Host: Nick Petrić Howe
Welcome back to the Nature Podcast. This week, the rising problem of people living in flood prone areas…
Host: Shamini Bundell
And a fairer way to get citizens participating in policymaking. I’m Shamini Bundell.
Host: Nick Petrić Howe
And I’m Nick Petrić Howe.
[Jingle]
Interviewer: Nick Petrić Howe
First up on the show this week, so far, 2021 has been a year of extremes. Climate change is making the likelihood of extreme weather events much higher. We covered heat waves recently, which have been ravaging much of the world in recent weeks, along with wildfires. But the devastating risk of floods is also growing.
Everything on the ground floor is literally gone. My kids’ toys, everything is gone.
I live on the ground floor flat so I’m currently homeless. The whole flat is wrecked.
Interviewer: Nick Petrić Howe
These are the voices of residents of London, less than ten miles from where I’m sitting. Last week, torrential rain descended on the city for the second time in as many weeks, with the London Fire Brigade responding to more than 1,000 calls as people became trapped or needed to escape flooded homes. But that, frankly, is the tip of the iceberg. In Belgium and Germany, hundreds of people lost their lives in floods in the past couple of weeks. China, India, and Bangladesh have seen deadly and devastating floods too.
Interviewee: Beth Tellman
China, Germany, India and Bangladesh, these are all countries where we expect an increase in the proportion of population exposed to floods by 2030, and I think we’re already starting to experience that today in 2021.
Interviewer: Nick Petrić Howe
This is Beth Tellman, a flood researcher who this week has published a study about flood risk in populated areas around the world. It’s well understood that the risk of floods is growing due to climate change and increasing population. However, much of this understanding has been based on models which involve a lot of uncertainty. So, in her new study published this week in Nature, Beth took a different approach. Rather than modelling where water might be, she directly observed where water was and thus how many people were affected. Using satellite data capturing images from all over the Earth twice a day, Beth and her colleagues identified 913 flood events over the course of 18 years. With this, Beth is able to show that we’ve been underestimating how many people are at risk of being inundated by floods.
Interviewee: Beth Tellman
We see that up to 86 million people newly reside in these observed inundated areas. What this represents is an increase in the proportion of global population exposed to flood nearly a quarter – 20-24% – and that’s ten times higher than previous estimates that have used flood models.
Interviewer: Nick Petrić Howe
The data showed that populations were increasing at a much higher rate in flood prone areas than elsewhere. To get this new data, the team had to develop an algorithm that took infrared data from the satellites in order to figure out which bits of land were covered in water. They also had to exclude known locations of permanent water. This way of mapping was able to fill in some key gaps in flood models, including flood predictions for parts of the world, such as the global south, where there are sparse data but excellent satellite imagery. Beth’s satellite dataset was also able to determine which regions of the world had the greatest increases of populations in flood-prone areas.
Interviewee: Beth Tellman
So, some of the places where we saw the largest increase were in southern Asia and southeast Asia, India, Bangladesh, and sub-Saharan Africa also in particular. There were several countries that had some of the highest rates of population increasing faster in floodplains than in other parts of the country.
Interviewer: Nick Petrić Howe
This precise geographical information could target interventions for floods because, of course, with climate change, things are only going to get worse. Speaking of, using the data they had on past floods, Beth was able to project populations with flood risk in 2030.
Interviewee: Beth Tellman
And unfortunately we see places like the US and Europe are going to experience larger floods than they’ve adapted to in the past, and it changes the trend. Places that have previously had higher populations outside of floodplains, that will no longer be the case and they will start to have higher population growth in places that have flooded before, and their land use, zoning plans and infrastructure are simply not going to be enough because floods are spanning into places where there’s higher density human population.
Interviewer: Nick Petrić Howe
With the forecasted impacts of climate change, planning will be vital to mitigate the worst impacts of floods, as Brenden Jongman, a disaster risk management specialist, explains.
Interviewee: Brenden Jongman
A critical part here that really comes from Tellman’s conclusions – we need to plan better. When people move to cities around the world – they may move from the countryside to a city – we need to make sure that the new developments or the new urban developments really happen in areas that are safe from flooding and not in those areas that are flood prone. And if they do happen in those flood prone areas, we need to make sure that we plan well and that the houses that are being built are resilient to potential flooding in that area.
Interviewer: Nick Petrić Howe
As Beth mentioned before though, some of the worst impacts of flooding have been in southern and southeast Asia, where many nations are low- and middle-income, so may not be as able to invest in mitigation strategies as wealthier nations. Additionally, some people build on floodplains as they don’t really have a choice.
Interviewee: Beth Tellman
One of the really unfortunate facts about some of the trends of populations growing in floodplains is that many people who are building there build on floodplains because they have no other option. In many countries, floodplains become places that are restricted for formal development. However, people that cannot pay for land and to build a home on the formal market are often forced to build informally or illegally in floodplains because it may be the only land available.
Interviewer: Nick Petrić Howe
There are potential solutions though. Beth has suggested that insurance and pay outs from nations with historically the biggest contributions to climate change could help here. Brenden also works with the World Bank that invests in development and offers expertise in these areas. Another solution that could work for some nations is natural flood control, things like protecting mangrove forests or wetlands.
Interviewee: Brenden Jongman
Sometimes you have a coastal city with a lot of space available and there’s an existing mangrove area, for example, that’s been degraded. In many of those cases, we see that actually investing in regaining that mangrove, restoring it and using that as a flood buffer is more cost effective than doing it through green infrastructure solutions that are often more short-lived and they are also more expensive.
Interviewer: Nick Petrić Howe
There are some limitations to Beth’s study. For example, clouds tend to occur when it’s raining heavily and so obscured some flood events from the satellite imaging. And some urban floods can cause a lot of damage but are only around 1-metre-high in very densely populated areas, which is again hard for the satellite to pick up. Beth, though, is confident these limitations can be overcome using new technologies, like satellites that can ‘see’ through clouds, and with that we can develop better planning for flooding, based on better data.
Interviewee: Beth Tellman
There’s so much work to do. Mapping nearly 1,000 floods seems like a lot, but it’s just the tip of the iceberg. If you look at other studies that have used, for example, Twitter or news media to try to estimate the number of events that have happened in the past 20 years, we think that number is probably more like 50,000 events, so we are an order of magnitude away, or may two, from being able to capture and map all of the floods in the world.
Interviewer: Nick Petrić Howe
That was Beth Tellman from the organisation Cloud to Street and the University of Arizona. You also heard from Brenden Jongman from the World Bank. To learn more about flood risk, have a read of Beth’s paper and a News and Views article written by Brenden. We’ll put links to them in the show notes.
Host: Shamini Bundell
Coming up, we’ll be finding out about the tricky problem of being fair and random in picking citizens to help craft policy. Right now, though, it’s time for the Research Highlights, read by Dan Fox.
[Jingle]
Dan Fox
A research game called Big Robber could help explain why someone might happily steal money from a group of people but then make a charitable donation to an individual. A team of researchers asked 640 people to play a game they devised called Big Robber, where a robber decides whether to steal up to half of the earnings from a group of the other participants. 80% of robbers took at least one third of the money, and 56% took half – as much as they could. Only 2% of individuals refused to rob. However, when the same participants took part in games where they were asked to decide how much of a small sum of money they should transfer to another player, they found that most people who took part gave some money away. The findings suggest that, if given the opportunity, people will harm a large group in exchange for a high reward, but the same individuals can be compassionate one-on-one. Take a look at that paper in full in Nature Human Behaviour.
[Jingle]
Dan Fox
Researchers have found that smoke from wildfires causes the formation of large numbers of new water droplets inside clouds, brightening the clouds and potentially affecting local weather and climate. A team of scientists flew a research aircraft through the clouds and smoke plumes over wildfires in western US, gathering data on the size and distribution of airborne particles. Smoke particles can serve as seeds around which water condenses. They found that compared with clouds in smoke-free skies, clouds in smoky skies formed roughly five times as many droplets, and that these droplets were much smaller – about half the size of those in clear skies. These smoky clouds are expected to reflect more light and produce less rain than clouds in clean air, and may cause other more complex effects due to the warming impact of the smoke itself. Read that research in full in Geophysical Research Letters.
[Jingle]
Host: Shamini Bundell
Next up on the show, reporter Benjamin Thompson has been looking into a new paper in Nature that hopes to help make policy decisions more democratic, by creating an algorithm that selects people to appear in so-called ‘citizens’ assemblies’ in the fairest possible way. Here’s Benjamin with the story.
Interviewer: Benjamin Thompson
Around the world, many governments are looking to democratise the way in which decisions are made by getting regular people – not elected officials – to come together in bodies known as citizens’ assemblies to help inform and affect policy decisions. These assemblies are designed to be a microcosm of a larger population, with panel members selected to be representative, whether that be by age, gender, ethnicity, beliefs or so on – people who have different backgrounds and opinions, who can learn about an issue and discuss it from different positions. They’ve been used, for example, in Ireland where a citizens’ assembly made recommendations for changing the country’s abortion law, and here in the UK, where one was formed to discuss getting to net zero emissions by 2050. But putting panels together is easier said than done for the organisations that convene them. This week in Nature, a new paper has been published that describes a way to do it as fairly as possible, by using some clever maths. Typically, convening is a two-step process, as Bailey Flanigan, from Carnegie Mellon University in the US, explains.
Interviewee: Bailey Flanigan
What they generally do is they send out a bunch of letters to people in the population and they select those people randomly. And then the people in the population who receive these letters can respond and say, ‘Yes, I would totally be willing to join a citizens’ assembly.’ And so, the people who respond affirmatively form what’s called a ‘pool’ of volunteers, and then from this pool, the panel is ultimately selected. And the reason that you need to form this pool first is that not everybody in the population will agree to participate, so you can’t just randomly invite people and assume that they’ll all show up, and because you really want a representative panel, you have to form this group first and then try to pull a representative panel out of that group.
Interviewer: Benjamin Thompson
And this is tough. Some groups, such as college graduates, are more likely to accept the invitation, and a panel exclusively made up of people from this demographic isn’t exactly representative of the wider population. So, organisers often have to put several demographic quotas in place to try and ensure that a panel mirrors wider society as closely as possible. A 40-person panel needs to contain between 19 and 21 women, say. But hitting these quotas while at the same time choosing people to be on the panel at random is quite the puzzle, so organisers apply algorithms to do it.
Interviewee: Bailey Flanigan
Meeting these quotas even without thinking about the randomness is actually a fundamentally, mathematically difficult problem. You can imagine that it’s really hard to find a group of people that satisfy all of these quotas simultaneously, and so the existing algorithms, what they do is they basically build the panel person by person. They say, ‘Okay, which quota is in the highest demand right now and which quota are there the fewest people remaining in the pool of volunteers that can satisfy that quota?’ Let’s say you still need a lot of women and there’s very few women left in the pool right now. Then you probably want to take a woman to make sure that you don’t run out of space or you don’t run out of women.
Interviewer: Benjamin Thompson
Although these are impressive feats of maths, this method isn’t without its issues. Because of the balance of participants in the pool and the quotas to satisfy, for some people, the chances of getting selected on a panel is really small. And this is where Bailey’s work comes in. She and her colleagues describe a new algorithmic approach that tries to make the selection process as fair as possible, giving everyone the best chance of being involved in a citizens’ assembly panel. Rather than just find a single panel that satisfies the conditions, their method comes at it from a different angle.
Interviewee: Bailey Flanigan
And so, what we do is we actually find a bunch of panels that satisfy the quotas and then we calculate what is the fairest distribution over these panels, and what I mean by that is we assign a probability to every single panel that we find in a way that gives everybody in the pool as equal chance as possible of appearing on the panel. So, let’s say for example we find a set of 1,000 panels and a lot of people are on many of them but there’s maybe one person who’s only on one. Then in order to give that person enough probability, we should probably select that one panel with higher probability than some of the others ones. So, the optimality of our algorithm really comes from the fact that we’re fine-tuning the probability with which we select each of those panels and then we draw one out of that hat, one of the panels that we found, with the probabilities that we computed, and then this means that we’re being as fair as possible to the individuals in the pool.
Interviewer: Benjamin Thompson
So, by making a bunch of different panels that all satisfy the quotas and weighting them, the team can maintain randomness and make sure that everyone has a fairer chance of being selected. This method performs better than pre-existing algorithms and works over several different defined measures of fairness. Mark Warren from the University of British Columbia in Canada researches democratic theory and practice and has written a News and Views article on the new paper. He says that fairness is paramount for democratic processes.
Interviewee: Mark Warren
The way that democracy works is that people invest in governments, in political systems, that they believe are fair, but our kind of legacy solutions of representative democracy are subject to all sorts of known biases. So, if we want democracy to succeed, we need to keep thinking about introducing fairness into political institutions. So, it seems to me that this paper helps us down this path to deepening democracy, in particular through representing publics in a way that is much more accurate and much more credible than the others institutions that we put under the democracy umbrella.
Interviewer: Benjamin Thompson
Mark also says that trust in a process and transparency are vital as well, if people are to buy into a system. And this can be an issue when people don’t understand how an algorithm makes decisions. Bailey says the team is keenly aware of this and are working to make the system as transparent as possible. For example, they worked with an organisation to transform the process into a live TV lottery where everyone could see the chances they had of being selected on a panel. Other organisations have adopted this new algorithmic approach in their selection processes as well, and so far it has been used to set up over 40 citizens’ assemblies. It’s also available for anyone to use online. But there’s more work to be done. As governments and other organisations look to further democratise how decisions – both large and small – are made, it seems like the use of citizens’ assemblies is set to grow. And as Bailey and her colleagues work to hone their algorithm, she says it’s imperative they continue to collaborate with the people who organise them to ensure that participants are chosen as fairly as possible.
Interviewee: Bailey Flanigan
Something that has been so fruitful throughout this process and will continue to be fruitful is the questions and the feedback of practitioners who are actually working in this area, right? Because while we come with some of the math and the computer science, what we really lack is the on-the-ground experience, and as we move forward with this algorithm and maybe make tweaks to it or implement new versions, what we’re really looking to do is talk with practitioners and hear what they think about the algorithm. Are there parts of the algorithm that make them uneasy, that they don’t like? These kinds of questions and concerns really direct our work and pose really interesting technical problems that come directly from the people who we really want to help.
Host: Shamini Bundell
That was Bailey Flanigan from Carnegie Mellon University in the US. You also heard Mark Warren from the University of British Columbia in Canada. You can find links to Bailey’s paper and Mark’s News and Views over in the show notes.
Host: Nick Petrić Howe
Finally on the show, it’s time for the Briefing chat, where we discuss a couple of stories that have been featured in the Nature Briefing. Shamini, what have you been reading this week?
Host: Shamini Bundell
So, yeah, I saw in the Briefing there was a Guardian article highlighted about a sort of mysterious disease that’s killing coral in the Caribbean, and some recent research that is sort of shedding some light on this mystery.
Host: Nick Petrić Howe
Oh, poor corals. Every time I hear, it’s something else affecting them, whether it’s climate change or ocean acidity. So, something else is going wrong? What is this disease? What insights does this paper give us?
Host: Shamini Bundell
Yeah, so this disease has been known for a while. They first spotted it in 2014, but it’s pretty bad, and, yes, as you noted, coral has been having a tough time of it. There’s coral bleaching, there’s climate change and pollution. This does seem to be a sort of specific infection and a very deadly infection. It’s spreading faster, it has an unusually high mortality rate. This article says it’s potentially the most deadly disease ever to affect corals, and specifically one thing that was highlighted that I didn’t realise is that coral bleaching is when the little polyps that live inside and make the coral exoskeletons expel their photosynthetic bacteria that help them survive. So, coral bleaching, they’re really unhappy and it’s really bad for them and could lead to death. But this disease just flat out kills them, and what you’re left with is patches of coral that is just dead, that is just that exoskeleton.
Host: Nick Petrić Howe
Yeah, well, that sounds terrible because coral bleaching in some ways is sometimes reversible, but this is a very permanent end and, as I guess as we all sort of understand these days, finding the origin of diseases is quite tricky, but are there any insights into where this disease has come from?
Host: Shamini Bundell
So, it’s funny how they’re kind of having to look for clues from its spread. They don’t actually know what causes it yet. It could be a bacteria but it could also be a virus, maybe even a chemical. But looking at the way it’s spread through this region, sort of from Florida and down and then into Mexico, that has actually been a key clue for them. So, one clue is that that’s kind of in the opposite direction from the ocean currents. It’s not what you would expect. If you think about the spread of lionfish, which are an invasive species, they will sort of follow the natural flow of the ocean, whereas this one doesn’t. The other thing is that it sort of seems to jump about a bit, suddenly appearing some way away. And so, what the researchers doing this assessment did is look at where this disease is most prevalent in the reefs and compare that to where to main shipping routes and commercial ports are, and they found a link between humans driving ships around and the worst hit places.
Host: Nick Petrić Howe
Well, I guess it’s hard for cause and correlation but do they think it’s ships spreading it around or are they in some way causing it?
Host: Shamini Bundell
So, the theory is that this is another issue caused by ballast water. So, big ships transporting cargo around, particularly when they don’t have any cargo and they would otherwise be very light, take on sea water to stabilise them, and it’s been known for a while now that this sea water can transport invasive species. So, for example, zebra mussels were sort of spreading everywhere and ships’ ballast was one of the places they thought it was coming from, and they actually introduced a bunch of regulations about where ships could release their water ballast, so saying sort of doing it somewhere away from shore and away from reefs, and also requiring ships to log where they’re taking this water from and where they’re transporting it to. And that has actually also helped them kind of identify certain places where it looks like ballast from Florida then going to the Bahamas has potentially caused an outbreak of that disease right there.
Host: Nick Petrić Howe
And you mentioned regulations on this sort of taking on and dropping off ballast water. Is that something that could be done to stop the spread of this disease?
Host: Shamini Bundell
Yeah, so in 2017, the International Maritime Organization had this sort of particular rule that started which said you have to discharge your ballast water 200 nautical miles from shore in water at least 200 metres deep before you enter the port so that you’re not bringing in harmful pathogens. But one potential issue is that certain areas and countries don’t necessarily have enough resources to enforce that.
Host: Nick Petrić Howe
Okay, so maybe that’s not always the solution. What else are people doing about it?
Host: Shamini Bundell
So, it’s definitely, yeah, something that’s been relatively recent. It’s spread really far and is sort of a notable problem that people are looking into and, for example, the Bahamas has a huge spiny lobster export business that brings in a lot of money and employs a lot of people, and that’s something that if the reefs die, that’s going to impact the lobster industry, so they’ve got a really good sort of motivation for sorting it out. There’s a national taskforce. One of the things is kind of looking into treating the disease. So, there is actually a photo in this Guardian article of someone physically putting antibiotic ointment onto a coral that’s affected by this disease. They haven’t necessarily found a sort of proper solution to that yet. There is no cure. But it might also be possible that if we can tackle the human causes of this, the ships, the ballast water, the shipping routes, as part of a sort of bigger push, then the reefs will be able to heal themselves.
Host: Nick Petrić Howe
Well, hopefully there’ll be some good news for coral in the future. But my story this week is about gender bias in citations in medical research. So, I was reading a news article in Nature that was written about a study published in JAMA Network Open that has found that, in elite medical journals, women are substantially less likely to be cited in similar articles than those authored by men.
Host: Shamini Bundell
So, yeah, I feel like it definitely fits a pattern. What are the stats on how disadvantaged am I if I publish, with my obviously like advanced medical research that I do, if I were to publish in a medical journal, how much disadvantage would I be at in terms of other people citing my paper?
Host: Nick Petrić Howe
So, this study looked at 5,554 articles published in a few top medical journals between 2015 and 2018, and they looked at whether the primary author or the senior author or both were women. In the case of the primary author, so the first one, there would be around a third fewer citations that if it was a man. And in the case of the senior author, it would be around a quarter less citations than if it was a man. And if both of those were women, it would be about half as much. So, certainly a significant amount less citations for female authors.
Host: Shamini Bundell
Wow, so, yeah, I’m surprised by that number. It wasn’t that I was sort of particularly optimistic but like potentially up to 50% sort of less likely to be cited seems even bigger than what I would have thought. What were the researchers’ sort of comments on that?
Host: Nick Petrić Howe
Well, the author of this new study says they doubt that it’s intentional. They think it’s to do with the visibility of women authors. So, we know from past studies that male researchers are more likely to speak at conferences, for example, and they’re also more likely to promote themselves on social media, so it could be that when people are thinking of papers, when they’re writing their own paper and they want a cite something, they may just think of these authors that they know more or have seen speak at conferences or have seen their articles on social media. And, in general, it seems like female researchers have a smaller professional social network than their male counterparts.
Host: Shamini Bundell
Oh, that actually makes a lot of sense because I was kind of thinking, when I’m looking for papers to cite, would I be going and checking the first names of the authors and looking them up and seeing whether they were women or not. So, it’s part of this whole big structural disparity between men and women, and this is sort of just one sign and symptom of it.
Host: Nick Petrić Howe
Yeah, it seems like it’s more of a systemic issue. I mean, another researcher who was interviewed for this article, who wasn’t part of this new study, said as well that part of it is male researchers are more likely to cite themselves, so that could contribute to the fewer citations. And also, male authors are more likely to use words such as ‘novel’ or ‘promising’, which are words that are quite catchy and can sort of catch the interest and attention, so there could be some sort of implicit biases as well. And also, women tend to get fewer awards and they tend to be featured in newsletters left often than their male counterparts, and there are fewer women in these fields in general.
Host: Shamini Bundell
Well, yeah, it sounds like it’s all huge interlinked problems with various causes and not many solutions. Was there any sort of silver lining of this paper about how to tackle it?
Host: Nick Petrić Howe
Well, the study author was optimistic, and they were keen to point out that the number of studies published by women is growing. So, the number of women who were first authors in 1994 was 27% and in 2014 it was 37%, and the number of women who have been invited to conferences to speak as well is starting to rise. So, things are shifting, but it’s quite a long and entrenched problem and unfortunately it might be a long time before things reach parity.
Host: Shamini Bundell
Yeah, I guess it’s good to be able to quantify the problems that still remain even as people focus on improving it, so that’s encouraging.
Host: Nick Petrić Howe
Yeah, and the authors of this study also wanted to spark conversations. With this study they wanted to start the conversations in corridors between researchers so they start to think about who they’re citing and why they’re citing them. So, hopefully, with studies like this, there’ll be more of a conversation and researchers can start to think about who it is they’re citing and why they’re citing them.
Host: Shamini Bundell
Great, well, thank you very much for telling us about that, Nick. And listeners, if you want to hear more on those stories and others, you can sign up for the Nature Briefing, which is an email newsletter every weekday, and we’ll put a link to the two stories we talked about and to where you can sign up to the Briefing in the show notes.
Host: Nick Petrić Howe
That’s all for this week. As always, you can keep up with us on social media. We’re on Twitter – @NaturePodcast. Or for the original social media – email – you can reach us at podcast@nature.com. I’m Nick Petrić Howe.
Host: Shamini Bundell
And I’m Shamini Bundell.