The Impossibles: Vera C. Rubin Observatory

00:00:03:05 – 00:00:25:15
Leanne Guy, PhD
We do analysis on different time scales. From when the shutter closes on the camera. We take an image, the shutter closes. The data is collected on top of a mountain in Chile. It goes down to the coast to a town called La Serena. Then it goes up through Florida, across the US to the data centers on the ESnet.

00:00:25:17 – 00:01:01:18
Leanne Guy, PhD
We then need to run prompt processing. Then, we need to package them up into a little alert packet, which is roughly 84kB in size, and then transmit these out over distribution networks throughout the whole world. We need to get information to them within sixty seconds of the shutter closing.

00:01:01:19 – 00:01:35:12
Leanne Guy, PhD
So what is Rubin, Rubin Observatory? So an observatory provides you with a building, with a telescope, with mirrors, with infrastructure, with networks, with storage, with toilets, with cafeterias. That’s an observatory. It’s a facility. It was decided to name the facility, after Vera Rubin, who was an astronomer in the early mid 20th century. Her research basically gave us the first real indications that maybe dark matter did exist.

00:01:35:14 – 00:02:00:23
Leanne Guy, PhD
I’m Leanne Guy, and I’m the Data Management Project Scientist at the Vera C. Rubin Observatory. The Vera C. Rubin Observatory is a new, state of the art observatory facility being built in the Andes in Chile on a mountaintop. On a ridge called Cerro Pachón. The facility itself is very impressive, if you look around you, it’s a desert, but it’s a very life filled desert.

00:02:01:01 – 00:02:24:11
Leanne Guy, PhD
It’s not like the Sahara. There’s a lot of life in that desert. The air is really, really clear. It’s really dry, and you can see for such a long distance. It’s just really beautiful. It’s very peaceful, very serene. Looking up into the sky at night from Cerro Pachón, standing next to the Rubin Observatory is absolutely spectacular.

00:02:24:12 – 00:02:43:22
Leanne Guy, PhD
It’s really hard to describe. It reminds me of when I was a child growing up in rural Australia, when I used to go out with my little cardboard star wheel and look up at the sky and try to identify the constellations in the stars. I think, probably the most dramatic difference between the northern and the southern sky is that you can really see the Milky Way from the southern sky.

00:02:44:00 – 00:03:10:00
Leanne Guy, PhD
And when you see this band of the Milky Way across the sky, and you can see the Southern Cross, which is really iconic, it’s just spectacular. You might ask, why do we want to build another observatory? Because there are hundreds of observatories on Earth already. So what’s so special about this one? What’s different? What’s particularly different about Rubin Observatory is a combination of the very wide field of view, that is enabled by the unique mirror three mirror system that has been designed for the telescope.

00:03:10:02 – 00:03:53:10
Leanne Guy, PhD
That, combined with a very fast moving telescope that allows us to slew to an adjacent field in only four seconds. A very fast readout system that allows us to read out images from a camera very rapidly, so that the camera and the system is ready to take another image, and then a fully automated data processing system that can transfer this data from Chile to our data facility in California, process that data and send out alerts to objects on the sky that have either moved, changed in position or brightness in under sixty seconds from the shutter close. Rubin Observatory will be complete in about two years and ready to start its first mission.

00:03:53:12 – 00:04:17:20
Leanne Guy, PhD
And its first mission is the Legacy Survey of Space and Time. This is an automated survey that will image the southern sky once every three nights, which means we’ll have a full picture of the entire southern sky once every three nights. The reason we can do this so rapidly in three nights is thanks to the very wide field of view of Rubin Observatory, combined with the very fast moving telescope and the fast camera readout system.

00:04:17:22 – 00:04:43:03
Leanne Guy, PhD
Modern astronomy has changed a lot recently in that, many telescopes will now conduct surveys as opposed to traditional telescopes, whereby you would think, what objects do I want to observe? You take that telescope, point it at the objects, take some images, study those images, try to detect objects in those images, do an analysis, and then come to some conclusions about the set of objects that you’d studied.

00:04:43:05 – 00:05:11:06
Leanne Guy, PhD
The Legacy Survey of Space and Time will use the Rubin Observatory to relentlessly observe the sky for ten years. So by the end of ten years, we will have observed every position on the sky, roughly on average about eight-hundred times. So this is, this is an enormous amount of data. And the fact that we’re able to go back and repeatedly image the same position on the sky eight-hundred times allows us to make a dynamic movie of the night sky.

00:05:11:08 – 00:05:32:04
Leanne Guy, PhD
At the end of the survey, we will be able to stack those images and see very, very deep into the universe, very, very far back in time. And so this allows us to see much deeper than any other telescope has ever seen. We anticipate that everyone, every astronomer in the world is going to want to work with this data set.

00:05:32:06 – 00:05:51:20
Leanne Guy, PhD
So traditional astronomy was data starved, very small data volumes, easy to work with. Survey astronomy, which is what the LSST is, have to work with massive data volumes. And so instead what they do is map the whole sky to catalog everything they can see. And so what you get coming out of a survey telescope is, millions of images.

00:05:51:20 – 00:06:13:05
Leanne Guy, PhD
And then you have to process those millions of images and you get millions of objects detected, billions of objects, even, in these images. One image is 6.4GB. That’s one image. That data has to be transferred from the summit in Chile. It has to be stored, archived, backed up. It then has to be available for processing. It has to be analyzed.

00:06:13:07 – 00:06:39:08
Leanne Guy, PhD
It has to pass through many, many complex algorithms in order to detect and measure the characteristics of the objects in those images. And then all those resultant data products also have to be archived, then made available to a wide community of scientists to do science. The processing generates more files because it generates new data products from the raw data products, and these data products that are generated are very heterogeneous and of a wide range of file sizes.

00:06:39:13 – 00:07:02:05
Leanne Guy, PhD
What you get at the end is an enormous data set, petabyte scale data sets. The LSST for example, will generate 20TB of data per night. How do you manage a massive great data set? How do you store it? How do you process it? How do you analyze it? How do you develop algorithms to find that needle in a haystack? What questions can you ask of this great big data set?

00:07:02:07 – 00:07:07:18
Leanne Guy, PhD
And how do you find new science within such large data volumes?

00:07:07:20 – 00:07:27:09
Leanne Guy, PhD
The design of the survey has been driven by four key science goals. These are probing dark energy and dark matter. Taking an inventory of the solar system. Exploring the transcient optical sky, and mapping the Milky Way. So maybe one of the great challenges of the data management system is to keep up with the data taking rate of the survey.

00:07:27:10 – 00:07:43:19
Leanne Guy, PhD
So first of all, there’s a very large data vault. There’s there’s the speed that we have to read out the data. So we need a very fast readout system from the camera in order to read out that data rapidly, in order for the telescope to be able to slew to the next field in four seconds and be ready to take a new image and read that image out.

00:07:43:19 – 00:08:09:06
Leanne Guy, PhD
So if we don’t keep up with the image readout, then we can’t continue to take new images. Then we won’t be able to survey the sky as rapidly, and then we’ll miss out on observing really interesting time domain phenomena like moving objects, asteroids, supernova, and one of the things that we want to do initially with this data, as it comes immediately off the telescope, is to take every image, do some basic calibration, and then detect objects in those images.

00:08:09:08 – 00:08:26:12
Leanne Guy, PhD
The first question, the first thing we want to do is say, are there any objects in this image that have changed in position or brightness? Are there any transient-like phenomena that we have detected in these images? Has a supernova just exploded? Has an asteroid moved through the field? And then we put together what’s called an alert packet.

00:08:26:14 – 00:08:45:12
Leanne Guy, PhD
This is a tiny packet of data. It’s about 84kB in size. So the idea here is that we want it to be small, and we want to get this out rapidly. We expect to generate approximately 10,000 alerts per image. And with approximately a thousand images per night that will result in 10 million alerts per night. And we do this in quasi real time.

00:08:45:14 – 00:09:04:12
Leanne Guy, PhD
And the reason for this is that the community take those and they schedule follow up observations on all the telescopes. There’s a whole program of work whereby the community, is setting up a network to automatically trigger other observatories to go and follow up on interesting LSST events. And the reason we want to do it so fast, the reason we want to do it in quasi real time, is that we don’t want to miss the event.

00:09:04:14 – 00:09:24:10
Leanne Guy, PhD
If your supernova goes off, you don’t want to come back a week later when all the exciting stuff has happened. So you want to catch these transient variable moving objects as fast as we can. So basically with LSST, discovery is only the start.

00:09:24:12 – 00:09:44:22
Leanne Guy, PhD
Well, what are we hoping to find with the Legacy Survey of Space and Time? One of the four science drivers of the Vera C. Rubin Observatory is probing dark energy and dark matter. So it’s a key science driver of this observatory. There are many theories as to what dark matter is, and there is some work at the LHC into trying to understand the nature of dark matter.

00:09:45:00 – 00:10:08:15
Leanne Guy, PhD
It’s currently inconclusive. In the 1970s, Vera Rubin studied the orbital motion of stars in the outer part of galaxies and found that their speeds did not decline as would be expected, and this indicated that there must be more matter contributing to the gravity of a galaxy than can be seen. We see gravitational influences that are not understood and cannot be explained by existing matter in the universe.

00:10:08:17 – 00:10:34:18
Leanne Guy, PhD
So when Vera Rubin was looking at the galaxies, she noticed that they, their speeds were not declining and that would be expected by Newton’s laws. That means that there must be more mass, somewhere, in order to contribute to the gravity of the galaxy. Basically, there’s got to be something. There’s a gravitational… There’s a gravitational phenomenon that’s not understood.

00:10:34:19 – 00:10:57:20
Leanne Guy, PhD
There are things that we know we want to explore. Asteroids, for example, as part of our solar system. So LSST will allow us to detect, or know, if an asteroid is coming too close to Earth, give us an early warning and hopefully if we can do something about it. But we also hope that this survey telescope has been designed to open up the possibility of discovering things that we’ve never even thought about.

00:10:57:22 – 00:11:17:11
Leanne Guy, PhD
We will make a data release roughly once every year. When these data releases are made. There’s going to be enormous interest from the community in accessing the new data products that come as part of that data release. So we expect to see a spike in activity. Using the cloud helps us to elastically scale, to provide the resources to the community, to enable them to do science with every new data release.

00:11:17:13 – 00:11:36:10
Leanne Guy, PhD
So the survey will continue to run until 2035, and then following survey completion, there’ll be a year or two to wrap up the final processing. So the final catalog won’t be released till something like 2036 or 2037. And then after that, there will be still years of exploitation of this data that can be done by the scientific community.

00:11:36:16 – 00:11:41:10
Leanne Guy, PhD
This data set will probably still be relevant when I’m dead.

00:11:41:12 – 00:11:45:06
Interviewer
Before that happens, what are you going to do after this project?

00:11:45:07 – 00:11:51:05
Leanne Guy, PhD
Oh my goodness, retire? I don’t know. It’s really hard to say that because when something interesting comes up, you just want to do something interesting.