Skip to main content
U.S. flag

An official website of the United States government

Eyes on Earth Episode 128 – 2024 EROS Fall Poster Session

Right-click and save to download

Detailed Description

In this episode of Eyes on Earth, we mingle at the 2024 EROS Fall Poster Session. A poster session is essentially a way for scientists to share their work with their colleagues in a public forum. About 30 posters were on display in the EROS atrium from EROS staff and several students from South Dakota State University and the University of South Dakota. We talked to a few of them to get quick summaries of their research.

Details

Episode:
128
Length:
00:29:39

Sources/Usage

Public Domain.

Transcript

TOM ADAMSON:
Hello everyone, and welcome to another episode of Eyes on Earth, a podcast produced at the USGS EROS Center. Our podcast focuses on our ever changing planet and on the people here at EROS and across the globe who use remote sensing to monitor and study the health of Earth. My name is Tom Adamson.

For today's episode, we mingle at the 2024 EROS Fall Poster Session. A poster session is essentially a way for scientists to share their work with their colleagues in a public forum. About 30 posters were on display in the EROS atrium from EROS staff and several students from South Dakota State University and the University of South Dakota. The posters covered a wide range of topics. Let's start with some of the students.

Locally in South Dakota, invasive species is an area of concern. Two students are studying invasive carp in water bodies in the region, and another student is measuring water loss and gain across the Prairie Pothole Region.

Go ahead and introduce yourself.

NATALIE LIBERATI:
My name is Natalie Liberati. I'm a graduate student at South Dakota State University in the natural resource management department. I'm going to be looking at the Landsat dynamic surface water extent data to evaluate flooding and how that could facilitate carp movement into adjacent watersheds. So currently, the James River in South Dakota is invaded with carp from the Missouri River. We're looking at which connections that flooding is creating, which ones are most at risk for carp invasion. So a couple of my images here. I looked at the discharge data from the James River from a specific gauge. And we have data ranging from no flooding at all to high flooding events. And I have an image here showing how these specific wetlands that are adjacent to the river in a non flood year, they're not connected. When it does flood, the water from the river then spills over and connects these wetlands to the river that are usually isolated. So they have not been subject to carp invasion until now. They could potentially become invaded.

ADAMSON:
Can you explain a little bit more what dynamic surface water extent data is?

LIBERATI:
So the DSWE, or dynamic surface water extent data, is satellite imagery of surface water. It's just showing me where water is present on the landscape, and I can distinguish them from what is a permanent water body versus what could be flood water using different water body polygons as well, from the USGS website and the National Hydrography Dataset. So it's showing you where surface water is present and then you can take that data and manipulate it to figure out what's flood water and what's not flood water. So we're using the dynamic surface water extent data to do that in ArcGIS. And we're looking at multiple metrics of flood water connection. So the surface area of each connection, how many habitats are being connected to the James River, the type of habitat being connected, so whether it's a lake, a pond, stream, a wetland, and where these habitats specifically are in the watershed.

ADAMSON:
Well, that makes sense. Do you have any idea how someone could use this data?

LIBERATI:
Yes. So we're hoping that with the results I come up with, this would give managers the ability to look at these connections and see which ones are more important to block carp movement with. So if there's like a really prominent permanent wetland that's becoming connected to the river, they might be able to use deterrents to block carp movement into that wetland that's usually isolated, especially if it's really good habitat for them.

ADAMSON:
Well, thank you.

LIBERATI:
Thank you.

ADAMSON:
Go ahead and introduce yourself for us.

BRENDEN ELWER:
I'm Brenden Elwer. I'm a master's student at South Dakota State University.

ADAMSON:
Okay. Sounds good. Can you summarize your study for us here?

ELWER:
Yeah, absolutely. So I'm working with using habitat quality as a tool for informing invasive species management. So in North and South Dakota, two invasive species are threatening lots of water bodies are the invasive silver and bighead carp. They compete with native species like walleye, perch, and potentially even up to species like waterfowl through food competition. My work is focusing on informing preventative actions. The tool we're using for that is habitat quality and looking at the way that that might shift through space and time, as most of these invasive individuals are going to spend spend their time in the highest quality habitat. So knowing where that quality habitat is at can make these preventative actions more efficient and more effective at detecting individuals as they become invaded into these new water bodies.

So to do that, we're going to be using an energetics based model. So it takes food consumption, so energy the fish consume, versus what they expend based on temperature, flow, and other environmental characteristics. So I actually went and physically collected temperature, Chlorophyll A, which is a surrogate for phytoplankton, zooplankton. And those are two food resources for these silver and bighead carp as well as velocity in the rivers I sampled. And so using that we can predict how well the fish will survive, grow, reproduce, using this energetics based model. However, I can't collect hundreds if not dozens of samples on some of these water bodies. So using remote sensing data, we're going to try to make this on a very fine scale. So using Landsat trying to get this on a 30 by 30 meter scale. So using temperature from Landsat we can divide our lake into, or wetland or river, into these 30 by 30 meter grids that we can assign a temperature to. And then the last one is looking at that Chlorophyll A, looking at reflectance values, whether it be from just the Landsat surface reflectance or we're going to explore the aquatic reflectance product that's being developed, or in the provisional stages, to try to predict Chlorophyll A using regression based off satellite and what we physically observed to assign those values. We can run our energetics model for each individual cell, that based on that output, we can map out habitat quality to see how those patches of quality habitat, if there are patches, shift through space and time to inform these preventative actions for managing agencies.

ADAMSON:
Okay, that sounds really good. You can't use your ground studies or satellite by themselves. What's the real value in using both combined?

ELWER:
So the both combined there, it just it helps us-- it's helping us get that finer resolution where taking five samples out of a 6,000 acre lake versus upwards of 50,000 for however many it comes to, you can just refine that spatial scale down and make it add more detail to what we're trying to do.

ADAMSON:
Thank you for doing this.

Okay. Go ahead and introduce yourself.

MADISON DEJARLAIS:
Hi, I'm Madison DeJarlais, and this is my preliminary time series analysis of surface water change in the Prairie Pothole Region.

ADAMSON:
Describe really quick what the Prairie Pothole Region is. I mean, pothole sounds like a funny term. So what's going on here?

DEJARLAIS:
Okay, so the Prairie Pothole Region is one giant, they do call it kind of a wetland complex, but it is overall hydrology. So there's a lot of lakes, rivers, streams, all these things. And it's from glaciation. So it is a direct byproduct of the glacial period that we have. And all of these wetlands, all of these bodies of water are actually kind of connected and feed off of each other in this huge area which crosses Manitoba, Saskatchewan, Alberta, North Dakota, even into Montana and Minnesota, Iowa. So this is the beginning steps of my actual thesis project, in which I am analyzing the Prairie Pothole Region to try and see what surface water change is doing overall using Landsat. So I'm comparing, in this specific instance, '84 to 2023 to try and get those bookend dates to see if there is a trend that's identifiable. And there does appear to be one. However, it is important to note that it is the end dates, so there could be other things kind of impacting this trend, whether it's the actual trend itself or not. So far, we're able to see that there is a loss happening on the Manitoba end, as well as the Iowa
end of the Prairie Pothole region, and we're seeing gain within Manitoba and the Dakotas, which this does actually track with a lot of climate data, which I have in my literature, of where we're seeing snowfall increases and loss of precipitation otherwise. And I'm using Google Earth Engine and the modified normalized difference water index--MNDWI. And that's just kind of seems to be the best fit for identifying water overall with the most consistency.

ADAMSON:
Yeah, this is a pretty large area. So it's a lot of Landsat data. How do you manage that?

DEJARLAIS:
Google Earth Engine is very helpful. It is very, very, very helpful. I was able to create a cloud-free composite over the growing season, which then I apply that MNDWI. And so I get my binary, which can be exported as one, kind of, image.

ADAMSON:
The image looks really neat. You know, we can't see that on audio here. But and you've zoomed in to some specific areas to show those water gains, water losses. So it's really nice to see it at that level.

DEJARLAIS:
Thank you.

ADAMSON:
What are the next steps?

DEJARLAIS:
So the next step for this thesis project is actually going to be looking at it on a pixel by pixel basis. So I have an R package that's able to track the trend rather than looking at the bookend dates. So from year to year we'll be able to see if it's consistent gain, consistent loss, or if there is a trend to it where it's oscillating every five years, or so on and so forth. So that'll be the next step is looking from year to year.

ADAMSON:
That's even more data to work with. This isn't just 1984 and 2023. Everything in between now.

DEJARLAIS:
Everything in between. Yep. Going all the way through it year to year.

ADAMSON:
That's going to be great. Thank you.

DEJARLAIS:
Thank you very much.

ADAMSON:
Another student has his eyes on a different part of the world. A University of South Dakota student is mapping grassland over Mongolia using cloud computing, because he is using 34 years worth of Landsat data. His study can help keep local livelihoods sustainable.

Go ahead and introduce yourself for us.

ABHINAV CHANDEL:
Hi, I'm Abhinav. I am from University of South Dakota. I am working in sustainability program. So my project here is on analyzing grasslands. So currently I'm focusing on Mongolian grasslands and trying to analyze biomass and canopy cover for a time series from 1993 till 2023. So I'm currently using some machine learning models and a Google Earth Engine platform, which is cloud computing, helping me to play around with huge amount of Landsat data. So my whole area is getting covered by around 130 Landsat tiles. So I'm just using Google Earth Engine to process all this data.

ADAMSON:
It's a lot of data that you have to work with.

CHANDEL:
Yeah. So like if I have to do it in a traditional way, so I have to download all this Landsat tiles. And each tile goes for like, one GBs to like two GBs. So getting that much amount of data for a single time stamp will be huge, because I'm working for 34 years. So that will be something humongous.

ADAMSON:
What makes Landsat ideal for this large of an area?

CHANDEL:
The main idea was because since last ten years, there's been a lot of extreme weather events which are happening. So due to that in last year itself, 40 to 50% of the livestock in this area died. And in grasslands, we know like lot of things are depending on grazing, and people earn a lot of from that--only that is the only way they have their bread and butter from. So that's where, since 1990 gave me an ideal way like up till 30 years how things have changed. And then what's the reason behind this last ten years where we are experiencing a lot of extreme events.

ADAMSON:
Can you describe real quick what aboveground biomass is exactly?

CHANDEL:
When we talk about green covers, it's either trees, grasses, or croplands. So croplands, from the sowing phase till the harvesting phase, the biomass keeps on increasing. So it's the amount of carbon they store in. And that's like really feeds into the mass which they have. And when we accumulate that whole number it becomes aboveground biomass.

ADAMSON:
Do you have a good idea of who will find this valuable? Who will use this?

CHANDEL:
So I think the the government obviously, first of all, and some kind of NGOs are also working in this area who want to like improve this scenario, like where grassland is being decreased over the years. So they can find this useful to identify which areas they can put in their efforts to plant more and, like, revive the green cover so that the livelihood of the people over there still stays the way it was since they, traditionally how it was, they can restore all that stuff.

ADAMSON:
Keep that sustainable. Okay. Thank you for your time.

Now we'll turn to EROS staff and see what they're working on. The USGS EROS recently released a new generation of land cover mapping in the Annual NLCD products. NLCD is the National Land Cover Database. A big part of that is making sure that land cover data is accurate.

Go ahead and introduce yourself.

JO HORTON:
My name is Jo Horton. I'm a land remote sensing scientist at the USGS EROS, and I'm working on the Annual NLCD project.

ADAMSON:
Go ahead and summarize what you have on your poster here.

HORTON:
Oh, my poster is focusing on the part of the work that I do for Annual NLCD, which is collecting the reference dataset that'll be used to validate the land cover and land cover change products. So basically it covers the collection of data. We're collecting up to 10,000 points across the entirety of CONUS, which is the lower 48. And we're collecting annual data for all the land cover, land use, and land change processes from 1984 through 2023. And then that data that we collect, which is considered like ground truth, will be compared to the Annual NLCD products at those same points, at the same locations, and the same time to see how they compare and see how accurate the Annual NLCD land cover products are.

ADAMSON:
It's really important that NLCD be accurate. So this is a big deal, right?

HORTON:
This is a very big deal. NLCD has been the standard for land cover mapping in the United States for decades at this point. And this new product with the annual time series, we really want to make sure that people understand how accurate it is, how consistent it is, and have the confidence they've always had in NLCD continue. The thing that I really like is, I mentioned we're collecting those 10,000 points, every point has at least one interpreter that looks at it and collects that data from 1984 to 2023. But what we're also doing is taking 50% of those points and assigning them to a second person who does an independent duplicate interpretation, and then we compare those labels and we can get  information not only on interpreter consistency and interpreter agreement for the collection of the reference data, but we can also use those metrics for quality assurance, identifying edits or corrections in the final dataset. And it can be used for all sorts of other research related to reference data collection consistency, which  is an interesting area of research.

ADAMSON:
Okay, that sounds great. Thank you, Jo.

Research scientist Hua Shi had an interesting poster about how the urban heat island effect, warmer temperatures in cities, affects vulnerable populations the most, and how his study can help inform city planners on decision making.

Okay, go ahead and introduce yourself.

HUA SHI:
My name is Hua Shi, and I'm the research scientist at ... Contractor to USGS EROS. We used Landsat to derive land surface temperature. And because urbanization in the urban area have higher temperature, right, in non-urban area have low temperature. So when we calculate the intensity, that defined by urban temperature minus non-urban, you can show that the intensity of heat island. We know urbanization coupled with climate change drive the land surface temperature, LST, higher by intensifying the heat waves, resulting increased heat-related energy costs, illness, and mortality. This study investigate whether and how urban-induced warming varied across seven national climate assessment regions. So we selected eight urban center to represent these NCA regions. The figure one, you can see--

ADAMSON:
Several different cities that you focused on to kind of analyze this urban heat data that you have, right?

SHI:
Yeah.

ADAMSON:
And what's the most interesting thing that you found out in this?

SHI:
Most interesting part of this study is how it shows the complex relationship between urbanization, climate change, and socioeconomic difference. It is revealed that urban-induced warm varies significantly, depend on the regional climate and neighborhood. The study focused on the different socioeconomic factors like income, diversity, and their correlation with heat waves. And this understanding can help inform urban planning to reduce the heat waves, especially in vulnerable communities, leading to more equitable strategy to address the challenge of urbanization and climate changes.

ADAMSON:
Yeah, these vulnerable populations in cities mostly get these heat intensity areas. What can we do to help alleviate that?

SHI:
Decision maker or city planning for future, they need like consider this like a more like increased green spaces and water body--make like, artificial lakes or ponds, you know, can cool down this kind of thing.

ADAMSON:
That sounds really good. This can help inform those decisions. Okay.  Thank you, Hua.

Senior scientist Lei Ji mapped historical agricultural water use in the U.S. in irrigated cropland. Okay, well, I'll let you introduce yourself.

LEI JI:
My name is Lei Ji. I'm a senior scientist, work with ASRC, contactor to USGS EROS Center. Go ahead and summarize your poster for us. Okay. Oh, my poster here presents the study about how to map historical agricultural water use across conterminous United States using remote sensing method.

ADAMSON:
Okay, this is across the entire conterminous U.S., the 48 states.

JI:
Yeah. In the United States, 15% of croplands are irrigated land. So traditionally, the estimation of groundwater is based on reports, survey that's received from farmers. Data is collected and compiled by government agency. So this the work is, kind of, very time and cost consuming. So now we're working on the using the remote sensing method to create a dataset that is called blue water evapotranspiration, in short, just blue water ET, product to map water consumption.

ADAMSON:
You called it blue water evapotranspiration. Evapotranspiration just refers to water that's evaporating and transpiring off of the land. So then blue water evapotranspiration, that is water use for irrigated cropland, is that correct?

JI:
That's correct. Yeah.

ADAMSON:
I like how you have this figure here that shows a few examples in different parts of the country. I see darker green and lighter green. What do the different shades of green mean in the imagery here?

JI:
The darker green means higher water use from irrigation water by the crops.

ADAMSON:
Okay.

JI:
So lighter means the less water use by crops only for irrigation water.

ADAMSON:
Okay, so we can visualize that here in the imagery. What other data was this using?

JI:
We have surface temperature, vegetation indexes, and the phenology and the climate data also. And soil data. So all that data input and calculate the final part out is blue water evapotranspiration.

ADAMSON:
That sounds good, Lei, thank you.

JI:
Thank you so much.

ADAMSON:
Besides the more than 52-year archive of Landsat data, EROS also has vast holdings of historical aerial images and declassified satellite imagery. And we keep adding more to the collection. Drone images? Yep, we've got that too.

Okay, go ahead and introduce yourself.

BRENT NELSON:
My name is Brent Nelson. I'm the work manager for the DMID operations group. We were responsible nine years ago for the scanning of the Oregon aerial film. It's where our project started. We bid it out nine years ago and got that for about 232,000 frames. And through time now, we have now scanned all of the Bureau of Land Management film on our scanners in the digital lab, which totaled just under 970,000 frames of film that was scanned. And so we also took some of the imagery that they had had scanned by other outside vendors and integrated that into our database. Had to change the resolution on it. So we now have just a little over a million frames of BLM film that is available to the world through EarthExplorer.

ADAMSON:
What are the dates of these images then? How far back does it go?

NELSON:
This film goes back to about the 1949 era. So it's a lot of historical data.

ADAMSON:
This is useful for all kinds of land change studies.

NELSON:
There's already been quite a bit of demand for it. And we have all that film now boxed up, ready to ship--38 pallets of film--to ship to NARA for long-term storage. They have went through all their offices, their places, and they said, we now all have all of their film and it's all scanned. And for nine years, this was virtually a 6-day, 24-hour operation to get this taken care of. So this was, for the digital lab, for DMID, and the Center, a major undertaking.

ADAMSON:
Yeah. We've got pictures of the multitudes of rolls of film and pallets which are well, now that they're scanned, they can go back to long-term storage, right?

NELSON:
Correct. They'll go to Kansas in the NARA salt mines and be stored there. Because most of this is color. So it'll be frozen for long-term storage. And because it's all digitized, BLM now has access to all of it, whether it be for research or legal issues. It's all there. None of it was modified in any which way. I mean, it was all scanned as specific requirements as it is.

ADAMSON:
Is all of the high-res available to download from EROS right now?

NELSON:
Yes. All of this is high-res, scanned at 14 microns. And it is all available free of charge through EarthExplorer.

ADAMSON:
Very nice. Hey, thank you, Brent.

Okay, go ahead and introduce yourself for us.

TIM SMITH:
Hi, I'm Tim Smith. I've been at EROS since 1979. I'm the archive task lead for the Data Management Information Division.

ADAMSON:
Sounds good. What do we have on your poster right here?

SMITH:
This is highlighting the declassified satellite program from 1960 to 1984. This imagery was all the spy world that we used to worry about, and you had to have a classification even be able to use this imagery. And then in about 1992 time frame it was released. And then we've had subsequent declassifications over the years. This being the highest, most worldwide coverage of the Declass Hexagon program.

ADAMSON:
Okay. So the latest release is from Hexagon. How many images came in that release?

SMITH:
That's right around 600,000 images that we're sitting on, right around 14,000 reels of film that came in. This is second generation imagery. All the original film is at the National Archives in College Park.

ADAMSON:
One of the more interesting photos that you have on display here, of course, we have things like pressure ridges in Antarctica. Very interesting. There's a lake in South Dakota. I also see one that says Cuba missile site. So that was very interesting in the '60s. This is declassified now. It is just seen as more valuable to the public for research.

SMITH:
That's correct.

ADAMSON:
What do we do with it here?

SMITH:
From here, we're making it available, making sure we can have this imagery digitized to create the browse resolution images that are posted on EarthExplorer. And then from there, people are using that imagery to select what images need to be scanned at a higher resolution. And once that's scanned, we keep that image forever.

ADAMSON:
Okay, and then it's out there, high resolution, anybody can download it. This has become public domain information, right?

SMITH:
Yes, it has. Yes. It works out very well that way. And again, it marries up nicely with all the other national land remote sensing data archive imagery, of which Landsat is a part of that too. We have that congressional mandate to keep that imagery forever and make sure the public can get access to it.

ADAMSON:
Thanks, Tim.

SMITH:
Thank you.

ADAMSON:
Go ahead and introduce yourself for us, please.

BRENT JOHNSON:
My name is Brent Johnson. I work at EROS Data Center here.

ADAMSON:
Summarize for us real quick what you have on the poster here.

JJOHNSON:
Okay, the poster is quantifying and telling us what we have archived for
uncrewed aerial systems.

ADAMSON:
Okay. UAS is how we have it on here.

JOHNSON:
Yep. UAS. We started off collecting, gathering data, UAS data, for different government agencies that were using drones. That's what it is. And now we've combined, you know, concatenated that. Now it's just USGS we're gathering stuff for.

ADAMSON:
We're distributing that from here at EROS?

JOHNSON:
We are distributing it and archiving this data.

ADAMSON:
Okay.

JOHNSON:
There was an edict, or a mandate, by the government saying that any federal dollars used for federal projects would be made public and then archived, so it could be retained and pulled back out if they need it. We made this so that it's dynamic in the sense that not only imagery, but because UAS are now used for volcanic eruptions, taking gassing, readings in the plumes of the volcanoes. The big thing now is lidar data, which is in this case, we're looking at yellowscans people have used. That's the cheapest and most reasonable sensor they can hook on to the UAS to gather lidar data. It's elevation models, point clouds. So many things are becoming discovered now, especially archeology uses it a lot--lidar. And they're discovering so many ancient different things out there.

ADAMSON:
How many of these UAS images do we have then in our archive?

JOHNSON:
The total UAS imagery, we call it raw, that's coming from right off the camera, is 806,712 in our archive; however, that doesn't mean that's all that's out there.

ADAMSON:
That's what we've gotten here so far anyway. Are we adding to it?

JOHNSON:
We're hoping to add many, many more, yeah, obviously.

ADAMSON:
Very nice. This is all part of what we do, studying land change with whatever means we have.

JOHNSON:
Remote sensing, different remote sensing, land change and the systems that are used to gather that data.

ADAMSON:
This is really cool. Thank you, Brent.

JOHNSON:
You bet. Thank you, Tom.

ADAMSON:
Thank you for joining us on this episode of Eyes on Earth, where we touched on a lot of research that uses remote sensing data. The university students and EROS scientists who attended had lively discussions and learned a lot from one another.

Check out our social media accounts to watch for all future episodes. You can also subscribe to us on Apple and YouTube podcasts.

VARIOUS VOICES:
This podcast, this podcast, this podcast, this podcast, this podcast is a product of the U.S. Geological Survey, Department of Interior.
 

Show Transcript
Was this page helpful?