Amen to this.
When tools to produce refined-looking graphics are only accessible to, or usable by, professionals and/or when expert-to-public translation leads to inaccuracies, a fear of elegant-looking graphics, and a concomitant exploratory-explanatory divide is understandable. The divide often leaves behind the aforementioned idea that exceptionally fancy graphics, and the time invested to make them, are only for public consumption, and not really useful for serious scientists or others pursuing deep quantitative analysis.
As the “democratization of data” continues, more and more services are making data behind both scholarly journal figures and public outreach graphics freely accessible. These open data sets represent a wealth of new information that researchers can combine with more traditional data acquisitions in their inquiries. If it’s quick and easy to get the data behind explanatory graphics, scientists will use those data, and learn more.
To generalise on that, take a particularly obvious lesson from nudge theory : software needs to be easy to use if people are going to use it. Visualisation is innately fun, but installing software is not. Interface design really matters - I'm far more likely to experiment with something if the basics just involve pressing a button. If have to pause to write even ten lines of code, well, I'm not going to do that. Case in point : HI source extraction. If the only software available doesn't let you record detected galaxies very easily, you might go away thinking that human-based detection is very difficult. In reality it is not, it's simply the lack of a sensible recording interface that makes it tedious. Humans are great at this, but they get bored by having to write down long numbers. Visualisation software should make it as easy as possible for humans to do what they're good at and ease the burden of the less interesting tasks.
This problem is particularly acute in fields where everyone writes their own code. Also, on a related point, visualisation software should have a freakin' GUI. No, I don't want to have to type commands to generate a plot, that's just plain silly. Code should be used to manipulate data, and not - wherever possible - be used to visualise it. Major caveat : it should always be possible to access the underlying code for experimentation with non-standard, custom techniques. Modern versions of Blender make it very easy to access the appropriate commands to control each module, thus giving the best of both worlds.
Sometimes, even though tools permit easy explanatory-exploratory travel, sociology or culture prohibits it. By way of a very simple example, consider color. To a physicist portraying temperature, the color blue encodes “hot,” since bluer photons have higher energy, but in popular Western culture, blue is used to mean cold. So, a figure colored correctly for a physicist will not necessarily work for public outreach. Still, though, a physicist’s figure produced in an exploratory system like the one portrayed in Figure 2 would work fine as an explanatory graphic for other physicists reading a scholarly report on the new findings.
It might be interesting to have some app/website that lets people play with the raw data behind images to compile them themselves. After a while, you start to lose the bias against thinking that what you can see with your eyes is an especially privileged view of the Universe, e.g. http://www.rhysy.net/the-hydrogen-sky.html
In 2006, no one quite knew what a “data scientist” was, but today, those words describe one of the most in-demand, high-paying, professions of the 21st century. Data volume is rising faster and faster, as is the diversity of data sets available – both in the commercial and academic sectors. Despite the rise of data science, though, today’s students are typically not trained–at any level of their education–in data visualization. Even the best graduate students in science at Harvard typically arrive completely naive about what visualization researchers have learned about how humans perceive graphical displays of information.
Over the past decade or so, more and more PhD students in science fields are taking computer science and data science courses. These courses often focus almost entirely on purely statistical approaches to data analysis, and they foster the idea that machine learning and AI are all that is needed for insight. They do not foster the ideas that one of the 20th century's greatest statisticians, John Tukey, put forward about visualization: 1) having the potential to give unanticipated insight to later be followed up with quantitative, statistical, analysis; or 2) that algorithms can make errors easily discovered and understood with visualization.
Exactly. It's true that human pattern recognition is fallible. However, it's at least equally true that statistical analyses can be fallible too. Having an objective procedure is not at all the same as being objectively correct. Working in concert, visualisation and statistical measurements are more than the sum of their parts. Finding a pattern suggests new ways to measure data, which in turn forces you to consider what it is you're actually measuring.
https://arxiv.org/abs/1805.11300
Wednesday 30 May 2018
Thursday 24 May 2018
Hydrogen rendered as disco floor lights
A flight through VLA data of Milky Way hydrogen. This is an attempt to show a solid surface changing imperceptibly into a diffuse volume. There are 240 planes, each with a different slice of the data, with transparency controlled by colour. Each has a build modifer that randomly increases the rendered faces, with staggered start times increasing outwards from the two initial meshes. The jumping in colour at the end is unintentional, but I kinda like it.
Tuesday 22 May 2018
The ALFALFA sky in full colour 3D
Someone on YouTube asked for a side-by-side 3D version of the ALFALFA sky. So, here it is. You can view it with 3D TVs, headsets, projectors and stuff I guess. Or visually if you can see stereo pairs.
This is the same version as previously, using the ALFALFA 70% complete catalogue (22,000 galaxies). The 100% catalogue (31,000 galaxies) is now online , so that will be my next little project once I finish with the abstract stuff. Ideally that will be in 360 3D if I can figure out how to get around Blender's silly Cycles image texture limit (or maybe it's already fixed...).
https://www.youtube.com/watch?v=YkR8E6x_2mo
This is the same version as previously, using the ALFALFA 70% complete catalogue (22,000 galaxies). The 100% catalogue (31,000 galaxies) is now online , so that will be my next little project once I finish with the abstract stuff. Ideally that will be in 360 3D if I can figure out how to get around Blender's silly Cycles image texture limit (or maybe it's already fixed...).
https://www.youtube.com/watch?v=YkR8E6x_2mo
Wibbly-wobbly gassy-wassy
This one uses all-sky HI data to displace a sphere as well as influence its colour.
Monday 21 May 2018
Exploiting the weirdness of data clipping
This one exploits normally annoying rendering artifacts. Blender's realtime display can only show objects from a single side, so for nested transparent spheres you can normally only see their interior or exterior, not both. But by clipping out all the lowest transparency values, part of the other side is revealed. Couple this with a wide-angle lens and changing viewpoint and get quite a lovely mess - a very different effect again compared to the previous straight-line flight.
Testing ESO's review process
Peer review is a waste of everyone's time, study finds
That's the misleading clickbaity headline I'd give to this mostly dry, technical, very detailed paper on whether reviewers are accurately able to evaluate proposals or not. Alternatively :
Yo Dawg, We Herd You Like Reviews, So We Reviewed Your Review Of Reviews So You Can Review While You Review
Aaaannnyway.....
This paper looks at the review process used by the European Southern Observatory to allocate telescope time. There's some good stuff in here, but my goodness it's buried in a lot of (I think unnecessarily complex) statistics. They investigate what they call the "True Grade Hypothesis", which is something familiarly Platonic :
For any given proposal a true grade does exist. The true grade can be determined as the average of the grades given by a very large number of referees.
Which is basically what everyone is assuming throughout the whole process.
The first part of hypothesis is obviously debatable, as it implicitly assumes that an absolute and infinitely objective scientific value can be attached to a given science case. It does not take into account, for instance, that a proposal is to be considered within a certain context and cannot be judged in vacuum. Most likely, a proposal requesting time to measure the positions of stars during a total solar eclipse would have been ranked very highly in 1916, but nowadays it would probably score rather poorly. The science case is still very valuable in absolute terms, but is deprived of almost any interest in the context of modern physics.
The second part of the hypothesis is also subject to criticism, because it implicitly assumes that referees behave like objective measurement instruments. This is most likely not the case. For instance, although the referees are proactively instructed to focus their judgement [no-one will ever convince me you can spell it "judgment", that is plainly ludicrous] on the mere scientific merits, it is unavoidable that they (consciously or unconsciously) take into account (or are influenced by) other aspects. Among these are the previous history of the proposing team, its productivity, its public visibility, the inclusions of certain individuals, personal preferences for specific topics, and so on.
What they find is that the TGH does seem to work, but the scatter is very high. Even for the supposedly best and worst proposals, the refereeing process is only a bit more consistent than random, and it's not really any better at all for the average proposals... in terms of grading, though it does do quite a lot better in terms of ranking.
Even so, their results agree with previous findings that the success of an observing proposal is about half down to the quality of the proposal and half due to luck. However, if you add more referees, results do converge. There might be an optimum number of referees to balance the benefits against the extra resources needed.
Of course a major caveat here is that observers write proposals knowing that they will be reviewed. I'd be any sum of money you like that if you were to peer review proposals that the authors were told weren't being peer reviewed, the peer review system would perform twenty bajillion times better than chance alone.
This is the first paper of two, focusing on the review stage when referees individually analyse their assigned proposals. The second paper will look at the effects of the committee meeting where each reviewer explains their reasoning to the others and the final ranking.
https://arxiv.org/abs/1805.06981
That's the misleading clickbaity headline I'd give to this mostly dry, technical, very detailed paper on whether reviewers are accurately able to evaluate proposals or not. Alternatively :
Yo Dawg, We Herd You Like Reviews, So We Reviewed Your Review Of Reviews So You Can Review While You Review
Aaaannnyway.....
This paper looks at the review process used by the European Southern Observatory to allocate telescope time. There's some good stuff in here, but my goodness it's buried in a lot of (I think unnecessarily complex) statistics. They investigate what they call the "True Grade Hypothesis", which is something familiarly Platonic :
For any given proposal a true grade does exist. The true grade can be determined as the average of the grades given by a very large number of referees.
Which is basically what everyone is assuming throughout the whole process.
The first part of hypothesis is obviously debatable, as it implicitly assumes that an absolute and infinitely objective scientific value can be attached to a given science case. It does not take into account, for instance, that a proposal is to be considered within a certain context and cannot be judged in vacuum. Most likely, a proposal requesting time to measure the positions of stars during a total solar eclipse would have been ranked very highly in 1916, but nowadays it would probably score rather poorly. The science case is still very valuable in absolute terms, but is deprived of almost any interest in the context of modern physics.
The second part of the hypothesis is also subject to criticism, because it implicitly assumes that referees behave like objective measurement instruments. This is most likely not the case. For instance, although the referees are proactively instructed to focus their judgement [no-one will ever convince me you can spell it "judgment", that is plainly ludicrous] on the mere scientific merits, it is unavoidable that they (consciously or unconsciously) take into account (or are influenced by) other aspects. Among these are the previous history of the proposing team, its productivity, its public visibility, the inclusions of certain individuals, personal preferences for specific topics, and so on.
What they find is that the TGH does seem to work, but the scatter is very high. Even for the supposedly best and worst proposals, the refereeing process is only a bit more consistent than random, and it's not really any better at all for the average proposals... in terms of grading, though it does do quite a lot better in terms of ranking.
Even so, their results agree with previous findings that the success of an observing proposal is about half down to the quality of the proposal and half due to luck. However, if you add more referees, results do converge. There might be an optimum number of referees to balance the benefits against the extra resources needed.
Of course a major caveat here is that observers write proposals knowing that they will be reviewed. I'd be any sum of money you like that if you were to peer review proposals that the authors were told weren't being peer reviewed, the peer review system would perform twenty bajillion times better than chance alone.
This is the first paper of two, focusing on the review stage when referees individually analyse their assigned proposals. The second paper will look at the effects of the committee meeting where each reviewer explains their reasoning to the others and the final ranking.
https://arxiv.org/abs/1805.06981
LAB vortex
A combination of techniques I've used before individually. LAB data is here rendered as a series of nested spheres with colour based on the velocity. Blender's realtime view is used, as that clips faces pointing away from the camera. Finally it also uses an extremely wide-angle lens, which gives the distortion affects at the edges.
Friday 18 May 2018
Non-spectral lines
OK, so it's not the most intuitive or informative way of looking at radio astronomy hydrogen data. But it is pretty, so there.
Thursday 17 May 2018
A 3D hydrogen Matrix
This one took a while to figure out. A Matrix-style wall of text shows the flux values in progressive velocity channels through a data cube. Both the position of each text object and its value (scrambled slightly to give a more Matrixy-look) are derived from the data. Most of the data is noise, but the brightest regions correspond to galaxies detected in neutral hydrogen. These data values are shown as persistent text objects at a position corresponding to their velocity channel, and are sampled more fully than the rest of the noise. Also there are various reflecting planes at work to make it look cool.
Tuesday 15 May 2018
More criticism of that galaxy without dark matter
More criticism of that galaxy without dark matter again. Unfortunately, while some other critiques are legitimate, this one is just plain silly.
The first problem we see is that, beside the positional proximity in the sky, there is no other evidence that the globular clusters used to derive the velocity dispersion are physically bound to the galaxy.
Umm... you mean other than the fact they're at an almost identical velocity to the galaxy ? Because that's pretty compelling evidence. I'd say it's virtually certain that they're associated.
Neither is there any evidence that the system is virialized and isotropic.
Other than the fact that it appears to be spherical.
What if, because of anisotropy, the line of sight velocity dispersion is not representative of the true value... it could perfectly be the case that the globular cluster system studied by van Dokkum and collaborators forms a flattened disk with substantial rotational support and is being observed close to face-on, as the image of the galaxy itself might suggest.
It could, but disc galaxies rarely if ever look that smooth. Possible though.
And what if the system is not virialized? Of course, if not virialized, the velocity dispersion of the clusters is not representative of the mass of the galaxy and the claim is invalidated.
Sure, but then why are the velocities of the clusters so damn similar to the galaxy ? That makes no sense.
Thus, according to van Dokkum and collaborators, this galaxy is unique in at least three different aspects: it is the only galaxy know not to contain dark matter, it is the only galaxy know to have such an extremely large number of globular clusters, and it is the only galaxy hosting a population of globular clusters not obeying to the universal luminosity function. All this supported only by their own claim that the selected clusters are physically related to the galaxy.
Oh lordy. This is trying to say that the galaxy is too weird to exist, so therefore it doesn't. Well, it does. And the globular clusters are clearly associated with it. You can't dismiss weird objects just because you don't like them. Outliers do occasionally overthrow theories, but more often they just point to incompleteness : implicit assumptions that were previously made that don't really say much about the fundamental mechanism proposed by the theory.
Any scientist knows that doing statistic with such a small number of measurements is dangerous, and not sufficient to base an exceptional claim on. Nevertheless, the authors went on to calculate the dispersion and instead of doing the strait forward calculation, they used a biweight approach that basically kicked out one cluster (as any one can see in fig 3b of their Nature paper). They basically neglected the velocity of cluster 98, precisely the value that disagrees with their claim!
Right, but these criticisms have already been made and the author's have responded (reasonably well, in my opinion).
'Tis true that the small numbers are a concern. But the other globular clusters are so close to the systemic velocity of the galaxy, it's damn hard to believe that one outlier is anything other than an outlier. There's also a whiff of too many explanations here : a determination to show that because there are so many other possibilities, it's surely more likely that one of them must be correct and therefore downgrade the probability of the original claim. This is a lot like climate denier strategies to say that global warming is either a hoax, natural but not significant, natural but beneficial, etc... anything besides the claim that warming is real, artificial and harmful.
Of course the problem is that having lots of crappy alternatives is no match so a single really good option. Here, the authors first imply the globular clusters aren't associated with the galaxy, then that they are associated but the system isn't stable, or is stable but measured from the wrong angle, or is a stable sphere after all but the statistics are wrong. This is not a sensible critique, especially because none of them are discussed in any detail. Without some estimate of the probabilities of the alternatives - even just a very hand-wavy one - then this amounts to little more than making stuff up. It's been long enough since the original to do some crude estimates of the different options. The original paper may not be a watertight piece of statistical analysis, but it's a damn sight better than not doing any at all.
Having said all the above, we cannot avoid to ask ourselves how could such a result have been published at all, wondering how a similar paper would have been received if the conclusion were the other way around. This points to the responsibility of journals that at present adopt standards orders of magnitude lower to publish results favouring the dark matter hypothesis compared to the ones required to papers claiming the opposite. Our sad conclusion is that science cannot progress this way.
Utter nonsense. The critics didn't even bother to check that these criticisms have already been raised. I was going to say that this kind of approach belongs on a blog, but van Dokkum's blog is far better than this !
Much more interesting, I think, is that the velocity dispersion of this object is so low it's amazing it's (apparently) spherical at all - there hasn't been time for the stars to have more than a couple of orbits. No-one seems to want to mention that.
https://arxiv.org/abs/1805.04817
The first problem we see is that, beside the positional proximity in the sky, there is no other evidence that the globular clusters used to derive the velocity dispersion are physically bound to the galaxy.
Umm... you mean other than the fact they're at an almost identical velocity to the galaxy ? Because that's pretty compelling evidence. I'd say it's virtually certain that they're associated.
Neither is there any evidence that the system is virialized and isotropic.
Other than the fact that it appears to be spherical.
What if, because of anisotropy, the line of sight velocity dispersion is not representative of the true value... it could perfectly be the case that the globular cluster system studied by van Dokkum and collaborators forms a flattened disk with substantial rotational support and is being observed close to face-on, as the image of the galaxy itself might suggest.
It could, but disc galaxies rarely if ever look that smooth. Possible though.
And what if the system is not virialized? Of course, if not virialized, the velocity dispersion of the clusters is not representative of the mass of the galaxy and the claim is invalidated.
Sure, but then why are the velocities of the clusters so damn similar to the galaxy ? That makes no sense.
Thus, according to van Dokkum and collaborators, this galaxy is unique in at least three different aspects: it is the only galaxy know not to contain dark matter, it is the only galaxy know to have such an extremely large number of globular clusters, and it is the only galaxy hosting a population of globular clusters not obeying to the universal luminosity function. All this supported only by their own claim that the selected clusters are physically related to the galaxy.
Oh lordy. This is trying to say that the galaxy is too weird to exist, so therefore it doesn't. Well, it does. And the globular clusters are clearly associated with it. You can't dismiss weird objects just because you don't like them. Outliers do occasionally overthrow theories, but more often they just point to incompleteness : implicit assumptions that were previously made that don't really say much about the fundamental mechanism proposed by the theory.
Any scientist knows that doing statistic with such a small number of measurements is dangerous, and not sufficient to base an exceptional claim on. Nevertheless, the authors went on to calculate the dispersion and instead of doing the strait forward calculation, they used a biweight approach that basically kicked out one cluster (as any one can see in fig 3b of their Nature paper). They basically neglected the velocity of cluster 98, precisely the value that disagrees with their claim!
Right, but these criticisms have already been made and the author's have responded (reasonably well, in my opinion).
'Tis true that the small numbers are a concern. But the other globular clusters are so close to the systemic velocity of the galaxy, it's damn hard to believe that one outlier is anything other than an outlier. There's also a whiff of too many explanations here : a determination to show that because there are so many other possibilities, it's surely more likely that one of them must be correct and therefore downgrade the probability of the original claim. This is a lot like climate denier strategies to say that global warming is either a hoax, natural but not significant, natural but beneficial, etc... anything besides the claim that warming is real, artificial and harmful.
Of course the problem is that having lots of crappy alternatives is no match so a single really good option. Here, the authors first imply the globular clusters aren't associated with the galaxy, then that they are associated but the system isn't stable, or is stable but measured from the wrong angle, or is a stable sphere after all but the statistics are wrong. This is not a sensible critique, especially because none of them are discussed in any detail. Without some estimate of the probabilities of the alternatives - even just a very hand-wavy one - then this amounts to little more than making stuff up. It's been long enough since the original to do some crude estimates of the different options. The original paper may not be a watertight piece of statistical analysis, but it's a damn sight better than not doing any at all.
Having said all the above, we cannot avoid to ask ourselves how could such a result have been published at all, wondering how a similar paper would have been received if the conclusion were the other way around. This points to the responsibility of journals that at present adopt standards orders of magnitude lower to publish results favouring the dark matter hypothesis compared to the ones required to papers claiming the opposite. Our sad conclusion is that science cannot progress this way.
Utter nonsense. The critics didn't even bother to check that these criticisms have already been raised. I was going to say that this kind of approach belongs on a blog, but van Dokkum's blog is far better than this !
Much more interesting, I think, is that the velocity dispersion of this object is so low it's amazing it's (apparently) spherical at all - there hasn't been time for the stars to have more than a couple of orbits. No-one seems to want to mention that.
https://arxiv.org/abs/1805.04817
Friday 11 May 2018
Thursday 10 May 2018
A hydrogen infinity sphere with extra colours
As with the previous gif, but now with extra colours. Needs some extra frames to make the interesting bit longer.
A hydrogen infinity sphere
This one is driven more by an effort to make a pretty picture than display the data in some fancy way. The main sphere has a texture of all-sky hydrogen data, so there's nothing clever about wrapping that to a sphere. But inside the data sphere is another sphere which is purely reflective, and outside the data sphere is another reflecting sphere which is included in the raytracing but not rendered. So the effect is of a sphere with an infinite series of reflections of the data... well, OK, probably more like ten. The raytracing depth is limited and the reflectivity of the spheres isn't quite perfect, so successive reflections appear darker.
It might be fun to transition this one to a volumetric spherical display.
The scientific business model (but only sometimes)
An unfortunate reality. Some significant caveats however :
- Some journals require no or minimum payment per paper, funded instead by institutional subscriptions.
- Payment for access also funds the process of finding the reviewers. This isn't such a minor thing (witness that peer-reviewed paper about the time travelling alien octopus).
- Access doesn't always require payment, since virtually identical versions can be found on arXiv.
- While reviewing is done by other scientists, it's not really the case that they review it for free. Most institutes accept that reviewing is an essential and unavoidable part of the job, so reviewing is done on work time. Hence reviewers are getting paid for this - they're sacrificing their own research time, not salary.
Will No-One Rid Me Of This Turbulent Sphere ?
This is my ninth paper as first author. The title was suggested by the ever-astute Robert Minchin (see the acknowledgements section, and if you've never heard of Thomas Becket you should consult wikipedia), and the hard work of the simulations was done by the second author.
While there's a full blog post on this for for the enthusiasts who despise reading dry academic texts, here's the short version. Back in 2012 we discovered these eight dark clouds of hydrogen floating around in the Virgo cluster. At first glance there's nothing too odd about that, quite a few of these are already known - some of which are a lot more impressive than ours. But ours are weirder. They're quite compact but six have high velocity dispersions, meaning they should be rapidly disintegrating. This is especially odd considering they're literally miles away from normal galaxies - if they were ripped out of galaxies, as is the conventional explanation for such clouds, then where are the expected streams of gas ? It's hard to see how you could rip out a compact gas cloud from a galaxy without a lot more debris spread all over the place, and we didn't find any.
Subsequently we confirmed this in great and tedious detail with a long series of simulations looking at tidal encounters between galaxies. We found that this mechanism has extreme difficultly in producing features matching the three key parameters of the clouds : their physical size, isolation, and high velocity dispersion. Oh, isolated clouds with dispersion < 50 km/s are common as muck, and those of < 100 km/s are just a bit unusual, but those > 100 km/s are so rare that it's just not a sensible interpretation for the real objects. Even with the important provision that weird objects imply weird physics - the rate at which they should form is just too dang low.
Then in 2016, another explanation was proposed by a certain Burkhart & Loeb. Their idea was that the clouds are in pressure equilibrium with the hot, thin intracluster gas. This gas pervades a large fraction of massive galaxy clusters and isn't bound to any individual galaxy. Though it's very low density, it's also extremely hot, so its pressure can be significant. They made an analytic calculation which uses the measured velocity dispersion of the clouds to estimate their internal pressure, X-ray data to get the external pressure of the surrounding gas, and then by equating the two they can find the expected size of the clouds. Which, interestingly, is in good agreement with the observations (limited though they are).
Now, if the velocity dispersion of the clouds was lower, that would be fine and we would have stopped there. The problem is that it's so high that the internal pressure of the clouds, in that model, has to come from bulk motions - i.e., turbulence (if the dispersion came from thermal pressure the clouds would have to be so hot they couldn't possibly remain atomically neutral, as they actually are). And that's a completely different state from thermal pressure, which acts uniformly. Turbulence implies that different bits of the clouds are moving in different directions, so it's hard to see how the external pressure could ever help them remain stable.
So we set up this series of simulations to test how long the clouds could match the observations in this scenario. Our models have spherical gas clouds set to be compatible with the observations, embedded in a hot, low-density gas to mimic the intracluster medium. We varied the mass of the gas clouds, their velocity dispersion, and more detailed parameters of the structure of their velocity field. Given the mass of the clouds, we know that their velocity is so high they can't be gravitationally bound, so we don't expect any equilibrium state to develop - what's much harder to predict is just how long they survive, given the external pressure helping to hold them together. It's even harder to guess how long they'd maintain that very interesting high velocity dispersion.
Our result is... not very long at all. Well, they do survive in the crude sense that there's still gas there, but the dispersion drops very rapidly. The clouds tend to either immediately explode, initially collapse and then slowly disperse, or are rapidly heated by their own self-interactions to the point where they'd no longer be detectable (you can watch one of the simulations here : https://plus.google.com/u/0/+RhysTaylorRhysy/posts/Buvd5QH9NR3). The high velocity dispersion lasts no more than about 100 Myr, which is so short in this context that it's safe to say the surrounding gas hasn't really helped at all. At most, all it's done is change the manner of their demise, not postpone it.
Could the model be saved or is something else at work ? Well the original model is dead in the water : turbulence is just not at all the same as thermal pressure. But in combination with other physics - rotation, or that chronic astronomical bogeyman known as magnetic fields maybe it would play a more significant role. It might help, but personally I doubt it : there's no good mechanism to explain where the velocity dispersion comes from, or what maintains it.
The most radical possibility of all is currently the front-runner to explain these objects : they're rotating, and bound together by a massive dark matter halo, and their gas density is so low they're not able to form any stars. That naturally explains the high "dispersion" of the objects and how it's maintained, and fits with chronic problem that simulations predict far more galaxies than are observed. Of course, it has its own problems, like how such low-density cool gas could survive as it moves through the hot cluster gas, but that's for a future paper.
https://arxiv.org/abs/1805.03414
While there's a full blog post on this for for the enthusiasts who despise reading dry academic texts, here's the short version. Back in 2012 we discovered these eight dark clouds of hydrogen floating around in the Virgo cluster. At first glance there's nothing too odd about that, quite a few of these are already known - some of which are a lot more impressive than ours. But ours are weirder. They're quite compact but six have high velocity dispersions, meaning they should be rapidly disintegrating. This is especially odd considering they're literally miles away from normal galaxies - if they were ripped out of galaxies, as is the conventional explanation for such clouds, then where are the expected streams of gas ? It's hard to see how you could rip out a compact gas cloud from a galaxy without a lot more debris spread all over the place, and we didn't find any.
Subsequently we confirmed this in great and tedious detail with a long series of simulations looking at tidal encounters between galaxies. We found that this mechanism has extreme difficultly in producing features matching the three key parameters of the clouds : their physical size, isolation, and high velocity dispersion. Oh, isolated clouds with dispersion < 50 km/s are common as muck, and those of < 100 km/s are just a bit unusual, but those > 100 km/s are so rare that it's just not a sensible interpretation for the real objects. Even with the important provision that weird objects imply weird physics - the rate at which they should form is just too dang low.
Then in 2016, another explanation was proposed by a certain Burkhart & Loeb. Their idea was that the clouds are in pressure equilibrium with the hot, thin intracluster gas. This gas pervades a large fraction of massive galaxy clusters and isn't bound to any individual galaxy. Though it's very low density, it's also extremely hot, so its pressure can be significant. They made an analytic calculation which uses the measured velocity dispersion of the clouds to estimate their internal pressure, X-ray data to get the external pressure of the surrounding gas, and then by equating the two they can find the expected size of the clouds. Which, interestingly, is in good agreement with the observations (limited though they are).
Now, if the velocity dispersion of the clouds was lower, that would be fine and we would have stopped there. The problem is that it's so high that the internal pressure of the clouds, in that model, has to come from bulk motions - i.e., turbulence (if the dispersion came from thermal pressure the clouds would have to be so hot they couldn't possibly remain atomically neutral, as they actually are). And that's a completely different state from thermal pressure, which acts uniformly. Turbulence implies that different bits of the clouds are moving in different directions, so it's hard to see how the external pressure could ever help them remain stable.
So we set up this series of simulations to test how long the clouds could match the observations in this scenario. Our models have spherical gas clouds set to be compatible with the observations, embedded in a hot, low-density gas to mimic the intracluster medium. We varied the mass of the gas clouds, their velocity dispersion, and more detailed parameters of the structure of their velocity field. Given the mass of the clouds, we know that their velocity is so high they can't be gravitationally bound, so we don't expect any equilibrium state to develop - what's much harder to predict is just how long they survive, given the external pressure helping to hold them together. It's even harder to guess how long they'd maintain that very interesting high velocity dispersion.
Our result is... not very long at all. Well, they do survive in the crude sense that there's still gas there, but the dispersion drops very rapidly. The clouds tend to either immediately explode, initially collapse and then slowly disperse, or are rapidly heated by their own self-interactions to the point where they'd no longer be detectable (you can watch one of the simulations here : https://plus.google.com/u/0/+RhysTaylorRhysy/posts/Buvd5QH9NR3). The high velocity dispersion lasts no more than about 100 Myr, which is so short in this context that it's safe to say the surrounding gas hasn't really helped at all. At most, all it's done is change the manner of their demise, not postpone it.
Could the model be saved or is something else at work ? Well the original model is dead in the water : turbulence is just not at all the same as thermal pressure. But in combination with other physics - rotation, or that chronic astronomical bogeyman known as magnetic fields maybe it would play a more significant role. It might help, but personally I doubt it : there's no good mechanism to explain where the velocity dispersion comes from, or what maintains it.
The most radical possibility of all is currently the front-runner to explain these objects : they're rotating, and bound together by a massive dark matter halo, and their gas density is so low they're not able to form any stars. That naturally explains the high "dispersion" of the objects and how it's maintained, and fits with chronic problem that simulations predict far more galaxies than are observed. Of course, it has its own problems, like how such low-density cool gas could survive as it moves through the hot cluster gas, but that's for a future paper.
https://arxiv.org/abs/1805.03414
A galactic blanket
Here's one I left rendering overnight. This is a different data set than the previous ones, looking at Milky Way hydrogen over an area about 4 times larger than the the other cube. It doesn't have as many frequency channels so I interpolated some extra ones for a smoother effect. This is rendered as a conventional surface, which works quite well for this naturally-landscapey data set.
Scanning a galaxy
And another one. To recap, this is a plot through an HI data cube, where instead of plotting a map of the brightness of the gas, the measured flux corresponds to the height of the line. Each line is a horizontal slice through the data at a single frequency channel, effectively a plot of flux as a function of right ascension (NOT a spectrum !). All I've done here is have each line reveal itself successively, adding a material to make the newly-revealed lines brighter and the older ones faded.
Next up will be to try and simultaneously animate both the appearance of the lines and their corresponding spectral channel, so that the entire surface will be constantly in motion.
Wednesday 9 May 2018
Bandpass lines in the Milky Way
And another one, this time with lines. It should be interesting to animate the sequential appearance of each line with the material darkening after the initial appearance... first all drawn from one channel, then at some point animating the channels as well.
(if you've no idea what this is, you're just gonna have to scroll through this category cos I'm not typing the whole thing again)
An interactive grant proposal system
You may remember that a few weeks ago I described this UAE "Space Settlement Challenge" seed grant. I was co-I on a proposal which was not successful - we were ranked 74th out of 260, which isn't bad going, but not enough to make the cut. There are still quite a lot of things I like about this funding system, and a few I don't.
The basic idea is that, after some initial rejection of illegible proposals (e.g. not written in English and relating to the topic at hand), everyone who submits a proposal is also a reviewer. Each proposal is randomly assigned three reviewers in a double-blind system (neither side know who the others are - all they see is the proposal and the reviewer comments). Proposals have to follow a template answering in 200 words (this is only a very small grant, and not a big deal at the end of the day) for each section, which are as follows :
- Summary : Write a brief non-technical description of the project or research, work to be done, and significance of your idea.
- Previous work : Describe what is known about the topic and what research has been conducted previously, if applicable.
- Impact : Describe what is new and innovative about this project or research, what is the broader impact, who and how it will benefit.
- Aims : Describe in detail the specific objectives of your project or research.
- Methods : Provide a detailed work plan of the activities you plan to carry out.
- Budget.
All three reviewers assign a score of 1-6 for each section (6 is the highest) along with comments. Then the scores for all proposals are sorted, the very highest are straightaway accepted, the worst are rejected, and the middle ground get a chance to re-submit a refined version, accounting for the feedback they received. Then those proposals get reviewed again.
This is an extremely fast way to evaluate a large number of proposals, and has some attempt to reduce reviewer bias thanks to anonymity. The refinement stage in particular is an excellent idea, because it's usually extremely frustrating to receive reviewer's comments without being able to respond to them.
What needs some considerable work here, however, is clarity. For example, in this case there was enough money available for about 30 fully-funded proposals up to the maximum budget, but presumably a few more if the best proposals asked for less than the maximum permitted funds. But 260 >> 30, so is there any need for refinement ? There were surely more than 30 highly scored proposals here, so you could select enough to consume all the available funding anyway. Also, inside the system there's a "cutoff" value (37), but it's not clear what it means. My guess would be that it means 37 proposals were funded. Perhaps they anticipated a much smaller competition pool and would fund all proposals above a certain score (after refinement) and reject any below this even if they still had available funds.
Also, the description of what's expected is far too short and open to misinterpretation. For example, we presumed that the methods section referred to logistics - how we'd collaborate as a group, communicate, etc. Two of the referees basically agreed, but the third interpreted this to mean scientific methods - and justifiably gave us a very low score.
The other issue is the amount, if any, of external oversight. It's implied that there must be some, since proposals are first confirmed if they're eligible before proceeding to the review section. Yet one of the proposals we reviewed wasn't about space at all (it didn't even mention the word !) and had references to figures that didn't exist. It was clearly garbage and never should have made the initial cut. And while getting everyone to review each other is an interesting idea, is anyone reviewing the reviewers ? Two of ours were overall positive but the third wholly negative - even on the extremely clear and detailed budget section. Someone ought to check if some reviewers only give low or high scores
Oh well, there's tonnes of other projects to be getting on with anyways.
https://www.guaana.com/funding/grants/mbrchallenge/details
The basic idea is that, after some initial rejection of illegible proposals (e.g. not written in English and relating to the topic at hand), everyone who submits a proposal is also a reviewer. Each proposal is randomly assigned three reviewers in a double-blind system (neither side know who the others are - all they see is the proposal and the reviewer comments). Proposals have to follow a template answering in 200 words (this is only a very small grant, and not a big deal at the end of the day) for each section, which are as follows :
- Summary : Write a brief non-technical description of the project or research, work to be done, and significance of your idea.
- Previous work : Describe what is known about the topic and what research has been conducted previously, if applicable.
- Impact : Describe what is new and innovative about this project or research, what is the broader impact, who and how it will benefit.
- Aims : Describe in detail the specific objectives of your project or research.
- Methods : Provide a detailed work plan of the activities you plan to carry out.
- Budget.
All three reviewers assign a score of 1-6 for each section (6 is the highest) along with comments. Then the scores for all proposals are sorted, the very highest are straightaway accepted, the worst are rejected, and the middle ground get a chance to re-submit a refined version, accounting for the feedback they received. Then those proposals get reviewed again.
This is an extremely fast way to evaluate a large number of proposals, and has some attempt to reduce reviewer bias thanks to anonymity. The refinement stage in particular is an excellent idea, because it's usually extremely frustrating to receive reviewer's comments without being able to respond to them.
What needs some considerable work here, however, is clarity. For example, in this case there was enough money available for about 30 fully-funded proposals up to the maximum budget, but presumably a few more if the best proposals asked for less than the maximum permitted funds. But 260 >> 30, so is there any need for refinement ? There were surely more than 30 highly scored proposals here, so you could select enough to consume all the available funding anyway. Also, inside the system there's a "cutoff" value (37), but it's not clear what it means. My guess would be that it means 37 proposals were funded. Perhaps they anticipated a much smaller competition pool and would fund all proposals above a certain score (after refinement) and reject any below this even if they still had available funds.
Also, the description of what's expected is far too short and open to misinterpretation. For example, we presumed that the methods section referred to logistics - how we'd collaborate as a group, communicate, etc. Two of the referees basically agreed, but the third interpreted this to mean scientific methods - and justifiably gave us a very low score.
The other issue is the amount, if any, of external oversight. It's implied that there must be some, since proposals are first confirmed if they're eligible before proceeding to the review section. Yet one of the proposals we reviewed wasn't about space at all (it didn't even mention the word !) and had references to figures that didn't exist. It was clearly garbage and never should have made the initial cut. And while getting everyone to review each other is an interesting idea, is anyone reviewing the reviewers ? Two of ours were overall positive but the third wholly negative - even on the extremely clear and detailed budget section. Someone ought to check if some reviewers only give low or high scores
Oh well, there's tonnes of other projects to be getting on with anyways.
https://www.guaana.com/funding/grants/mbrchallenge/details
Tuesday 8 May 2018
The hydrogen Matrix
I'm fascinated by the idea of rendering HI data cubes as landscapes. But they're not real landscapes; how you interpret them depends on how visualise them. Hence, here's M33 in my first attempt at a Matrix-style rendering. Each frame of the animation shows a single frequency channel where both the offset and value of the text indicate the flux at that position (the Matrix font for some reason renders a minus sign as something like a 7, and the formatting doesn't round negative numbers correctly, hence there are lots of apparent values of 70). It needs some more frames interpolated to slow it down but this is just a first test.
Saturday 5 May 2018
The roiling seas of Triangulum
One more... this time with the colour and height both depending on intensity. The colour scheme has a higher contrast than I'd like, but that's mainly due to the nature of the data and its highly dynamic dynamic range. Still, I like the way a varying landscape emerges from a sea of noise...
The hydrogen foothills of the Milky Way
And another one. More background for those who have no idea what's going on...
HI data cubes map the atomic gas content across the sky. As well as the "brightness" (actually density) of the gas, they also record what frequency it's emitted at. The frequency depends on how fast the gas is moving towards or away from us, which is very roughly correlated with its distance. The correlation gets better when you start to go to distant galaxies, but is pretty meaningless within our own galaxy - its own rotation dominates.
There are lots of ways to visualise the data depending on what you want to do with it. Normally you'd pick some fixed velocity and then make a 2D map of the density. Or you can render lots of different velocities at once and get a 3D image In this case I've used the density to set both the colour but also the height of each pixel, creating a landscape. Each pixel has the same frequency.
What this actually is is part of the gas in the Milky Way, with more dense regions shown as blue peaks and less dense ones shown as red troughs. You can also see some perfectly linear features - some very narrow, others broad. These happen because our survey isn't really designed for detecting signal across the whole field of view. While the telescope calibrates the data to some extent, we also rely on the signal only occupying a small fraction of the view at any given frequency. We use this to estimate the zero level of the data, and for pretty much every other data set this works extremely well. But in the Milky Way, where there's gas all over the place, it doesn't, hence the data levels are at best unreliable and at worst meaningless. The linear shape of these artifacts occur because we construct the data by making narrow scans across the sky, so if something changes from one scan to another (e.g. if there's a particular big, bright bit of gas in one scan but not so much in an adjacent one), they look noticeably different.
The reason I started trying this was because most of the sources are much less interesting : they're just unresolved points. Knowing the shape of the telescope beam (a Gaussian) and the maximum flux, it should be possible to remove them (hopefully revealing any more extended sources that might be present). I've done this before, but this time it wasn't working so well. The Gaussian model of the source looked like a good approximation of the real source, but it's hard to tell exactly how close it is by comparing colours. It's much easier to do this by comparing surface heights, which let you compare the data over the 2D area rather than pixel by pixel. But now I've got distracted by how pretty the data looks... well, it is a Saturday, after all. :P
Next up : animate the frequencies to generate a surface that changes both shape and colour...
The Mountains of Triangulum
I thought it might be fun to take an HI data cube and plot each slice as a surface, rather than a map. This is the M33 galaxy at a particular frequency where the height, rather than the more usual colour, indicates brightness.
More to come when I can get the frequencies to animate.
Subscribe to:
Posts (Atom)
-
Of course you can prove a negative. In one sense this can be the easiest thing in the world : your theory predicts something which doesn...
-
In the last batch of simulations, we dropped a long gas stream into the gravitational potential of a cluster to see if it would get torn... -
Another day, another paper on how exciting Ultra Diffuse Galaxies are. At first, these large, faint galaxies were just wholly unexpected, a...