The Biggest Sky Survey Ever Taken: Exploring the Universe with the Rubin Observatory

evening everyone my name is Andrew frack Noi I'm the emeritus chair of astronomy here at Foothill College and it's a pleasure for me to welcome everyone here in the Smithwick theater and everyone viewing us on the web to this lecture in the 20th annual Silicon Valley astronomy lecture series at Foothill College these programs are co-sponsored by four organizations the Foothill College physical science math and engineering division the Astronomical Society of the Pacific the SETI Institute and the University of California observatories which includes the venerable lick observatory on Mount Hamilton in San Jose and we're delighted to have their help in publicizing and sponsoring these outreach events tonight's speaker is dr. Phil Marshall dr. Marshall is a senior staff scientist at Stanford SLAC National Accelerator Laboratory his research interests are in observational cosmology using gravitational lensing what he does is to weigh galaxies and measured the expansion rate of the universe not bad as a job description he is a member of several collaborations developing new methods to take advantage of the enormous flow of data that will be coming back from the Rubin observatories legacy survey of space and time he received his PhD from the University of Cambridge where he first got interested in techniques for measuring astronomical objects that we may not be able to observe directly and he is currently the deputy director of operations for the VRC Reuben Observatory and that indeed is his topic for tonight an upcoming project which everyone in the astronomical community is looking forward to the biggest sky survey ever taken exploring the universe with the Rubin Observatory ladies and gentlemen please help me welcome dr. Phil thank you for all being here it's nice to we get to spend the evening together okay so as Andrew said I'm here from the Rubin observatory and slack National Lab and I wanted to tell you a little bit about where we are with this exciting new project it's been under construction for a while now but we're getting very close to to its operations period where we turn the telescope on and start observing the sky I also wanted to show you a little bit about the research we've been doing at SLAC to use the data coming from the Rubin Observatory and along the way I thought it'd be good to try and understand a little bit what it is to be an explorer what does it mean to be exploring the universe in the 21st century so you might ask well why deny and where did I get this idea from to base the talk on this theme so I got the idea for this talk the exploring the universe from talking to my wife actually she asked me when we were dating what is it that you do you do why do you why do you do it and I spent some timing trying to explain it to her and she said oh you're like Magellan I thought well what am I really like Magellan I'm not sure um but then I did some searching around found this interesting quote from Edwin Hubble an astronomer in the early 20th century and he wrote or said equipped with his five senses man explores the universe around him and caused the adventure science so I think it's not unfair to think of scientists as explorers but let's let's get into this a little bit more where do we start with doing cosmology with observatories like Rubin we start with some good questions so the questions I'm interested in perhaps you are interested in them to include where did we come from how does the universe work and what else is out there and these aren't very different from questions that explorers have been asking since the beginning of history so here's a famous explorer here is Magellan he started with a different question his specific question I think was something like is there a way to reach the Philippines by sailing west but that was a specific kind of question that he would have written in his proposal I think really he was just interested in what was out there so what do explorers do well they propose he had to propose to get funding to go on this voyage of discovery I did a lot of preparations and then he set out his one of the ships from his flotilla this is the Victoria this is a special one he left in 1519 in five ships with 273 crew on a very famous voyage the first one round-the-world only 18 of them returned three years later on this one ship so I'm hoping that one way in which I'm not like Magellan is that at the end of the LSST survey we'll all still be alive okay so what did Magellan and and and his his crew discover well what they did was set out to find a way to get the Philippines by sailing west and they found it so here on this map in the red circle down here this is the Straits of Magellan this is what allows you to sail safely around South America and emerge in the Pacific which never been done before so we can think of this map as the sort of it's the main result of Magellan's mission made by a Diogo Ribeiro and it shows it showed everybody in the world a new path forward and when you get through them the Straits of Magellan you get to the Pacific there's this whole ocean of possibilities out there then no one didn't people didn't realize was there this interested you asked well what was what was the Magellan's motivation his the man in his own words the sea is dangerous in its storms terrible but these obstacles have never been a sufficient reason to remain ashore unlike the mediocre intrepid spirit seek victory over those things that seem impossible he's a product a product of his age Ferdinand Magellan back when being macho was still cool he goes on it is with an iron will that they embark on the most daring of all that endeavors to meet the shadowy future without fear and conquer the unknown I think that was really what was driving gentleman he wanted to see what was out there okay so what is exploring look like in the 21st century there are places on the earth that we haven't been to mainly the depths of the oceans the very deepest parts you hear talk of manned missions to Mars getting further out into the solar system but there are other ways that we can explore the universe as you know staying here on the earth and here's some some good words to set us off witness from Marcel Proust the real voyage of discovery consists not in seeking new landscapes but in having new eyes so I'm here to tell you about a new pair of eyes we're getting this is the Rubin Observatory we can we should think of it as the biggest digital camera in the world mounted on an 8 meter Class telescope that will feed a data facility that will automatically process all the images coming out of this big telescope the world's biggest camera and that data set 500 petabytes we crunched in order to make measurements that astronomers all over the world can use five hundred petabytes by the way is a unimaginably large amount of data it's enough to store a copy of every book every ever written in every language currently in use ok so where are we with the Rubin Observatory here's the summit facility on Serra push on its a mountain in Chile here's where we are so far you can see the the back of the building where we'll do maintenance of the telescope and the camera and so on you can see the dome that's their open structure that's gradually being closed in this is a recent photo of the panels on the dome going up so it's a big facility here's a slice through interesting things to point out inside the observatory here you can see down here this is the control room this is where the observing specialists will will operate the telescope during the night here's the clean room in the back where we will maintain the camera I'll tell you a bit more about in the MIT in a minute we have our own mirror cleaning facility here and in fact that's already in action preparing the main mirror whose name is M 1 M 3 I'll explain why in a second other things to notice the telescope in its dome up here on the left and here's a great big elevator for moving the mirror up and down when we take it off to clean it and put it back on again so the telescope that's an in sitting in the dome that will be taking the images it's called this Simoni Survey telescope it weighs 300 tons but can slew 90 degrees in 9 seconds so just to give you a I like to act this out so 90 degrees in 9 seconds looks like this 1 2 3 4 5 6 7 so 4 if 9 for a 300-ton telescope it's pretty fast why does it need to go that fast was because we have a lot lot of sky to cover and we want to keep imaging it that's what the Simoni telescope does it repeatedly images the sky it will take a deep image 20 gigabytes worth of data every 40 seconds on average and we'll do that for 10 years so inside the telescope if we take away all the the mounts and the hexapods and all of that stuff and just leave you with the optical elements the mirrors you can see here that there are two mirrors that are actually three mirrors this big one down here is the primary on the outside the outer annulus and then in the middle part is the tertiary and mounted above it there's a secondary mirror okay what does this mean maybe better to show you a picture over here of what the light is doing as it comes into this ammonia telescope so imagine a star some light coming in bounces off the primary up to the secondary down on to the tertiary and then into the camera which is the size of a car mounted upside down hanging on top of this mirror here is the mirror in under when it was under construction in Arizona at the mirror lab at the University of Arizona in the steward observatory and you can see here how it's been made it's a parabolic power Baalak mirror that was first cast and then a spherical tertiary mirror was ground out of the middle took quite a while to ground grind all their glass out in fact it took seven years to build this mirror here's me at least younger me this was me in 2008 after one year of mirror fabrication it was finished in in 2014 so I was already working on on the project then and still going now that's how awesome it is okay so where is m1 m3 now well it's actually installed in that cleaning facility it's been being prepared for observations it arrived last year shipped on a container ship here's a fun photo this is the primary mirror of the Simoni telescope being taken through the the tunnel at the Picaro dam on the way up to the syrup summit at sarah pitch on and you can see it just fits through the tunnel by design is a sense of scale for you here is the mountain sarah pitch on up there on the right you can see that's the Dome of the Rubin Observatory over here you can see the Gemini South telescope same mountain and here you see m1 m3 in convoy being taken up the road on the way to the summit okay so this Simoni telescope has one instrument one instrument only and its name is Ellis st cam this is the biggest digital camera in the world I mentioned briefly was it mean to be the biggest camera in the world well it has 3200 megapixels I forget where we're up to with our phones is it 10 megapixels now this is 3200 another number that's hard to visualize so I'll spend some time trying to help us understand what it means to be that big where are we now this is the focal plane with all the sensors or the CCDs that collect the light mounted and installed in the cleanroom at SLAC this is where we're building the camera just up the road their sensors are all installed and we're testing the refrigeration system this camera has work at low temperature and pretty soon we'll have LSST cams first image so watch out for a press release about that very fun in the cleanroom at SLAC you can see here the the camera is pointing downwards that's how it will will be for most of its life thanks it's mounted above the tertiary mirror looking down with the nice guy up there oh and remember I said it was the size of a car hung upside down over over the mirror you can get a sense of scale in this picture on the right-hand side this is the cleanroom at SLAC where the cameras being assembled that's part of the camera body and by the time it's all been put together all the various pieces the back of the camera will almost touch the ceiling I used to think that it must be something like like a mini or something like that when we said the camera was the size of a car is more like a sedan okay so we're on schedule and due to start a 10-year survey in October of 2022 so that's the the date to remember survey start there'll be first light before then and we'll do some commissioning observations and so on but the survey will start October 2022 okay so just like Magellan we have a crew of hundreds hundreds of scientists engineers and support staff not so good in a storm in the South Atlantic but great for putting together a an observatory of this scale so here's just a sense of the people who have who are currently building the Rubin observatory and here are people gathered together from universities all around the world who are planning to use the Rubin Observatory to do cosmology and in the second half of this talk I'll tell you a little bit about some of the science that we're planning with it okay so the Rubin Observatory is designed the way it is in order to be a very fast Survey telescope that's its main characteristic has a big mirror to collect a lot of light we go through those three reflections in order to make a very good image over a wide field of view and so we can take big images and then cover the sky very quickly we slew fast to go from one place to another very quickly we're going to build up as an image of the night sky every few nights and then we'll do it again and again so here's a picture on the right-hand side of the sky coverage roughly haven't fully optimized this yet but it will look something like this which what you see here is all of the southern sky covered with as you can see on the color scale down here something like 150 to 200 images over ten years in each of six filters so what does this mean it's you can think of it as a Technicolor movie of the night sky taken over ten years where we go back every every few nights and take another image of each patch of sky and so if you imagine just focusing in on one galaxy say we'll be able to see how that galaxy varies over ten years all the things that could happen in that galaxy we can start to study very exciting so an 825 frame movie and six filtered Technicolor and the name of this 10 year survey is the legacy survey of space and time ok so I didn't mention this at the start I thought I'd try and do this all in a straight face you might know of the Rubin Observatory as the LSST the large scale the large synoptic survey telescope it was renamed earlier this year in honor of Vera see Rubin but we wanted to keep the acronym because we had so many things named LSST so now the survey is LSST and the observatory is Rubin we might consider I think doing a second survey after 10 years if we have enough scientific reasons to do that that would have a different name it wouldn't be LSST every something else okay so biggest camera in the world what does it mean to have such a big camera what is 3200 megapixels mean well here's a picture of the night sky over which I've laid layout of the camera focal plane so you can see for scale here the moon is in the top right hand corner so for professional and astronomy standards this is a huge field of view it's not so big for us with you know our eyes that can take in all the night sky at once or even for if you have a telescope at home you can easily get a big field of view bigger than the a necessities the surveys but for the images in the the survey what makes them special is not only are they wide field they also have very small pixels which means we can take high-resolution images and in fact if you wanted to print out one of the images from the survey at standard printer resolution 72 dots per inch you need a 70 by 70 foot piece of paper and three tennis courts to lay out all right so this is still somewhat head hurty let's let's try zooming in to get a feel for what will be in one of these survey images from the from the LSST survey so here's the full moon a bit easier to imagine don't forget though when you the size of the full moon is your the tip of your pinky at arm's length that's how big the full moon is anyway here we are zoomed in on the full moon now let's take the moon out of the way and here's what the night sky the deep night sky looks like behind the full moon so you can see some nearby galaxies here here's one everything with a funny sort of hole in the middle where the detector is saturated that's a bright star this isn't an image from the Rubin Observatory of course but not on the sky yeah this is an image from the Canada France Hawaii telescope legacy survey but it's not not very different from what we expect to get from the Rubin Observatory okay so let's zoom in a little bit more this is a patch this guy the size of the full moon oh yeah here we go this is a hair held up at arm's length for orientation let's zoom on this little patch of sky down here happens to be very interesting here's the hair again see how wide that is and you can see right in the middle there is a very interesting object this is a gravitational lens so it's a group of four massive galaxies that happens to be sitting in front of a distant faint blue galaxy and it's acting like a very badly descent but very badly designed lens the light is coming both ways or several ways around this massive object and being refocused towards us so we see multiple images of the background galaxy so I put this up here because I like to study these I think they're very interesting objects they're very useful for cosmology and I'll come back to that in a bit later so that's what it means to have a wide field of view and small pixels and this is why we're going to end up with hundreds of petabytes of data okay so what would we do at the data facility and the Reuben Observatory well we take in these images as they're as they're taken they'll be piped up to us over special internet cables from South America and within a minute or so we'll have reduced these images compared them with the previous image or an earlier image of that same patch of the sky and made automatic measurements of things that have changed and then once a year we'll go through all of the images and update our measurements of everything including the things that haven't changed so the key here is the automatic nature of the measurement so we'll have computer programs going over these images measuring the stars and the galaxies measuring their brightness their size their color all these things and then putting them into database here's a view of a database and this database will be available to all astronomers in the US and many astronomers around the world to do science with a good fraction of it will also be available to anybody to do to do science with and I'll come back to that towards the end okay so what about some of the science we can do with the images coming from the Rubin Observatory these LSST images okay so the survey will provide us a census of the solar system we have eight planets check there's a whole bunch of other things that we can measure in the solar system and the survey will contain information about 5 million main-belt asteroids 300,000 Trojans there the asteroids that are in Jupiter's orbit 100,000 near-earth objects they're the ones that we might have to worry about like the dinosaurs did and 40,000 Kuiper belt objects which I'll come back to you in a second so all of these are automatically detected as moving objects in one frame of the movie they're here in the next frame they're over here so what can we do with these kuyper Bell objects well turns out that if you look carefully at the orientations and the spins of the eight planets in our solar system there's some weirdnesses about them you they're not quite what you'd expect some of them are tipped on their side some of them are rotating the wrong way and it turns out to explain those odd motions of planets in the solar system one way to do it is to have a ninth planet that used to live in the solar system and has been ejected at some point in the solar system's history where is it now well it would it's predicted to be somewhere outside the solar system but not very far outside the solar system and in fact in the Kuiper belt so Planet 9 we can make we can take these models that explain the weirdnesses in the planets orbits and orientations and make predictions for where Planet 9 could be roughly but we can't make very precise predictions of where it is we can just say roughly where it is what you need is a big Sky Survey to go looking for something bright crossing crossing the images so this is something we can do with the other system images we can go looking for Planet 9 we think it's out there we'll go see let's go a bit further out into the the galaxies here's a notion of possibility once you start taking multiple images of the night sky every few nights you start to build up a sense of how everything is varying in brightness over time we call it their light curve you can imagine you know there are stars that get brighter and fainter and brighter in some periodic way there are things that stars and other things that explode in interesting ways that we can study here's a chart that captures all the explosions we currently know about types of explosions there's intrinsic brightness increasing that way and the duration of the explosion increasing that way so we know about Novi for example nuclear explosions on the surfaces of stars they're over here they last about three months and they're about that right if you go up the chart you get up to super novae things that are even brighter that's where the whole star blows up and you can see over here there's all this space where we just haven't looked and we haven't looked because we haven't been able to take repeated images of enough sky to find them yet so we'll get to go see with the OSS team so right the start I showed you this map that Ribeiro had made Magellan's map maker the main result I called it of the Magellan's voyage well we can also think about maps we can make with the LSST images it's good to remind ourselves that maps are they're not data they're not images themselves they're sort of simplified representations of images when you when you dial up Google Maps it's fun to look at the image behind it you want to see your neighbor's house or something but if you actually want to make predictions about how to get somewhere you want to look at the map view so we do the same thing in astronomy we have all the images of the night sky but actually the automatic measurements filtered in various ways allow us to build up simple of use of that data maps of the universe at different distances or different scales and we can learn a lot about what's what's there and where we should be looking for other things so what does it mean inside our galaxy mapping the Milky Way here's what happens not for the Milky Way with your image too relatively nearby galaxies and you go from a shallow image to a deeper image you start to see all these features in the starlight from those galaxies where do these come from well we think these are trails of stars left behind after the two galaxies have merged into one and then in the third Galax has merged in to make this galaxy even bigger and every time you get a merger you get a trail left behind and so images like this allow you to do sort of fossil studies of the history of galaxies how were they built up and we can do the same with our own galaxy we can look for streams of stars around our own galaxy here's an early study of this with a shallower survey the Sloan Digital Sky Survey with the LSST images will be able to go deeper and find more streams map out more of the history of our own galaxy and start to understand a bit more about where did we come from how was our galaxy made so seems like a good place to pause for a second and say thank you to the inspirational scientist after whom the Rubin Observatory is named so here's Vera Rubin in 1975 working on a very interesting question which is how faster the stars move in other galaxies like our own galaxy and so she made very accurate measurements of the speeds of stars in galaxies both near the center of the galaxy and then as you move further away from the center and what she found was that the stars out here a long way from the center of the galaxy were still there and moving very fast but still there is important if you're moving very fast and you don't escape that means there's something stopping you escaping the only thing we know about that stops stars escaping from galaxies is gravity and so if it's gravity holding these stars in then what's doing that gravitating if you add up all the mass that you can see here in stars in in the galaxy it's not enough it's not nearly enough to keep these high speed stars from escaping and so what Vera Rubin had found was the first evidence for dark matter in galaxies remarkable thing so galaxies are heavier than they than they look by mapping out the streams of stars in our Milky Way the leftovers from mergers that made up our galaxy we can start to ask what is the dark matter made well how does this work once you've found a stream you can look for disruptions to that stream imagine a stream of stars across the sky and imagine some time in our history a dark clump of dark matter flew across that stream you might expect to see a little kink in this in the stream and in fact that's what models predict here's some a model up at the top maybe you can see that that little kink in in black along the way if we find something like that it will tell us something about what kind of matter dark matter is cold dark matter doesn't interact very much it will have a different density profile for this clump of dark matter that's flown through the stream it's a different kind of dark matter there might be a fuzzier cloud of data that flew through the stream and it wouldn't make such a pronounced kink so by studying the detailed structure of these streams we can hope to say something about what dark matter is and make Vera Rubin proud ok Vera Rubin I said she was the inspiration of scientists here she is running the show on an observatory floor one of the very famous things she said there is no problem in science that can be solved by a man that cannot be solved by a woman next time in your life if you happen to have the thought that maybe a man could do a job better than a woman you should remember this photo his Vera Rubin telling you that you're wrong ok let's move further out outside our own galaxy out into the extra galactic sky ok so moving outside our own galaxy let's see what the LSST can do for us in extra galactic astronomy well just like mapping out the stars in our own galaxy we'll be able to map out the galaxies outside the Milky Way and we'll be able to do that to great distances looking further and further back in time we'll map out the positions brightnesses colors and shapes of several billion galaxies what's this useful for well you see here that we expect galaxies to be clustered into clusters of galaxies but also clustered into filaments between clusters of galaxies we expect to be able to see that why do they do that will they're tracing out where the dark matter is where the dark matter has flowed together under gravity in order to make something like this so this is the underlying dark matter structure in this simulation and I previously I was showing you the galaxies in that simulation how could we get at the where the Dark Matter is directly well it's very hard it's dark but what it does do as Vera showed it gravitates so rather than looking for at the speeds of stars under the gravity of dark matter we can look at the deflection of light under the gravity of dark matter so remember those gravitational lenses that I showed you a minute ago all galaxies in the universe are acting as gravitational lenses these filaments of dark matter are acting as gravitational lenses maybe not strong ones that caused multiple imaging but weak lenses that cause little distortions in the positions and especially the shapes of background galaxies so if we look for coherent distortions in the shapes of background galaxies caused by the filamentary structure of dark matter in the foreground we can start to understand something about that filamentary structure first is it there second how has it formed as it formed in the same way that our simulations would predict and has it formed at the same rate that our simulations would predict we can start to get at that by making these weak lensing Maps so here's a picture of what's going on light from a background blue galaxy flying through the universe past all these filamentary structure of dark matter and as you can see as the light ray travels from left to right the universe is changing it's getting clumpy as the dark matter falls together over time and so we can try and make maps at different distances from us of the dark matter and watch this cosmic history of structure forming amazing thing made some progress on this already it's one of the ways that we've been studying the universe in its expansion history and you probably all know this but it's good to remind ourselves that there's something odd about the cosmic expansion the expansion of the universe we've known for a long time that the universe is expanding that it started in some hot Big Bang expanded very rapidly and then and then started to slow down as the stars and galaxies formed over the course of billions of years the unexpected thing was that sometime in the recent past a few billion years ago that expansion rate seemed to instead of slow speed up again so the expansion of the universe is accelerating and no one knows why I put this picture up to try and illustrate just how weird this is it's as if you threw a baseball in the air and despite you knowing or thought you knew that under gravity what goes up must come down the same is not true of the universe so this is perhaps the biggest mystery in physical science and we should do everything we can to try and understand it one way to understand it is to come up with ideas for what could be going on that's what the theorists do what can observers do we can try and make the most accurate measurements we can of the expansion in order to rule out some of the ideas most of the ideas hopefully all but one of the ideas that the theories come up with so that's our program how do you measure acceleration in the universe though it's hard enough to measure distance how do you measure acceleration well if you think about this San Francisco Street how would you measure the acceleration well if you remember high school physics or something all that time you spent rolling trolleys down hills and timing things remember that the steeper the slope the faster the trolley fell you didn't even need to go to high school for this use roll bones down hills and you know that the steeper the grade the higher the acceleration so if you want to understand acceleration one one way you can do it is to try and estimate the the gradient that the ball is rolling down you could look at the ball directly you could study the dynamics of falling objects and we do that with for example star in galaxies the dynamics of falling objects orbiting objects we can also do it by galaxies falling together under gravity that tells us something about the acceleration going on but we can also we can also do something else we can try and estimate this gradient directly what does it mean gradient in the universe we can try and estimate the relationship between distance and apparent speed how would you do it in San Francisco well if you knew that all these cars headlights were the same brightness you could estimate how far away they were and then you can look at how high up the hill they seem to be and you could make some estimate for how steep the hill was hopefully that's plausible the corresponding thing in observational cosmology is to use certain objects as standard candles where we think we know what luminosity they are they what intrinsic brightness they have and then use those to measure distance and compared with their apparent speed going away from us their redshift all right so standard candles LSST the survey will contain hundreds of thousands of of supernovae that we can use to do this type 1a supernovae it turns out all look roughly the same their standard candles or at least standardized herbal candles we can use them to measure distance if you have several thousand new supernovae detected every night you can start to build up a big enough number that you can ask questions like is the acceleration the same in that direction as that direction is the universe expanding in the same way over there as it is over there I like to point down because I have to remember that the surveys been taken in the south okay so how well do we think we can do combining all the different ways we have of measuring cosmic acceleration with the LSST survey data we think we can get to two percent precision and also do cross checks to make sure we haven't made any mistakes so that's a big big program ten-year program and you should get people to come in here and give you updates every now and then okay so I wanted to spend just a few minutes telling you our little piece of this program at slack actually we have several pieces but let me let me be honest my little piece with the people that I work with directly so we use these strong gravitational lens to measure distance in the universe each one it turns out can be a very accurate measurement of distance and it's because when you have this multiple imaging effect you see multiple images of the same background object if that object is varying in time flickering then then you can imagine that you see the flickerings at different times in this image compared to this image just because the light is traveling on different paths and so there's a delay between the images we can measure that time delay and it turns out we can use that to measure distance now LSST will provide measured time delays for hundreds of lens systems and information about these massive galaxies that are doing the lensing for thousands more so here's a preview if you like of the kinds of lenses we expect to be able to find in the LSST survey images now it turns out you not only have to measure the time delay between these flickerings you also need a very good model for the mass that's doing the deflection you have to understand where the mass is where the dark matter is in this galaxy in order to be able to use it to measure distance and it turns out to do accurate cosmology using these lenses you have to spend a lot of time modeling each lens at the moment it takes us several weeks computers running full speed to do the to do the modeling analysis so we started asking said well how can we scale this up to LSST the size of the LSST data set makes us think differently and this is going to be true for men astronomers I think we have to think differently about the way we do our analysis so since we're in Silicon Valley we've been trying to learn from our colleagues down the road and we're using artificial intelligence to do superfast lens modeling you're probably aware by now that the tech companies in Silicon Valley have been using neural networks deep learning networks to do facial recognition you make a neural network and you show it millions of images of faces and gradually the network learns what faces are such that it when it's shown a new photo it can tell you who it is that's a classification problem that they're that they're solving basically solved now so we're doing the same thing except that we're not interested in the classification problem this type of lens versus that type of lens what we want to know is what are the model parameters of each lens so that we can make predictions with that model of what the time delay should be in a given in a given cosmology so we make millions of synthetic lenses where we know that true parameters and we train the network to be able to just look at a new image and tell you what the model is that's really cool so how is this going jiwon Park and Sebastian agna korenna they're grad students at Stanford working on this along with Simone Berra and Erin Rubin and me so far so good we're starting fairly simple we have moderately idealized conditions but we're finding we can do accurate cosmology with competitive precision about ten thousand times faster than current methods so this is this is going to allow us to do new things that we hadn't thought of before now that we can do things super fast that's just a little taste of what we're doing that's like okay so the Rubin Observatory will provide many different maps for others to follow it's making us think differently it's giving us new views of the solar system of the Milky Way of the extra galactic universe those maps for others to follow those others could include you how is this going to work well one of the things we'll be doing with the Rubin Observatory is making it possible for anyone to take part in the science so we have a system where professional astronomers can design citizen science projects for anyone to help out with so for example I work on one project called space warps this is to find new gravitational lenses in deep sky images it's been pretty successful so far that the volunteers who spend their time looking at images saying lens or not they turn out to find different kinds of lenses than the than the machines do which is pretty cool so there'll be more projects like this with the Rubin Observatory data we teamed up with the Zooniverse to make a system where where anybody in fact can make a make a project and so one fun thing is are we going to see citizen LED citizen science projects I think that would be great but it makes us think what does it take to be a scientist you might be sitting in the audience thinking well I I could never do that but what does it take really to be a scientist so let's ask let's ask era so here she is as an undergraduate at Vassar College in the 1940s but in one of her memoirs she writes that her childhood bedroom had a window that faced north and at about age 10 she started watching the stars go across the sky and by age 12 she'd prefer to stay up and watch the stars and go to sleep and she started learning inspired by this going to the library and reading it was just nothing as interesting in my life as watching the Stars every night that's how inspiration when she found the night sky some of you might recognize that feeling she found it a remarkable thing she says you could tell time by the Stars throughout the night I would memorize where each meteor went so that in the morning I could make a map of their trails even then she says I was more interested in the question and then the answer I decided at an early age that we inhabit a very curious world so what does it take to be a scientist I think two things first thing is you have to be curious you have to want to know how the universe works and the second thing is you need to never ever give up okay so what will we do for helping budding scientists young and old get to know the LSST data start thinking about how to do science with the LSST data how to do astronomy as citizens well here are some tools that will go out into classrooms to introduce students to the data introduce concepts to do with it to do with astronomy there's an awful lot of teaching we can do with the LSST data we'll have a sky viewer that allows you to browse around the night sky and learn about galaxies and stars and go and see their what's happened to your favorite objects lately try and understand these light curves and maybe make some contributions but certainly give everyone a sense of the adventure that we're on ok so his is Vera Rubin again she gets the last word so she says the joy and fun of understanding the universe we bequeath to our grandchildren and to their grandchildren that's what we're going to give them and with over 90% of the matter in the university were to play with even the sky will not be the limit ok thanks very much for listening [Applause] so you were talking about megapixels gigapixels resolution and technical parameters but the atmosphere gives us certain distortion so I would like to ask what's the like a balance trade-off between going the pixels size like the resolution wise versus the distortion so what's the limit we can go taking into account the atmosphere and for all those six color filters yeah thank you so it's true I said we have a lot of pixels which allows us to make high resolution images but really the way we should think of it is the the atmosphere as you say sets a limit to what what resolution we can get we go to high mountains in Chile which are very dry very high they give us very good image quality that really limits the resolution so the median seeing as we say is 0.7 arc seconds in syrup a charm what that means is that there will be each star in a typical image will appear as a round blob roughly three and a bit pixels across and in fact we knew what the image quality at the site would be when designing the camera and the optics and we chose the pixel scale to be such that every star in a typical image would be three and a bit pixels across so that the design of the telescope matches the quality of the imaging at the site now a typical image will be as will be that resolution but sometimes they're not so much the weather but the conditions in the atmosphere maybe it's a bit windy at high altitude or something will be such that the images will be blurrier than that and the images won't be as good as 0.7 arc seconds on the other hand some of them sometimes you'll get a very still night and you'll get an image that is better quality than 0.7 arc seconds so one way to think about the LSST Survey data set is that you're sampling the distribution of resolutions and if you want if you wanted to the best possible image you could just take the images that had very high resolution where the atmosphere was still and make an even higher resolution image but roughly speaking that that's what the trade-off is we chose the the design the optics and the camera such that stars would appear to be typically three and a bit pixels across there's a bit of variation from from one filter to the other the the zband tends to be better quality than the in the bluer bands but roughly speaking the three pixels across in the eye band gives us a pretty good range of resolutions so two quick questions one is what visual magnitude can you achieve in your typical image so in a single visit we get down to about twenty fourth magnitude did you hear the wow that's pretty do you think that up is recording thank you the second question is you show the field of view in that picture what would the Hubble feel the view be compared to that I imagine it's pretty tiny it is pretty tiny I used to you see if I can remember so the Hubble field of view is several I forget the number sorry several arc minutes across and the full moon is thirty arc minutes across so if you think of the the eye in the man of man on the moon that's about the survey camera on HST roughly I might be a little bit off that but roughly speaking thank you hello thanks to the talk how would you describe the light sensitivity of this sensor array how would I describe the light sensitivity of the array it's it's incredibly good but I mean these are very advanced sensors but they're not they're not sort of super special and I'm qualifying this a bit because I know that there are sensors out there that are that are better behaved in some respects than others what we're doing is we're using new sensors that very thick a very deep wells can can can do the imaging very reliably but there are better sensors in other cameras what's different about the LSST cam is that the scale of it so we have very good sensors we have a lot of them and we need to keep all of them going for for a long time so it has to be reliable so I would say we have better than better than average light sensitivity but we're going for we're going for scale and reliability what would it be so could you if we could see anything beyond the Andromeda galaxy what would it be if we could see anything beyond the Andromeda galaxy so the Andromeda galaxy for the rest of people in the room that's our twin galaxies right that's how we should think of it and dramatize our nearest big neighbour it's coming towards us and so on a member of the local group it turns out we can see many galaxies beyond the local group there's a local local cluster and then beyond that as I was saying there's billions of galaxies out there in other groups and clusters further out and it's interesting to then compare the galaxies in our local group with other groups that are like ours and start to understand a bit more about how special our galaxy is so far it seems like it's not a very special galaxy but maybe we'll we'll learn more by making more and more comparisons with more and more galaxies as we look further and further out thank you um thanks for your question hi I have a question about out beyond the andromeda galaxy so since you'll be able to look at the large-scale structure of the universe and it's a space and time survey and goes back in time as you go further in space and you're gonna be working with theoretical models my question is once you've sort of used the actual data set to home in on some nicer you know theoretical models of the large-scale structure of the universe and its formation will you then be able to extrapolate what the universe looked like at different times if you didn't have to worry about the speed of like in other words if you had like a God's eye view and we're taking a snapshot where time wasn't a factor but like this is what the whole universe visible universe looked like at five billion years and ten billion years and thirteen points will you be able to do something like that with any accuracy yes I think we will so we're all so so in the patch of sky that we're looking at and remember that's half the skies it's a big patch in that patch we'll be able to make 3d models of the large-scale structure at different distances from us and as you say hence that at different times but of course we'll still it will still only be able when we're doing that we have to think sometimes based on the same guy here's the the bit of the universe that we'll see as it was a few billion years ago and then here's the bit of the universe that we'll see as it was ten billion years ago and then we can maybe make another one a bit further out than that and so that's showing us a patch like this in a sort of shell around us so that that's a lot of information about what the universe was was doing at that time and how the universe behaves in general but if you think about it it's still a fairly small fraction of the of the universe I think what's what's going to be interesting is to take that take that information in that shell and use it to constrain a model of the entire universe where we have the CMB that the microwave background that constrains what it was like in the very early early universe in a shell very far out and all the way around and then we'll have more information at in half the sky and these other shells and I think we want to make a model that explains all of those data and starts to show us what the 3d structure of the universe was like at any time and I agree I think that was a super cool project and we made a small start on it we started with the CMB data and we just looked at the largest scale fluctuations like roughly how much mass is over here compared to roughly how much math is over here and then we started to try and play that model forward in time to say well what would that must be doing ten billion years later and so we're very low resolution at the moment what we'll get from L from the LSST data is a much sharper view everything will kind of come into into into picture into better crisper resolution and it will be enormous fun to make that movie and play it for ourselves thank you your graph of the duration of the brightness versus the magnitude of the brightness last leads me to ask the question if you got a resolution of only a few days is that going to be useful for doing something like searching for optical SETI signals let's see so it's true we'll come back on average M on average every few nights but it turns out we won't be very strict about that actually we'll come back to any given patch of sky we'll take a 30-second its effectively a 30-second exposure it might be a pair of 15-second explosions but anyway a single visit turns out we'll come back to that patch of sky 30 minutes later and the reason for that is to try and track the fast-moving objects 30 minutes is about what you need for to see something move within our field of view and be able to link them together so we'll do it every 30 minutes we may then come back the same night a few hours later and then a few days later we'll come back and do the same thing but it won't always be strictly three days sometimes it'll be two days sometimes it'll be dey's occasion you'll be one day and so we'll start to build up for irregular sampling in time and that's what will allow us to to search for faster variations now your specific application how much will it help with the search for extraterrestrial intelligence I have to admit I don't know I've been puzzling more about you know explosions of stars and things like that where I think there is a lot of unexplained and and as yet undiscovered explosions likely to be out there for us to study but it could be that we learn something interesting about 30 as well thank you oh no now I'm on camera saying that it could be interesting so just some information that may be of interest your mirror was done at the mirror lab at the University of Arizona and Tucson they offer public tours several days a week and I've taken the tour it's absolutely fascinating anyone who's in the Tucson area highly recommended thank you that's a great comment in fact just as another teaser for people who are thinking about going on holiday to Tucson and taking a tour around the lab what's what was most fun for me was how they make these parabolic parabolic mirrors turns out if you have a bowl of liquid and you new spin it the surface of the liquid just under gravity and in centrifugal force makes a parabola and they do that with the molten glass so they have these these big vats that are turning and maintaining a nice parabola as the glass very slowly cools and sets it's so cool I agree definitely worth a visit thank you I did a back-of-the-envelope calculation and I think you're going to be pulling down about 50 terabytes a day I think that's right yeah can you discuss your storage format do you remember those floppy disks in the 8th you're not using those okay so in another das text format either so that like I said that the images will will stream over special internet up to the data facility in the in the states for processing and we'll keep those images stored on on disk while the processing is happening eventually I forget what the plan is after I've to one or two data releases will start to archive on tape which is still the best sort of high-density format for long-term storage but in the meantime we'll keep things on disk and as separate images of the directory as separate images in their directory I suspect not because I think we'll hit the limits on the number of files you can you can fit in a in a directory so I think there'll be a we'll use a file format that allows multiple images to be stuck together thank you okay so thanks again there's a somebody who lives nearby here who's put an awful lot of communication satellites up in orbit how is this project of his going to affect what you're going to be doing okay so for those of you haven't been following questions about these low-earth orbit satellites that companies like SpaceX and others are putting up what their goal is is to provide Internet to people all around the world who may not have coverage so it's very worthy cause but these low-earth orbit satellites look very bright maybe you've seen them launched in constellations of you know you see them as streams of satellites going across the sky very bright at you see them at dusk and dawn because they're in low-earth orbits so that's the it's kind of like the space station you only see them at dusk and dawn but the question is how what impact will that have on the legacy survey of space so the short answer is we don't know yet but we're working on it with SpaceX so the chief scientist for the Reuben construction project is working closely with them in fact the last launch that SpaceX did one of the satellites had been darkened they painted some of the surfaces to try and reduce the brightness and we're taking observations of those of those satellites in that particular little constellation we'll have more results on that later this spring so the you can see where we're going with this we're trying to understand first how bright they are and what we can do to make them less bright and then on our side we can ask questions like how could we take the observations how can we process the data in such a way to mitigate the effects basically what you what you'd see in analysis T image is a is a very bright streak across the image we think that in a current planned exposure time it wouldn't saturate the image but it would be very bright and so we would have to do something to try and get rid of that band or at least cope with it so we started thinking about things we can do do with the data it's a it's a good question I was think about this in the car on the way up here because I thought you might answer my might ask this and I remembered another story from from Vera Rubin actually about about similar as simple similar things when she started worker and Observatory I forget which one it was sorry she started work in Observatory she was the only female astronomer there and her new male colleagues said to her well you're going to have some problems using the facilities we never had a woman astronomer for we only have many have men's rooms and so what she did was went back to her office and cut out a triangle of paper and went back to the door of one of the men's rooms and just stuck it on the picture picture of the man and then said to a colleagues look now you have a ladies so this is why I think you know to be a scientist you never ever give up things like this come up all the time there are new problems to solve and we're try and solve them thank you two quick questions that what percent of the sky are you doing the survey on because it's only the southern hemisphere and not all of it and the second one is why slack I mean I don't see them as the the forefront in astronomy so first part it's about 50% not exactly 50% because we haven't fully optimized the survey strategy yet it might be better to edge a little bit into the northern sky and cover more sky but if you do that of course you can't go as deep in the southern sky so there's a trade-off to understand and we're taking taking input from all our over the Astronomy community as to how we should optimize the survey the second part of your question why slack so it's interesting a few years ago the Department of Energy that provides funding to slack to do its research they did a they followed the high energy physics his community's own survey of the physicists and what they wanted to do over the next few decades it's called the Snowmass process it's something we go through every every few years and it turned out the particle physicists the high-energy physicists were interested in cosmology you can do fundamental physics by studying the night sky this property of the universe that it's expanding points towards there being a new force or a new energy field that's causing this acceleration and that's of great interest to high-energy physicists including those at SLAC in fact we have high-energy physicists who previously worked on the babar experiment doing particle physics who've moved into cosmology in fact Aaron rumor is one of them I showed here showed his name he's working with me on the AI for lens modeling so we should think of cosmology as part of high-energy physics it's getting at the fundamental physics of the universe in a way the particle physicists are interested in and that's why slack is is interested in doing this and perhaps another question what's the budget the budget for this project so it's to build the Rubin Observatory is between four and five hundred million dollars to operate it is going to be between seven and eight hundred million dollars it could be less of time the way you should think about the cost of the Rubin observatory is a billion dollar project another way to think about billion dollar projects or other billion billion dollar projects you could have in mind new run a new runway at a major airport that's a billion dollar project or one election for one individual yes exactly exactly exactly I think two more questions okay one from this time with no second so how are you going to be affected by correct for take advantage of the displacement from the Earth's orbit the displacement from the Earth's orbit how do you mean the earth is moving in the circle okay so I'm sorry I've got jargon on the brain I know that as parallax so we can use the fact that the the earth is moving around the Sun from one side of the orbit to the other every six months we can use that to get a handle on how far away nearby stars are so called parallax distances we'll be able to measure parallax distances to many more stars than we have been before because we're able to look so so faint in any given image and so this is part of mapping the Milky Way it will give us a allow us to go beyond what the Gaia satellite has done extend that model out to greater distances and get a much more precise picture of the of the Milky Way in all its stars and allow us to do better at this sort of fossil record of how the Milky Way was formed yeah it's it's true we should be able to do extremely well at parallax distances and go beyond the Gaia satellite well since it's the last one thank you for a great talk I'm just curious you talked about making 3d models of the unit gallica billions of galaxies is what you said this is an optical telescope right so it's not spectrographic you can't measure redshift even though you have six as I understand it but please tell me so how would you know if there's no variable in some galaxies that you see out there how could you determine its distance good so I think I would quibble with we can't measure redshift I think we will be able to measure redshift okay as long as you kind of go with me a bit on what it means to measure something so to measure the redshift of a galaxy normally you take a spectrum and you'd see all the emission lines and absorption lines and you'd match up the emission lines with where they're supposed to be and you'd get a very accurate redshift and you'd say ha I know how far away that is once I've assumed how the universe is expanding so forget about what that cosmology stuff for now let's assume we know how the University expanding we can measure redshift and their estimate distance and then start to build up 3d models of of the universe as you say so we don't have a spectrograph we can't map out a spectrum in very fine wavelength bins that allow us to see emission lines and all the features but you can think of six filters as giving you a very low resolution spectrum you have a brightness point here and one here and one here and one here and one here and that shape of the galaxy spectrum you can infer the redshift from that low resolution spectrum made up of the six wave bands now now it's only a question how accurate can we do how accurate can we be at measuring redshift that depends on what kind of galaxy is a little bit but we think we should be able to get photometric redshifts that's what we call this photometric redshifts to 5% for the EZ galaxies maybe 10% for the hard ones and occasionally there'll be catastrophic errors where we completely get it wrong but there are ways to get around that by looking at the neighbors of these galaxies you know you expect galaxies to cluster together and so you'd be surprised to see a very nearby one in amongst a lot of distant ones so I think we will be able to do well with photometric redshifts well enough to to support our big statistical understanding of things like explosions and variability in the universe so I think we can measure redshift we just have to cope with that uncertainty thank you [Applause]

Channel: SVAstronomyLectures

Views: 9,679

Keywords: astronomy, science, astrophysics, science news, telescopes, astronomical survey, Rubin Observatory, LSST, Phil Marshall, universe, asteroids, supernovae, dark matter, Galaxy, big data

Id: 59xvUJsqQVU

Length: 69min 16sec (4156 seconds)

Published: Fri Apr 03 2020