Monitoring Arctic Permafrost with Satellites, Supercomputers, and Deep Learning

Arctic researchers and remote sensing experts use AI and HPC to characterize large, unexplored parts of the Earth

Peer-Reviewed Publication

UNIVERSITY OF TEXAS AT AUSTIN, TEXAS ADVANCED COMPUTING CENTER

Automated ice-wedge polygon prediction
IMAGE: PROGRESS OF AUTOMATED ICE-WEDGE POLYGON PREDICTION FROM SUB-METER RESOLUTION COMMERCIAL SATELLITE IMAGERY. SO FAR, RESEARCHERS HAVE MAPPED MORE THAN 1 BILLION INDIVIDUAL-ICE WEDGE POLYGONS ACROSS THE ARCTIC TUNDRA. IN ADDITION TO POLYGON OUTLINE, EACH DETECTED ICE-WEDGE POLYGON COMES WITH A SUITE OF ANALYSIS-READY ATTRIBUTES, SUCH AS ICE-WEDGE POLYGON TYPE, SIZE, LENGTH, AND WIDTH. view more  CREDIT: CHANDI WITHARANA, UNIVERSITY OF CONNECTICUT

Permafrost — ground that has been permanently frozen for two or more years — makes up a large part of the Earth, around 15% of the Northern Hemisphere.

Permafrost is important for our climate, containing large amounts of biomass stored as methane and carbon dioxide, making tundra soil a carbon sink. However, permafrost’s innate characteristics and changing nature are not broadly understood.

As global warming heats the Earth and causes soil thawing, the permafrost carbon cycle is expected to accelerate and release soil-contained greenhouse gases into the atmosphere, creating a feedback loop that will exacerbate climate change.

Remote sensing is one way of getting a handle on the breadth, dynamics, and changes to permafrost. “It’s like a virtual passport to see this remote and difficult to reach part of the world,” says Chandi Witharana, assistant professor of Natural Resources & the Environment at the University of Connecticut. “Satellite imaging helps us monitor remote landscape in a detailed manner that we never had before.”

Over the past two decades, much of the Arctic has been mapped with extreme precision by commercial satellites. These maps are a treasure trove of data about this largely underexplored region. But the data is so large and unwieldy, it makes scholarship difficult, Witharana says.

With funding and support from the U.S. National Science Foundation (NSF) as part of the “Navigating the New Arctic” program, Witharana, as well as Kenton McHenry from the National Center for Supercomputing Applications, and Arctic researcher Anna Liljedahl of the Woodwell Climate Research Center, are making data about Arctic permafrost much more accessible.

The team was given free access to archives of over 1 million image scenes taken in the Arctic. That’s a lot of data — so much that traditional analysis and features extraction methods failed. “That’s where we brought in AI-based deep learning methods to process and analyze this large amount of data,” Witharana said.

One of the most distinctive, and telling, features of permafrost are ice wedges, which produce recognizable polygons in satellite imagery.

“The ice wedges form from the freezing and melting of soil in the tundra,” said Liljedahl. “Some of them are tens of thousands of years old.”

The shape and dimensions of ice wedge polygons can provide important information about the status and pace of change in the region. But they short-circuit conventional analysis.

“I was on Facebook some years ago and noted that they were starting to use facial recognition software on photos,” recalled Liljedahl. “I wondered whether this could be applied to ice wedge polygons in the Arctic.”

She contacted Witharana and McHenry, whom she had met at a panel review in Washington, D.C., and invited them to join her project idea. They each offered complimentary skills in domain expertise, code development, and big data management.

Starting in 2018, Witharana began using neural networks to detect not friends’ faces, but polygons from thousands of Arctic satellite images. To do so, Witharana and his team first had to annotate 50,000 individual polygons, hand-drawing their outlines and classifying them as either low-centered or high-centered.

Low-centered ice wedge polygons form a pool in the middle of the ridged outer part. High-centered ice wedges look more like muffins, Liljedahl said, and are evidence of ice wedge melting. The two types have different structural hydrological characteristics, which are important to understand in terms of their role in climate change, and to plan future infrastructure in Arctic communities.

“Permafrost isn’t characterized at these spatial scales in climate models,” said Liljedahl. “This study will help us derive a baseline and also see how changes are occurring over time.”

Training the model with the annotated images, they fed the satellite imagery into a neural network and tested it on un-annotated data. There were initial challenges — for instance, images trained for Canada were less effective in Russia, where the ice wedges are older and differently shaped. However, three years later, the team is seeing accuracy rates between 80 and 90%.

They have described the results of this research in the ISPRS Journal of Photogrammetry and Remote Sensing (2020), the Journal of Imaging (2020) and Remote Sensing(2021).

After showing that their deep learning method worked, they turned to the Longhorn supercomputer, operated by the Texas Advanced Computing Center (TACC) — a GPU-based IBM system that can perform AI inference tasks rapidly — as well as the Bridges-2 system at the Pittsburgh Supercomputing Center, allocated through the NSF-funded Extreme Science and Engineering Discovery Environment (XSEDE), to analyze the data.

As of the end of 2021, the team had identified and mapped 1.2 billion ice wedge polygons in the satellite data. They estimate they are about halfway through the full dataset.

Each individual image analysis involves pre-processing (to improve the clarity of the image and remove non-land features like lakes), processing (where polygons are detected and characterized) and post-processing (reducing the data to a manageable scale and uploading it to a permafrost data archive). In addition to identifying and classifying ice-wedge polygons, the method derives information about the size of the wedge, the size of the troughs, and other features.

The individual analyses can be performed in less than an hour. But the sheer number of them make it unfeasible to run anywhere but on a large supercomputer, where they can be computed in parallel.

Recently, Witharana and collaborators benchmarked their workflow to find the optimal configuration to run efficiently on supercomputers. Writing in Photogrammetric Engineering and Remote Sensing (PE&RS) in 2022, they evaluated four workflow designs on two different high performance computing systems and found the optimal setup for high speed analysis. A separate 2022 study in PE&RS explored the efficacy of different image augmentation methods (such as changing the hue or saturation) when applied to deep learning convolutional neural net algorithms to recognize ice-wedge polygons from commercial satellite imagery. (Both projects were presented at the American Geophysical Union Fall Meeting in December 2021.)

“Every year, we get an almost near real-time pulse meter on the Arctic in the form of sea ice extent,” Liljedahl said. “We want to do the same with permafrost. There are so many rapid changes. We need to be able to really understand, and communicate, what’s happening in the permafrost.”

The ice wedge data will be available for rapid analysis on the new Permafrost Discovery Gateway, which will “make information about the Arctic more accessible to more people,” Liljedahl said. “Instead of having to wait 10 years to learn about something, they can learn about it right away and explore it directly through their own experience.”

Another important phase of the research project will come when the researchers analyze satellite imagery representing different years and times of year. Comparing the state of the ice-wedge polygons can show trends and trajectories, such as how fast the landscape is changing, and where those changes will cross paths with settlements or infrastructure.

“This is a perfect example of how previous investments in computing infrastructure, combined with new understanding of deep learning techniques, are building a resource to help with an important issue in the Arctic,” said NSF Program Director Kendra McLauchlan.

“Plato said, ‘Man must rise above the Earth — to the top of the atmosphere and beyond — for only thus will he fully understand the world in which he lives,'” Witharana said. “Earth observation technologies enable us to see how climate change is happening and how even the land is changing. It’s the main tool to observe, monitor, predict and make decisions to prevent a negative impact on fragile regions.”


JOURNAL

Photogrammetric Engineering & Remote Sensing

DOI

10.5194/isprs-archives-XLIV-M-3-2021-175-2021 

METHOD OF RESEARCH

Computational simulation/modeling

SUBJECT OF RESEARCH

Not applicable

ARTICLE TITLE

An Optimal GeoAI Workflow for Pan-Arctic Permafrost Feature Detection from High-Resolution Satellite Imagery

ARTICLE PUBLICATION DATE

4-Jan-2022

From EurekAlert!

4.8 5 votes
Article Rating
42 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Pat from kerbob
February 22, 2022 6:19 pm

Since it was very much warmer many times in the last 8000 years, why didn’t runaway methane release and global warming occur then?
The tree line has been much further north

I just want to know

Scissor
Reply to  Pat from kerbob
February 22, 2022 7:25 pm

That’s one of those questions that cannot be answered (or even asked) as it would ruin the conjecture.

Opus
Reply to  Pat from kerbob
February 22, 2022 7:27 pm

The fact checkers have determined you have no need to know.

Graemethecat
Reply to  Pat from kerbob
February 23, 2022 6:52 am

The one question which destroys the whole CAGW narrative. Never will you get an answer.

Tom Halla
February 22, 2022 6:41 pm

Cute. Remotely sensing permafrost. I do wonder how rigorously they will validate their method, and how.

Trying to Play Nice
Reply to  Tom Halla
February 23, 2022 5:18 am

They’re using Deep Learning. There is no need to validate. /sarc off

mal
Reply to  Tom Halla
February 23, 2022 7:48 am

Modern science has devolved to computer games. It looks to the bulk of so called climate scientist have little experience with weather, you know the kind of weather when you do field work.

JPSF
February 22, 2022 7:06 pm

Maybe the AI computer models they build will answer your question….or maybe not.

Joao Martins
Reply to  JPSF
February 23, 2022 3:20 am

If AI will not do, then better try NS (Natural Stupidity)…

Last edited 4 months ago by Joao Martins
archer
Reply to  JPSF
February 23, 2022 1:22 pm

Oh, they already have the answer, they’re just trying to figure out the question.

Lil-Mike
February 22, 2022 7:32 pm

Here’s what I don’t get … granted I’m not a ‘Permafrost Expert’ but, a geo working around permafrost in Alaska.

Permafrost is not a binary thing. Permafrost I see is in patches bordered by non-permafrost.

Permafrost areas are characterized by sparse stunted black spruce forest with mosses. Non-permafrost is characterized by birch-alder and large healthy tall robust black spruce forest.

These stands are melted unevenly, but only during the summer months. Everything refreezes in October. Some zones melt a little in the summer, some melt a lot. When a zone melts, biological processes begin and run through the summer into fall. This means, yes carbon is being released, but it also means that carbon is also being re-sequestered.

But this is being sold as a binary change, when clearly it is not.

Ron Long
Reply to  Lil-Mike
February 23, 2022 2:44 am

Good comment Lil-Mike, especially about the tree variation. See my comment below for another geologists experience with permafrost in Alaska. Good luck.

markl
February 22, 2022 7:38 pm

So what will they compare their ice wedges to? How long will it take before they understand the significance? Interesting approach though.

Michael Elliott
Reply to  markl
February 22, 2022 8:04 pm

Having obtained data from AI & their Super computer, have they gone to any of these areas to compare the real World data with the remote data.

Otherwise the whole exercise is completely pointless.

Anyway why bother about 8000 year old CO2 being released when both India & China are busy putting

vast amounts of this nasty stuff into the atmosphere .

Michael VK5ELL

Steve Case
Reply to  markl
February 22, 2022 8:28 pm

 How long will it take before they understand the significance?
__________________________________________________

They knew that answer before they started: “Worse than previously thought!

Allan MacRae
February 22, 2022 8:07 pm

More global warming…

By the end of 2020, the climate doomsters were proved wrong in their scary climate predictions 48 times. At 50:50 odds for each prediction, that is like flipping a coin 48 times and losing every time! The probability of that being mere random stupidity is 1 in 281 trillion! It’s not just global warming scientists being stupid.

But no sensible person makes a 50:50 prediction – at 60:40 the odds against being this wrong are 1 in 13 quintillion; at 70:30 the odds against being this wrong are 1 in 13 septillion.

Climate doomsters have not been telling the truth. Ever! Quelle surprise! 🙂

RECORD COLD ENGULFS THE U.S.; HISTORIC FLURRIES CONTINUE TO BATTER JAPAN; + “ONCE-IN-A-LIFETIME” SNOW SWEEPS NAGALAND, INDIA
February 22, 2022 Cap Allon
Some 30 million Americans are currently in the path of a major winter storm — prepare.

FIERCE COLD WAVE SWEEPS EAST ASIA, CHINESE CITY SEES FIRST SNOW SINCE 1893; + “DANGEROUSLY COLD TEMPERATURES” TO HIT U.S. AND CANADA
February 21, 2022 Cap Allon
The COLD TIMES are returning…

KANSAS CITY BREAKS 129-YEAR-OLD SNOWFALL RECORD; DENVER OFF TO ITS SNOWIEST START TO A YEAR ON RECORD; + HISTORIC ACCUMULATIONS CONTINUE TO BUILD IN JAPAN
February 18, 2022 Cap Allon
“Snowfalls are just a thing of the past” — the AGW narrative.

RECORD COLD VALENTINE’S DAY; GREAT LAKES ICE SEES MONSTROUS RECOVERY; NEWLY DISCOVERED DATA REVEALS SEA TEMPERATURES AT THE GREAT BARRIER REEF HAVEN’T INCREASED IN 150 YEARS; + HUGE FARSIDE CME
February 16, 2022 Cap Allon 
“We may have dodged a bullet,” writes Dr Tony Phillips.

CANADIAN TRUCKS RUNNING DAY AND NIGHT TO CLEAR RECORD-SMASHING SNOWFALL AS NORTHERN HEMISPHERE SNOW MASS PUSHES HIGHER
February 15, 2022 Cap Allon
While the words “truckers” and “Canada” evoke a certain imagery these days (of a battle against tyranny), hundreds of dump trucks are embroiled in a different kind of fight right now, one versus the elements, versus the unprecedented accumulations of SNOW.
 

Last edited 4 months ago by Allan MacRae
Tom Abbott
Reply to  Allan MacRae
February 23, 2022 1:36 am

It looks like the whole northern hemisphere is experiencing the same cold weather. I guess we should call this “global”, shouldn’t we? Or should we switch to “hemispheric”? Seasonal? Since temperatures are never really global. When it’s cold in the northern hemisphere, its warm in the southern hemisphere and vice versa.

Allan MacRae
Reply to  Tom Abbott
February 23, 2022 2:30 am

The Southern Hemisphere has also experienced extreme cold and recent crop failures. See Electroverse.net

OweninGA
Reply to  Tom Abbott
February 23, 2022 5:56 am

Ahh, but the warmunists will point to the current Southeast US WEATHER and say see it is all going to burn. We will likely see 80+ here in Georgia today, but 200 miles north of us is seeing very cold weather. It is all to the positioning of the jet streams in their meander and we will be back to our more usual 60s by Saturday. They will forget that the last half of January was 10-15F below usual for the area.

mal
Reply to  OweninGA
February 23, 2022 7:52 am

Arizona may end up in record breaking cold today but then again maybe not. It not suppose to get out of the fifties. The lows tonight are suppose to be cold for us.

Allan MacRae
Reply to  Allan MacRae
February 23, 2022 5:14 am

CHINA’S GAS PRICES SOAR TO RECORD HIGHS AS COLD WAVE DRAINS SUPPLIES; SUMMER CHILLS STRIKE TASMANIA; + HUNDREDS OF LOW TEMPERATURE RECORDS FALL ACROSS THE U.S.
February 23, 2022 Cap Allon
It’s been a glacial start to the week for many Americans as winter 2021-22, rather than abating as the calendar nears March, has instead intensified… GSM.

Allan MacRae
Reply to  Allan MacRae
February 24, 2022 12:53 am

https://wattsupwiththat.com/2021/12/28/unsung-zeroes-the-top-10-under-reported-climate-flops-of-2021/#comment-3420346
[excerpt] 
I accurately predicted the current British energy crisis in 2002 and in greater detail in 2013. January and February 2022 will be worse.

https://wattsupwiththat.com/2013/10/31/blind-faith-in-climate-models/#comment-1130954
AN OPEN LETTER TO BARONESS VERMA
British Undersecretary for Energy and Climate Change, 31Oct2013

Yup – as predicted.
https://wattsupwiththat.com/2022/01/03/californias-sierra-nevada-sets-all-time-december-snow-record/#comment-3424429
 
The killer global cold will come in January and especially February 2022 – locked in by the cold Nino34 SST in October 2021. Nino34SST leads UAHLTglobal by ~4 months.
 
The difficult-to-predict southward descents of the Polar Vortex will decide who lives and who dies. I am particularly concerned about the UK and Germany, two old enemies joined in their struggle for winter survival, their vital energy systems sabotaged by toxic green-energy false propaganda. Lenin and Goebbels would be proud – their lessons of false propaganda have been well-learned by their modern pseudo-green acolytes.

Len Werner
February 22, 2022 8:20 pm

I just had to–are we talking wedgies in iceholes now?

AndyHce
February 22, 2022 9:37 pm

I didn’t see one word about sizes of these “ice-wedges polygons”. Am I just not reading well enough? Is there some logical reason they are called wedges?

While I haven’t been keeping track, I know I’ve seen at least two different papers based on close, at the source, study that said thawing permafrost seems to be be a net methane sink. Just too inconvenient to acknowledge or is there actual work that refutes the claim?

Tom Abbott
February 23, 2022 1:06 am

From the article: “As global warming heats the Earth and causes soil thawing,”

That’s right, start this article out by making claims about the future you can’t possibly know will happen. You are speculating that warming will continue. There’s no guarantee it will.

Tom Abbott
February 23, 2022 1:10 am

From the article: “As global warming heats the Earth and causes soil thawing, the permafrost carbon cycle is expected to accelerate and release soil-contained greenhouse gases into the atmosphere, creating a feedback loop that will exacerbate climate change.”

Pure speculation. There’s no evidence that the “carbon cycle” will accelerate and there is no evidence increased CO2 will cause such an acceleration.

These “scientists” are assuming way too much. Typical for alarmist climate science. This is what it is made of.

DMacKenzie
Reply to  Tom Abbott
February 23, 2022 8:20 am

The vegetative matter in the permafrost is plant material that previously grew by absorbing CO2 out of the atmosphere when the climate was warmer. Over one glacial cycle, 25K years or 100K years, permafrost will be “net zero”….

Tom Abbott
February 23, 2022 1:15 am

From the article: ““That’s where we brought in AI-based deep learning methods to process and analyze this large amount of data,” Witharana said.”

They put the data through a computer and then call it by fancy names to make it sound more important and innovative. Sheesh. Got to have the hype.

Peta of Newark
February 23, 2022 1:24 am

Why or how is melting permafrost ‘breaking drains
1/ Is it beyond belief that the stuff going down the drain (liquid water based stuff) is doing the melting. Especially if it came from bathtubs, dishwashers, laundry machines, showers, kitchens etc
2/ Like wise and different, maybe the permafrost is causing the stuff in the drain to actually freeze – and it is that which is ‘breaking the drains‘?

I do, being a self appointed Soil Scientist, love the disconnect between frozen Arctic Soils and the sort of non-frozen soils found elsewhere

Especially why do Arctic soils emit CO2 when they warm up while already warm soils seemingly do not

I get worse, much worse, as they reveal that they haven’t a <expletive> clue how soils and plants interact and work
If the warming permafrost was going to release CO2 and CH4, it is The Bacteria within the soil that are doing the emitting.
They are only really going to get to work when the temperature of the soil gets above 5 Celsius.
Strangely, that is also the temperature at which plants, which capture CO2, come out of their ‘winter dormancy’
Thus, is it beyond another belief that the plants on the surface will capture the CO2 as it is being released from the soil below them.
(There’s No Possibility that that is why the stomata on all plants’ leaves are on the underside of the leaves – that plants expect their food to be coming up at them and not down at them?
Nah. Thought not)

So we get to the real conundrum, how did the permafrost come to have so much carbon within it?
Yet again, is it not possible that the plants capture more carbon than is released from the soil beneath them.
Surely Shirley that has got to be true else there’d be no ‘carbon rich permafrost

Positive Climate Feedback
Words fail, what can be done with people as dumb as that?

There is completely nothing of scientific value in these folks’ work what-so-ever.
What sort of school did they go to, What Is Happening Here….

Tom Abbott
February 23, 2022 1:25 am

From the article: ““Earth observation technologies enable us to see how climate change is happening and how even the land is changing.”

Really? I thought your analysis wasn’t finished yet?

You appear to be assuming things not yet in evidence. Again.

What is it with you “climate scientists” anyway?

Ron Long
February 23, 2022 2:40 am

“METHOD OF RESEARCH-Computational simulation/modeling” versus my personal experience fighting against permafrost? In 1967 as a Geology Student/Summer Field Assistant, on the Ignaty Creek mercury prospect, on the lower Kuskokwim River in Alaska, just down-river from Red Devil, I participated in a fight against permafrost, which was hindering exploration of the prospect. The US Bureau of Mines jumped in and ran a field trial with a Bulldozer, water cannon (from Jon Miskovich, of Flat Alaska, inventor of the IntelliGiant), and finally, my favorite, EXPLOSIVES! Focused attacks by all 3 techniques failed to make any economically viable headway against the frozen ground, so the final result: PERMAFROST 1, PROSPECTORS 0.

Joao Martins
Reply to  Ron Long
February 23, 2022 3:29 am

In that distant past you were lacking “artificial” intelligence and were only using the “natural” stuff…

tom hewitt
Reply to  Ron Long
February 23, 2022 11:58 am

Commercial drilling and soils testing contractor from Alaska. First of all, the word “Arctic” generally refers to the area above the Arctic Circle, currently 66°33′48.9 north latitude. At least in Alaska there is a great deal of permafrost at a lower latitude. The lower Kuskokwim at Red Devil would be at a latitude of 61 degrees, 72′, a little less than 400 miles south of the Arctic Circle.

The permafrost at Prudhoe Bay is continuous to a depth of 1600 feet, which requires that oil field drilling mud be almost the thickness of peanut butter to prevent hole erosion and casing cement that produces no heat during hydration to prevent blow-outs.

Tom.1
February 23, 2022 3:12 am

Measuring changes in something like permafrost is a legitimate area of scientific inquiry. So is monitoring those changes over time. I see no harm in it. The MSM and a lot of others don’t really care about science, but not doing science cannot help that.

mal
Reply to  Tom.1
February 23, 2022 7:55 am

Without proper field work to check the number the satellites are giving you, the so called research is junk. Now what are the odds they will go and check their work on the ground? My guess is slim and none. The bugs on the tundra during the summer from my understanding is similar to the Minnesota north woods. You hand will turn black from the mosquitoes on it and said mosquitoes are the least of your problems. It the flies and ticks that do most of the damage.

Last edited 4 months ago by mal
Joao Martins
February 23, 2022 3:18 am

Monitoring Arctic Permafrost with Satellites, Supercomputers, and Deep Learning
Most probably this question is the outcome of my intellectual limitations:

Is “Deep Learning” what we in the (not distant) past called acquiring a shallow knowledge?

February 23, 2022 3:59 am

“As global warming heats the Earth and causes soil thawing, the permafrost carbon cycle is expected to accelerate and release soil-contained greenhouse gases into the atmosphere, creating a feedback loop that will exacerbate climate change.”

Taking the assumption of warming permafrost at face value, is this really true? How do they think all that Biomass got there in the first place. I think if anything these areas would rapidly become sinks as the grass lands would spread probably every bit as fast as the permafrost would melt.

Reply to  bob boder
February 23, 2022 10:58 am

My take away from this study is that it is a natural process in the “green-house gas cycle which is much greater than anthropogenic contributions. A “net-zero” solution is stupid.

Michael E McHenry
February 23, 2022 6:41 am

In the “Dune Trilogy” they have destroyed all the computers/AI. I am beginning to understand why

Howard Dewhirst
February 23, 2022 10:55 pm

Didn’t most of the permafrost melt during the Holocene maximum? And was there a methane Spike?

Jeff Richard
February 25, 2022 4:06 am

Deep Learning is a black box that can’t be deciphered because you don’t know what the weights are actually locking onto.

Nolan Parker
February 26, 2022 3:25 am

Permafrost is important for our climate, containing large amounts of biomass stored as methane and carbon dioxide, making tundra soil a carbon sink.

I can’t help but hear the voice of an elementary school student reading this crap out loud.

Not One, zero, ZIP, not one of the
Eaux MuGawd the world is burning
Warming Warnings has actually happened in the forty years, and two End of Times ,twelve year s to get it right,, periods..

I hope you are triple jabbed and double masked.

%d bloggers like this: