Quirks and Quarks

00:00 If you're listening to this, I already know you have great taste in podcasts. But maybe, if you're like me, you still wonder if you're missing out on the best stuff. That's where the Sounds Good newsletter can help you out. Every other Thursday, the audiophiles at CBC Podcasts highlight one must-hear show and lots of other new and noteworthy titles. They do conversation starters, they do hidden gems, and they also tell you about the stuff they love that they didn't make. Go to cbc.ca slash soundsgood to subscribe. 00:31 This is a CBC Podcast. 00:38 Hi, I'm Bob McDonald. Welcome to Quirks and Quarks. On this week's show, finding a fossilized feather inside dinosaur poop. 00:47 You know, everybody poops. It's part of life. And now, you know, even 66 million years later, a poop can tell us so many wonderful stories. 00:55 And using tiny trackers to follow elusive chimney swifts. 00:59 I thought, oh, urban area, lots of phones, chimney swifts, really lightweight tags. Let's see what we can do with these birds. 01:07 Plus, mathematicians grappling with AI, climate change destabilizing mountains, and what burying 2,000 pairs of underwear reveals about soil. All this today on Quirks and Quarks. 01:25 Just over 66 million years ago, a massive asteroid hit our planet and abruptly ended the age of the dinosaurs. But as we now know, some dinosaurs did survive. 01:37 Namely, one particular lineage of avian dinosaur which eventually became the modern birds we know and love today. 01:46 Why this one lineage survived and the other 99% of dinosaurs were wiped out is a big mystery. But we now have a new clue in this case in the form of an impeccably preserved fossilized feather found in a cherry-sized piece of dinosaur poop. Dr. Jing May O'Connor was part of the team analyzing this feather. She's a paleontologist at the Field Museum of Natural History in Chicago. Hello and welcome back to our program. Thank you so much for having me. Everybody loves fossil poop, apparently. 02:17 Apparently. So tell me about this fossilized feather. How was it found? 02:22 It was found thanks to a lucky break. So my colleague Dave DeMar was in the Badlands of the Hell Creek Formation in Montana. And so he's, you know, you want to be close to the rock surface to find really good fossils, especially small ones. So he's creeping around and he sees this weird, unusual nodule. And so he picks it up and kind of like flips it over. And there was a break in this piece of coprolite that exposed a feather inside. And so, you know, we've never found feathers in coprolites before. And if it wasn't for this break exposing the feather, we probably would go on blissfully unaware that coprolites are such wonderful potential sources of fossilized soft tissues. 03:01 Now, when you say coprolite, that's a science term for dinosaur poop. 03:06 Any kind of fossil poop, yes. 03:09 Now, do you know what animal left the poop behind? 03:13 We can narrow it down. So it's too girthy and has an unusual shape to be a crocodilian. And the only other large enough carnivores are theropod dinosaurs. So there's four contenders. But I think the most likely is a Tyrannosaurus rex or a Nanotyrannus. So not a full-grown T-Rex, like an immature, smaller individual. 03:33 How rare is it to find feathers from the age of the dinosaurs like this? 03:38 Typically, they're very rare. You have to have just the right conditions to fossilize them. So actually, all the fossil sites that have those just right conditions that are showing us what feathers were like in the Mesozoic, they're almost all from the early Cretaceous. Our understanding of feathers in the late Cretaceous is extremely limited. So, you know, the Hell Creek is one of the best known, you know, fossil-bearing geologic units in the world, right? You know, it's where T. rex and Triceratops heridus, like all the big famous dinosaurs that we all know are from there, right? 04:09 And after over 110 years of prospecting, this is the first time we've ever found a feather. So feathers are rare, especially in the late Cretaceous, and that makes this discovery extra exciting. 04:21 How well preserved is this feather? Yeah. 04:24 It's pretty darn perfect. It's the best preserved feather from the Mesozoic. The previous best feathers were all preserved in amber. But in amber, you still get some distortion in the resin. So you're not able to see certain very fine detailed features. This feather is just incredible. It's really a beautiful feather. Wow. 04:48 Okay, so you have this preserved feather inside a piece of dinosaur poop. How did you confirm what kind of animal it belonged to? 04:56 Yeah, this is actually the really fun part, because if we can't say who that feather belongs to, then we can't understand the significance of what we see in that feather in an evolutionary context, right? So we have to figure out whose feather it is. So luckily, in this little part of the coprolite, there were feathers, but there were also a couple bones. And these two bones allowed us to make the hypothesis that it's probably a hesperonithoform bird. So Hespernithophores are birds that have been known since 1872. This is the first time we've ever seen their feathers. 05:27 They're secondarily flightless, like foot-propelled diving birds. So they're aquatic. So then we went one step further to test that hypothesis by checking if the feathers had aquatic specializations. So we were able to see that these feathers are indeed better adapted for aquatic locomotion than any other feathers we've ever seen in the Cretaceous before. You know, it's like a preponderance of evidence that allowed us to make this hypothesis. We definitely don't have a smoking gun. And I'd be the first person to say that it's possible we're incorrect. 06:00 But, you know, I think we've made a pretty strong argument and it was really fun piecing it together. So what did this feather tell you? 06:07 We saw two features that are present in modern feathers that we've never seen in a Mesozoic feather before. And it has to do with the shape of the cross-section of the rachis. It's square in this feather, whereas it's oval in other feathers that we've seen in more primitive birds. And also this rachis is filled with tiny little air cells. And so these are features that are present in all modern birds. So if we couldn't identify the feather, we would just say, okay, now we know, thanks to discovery, that modern feathers were present in the latest Cretaceous. 06:39 But because it's hespernithiform bird, we can say that The modern feather was present in Cretaceous, but the modern feather evolved outside of modern birds. But what we also see, because there's multiple feathers preserved in this coprolite, we see modern feathers alongside primitive feathers. And this is a combination that we've also seen in other primitive birds, where the feathers that make up the wing are more advanced than the feathers that are on the body. And we think that this combination might be one of the reasons why all these groups of birds went extinct. 07:12 Well, take me through that. Why would some birds survive and others go extinct? 07:17 The Mesozoic was a greenhouse world. So during most of the Mesozoic, the polar regions were ice-free. So it was much warmer, right? And, you know, it's this asteroid impact that triggers this mass extinction. But what really caused the mass extinction is the impact winter that resulted from the asteroid collision. So during this impact winter, global temperatures dropped significantly. And also you now have a time of resource scarcity because the sun is blocked out and so like the ecosystems are collapsing, right? So these birds that had, you know, these primitive body feathers that were probably not as good at insulating them, that wasn't a problem when it was warm in the greenhouse world. 07:58 But suddenly you have rapid climate change. It's much colder. And your feathers that aren't as good for insulation mean that you need more energy to stay warm. But there's no food through which to get that energy, right? So this is a combination that makes organisms very susceptible to extinction. So our hypothesis is that the birds that do survive, they had modern wing feathers, but they also had modern body feathers. And that those body feathers, which were better for insulation, then allowed these birds to survive through this impact winter. 08:32 So to test this, what we'd really have to do is find a modern bird at the end of the Cretaceous and be able to look at what its feathers look like. It's probably never going to happen, unfortunately, but I'd never say never. But the reason I am happy to share this hypothesis is because we've seen this before in enantiornithines preserved in amber, where they have wing feathers that are pretty modern looking, missing these two features that we see in the coprolite feather, but like feathers good enough for flight. But then their body feathers are like really weird and definitely like very obviously not good for insulation. 09:03 So I had already three years ago put forth the hypothesis that that's why they went extinct. So, you know, it's starting to like the data is starting to suggest that plumage differences was a major contributing factor in the selectivity process. I'm not saying it is the one reason why some birds survive and some don't, but I think it contributed definitely. 09:28 So are you suggesting that we scan more dinosaur poop to answer some of these questions? 09:34 Oh, absolutely. I'm not saying that we should scan all the dinosaur poop, but any coprolite that is really well preserved. Like this one, it just really retains its shape, which means that it was buried very rapidly. And so if you have rapid burial, then you're going to have a higher chance of having this exceptional preservation. So, yeah, very exciting. 09:54 Poop as a resource. 09:56 Yeah. Well, I mean, you know, poop is a resource to so many organisms. You know, it's food for organisms like scarab beetles lay their eggs in it. Like, you know, I mean, like we think of poop as dirty and shameful, but, you know, everybody poops. It's part of life. And now, you know, even 66 million years later, a poop can tell us so many wonderful stories. 10:16 Dr. O'Connor, thank you so much for your time. My pleasure. Dr. Jing Mei O'Connor is the associate curator of fossil reptiles at the Field Museum of Natural History in Chicago. Music An announcement from OpenAI at the end of Labor Day weekend this year is creating an existential crisis in the field of advanced mathematics. 10:48 For decades, mathematicians have been working to solve this problem. They haven't been able to crack it. But AI, it just did it easily. OpenAI sending this out, saying we're sharing a solution to the Navier Stokes Millennium Prize problem, one of the deepest problems at the frontier of mathematics. 11:06 The proof was produced by an army of OpenAI agents using an unreleased model, significantly more capable than the Astra model the company just put out. 11:18 It took about 10,000 of these AI agents a staggering 88 hours to solve and cost millions of dollars. 11:27 Now, this wasn't the first mathematical proof that AI came up with, but it's certainly the biggest and most controversial. 11:35 Quirks and Quarks producer Sonia Biting has been following the story and spoke to some of the mathematicians at the heart of this issue. She's with us now to tell us about it. Hi, Sonia. Hi, Bob. So first, set the scene for me. How did we get to this point where an AI model can solve a problem that mathematicians have been pursuing for close to a century? 11:55 Well, 2026 has been a crazy year for mathematics. I would say the story really started back in May when OpenAI announced they'd solve the most well-known problem in a subfield of math known as combinatorial geometry. This was one of the numerous puzzles posed by the Hungarian mathematician Dr. Paul Erdos that OpenAI was celebrating in this announcement. 12:20 I mean, this sounds like too good to be true. We turned it to a few of the Erdos questions that many, many people are interested in. 12:28 And to our surprise, it came back with a solution to one of the most important Erdos questions. 12:35 That was the first major AI breakthrough in math. And it was really the first drop of what became a torrent of math breakthroughs. 12:44 August 1st, OpenAI published a 249-page manuscript describing 10 new results across mathematics and theoretical computer science produced by Astra, solving problems that have been stuck for decades. 12:57 So this summer, I was seeing all of these announcements coming out about these huge math breakthroughs. And I thought that's what we'd be talking about on today's show, how AI was changing the way math is being done. But on Labor Day weekend this year, this story took a really sharp turn. 13:17 Why? What happened? 13:19 Well, let me back up for a minute. There are seven pinnacle mathematics problems called the Millennium Prize problems. One is the Navier-Stokes problem. It looks at equations that describe how liquids or gases flow in the water or air, and it asks, Do these equations stand up in all situations or do they sometimes become unreliable? And there are a lot of mathematicians working on these. One is Dr. Tristan Buckmaster from New York University. 13:51 Late that Sunday night of Labor Day weekend, Dr. Buckmaster released a set of three mathematical proofs related to the Navier-Stokes equation and a statement that is shaking the very foundations of math. 14:06 Why is that? 14:08 Well, Dr. Buckmaster wrote how by the first days of September, he and his collaborator were well on their way to solving the main Navier Stokes problem. Then when OpenAI got wind of it, the company swooped in with their massive computing resources and beat them to it. Now, Dr. Buckmaster ended up speaking with OpenAI when he found out that they'd solved it. When he asked them how, he said their answer raised a bright red flag. 14:37 I start asking, like, when did you start working on this? I mean, he wouldn't answer. Eventually, it became clear that I only started working on this after the rumor. 14:49 Well, why is that timing important? 14:52 Well, like a lot of mathematicians who've been using AI models to accelerate their research, Dr. Buckmaster was using chat GPT to advance his Navier-Stokes work, and he used it to help solve the Euler problem that was one of those three proofs. The Euler problem is similar to Navier-Stokes, but a lot simpler, and it's like a stepping stone to the big problem. 15:14 Euler is the same equation as Navier-Stokes, except there's no viscosity. So viscosity is the internal friction in the fluid, and Euler is the essence of the whole, the fluid. The main difference between the two is that you add friction. 15:31 He and his collaborator came up with that proof in mid-August, but he had been using a commercially available OpenAI model for about a year to work on these problems while his collaborator was working with Anthropics models. So Dr. Buckmaster wonders if his prompts may have given OpenAI's new model an edge that allowed it to beat them to the proof. Now, there are a lot of paths to potentially solving the main Navier-Stokes problem, but the path OpenAI took was similar to the one Dr. Buckmaster and his collaborator were on. 16:07 So is the issue here that Dr. Buckmaster was working on this problem and he was using ChatGPT, but then OpenAI came along, they own ChatGPT, so they had access to his information and they used it to solve the problem first? 16:23 Well, that is what he suspects. It's circumstantial evidence, though. A bit of he said, he said. But yes, that is what he's alleging, that the work they'd been doing inside chat GPT was either intentionally or inadvertently used to leapfrog their own work. 16:40 So what has OpenAI said about all of this? 16:43 Well, first, OpenAI said that no specific user data was accessed to solve the problem, but added that they, and I'll quote, couldn't rule out that de-identified data derived from their usage of our products helped improve our models, end quote. Then the next day, they clarified and said, and I will quote again, we can say categorically that it is impossible for Dr. Buckmaster's codex prompts over the last two months to have influenced the system in any way, including training. 17:16 Dr. Buckmaster says he doesn't believe them. And after I spoke with him, he told me by email that he actually had been using three separate chat GPT accounts over the past year. And he only just realized this week that he didn't opt out of a setting that allows open AI to use conversations for training. in one of those accounts. He said the improve the model for everyone setting in the chat GPT interface he was using is also super confusing because you have to go to the web version to set it. 17:47 And if you slip up at any point, which he said he can't be sure hasn't happened with his other two accounts, it can swallow up everything you've done. 17:57 But he is not the only mathematician who was upset with how OpenAI has been operating in this space. 18:04 What do you mean? What happened? 18:06 Well, once word got around the math community about this issue with Dr. Buckmaster, a different mathematician started to rethink a conversation he had with OpenAI earlier in the summer when it released the 10 math proofs on August 1st. Dr. Andreas Thome at the Technical University Dresden in Germany had been working in a branch of math that focuses on infinite symmetry groups. The techniques he had come up with to probe this area of mathematics were key to one of those 10 math proofs that OpenAI solved. 18:40 But initially, he wasn't credited for his contribution. 18:44 So in the beginning, their public announcement said that there wasn't any progress in the last 10 years and they would really be for decades. I think that's what they actually said. And so I think back in the beginning of August, we complained about that, that that's kind of not really reflecting the importance of our work in their work. And then they actually changed his public announcement. 19:04 OpenAI did later credit his contribution, but he had also been using chat GPT for his research. So at the time, he was curious if they had access to his chats because he said there's only a handful of people on this planet that understand the exact theory he's working on. So he asked them. 19:24 I asked him whether there would have been some direct access to our chats so that they could really extract somehow the exact strategies that we were following. That would have been really a scandal, I would say. 19:35 Even if you allow it to be used for training data, that doesn't mean that we can really go to the personal chats, like word by word. I mean, you... 19:45 And that was the second question. So has our research process been used as training data for their new model that they were testing so that it could proceed much faster to solve this problem? 19:58 He said OpenAI answered no to his first question about whether their conversations were being used for training, but did not answer if his research process was made available to the model that made the proof. 20:11 Wow. So how is the math community reacting? 20:16 Well, mostly they're pretty upset. To borrow a phrase from our Prime Minister, Mark Carney, the field of mathematics is undergoing a rupture, not a transition, because of AI. And their concerns center around three main issues. One is the concern that a big company like OpenAI or Anthropic might use their inputs into these powerful AI systems to leapfrog discoveries that humans are on their way to solving. So they're concerned that if rumors start swirling that mathematicians are close to a proof, that these companies with their massive computing resources will beat them to it. 20:54 And we know that happened with the Navier-Stokes equation. And the second issue is about giving credit where credit is due. And this was an issue for Dr. Andreas Thom and Dr. Buckmaster. Neither of them were initially credited for how their work may have helped OpenAI solve their respective problems. 21:13 And the third major issue has to do with the loss of human understanding with these AI math advances. Here's Dr. Tom again. 21:21 This digestion process, it takes a long time. And humans have to kind of see through the difficult concepts that are involved in a particular argument to see the structure of the argument. What does it actually mean? So how to think of it, right? Not just how to formally prove it. That maybe a computer can do very efficiently these days. 21:42 But what does it actually mean? So how can we put it in the landscape of existing mathematics as a new object? 21:49 Bob, I read this really great analogy that really illustrates why this is important. Imagine we're on an island and there's something we need to get to in the middle of the island. And the way the AI systems work, it's like they can teleport themselves directly to that location. But then how do the rest of us get there? What path can everyone else take to get to that same place? And it's a similar thing happening with these math proofs. Just because AI solved them doesn't mean anybody else will be able to learn anything from it. 22:20 This sounds like the classic black box. I don't know how it does it. It just does it. 22:25 Well, these are the ultimate black boxes, too. 22:28 Yeah. 22:29 So what happens now with the Navier-Stokes problem? 22:32 Well, some mathematicians are now going over OpenAI's proof, which is proving to be an extremely challenging task. I spoke with Dr. Scott Armstrong, who's a mathematician at NYU and is also affiliated with the University of Somme in France. He's going over it now with the intention of rewriting it in a way that other mathematicians will be able to understand. 22:54 If you read their paper, the first few sections are pretty nicely written. And you're like, wow, they did a great job. And then you get to the part where the details come and then it gets really, really hard to read suddenly. 23:05 For some context, Dr. Armstrong says this inability to understand proofs isn't just an AI problem, though. This is something that's been plaguing the mathematics field for a long time. 23:16 Mathematicians are famous for like spending three years to prove something and then what they write is just you can't penetrate it. They're like, why can't they explain how they thought of it? So this is always what mathematicians have done. And no one has ever complained about it. I mean, we've complained about it a lot, actually, but no one has really thought it's a big crisis until this week. 23:38 So what do mathematicians want? 23:41 Mathematicians, they're seeing value in these AI models, but they have serious concerns. They want more transparency about whether or how these AI models have access to their chats. And they don't want AI companies to rush through these complex problems. to take their time and do their due diligence so that if anyone needs to be credited, that they get their proper acknowledgement. Dr. Armstrong says he thinks he can get through the Navier-Stokes proof in a few months' time, so he's one of the very few mathematicians who thinks it's great to have more knowledge with these AI proofs, even if it does take a while to understand it. 24:18 Well, thanks for telling us about this, Sonia. 24:21 Thanks, Bob. 24:23 Sonia Biting is a producer with Quirks and Quarks. 24:27 I'm Bob McDonald, and you're listening to Quirks and Quarks on CBC Radio 1 and streaming live on the CBC News app. Just go to the local tab and press play wherever you are. 24:38 Coming up later in the program, why Swiss scientists ask hundreds of people to bury 2,000 pairs of cotton underwear, then dig them back up weeks later. 24:50 Some instances, there was only the synthetic threads at the waistband left. The underwear was completely eaten. 24:57 And in other cases, the underwear still looked like almost new. 25:03 What if I told you that most of history's underwear wasn't remotely close to the underwear you're wearing right now? Hi, I'm Greg Jenner, host of the You're Dead to Me podcast, and we're back with a brand new season. And if you're a fan of fashion trends of the past, you'll love our newest episode on the history of underwear. We explore how our underwear has changed throughout history and whether TV depictions of Regency-era ladies in their lacy underwear are truly accurate. Listen on BBC.com or wherever you get your podcasts. 25:35 Biologists often have to travel far and wide to do field work on the animals they study, sometimes venturing into remote wilderness or trekking deep into forests to collect their data. But our next researchers didn't have to do any of that. Instead, they traveled up to urban rooftops around Guelph, Ontario to figure out how to stick teeny tiny tags on chimney swifts. 26:01 It was a pilot program led by Natasha Barlow from Birds Canada to test whether new tracking technologies would work on these fast and nimble birds and give scientists another tool to help this threatened species thrive. 26:16 Producer Amanda Buckowitz spoke with Ms. Barlow about her adventures. 26:22 Chimney swifts are part of our incredible urban wildlife. When you're on a patio or you're walking downtown, you might be hearing this really high-pitched twittering sound. You can look up and you can probably see some of these chimney swifts flying around overhead. 26:44 They are kind of a story of resilience because they used to nest and roost in mature forests and caves. And then when the forest started being cleared, they had to adapt. And they readily adapted to urban areas, brick and mortar structures like chimneys, some wood barns. And then now we're not using chimneys anymore. And so when those chimneys are being taken down, we're asking them to adapt again. And so chimney swifts are a threatened species. 27:17 They've declined by about 90% across Canada since the 1970s. And so they are one of the species that we're trying really hard to understand and appreciate and then also conserve. 27:31 I'm Natasha Barlow. I'm a programs biologist with Birds Canada, and I'm primarily responsible for our aerial insectivore programming in Ontario, which is a fancy word for birds that fly around and catch insects while they're flying. We surprisingly know little about chimney swifts. We don't really know why they're choosing certain habitats. We don't even know really where they go when they place their nest in the chimney and then where do they forage. 28:02 We know that they can fly sometimes as high as over 2,000 meters high and mean flight speed of around 46 kilometers per hour. So they're fast and they're nimble. which makes them hard to track and that hinders some of our effective conservation measures, which is why we're trying to uncover some of that. 28:24 We came up with the idea to track these birds because of seeing a presentation on these Bluetooth tags on monarch butterflies. 28:35 They're really, really lightweight. They are about the size of a grain of rice. 28:39 And they use Bluetooth kind of like an air tag where if you attach it to an animal and it flies over a phone that is connected to Bluetooth and has cellular service, it picks up that animal. And so when I saw that on the Monarch Butterflies, I thought, oh, urban area, lots of phones, chimney slifts, really lightweight tags. Let's see what we can do with these birds. 29:12 When we came to the first site that we were tagging at, so thankful that we were able to work with an apartment complex that actually has a really long history of caring about the chimney swifts in their chimney. 29:28 We go up the elevator, we go to the rooftop, and we put a net inside of the chimney. 29:36 Let me know when I can start to lower. 29:40 It looks a little bit like a basketball net where it's a rectangle. There is a hoop almost, a net that is hanging down inside of the chimney. And then we walk away from the chimney and we wait until the swifts go in. 29:55 There's a bird thinking about going in. 29:59 Oh, making another round. 30:00 Oh, almost. 30:03 When it starts getting near dust, you start having a bunch of chimney slifts coming through. They're flying overhead. They're catching insects. You can hear them chittering and communicating with each other. So it's a really cool experience just almost being at eye level sometimes of these birds that are usually so high in the sky. We're like, oh my goodness, this is the time. Oh my goodness, is this going to happen? And it just dropped in. 30:30 There it goes. It's in. It was just absolutely incredible. It was a rush. We had our first bird in hand and then we got to work. 30:40 With the first bird, we put it in a little bird bag. It keeps it warm and more calm and brought it back to our banding station. Then once we get back to the station, which is a really short walk, we remove the bird from the bag and we start looking at it. It's the first time any of us have seen a chimney swift in the hand. 31:03 So this is a second year bird. 31:07 You want to look at this just for learning. 31:11 And that's my plumage. 31:13 It's really fascinating having one of these birds that kind of looks like a small dinosaur in your hand. Yeah. 31:19 146 grams with the bag. 31:23 126 for wing. 31:25 Then once we've taken all of our data, once we've figured out how much it weighs, how much fat it's carrying to see if it is a really good body condition or not, we take a uniquely ID'd aluminum band and we put it on the leg. 31:44 And then with this bird, it was our first time tagging the swift. So essentially what you do is you trim a little bit of the feathers on the back so there is a tiny little patch of skin. 31:56 We'll see if it works. I brought my mustache scissors. 32:00 Excuse me? I brought my mustache scissors. They should be better. Oh, mustache scissors. That's so funny. 32:04 And then you just do some really light glue, attach the tag to the back. It's okay. We did this. 32:13 Yeah. 32:14 And then once it was all finished, we put the bird inside of a butterfly enclosure. 32:23 Okay. Yeah. 32:25 Clinging well. 32:26 She's going to have to leave the door. All right. 32:28 Let's release him. 32:30 And then once we saw that she was fine, we let her go. 32:38 She kind of just flew off into the dark. 32:44 I, of course, could not sleep. It was so exciting. Woke up really early the next morning, went on the portal online from the manufacturer, and you can essentially see real time if the data is working. And so because it's such fast and new and emerging conservation technology with these tags... they give you almost minute by minute information about where these birds are moving. It was just absolutely incredible being able to see how she was able to forage and move around and use some rivers. 33:18 And so by even tagging a couple birds to start, we already know more about the local scale movement, where these birds are going, if they're using waterways, if they're staying around their nest site, how far they're going. from only using these tags alone. 33:38 This was definitely a pilot year and we got way more out of it than we expected. We were trialing this new tag. We were catching chimney swifts for the first time. And so we were really excited to know that it works. It really works. So our main goal is it works. The tags work and building on that is going to be incredible. 34:08 That was Natasha Barlow, Programs Biologist with Birds Canada. And you also heard the voices of Dr. Matt First, a research ecologist and observatory manager, and Mr. Gabriel Evans-Cook, an Ontario Aerial Insectivore Program Biologist, both with Birds Canada. 34:42 On August 26th, a rock and ice avalanche unleashed catastrophic flooding across Nepal. More than 1,300 people were swept away, and thousands are still missing. 34:56 That disaster, scientists now say, was triggered by several factors that developed over years, including climate change, according to a report published this week by World Weather Attribution. a group that analyzes the role of human-caused warming in weather disasters. 35:15 So while experts say there's no one single cause to this tragedy, they agree that this event was decades in the making. Dr. Dan Sugar is a geomorphologist at the University of Calgary who's also been studying the events that led up to the flood. 35:32 Dr. Sugar, welcome back to our show. Thanks very much. 35:34 Happy to be here. 35:36 Now, you've studied satellite images of this region both before and after the disaster. Can you walk us through what they show? 35:43 So the morning of the disaster, I was able to see where this amazing flood originated from by tracing the erosion, sort of the scars of this incredible event up through the Himalayan valleys to the the north side of a mountain called Langtang Lirung, where we saw that, initially anyway, we saw that a piece of a glacier had essentially vanished. 36:11 Wow. Well, just give us a sense of scale from the satellite images. I mean, how large of a disaster was this? 36:18 It's almost hard to comprehend. 36:20 The scale of this event is... 36:25 almost literally off the charts in terms of the power of the flood and the scale of devastation in terms of number of people killed as well is almost unprecedented in a mountain disaster like this, at least in recent decades. 36:42 The landslide that triggered this event found fell sort of the maximum distance before it hit the valley bottom and began to kind of break up was probably just over about a kilometer. So quite a vertical fall. And as you can imagine, when 100 million cubic meters of boulder bedrock glacier ice hit the ground after falling a kilometer, it makes a tremendous, tremendous bang. 37:11 So now we have this report by the World Weather Attribution. It's the first scientific study on this event so far. What has that added to our understanding of the event? 37:21 You know, the report offers a few interesting tidbits, in particular that snowfall in the region... this past winter was substantially higher than average, which may have contributed a lot of water to the system, or there may have been a lot of water in the system as a result, which, you know, if enough of that had been sort of plowed by this landslide as it moved down the slopes, that could have certainly contributed to these incredible floods that we saw. 37:52 Oh, I see. The extra snow provides meltwater. 37:55 Yeah, exactly. 37:56 What other factors were involved that could have led up to this? 38:00 So, you know, when we think of climate change in the high mountains, two main things that immediately come to mind are glacier thinning, glacier retreat, and permafrost thaw or permafrost degradation. And so... Glacier retreat often is a contributing factor to these sorts of high mountain landslides. 38:23 In this case, though, the glacier retreat is a little bit trickier to pin down. You know, certainly... 38:29 human-caused warming has thinned the glaciers in this area substantially. Whether that had enough of a, played enough of a role in sort of debuttressing, removing support from those slopes, that remains to be worked out. 38:45 But the permafrost changes are one that I think is very interesting. So, you know, permafrost exists in both the high latitudes, you know, up up north, but also at the high altitudes. And so if the ground, whether it's soil or bedrock, remains frozen below zero for more than two years, and we call it permafrost. Now, as the climate warms, the latitude or the altitude above which permafrost exists, that is changing. 39:17 And so the If a landslide occurs right around where that permafrost sort of threshold is, then we might begin to think, okay, well, maybe thawing permafrost had something to do with this. Think of a nice hot summer day and you're down at the beach and your kids or your grandkids are building sandcastles and the dry sand doesn't really, you can't build much of a castle with that. 39:47 And so you get the kind of wet sand, and then that will make a nice steep-sided castle. But if the sand gets too wet, then the water pressure pushes those grains of sand apart, and the whole thing collapses. And so bedrock is a little bit different from sand. It's a lot stronger. But nevertheless, liquid water can play a similar kind of role. And so in this... 40:12 particular case, any big cracks in the bedrock, even if it's permafrost, if there's an injection of liquid water pretty quickly into those cracks, that certainly could destabilize that slope. And so one of the first things, aside from the glacier breaking off, one of the first things that we could see very, very early on in the satellite imagery was that the glaciers in the valley were were, the day before, were very bright white, you know, covered in actual snow, whereas the day of, they were much darker covered, grayer. 40:49 And so that suggests that the day before was maybe quite warm and melted a lot of that snowpack, which may have delivered, sort of made available and delivered liquid water to the slope that failed. That would certainly make a lot of sense. 41:06 Is there any way that you could see precursors, that you could monitor these mountain ranges and give the people in the valleys more warning? 41:16 The truth is that this sort of analysis is very hard to do in steep mountains. In steep mountains, whether it's in the Himalaya or the Rockies or the St. Elias Mountains and Yukon, the slopes everywhere are moving. And so one of the big challenges is knowing which one is going to be a problem. It would have been very difficult, if not impossible, to forecast that this one would be a problem three weeks ago. 41:44 If we had had the technology to have real-time satellite monitoring as well as from a network of seismometers across the Himalaya, maybe then... 41:56 We, the scientific community, the disaster risk reduction community, could have seen that something was up on that slope, but it's a very difficult nut to crack. That's not to say we shouldn't try. I think we really should. Not just in the Himalaya, but here at home in Canada as well, where we really know almost nothing about slope stability. 42:18 Well, yeah. What about here in Canada, where we have our Rockies with glaciers sitting on top of them? 42:23 We've got the Rockies, the Selkirks, the Monashies, the Coast Range, etc. And we don't tend to have the same sort of level of settlements right up against the mountains as is much more common in the Himalaya. But, you know, we do have a lot of people that live in mountain environments and we've got a lot of infrastructure, whether that's highways or pipelines or whatever. And... We really don't know very much about what is moving and where might the next big one happen. 42:58 And we're certainly not immune to it. We've had several in the last couple of decades. Luckily, the damages have been relatively low, but they're there. 43:09 And across High Mountain Asia, we've had about one of these sorts of disasters per year for the last decade or so. And this is a trend that I think is not a blip. It's not an anomaly. I think this is what we expect with climate change, even if we haven't determined that this one in particular is linked explicitly to climate change yet. But this is the sort of thing we expect. And we'll see these sorts of things in North America as well. 43:40 And, you know, we would probably be even less prepared than Nepal. 43:44 So are you saying that we can expect more of these events in the future? 43:49 I suspect so, yeah. In the high mountains that are changing, you know, very rapidly due to climate change, but also in the Arctic, you know, across northern North America, but also in Scandinavia and Russia and Asia, we will... likely see these kinds of very large landslides occurring that present a little bit of a different risk, similar to this Alaskan one last year, where these things occur over bodies of water, like fjords, they could trigger very large tsunamis, which if a cruise ship or a freighter going through the Northwest Passage, for example, is in the area, then that could be a real catastrophe. 44:31 Yeah. 44:33 Dr. Sugar, thank you so much for your time. 44:36 Thank you, Bob. Dr. Dan Sugar is an associate professor in the Department of Earth, Energy, and Environment at the University of Calgary. 45:00 Back in 2015, a group of Ontario farmers had a wacky idea, encouraging people to bury their underwear in their yards to watch how fast it degrades. It inspired a campaign across Canada that, understandably, became very popular. 45:15 Here's Alan Kruzel, a board member with the Soil Conservation Council of Canada. 45:20 I was part of that group at the time. I said, geez, you know, we should look at picking that national. So we built a national campaign that we launched in 2017 called Soil Your Undies. And it was just a marvelous hit. I buried undies with all kinds of different people. We did it with senators. We did it with deans of universities, school kids, garden clubs. You have no idea how much fun a group of little old ladies had with holding up soiled undies. 45:48 It was initially meant to raise awareness of microbial life in the dirt, but the idea soon captured people's attention around the world, including Swiss soil ecologist Dr. Franz Bender. He and his colleagues were inspired to launch a similar project in Switzerland, only this time, to make it scientific, they recruited 1,000 citizen scientists, standardized the process, and used the information they gleaned from thousands of degraded undies to create a soil health map of the country. The work was recently published in a new study, and this month it also won the team an Ig Nobel Prize, an award for science that first makes you laugh, then makes you think. 46:30 Dr. Bender works at the Swiss Federal Research Institute Agriscope in Zurich. Hello and welcome to our program. Hello. 46:38 First of all, I have to ask, why underpants? 46:43 That catches people's attention. When people hear about underwear, they ask, why underwear? They start to laugh and smile, and then we make them think about soil and catch their attention. 46:55 Well, besides just getting attention for your project, what is it about underpants that makes them such a good indicator of the health of the soils? 47:03 Yeah, underpants also have very practical advantages. So when you use 100% cotton underpants, cotton is a natural material. It's made of cellulose. So that's something that soil organisms can decompose, that they can break down. 47:17 And another very important feature of underpants, that they have a waste band made of rubber. And also the seams are usually made of synthetic material. So when you bury them in the soil and the cotton gets eaten up, You still will find the skeleton of the underpants, so to say. If it would have been just a piece of cloth, you wouldn't know if you found it or if it was eaten up. 47:40 Ah, so this is not soiled or used anywhere. 47:43 Exactly. This would bias the results because, you know, depending on how long they have been worn and how they have been worn and who wore them, this might affect how much soil organisms like to eat them. 47:56 So take me through what you had the people actually do, your volunteers. 48:01 Yeah, so every volunteer got shipped a participant package containing two pairs of underwear and 12 tea bags, because there's also a method called the tea bag index, which looks at the decomposition of tea. So there are very specific tea bags made of synthetic material, and the tea inside gets decomposed, but not the bag. So it's a scientifically more established method to look at soil decomposition processes. So we used that as a quality control. And we also created a smartphone app where people could take pictures and they get uploaded to a map and where they could also add or enter all sorts of background information on the site where the underwear was buried. 48:42 So we knew what kind of land use it was. Was it a private garden, an arable field, a forest? How was it managed? Was it fertilized? Were pesticides used? And all these sort of things that helped us interpret the results at the end. 48:56 Oh, I see. So you included teabags to compare how the underwear performs compared to the standard test, which is the teabag. 49:04 Exactly, yeah. 49:05 How long were the underwear buried? 49:09 Yeah, so one pair of underwear and one set of teabags was taken out of the soil after four weeks, and the second set of underwear and teabags was taken out after eight weeks. 49:20 Eight weeks, wow. So what did people find when they dug up the old underwear? 49:26 Yeah, depending on how the soil was. Some instances, there was only the synthetic threads at the waistband left. The underwear was completely eaten. 49:36 And in other cases, the underwear still looked like almost new. 49:40 Well, what type of soil did the most damage to the underwear and what type of soil did the least damage to the underwear? 49:47 So we found in private gardens, so private vegetable gardens and flower beds, decomposition was highest. So that's where people usually use a lot of compost. They also fertilize the soil. Maybe they also irrigate the soil. So these are conditions, a lot of nutrients, a lot of organic matter, moisture that promotes soil biological activity. And we found the least decomposition in lawns. So that's mostly just grass, often a monoculture, maybe herbicides are used. 50:16 And then arable fields and grasslands were somewhere in between. 50:20 Why do you think there was such a difference? 50:22 because of the conditions that promote decomposition that helps soil organisms thrive, which can be nutrient concentrations, but also organic matter in the soil. So that was clearly the highest in the private gardens. 50:36 And yeah, organic matter is really important to store water in the soil. Soil organisms can break it down. It also provides a lot of different niches for soil organisms to live in. 50:49 So there's really an active biological community, especially when it's compost, there's material that is actively broken down by soil organisms. So there's a community there that is already used to break stuff down. So that's why it seems to be particularly working well there. 51:05 Wow. Well, aside from the fun factor of this project using underwear, I mean, it was a serious scientific endeavor. So tell me about the soil health map. Why did you want to do that? 51:18 Well, it's incredibly difficult to study soils. They're underground. And once you take a soil sample to the lab, you destroy soil structure. Also, it harms the organisms. They might not be doing the same thing that they would do in an intact soil. So that's why there's also not much data on soil. 51:36 And so this citizen science approach was also a big chance for us to really get a lot of soil samples from all over the country. So we had close to 900 soil samples that we were analyzing to look at chemical, physical characteristics of the soil and can really give us interesting insights that we didn't have before. 51:58 Well, why is it important that we track the health of the soil? 52:02 Well, soil provides 95% of our nutrition, of our food. It grows directly or indirectly in soil. They provide clean drinking water for us. They store carbon in the soil. So soil contains more carbon than the atmosphere and all the living biomass on the planet together. 52:20 And they're the biggest recycling enterprise on earth. So all the organic material, all the plants and trees and animals, once they die, they get broken down by soil organisms and the nutrients and resources in the material get recycled and enable the growth of new life. So soil is really important to keep the cycle of life going and we all depend on soil. And soils are really threatened worldwide, so we lose them at an alarming rate. 52:51 In the European Union, it's about 60% of agricultural soils actually degraded through compaction or construction activities. So they're covered under a layer of concrete. For example, in Switzerland, it's eight soccer fields per day that are covered under concrete. And once it's covered, basically there's no oxygen, there's no water, so it more or less dies off. 53:12 Then we pollute the soil with microplastics and chemicals like pesticides and PFAS and all other sorts of chemicals where we don't even know what this will do to soil and the soil organisms living therein. So it's really important that we understand our soils better and know how we can protect and maintain them in order to survive. 53:38 So do you think underwear soil tests will become a common practice moving forward? 53:43 Well, we hope to get more people to get their hands dirty and bury some underwear in the soil because it really helps to understand or learn about soil processes and learn that it's a living system underneath and they really should value it more than they actually do. 54:03 And yeah, take care of the soil. 54:06 Well, congratulations on your Ig Nobel Prize, and thank you so much for telling us about it. 54:11 Thank you. It was a pleasure. 54:14 Dr. Franz Bender is the team leader for agroecological assessments at the Swiss Federal Research Institute Agroscope in Zurich. 54:25 And that's it for Quirks & Quirks this week. If you'd like to get in touch with us, our email is quirks at cbc.ca. Our webpage is cbc.ca slash quirks, where you can check out our past episodes and find out more information about the research we covered in the show. You can also follow our podcast, get us on SiriusXM, or download the CBC Listen app. It's free from the App Store or Google Play. 54:53 Quirks and Quarks is produced by Sonia Biting, Rosie Fernandez, and Amanda Buckowitz. Our senior producer is Hannah Hoag. I'm Bob McDonald. Thanks for listening. 55:06 For more CBC Podcasts, go to cbc.ca slash podcasts.