of trillion-solar-mass galaxies at z~6.5 seems a bit problematic, doesn't it?

So the writer of the paper is wrong and you have incredibly enormous galaxies at z=6.5. What's the problem? Galaxy formation is something that no one has any clue about, so if you argue that some one's scenario of galaxy formation is bogus, that's not going to bother me or anyone else right now.

You'll only run into problems if you end up arguing that whatever caused galaxy formation had structures that were wildly different from the initial distribution of matter.

 

[twofish-quant]

Right now the really big focus of interest in cosmology is in the "dark ages". The period between CMB and the first observable galaxies. Since you can't see anything, you don't have any data, since you have no data, you can make up lots of stuff most of which is wrong.

Data is coming in, so we'll have a much better idea of what's going on in about a year or two.

About how bad science journalism is in relation to cosmology, there is this very unhealthy obsession over "what happened before the big bang" and barring something really weird, I doubt there is going to be much progress on that over the next two years. Also journalists have this other obsession with "young maverick overturns deep established truths" (which I think is a product of the Vietnam War era). This causes science journalists to tend to ignore parts of science where there are no established truths. Right now, you can make up any galaxy formation scenario you want, and people will nod their heads because there is a huge lack of data. Once more data comes in, most of those scenarios will end up in the trash can.

 

[turbo]

Been tried. Look up "modified Newtonian dynamics" (MOND), there are problem several dozens (if not hundreds) of papers that are trying to do that. The problem is that what you end up with looks (at this point) to have more tooth fairy than LCDM.

In MOND, the constant G still is a constant as far as I know (there may be other flavors), though it is assumed that the forces derived from Newtonian dynamics do not hold in low-acceleration regimes. Of course MOND doesn't work on cluster scales, so the range of applicability is limited.