Occam's Machete: Cosmology

Much of modern cosmology is built on the idea that the universe is uniform everywhere and in all directions. If it wasn’t uniform that would be a pretty interesting discovery. So far there is precious little evidence that contradicts nearly perfect uniformity at large scales. But it may be very hard to detect non-uniformity. There are several theories about the cosmos beyond our universe that would seem to predict effects we could look for. The telltale signature of some of these theories would be circular anomalies in the CMB and a few have been posted about on this blog here, here, and here.

For example what if the cosmos is filled with many big bangs and our universe is only one bubble in a foam of universes? That is the central question in the idea of Eternal Inflation. A lot of people have thought about the implications of such a model and many have claimed to have seen evidence of such universe collisions. Usually when such claims are made established cosmologists come down hard on the observations as not supported by the data. But how do we decide if an observation is supported or not?

A team comprised of members of the University College London, the Imperial College London, and The Perimeter Institute have tackled the question. They start with the assumption that a hypothetical collision would result in some sort of disk-like structure in a 2D picture of the universe like the WMAP image of the CMB. They then simulate the structure and try to find the most conservative measure for detection. In other words they try to find the 2-sigma rule for detection of a disk. They also define a Bayesian parameter metric for the probability that a detected anomaly is real and not just a trick of chaos. The full paper is here.

This picture shows four of some 10 candidates for possible collision sites in the WMAP data that they found of interest. Two of these sites have been noticed before, one is the famous ‘cold spot’ visible to the naked eye (at least if the naked eye could see microwaves) and another, a hot spot, is described here. Their conclusion is that there is no 2-sigma observation of a disc but that there are some candidates that are never the less significant. Here is a picture of the candidates.

I can’t help thinking that the CMB is a snapshot taken at the time of inflation at the very beginning of the universe when it grew to almost its present size. There are other observations of the universe that reflect effects that have happened since inflation slowed down. For instance speed and direction of movement as in ‘dark flow’ or other observations of behavior seen since the universe emerged from the dark time with the formation of the first stars. I wonder if similar techniques could be applied to that data because if there were a collision during inflation, it would have continuing for the rest of the 13 some odd billion years the universe has been around and should be observable in other ways. Even without using observations of more recent events the authors look forward to applying their techniques to the coming Planck satellite CMB data.

This paper is good for taking a step towards refining how to see order in the chaos and sheds light on how to measure the difference.

There is a spot in the sky at the edge of the observable universe where galaxies are moving towards something that cannot be observed and is probably beyond the edge of what we can observe. I’ve written several times (here and here) about dark flow since it was observed in 2008 by Kashlinsky and about one possible explanation for it here. In the time since dark flow was originally observed, the Kashlinsky data has been vetted pretty well.

But what is causing it? Kashlinsky thinks that a dense region of the universe exists just outside our observation horizon and imagines that this implies a non-uniformity to the universe, one that is great enough to call into question the Lambda Cold Dark Matter model. This model leads to a very uniform universe caused by early hyper inflation, just after the big bang. The CMB shows a pretty uniform universe in agreement with LCDM. So Kashlinsky is casting doubt on LCDM, basically saying that the universe may be very lumpy we just can’t see the lumps in the CMB from here.

Enter Dai et al who use a fundamentally different method to measure the flow of individual objects. They examine the type 1A supernova data to see if there is a relationship between the very far away novas (those with a high red shift) and their velocity towards the path of dark flow; they find no correlation, which means that the furthest away novas aren’t really moving faster than the closer ones in the path of the flow. Furthermore they find the likely velocity of the flow to be much smaller than the Kashlinsky suggests. Kashlinsky measured an aberrant bulk flow of more than 600 kilometers a second, while Dai et al found only 188 kilometers a second. Dai et al observe that this velocity is very close to that predicted by LCDM.

So what is one to conclude? Well Dai is measuring individual objects to see if there is a generalization to make about their movement and Kashlinsky is looking at the movement of great globs of matter. Also the methods they use are different. The main thing that the new data questions is whether the observations imply LCDM is wrong. Even within LCDM predictions something is going on out there that has yet to be explained very well. So we continue to observe and perchance to speculate…

Image courtesy of universe-review.ca which (though really slow) has an incredible array of photos and illustrations.

I love anomalies, the thrill of thinking that half of what we know is wrong and the other half is suspect. They hold the promise of new discoveries and a newer better understanding of reality. Unfortunately the vast majority of observations in cosmology, which appear to break the rules, do not stand the test of deep scrutiny.

I have written about the Cosmic Microwave Background (CMB) radiation several times, usually in the context of one anomaly or another. Most of them have since been deemed to be questionable. We are talking about tiny variations from the ideal random fluctuations and the sensitivity of the experiments is often pushed beyond its limits. It is not surprising that many conclusions drawn from the data are, well somewhat speculative. That said who would fail to find their possibilities intriguing?

Two new papers have been published describing ripples seen in the CMB. The most recent arXiv post by Stephen M. Feeney, et al, is based on some implications of Eternal Inflation. The model states that our universe banged then inflated quickly and so do other universes. As other universes (or false vacuum bubbles in the jargon) blow up they may slam against ours causing bruises in the CMB. They analyze data from WMAP with special software that looks for the telltale signs of these bruises.

In this image they show an idealized collision, the temperature modulation, a high needlet response, and results of edge detection in the CMB. Using these techniques they have found four candidates for primordial collisions. Check it out here: arxiv.org/abs/1012.1995

Another paper coauthored by the renowned Mathematical Physicist Sir Roger Penrose takes a different starting point for its analysis. Penrose is a proponent of Cyclic Inflation rather than Eternal Inflation. Cyclic Inflation starts from the question of why the beginning of the universe had such low entropy and postulates that at the end of the universe there are only black holes and that they evaporate, somehow removing the entropy from the universe and leaving it in an extremely low entropy state from which another big bang will start the whole rising entropy cycle anew.

He and his collaborator see evidence of concentric circles in the CBM which they imagine may have come about from the merger of ultra massive black holes that existed before the big bang. Check in out here: arxiv.org/abs/1011.3706

The first wave of anomalies (purportedly) seen in the CMB came from analyzing the data with little pre supposition about what the anomalies would look like. We now have at least two examples which start from an existing theory and try to see if there is evidence to support the theory. Both methods are valid but they have to be truly supported by the data and only time and those blessed second guessing trolls who bash through the data looking for mistakes will sort that out. And, of course those trolls will have the Planck data soon. In the meantime we have something to pique our imagination.