More about CANDELS at the AAS

As mentioned by Liz McGrath in her recent blog post, the American Astronomical Society held its 223rd meeting  in early January. Altogether there were 37 presentations that had the word CANDELS somewhere in their title, author list, or abstract, and probably several others that used CANDELS under the radar. I thought it worth sharing a few highlights.

The meeting opened with the Kavli lecture by Bob Williams, the former director of the Space Telescope Science institute and president of the International Astronomical Union, who recounted the history of the original Hubble Deep Field and the science legacy that flowed from that first observation. One of the things he emphasized is how the making the data completely non-proprietary helped build the scientific momentum. Instead of observing separate spots in the sky and keeping the data to themselves, astronomers interested in the distant universe were suddenly much more willing to point telescopes to the same spot and share the data. CANDELS carries forward the legacy of this culture change; all of the Hubble data and most of the data from other telescopes are available to everyone.

A figure from Steve Finkelstein's talk, showing the number of galaxies found at redshifts
z=7 and 8 from the ultra-deep survey fields (blue) and CANDELS (red). Absolute magnitude
is a measure of brightness, with smaller numbers (more negative) indicating brighter galaxies.
CANDELS is great for finding brighter (rarer) galaxies, while the ultra-deep fields
excel at finding fainter (but more common) galaxies.

On Monday afternoon, the CANDELS and CLASH projects were featured in a special session. This was an opportunity for astronomers from both projects to give a top-level summary of the results. The CANDELS talks were given by Jeyhan Kartaltepe, Steve Finkelstein, Yu Lu and Steve Rodney. Jeyhan summarized the results on galaxy morphology, with particular attention to the CANDELS visual classifications and the evidence that ultraluminous infrared galaxies are associated with galaxy collisions and mergers, even at high redshift. Steve Finkelstein summarized the state of research on the most distant galaxies. He showed a nice diagram that helps to put into perspective the contributions of the "deep but narrow" fields (the Hubble Ultradeep field and associated parallel observations), and CANDELS. He also talked his discovery of what is currently the record-holder for the most distant spectroscopically confirmed galaxy, at redshift z=7.51. Yu Lu summarized the effort of part of the CANDELS theory group to compare semi-analytical models of galaxy formation. This is the most detailed side-by-side test of such models yet carried out and will be of great utility in helping us understand what we are learning from CANDELS. Finally, Steve Rodney partnered with Or Graur of the CLASH team to give a summary of the supernova program, including the first estimates of the evolution of supernova rates from both surveys.

A figure from Guillermo Barro's talk about the evolution of compact galaxies at high redshift.
The vertical axis shows how active the galaxy is in forming stars. Specifically, it shows
the ratio of each galaxy's star-formation rate, to its total stellar mass. The horizontal axis
shows how densely packed the stars are together. Dense galaxies are to the right,
diffuse galaxies are to the left. Guillermo is finding that galaxies tend to make the
transformation from being diffuse to being compact while they are still forming stars;
the quenching of star formation happens later.

The next day, CANDELS figured prominently in the Hieneman prize lecture given by Rachel Somerville. Rachel has been one of the leaders of the theory effort in CANDELS and recounted the progress in trying use the observations to get at the detailed physics of galaxy formation. One of the more recent results she showed in her talk was the success that she and Lauren Porter have had in matching the trends that Guillermo Barro and Christina Williams see in the evolution of compact galaxies. Both Guillermo and Christina spoke that morning in a session about galaxy surveys that was almost entirely populated by CANDELS talks.

In addition to the talks, there were quite a few posters. Posters are tacked to display boards in a big exhibition hall, and stay up all day. On the plus side, this gives people an opportunity to wander by and discuss the research, which is harder to do in a session packed with 5-minute talks. On the minus side, there are so many posters that it impossible to look at them all. CANDELS was very fortunate to have most of our posters on the first day located right near the entrance to the hall, so there was lots of opportunity for people to see them.

It was also great to see work by other groups using the CANDELS data. One of the press releases from the meeting was from Garth Illingworth, discussing finding some suprisingly bright galaxies in the CANDELS GOODS-N data that might be at redshift z>9. These are not yet spectroscopically confirmed, but if they are really at high redshift, then they indicate that star-formation was already proceeding vigorously when the universe was only 500 million years old.

An Update on Mature Galaxies in the Early Universe

One of the goals of CANDELS is to document how galaxy shapes evolved over about 13 billion years of history.  Avid readers of this blog might recall our posting about the presence of mature galaxies 11 billion years ago. In that post, we explained that a beautiful montage of galaxies that accompanied a recent European Space Agency press release gave a misleading impression of the scientific finding.

We're delighted to say that the ESA press officer understood our concerns immediately and offered to produce a version with images of galaxies that are from the CANDELS data themselves. That image is shown below. The leftmost panel is identical with the original press release, but the center panel now uses images of galaxies at redshifts between z=0.3 and 0.7 and the rightmost panel uses galaxies between z=2 and 2.7.

An updated image of the evolution of the Hubble sequence over the past 11 billion years. From ESA.

At each epoch, the illustration is intended to show the Hubble tuning fork. The four leftmost postage-stamps show early-type galaxies, known as ellipticals and lenticulars (or S0 galaxies). The tuning fork then splits to show the late-type spiral galaxies. The top four images show normal spiral galaxies while the bottom four panels show barred spirals (galaxies with a bar-like feature in the middle).

While this is still not a perfect representation of what is happening to galaxy shapes, there are several things to notice.

• The galaxies 4 billion years ago and 11 billion years ago were smaller. Alas, this image still doesn't have the relative scales exactly correct between the epochs, but the sense of the evolution from smaller to larger is consistent with detailed measurements.
• The early-type galaxies 11 billion years ago look sort of similar to those today, albeit smaller. They had about the same round, ball-like shape with nearly uniform color. They tend to be redder than the late-type galaxies.
• The late-type galaxies 4 billion years ago look pretty similar to today. They have spiral features and you can find examples that have bars in their centers.
• The late-type galaxies 11 billion years ago look a bit different. It is very hard to find any convincing examples of barred spirals, and the ones that don't have bars look more disordered than their present-day counterparts.
• The greenish tint of the galaxies in the center panel isn't real. These images are composites through different filters, and the color balance has not been tweaked in a meaningful way. When measured in detail, galaxies in the past tended to be bluer than today, which is a result of the fact that they contained more hot, young stars and less dust.

While the illustration here gives a better illustration of how galaxies are transforming their shapes, it doesn't illustrate some of the more interesting features like bright clumps or mergers. Also it doesn't tell you anything about the relative numbers of galaxies of different shapes. It turns out that the early-type galaxies were quite rare 11 billion years ago, while they are much more common today, for example. 

Finally, in spite of our quibbles about the artwork, we would like to give a shout-out to the Hubble outreach team at ESA for phenomenal work over the years. Most recently, they were recognized for their efforts with a Parsec award "Fact behind the Fiction" award for their brilliant Hubblecast series. 

Were Galaxies Really Mature 11 Billion Years Ago?

Nostra Culpa

Yesterday, the European Space Agency put out a press release on CANDELS research that has been widely covered by various media. Sometimes the scientific message gets garbled in press releases, in spite of best intentions all around. In this case, the wording of the press release is pretty good about conveying the key message (although some of the subtleties get lost). But the gorgeous picture that accompanies the release is more wrong than right. Alas, there was a bit more "artistic license" taken with the press-release artwork than there should have been, and we failed to put the brakes on before the press release went out. The pictures of the galaxies marked as "11 billion years" were not taken from the CANDELS images, and are not at the right redshift (redshift z=2.5 corresponds to a lookback time of 11 billion years). If you would like to see what the images really look like, read on.

This is the gorgeous image that accompanied the ESA CANDELS press release. If only it were closer to the truth! The images further down in this blog post show what galaxies 11 billion years ago really look like in the CANDELS images.

What did galaxies look like 11 billion years ago?

Cutouts from Lee et al.
Galaxies on the right are
forming stars rapidly.
Galaxies on the left are
not. 

The basic conclusion of the papers on which this story was based is that we can start to see the dichotomy between star-forming galaxies being "disk like" and non-star-forming galaxies being "spheroid like" already being set into place 11 billion years ago. That's the central message and is what we are finding in the CANDELS survey in the various samples studied by Tao Wang, Eric Bell, BoMee Lee, Alice Mortlock, Victoria Bruce and others.

We wish we could see galaxies 11 billion years ago with the sort of clarity shown on the press-release image, but unfortunately even with Hubble we can't see that level of detail. To the extent that we can distinguish detail they look (a) smaller and (b) generally bluer, and (c) less well-ordered than present-day galaxies.

So -- yes we can see spheroids and disks emerging as separate sequences -- but it is not the case that the galaxies look like their present-day counterparts, which is the impression that you get from looking at the press-release image.  Arjen van der Wel's quote in the press release that they look "remarkably mature" is reasonable when you consider that they might have all looked like train-wreck mergers-in-process.  But some of the news stories based on the press release are taking that to mean they are completely mature, which is certainly not the case.

Illustration of the morphologies of massive galaxies at redshift z~2 using CANDELS cutouts from Wang et al. (2012). The upper-left insert color codes galaxies on the same axis scale, with blue being the star-forming galaxies and red being the less-star-forming galaxies. You can easily see that the less-star-forming ones tend to look more spheroidal.

The light reaching us from galaxies at a redshift z=2.5 took about 11 billion years to get here. So we are looking back in time to when the universe was only about 2.7 billion years old. Below and to the right are some figures from some recent papers by CANDELS and other teams that show images of galaxies at about this redshift. Perhaps this gives a more accurate impression of what we can say and can't say about the comparison between nearby galaxies and their distant progenitors.

A montage of "Milky-Way progenitor" candidates from the CANDELS images put together by Pieter van Dokkum and the 3D-HST team. 

The End is Just the Beginning

Hubble's last observation for CANDELS is scheduled for August 10, the end of next week. Not counting the images taken next week, Hubble has taken three thousand four hundred and fifty eight pictures for CANDELS over the past three years. So now what?

The upcoming schedule for the last few CANDELS images.

Turning photons into numbers

Hubble's cameras have detectors that capture the minute amounts of light from distant stars and galaxies and convert that light into a tiny electrical signal. That electrical signal is recorded digitally as a set of ones and zeros. Hubble beams this data back to earth with a radio antenna every few hours. It will typically be several hours to a day before we have the data on our computers.

Turning the numbers into images

Even with fancy computers and software, it takes us about a month to take the raw images that are radioed back to earth from Hubble, and carefully assemble them into science-quality mosaics. This process involves identifying and masking artifacts left by charged particles in the solar wind, or trails of passing satelites or space junk, and subtracting off any scattered light that reflected into the camera from the earth's surface. We typically take several raw images of each patch of the sky, shifting the telescope a tiny bit between images. So we can almost always tell what is a real star or galaxy, and what is an artifact. We've been keeping up pretty well with the influx of data. If you are patient, you can download our processed images from the Mikulski Archive web site.

Turning the images back into numbers

In the end, we expect to detect about 250,000 galaxies in the CANDELS images. To use the images to try to make progress in understanding galaxy evolution, we need to measure the properties of the galaxies: their brightness, their colors, their sizes, and their shapes. We need to use those measurements to try to infer something more fundamental about the galaxies: their star-formation rates, their stellar masses, their ages, the amount of interstellar dust that they contain, and whether or not they harbor a central black hole.

In other words, we need to turn the images back into numbers.

We have a lot of specialized software to help us do this. To detect the galaxies, we use a customized version of a computer program called SExtractor, which identifies faint smudges of light and tries to make a semi-intelligent decision about whether two adjacent smudges are part of a single galaxy or two separate galaxies. That is, it segments the images into different regions, assigning most of the pixels to "background sky" and a few percent of them to separately-detected stars or galaxies.  It took quite a bit of work to get to the point where we were reasonably satisfied with how it was doing this. Nonetheless, there are still a few percent of the sources that have either been poorly "deblended" by the software, or are not real (the most common offender being scattered light from a bright star). At this stage, we can flag most of the artifacts. We're stuck with SExtractor's image segmentation for now, which forces us to worry about the issue for almost every CANDELS paper. If anyone would like a nice image-processing challenge, improving the segmentation step -- using all of the available color information from the multiple images -- would be a big step forward.

The image segmentation step. The picture to the left shows a small slice of the CANDELS image, with two bright galaxies and a lot of fainter ones. The image to the right illustrates how SExtractor segments the image into different galaxies, which sometimes overlap. Human judgment doesn't always agree with what this software does. Sometimes it becomes more obvious how to break up the objects if you look at full color images instead of black & white images. SExtractor only works in black & white.

SExtractor not only detects the galaxies, it measures their sizes, shapes and brightnesses. It does this very quickly and reasonably precisely. We have done thousands of experiments inserting artificial galaxies into the images to quantify the accuracy and precision of these measurements. These experiments allow us to determine some statistical corrections to SExtractor's measurements, so that we can infer more accurate brightnesses of the galaxies (for example).

Having measured the Hubble images, we next need to make use of data from other observatories. Primarily these are large ground-based telescopes like the VLT in Chile and the Spitzer infrared telescope. These images are not nearly as sharp as the Hubble images, so unfortunately many of the galaxies are blended together. However, the Hubble images tell us where almost all of the galaxies are.  So we can cut out the individual sources from the Hubble images, blur them to match the image quality of the other telescope, and then ask the computer to tell us what combination of brightnesses of the blended sources best reproduces the blended image. We use a program called TFIT to do this. As was the case for SExtractor, we have done extensive tests with artificial galaxies to convince ourselves that TFIT is doing the measurements correctly. By and large we are satisfied that it is doing an excellent job, but we have a relatively long wish-list of things that we would like to improve, both to remove systematic biases at the few percent level in the brightnesses of all the galaxies, and to deal with the very rare "problem cases" where the answers don't make sense.

This animation starts with an X-ray image from the Chandra telescope, 
showing two sources which probably harbor central black holes. It then
transitions to the Hubble images, where you can see that some of the
galaxies have spiral features and others don't. It then transitions to the
infrared images from the Spitzer, and finally the Herschel observatory.
You can see that the resolution of the infrared images is much poorer
than the Hubble images -- the galaxies are all blended together.
Nevertheless, you can make out that some galaxies are "bluer" than
others at infrared wavlengths. Looking at the last image, you can convince
yourself that the brightest source of far-infrared radiation in this image
(which comes primarily from heated dust) is probably the galaxy just
below the center, which is also an X-ray sources.

The output of this is a "Multi-wavelength Photometry Catalog," which is one of the most useful products for scientific study. We are striving to make this as reliable as possible because much of the science depends on that.

The multi-wavelength catalogs can be used to estimate "photometric redshifts," which tell us roughly how far away each galaxy is. The brightnesses and colors of each galaxy can be used to infer the stellar mass of each galaxy, the star-formation rates, dust content, and ages of each galaxy.  Our simulations tell us that the stellar mass estimates are pretty accurate (generally to within about a factor of two of the truth), but the estimates for the other quantities are pretty shaky. A lot of work is going into trying to quantify the biases and uncertainties, and perhaps find some way to improve the estimates.

The other type of measurement is galaxy shapes and sizes. We are going at this in a whole variety of ways. We use a computer program called GALFIT to fit a smooth light profile to each galaxy and extract some basic numbers that characterize the size and shape of the galaxy and gives us a measurement of how centrally concentrated the light is in each galaxy.  We have some other programs that make estimates of concentration without assuming a smooth light profile, and that measure the asymmetry of the images. Teams of people both in CANDELS and in Galaxy Zoo are inspecting the images and classifying the the galaxies into various categories.  We also have some software that tries to identify separate clumps of light within each galaxy. We need to calibrate all those measurements by inserting artifical galaxies into the images. This has been done for GALFIT, but it is just getting started for the other measurements.

Finally, there are measurements of correlations between galaxies. Galaxies have a propensity to cluster together. We can measure this quantitatively, but we have to carefully account for the fact that the images don't all have the same exposure time. In some patches of sky we can detect fainter galaxies than in other patches of sky. Once again, we resort to inserting artificial galaxies into the images and recovering them to quantify how our detection limits vary across the full survey. The clustering estimates make use of the photometric redshifts, so we also need to do lots of simulations to understand the effect of photometric redshift errors on the clustering measurements.

So you have a bunch of numbers. Now what?

This is where the fun really starts. This is where we can try to give rigorous answers to the scientific questions posed in our original proposal.  We would like to know, for example, how galaxies build up their masses over time. To address this, we can collect the galaxies together in different redshift slices and we can estimate how their stellar masses change as the universe gets older. We can compare these evolving stellar masses to our estimates of the star formation rates to see whether they are consistent. If they aren't, then we must be doing something wrong. Either we are missing some galaxies (for example, because they are obscured by dust), or we are not measuring the mass or the star-formation rates correctly. This kind of consistency check is essential. It's how we gain confidence that we really understand what we are seeing.

We can ask whether galaxies that have active nuclei -- black holes that are surrounded by hot gas that emits x-rays -- look any different from galaxies with the same stellar mass that don't emit X-rays. So far,  somewhat suprisingly, the answer is that they look the same. Now that we have all of the data in hand, we can look more closely with better statistics. It's possible that some classes of galaxies with active nuclei (perhaps the brightest ones) look different, for example.

We can look at how galaxies grow in size. This has been a long-standing puzzle. We know that galaxies were smaller in the past. For galaxies that are continually forming stars, this is not really a problem -- the later stars presumably form in the outskirts, making the galaxies bigger over time. However, galaxies that have stopped forming stars (become quiescent) already when the universe was only 3-5 billion years old are much smaller than quiescent galaxies today. The suspicion is that such galaxies merge together, becoming bigger without forming many more stars. But that suspicion has been hard to verify because we simply haven't had surveys that are large enough to accurately measure the number of galaxies that are in pairs or in the process of merging.

There are lots and lots of other questions, outlined in our original science goals for the survey. Sometimes the work involves comparing directly to theoretical predictions -- for example there are some beautiful predictions for the evolution of star-forming clumps that we can begin to test by dissecting the images. Sometimes there are unexpected discoveries.

So even though the last images are going to be here next week, we still have about two full years of work to make all the measurements, produce the catalogs, and really dig in and try to make sense of what we are seeing.

In communicating our findings to fellow astronomers and to the general public, we often strive to find a very clear diagram or figure to make the statistical evidence readily apparent. In fact, these diagrams are often the route to fame for an astronomer. Every beginning astronomy student learns about the Hubble diagram or the Hertsprung-Russell diagram. So maybe in the end, having converted the photons into numbers, the numbers into images, and the images back into numbers, we need to very cleverly turn these numbers back into images. Yes it seems silly, but that's often where we get to exercise our scientific creativity, and it's also often how we gain the most insight into the workings of the universe.

Exciting News from the Planck Satellite

We live in a weird universe. It had a beginning. It is enormous, probably infinite, in extent. And yet at least our observable part of it started out in a volume smaller than the head of a pin 13 to 14 billion years ago. In the vast unobservable part of the universe, there may well be an exact copy of you reading this blog. The universe is getting bigger, but the expansion rate started out fast, slowed down, and is now speeding up again. Is that weird enough for you? No? How about all that and lopsided? Read on.

All this would sound crazy if we didn't have a lot of evidence that tells us that this is the universe we inhabit. That evidence includes:

• The observation that almost all galaxies in the sky are moving away from us, and that the ones that are further away are moving away from us faster. 
• The fact that we cannot find stars or galaxies that look like they are much older than about 14 billion years old (plus or minus a few billion). 
• The observation that relatively pristine gas in the universe consists of about 73% hydrogen and 25% helium. This ratio was set in place during the few minutes during cosmic expansion when the universe was cool enough for protons and neutrons to exist, but hot enough and dense enough for them to be continually bumping into each other.

And the evidence includes measurements of the Cosmic Microwave Background (CMB) radiation - radio waves that are all around us. The CMB accounts for a very small fraction of the static you can hear on an FM radio if you tune between channels. These radio waves originated when the universe was only 380,000 years old and was still permeated by a dense fog of electrons that were bumping into each other too often to be bound to atoms.

The big news this week is the release of the first cosmology results from the Planck satellite, which is the latest in series of increasingly sophisticated observatories designed to measure the statistical properties of this radio noise from the early universe. I'm going to comment on a few of interesting results in the 29 papers that were just submitted for publication. I am not an expert on the measurements and I have only just begun reading the papers. I'm aiming here for a set of highlights that are a bit more detailed than the press release, but easier to digest than the papers. There are some nice blog posts at Bad Astronomy and Starts with a Bang as well. If you are an aficionado, feel free to comment if you find that I've botched something, or missed something important!

The Data

The map of the Cosmic Microwave Background from the Planck satellite, 

after removing the foreground sources of radio emission. 

From the European Space Agency.

The Planck satellite is a radio telescope that scans the sky repeatedly. By comparing the radio power it receives when pointing in different directions, it can build up a map of the temperature fluctuations in the gas that permeated the universe when it was only 380,000 years old. These temperature fluctuations are tiny -- less than one part in 10,000. The engineering feat of creating devices to measure this is quite amazing. There are also other sources of radio emission in the sky, and it has taken a lot very careful work to measure these sources and account for their effect on the Planck measurements. Once these "foreground" sources are removed, the result is a map of the surface of last scattering. The places where you see red are just a tiny bit hotter than the places where you see blue. These hotter and cooler spots agree remarkably well with the previous best measurements from the Wilkinson Microwave Anisotropy Probe (WMAP) satellite.

A blowup of a small portion of the CMB map. The new map on the right is from Planck. The previous best map from the WMAP satellite is on the left. The Planck measurements agree extremely well, but obviously provide more detail. From the European Space Agency.

You can see by eye that these maps are not uniform. There are fluctuations from pixel to pixel. But these aren't completely random. There are somewhat coherent patches that are bluer and other patches that are redder. The tool that astronomers use to quantify this is the angular power spectrum. The basic concept of measuring the angular power is something like the following. Draw a little circle on the map. Measure the average temperature in that circle. Do that in a bunch of other little circles that are the same size and compare the results. If they are all the same, write down the number zero. If they aren't,  you can quantify the dispersion of the measurements using the standard deviation or some other statistical quantity. Now vary the size of these circles. You will find that the dispersion in these mean temperatures depends on the size of the circle. Plot up these statistical results as a function of the angular separation and you have an angular power spectrum, more or less. (In practice, the fluctuations are measured using Fourier analysis and spherical harmonics). The figure below shows the angular power spectrum from Planck.

The angular power spectrum of temperature fluctuations measured by Planck. Small angular separations are on the right, large angles are on the left. The red points show the Planck measurements. The green curve is the prediction from the standard model of cosmology -- that weird model described at the opening of this blog post. (If you want to understand what causes the wiggles, I highly recommend reading  Wayne Hu's web site and looking at some of his animations.) A couple of amazing things to note. (1) Every point has an error bar. Most of them are tiny. This is  a great measurement. (2) The predictions from the cosmological model go right through the data. The shaded green region at large angular scales shows the uncertainty due to the fact that we can only observe from one vantage point whereas the models predict the fluctuations that would be observed for an ensemble vantage points placed randomly throughout the universe. From the European Space Agency.

What Does It Mean?

The standard (weird) model of cosmology is expressed as a series of mathematical equations, which include physical constants that we know quite well (for example the speed of light), and include other quantities, like the density of matter in the universe, that we don't know as well. These quantities are known as cosmological parameters. There are seventeen parameters involved in the model represented by the green curve in the angular power spectrum plot. Having adjusted the parameters to fit the data, it's then possible to infer other quantities -- like the current rate of expansion of the universe, its age, and the mean density of matter and energy. 

So here are a few highlights of the Planck findings.

The universe is expanding more slowly than we thought.

The present-day expansion rate of the universe is known as the Hubble Constant. The Planck estimate is only about 90% of the expansion rate that we measured a couple years ago with the Hubble Space Telescope. This is a bit surprising given the estimates of the uncertainties from both measurements. We usually learn something when two sets of precision measurements don't quite agree. So it's going to be interesting to try to understand what the source of the discrepancy is. 

The universe has only three flavors of neutrinos.

One of the interesting inferences from the previous, higher, expansion rate was that the model that provided the best fit to all the data required more than three types of neutrinos. Neutrinos are very light subatomic particles that carry no charge and rarely interact with atoms. There about 100 million neutrinos passing through you right this second. The standard model of particle physics has three types of neutrinos. The previous cosmological observations suggested there were four. Because this would have a profound impact of our understanding of the forces of nature, the measurements created a lot of interest in a hypothetical particle called a sterile neutrino, which interacts with normal matter only through gravity. If the sterile neutrino had mass, it could be part of the dark matter that pervades the universe. It would be warm dark matter, and its existence might help resolve some of the problems that we encounter when trying to explain galaxy properties using just cold dark matter. If Planck is right, we will need to look elsewhere for solutions.

The universe is flat.

It's that time of year, so imagine yourself on the surface of a giant basketball, but one that is poorly inflated. If you start walking one direction, you will eventually end up where you started. If you count your steps, you can estimate the size of the basketball. Now imagine someone inflates the basketball. It's still more or less spherical, but it will take more steps to arrive back at your starting point. The number of steps gives you the size. The fact that you ended up where you started, tells you that the basketball is a closed surface. Another thing that tells you this is that two parallel lines will eventually intersect -- like those two lines that meet in the middle of the picture. 

The universe is harder to picture because the "surface" has three dimensions, not just two. If the universe is closed, a beam of light sent out in one direction will eventually (billions of year later) return to you. In the CMB measurements, changing the curvature of the universe changes the position of the peaks in angle on the sky. The Planck observations confirm what previous measurements were telling us, that the universe is spatially flat. Two lines that start out parallel won't intersect, and they won't drift apart. Flatness is one of the consequences of the early, rapid expansion of the universe called inflation. So it is good to see this holding up. 

The universe has a bit less Dark Energy and a bit more Dark Matter than we thought.

We don't know what Dark Matter and Dark Energy are. Dark Matter is probably some kind of subatomic particle that interacts with normal matter primarily through gravity. We think it is there because we can measure its gravitational influence on large scales in the universe. There are many, many observations that are best explained by invoking dark matter. But we still don't know what it is. Dark Energy is even stranger. It has been invoked to explain the observation that the universe has started to expand faster. It is probably has some link to Vacuum Energy in particle physics, but there is no consensus among physicists on how it is related. The best fit to the wiggles in the Planck power spectrum suggests that the universe consists of 68% Dark Energy, 27% Dark Matter and 5% normal matter (what all of the stars and galaxies and people in the universe are made of). The previous best estimate was 72% Dark Energy.

Dusty galaxies can be detected by their mass.

This result is interesting to those of us who study distant galaxies. There is a fluctuating background at far-infrared wavelengths that has nothing to do with the CMB. It probably arises from dusty galaxies, but we don't know much about them, other than that they are very numerous and probably very distant. The dust absorbs the optical and ultraviolet radiation emitted from stars and re-emits it in the infrared. Planck can measure this fluctuating background in its short wavelength channels, and by correlating with the radio maps, can see the statistical effect of gravitational lensing of the CMB by the mass associated with the dusty galaxies. The average dark-matter mass inferred from the measurement is only about 3% of the mass of the Milky Way. I find this a bit surprising because we tend to find that massive galaxies are very dusty and low-mass galaxies are not. On the other hand, the uncertainties of the estimate are quite large. The constraints might get tighter with more analysis.

This is a cool measurement. In addition to making radio maps, Planck can measure far-infrared radiation that comes from dusty galaxies. The maps in the right panel show the fluctuations in the infrared brightness of the sky at wavelengths of 350 and 545 microns due to these dusty galaxies. You can't see the individual galaxies, just their combined effect. The mass of each of these individual galaxies can imprint a gravitational lensing effect on the CMB. It's possible to detect that effect (the diagrams on the left and center) by stacking the CMB maps at the positions of the peaks (left) and troughs (center) of the far-infrared map. By measuring the effect, it's possible to infer the mass and distance of the galaxies responsible for the far-infrared radiation. Paper XVIII.

The Universe is lopsided?

This is the result that might engender the most discussion, because it suggests the universe might be even weirder than we thought. A fundamental assumption of cosmology is that this vast universe looks statistically the same in all directions. There is no preferred direction. There was a hint in the earlier data from the WMAP satellite that there might be a preferred direction. That hint has gotten much stronger with the Planck data, which confirms both the effect and the direction. One interpretation is that our entire galaxy is moving in one direction faster than we think it should if we add up the pull of gravity from all the other galaxies in that direction. But that interpretation doesn't fit all the data, and there are weirder possibilities that involve changing some of our assumptions about the cosmological model.

And one other tidbit that's going to cause problems.

The Planck papers have a lot of co-authors. The the authors are listed alphabetically after the words "Planck Collaboration." The first author is almost always Peter Ade of Cardiff University. It's rare (but not unprecedented) in astronomy to construct the author lists this way, because the tradition has been to recognize the people who have contributed most to a particular paper (even if it is from a giant collaboration like CANDELS) by listing their names near the top of the author list. By bucking tradition, the Planck collaboration is acknowledging  that it often impossible to judge who made the most significant contributions to a particular paper, particularly if it depends on the technical efforts of many, many other people. On the other hand, it is going to be interesting to see who gets invited to the conferences!

Star Formation in the Mountains

A view of the mountains surrounding Sesto, Italy. Photo by Dale Kocevski.

No two snowflakes are alike, and yet forecasters are pretty good at predicting snow. No two mountains are alike, and yet geologists can tell us quite a lot about how mountain ranges form and erode. Similarly, no two galaxies are alike and yet astronomers would like to understand how galaxies as a whole form and evolve. So what better place to talk about this topic than in a snowy mountain range! Last week a group of about 40 astronomers met in the small town of Sesto, Italy, nestled in the Dolomite mountains right near the Austrian border. The title of the workshop was "Star Formation Through Cosmic Time," and the focus was on trying to link together what we are learning about star-formation in very distant galaxies from Hubble observations like CANDELS to observations at infrared wavelengths from the Spitzer and Herschel observatories. This is important because more than half of the energy produced by stars in distant galaxies is absorbed by dust and re-emitted as infrared radiation. 

Most galaxies seem to form stars at a rate that is proportional 

to the number of stars that they already have. Some astronomers 

are calling this the main sequence of star-forming galaxies. 

Other  galaxies fall off the sequence. The red and dead ones

or quenched or quiescent ones aren't forming many stars at all. 

On the other hand there are some galaxies forming stars at much 

higher rates, which we call starbursts. Then there are a few galaxies 

that are still forming stars, but at lower rates than on the

main sequence. These populate the green valley,  although shutting 

down star formation isn't the only way to end up with greenish colors, 

so the green valley is sort of a hodgepodge of various

kinds of galaxies. 

A lot of the discussion at the meeting centered on the "Main Sequence of Star-Forming Galaxies" and on the galaxies that depart from that sequence. The "main sequence" is a term that was coined by CANDELS team-member Kai Noeske a few years ago and seems to have caught on. He noticed that most galaxies that are forming stars are forming them at a rate that is roughly proportional to their existing stellar mass. We don't understand in detail why this should be the case, so one item on the agenda was to discuss the evolution of this main sequence and the link between galaxies on the main sequence and galaxies that fall off it. The galaxies that fall off it fall into two classes: those that are forming stars at much higher rates ("starburst galaxies"), and those that have more-or-less stopped forming stars. Several people at the meeting talked about the starburst sequence. Depending a bit on how you define it, it looks like starbursts account for about 10-15% of all the cosmic star formation. I'm not sure the evidence that there are two separate sequences is all that compelling, but it is impressive that by assuming there are two sequences, it is possible to explain the evolution of the infrared luminosity function of galaxies, and to infer something about the evolution of the gas and the evolution of the heavy elements in galaxies. This is very handy because it can help inform us what to expect (and what to go look for) with two powerful radio telescopes that are just coming online, the JVLA and ALMA.

There were several talks about the ability of theoretical models to explain these two sequences. Currently, they seem to get the qualitative behavior right (there is a main sequence), but the quantitative behavior wrong (e.g. the proportion of stars forming in starbursts was about a factor of two too low in one of the models discussed). The failures of the model are almost certainly connected to the feedback of energy into the gas that is too cool to form the stars. This feedback can come from the stars themselves, particularly when they explode as supernovae, or from gas funneling into central massive black holes in the centers of galaxies. Supermassive black holes probably go through periods when they are not accreting a lot of gas, and other periods when they are. When they are actively accreting, they are called Active Galactic Nuclei (AGN) and emit a lot of high-energy radiation such as X-rays. However, if they are surrounded by dust, those X-rays can be absorbed and re-emitted as infrared radiation. There were discussions about new ways to identify AGN using infrared radiation as well as discussion about the properties of the host galaxies surrounding the AGN. CANDELS observations have revealed that distant AGN hosts don't really look any different than galaxies that are not hosting AGN, so that probably means that whatever is causing the gas to funnel into the black hole is not affecting the overall shape of the galaxy. That's a bit of a problem because it seemed quite likely that mergers between galaxies were a key way of getting the gas into the center.

Another interesting question is whether AGN prefer to be in star-forming galaxies or in galaxies that are shutting down their star formation. If feedback from AGN is important for quenching star formation, than one might expect that the galaxies that host AGN might look like they are starting to shut down. You might expect the "green valley" of galaxies in the diagram above to be populated by galaxies with AGN in their centers. The jury is out on this. Dale Kocevski showed evidence that the AGN hosts in CANDELS span the full range of star-forming activity that is seen in galaxies of the same mass. On the other hand another CANDELS member, David Rosario, showed evidence from far-infrared data that AGN hosts are drawn from a population of normal actively star-forming galaxies, and tend to avoid weakly star-forming, quenched or quiescent galaxies. So the observations are giving us somewhat contradictory information, and it is going to take some work to see how to reconcile these results.

One of the things that CANDELS provides is a good way to find and study quenched or quiescent galaxies at great distances. Several talks focused on the numbers of these galaxies. We are now finding massive quiescent galaxies when the universe was just a few billion years old. These galaxies are much more compact than massive non-star-forming galaxies today, so a couple of questions arise: (1) can they grow into their high-mass cousins by just acquiring stars in their outskirts by merging with surrounding galaxies and (2) can we find galaxies on the star-forming sequence that have enough stars jammed into their centers to be the likely progenitors of the quenched galaxies. The answers to these questions are tentatively yes: the densities of stars in the centers of the very distant quenched galaxies are pretty comparable to the central densities today, so adding stars to the outskirts probably works. And there appear to be enough compact galaxies on the star-forming sequence to form the galaxies on the quenched sequence if the star-formation shuts down on a reasonable timescale. On the other hand, CANDELS observations are finding fewer quenched low-mass galaxies than theory predicts, so that may be a problem. 

Harry Ferguson talked about some of the difficulties of inferring the
star-formation histories of high-redshift galaxies. Photo by Dale Kocevski.

There was also a lot of discussion about the star-forming histories of galaxies. We can estimate the stellar masses of galaxies in a variety of ways, and lots of checking suggests that these measurements are pretty robust; for an individual galaxy the estimates based on existing data are probably within a factor of two of the true value. Estimating star formation rates is much more difficult, but if you have information from the far-infrared together with infrared from the ultraviolet part of the spectrum, then it is also possible to make pretty good estimates. So putting those together, astronomers can estimate the total number of stars forming per year, and can do this at various "lookback times" from the present day to about 12 billion years in the past. Astronomers can also estimate the amount of stellar mass present at each of these lookback times. Now the stellar mass at later times ought to agree with what we infer from the rate of star formation at earlier times. This has been a problem in the past, but it now looks like the problem has been resolved with better estimates of star formation rates and stellar masses. So that's good news. On the other hand, the very fact that these estimates agree means that there can't be a lot of galaxies missing from the census of either star-forming or non-star-forming galaxies. That's a bit weird because galaxies can disappear from the census pretty easily if they become very dusty, or fade enough between bursts of star formation. Some theoretical models predict a lot of bursting and a lot of dusty galaxies, so these models might need to be revisited. 

We can also look in detail at the measurements of galaxy colors and spectra and try to infer a bit more about their individual histories of star formation. A lot of discussion at the meeting was about the difficulties involved in doing this. Unfortunately, the current state-of-the art is that when you use all of the information provided in the spectrum to try to estimate the star formation rate, you probably get a worse estimate than if you ignore the optical and near-infrared portion of the spectrum and just use the information from the ultraviolet or the far-infrared (or better yet, both). This is probably because we don't have the correct star formation histories in our models, but we need to find a way to introduce more realistic star-forming histories without "over fitting" the data. 

The useful thing about small workshops is that people are more willing to admit what they don't understand. That tends to make for very fruitful discussion and provides the fodder for new projects. On that score, the meeting was very successful.