Header 1

Our future, our universe, and other weighty topics

Saturday, August 5, 2017

A Dubious Claim to Have a Map of a Never-Observed

As I described in this post, there is a pattern I observe repeatedly in science reporting:
  1. A scientific paper will be released making modest claims that may not be particularly interesting.
  2. That paper will be hyped and exaggerated by a press release published by some institution related to the scientific paper, perhaps a particular university or scientific group.
  3. Further hype and exaggeration will be done by the popular press, which is always eager to sensationalize the scientific, because of the Internet profit that results from an increased number of users clicking on a click-bait link.
By the time the average person reads the story prompted by the scientific paper, they will be given some idea that may not at all be justified by the original paper.

I saw such a thing going on in a recent announcement of a new map called a “dark matter map.” When I saw the story announcing this on bbc.com, I was quite surprised. The bbc.com story announced, “Researchers have released the most accurate map ever produced of the dark matter in our Universe.” But how can someone have a map of dark matter locations when dark matter has never been observed? All attempts thus far to make direct observations of dark matter have failed. Dark matter doesn't even have a place in the Standard Model of Physics, and no evidence for it has turned up at the Large Hadron Collider. 

I began examining the sources of these claims. The BBC story took me to the the web site of something called the Dark Energy Survey. On that site was a press release issued by Fermilab, a major scientific organization. The press release was entitled, “Dark Energy Survey reveals most accurate measurement of dark matter structure in the universe.” That is pretty much the same as claiming to have a map of dark matter in the universe. The Fermilab press release claimed that scientists had “precisely measured the shapes of 26 million galaxies to directly map the patterns of dark matter over billions of light-years, using a technique called gravitational lensing.” It then gave a link back to the Dark Energy Survey page and some papers released on that page.

This sounded very fishy to me, because observations of gravitational lensing are not equivalent to observations of dark matter. Gravitational lensing is a strange effect produced on light rays bent by the gravity of high concentrations of matter. Such matter can be any type of matter: either normal matter or possibly some type of dark matter. As Scientific American puts it when describing gravitational lensing:

According to Albert Einstein’s general theory of relativity, mass warps space, so a large amount of matter in the foreground of a galaxy can bend its light in a way that makes it look slightly squashed. This is true whether the foreground mass is made of invisible dark matter or ordinary matter.
So if you are claiming to have a map of dark matter made by observing gravitational lensing, you are doing something rather like announcing that you have a map of UFO landing sites made by observing small burnt patches in the forest. Such patches might be produced by hot UFO's that are landing, but they also might be produced by ordinary lightning flashes.

What I also thought was fishy was that the Fermilab press release gave us a visual which it tells us is a map of dark matter. But the map legend on that visual does not refer to dark matter, but is instead labeled “Density of matter.”  The map is below:

To further investigate whether both the BBC and Fermilab are guilty of exaggeration or hype or dubious interpretations, I tracked down the scientific paper that is the source of these claims. The paper is here. It is entitled “Dark Energy Survey Year 1 Results: Curved-sky Weak Lensing Mass Map.” This paper has the same visual that appears in the Fermilab story and the BBC story, so it seems to be the source of their stories.

I read the abstract of the paper. There was no mention made of dark matter. I tried searching for “dark matter” in the text. The only two mentions of dark matter were incidental mentions not claiming to have observed or mapped dark matter. The first mention was this:

Briefly, three flat LCDM dark-matter-only N-body simulations were used, with 10503, 26003 and 40003 Mpc3h−3 boxes and 14003, 20483 and 20483 particles, respectively.

Simulations? That's “make believe” stuff, not an observation of dark matter or a mapping of dark matter.

The second and last mention of dark matter in the scientific paper was the following turgid prose:

Galaxies are assigned to dark matter particles and given rband absolute magnitudes based on the distribution p(d|Mr) measured from a high resolution simulation populated with galaxies using subhalo abundance matching (SHAM) (Conroy, Wechsler
& Kravtsov 2006; Reddick et al. 2013), where d is a large scale density proxy.

This is not a statement making any claim about dark matter. The scientific paper makes no claim at all to have mapped or observed dark matter. It only claims to have observed gravitational lensing and made a map of “mass distribution,” which could be any type of matter, either regular matter or dark matter. The paper did have several visuals like the visual in the Fermilab press release (which calls the visual a map of dark matter). But in the scientific paper none of those visuals was described as a map of dark matter.

Rather than claiming to be making a map of dark matter, the paper claims to be creating a map of the “mass distribution of the universe,” a more general term referring to any type of matter, either regular or dark matter. The paper states in its first sentence: One way to map the mass distribution of the Universe is by using the technique of weak gravitational lensing.” Then in its conclusion the paper states the following:

Weak lensing allows us to probe the total mass distribution in the Universe. One of the most intuitive ways to visualize and comprehend this information is through weak lensing convergence maps,or mass maps....In this paper, we construct weak lensing mass maps for the first year of Dark Energy Survey data (DES Y1) using two independent shear catalogs. METACALIBRATION and IM3SHAPE, in the redshift range 0.2 < z < 1.3 and in the region overlapping with the South Pole Telescope footprint.

This is exactly how a paper would speak if it were presenting a map of the mass distribution of the universe (its total matter that is either regular matter or dark matter), not specifically a map of dark matter.

What we have going on here by bbc.com and Fermilab seems to be shameless hype and exaggeration, which includes the inaccurate claim that a map has been made of dark matter. The bbc.com and Fermilab press coverage refers us to the Dark Energy Survey. But when the relevant scientific paper is tracked down from the site of the Dark Energy Survey, we find that the paper did not actually make any substantive claim at all about dark matter, referring to it only in two passing references. The paper does not claim to have presented a map of dark matter, but merely claims to have made maps of mass distribution (a term that means the total amount of any type of matter, whether dark or regular). A Scientific American article describes the research correctly, saying it is a finding about "the distribution of matter," and making no claim that it involves  any type of map of dark matter. 

Described by those who believe in it as something "invisible," dark matter has never been observed, and you cannot make a map of something that has never been observed. The map these stories referred to is a map of mass distribution, not specifically dark matter. But, of course, if you describe such a map as a map of dark matter, that will result in more web traffic, because it sounds like something new and exciting (scientists have been making mass distribution maps for decades).

A lesson we may draw from this episode is perhaps that exaggeration and misinterpretation of scientific papers is not something done only by more sensationalistic sites like  DailyGalaxy.com; it is also something that can be done by the most mainstream and respected science-reporting sites such as the BBC site and the Fermilab site.