Spirals in Galaxies and Large Surveys by Prof. Karen Masters

okay let's get started thank you everybody for coming I'm delighted this today uh to introduce our colloquium speaker Professor Karen Masters who has come to us from haford college in Pennsylvania um Professor Masters was previously of the institute for Gra on cosmology in Portsmouth and before that Harvard and Cornell in Oxford uh she's an expert on exactic extragalactic astronomy in particular working with a bunch of large surveys and large collaborations so is the I'm G to get the wrong way around the pi for Galaxy Zoo yes and uh the spokesperson for sss4 obviously I'm never going to be the spokesperson given that yeah for sss4 so she's working a lot with these manga Galaxy SL digital Cisco survey citizen science and we're going to hear about some of that today and she's going to tell us about spirals in galaxies so please thank you Chris for a very nice invitation uh or introduction and uh for the invitation to come and speak today um so I'm going to talk about spirals in galaxies as Chris mentioned I have broad interests in Galaxy Evolution with a kind of focus on why morphology is interesting and what morphology can tell us from Big surveys um but in particular this talk uh came out of some work I did on my last sabatical which was a CO sabatical so I spent it in my basement um spending a lot of time uh well it's okay it was fine uh talking uh with Jerry salwood about spiral structure and and we wrote on this review article um so I'm really going to focus on observations of spirals um I'm well aware that there are some people in the audience who spend a lot of time thinking about the theory of SP spirals but of course U maybe not everyone does that so I do have some introduction to why spirals are interesting and some of the background about what we know about Spirals and then I'll kind of focus on observations I've been involved with with the projects I work on um so it is a little bit uh biased uh towards the work myself and my students have done but um hopefully you'll forgive me for that so spiral uh have been observed in Galaxies for almost around 200 years I would say I just always enjoy starting with this uh drawing ing of the wellp pool Galaxy done by an asomo from Ireland uh in the uh mid of mid mid part of the 1800s um comparing it it's the worldo Galaxy right so this is the Hubble image of the same galaxy um this is the first recorded uh observation we have of spiral structure in galaxies um and people are obviously very struck by this uh so there are suggestions that it was the inspiration for the spiral in the middle of Spar night um so these are beautiful objects and and fascinating objects with some really interesting dynamical physics if you've taken an astronomy class or taught sort of Astro 101 you've probably taught about the classification of galaxies using the Hubble tuning fork um good reminder um of the kind of big picture here of galaxies uh Edwin Hubble looked at uh photographic plate images of around a thousand galaxies and put together this classification scheme in the 1920s uh where we have elliptical galaxies so these smooth reddish the more massive galaxies on one side um and spirals on the other side with this Fork set by whether or not they have a galactic bar uh this this linear structure across the middle um the diagram here is Hubble's sketch um and the images are his example images but scss versions of them um so one other note on terminology I already mentioned elliptical tend to be red spirals tend to be blue you can see that actually in the scss images that I've put here um we also tend to use early type for things at this side of the diagram and late type for that side and one of my pet peeves is people then go on to say and Hubble thought this was an evolutionary sequence it's not true can find a footnote in one of his papers about how that was not true um and he was just using an analogy to late and early type stars to provide some names he never thought it was a sequence in time random aside to my pet peeve uh so you've probably either taught this or seen it in a class uh um since we're talking about spirals a little bit more detail about the spiral part of the sequence um so spirals are ordered from an sa an early type spiral to an sc or a late type spiral um based on uh two two or three different properties in Hubble's classification um and so sort of implicitly we're saying these properties correlate with each other right um so SAS uh in in Hubble's language have large bulges tightly wound spirals which are very clear and distinct uh whereas you move to sc's where you have small bulges Loosely wound arms um and the spirals tend to be patchier um another interesting side note so as as recently as 10 years ago Rona one of the real experts uh in traditional morphology is talking about how if there's a conflict here so if you have something with a bit a large bulge and loose arms or a small bulge and tight arms um you should classify it based on the arm winding um that is just not what's done these days in big surveys um this distinction between early type and late type well is almost entirely on bulge size uh in big surveys so just to note that those two things are different okay there's more complicated classification schemes for spiral so dealore in the 1950s had this sort of wheel thing where um we put Spirals and Rings um on a classification scheme uh this doesn't work very well sorry uh one of my favorite words in astronomy flocculent uh is uh meaning literally fleece or cloud-like these sort of very patchy spirals um as opposed to Grand Design the very large scale stretching across large radial range that typically symmetric um this was a classification or a separation developed by the eligens in the 1980s um and you see spirals all over the place so there's something very interesting and like generic about spirals in dynamical systems made of numbers of particles um so just some examples of spiral um in a protoplanetary disc in the ofus star formation region you can make spiral in the cream in your coffee cup uh Saturn's rings actually are extremely tightly round spirals in places um and of course hurricanes spirals a little bit of a spiral math now um so in nature most of these spirals are something called a log spiral um so you have the log of the radius is proportional to the asmal angle um and the uh the angle that this log spiral makes to the tangent to the circle at the same radius is called the pitch angle um and in a log spiral that pitch angle is constant with radius so it's one number um so you know Nautilus shells mandle BR sets all kinds of things of versions of log spirals mathematically so again just a reminder large pitch angle is very open spirals so 90 degrees would be spokes like in a bicycle wheel um and a pitch angle of zero is ring so if you've uh taken a galactic Dynamics class you'll heard of the winding problem um but just a reminder of what this is right galaxies rotate differentially very famously right they have flat rotation curves with radius so that's the flat circular velocity that means the inner part um does uh more rotations in the same amount of time than the outer part um so if spirals are material structures if they're like ribbons uh they W would wind up over time in a galaxy um so of course this was the original argument that they cannot be material they must be some kind of density wave a reminder of what a density wave is using an analogy uh with cars instead of stars um it's a really fun little video um that I a student of mine found once um where it's a circular well the cars are driving in a circle and you're going to see a little density wave get set up over here you can see right the density wave starts to move this this way actually but the stars are moving the cars are moving this way through it um so the idea is in a spiral galaxy the spiral arms are this kind of feature right the the stars are bunching up obviously they're not changing speed in the traffic jam like cars are um the density is very low right they're not interacting with each other um but the spirals are places in the Galaxy where the orbits of the stars are bringing them slightly closer together um so density waves um so this is another classic diagram talking about these sort of electrical orbits of stars in a galaxy um and you can start to orient them appropriately that that they end up looking like spirals um many people uh I think would agree that a large number of spirals have driven by uh hides interactions of galaxies um so the wellp pool galaxy is the Prototype of this with its companion um clearly connected to the spiral in some way um those also discussion of how the galactic bar might drive spirals in some cases it's definitely the case observationally the bar spirals much more likely to have two arms two spiral arms uh than other numbers of arms so there appears at least superficially to be a connection between bars and spirals um oh I find this one interesting and maybe everyone knows this but I didn't know this until I was working on the review um density waves should Trail which means the the spiral arm the sense of the arm is that the Galax is rotating this way right so the arm is sort of trailing behind the rotation um you can't have a density wave that goes the other way I think someone can correct me on that if I'm wrong um but it's actually a really hard measurement to make um we're good at rotation in galaxies observationally the doler shift works really well we can get the rotation really easily the tricky thing is determining the orientation of the Galaxy what the front is and what the back is um this diagram makes it look not tricky at all because these are beautiful examples of galaxies where you can tell because of the Dust lames um and so these are tocal sketches of these galaxies where you can see like this is the front because the dust Lane cuts across the Bulge and you can see a little dust Lane here so this is the front um but it's not as if you look at just a general image of a galaxy it's not that easy to tell um and it's certainly not something you can automate very well um so we've actually only done this in around about 100 galaxies um to determine observationally that they are actually trailing and almost all of hundred are trailing so there's no cause for concern um but I just think it's interesting that in today's sort of Big Data astronomy we only have this measurement for 100 galaxies yeah there are ones that are not or yeah and people sort of argue about them in the literature a little bit um why that might be but yeah there is at least evidence there's a couple of spirals where there appears to be some trailing and some leading arms which is very weird there's a lot of galaxies so there's always some weird ones yeah okay um but back to density waves um you know we said there might be some kind of effect of of environment driving spirals certainly there you know we know the galaxies interact a lot with their environment bars uh may have some kind of role to play in driving spirals in some galaxies but there are very isolated spirals that don't have bars um and they still have spirals so there must be a way to generate spirals fairly independent of you know other stuff going on um and one of the sort of interesting things that's happened in the last I guess 10 or more years is that in numerical simulations so this is a hydro uh simulation has stars and gas we're seeing the spirals actually are winding to some extent they are still density waves so they're not winding as fast as they would if they were material but um they are not uh they don't have a a constant patent speed with radius so they're not rotating at the constant patent speed they are winding up over time um so one way to generate that that that Jerry uh cwood has has worked a lot on these self- excited instabilities so overlapping modes uh in a galaxy um I'm not particularly a dynamicist um so I like to kind of think of physical analogies what this what is going on here um so I come sometimes think about like wave interference patterns you get a very coherent pattern from multiple waves overlapping U perhaps a better analogy is this kind of uh beautiful patterns you can get from uh applying sound waves to metal plates so you can get all kinds of interesting patterns know in a in a spiral galaxy of course right it it does have an edge but the edge kind of Fades off it is rotating so there's some you know frequencies which are special um but the idea is you can end up with density waves that wind up that are caused by these sort of overlap modes making patterns that change over time um but others can can probably explain it better than me and there's other ideas for how uh isolated spirals can form in in galaxies and there's a lot of galaxies in the universe so probably if it's physically plausible it's happening in a galaxy somewhere um so my summary sort of of the introduction here what we know about spiral arms and what some other questions are that I'm interested in you know we know they're density waves they're probably not static density waves exactly they may be winding up over time there is some evidence in the simulations for winding and I'll show you some observational evidence which might hint towards winding going on um there's probably a link to tides and maybe bars at least for some spirals but there are spirals in isolated galaxies that don't have bars um and I think you know in in observational extract astronomy the big question really is does it matter um do they actually have an impact on Galaxies do should we should we care about it in terms of the ution of the Galaxy um or are they just kind of a detail that's nice to look at um so um I've worked with the sonal sky survey for a really long time not 23 years to be fair um but a long time um and so uh I don't probably don't need to tell people near Princeton too much about the snow digital Sky survey but just a place in context that I'm working with these kind of big surveys where we have data for hundreds of thousands or in some cases even millions of galaxies um but spal are really complicated things um they're not a simple thing um to measure from um a lot of images of galaxies um so I don't know if you've ever seen Marshalls versus the machine um but it's a a cute uh kids movie where they defeat the machines by confusing them um because uh machine image uh recognition algorithms can get confused by similarl looking images um that are actually very different and obviously different to people um so the puppy versus bagel is another version of that you can find other examples on the internet um so I've been involved for a long time in leveraging the fact that people are really good at image recognition um if we can collect lots of people's opinions of what things look like um in some kind of statistical way we can use that for science um and we'll come back to machines getting involved uh later as well um so this is Galaxy Zoo which is a crowdsourcing project that's been running now for uh 16 years it's a a very very well established project um we put images of galaxies uh into a website and we invite people to classify what they see in their images of galaxies we try to keep the questions quite simple so it's not interpreting what you see it's just saying what you see um and we ask a lot of people per Galaxy so that we can get good statistics on the answers and we've done this for a lot of surveys now so the SSS surveys um all public HST surveys ukids is a British uh infrared survey which we have yet to publish that one but it's we have done the classifications um a lot of the images that have been on the site recently are from deal uh the dark energy camera Legacy surveys uh and Mike Wy has led a lot of that work um and this is where the machines are starting to get involved and be in the loop to help us accelerate what we can do with Galaxy Zoo um and we've also classified some simulated galaxies uh in Galaxy Zoo um to see how the population that the simulated that the simulations make compared to the population uh that we see in the real Universe oh yeah and right now we have some JCS images on the site and we're hoping to do more of the public or the to be public JBC Imaging surveys in the future um so the idea behind Galaxy zo was pretty simple um and this is before I became involved I started working with Galaxy zo about a year after it launched um it's really easy to tell people tell a person the difference between a spiral galaxy and an elliptical galaxy um you just tell them that and then you ask them to go classify them and you collect um you collect statistics on what what they answer this worked really really well and was really really popular um so it's a BBC News article from 2007 about seeking uh help looking at galaxies um became the which one is it second most emailed uh story on the BBC website for a time um the classifications per hour really took off um so uh after a few hours uh I think we think when the east coast of the US started to wake up and read the news it really started to take off um we were up to 40,000 classifications per hour in the first day of Galaxy Zoo but it crashed the scss servers um which again I wasn't involved at this point um so these days uh the images are hosted on a Cloud Server which can um react if there's some event that drives a lot of people to the site um so just highlighting the hope this is from the article The Hope is about 30,000 people might participate um and we like to thank our Volunteers in papers um so the first paper out of Galaxy Zoo had 160,000 volunteers um this is 200,000 volunteers another 85,000 volunteers we do count who did which phase so we can like have good statistics so this is not you know the sum overlap but so a lot of people are doing this um and and it it continues to be of interest today um we don't collect 40,000 classifications per hour anymore but we still get about 15,000 a day um so we can still classify pretty quickly sets of images and we like to collect 40 classifications per image something on that order so we have good statistics um and one of the reasons for wanting that large number is that uh we we now ask somewhat more complicated questions we don't just ask is this a spiral galaxy or an elliptical people turn out to be really good at that Division and so we ask more complicated things and this is where we start to get back to spirals um so this an example classification for this galaxy um if you care it has its classical morphology down here um all the different things um so you know the first question is is there uh do you see a disc or a feature in this galaxy and you have to say yes or no you'll notice that we say that 39.0 one9 people classified this galaxy so that's obviously not possible um and that reminds me to tell you about the post-processing we do so if someone comes in and says haha I'm going to mess with Galaxy zo and I'm going to do it all wrong um and I'm going to say that school teachers do use this in the classroom so probably does happen right um so we we actually do go through um with an algorithm and check if someone consistently disagrees with everyone else we downweight their contributions so that they don't um don't contribute so that's why we have non-intent numbers of people um and so then you know of these 39ish people 38 said they saw this s a feature so there's always one right but that's okay um if we have 38 out of 39 saying you see disc feature you could be pretty sure that you've got a disc feature um and if this split is closer the 50/50 you know that's hard to classify so it's not zero information if it's 50/50 right so we're starting to move to having a probability like number that goes from zero to one that tells you the whether or not the feature is present in the galaxy and probability like numbers are great for doing uh data science type things at the other end of the process right um anyway so we've got our 38 people saying they saw a feature of dis um they then asked if it's fa on again 37 of them agreed that this galaxy was was a face on dis um and then they're asked about bu spiral arms the the windiness of the arms how many arms there are about the size of the Bulge um and other things uh we actually have other more complicated questions that you can answer later um so as you can see you know when you start with 40 people we we come down to having 11 people say that this has two arms confidently um so there's a little bit of like dealing with the fact that it's a tree and you have to think about how many people out of how many people answered a specific question Okay so we've done a lot of science with this um you know we're a pretty small science team actually uh it's a lot of volunteers and a really big footprint um in the sort of Popular Science World but in terms of the the scientists working behind the site we've always been pretty modest in scale uh just a handful of people um so I'm pretty proud of the fact that over uh around about a decade we have 77 official team Publications um we're very open collaboration so if you have something you're interested in doing with Galaxy morphology uh please reach out reach out and we'd love to talk with you um we have a number of papers with a lot of citations which I'm really proud about um a lot of these are the data release papers which make sense a lot of people are using uh the Galaxy zo catalog to do their own science which is fantastic um but we also have papers about the um extra Galactic work um in particular a number of papers about the correlation between color and morphology we talked at the beginning about red ellipticals and blue spirals um but how well does that work in large populations finding interesting and odd things so computers are great at answering what you ask them to tell you um humans are curious and will sometimes tell you things you never asked them or never thought to ask them yeah to do color and morphology do you have them do some calibration step for the color of their monitor or oh we get the color in a in a the standard astrophysical way so the color is from apertures and photometry we don't ask people about color we're only asking people about morphology yeah um yeah and then we can find uh samples by morphology or we can look at the morphology of specific examples of galaxies and so some examples of that okay so back to Spirals um so reminder I I already showed this slide right about the classic spiral sequence and this implicit uh statement that um how tightly wound the spirals are should correlate with bold size so we asked those two questions separately in Galaxy Zoo so it seemed interesting to check um uh we didn't exactly ask about pitch angle we asked how tightly uh wound are the spiral arms appear um but we've done a number of checks um of how well this correlates with pitch angles measured in in more traditional ways um it turns out to be really really hard to measure pitch angles for a variety of reasons um but this does seem to work reasonably well probably as well as other methods um so this was our result um which we published in 2019 um so we plot on the y- axis uh our measure of pitch angle at some level um It's a combination of different answers from Galaxy Zoo such that Loosely wound spirals are down here and tightly wound are up here and managed to get a single number from zero to one um and then this is is also a bulge size score coming out of Galaxy Zoo so small bulges are close to zero and large bulges are close to one it's not exactly bulge the total um in the traditional way um there's just a random diagonal line plotted to guide the eye I guess um and then the red and the blue lines are the averages b a GES in the two directions um so we were surprised to see basically no correlation um and this is a volume limited sample of almost 5,000 spirals um that seemed very contrary to you know what we had learned about the spiral arm sequence so that's really interesting right and you want to dig into it a bit more one thing we did uh was separate it into whether or not the galaxis had bars um so I think this is just interesting slide um while there is no correlation uh evident in either which way around is it the unbarred side or the very strongly barred galaxies and there's a population in the middle that have ambiguous maybe weak bars um there's no correlation in either but there is a shift right so the unbarred galaxies tend to be more tightly wound at the same bulge size uh than the B galaxies um and uh we did this here this is also some nice work by Russ Hart who was a PhD student in Nottingham um did more traditional pitch angles as a function of uh the bar likelihood so strong bars over here no bars over here you can see pitch angle growing uh with bar probability um so there does seem to be a connection between bars the presence of bars and the pitch angle of spirals which is interesting um it's always more complicated with uh galaxies oh and this plot doesn't show up very well my apologies but you'll have to go read Petra's paper so Petra is an undergraduate haford working with me she's a senior this year so she'll be applying to phds this year um and what we did was the same thing that i' done in 2019 but splitting the spiral galaxy population into red and blue spirals um just using Color Optical color and massive and low mass spirals um and so basically uh if you have massive spirals either red or blue they have the opposite correlation to what you can read about in the Hubble sequence um if you go for low mass red spirals they have no correlation at all um if you go to low mass blue spirals the kind that are really visible in the local Universe in photographic plates for example of the kind that Hubble proberbly looked at um you do see a correlation so that kind of makes sense um you do see the expected correlation it's not very strong but it is there um what was interesting in this also is that blue spirals uh tend to have more loose arms which we'll come back to you um and there's also a mass dependence massive spirals tend to have tighter arms in the population um I promised uh I mentioned of this project uh so this was a PhD student I work with in Portsmouth uh uh who graduated in 2020 um this was a side project in Galaxy Zoo Tim now works uh as a data analyst um she's worked for a number of companies he's a very very talented data scientist um so he ran this project where we actually have did in browser fitting of models so the citizen scientist tweaked um a multi component uh model of a galaxy to fit uh the light um and the thing that we could do that's really hard to do in an automated fashion is try to fit the spirals um so Tim had a number of forget how big a sample is but he had a number of measurements or indicators of the Spiral pitch angle um one thing um that we Quantified and other people have done this as well but I think it was nice to see it done in this sample is how the pitch angle can vary quite a lot uh between arms in a single Galaxy so it's not necessarily a single number for a Galaxy um so this was that plot hello more what we're actually looking at this is one object that we're looking at it's multiple objects um and then the black lines are multiple people's uh indication of where the spiral is um and then the red is some optimal clustering of those lines that is constrained to be a log spiral actually there was no there was not enough statistics to suggest that it should be anything other than a log spiral so we went with log spirals as the simplest mathematical model okay do you Rectify attempt to rectify these images before you present so by rectifier you have to deproject them so make them be face on um which is definitely something which can introduce error um right and so the interarm variability that we found so so by that I mean the difference in pitch angle between the three red lines here um averaged at 11° um and so that's that uh this is the difference between the average for the Galaxy and then each individual arm as a function of the average for the Galaxy so the scatter here is about 11 degrees um Tim also looked at Trends with bulge size and agreed with our more simple analysis there was no Trend in this sample of pitch angle with bulge size um and this is bold strength from the model that we fit so from photometric decomposition um he also found no Trend with bar the presence of a bar um so that was disagreeing with our earlier work so that's interesting um but some interesting statistics um and one thing he also did was look at a comparison with the expected distribution of pitch angles so coent of of pitch angle um from this model uh of it's not really a model but it's a suggestion um from Pringle and dubs 2019 that if you have a population of spirals that are winding you should end up with a distribution of pitch angles that is uniform in the coent of a pitch angle and you'll have to quiz me later on the details of why that works uh because I'll have to remind myself uh but what we found was the distribution of pitch angles in the sample of galaxies matched that prediction pretty well so it was consistent with being uniform in the cotangents of pitch angle so plausibly evidence for winding okay shifting gear a little bit let talk about star formation um so spirals or density where waves in a galaxy um so there's two ways that should impact star formation or that could impact star formation um you may well have heard of the shenica relation right where you have more dense gas you have more star formation that's a really clear correlation between the density of gas and star formation so if you have an overd density in gas you should have more star formation in that over density the other thing that can happen um and that uh was I think first talked about in 1969 the density wave itself as it passes through um a gas cloud can create shocks and cause station to happen um so this this beautiful illustration of this happening in this uh NGC 5247 Grand Design spiral where the different like x's and dots and things are tracing out the spiral in different uh bands of light um and this is not work I've been involved in this is you you and ho uh Louis Ho's group uh work from L's host group um and so you can see that as you go from Far UV through near near UV blue as you get redder inside corotation which is the circle the gradient goes in One Direction and outside it goes in the other direction so this is a signature of this density wave passing through and generating star star formation and then as the Stars age they're progressively behind the where the spiral passes through the stars or if the stars are passing through the spiral the trend flips um and that flip will happen at PR rotation inside do we were talking about that earlier and should have looked that up but I don't remember probably is my guess yeah all right so what do we know about this from Galaxy zoo and the work that I've been involved in um so this is one nice result um looking at how spiral arm number correlates with Galaxy color so each of these little diagrams is a color Mass diagram so it's a little bit like mass star formation rate um which is a more like theoretical way of putting it but we actually observe color or something the that correlates with star formation um so up is red down is blue which means up is less star forming down is more star forming and we have you know normal masses um so all galaxies are the like grayscale which is the same in every one and then we're overplotting a contour of where you get spirals that have certain numbers of arms and you can probably see that they get Bluer as you get more arms so m 1 are kind of weird m equal 2 most spirals are m equals 2 like this is by far the largest population so this is most of the spirals um m equals 3 are much Bluer four five much Bluer so clearly a correlation with the color of the Galaxy and the number of arms although when Ras dug into this a lot more in his beautiful PhD work um it actually turned out to not really be a difference in Star formation rates it was more about dust obscuration and how much of the star formation was obscured a really uh neat piece of work that's just from photometry uh Chris mentioned I worked with uh SSS in the manga survey which is spectroscopy a lot so we're just going to have a little quick reminder of what we can learn about galaxies from spectroscopy from Optical spectroscopy um and then I'll tell you about some manga work that we're we're doing right now um so remember the spectrum of a galaxy is the sum of the Spectra of all of the stars in the galaxy so if you're looking at a Galaxy that doesn't have any ative Star formation going on you're just looking at light from Stars uh predominantly um so you have the Continuum is just the sum of the continua um and then you end up with absorption features um from the atmospheres of the stars um and so that's how you get your metallicities and your red shifts um there's no emission lines because there's no ionized gas in this galaxy because there's no star formation in this galaxy to first order um the other thing to notice is this uh break at 4,000 angstroms this is caused by a lot of different metal lines um in cooler stars and so it gets bigger and bigger the sort of jump you get at 4,000 anoms gets bigger and bigger as the Stell population ages and you lose the hot massive stars and you start to see more of the light from the low mass stars that live for a really long time let's switch to a star forming Galaxy instead so immediately you can see the Continuum is Bluer because it's actively star forming it still has O and B stars that are bright and dominate the light um you have a weak break at 4,000 angstroms you can still see the absorption features um but you also see these very uh distinctive emission features that are coming from the ionized gas recombination lines from ionized gas around star condition regions so really distinctly different and we're learning a lot about what kind of stars are in the galaxy from the Spectra um so manga uh is a spectral mapping survey uh so if you work with Spectra of galaxies most people uh the sort of the biggest sample is the sdss sample which put a single fiber on each Galaxy and for very nearby galaxies that single fiber is very much in the center of the Galaxy uh manga instead the biggest array had 127 of those fibers in a hexagonal array and so you get a Spectra at every point across the Galaxy okay uh so this turns out to be pretty complicated data um we call these data cubes uh so you get basically an image of the Galaxy at each spectral resolution element or you can do some analysis on the Spectra and generate maps of properties that come out of the Spectra um so here are some example maps of properties you can get out of a manga data Cube um so maps of the uh Hal Alpha emission that that emission line of hydrogen um maps of nitrogen emission line an oxygen emission line silicon emission line um the doler shift is really helpful in astronomy right so we also get U information about how the gas and the Stars rotate separately um and you can um and the random motion of stars you can also fit models to get information about the Stellar population so you fit uh uh Stellar population models and get Stellar masses metalis and so on lots and lots of information for lots and lots of galaxies um so in total manga did 10,000 galaxies um this is them on a color magnitude diagram and a star formation rate Stellar Mass um it's not it's not a complete sample of the local Universe it is a representative sample so we oversampled rare galaxies um but it's embedded in the bigger scss sample so there's a weights you can use to rec recreate a volume limited sample this is not 10,000 manga galaxies it's some subset of them but this is uh a cute uh little diagram that Carl westall made uh making the name of the survey in the G gri images and I'll flick through some other other properties from mango in the same order uh maybe oh I went to two there we go um this is the H Alpha flux of the same one so you can see where things are have a lot of ionized hydrogen and where things don't uh that d4000 measure which I said correlated with Stellar age pretty well so you can see the old stars and the young stars or galaxies with old Stars got with young Stars Stell velocity uh lots of things um so manga's been really really great and one of the big things that's come out of manga is lots of papers looking at gradients of properties from the Spectra um so to get a gradient in manga you put these elliptical isophotes over your Galaxy and then the average the measurement uh in in each of these elliptical rings and then you just plot the plot that uh as a function of radius this is actually not each of these lines is in a single manga Galaxy each of these lines is the average of a bunch of manga galaxies at the same mass um this is really nice work that shows how gas phase metallicity measured in the manga Spectra how the gradient of it uh is different for the Massa galaxies the red lines um and you get very shallow gradients for low mass galaxies you can also uh you know fit straight lines to this and then plot gradients uh of uh properties against something else so that's what this work is this is the gradient of DN 4000 so the gradient kind of like the rate of change of the age of the stars as a function of radius against what the central value is um some very interesting work so we thought it'd be interesting to add in some morphology here and this is some work that uh my student Emy Weiss has been doing um she graduated from bimar last year bimar and havord a this by Co agreement so I supervise a lot of bimar students for senior thesis work em is also applying for phds this year she's working at uh the Mariah Mitchell Observatory this year um and so we thought it'd be meet to do those metalis gradients uh as a function of spiral pitch angle and see if there's any impact the idea kind of came out of work some work by K Daniel who uh just recently moved from binmar to Arizona she's a dynamist um has some predictions that more Loosely wound spirals should drive more radial radial migration so the idea here is you know your Galaxy gets set up with a radial gradient um because you've got more star formation in the center so more metalis more metal rich in the center less metal rich in the outskirt and any kind of radial migration should shallow that gradient that's the idea um so we thought we'd look for that um and so that's uh the result is here it's a little bit complicated but um the which way around is it color is the arm winding um so if you just look at the dotted lines this is our lower Mass sample we split it by mass because there's clearly a mass dependence of these gradients and in the lower Mass sample whether they were tightly wound arms the green line uh medium or Loosely wound the pink really separated out these metalis gradients which I don't really know how to interpret yet but it's there in the data so I thought it was interesting um in the more massive spiral samples we didn't see any impact from spiral winding um but I think this is really interesting work and EM is currently working on writing this up for publication so U we didn't just do spiral winding we actually looked at other morphologies on the impact of them on these gradients but this was the most interesting effect um so we could do this radial averaging and just come up with gradients but there's actually a lot of really beautiful complexity in the manga data so again some maps from manga um compared to their images and you can see spiral structure popping out um in the maps of properties even this d4000 has some hints of spiral structure in it um from the age of the Stars I'm almost done I promise um uh so we wanted to go beyond gradients um and to do that we needed to be able to pick out the Spectra from manga in specific regions of the Galaxy so that would be an interesting crowdsourcing project and so we ran this galaxy Zoo 3D site 3D because of the data cubes from anganga um where I asked people to look at images of galaxies and draw on them where they saw certain features um so this ran um and we have uh done the analysis to get the results um and we end up with these sort of heat Maps which show you where the bar is in Orange where the spirals are in blue we also asked for the centers of galaxies and foreground stars because the foreground Stars mess up um a bunch of the pipeline um and this is all available we published it a few years ago and you can use it with the manga data to do things like this so this is uh the flux in h Alpha as a function of azoth around the galaxy in the spirals the blue so you can see the up and down of the spirals and in the bar in the orange um and all kinds of other things with it um the spiral masks are a little bit tricky to work with um I think I'll skip over this a little bit in the interest of time you have to pick a threshold level where you believe that it's a spiral uh but one thing thing we did do um again an undergraduate project uh looking at the amount of star formation there is in spirals compared to the star formation in the whole disc so we were able to show from this measurement that there's an excess of star formation in the spirals and actually an excess which grows with radius um so this would be no uh enhancement of star formation in the arms and you can see as a function of radius you get a larger and larger fraction of the star formation by area is in the spiral arms um and this is uh Mara kilada who's a sophomore at haford really uh really really fantastic student has been using these masks uh just this past summer to look at the mass excess uh in the spiral arms so we use a model for the stellam mass surface density fit to the Spectra which is uh published in uh uh Sanchez edl 2021 this pipe 3D model apply the spiral masks so you can't really see it too well but there's like a little faint blue line that is uh around the spiral arms in this galaxy that was our arm measurement we have a little white bit where we sort of just throw out the data because it's ambiguous and then we have an interarm measurement that's not surrounded um and so what Masa showed is that in both of these regions the stellamar surface density drops exponentially with radius but it drops more quickly in the interarm regions than the arm regions so you have this excess uh surface excess Stellar Mass surface density which grows with r um it goes from about 15% in the inner parts to 30% in the outer parts I think this is really cool yes photometric model fitting that is you're saying that the effective radius in the arms and interarm regions are different yeah that's very cool yeah interesting right yeah and so you're right this effective radius is the whole thing um but clearly they have different effective yeah so we we still have some like data cleaning to do before we'll publish this um make sure that the you know the masks are really good and that we have the right sample of galaxies and think a bit about the observational biases like the these are the spirals we can see which worries me a bit about this measurement but I think it's a really neat measurement and I hope we'll publish it soon yeah so that is uh the end of what I wanted to talk about with spiral I've definitely not covered everything we know observationally about spirals but just a flavor of things I've worked on you just end with the questions that we started the observational section with um you know there is again some evidence for this winding um there's a puzzle about isolated galaxies maybe or maybe not I don't know um and there definitely seems to be a link between the arms and star formation and Stellar uh Stellar Mass surface density excess um that we can measure really nicely with the manga data um I'll skip over I had a few backup slides which I will Skip and just end with thanks and some pictures of my research group [Applause] thanks would you care to comment on inclined Bridges between spiral arms of of stars and presumably gas in their mechanism of formation inclined ridges between adjacent spirals there are little bridges of comment on that I I don't know anything about that I've definitely uh I see I know what you mean but I I don't know abely what do they call them feathers kind of [Music] right data from J large arms these sub arms coming off for that yeah yeah I'm definitely working in a like Rift regime where those are not really clear um the Fang data is just beautiful for that kind of thing obviously a much smaller sample though much more nearby and bars can do is transport angular momentum so has anybody looked at correlating either AG activity or nuclear star formation with r yes yes yeah they're very much so there's definitely evidence that galaxies bars have enhanced nuclear star formation um and there's a lot of ongoing debate in the uh observational Community about whether there's a correlation between bars and AG just spirals no bars I mean that's a hard question to answer because the G the this galaxy that don't have spirals are just super different to the ones that do have spirals maybe that's also the case for spars now I'm saying it out loud but I mean there's you know there definitely you can find a sample of disc galaxies with Spirals and no bar you can definitely do that but your sample of disc galaxies that doesn't have spirals they're kind of weird right so um they're more like lenticulars and things like that um and then there's a big Mass difference as well so I think it would be hard to answer that question and it's the mass difference that's causing the debate I think Ag and bars where bars are more likely in more massive galaxies AG are more likely in more massive galaxies so if you just do a simple correlation yes there are more AG in galaxies that have more bars but if you do it more carefully controlling for Mass um it's less clear but maybe a little bit of a think yeah so you just mentioned passing lentic or szer what's your what are your thoughts of how they fit in people have been arguing this since since hble I think about how they fit into this do you wna you want to speculate or what what what is the current uh consensus if there is one yeah there's a really really nice paper by Amelia frasa Meli on this um she was a post talk who worked within manga um and um that maybe there so yeah big fan of morphology and I do think you can get physics out of morphology and like you know there are morphologies where you can get unique uh sets of galaxies lentic might not be that right you might end up actually with more than one way to make a lenticular and so Amelia's work actually was looking at differences between lenticulars and there are some lenticulars that may well be faded spirals so just spirals that I mean the slides I skipped were about red spirals red spirals are interesting and not that rare and shouldn't last for a long time so these are spirals that have no current active star formation um they should fade over time and then maybe they form a lenticular so that's just one route right but certainly not all lenticulars can be this so um there's also the interesting thing that a lot of elliptical galaxies show significant rotation so yeah it's a diss all the way down fact that we see so many of these things if they supposedly so shortly I don't think there's any tension I think it's just it's just a common evolutionary route yeah we don't think so it's a good question any final questions uh question about Galaxy Zoo um who is in charge of coming up with questions and follow up to that is if someone were to have kind of an idea for a type of question that is more specific is there a process for sort of proposing it that's a great question um we are very collaborative uh so we have uh bi-weekly telecons and when we have new sets of images we discuss what the good classification set would be uh increasingly our new sets of images are coming via some kind of agreement with the team that make the images and so that team have specific things they want to have uh in the classification tree um we also we have a number of like slide projects going so like quick morphology questions particularly like either or there's a a mobile app that is really good at like really quick either or questions so if for example you were just searching for shells in the of galaxies um you could be like do you see shells here yes no and that can be done really quickly in the mobile app um we've always been a fairly informal collaboration so as I said if you're interested in collaborating reach out and uh would like to talk so find the arms and the arm number by asking observers asking your scientist to trace the arms then taking some sort of mean um how well does that agree with what you would just get from forier analysis because which is of course closer to the dynamical theory so for sure um example you say most are really n equals to what does that mean in terms of the power I've been meing some I have never I have not measured that and I don't think I've seen that comparison done so directly but I've been meaning to do that because it's not that hard to do the fur analysis at some level right so um yeah and I always wonder you know this whole thing about the Milky Way does it have two arms or fourarms and that very much depends on the band you look at the kind of traces it's a good question and it's something we should check we the the arm number is not from tracing it's just from asking people what they see do you you know how many arms do you see in this galaxy what's interesting is that isn't necessarily a single answer to that either I said there's not necessarily a single pitch angle but the number of arms can change with radius so you can have really clear two arms in the center and then it's just a big mess in the outskirts so what you answer in that case I don't know but yeah yes so um you mention uh there are also some robotic uh ways to measure the things could you say more about that and also I mean now recently the development of the machine learning techniques and I wonder how do you think that will be useful for these spiral study yeah so this is what we have going right now in uh Galaxy Zoo so the crowdsourcing creates a training set which is fed to zubot is the name of our robot uh which is very good at classifying uh the normal stuff where we have a good training set um and then it's in the loop so stuff that doesn't have a high probability of a good classification gets fed by back to the slite um so this is working right now um this is Mike wy's work so Chris lart Mike was a a Oxford PhD student he worked at Manchester for a while he's now in Toronto so come over to this side of the Atlantic which is cool um and uh yeah so I refer you to to Mike's uh work on the machine learning in the loop at Galaxy Zoo um Mike's also been working some on image segmentation so using our our our masks from Galaxy zo 3D to try to train image segmentation um this stuff is getting better but generally will'll always need some kind of training setep so that's the sort of Niche we've been discussing in Galaxy zoo of working with it in an integrated way to get the training set a lot of the Galaxy Zoo classification uh not citations to Galaxy Zoo a lot of those are actually people using it to train a machine learn Sor state of these uh machine learn a new network based machine learning is that b still not as good as human being I mean that's it starts to get hard to quantify right that they're they can agree with humans to sort of 90% but if you set a population of humans to classify galaxies they will also agree with each other to 80 90% it's a really cool paper from the 90s where someone did that with a bunch of Galaxy classiers what about the more traditional I mean statistical way do the image process in TR like is there any new algorithm now to measure like without the human guidance just just to do some Statics yeah the P jungle or that identify spiral arms or something like that this is it is that progress on that we have to rely on human being still like to identify spiral for example it's a yeah I mean wouldn't it be it would be lovely right to be able to just feed the images to a computer and and out you get really reliable pitch angles I suspect it's going to be tricky for a long time um but and I think you know I do end up reading a bunch of machine learning papers actually was reading one not to do even with galaxies the other day for reasons to do with being a professor at a small liberal arts college I guess um and even that computer science professor like buried in the paper was a bit where they had people look at the class this the the sets that they classified which I was sort of like oh okay so there's still people looking at stuff in this computer science paper to figure out what's coming out of the machine learner um and I think that's one of the the sort of hidden secrets of all of these algorithms is the amount of human effort that went into training them um so I don't know I I'm a bit of a skeptic uh partly because spirals are so patchy and weird looking and we were talking earlier right what even is the pitch angle especially if you have these uh things where the gradient of the color and therefore maybe where the brightest light is maybe switches side from the inner part to the outer part of the spir how does that impact the pitch angle um the pitch angles are not the same for every spiral in a galaxy or not the same with function of radius is there even a single number for a pitch angle for a Galaxy so I think they are somewhat complicated objects that I mean it's like lots of places in astronomy probably all science right once you have a really detailed look at an object it's much more complicated than you thought Okay one minute any last question Matt I'm I'm curious about the human element here because humans are good at pattern recognition but certain patterns more than others so I mean if if you just showed them if you took a Galaxy and plot it in log R and angle space do they see do people see more than you just than if you just put it in in XY and there's thing you can do to like unsharp masking and pull out features right we've not tried like weird projections there is I mean you're just giving them the the raw image well we give them a processed image we usually give them a three-c color image with a nice scaling so that features I mean there's a there's a sort of art to picking the right three color scaling and the image depth that makes a nice image I mean as you mentioned it's people so we have to get them to actually engage with the site so there is a little bit of like the images it helps if the images are somewhat attractive although which is easy with spases but there are spin you know the Galaxy Zoo works so well it inspired this whole zooniverse of Citizen science projects and some of the projects in that are the data is less attractive um but if you Mo motivate the question appr appropriately and say you know we want to find exoplanets or something uh people still are engaging with it so maybe we could try uh different projections of galaxies if we think it's really necessary but computers are pretty good at picking out straight lines right so if you think you got log spirals in a in the right projection you have straight lines I don't know cor right there are a couple of spots left for dinner if anyone wants to sign up otherwise it's still to lunch and

okay let&#39;s get started thank you everybody for coming I&#39;m delighted this today uh to introduce our colloquium speaker Professor Karen Masters who has come to us from haford college in Pennsylvania um Professor Masters was previously of the institute for Gra on cosmology in Portsmouth and before that Harvard and Cornell in Oxford uh she&#39;s an expert on exactic extragalactic astronomy in particular working with a bunch of large surveys and large collaborations so is the I&#39;m G to get the wrong way around the pi for Galaxy Zoo yes and uh the spokesperson for sss4 obviously I&#39;m never going to be the spokesperson given that yeah for sss4 so she&#39;s working a lot with these manga Galaxy SL digital Cisco survey citizen science and we&#39;re going to hear about some of that today and she&#39;s going to tell us about spirals in galaxies so please thank you Chris for a very nice invitation uh or introduction and uh for the invitation to come and speak today um so I&#39;m going to talk about spirals in galaxies as Chris mentioned I have broad interests in Galaxy Evolution with a kind of focus on why morphology is interesting and what morphology can tell us from Big surveys um but in particular this talk uh came out of some work I did on my last sabatical which was a CO sabatical so I spent it in my basement um spending a lot of time uh well it&#39;s okay it was fine uh talking uh with Jerry salwood about spiral structure and and we wrote on this review article um so I&#39;m really going to focus on observations of spirals um I&#39;m well aware that there are some people in the audience who spend a lot of time thinking about the theory of SP spirals but of course U maybe not everyone does that so I do have some introduction to why spirals are interesting and some of the background about what we know about Spirals and then I&#39;ll kind of focus on observations I&#39;ve been involved with with the projects I work on um so it is a little bit uh biased uh towards the work myself and my students have done but um hopefully you&#39;ll forgive me for that so spiral uh have been observed in Galaxies for almost around 200 years I would say I just always enjoy starting with this uh drawing ing of the wellp pool Galaxy done by an asomo from Ireland uh in the uh mid of mid mid part of the 1800s um comparing it it&#39;s the worldo Galaxy right so this is the Hubble image of the same galaxy um this is the first recorded uh observation we have of spiral structure in galaxies um and people are obviously very struck by this uh so there are suggestions that it was the inspiration for the spiral in the middle of Spar night um so these are beautiful objects and and fascinating objects with some really interesting dynamical physics if you&#39;ve taken an astronomy class or taught sort of Astro 101 you&#39;ve probably taught about the classification of galaxies using the Hubble tuning fork um good reminder um of the kind of big picture here of galaxies uh Edwin Hubble looked at uh photographic plate images of around a thousand galaxies and put together this classification scheme in the 1920s uh where we have elliptical galaxies so these smooth reddish the more massive galaxies on one side um and spirals on the other side with this Fork set by whether or not they have a galactic bar uh this this linear structure across the middle um the diagram here is Hubble&#39;s sketch um and the images are his example images but scss versions of them um so one other note on terminology I already mentioned elliptical tend to be red spirals tend to be blue you can see that actually in the scss images that I&#39;ve put here um we also tend to use early type for things at this side of the diagram and late type for that side and one of my pet peeves is people then go on to say and Hubble thought this was an evolutionary sequence it&#39;s not true can find a footnote in one of his papers about how that was not true um and he was just using an analogy to late and early type stars to provide some names he never thought it was a sequence in time random aside to my pet peeve uh so you&#39;ve probably either taught this or seen it in a class uh um since we&#39;re talking about spirals a little bit more detail about the spiral part of the sequence um so spirals are ordered from an sa an early type spiral to an sc or a late type spiral um based on uh two two or three different properties in Hubble&#39;s classification um and so sort of implicitly we&#39;re saying these properties correlate with each other right um so SAS uh in in Hubble&#39;s language have large bulges tightly wound spirals which are very clear and distinct uh whereas you move to sc&#39;s where you have small bulges Loosely wound arms um and the spirals tend to be patchier um another interesting side note so as as recently as 10 years ago Rona one of the real experts uh in traditional morphology is talking about how if there&#39;s a conflict here so if you have something with a bit a large bulge and loose arms or a small bulge and tight arms um you should classify it based on the arm winding um that is just not what&#39;s done these days in big surveys um this distinction between early type and late type well is almost entirely on bulge size uh in big surveys so just to note that those two things are different okay there&#39;s more complicated classification schemes for spiral so dealore in the 1950s had this sort of wheel thing where um we put Spirals and Rings um on a classification scheme uh this doesn&#39;t work very well sorry uh one of my favorite words in astronomy flocculent uh is uh meaning literally fleece or cloud-like these sort of very patchy spirals um as opposed to Grand Design the very large scale stretching across large radial range that typically symmetric um this was a classification or a separation developed by the eligens in the 1980s um and you see spirals all over the place so there&#39;s something very interesting and like generic about spirals in dynamical systems made of numbers of particles um so just some examples of spiral um in a protoplanetary disc in the ofus star formation region you can make spiral in the cream in your coffee cup uh Saturn&#39;s rings actually are extremely tightly round spirals in places um and of course hurricanes spirals a little bit of a spiral math now um so in nature most of these spirals are something called a log spiral um so you have the log of the radius is proportional to the asmal angle um and the uh the angle that this log spiral makes to the tangent to the circle at the same radius is called the pitch angle um and in a log spiral that pitch angle is constant with radius so it&#39;s one number um so you know Nautilus shells mandle BR sets all kinds of things of versions of log spirals mathematically so again just a reminder large pitch angle is very open spirals so 90 degrees would be spokes like in a bicycle wheel um and a pitch angle of zero is ring so if you&#39;ve uh taken a galactic Dynamics class you&#39;ll heard of the winding problem um but just a reminder of what this is right galaxies rotate differentially very famously right they have flat rotation curves with radius so that&#39;s the flat circular velocity that means the inner part um does uh more rotations in the same amount of time than the outer part um so if spirals are material structures if they&#39;re like ribbons uh they W would wind up over time in a galaxy um so of course this was the original argument that they cannot be material they must be some kind of density wave a reminder of what a density wave is using an analogy uh with cars instead of stars um it&#39;s a really fun little video um that I a student of mine found once um where it&#39;s a circular well the cars are driving in a circle and you&#39;re going to see a little density wave get set up over here you can see right the density wave starts to move this this way actually but the stars are moving the cars are moving this way through it um so the idea is in a spiral galaxy the spiral arms are this kind of feature right the the stars are bunching up obviously they&#39;re not changing speed in the traffic jam like cars are um the density is very low right they&#39;re not interacting with each other um but the spirals are places in the Galaxy where the orbits of the stars are bringing them slightly closer together um so density waves um so this is another classic diagram talking about these sort of electrical orbits of stars in a galaxy um and you can start to orient them appropriately that that they end up looking like spirals um many people uh I think would agree that a large number of spirals have driven by uh hides interactions of galaxies um so the wellp pool galaxy is the Prototype of this with its companion um clearly connected to the spiral in some way um those also discussion of how the galactic bar might drive spirals in some cases it&#39;s definitely the case observationally the bar spirals much more likely to have two arms two spiral arms uh than other numbers of arms so there appears at least superficially to be a connection between bars and spirals um oh I find this one interesting and maybe everyone knows this but I didn&#39;t know this until I was working on the review um density waves should Trail which means the the spiral arm the sense of the arm is that the Galax is rotating this way right so the arm is sort of trailing behind the rotation um you can&#39;t have a density wave that goes the other way I think someone can correct me on that if I&#39;m wrong um but it&#39;s actually a really hard measurement to make um we&#39;re good at rotation in galaxies observationally the doler shift works really well we can get the rotation really easily the tricky thing is determining the orientation of the Galaxy what the front is and what the back is um this diagram makes it look not tricky at all because these are beautiful examples of galaxies where you can tell because of the Dust lames um and so these are tocal sketches of these galaxies where you can see like this is the front because the dust Lane cuts across the Bulge and you can see a little dust Lane here so this is the front um but it&#39;s not as if you look at just a general image of a galaxy it&#39;s not that easy to tell um and it&#39;s certainly not something you can automate very well um so we&#39;ve actually only done this in around about 100 galaxies um to determine observationally that they are actually trailing and almost all of hundred are trailing so there&#39;s no cause for concern um but I just think it&#39;s interesting that in today&#39;s sort of Big Data astronomy we only have this measurement for 100 galaxies yeah there are ones that are not or yeah and people sort of argue about them in the literature a little bit um why that might be but yeah there is at least evidence there&#39;s a couple of spirals where there appears to be some trailing and some leading arms which is very weird there&#39;s a lot of galaxies so there&#39;s always some weird ones yeah okay um but back to density waves um you know we said there might be some kind of effect of of environment driving spirals certainly there you know we know the galaxies interact a lot with their environment bars uh may have some kind of role to play in driving spirals in some galaxies but there are very isolated spirals that don&#39;t have bars um and they still have spirals so there must be a way to generate spirals fairly independent of you know other stuff going on um and one of the sort of interesting things that&#39;s happened in the last I guess 10 or more years is that in numerical simulations so this is a hydro uh simulation has stars and gas we&#39;re seeing the spirals actually are winding to some extent they are still density waves so they&#39;re not winding as fast as they would if they were material but um they are not uh they don&#39;t have a a constant patent speed with radius so they&#39;re not rotating at the constant patent speed they are winding up over time um so one way to generate that that that Jerry uh cwood has has worked a lot on these self- excited instabilities so overlapping modes uh in a galaxy um I&#39;m not particularly a dynamicist um so I like to kind of think of physical analogies what this what is going on here um so I come sometimes think about like wave interference patterns you get a very coherent pattern from multiple waves overlapping U perhaps a better analogy is this kind of uh beautiful patterns you can get from uh applying sound waves to metal plates so you can get all kinds of interesting patterns know in a in a spiral galaxy of course right it it does have an edge but the edge kind of Fades off it is rotating so there&#39;s some you know frequencies which are special um but the idea is you can end up with density waves that wind up that are caused by these sort of overlap modes making patterns that change over time um but others can can probably explain it better than me and there&#39;s other ideas for how uh isolated spirals can form in in galaxies and there&#39;s a lot of galaxies in the universe so probably if it&#39;s physically plausible it&#39;s happening in a galaxy somewhere um so my summary sort of of the introduction here what we know about spiral arms and what some other questions are that I&#39;m interested in you know we know they&#39;re density waves they&#39;re probably not static density waves exactly they may be winding up over time there is some evidence in the simulations for winding and I&#39;ll show you some observational evidence which might hint towards winding going on um there&#39;s probably a link to tides and maybe bars at least for some spirals but there are spirals in isolated galaxies that don&#39;t have bars um and I think you know in in observational extract astronomy the big question really is does it matter um do they actually have an impact on Galaxies do should we should we care about it in terms of the ution of the Galaxy um or are they just kind of a detail that&#39;s nice to look at um so um I&#39;ve worked with the sonal sky survey for a really long time not 23 years to be fair um but a long time um and so uh I don&#39;t probably don&#39;t need to tell people near Princeton too much about the snow digital Sky survey but just a place in context that I&#39;m working with these kind of big surveys where we have data for hundreds of thousands or in some cases even millions of galaxies um but spal are really complicated things um they&#39;re not a simple thing um to measure from um a lot of images of galaxies um so I don&#39;t know if you&#39;ve ever seen Marshalls versus the machine um but it&#39;s a a cute uh kids movie where they defeat the machines by confusing them um because uh machine image uh recognition algorithms can get confused by similarl looking images um that are actually very different and obviously different to people um so the puppy versus bagel is another version of that you can find other examples on the internet um so I&#39;ve been involved for a long time in leveraging the fact that people are really good at image recognition um if we can collect lots of people&#39;s opinions of what things look like um in some kind of statistical way we can use that for science um and we&#39;ll come back to machines getting involved uh later as well um so this is Galaxy Zoo which is a crowdsourcing project that&#39;s been running now for uh 16 years it&#39;s a a very very well established project um we put images of galaxies uh into a website and we invite people to classify what they see in their images of galaxies we try to keep the questions quite simple so it&#39;s not interpreting what you see it&#39;s just saying what you see um and we ask a lot of people per Galaxy so that we can get good statistics on the answers and we&#39;ve done this for a lot of surveys now so the SSS surveys um all public HST surveys ukids is a British uh infrared survey which we have yet to publish that one but it&#39;s we have done the classifications um a lot of the images that have been on the site recently are from deal uh the dark energy camera Legacy surveys uh and Mike Wy has led a lot of that work um and this is where the machines are starting to get involved and be in the loop to help us accelerate what we can do with Galaxy Zoo um and we&#39;ve also classified some simulated galaxies uh in Galaxy Zoo um to see how the population that the simulated that the simulations make compared to the population uh that we see in the real Universe oh yeah and right now we have some JCS images on the site and we&#39;re hoping to do more of the public or the to be public JBC Imaging surveys in the future um so the idea behind Galaxy zo was pretty simple um and this is before I became involved I started working with Galaxy zo about a year after it launched um it&#39;s really easy to tell people tell a person the difference between a spiral galaxy and an elliptical galaxy um you just tell them that and then you ask them to go classify them and you collect um you collect statistics on what what they answer this worked really really well and was really really popular um so it&#39;s a BBC News article from 2007 about seeking uh help looking at galaxies um became the which one is it second most emailed uh story on the BBC website for a time um the classifications per hour really took off um so uh after a few hours uh I think we think when the east coast of the US started to wake up and read the news it really started to take off um we were up to 40,000 classifications per hour in the first day of Galaxy Zoo but it crashed the scss servers um which again I wasn&#39;t involved at this point um so these days uh the images are hosted on a Cloud Server which can um react if there&#39;s some event that drives a lot of people to the site um so just highlighting the hope this is from the article The Hope is about 30,000 people might participate um and we like to thank our Volunteers in papers um so the first paper out of Galaxy Zoo had 160,000 volunteers um this is 200,000 volunteers another 85,000 volunteers we do count who did which phase so we can like have good statistics so this is not you know the sum overlap but so a lot of people are doing this um and and it it continues to be of interest today um we don&#39;t collect 40,000 classifications per hour anymore but we still get about 15,000 a day um so we can still classify pretty quickly sets of images and we like to collect 40 classifications per image something on that order so we have good statistics um and one of the reasons for wanting that large number is that uh we we now ask somewhat more complicated questions we don&#39;t just ask is this a spiral galaxy or an elliptical people turn out to be really good at that Division and so we ask more complicated things and this is where we start to get back to spirals um so this an example classification for this galaxy um if you care it has its classical morphology down here um all the different things um so you know the first question is is there uh do you see a disc or a feature in this galaxy and you have to say yes or no you&#39;ll notice that we say that 39.0 one9 people classified this galaxy so that&#39;s obviously not possible um and that reminds me to tell you about the post-processing we do so if someone comes in and says haha I&#39;m going to mess with Galaxy zo and I&#39;m going to do it all wrong um and I&#39;m going to say that school teachers do use this in the classroom so probably does happen right um so we we actually do go through um with an algorithm and check if someone consistently disagrees with everyone else we downweight their contributions so that they don&#39;t um don&#39;t contribute so that&#39;s why we have non-intent numbers of people um and so then you know of these 39ish people 38 said they saw this s a feature so there&#39;s always one right but that&#39;s okay um if we have 38 out of 39 saying you see disc feature you could be pretty sure that you&#39;ve got a disc feature um and if this split is closer the 50/50 you know that&#39;s hard to classify so it&#39;s not zero information if it&#39;s 50/50 right so we&#39;re starting to move to having a probability like number that goes from zero to one that tells you the whether or not the feature is present in the galaxy and probability like numbers are great for doing uh data science type things at the other end of the process right um anyway so we&#39;ve got our 38 people saying they saw a feature of dis um they then asked if it&#39;s fa on again 37 of them agreed that this galaxy was was a face on dis um and then they&#39;re asked about bu spiral arms the the windiness of the arms how many arms there are about the size of the Bulge um and other things uh we actually have other more complicated questions that you can answer later um so as you can see you know when you start with 40 people we we come down to having 11 people say that this has two arms confidently um so there&#39;s a little bit of like dealing with the fact that it&#39;s a tree and you have to think about how many people out of how many people answered a specific question Okay so we&#39;ve done a lot of science with this um you know we&#39;re a pretty small science team actually uh it&#39;s a lot of volunteers and a really big footprint um in the sort of Popular Science World but in terms of the the scientists working behind the site we&#39;ve always been pretty modest in scale uh just a handful of people um so I&#39;m pretty proud of the fact that over uh around about a decade we have 77 official team Publications um we&#39;re very open collaboration so if you have something you&#39;re interested in doing with Galaxy morphology uh please reach out reach out and we&#39;d love to talk with you um we have a number of papers with a lot of citations which I&#39;m really proud about um a lot of these are the data release papers which make sense a lot of people are using uh the Galaxy zo catalog to do their own science which is fantastic um but we also have papers about the um extra Galactic work um in particular a number of papers about the correlation between color and morphology we talked at the beginning about red ellipticals and blue spirals um but how well does that work in large populations finding interesting and odd things so computers are great at answering what you ask them to tell you um humans are curious and will sometimes tell you things you never asked them or never thought to ask them yeah to do color and morphology do you have them do some calibration step for the color of their monitor or oh we get the color in a in a the standard astrophysical way so the color is from apertures and photometry we don&#39;t ask people about color we&#39;re only asking people about morphology yeah um yeah and then we can find uh samples by morphology or we can look at the morphology of specific examples of galaxies and so some examples of that okay so back to Spirals um so reminder I I already showed this slide right about the classic spiral sequence and this implicit uh statement that um how tightly wound the spirals are should correlate with bold size so we asked those two questions separately in Galaxy Zoo so it seemed interesting to check um uh we didn&#39;t exactly ask about pitch angle we asked how tightly uh wound are the spiral arms appear um but we&#39;ve done a number of checks um of how well this correlates with pitch angles measured in in more traditional ways um it turns out to be really really hard to measure pitch angles for a variety of reasons um but this does seem to work reasonably well probably as well as other methods um so this was our result um which we published in 2019 um so we plot on the y- axis uh our measure of pitch angle at some level um It&#39;s a combination of different answers from Galaxy Zoo such that Loosely wound spirals are down here and tightly wound are up here and managed to get a single number from zero to one um and then this is is also a bulge size score coming out of Galaxy Zoo so small bulges are close to zero and large bulges are close to one it&#39;s not exactly bulge the total um in the traditional way um there&#39;s just a random diagonal line plotted to guide the eye I guess um and then the red and the blue lines are the averages b a GES in the two directions um so we were surprised to see basically no correlation um and this is a volume limited sample of almost 5,000 spirals um that seemed very contrary to you know what we had learned about the spiral arm sequence so that&#39;s really interesting right and you want to dig into it a bit more one thing we did uh was separate it into whether or not the galaxis had bars um so I think this is just interesting slide um while there is no correlation uh evident in either which way around is it the unbarred side or the very strongly barred galaxies and there&#39;s a population in the middle that have ambiguous maybe weak bars um there&#39;s no correlation in either but there is a shift right so the unbarred galaxies tend to be more tightly wound at the same bulge size uh than the B galaxies um and uh we did this here this is also some nice work by Russ Hart who was a PhD student in Nottingham um did more traditional pitch angles as a function of uh the bar likelihood so strong bars over here no bars over here you can see pitch angle growing uh with bar probability um so there does seem to be a connection between bars the presence of bars and the pitch angle of spirals which is interesting um it&#39;s always more complicated with uh galaxies oh and this plot doesn&#39;t show up very well my apologies but you&#39;ll have to go read Petra&#39;s paper so Petra is an undergraduate haford working with me she&#39;s a senior this year so she&#39;ll be applying to phds this year um and what we did was the same thing that i&#39; done in 2019 but splitting the spiral galaxy population into red and blue spirals um just using Color Optical color and massive and low mass spirals um and so basically uh if you have massive spirals either red or blue they have the opposite correlation to what you can read about in the Hubble sequence um if you go for low mass red spirals they have no correlation at all um if you go to low mass blue spirals the kind that are really visible in the local Universe in photographic plates for example of the kind that Hubble proberbly looked at um you do see a correlation so that kind of makes sense um you do see the expected correlation it&#39;s not very strong but it is there um what was interesting in this also is that blue spirals uh tend to have more loose arms which we&#39;ll come back to you um and there&#39;s also a mass dependence massive spirals tend to have tighter arms in the population um I promised uh I mentioned of this project uh so this was a PhD student I work with in Portsmouth uh uh who graduated in 2020 um this was a side project in Galaxy Zoo Tim now works uh as a data analyst um she&#39;s worked for a number of companies he&#39;s a very very talented data scientist um so he ran this project where we actually have did in browser fitting of models so the citizen scientist tweaked um a multi component uh model of a galaxy to fit uh the light um and the thing that we could do that&#39;s really hard to do in an automated fashion is try to fit the spirals um so Tim had a number of forget how big a sample is but he had a number of measurements or indicators of the Spiral pitch angle um one thing um that we Quantified and other people have done this as well but I think it was nice to see it done in this sample is how the pitch angle can vary quite a lot uh between arms in a single Galaxy so it&#39;s not necessarily a single number for a Galaxy um so this was that plot hello more what we&#39;re actually looking at this is one object that we&#39;re looking at it&#39;s multiple objects um and then the black lines are multiple people&#39;s uh indication of where the spiral is um and then the red is some optimal clustering of those lines that is constrained to be a log spiral actually there was no there was not enough statistics to suggest that it should be anything other than a log spiral so we went with log spirals as the simplest mathematical model okay do you Rectify attempt to rectify these images before you present so by rectifier you have to deproject them so make them be face on um which is definitely something which can introduce error um right and so the interarm variability that we found so so by that I mean the difference in pitch angle between the three red lines here um averaged at 11° um and so that&#39;s that uh this is the difference between the average for the Galaxy and then each individual arm as a function of the average for the Galaxy so the scatter here is about 11 degrees um Tim also looked at Trends with bulge size and agreed with our more simple analysis there was no Trend in this sample of pitch angle with bulge size um and this is bold strength from the model that we fit so from photometric decomposition um he also found no Trend with bar the presence of a bar um so that was disagreeing with our earlier work so that&#39;s interesting um but some interesting statistics um and one thing he also did was look at a comparison with the expected distribution of pitch angles so coent of of pitch angle um from this model uh of it&#39;s not really a model but it&#39;s a suggestion um from Pringle and dubs 2019 that if you have a population of spirals that are winding you should end up with a distribution of pitch angles that is uniform in the coent of a pitch angle and you&#39;ll have to quiz me later on the details of why that works uh because I&#39;ll have to remind myself uh but what we found was the distribution of pitch angles in the sample of galaxies matched that prediction pretty well so it was consistent with being uniform in the cotangents of pitch angle so plausibly evidence for winding okay shifting gear a little bit let talk about star formation um so spirals or density where waves in a galaxy um so there&#39;s two ways that should impact star formation or that could impact star formation um you may well have heard of the shenica relation right where you have more dense gas you have more star formation that&#39;s a really clear correlation between the density of gas and star formation so if you have an overd density in gas you should have more star formation in that over density the other thing that can happen um and that uh was I think first talked about in 1969 the density wave itself as it passes through um a gas cloud can create shocks and cause station to happen um so this this beautiful illustration of this happening in this uh NGC 5247 Grand Design spiral where the different like x&#39;s and dots and things are tracing out the spiral in different uh bands of light um and this is not work I&#39;ve been involved in this is you you and ho uh Louis Ho&#39;s group uh work from L&#39;s host group um and so you can see that as you go from Far UV through near near UV blue as you get redder inside corotation which is the circle the gradient goes in One Direction and outside it goes in the other direction so this is a signature of this density wave passing through and generating star star formation and then as the Stars age they&#39;re progressively behind the where the spiral passes through the stars or if the stars are passing through the spiral the trend flips um and that flip will happen at PR rotation inside do we were talking about that earlier and should have looked that up but I don&#39;t remember probably is my guess yeah all right so what do we know about this from Galaxy zoo and the work that I&#39;ve been involved in um so this is one nice result um looking at how spiral arm number correlates with Galaxy color so each of these little diagrams is a color Mass diagram so it&#39;s a little bit like mass star formation rate um which is a more like theoretical way of putting it but we actually observe color or something the that correlates with star formation um so up is red down is blue which means up is less star forming down is more star forming and we have you know normal masses um so all galaxies are the like grayscale which is the same in every one and then we&#39;re overplotting a contour of where you get spirals that have certain numbers of arms and you can probably see that they get Bluer as you get more arms so m 1 are kind of weird m equal 2 most spirals are m equals 2 like this is by far the largest population so this is most of the spirals um m equals 3 are much Bluer four five much Bluer so clearly a correlation with the color of the Galaxy and the number of arms although when Ras dug into this a lot more in his beautiful PhD work um it actually turned out to not really be a difference in Star formation rates it was more about dust obscuration and how much of the star formation was obscured a really uh neat piece of work that&#39;s just from photometry uh Chris mentioned I worked with uh SSS in the manga survey which is spectroscopy a lot so we&#39;re just going to have a little quick reminder of what we can learn about galaxies from spectroscopy from Optical spectroscopy um and then I&#39;ll tell you about some manga work that we&#39;re we&#39;re doing right now um so remember the spectrum of a galaxy is the sum of the Spectra of all of the stars in the galaxy so if you&#39;re looking at a Galaxy that doesn&#39;t have any ative Star formation going on you&#39;re just looking at light from Stars uh predominantly um so you have the Continuum is just the sum of the continua um and then you end up with absorption features um from the atmospheres of the stars um and so that&#39;s how you get your metallicities and your red shifts um there&#39;s no emission lines because there&#39;s no ionized gas in this galaxy because there&#39;s no star formation in this galaxy to first order um the other thing to notice is this uh break at 4,000 angstroms this is caused by a lot of different metal lines um in cooler stars and so it gets bigger and bigger the sort of jump you get at 4,000 anoms gets bigger and bigger as the Stell population ages and you lose the hot massive stars and you start to see more of the light from the low mass stars that live for a really long time let&#39;s switch to a star forming Galaxy instead so immediately you can see the Continuum is Bluer because it&#39;s actively star forming it still has O and B stars that are bright and dominate the light um you have a weak break at 4,000 angstroms you can still see the absorption features um but you also see these very uh distinctive emission features that are coming from the ionized gas recombination lines from ionized gas around star condition regions so really distinctly different and we&#39;re learning a lot about what kind of stars are in the galaxy from the Spectra um so manga uh is a spectral mapping survey uh so if you work with Spectra of galaxies most people uh the sort of the biggest sample is the sdss sample which put a single fiber on each Galaxy and for very nearby galaxies that single fiber is very much in the center of the Galaxy uh manga instead the biggest array had 127 of those fibers in a hexagonal array and so you get a Spectra at every point across the Galaxy okay uh so this turns out to be pretty complicated data um we call these data cubes uh so you get basically an image of the Galaxy at each spectral resolution element or you can do some analysis on the Spectra and generate maps of properties that come out of the Spectra um so here are some example maps of properties you can get out of a manga data Cube um so maps of the uh Hal Alpha emission that that emission line of hydrogen um maps of nitrogen emission line an oxygen emission line silicon emission line um the doler shift is really helpful in astronomy right so we also get U information about how the gas and the Stars rotate separately um and you can um and the random motion of stars you can also fit models to get information about the Stellar population so you fit uh uh Stellar population models and get Stellar masses metalis and so on lots and lots of information for lots and lots of galaxies um so in total manga did 10,000 galaxies um this is them on a color magnitude diagram and a star formation rate Stellar Mass um it&#39;s not it&#39;s not a complete sample of the local Universe it is a representative sample so we oversampled rare galaxies um but it&#39;s embedded in the bigger scss sample so there&#39;s a weights you can use to rec recreate a volume limited sample this is not 10,000 manga galaxies it&#39;s some subset of them but this is uh a cute uh little diagram that Carl westall made uh making the name of the survey in the G gri images and I&#39;ll flick through some other other properties from mango in the same order uh maybe oh I went to two there we go um this is the H Alpha flux of the same one so you can see where things are have a lot of ionized hydrogen and where things don&#39;t uh that d4000 measure which I said correlated with Stellar age pretty well so you can see the old stars and the young stars or galaxies with old Stars got with young Stars Stell velocity uh lots of things um so manga&#39;s been really really great and one of the big things that&#39;s come out of manga is lots of papers looking at gradients of properties from the Spectra um so to get a gradient in manga you put these elliptical isophotes over your Galaxy and then the average the measurement uh in in each of these elliptical rings and then you just plot the plot that uh as a function of radius this is actually not each of these lines is in a single manga Galaxy each of these lines is the average of a bunch of manga galaxies at the same mass um this is really nice work that shows how gas phase metallicity measured in the manga Spectra how the gradient of it uh is different for the Massa galaxies the red lines um and you get very shallow gradients for low mass galaxies you can also uh you know fit straight lines to this and then plot gradients uh of uh properties against something else so that&#39;s what this work is this is the gradient of DN 4000 so the gradient kind of like the rate of change of the age of the stars as a function of radius against what the central value is um some very interesting work so we thought it&#39;d be interesting to add in some morphology here and this is some work that uh my student Emy Weiss has been doing um she graduated from bimar last year bimar and havord a this by Co agreement so I supervise a lot of bimar students for senior thesis work em is also applying for phds this year she&#39;s working at uh the Mariah Mitchell Observatory this year um and so we thought it&#39;d be meet to do those metalis gradients uh as a function of spiral pitch angle and see if there&#39;s any impact the idea kind of came out of work some work by K Daniel who uh just recently moved from binmar to Arizona she&#39;s a dynamist um has some predictions that more Loosely wound spirals should drive more radial radial migration so the idea here is you know your Galaxy gets set up with a radial gradient um because you&#39;ve got more star formation in the center so more metalis more metal rich in the center less metal rich in the outskirt and any kind of radial migration should shallow that gradient that&#39;s the idea um so we thought we&#39;d look for that um and so that&#39;s uh the result is here it&#39;s a little bit complicated but um the which way around is it color is the arm winding um so if you just look at the dotted lines this is our lower Mass sample we split it by mass because there&#39;s clearly a mass dependence of these gradients and in the lower Mass sample whether they were tightly wound arms the green line uh medium or Loosely wound the pink really separated out these metalis gradients which I don&#39;t really know how to interpret yet but it&#39;s there in the data so I thought it was interesting um in the more massive spiral samples we didn&#39;t see any impact from spiral winding um but I think this is really interesting work and EM is currently working on writing this up for publication so U we didn&#39;t just do spiral winding we actually looked at other morphologies on the impact of them on these gradients but this was the most interesting effect um so we could do this radial averaging and just come up with gradients but there&#39;s actually a lot of really beautiful complexity in the manga data so again some maps from manga um compared to their images and you can see spiral structure popping out um in the maps of properties even this d4000 has some hints of spiral structure in it um from the age of the Stars I&#39;m almost done I promise um uh so we wanted to go beyond gradients um and to do that we needed to be able to pick out the Spectra from manga in specific regions of the Galaxy so that would be an interesting crowdsourcing project and so we ran this galaxy Zoo 3D site 3D because of the data cubes from anganga um where I asked people to look at images of galaxies and draw on them where they saw certain features um so this ran um and we have uh done the analysis to get the results um and we end up with these sort of heat Maps which show you where the bar is in Orange where the spirals are in blue we also asked for the centers of galaxies and foreground stars because the foreground Stars mess up um a bunch of the pipeline um and this is all available we published it a few years ago and you can use it with the manga data to do things like this so this is uh the flux in h Alpha as a function of azoth around the galaxy in the spirals the blue so you can see the up and down of the spirals and in the bar in the orange um and all kinds of other things with it um the spiral masks are a little bit tricky to work with um I think I&#39;ll skip over this a little bit in the interest of time you have to pick a threshold level where you believe that it&#39;s a spiral uh but one thing thing we did do um again an undergraduate project uh looking at the amount of star formation there is in spirals compared to the star formation in the whole disc so we were able to show from this measurement that there&#39;s an excess of star formation in the spirals and actually an excess which grows with radius um so this would be no uh enhancement of star formation in the arms and you can see as a function of radius you get a larger and larger fraction of the star formation by area is in the spiral arms um and this is uh Mara kilada who&#39;s a sophomore at haford really uh really really fantastic student has been using these masks uh just this past summer to look at the mass excess uh in the spiral arms so we use a model for the stellam mass surface density fit to the Spectra which is uh published in uh uh Sanchez edl 2021 this pipe 3D model apply the spiral masks so you can&#39;t really see it too well but there&#39;s like a little faint blue line that is uh around the spiral arms in this galaxy that was our arm measurement we have a little white bit where we sort of just throw out the data because it&#39;s ambiguous and then we have an interarm measurement that&#39;s not surrounded um and so what Masa showed is that in both of these regions the stellamar surface density drops exponentially with radius but it drops more quickly in the interarm regions than the arm regions so you have this excess uh surface excess Stellar Mass surface density which grows with r um it goes from about 15% in the inner parts to 30% in the outer parts I think this is really cool yes photometric model fitting that is you&#39;re saying that the effective radius in the arms and interarm regions are different yeah that&#39;s very cool yeah interesting right yeah and so you&#39;re right this effective radius is the whole thing um but clearly they have different effective yeah so we we still have some like data cleaning to do before we&#39;ll publish this um make sure that the you know the masks are really good and that we have the right sample of galaxies and think a bit about the observational biases like the these are the spirals we can see which worries me a bit about this measurement but I think it&#39;s a really neat measurement and I hope we&#39;ll publish it soon yeah so that is uh the end of what I wanted to talk about with spiral I&#39;ve definitely not covered everything we know observationally about spirals but just a flavor of things I&#39;ve worked on you just end with the questions that we started the observational section with um you know there is again some evidence for this winding um there&#39;s a puzzle about isolated galaxies maybe or maybe not I don&#39;t know um and there definitely seems to be a link between the arms and star formation and Stellar uh Stellar Mass surface density excess um that we can measure really nicely with the manga data um I&#39;ll skip over I had a few backup slides which I will Skip and just end with thanks and some pictures of my research group [Applause] thanks would you care to comment on inclined Bridges between spiral arms of of stars and presumably gas in their mechanism of formation inclined ridges between adjacent spirals there are little bridges of comment on that I I don&#39;t know anything about that I&#39;ve definitely uh I see I know what you mean but I I don&#39;t know abely what do they call them feathers kind of [Music] right data from J large arms these sub arms coming off for that yeah yeah I&#39;m definitely working in a like Rift regime where those are not really clear um the Fang data is just beautiful for that kind of thing obviously a much smaller sample though much more nearby and bars can do is transport angular momentum so has anybody looked at correlating either AG activity or nuclear star formation with r yes yes yeah they&#39;re very much so there&#39;s definitely evidence that galaxies bars have enhanced nuclear star formation um and there&#39;s a lot of ongoing debate in the uh observational Community about whether there&#39;s a correlation between bars and AG just spirals no bars I mean that&#39;s a hard question to answer because the G the this galaxy that don&#39;t have spirals are just super different to the ones that do have spirals maybe that&#39;s also the case for spars now I&#39;m saying it out loud but I mean there&#39;s you know there definitely you can find a sample of disc galaxies with Spirals and no bar you can definitely do that but your sample of disc galaxies that doesn&#39;t have spirals they&#39;re kind of weird right so um they&#39;re more like lenticulars and things like that um and then there&#39;s a big Mass difference as well so I think it would be hard to answer that question and it&#39;s the mass difference that&#39;s causing the debate I think Ag and bars where bars are more likely in more massive galaxies AG are more likely in more massive galaxies so if you just do a simple correlation yes there are more AG in galaxies that have more bars but if you do it more carefully controlling for Mass um it&#39;s less clear but maybe a little bit of a think yeah so you just mentioned passing lentic or szer what&#39;s your what are your thoughts of how they fit in people have been arguing this since since hble I think about how they fit into this do you wna you want to speculate or what what what is the current uh consensus if there is one yeah there&#39;s a really really nice paper by Amelia frasa Meli on this um she was a post talk who worked within manga um and um that maybe there so yeah big fan of morphology and I do think you can get physics out of morphology and like you know there are morphologies where you can get unique uh sets of galaxies lentic might not be that right you might end up actually with more than one way to make a lenticular and so Amelia&#39;s work actually was looking at differences between lenticulars and there are some lenticulars that may well be faded spirals so just spirals that I mean the slides I skipped were about red spirals red spirals are interesting and not that rare and shouldn&#39;t last for a long time so these are spirals that have no current active star formation um they should fade over time and then maybe they form a lenticular so that&#39;s just one route right but certainly not all lenticulars can be this so um there&#39;s also the interesting thing that a lot of elliptical galaxies show significant rotation so yeah it&#39;s a diss all the way down fact that we see so many of these things if they supposedly so shortly I don&#39;t think there&#39;s any tension I think it&#39;s just it&#39;s just a common evolutionary route yeah we don&#39;t think so it&#39;s a good question any final questions uh question about Galaxy Zoo um who is in charge of coming up with questions and follow up to that is if someone were to have kind of an idea for a type of question that is more specific is there a process for sort of proposing it that&#39;s a great question um we are very collaborative uh so we have uh bi-weekly telecons and when we have new sets of images we discuss what the good classification set would be uh increasingly our new sets of images are coming via some kind of agreement with the team that make the images and so that team have specific things they want to have uh in the classification tree um we also we have a number of like slide projects going so like quick morphology questions particularly like either or there&#39;s a a mobile app that is really good at like really quick either or questions so if for example you were just searching for shells in the of galaxies um you could be like do you see shells here yes no and that can be done really quickly in the mobile app um we&#39;ve always been a fairly informal collaboration so as I said if you&#39;re interested in collaborating reach out and uh would like to talk so find the arms and the arm number by asking observers asking your scientist to trace the arms then taking some sort of mean um how well does that agree with what you would just get from forier analysis because which is of course closer to the dynamical theory so for sure um example you say most are really n equals to what does that mean in terms of the power I&#39;ve been meing some I have never I have not measured that and I don&#39;t think I&#39;ve seen that comparison done so directly but I&#39;ve been meaning to do that because it&#39;s not that hard to do the fur analysis at some level right so um yeah and I always wonder you know this whole thing about the Milky Way does it have two arms or fourarms and that very much depends on the band you look at the kind of traces it&#39;s a good question and it&#39;s something we should check we the the arm number is not from tracing it&#39;s just from asking people what they see do you you know how many arms do you see in this galaxy what&#39;s interesting is that isn&#39;t necessarily a single answer to that either I said there&#39;s not necessarily a single pitch angle but the number of arms can change with radius so you can have really clear two arms in the center and then it&#39;s just a big mess in the outskirts so what you answer in that case I don&#39;t know but yeah yes so um you mention uh there are also some robotic uh ways to measure the things could you say more about that and also I mean now recently the development of the machine learning techniques and I wonder how do you think that will be useful for these spiral study yeah so this is what we have going right now in uh Galaxy Zoo so the crowdsourcing creates a training set which is fed to zubot is the name of our robot uh which is very good at classifying uh the normal stuff where we have a good training set um and then it&#39;s in the loop so stuff that doesn&#39;t have a high probability of a good classification gets fed by back to the slite um so this is working right now um this is Mike wy&#39;s work so Chris lart Mike was a a Oxford PhD student he worked at Manchester for a while he&#39;s now in Toronto so come over to this side of the Atlantic which is cool um and uh yeah so I refer you to to Mike&#39;s uh work on the machine learning in the loop at Galaxy Zoo um Mike&#39;s also been working some on image segmentation so using our our our masks from Galaxy zo 3D to try to train image segmentation um this stuff is getting better but generally will&#39;ll always need some kind of training setep so that&#39;s the sort of Niche we&#39;ve been discussing in Galaxy zoo of working with it in an integrated way to get the training set a lot of the Galaxy Zoo classification uh not citations to Galaxy Zoo a lot of those are actually people using it to train a machine learn Sor state of these uh machine learn a new network based machine learning is that b still not as good as human being I mean that&#39;s it starts to get hard to quantify right that they&#39;re they can agree with humans to sort of 90% but if you set a population of humans to classify galaxies they will also agree with each other to 80 90% it&#39;s a really cool paper from the 90s where someone did that with a bunch of Galaxy classiers what about the more traditional I mean statistical way do the image process in TR like is there any new algorithm now to measure like without the human guidance just just to do some Statics yeah the P jungle or that identify spiral arms or something like that this is it is that progress on that we have to rely on human being still like to identify spiral for example it&#39;s a yeah I mean wouldn&#39;t it be it would be lovely right to be able to just feed the images to a computer and and out you get really reliable pitch angles I suspect it&#39;s going to be tricky for a long time um but and I think you know I do end up reading a bunch of machine learning papers actually was reading one not to do even with galaxies the other day for reasons to do with being a professor at a small liberal arts college I guess um and even that computer science professor like buried in the paper was a bit where they had people look at the class this the the sets that they classified which I was sort of like oh okay so there&#39;s still people looking at stuff in this computer science paper to figure out what&#39;s coming out of the machine learner um and I think that&#39;s one of the the sort of hidden secrets of all of these algorithms is the amount of human effort that went into training them um so I don&#39;t know I I&#39;m a bit of a skeptic uh partly because spirals are so patchy and weird looking and we were talking earlier right what even is the pitch angle especially if you have these uh things where the gradient of the color and therefore maybe where the brightest light is maybe switches side from the inner part to the outer part of the spir how does that impact the pitch angle um the pitch angles are not the same for every spiral in a galaxy or not the same with function of radius is there even a single number for a pitch angle for a Galaxy so I think they are somewhat complicated objects that I mean it&#39;s like lots of places in astronomy probably all science right once you have a really detailed look at an object it&#39;s much more complicated than you thought Okay one minute any last question Matt I&#39;m I&#39;m curious about the human element here because humans are good at pattern recognition but certain patterns more than others so I mean if if you just showed them if you took a Galaxy and plot it in log R and angle space do they see do people see more than you just than if you just put it in in XY and there&#39;s thing you can do to like unsharp masking and pull out features right we&#39;ve not tried like weird projections there is I mean you&#39;re just giving them the the raw image well we give them a processed image we usually give them a three-c color image with a nice scaling so that features I mean there&#39;s a there&#39;s a sort of art to picking the right three color scaling and the image depth that makes a nice image I mean as you mentioned it&#39;s people so we have to get them to actually engage with the site so there is a little bit of like the images it helps if the images are somewhat attractive although which is easy with spases but there are spin you know the Galaxy Zoo works so well it inspired this whole zooniverse of Citizen science projects and some of the projects in that are the data is less attractive um but if you Mo motivate the question appr appropriately and say you know we want to find exoplanets or something uh people still are engaging with it so maybe we could try uh different projections of galaxies if we think it&#39;s really necessary but computers are pretty good at picking out straight lines right so if you think you got log spirals in a in the right projection you have straight lines I don&#39;t know cor right there are a couple of spots left for dinner if anyone wants to sign up otherwise it&#39;s still to lunch and

Transcript for:Spirals in Galaxies and Large Surveys by Prof. Karen Masters

Transcript for:
Spirals in Galaxies and Large Surveys by Prof. Karen Masters