Suggested Searches

Adam Moreno Talks About Supercomputing for Global Systems

Season 1Oct 19, 2017

A conversation with Adam Moreno, a postdoctoral fellow working with the NASA Earth Exchange at NASA’s Ames Research Center in Silicon Valley.

Adam Moreno

Transcript

Adam Moreno

Host (Matthew Buffington): You’re listening to the NASA in Silicon Valley podcast, Episode 64. Today for the intro I have Frank joining me again. Frank, tell us a little bit about our guest today.

Frank Tavares: Hey, Matt. Sure! Today our guest is Adam Moreno. He was actually a fan of the podcast before coming to NASA. So that’s kinda a cool thing to see that cycle of the podcast bringing people in, and now putting that all back out there.

Host: It was interesting after we recorded the session that he was like, “Oh yeah, I’ve listened to all the episodes before, so this is kinda surreal.” So it’s nice coming full circle.

Frank Tavares: Totally, totally. So yeah, Adam is a computer scientist by trade who sits at the crux between supercomputing and Earth Sciences. Which is I think a really cool combination of skillsets because NASA does so much research on Earth sciences and gathers a huge amount of global data. So what Adam does is he works with a program called NEX, the NASA Earth Exchange, that works to bring those datasets and make them public, so any researcher anywhere in the world can make use of all that data that NASA collects. So especially in this day and age, having access to all of that information about the Earth’s climate systems and weather and all this different stuff can be really, really important.

Host: Excellent! So we don’t want to spoil it too much, but a little bit of housekeeping before we jump on in. Folks who are listening, who want to participate, we’re using the hashtag #NASASiliconValley on any social media platform you can think of. Also, we have a phone number, that’s (650) 604-1400. Give us a call, leave us a message, we’ll see how we can integrate that into the episodes. Also, obviously if you’ve already found us, we would love reviews, comments, thoughts, suggestions, all that stuff. You can find us on all of the major podcast platforms. Or you can grab our RSS feed and plug into your favorite podcasting app. But for today’s episode…

Frank Tavares: Let’s hear from Adam Moreno.

[Music]

Host: Tell us a little bit about yourself. You haven’t been here that long, but what brought you to NASA? How did you end up in Silicon Valley?

Adam Moreno: Yeah, so I’m from a small, rural town in northeastern Oregon. Then I went and got my bachelor’s in computer engineering from Oregon State. I went into the Peace Corps for a couple years, then went to graduate school for my master’s at University of Montana. There, I was fortunate enough to have an advisor who, super smart, super nice guy, and has been working with NASA for years now, decades now, on some of their Earth-observing satellite missions.

While I was there, he, I guess, got an invitation from NASA Ames to send some of his students to come here and learn how to use the supercomputer that we have here to be able to do cool Earth science.

Host: So even thinking back, I’m imagining spending time in Montana, it’s called Big Sky for a reason. But even imagining like Peace Corps, I mean I know from traveling overseas, if you go to some countries that are developing countries, not as much light pollution. I’m thinking Montana, not as much light pollution. So you’d just be like sitting back and looking at the stars, like just blow your mind.

Adam Moreno: Yeah, that’s right. But you know actually I think it’s interesting because if you ask most people what they think about NASA, or what comes to mind when they think about NASA, the first thing that comes to mind, of course, is space. You know, astronauts, space shuttles, and of course we do all that cool stuff. But we also do a lot of cool Earth science as well.

Host: Absolutely.

Adam Moreno: And so actually, in the Peace Corps, this is where I kind of changed gears. Because I had a degree in computer engineering, but then I went into the Peace Corp, and I went –

Host: That seems kind of rare, going from computer engineering to go to the Peace Corps. Where did you go? What did you do?

Adam Moreno: Well, so I went to Paraguay. I was sent out into the jungle in Paraguay to a little community of like 300 people. I built my house there. I worked with the schools, farmers, students, all that sort of stuff. But what I think actually kind of changed my path from maybe a traditional engineering path to doing Earth science sort of stuff is just, you know, the majority of people who live out in the countryside in South America are subsistence farmers, right? They live off of what they grow. If they don’t grow anything, their family doesn’t eat.

So things like droughts and heatwaves –

Host: High stakes.

Adam Moreno: Yeah, they’re a matter of life and death, really. So when I came back I realized I want to work on something that can help the majority of the world that are so tied to the environment that it’s a matter of life and death. So I came back and I decided that I wanted to study climate-related issues, ecology-related issues. So at University of Montana I studied forest ecology and ecological modeling. And ecological modeling just means taking an ecological system, like a forest, turning it into mathematical equations, and then turning those equations into a computer model that you can then use to –

Host: Yeah, and this is one of the cool things that NASA does, is the data collection, then using the scientific community to turn data into knowledge. But gathering that data, it’s not just from satellite-based telescopes. There’s a lot of data collection on the ground. People always tend to think of the larger climate, the big picture, mother Earth perspective, but a lot of that data is used for agriculture.

Adam Moreno: Yeah, that’s right.

Host: Even here in the United States, it’s crucial information for, you know, even disaster response. You’ve got to know how the Earth is reacting. That data that NASA collects pays dividends for all kinds of groups down the road.

Adam Moreno: Yeah, that’s right. And that’s NASA’s job, right? To understand this huge, massive, complex system that we’re living on, and in, and that we depend on. I mean it hasn’t been that long that humans have been able to actually study the whole entire globe. It wasn’t until the ’70s that NASA and USGS sent up the first satellites dedicated to monitoring the environment, actually. And also you need big computers to be able to understand what’s going on, and that also hasn’t been that long. So actually this is all relatively new science, being able to study continents, entire continents, or the entire global system altogether.

But yeah, so that’s what I do here. But that was my master’s. I did my master’s in Montana, then for my Ph.D. I went to Vienna, Austria, and I did my Ph.D. over there. My last year of my Ph.D., I came to AGU, the American Geophysical Union, here in San Francisco. I knew I wanted to come NASA. I’d been here before, I had some exposure. I was like, “Okay, this is where I want to be.” So I met with Rama Nemani, who is now my mentor here at Ames, and I just asked him. I was like, “I want to come be in your group, how can I make this happen?”

He pointed me towards this fellowship opportunity that all NASA bases have. So I applied for it, and I got it, and now I’m here.

Host: And you said, it hasn’t even been quite a year yet.

Adam Moreno: No, not even a year.

Host: Oh wow. And we were talking right before we came on doing this that you had actually listened to the podcast before even joining up.

Adam Moreno: Yeah, that’s right.

Host: “Well, I’m going to be working there, better find out what people are doing.”

Adam Moreno: Yeah, that’s right. So, my last year of my Ph.D., I knew I was coming. I wasn’t really nervous, I was just kind of concentrating on my dissertation and defending, but then I stumbled across this podcast here. That’s when I got nervous, actually. Hearing about all the cool stuff all the smart people here are doing, and I was like, “Oh okay, coming to the big leagues now.”

Host: Oh, fun. So you’re working right now, is it called Earth Exchange? Am I saying that right?

Adam Moreno: The NASA Earth Exchange, yeah. I work in a group called the NASA Earth Exchange, or NEX, and it’s made up of Earth scientists and software engineers. So actually it’s kind of a hybrid group of the Earth sciences division and the advanced supercomputing division. What’s NEX’s goal is, is just facilitating Earth scientists to be able to use all of NASA’s facilities, like the supercomputer, and getting access to all of the data that exists in the world actually.

So we try to bring in massive datasets, bring it in to one place, then allow the NASA scientists, especially throughout all of the bases at NASA to come and be able to use the facilities here. And I think the longer-term goal is to make it even more accessible to the general public at large. You know, NASA has 10 to 12 bases, however you want to count it.

Host: Different centers and locations.

Adam Moreno: Right, different centers. You know, there’s a lot of people doing all kinds of science there, but maybe they don’t all have direct access to the supercomputer like we do here at Ames. So NEX’s idea was just to help facilitate that.

Host: Help facilitate that because it’s not only the researchers here, it’s also NASA, as funded by the taxpayers, goes through great effort to get that information also available to the scientific community. Because we don’t have to do it on our own, we can reach out to the larger group. You almost kind of crowdfund, or you know, crowdsource, I guess. Have everybody working and looking at these datasets and you just see things as they pop up.

Adam Moreno: That’s right. One of the great things that NASA Earth science does is they give out their data to the entire world for free, actually. This actually spurs science and innovation all over the world. Because as scientists here in the United States, maybe we don’t focus on small countries like Paraguay or something, but if I’m an ecologist in Paraguay, you know, Paraguay can’t afford to have a space program, but they can get access to NASA’s data to help them understand their ecosystem over there. That’s one of the great things NASA does, is give away all of the data that we have processed, ready to use, for free to the entire world.

Host: Yeah, not too long ago, we had one of the exoplanet specialists in here, and they were talking about – of course, traditionally when you think of NASA, you think of rocket launches, but you also think of looking out into the universe, the solar system and beyond – but they were telling this story of how looking at a faraway star and how a planet will transit in front of it, and then you can confirm, “Yes, okay. There is an exoplanet around that star.” You can see how that planet pulls on the gravity to get another idea of like, you can start deducing the density of that planet.

But then she was also saying how if you capture that light right as it gets through the ring around that star, you have light that passes through that planet’s atmosphere, we can look at that atmosphere, you put it out on the spectrum and you can start figuring out, “Oh, this is mainly made of oxygen, nitrogen. This is more carbon,” or whatever the molecules or the properties. So you can really start to understand exoplanets and understand what they’re making up. And I immediately think of an analog of a planet to compare it to, is the one we’re sitting on.

So as much as we can understand what this Earth is doing, and what it’s made of, then that helps us be able to compare it to those exoplanets so we can confirm, are they really Earth-like planets out there? And we know that because of what we know about Earth.

Adam Moreno: Yeah, no, that’s right. That’s another reason why NASA does Earth science. Because we need to understand what’s happening on Earth to be able to understand the other planets. And actually one of the great things about being an early-career scientist here at NASA, is we have this early-career network here. Part of that is we take tours of different places around Silicon Valley, we also have these happy hours and stuff. So during the happy hours I’ve been able to talk with other scientists who are not Earth scientists, you know, they’re studying planets or whatever.

Host: This is the newcomers to NASA.

Adam Moreno: Yeah, and they’re doing the cutting-edge stuff. There is a lot of overlap. There are a lot of people who are doing spectroscopy on other planets, and the same methods that they use are derived maybe from the methods that were developed first to study the Earth, actually.

Host: And I think some of your work, talking about the supercomputers and software, it’s also one of those natural ways to kind of pull things together. It seems like, whereas stuff may have been silo-ed. Like okay, you have Earth science over here, you have biosciences in one area, astrophysicists, or even aeronautics, but it seems like a commonality between a lot of these is becoming computing. Because everybody who is doing research, the supercomputer can help make better sense of the data.

Adam Moreno: Yeah, that’s right. Getting data from one of the satellites about Earth, that’s super cool, and you can learn a lot just by looking at that data. But in Earth science, often what we do is we take data from all sorts of different sources. Maybe several different satellites.

Host: Okay. They have different instruments on them. They’re looking at different things.

Adam Moreno: Right, they’re looking at different things. Maybe one is looking at the oceans, one is looking at the forests, atmospheres, cryosphere, so on and so forth. But then we also couple that with data on the ground. So in my case, I study forests, so I’ll take forest inventory data where people are on the ground measuring forests, along with weather station data, and soil data, and then we try to put that all together to make something even greater than the sum of its parts. Through maybe a big computer model that we need to supercomputer to be able to crunch all of those numbers to get a good picture of what’s happening on a continent, or on a global scale, things like that.

Host: So talk a little bit about the forests. Especially having lived in Paraguay for a while, in a heavily forested area. What happens there can affect the entire planet in a lot of times, or the areas around it. So talk a little bit about how the forest in general play into the global –

Adam Moreno: Right, yeah. Why is a forest ecologist at NASA?

Host: That’s a good question.

Adam Moreno: Yeah. There are several cycles that make up the global system. There’s the water cycle, the carbon cycle, energy cycle, nitrogen cycle, all these different cycles. Things that are moving around the Earth and make up really the system that we call the Earth. So if we’re talking about something like the carbon cycle, global forests make up the largest terrestrial component of the carbon cycle. Half of the carbon that goes into the atmosphere gets sucked down by forests, so large-scale changes to the forests can change the carbon cycle, that could then change the climate cycle and have repercussions all over the world. Same with the water cycle and all of that sort of stuff.

Host: Everything kind of interacts, and ebbs and flows.

Adam Moreno: Everything interacts, yeah.

Host: Like the food chain, to a certain extent. You tweak one thing and then other stuff happens, even if you don’t understand it right away.

Adam Moreno: That right.

Host: Things are all interconnected and interwoven, very complex.

Adam Moreno: Yep. Right, so I kind of say, well, a computer is a very complex system. It takes a lot of engineering and science and math to understand or build this sort of system. But the actual biggest and most complex system on Earth, is the Earth itself. And we didn’t engineer it, so that makes it even harder. So actually, we’re just trying to figure out how this system works. There’s still a lot of work to be done. My little component of figuring out this system is figuring out how atmospheres and forests interact with one another.

Host: Okay. So you talk about studying forests, doing that research, and then playing that into being a computer scientist. This is an option for all computer scientists, “Oh, you could also go study forests.”

Adam Moreno: That’s absolutely true. And we need more people with technical abilities to come and be able to make sense of this huge amount of complex data that we have.

Host: How do you do that? What do you? What is your day-to-day? Are you running this data through new software, through databases, putting it, visualizing it in different ways where you come up with newer ideas instead of it just being numbers on a spreadsheet? Talk about that.

Adam Moreno: Yeah, so my day-to-day is getting data from all different sources, satellites or the ground, and then transforming that data into something I can use. Maybe it’s a map of a particular area, like the United States that I want to study. And getting all of those datasets to match up, right? They need to be able to be comparable and usable no matter if they came from satellites of from the ground, or what have you.

Then I write a lot of code that crunches those numbers in the supercomputer. Maybe I’ll have a forest model that will get productivity estimates of the forest, given its reflectance from the satellite, given its climate, given the type of forest there is on the ground, and we want to know, “Well, how productive is vegetation all over the globe?” for example. That sort of stuff goes through the computer.

My ultimate goal right now, while I’m here at NASA, that kind of got me here, is I’m hoping to develop an early warning system where we can monitor all of the forests throughout the United States, and pinpoint forests that are extremely vulnerable to a large-scale mortality event, whether it’s from fire, or beetles, or even just mortality without any sort of agent of death like fire or beetles. And also understanding why those forests are vulnerable, and then giving some recommendation of what can be done to maybe prevent this large-scale mortality event.

So, to do that, what I’m doing is, I’m taking all of this data that I’m getting from the satellite, from the ground, and I’m writing code that will hopefully pull out equations, essentially, of the physics on how climate limits forests.

Host: Oh, interesting, okay. What are the rules?

Adam Moreno: Yeah, what are the rules, exactly. What are the rules? Because it’s obvious, climate dictates what sort of forest can be where, right? So, we have a different sort of forest here in California, in Silicon Valley, than we have up in the Pacific Northwest, in Washington or Oregon, and that’s because the climate is different. The climate dictates all sorts of things within the forest. It dictates how tall the forest can get, how big the trees can get, how many trees can be in a forest, and things like that.

But, as of yet, we don’t really understand how it limits all of those, or what this curve looks like, depending on your temperature and precipitation, what sort of trees could exist in an area. So I’m trying pull out these biophysical limitations that climate places on forests that I can then use to develop some sort of system that can recognize when a forest is out of balance with its climate system.

Host: You said it a handful of times, but talk a little bit about writing the code. Because it’s not like – I’ll probably date myself – it’s not like you’re opening up Delphi. It’s not like you’re just powering up your PC and just writing this stuff, because a supercomputer, I’m imagining it does not run on a Microsoft or an iOS system.

Adam Moreno: No, that’s right.

Host: What all goes into that?

Adam Moreno: Yeah, so the supercomputer we have here at Ames is called Pleiades, and it has its own building that takes up this whole entire – it’s like a city block essentially. It even has its own cooling tower, it’s so big. Why do we have supercomputers? Well, number one, all of the data that we need to understand the global system could not fit on a smaller computer, on a desktop computer. But then also the computations that it takes, if we were to try to do it on a desktop computer, it would take three years or something like that.

Host: For it to render –

Adam Moreno: Yeah, or just to run all of the calculations on all of the data. So essentially it’s impossible on a normal computer. But then if you write it correctly and you utilize the architecture of the supercomputer, then maybe it will take a day, or something like that. So, essentially a supercomputer makes impossible calculations possible.

Host: Oh wow.

Adam Moreno: Yeah, so it’s not just normal programming when you’re programming on a supercomputer.

Host: Yeah, HTML, you’re screwing around.

Adam Moreno: Yeah, right. So often, I think, scientists who have not worked with supercomputers, they will think, “Oh, okay, I have this little model, and it works well on studying, let’s say, forests, and it works well on like a watershed, or just one particular forest. Well, shoot, if I just have access to the supercomputer, I’ll just run it on the supercomputer and boom.”

Host: Do your supercomputing magic.

Adam Moreno: Yeah, and then boom, I have a whole global simulation. But that is not how it works. I guess I kind of liken it to if you’re chopping wood and you’re stacking wood, okay?

Host: Okay.

Adam Moreno: If you have a little bit of wood you can do it yourself and it’s fine. You chop wood, you stack wood. But now let’s say I give you a huge, huge warehouse full of wood and I tell you I want one big stack of wood. And I say, but I’ll also give you as many people as you want. I’ll give you a million people if you want a million people. Well, okay, if you just say, everyone cut and stack wood –

Host: Everyone start, like brute force. Let’s start chopping wood.

Adam Moreno: Right, you don’t know where the wood’s going. Everyone cuts wood in different ways, whatever. It doesn’t work. Now all of a sudden you need to organize the workers. You need to put some of the workers chopping wood, some of the workers moving wood, some of the workers stacking wood, and you need somebody organizing the stack of wood. Things like that.

Host: Let alone lunch breaks and payroll.

Adam Moreno: Yeah, those sorts of things. Anyways, the point is that the end goal is the same, cut wood, stack wood. But since the scale is so different, and you have so many workers, the whole process chain is completely different. And it’s the same with supercomputing. You have a small process and the end result is the same but just on a bigger scale, but getting from A to B is a completely different story.

Host: Excellent. So you’re working as the fellow. You finished your Ph.D., right?

Adam Moreno: Yeah, I finished my Ph.D.

Host: So now this is the postdoc.

Adam Moreno: This is the postdoc, right.

Host: What do you see in the future? What are you hoping to work on down the road?

Adam Moreno: Well, this early warning system is a big dream that I have that will take many years to complete. I hope to either stay on here at NASA, or go on and become a professor or something. I like teaching as well, I like mentoring, so I don’t know, we’ll see. I think that’s part of the postdoc experience, is really understanding where you want to go with your next steps of your career.

Host: So, for folks who are listening who have questions for Adam, we are on Twitter @NASAAmes, and we’re using the hashtag #NASASiliconValley. So, questions come in, we’ll loop them on over to you and see if you can get some answers back and forth.

Adam Moreno: Yeah, that would be great.

Host: Thanks for coming on over.

Adam Moreno: All right, thanks.

[END]