Eyes on Earth Episode 112 - Landsat in Popular Media

[00:00] If you're a producer, if you're a writer, and you go out on the Internet, it's pretty easy to find Landsat. It's pretty easy to understand at a very basic level what it does, and I think they just probably drew some inspiration from... [00:13] some sort of basic research and then just decided to go with it. Sounds cool, landsat. [00:19] and I think it fit the purposes of the plot. They just took it and ran with it and turned it into something much bigger. [00:30] Hello, everyone, and welcome to this special episode of Eyes on Earth, a podcast produced at the USGS Eros Center.

[00:39] As the saying goes, all publicity is good publicity. But when our Landsat satellites, or Eros, are mentioned in popular media, [00:47] movies and books don't always get it right. [00:49] which is a bit unnerving for a center that's focused on science. [00:52] Thank you. [00:53] So today we're correcting the record and having a bit of fun with it too. [00:58] My name is Sherry Levese. And I'm Jane Lawson. And we're hosting today's episode. [01:03] I'll be sharing the fiction. [01:04] And I'll be the fact checker and explain how Landsat data are used in real life.

Let's get started with Keith Masbach. [01:11] who was the inspiration for this podcast. [01:13] Keith, who serves on the Department of the Interior's Landsat Advisory Group, hosted a video looking at how satellites are depicted in films. [01:22] Welcome, Keith. Hi, Shari. Hi, Jane. You have a very long resume. [01:28] Can you summarize some of your background, especially your connections to Landsat? [01:33] I was very fortunate to spend time [01:36] as a military officer in the Army. [01:39] And that's where I first got introduced to remote sensing and uses of remote sensing various other missions.

And then I did that as a defense officer. [01:48] an intelligence executive as well and got even more involved. [01:51] And it was really during my later time running a nonprofit associated with the industry that I was appointed to the National Geospatial Advisory Committee. [02:01] The Landsat Advisory Group falls under that and then ultimately chaired that committee during my tenure. And that's when I really got exposed to Landsat and a lot of the civil applications. [02:14] more of both our classified remote sensing systems and systems like Landsat. [02:20] In your video, you had something to say about Landsat in the movie Battleship.

Explain how the depiction of Landsat wasn't quite spot on. [02:29] Yeah, I was very fortunate. I'm on the advisory committee of the International Spy Museum. [02:34] occasionally collaborate with Wired magazine. And it was just so fun to do. They asked if we would take a look at depictions, and [02:42] remote sensing, spy gear, [02:46] and say, hey, look, this is what looks real, this is what looks phantasmagorical, [02:51] In that particular scene that I talk about, they are using the Landsat system as a deep space communication system, which, you know, ultimately, I hate to give away the plot of this super well-panned movie.

[03:07] They use the Landsat deep space system to communicate out to a planet that they believe in. [03:13] could sustain life similar to earth. So, [03:16] They've really just taken Landsat and turned it into a communications satellite for deep space. [03:22] So is that really... [03:24] possible in real life. [03:26] No, that's just simply not what Landsat is used for. And one of the more... [03:31] comical pieces of that entire video I was able to do. [03:35] And also, [03:36] Landsat isn't deep space, right? [03:39] No, by no means is the deep space.

When we think of remote sensing, you know, we tend to think of more of a low Earth orbit. They are the satellites that tend to be closest to Earth because of the nature of what they do. [03:52] and the physics of the size of optics. [03:56] and what they need to discern on the Earth. So you want to have them flying nice and as low as they can without incurring too much drag and getting pulled back into the Earth. [04:07] Okay, we know that Battleship isn't 100% accurate about how Landsat can help in battle situations, but in real life, Eros has supported military missions.

[04:18] Longtime Aeros employee Ron Ristey has some familiarity with this. Hi Ron, please share a bit about your years at Aeros. [04:24] My years at Arrows. I retired after 47 plus years. Started out here as a research analyst, and from there I went into supervisor for customer services, and then went over into the science department and did a lot of work in support of military operations, emergency operation coordinator, working with people within the center that had their own military operations. [04:54] standpoint of what we did is we provided support to different organizations around the country as the troops went into Iraq or Afghanistan, military support and mapping particularly those countries before our troops went in.

As we expanded on the use of Landsat data for the military, it was more used for mapping of the lay of the land before our troops went in. So [05:24] when they were trying to find the different terrorist activities, and then we would provide that Landsat information to them so that they knew where there were feet on the ground. It allowed them to be able to look at mountainous areas or if there may be any potential cave areas that they had to worry about or stuff like that. And the other thing is that when we were doing work in Baghdad, [05:54] We sent people over there to support their operation and trying to help them build up their own facility and use of Landsat data.

One of the things using the Landsat data to map Afghanistan or Iraq is that if you saw it in the newspaper or heard about it on the news, we had already provided information to the support of the troops six to nine months ahead of time. [06:24] discussion at the time that they would, as our troops come in, that they would blow up the dam of the Tigris River and flood it. And of course, we needed to do what we call a workaround. And one of the things that our hydrologists here at the center worked with our scientists and looking at what was the depth of the water, how much water was contained behind the dam.

And if the troops did go in and they blew it up, how concerned would it be for them to be, [06:54] in danger of the area. And so we were able to determine that if they did blow the dam up when our troops were in that area, the water level would only be about ankle deep. So there was no substantial threat to our troops as they were going through the Iraq area. So Landsat has been used to support military operations, just not in deep space. So Keith, another movie that misses [07:24] Some of the characters in the movie are part of a corporation called Landsat, complete with matching Landsat jumpsuits with their own logo, who used satellites to help find Skull Island.

Do you have any commentary about that? [07:41] that Landsat has turned in, to your point, to an entire corporation. It's not just a satellite, but [07:47] It's people talking about applications of the downstream uses [07:51] how they're using it to track down Skull Island and deal with... [07:55] with Kong, again, another super critically acclaimed movie, not necessarily at my alley. [08:02] I think the issue there is if you're a producer, if you're a writer, [08:10] And you go out on the Internet, it's pretty easy to find Landsat. It's pretty easy to understand at a very basic level what it does.

And I think they just probably drew some inspiration from... [08:20] you know, some sort of basic research and then just decided to go with it. Sounds cool, Landsat. [08:26] And I think it fit the purposes of the plot. They just took it and ran with it and turned it into something much bigger. Is there any idea that you think that actually movies kind of get right about the way that satellites are used and in particular Landsat? What I'd say is, you know, commercial remote sensing. [08:46] That's come a long way since the earliest days of Landsat as a government system.

And the amount of innovation that's going on in the commercial space around all kinds of phenomenologies from electro-optical technology, [08:58] It's a synthetic aperture radar. [09:00] to infrared and soon LIDAR. [09:04] I mean that there is this rich amount of data and information available out there. [09:09] we have this chronicle of the earth and we talk about [09:13] climate change and what it's doing to the earth and how the earth is changing under those stresses. And because we have this story, because we have [09:23] the ability to have a slider bar back into time.

[09:27] This is an objective... [09:29] of where we were, where we are, and gives us the data really to understand and analyze and think about where we're headed into the future. [09:39] And that investment by the United States government over time just has a remarkable return. [09:47] for everyone on this planet. [09:50] So Keith, thank you so much for your time. [09:52] There is a real-life crossover between the fictional Landsat Corporation and NASA, which builds and launches the satellites before handing them over to USGS. [10:01] Mark Evan Jackson, the actor who played Landsat Steve in Kong Skull Island, helped publicize the Landsat 9 launch in 2021 at Vandenberg Space Force Base in California.

[10:13] He also narrated a series of videos by NASA about the real Landsat program, [10:18] called Earth Matters, Landsat Continuing the Legacy. [10:22] Another note, the movie Kongskill Island is set in 1973, soon after the launch of the first Landsat. [10:30] Here's another point where the movie mirrors real life. [10:33] In 1973, Landsat really did discover a new island off the shore of Canada. [10:38] It was first named ERTS Island, or ERTS, after the original acronym of the Earth Resources Technology Satellite, and then changed to Landsat Island in 1979 to reflect the satellite's 1975 name change.

[10:56] Moving on to popular literature, there are actually two Michael Creighton books that became movies in which Landsat and even Eros play a role. [11:08] Congo. [11:09] In The Lost World, written in 1995, the bad guys use heat signatures to find nesting sites for dinosaurs so they can steal the eggs. [11:18] Here's a quote of the conversation between two characters in the book. [11:22] I get it, King said. These red circles are infrared signatures. [11:27] Yes, Dodgson said, big animals leave big signatures. I got all the satellite flybys over this island for the past few years and mapped the location of heat sources.

And the locations overlapped from pass to pass, which is what makes these red concentric marks, meaning that the animals have to be located in these particular places. [11:47] Why? [11:48] He turned to King, because these are the nesting sites. [11:52] To vet this idea, we are turning to George Shen, who knows a lot about Landsat's thermal capabilities. George, tell us a little bit about your background at Eros. [12:01] My background at Aeros, actually, I use the remote sensing data to study land cover and land use change in the S. states.

And so that's many. The data I've mainly used is from the Landsat. And in recent years, we're focusing on using Landsat thermal data to study the land surface thermal features, especially in the urban area. [12:31] Thank you. [12:31] So is this idea of finding animals, even very large animals like dinosaurs, using Landsat thermal data at all realistic? [12:40] I think for the large animals, usually those animals can produce certain [12:48] footprint, and it should be large enough, we can track. And so if that happens, for example, if you change the land surface features, a lot of time you're going to change the thermal features.

And definitely thermal remote sensing data can capture that kind of a change. [13:09] That sure surprises me. I thought that maybe Michael Crichton was making this up. But you're saying that if the animals are perhaps located in one spot frequently, do you have an idea of how big it would have to be in order for it to be picked up by Landsat? [13:28] Let's have the spatial resolution, thermal resolution that's varied from the 100 meter to 60 meter. But when we use that data, we convert that one into the 30 meter resolution.

That means if anything is large enough in that pixel, the pixel can capture that one. But for the thermal, that's different from the visible. [13:58] the signature in that pixel. But a thermal one is, even though the object is not large enough in that pixel, but the thermal feature will be changed. And then if you compare with the pixel by pixel, definitely you're going to see the difference, because the animal in that pixel, we're going to generate... [14:23] the different thermal features than the surrounding area without animals. [14:29] And that sort of relates to the urban heat island work that you do.

Exactly. Exactly. [14:35] What does the temperature difference have to be to notice it? Good question. Yeah. That's depending on if, for example, if animals stay in that pixel without much vegetation, you definitely can see the difference between the long vegetated animal controlled pixel and the vegetated pixel. [15:05] But if that is mixed with animal and vegetation, sometimes it's a little bit difficult to separate them. [15:14] But if you're lucky, when you take the photos, you take the image, and the time is the animal staying in that area, then you can definitely see the difference between those pixels.

And a lot of the time, you know, visible image, you cannot see clearly, but you definitely can see the signature from the thermal. [15:40] All right, Jane. [15:44] it was seeing dinosaur heat signatures from space. Unbelievable. Unfortunately, there are no real life dinosaurs to test this theory on, only fossils. But believe it or not, Landsat data actually can be used to find fossils. [15:59] both from the time of the dinosaurs and more recent eras. Here to describe his work is Robert Anamone, a paleontologist at the University of North Carolina, Greensboro.

Bob, welcome to Eyes on Earth. What made you start thinking about using Landsat data for paleontological purposes? [16:17] You know, the ways that paleontologists look for fossils really haven't changed all that much since the early days of the field in the 19th century. We gather a team, we choose a place where we know or suspect that the right kind of sedimentary rocks of a particular age are exposed at the Earth's surface, and then we walk long distances with our eyes on the ground searching for evidence of fossils that are eroding right out of the surrounding sediments.

Since the places where fossils are found tend to be remote and are often vast in extent, fieldwork can be expensive and time-consuming. [16:47] The Great Divide Basin of southwestern Wyoming, where I've been working since the 1990s, has an area of about 10,000 square kilometers. My team experienced work. [16:57] serendipity in [redacted address]. We were driving to some established localities that we hadn't visited for a few years, and we took a wrong turn on a two-track road that led us to a series of sandstones that we had never seen before. So I decided to park our vehicles and take a look for fossils.

Within about an hour of crawling around and up and down these sandstones, we realized we had stumbled upon one of the richest early Eocene fossil sites in the American West. It's about 50 [17:26] This one site has yielded 4,500 of a total of 13,000 fossils that we found in the Great Divide Basin since the 90s. So it was this serendipitous experience, really, that made me start to think seriously about developing predictive models for the location of fossil sites. [17:43] So we use, you know, thanks to Landsat, we use remote sensing technology and artificial intelligence, or if you prefer, machine learning approaches.

And what we're doing is searching for outcrops of sedimentary rock that are likely to yield fossils. So we're not using the technology to actually find the fossils. We still have to do that by surveying on foot. The trick really is choosing which outcrops of rock are most likely to yield fossils. And that's where Landsat and machine learning comes into the equation. [18:11] So I've worked in collaboration with paleontologist Glenn Conroy of Washington University and geographer Jay Emerson of Western Michigan University. And together we've developed and field tested or ground truthed several predictive models for locating mammal fossils from the Eocene in the Great Divide Basin in Wyoming.

So we start with Landsat imagery of this large basin. It straddles the continental divide in the southwestern part of the great state of Wyoming. [18:41] Yeah. [18:53] And we look at this all in a GIS sort of software environment. So in our initial published model, which I like to talk about, we used an approach to the satellite imagery that's known as supervised classification. We use this to identify a number of different land cover types, for example, forest, wetlands, sand dunes, grasslands, but also including previously identified fossil localities. [19:23] electromagnetic radiation from the Landsat 7 ETM Plus sensor.

[19:27] that we got from you guys, again, spanning the visible to the infrared parts of the spectrum. So the neural networks that are churns and figures out how to distinguish between these different land covers, and then it classifies... [19:40] every pixel in the entire basin, all 10,000 square kilometers. So basically it's this multivariate spectral signature that distinguishes between sandstone exposures or mudstone exposures that have a higher probability of having fossils and that then become our sort of high-priority objects for prospecting. [20:03] This has completely changed our approach to fieldwork, our sort of workflow in the field.

Now we use these maps to guide our day-to-day fieldwork. We only survey in places where the map tells us there's rock that looks very similar to other places where we've successfully found fossils in this basin. And our ground truthing surveys indicate that this really works. [20:33] as sort of high priority. And what this does in effect is it kind of reduces the aerial extent of what we have to prospect by an order of magnitude or more. I mean, rather than feeling like we have to cover 10,000 square kilometers, this map now tells us which places we should really focus on, which makes our limited time in the field, we go back to the field for two to three weeks every summer.

It makes us more efficient, more effective, and maybe more likely to turn up the next really [21:03] discovered in [redacted address]. So we can use Landsat to find fossils, but can Landsat data tell the difference between ancient jungle and more recent growth? In Congo, the scientists in search of diamond mines use reflectance to identify really old growth forests versus forests that's only 500 years old. [21:27] Here's the relevant quote from the book. Albedo was technically defined as the ratio of electromagnetic energy reflected by a surface to the amount of energy incident upon it.

[21:38] If there was a lost city, what signature might appear in the vegetation? Late secondary jungle. The untouched or virgin rainforest was called primary jungle, huge hardwood trees. However, if the primary jungle was cleared by man and later abandoned, an entirely different secondary growth took over. The dominant plants were softwoods and fast-growing trees. [22:08] and it could be graded by age. [22:10] On this topic, we're tapping the expertise of Chris Barber, who is a USGS research physical scientist at Eros. Chris, are remote sensing scientists truly able to differentiate between different levels of growth in forests or jungles in a way similar to the plot of Congo?

[22:28] Well, yes and no. It's actually a good problem. And most of the problem described in Congo is spot on, this idea that there's primary forest and secondary forest. And after human disturbance, it's kind of this fast growing secondary forest that comes in. [22:58] to that real kind of mature primary forest. And we do have techniques in road science and that we can tell the difference between that young forest and kind of 20, 30, 50 years, up to 100 years, 50 years or 100 years or more. But when you start to get out to 300 years or 500 years or more than that, then that's just fiction.

We're looking at a much smaller timescale for what is reality. [23:21] But yeah, the idea that we could tell 300 from 500 a year, that's just – no. 20 from 50 maybe. [23:29] To wrap things up, let's call attention to Crichton's two shout-outs to Landsat and Eros in Congo. One, the diamond-seeking company is called Earth Resources Technology Services,, or ERTS. ERTS, as we mentioned earlier, was the original abbreviation for the Landsat satellite mission. And two, the epilogue for the book actually sends a major character to Eros, known as the Eros Data Center, or EDC at the time.

[23:55] The book reads, on October 23, this would be 1979, Karen Ross resigned from Ertz to work for the S. Geological Survey, EDC, in Sioux Falls, South Dakota, where no military work was conducted and no field work was possible. She has since married John Bellingham, a scientist at EDC. [24:14] Chris, we love it that they mentioned the EDC, but is it quite accurate that no fieldwork was possible? No, that's complete fiction. We have people at what is now known as Eros doing fieldwork right outside our offices in South Dakota and throughout the state and all over the

S. and internationally for wildfire impacts and food security and land cover and land cover change. So we do fieldwork all over the United States and all over the world today. [24:44] We'd like to thank all of our guests for participating in this podcast separating science fact from science fiction. And thank you to the listeners for joining us for this episode of Eyes on Earth. Check out our social media accounts to watch for our newest episodes. You can also subscribe to us on Apple Podcasts. [25:04] This podcast. [25:09] Is a product of the

S. Geological Survey Department of Interior. [25:14] Thank you.