Connecting the Dots: How Climate Detectives Link Human-Caused Environmental Change to Migration

This transcript was generated using AI and may contain inaccuracies. If you notice an error, feel free to email [email protected].

CHAPTERS 

[00:02:01]: How extreme weather events influence displacement and immobility 

[00:04:14]: Attribution science and methods for linking climate change to impacts 

[00:05:59]: Challenges in attributing slow-onset events such as drought 

[00:07:11]: Climate, conflict, and livelihoods in the Somalia case study 

[00:12:40]: Counterfactual scenarios and their role in attribution analysis 

[00:18:29]: Legal applications of attribution science in climate-related cases 

TRANSCRIPT
[00:00:05.11] 

There have always been hurricanes. There have always been droughts. The seas have risen and fallen countless times over the millennia that our planet has existed. So when a new hurricane comes, or when it won't rain for weeks on end, how do we know whether or not it's the result of human caused climate change? And when people are forced to move, how do we know that it's those extreme weather events that they're fleeing? This is Changing Climate, Changing Migration. Coming to you from the Migration Policy Institute in Washington, D.C. this is a podcast dedicated to examining and exploring what climate change means for migration. My name is Julian Hattem. I am your host and I am the editor of MPI's online magazine, the Migration Information Source. Today on the program, we're asking about how we can attribute human movement to specific extreme weather events and how we can attribute specific extreme weather events to human caused climate change. In other words, when is it climate change and when is it just the weather? I am happy to have as our guest today Lisa Thalheimer. Lisa is a climate scientist and a research scholar at the International Institute for Applied Systems Analysis in Austria, and she's done a lot of work on what is known as attribution science.

[00:01:25.24] 

Lisa, thank you so much for coming on.

[00:01:27.22] 

Thank you, Julian. Very happy to be here.

[00:01:30.03] 

So correlation is not causation, right? The fact that a period of extreme weather occurs does not mean that everyone who moves during that time is displaced or migrating because of that single incident of extreme weather. You've done a lot of research on attributing displacement and migration to environmental change and specifically human caused climate change. Walk us through it. How do we know whether a particular incident of movement or particular movement trend is the result of a particular environmental event? And how do we trace that linkage?

[00:02:00.18] 

So extreme weather, and if I say extreme weather, I mean droughts, storms, floods, all of these really extreme types of weather can trigger a range of human responses. The one side is that, you know, people migrate voluntarily, like think of a drought. They try to find water resources, other ways of livelihoods. But then there are also these more abrupt responses we see in the context of floods, for example, where people get displaced and they have no other options, no agency than actually to move. So this is one side of the coin which is probably more familiar from news, also in the context of hurricanes. Then there's also another side to the same coin where people want to move, but they're unable to move when they're hid from extreme weather events. And this has been a predominant push in the at least science community to also recognize what's referred to as involuntary immobility. And I say this because it really has implications for how we link extreme weather and human responses. Also on the attribution side, and I also say this because in the aftermath of a flood, storm or drought, humanitarian organizations, non governmental organizations, even governmental organizations assess disaster areas.

[00:03:23.21] 

And often what's counted are the number of damaged houses. And then an estimate of people displaced is kind of put in a model in a database. But what we actually don't know is how many people are unable to to leave their houses when hit by an extreme weather event. And we've done a perspective piece on this in Nature Communications published in March 2025. And I think it's really important to recognize this also from linking extreme weather, climate change, and also these human responses that are less predominantly discussed amongst the community of policy and also science.

[00:04:01.16] 

Let's keep going then, I guess how do we know whether a specific drought or flood or hurricane or whatever is a product of anthropogenic human claws? Climate change vs normal typical variability in the weather?

[00:04:13.20] 

Yeah. Attribution science brings actually the tool set for that. It's a rapidly growing field within climate science. And it, to put it in a nutshell, rigorously measures how climate change affects the frequency and intensity of extreme weather events. And kind of, here's how it works. So scientists track changes in the environment, the ocean, glaciers, forests and the atmosphere. And that's basically answering what is happening. And where attribution science comes into is the why. So did climate change influence this drought, this flood, all sorts of extreme events you can think of. We as climate scientists also account for that many factors drive the frequency and severity of, of extreme weather events and impacts on populations, for instance, migration and displacement, but also infrastructure changes, the economy. So think of loss and damage in that. And attribution science checks a possible list of explanations using computer models and a lot of data, weather data, climate data, but also demographic, population data, in order to understand the kind of fingerprint of climate change within the disaster itself.

[00:05:32.07] 

Is the process the same? To nerd out a bit, is the process the same for different kinds of extreme weather events? I mean, if you're trying to attribute climate change impact on a drought versus a hurricane versus a wildfire, is it the same process or does the process change?

[00:05:47.02] 

You're pointing to a really important issue we're facing in our daily lives as climate scientists doing attribution studies. Unfortunately, we can't treat all extreme weather events equal. Droughts are much harder to assess on the one hand side, because we don't really easily know when a drought starts like a flood, you see it immediately. A lot of water masses come and flood a house. You see the immediate effect and also the aftermath of it. But for a drought, it's a slow process, slowly rainfall is deteriorating, water resources deteriorating. But what's actually the point where you say like, okay, a drought has started, people have been impacted by it with drought and population impact. It's also an interesting angle to look at it from an attribution side of things, because people when hit by a drought, actually try to find ways to cope with, for instance, lack of rainfall or deteriorating water resources sources. So you won't see immediately an effect of drought on populations. You won't see like, okay, a drought started, people are moving. We'd much rather see coping mechanisms, adaptation measures in place, sometimes even known for centuries in some, for some populations, some communities.

[00:07:09.05] 

And an example for that is Somalia and East Africa, where we have a predominantly mobile population, custodial pastoralists, moving from, you know, water resource A to water resource B. However, with increasing droughts, both in severity and intensity and frequency, we see pastoralists moving to not so friendly other territories and using water resources or other natural resources. And that creates conflict. And this is really the aftermath of, you know, lagged effects of a drought. We, we see that they're trying to cope. But it's also different to, you know, maybe a country in, in the, in the global north, where we have maybe better infrastructure or infrastructure investments to help those populations affected by a drought, for instance, to cope with extreme weather. Populations in the global south don't have that kind of luxury to, to have, you know, alternative ways of infrastructure or livelihoods in general.

[00:08:16.21] 

I know you've done a bit of research on Somalia and kind of the Horn of Africa, East Africa in general. Can you talk about an episode of research you did in which attributing a specific climate extreme weather event to human caused climate change and kind of how you did that and what that looked like.

[00:08:32.21] 

Maybe I should take a step back and explain a bit the data situation both in terms of weather and population impacts in East Africa, specifically Somalia. So in East Africa we don't have perfect data in terms of weather and climate. In Somalia specifically, we have barely a weather station working. We as climate scientists are reliant on what's called reanalysis data from satellites and kind of processed weather observations. That's interesting and it's working for attribution studies. However, then we also have the big hurdle of how do we actually know that a specific extreme event, like a flood for instance, is occurring and actually impacting populations on the ground. There's not, you know, Somalia is impacted by and affected by conflict, by poverty in 2020-2022, also by a locust crisis. So there's just a lot going on on climate change is, is happening or might be happening on top of all these vulnerabilities and making people even more vulnerable. However, to carve out what's happening on the ground of a country that is not very accessible, to put it in very, you know, blunt terms, it's quite difficult. So we have a heavy reliance on, on data.

[00:09:57.05] 

And I want to put this out as a caveat because it's it nice to have attribution as a tool and go to tool for instance, to really understand both economic and non economic losses from climate change. However, we're so reliant on data that it's also incredibly important to talk to populations on the ground. And that's difficult, right? If a country is impacted by conflict from an accessibility point of view, but also as a researcher to go to a community, ask a couple of questions and then leave, that's not very satisfying for both sides, not for researchers and not for the affected community. So I want to put this out as, as a caveat before I start explaining or giving an example of what we've done in terms of studying the frequency intensity of extreme events in East Africa and Somalia and the human responses in terms of displacement on the ground. And when I say displacement displacement, I really mean displacement within the country. So not crossing any borders. So internal displacement. Yeah. We have observed long term changes in climate that essentially change the occurrence of extreme weather circumstances in East Africa. And in addition to the occurrence, climate change also affects the magnitude of these events.

[00:11:17.18] 

And sometimes this reflects in shifts in seasons. And for example, East Africa, Somalia usually knows time periods of two rainfall and two dry seasons. And this is kind of how agriculture also operates. But with the impact of climate change, local weather conditions also experience a shift which then causes more extreme dry and extreme wet weather. And this has implications for populations on the ground, both in terms of displacement but also in terms of agricultural output. So dry spells have major impacts on populations, for example, through, you know, farming and pastoral agriculture. And another example in this estimation are mobility outcomes. Again, very heavy focus on on data collected through local partners from UN institutions and the Norwegian Refugee Council. And in addition to economic losses with attribution science, we actually study whether climate change was a game changer in people getting displaced from extreme weather. One of the more tricky parts I found, in addition to using attribution for just, and I put just in quotes, an extreme weather event is to create what's referred to as a counterfactual scenario of displacement. So think of a world without climate change. That will be a counterfactual scenario in attribution science and attribution studies.

[00:12:50.12] 

But how would displacement look like if an extreme weather event had not occurred, or had occurred, but not in the same intensity, not in the same frequency, for instance? And this is really a tough job to do. We have limited data, let's say from 2019 to 2025. How do you know what's happening, you know, in 1960, for instance, especially in countries where census data is not that important, I would say, or readily available. So this is kind of a limitation I'm facing in my, in my current research. And I'm really reliant on talking to people and also for these people to create narratives of risk and how vulnerabilities change over time, to understand intervention points, how we can help them. In the absence of having the perfect answer, was climate change a game changer or not? But rather to shift the focus, how can we actually prepare populations ahead of the next extreme weather event?

[00:13:54.10] 

You've mentioned several things in there that I want to touch on, but I guess very briefly, I want to go back to this idea of counterfactuals, which I think is really interesting. In a, a low information setting or a setting like Somalia, where there's minimal amounts of data about people, about impacts, that's one challenge. But even in northern Europe and North America, where there's lots of data, this process is kind of the same, right? You make a counterfactual of what would happen in a world without climate change, and then you compare that to the world that did exist, and you kind of compare those two. Like, what does that look like? How do you, I mean, you're still kind of imagining a world that does not exist, right? What are the kinds of data that you would use for that to create this kind of imaginal imagination world? And how do you know that that's right?

[00:14:37.12] 

The first step would be to look at the climate data and weather observation and taking out all the greenhouse gas emissions that humans have put in the air since the Industrial Revolution. So that will be the kind of first step in attribution science to create a counterfactual with impact attribution. So looking at the impacts from climate change, so taking a step beyond attribution Science, normal attribution science would then also account for how would populations be shaped and in our case, migration movements look like without climate change, without a, an extreme weather event. And we can do this in multiple ways. One way I found particularly compelling is creating scenarios of how an extreme weather event would have looked like in a world without climate change. So taking a hypothetical scenario of for instance, a flood not occurring or the flood depth being lower than it was, creates different narratives for how a flood impacts or no flood impacts human mobility, in this case displacement. So with these kind of scenarios, we're able to reconstruct a model replicating what might have happened without climate change. In the absence of having perfect or good data, I'm also bringing this up because what we do in terms of these modeling exercises is putting out a message for policymakers.

[00:16:09.12] 

You know, without climate change, you would actually have more money, for instance, for investment in X, Y, Z. You know, and, and look at if, you know, a flood would have happened to a lesser extent, even then you would have more free money, so to say, to cope with, you know, I don't know, like agricultural impacts that haven't gone so well as you might have planned for in your, in your household budget and things like that. And I think this is really important message, what attribution study can do and cannot do and where there are limitations.

[00:16:41.04] 

I want to keep going on that kind of the message to policymakers or the, I guess the practical applications you mentioned there and you've mentioned previously in this conversation, the idea of losses, especially economic losses due to climate change. And when we talk about losses and losses and damages, we tend to talk about both economic and non economic. If you lost a house, that's X amount of dollars. If you lost a religious place of worship or something that can't quite be quantifiable. But with the economic losses, I guess is there. And as we look to a world in which the UN has started to propose this, a loss and damage mechanism in which individual countries or communities are repaid financially for the losses due to climate change as we look towards that world, is attribution science, is there a role for attribution science in, in that system, in that calculus, in that algorithm of payouts ultimately for losses and damages?

[00:17:33.03] 

Yes, absolutely. I think in three ways. One is from an analytical standpoint where we know that climate change means a warmer atmosphere which can cause for instance, higher intensity of rainfall, more severe flash floods. And in most cases we can identify an increase in likelihood and impacts of flood that results and then also impacts on infrastructure and so on and so forth. So I see as a researcher, I see attribution signs as a tool and also kind of policing and putting a dollar sign on a particular extreme event in the context or attributable to climate change. So really creating the causal link. And that brings me to my second point where I also see that attribution science is helpful on a legal side of things. In court, for instance, where when you can actually prove a causal link between two things, you have a case. Right? So there's this example from a Peruvian farmer who rightfully said climate change has been reducing the glacier amounts. And you know, it's impacting my livelihood, it's impacting my, my farming outputs in general. And as far as I know, the legal side of that or the case is still ongoing.

[00:18:52.03] 

But so far he has been quite successful in his case and making the energy company RWE in Germany responsible and holding that company responsible for the losses he and his community have occurred. And that's an important message that goes across building bridges from science to actually the legal branch of, you know, outputs of our daily lives. And that's creating a big message. And as a third point, I think in terms of policy, it also attribution science is also useful in terms of creating awareness and maybe that goes beyond the policy side of things also to the community to understand really the implications and the causal impacts of climate change goes a long way to also bridge economic and non economic losses that are, these are intertwined. It's not just like, you know, an extreme weather event just causes economic impacts, a bridge that is destroyed. But it also means that, you know, communities are affected and we need to take care of that because those people will have to rebuild that bridge. And if there's no one to rebuild the bridge, your economic, you will probably have an economic fallout, right? You can't transport your goods and so on and so forth.

[00:20:10.19] 

So I'm, you know, I'm creating like a very specific example.

[00:20:13.10] 

No, yeah, yeah, that's great.

[00:20:14.10] 

This is, this is something we've seen in the context of drought in Madagascar where the southern part of Madagascar is so dependent on a already poor infrastructure to put their, their, to bring their goods from the south to the north and ship it essentially across and, and, and do business, right? And those people in the global, in the global, in the grand suit of Madagascar were not able to put their, their agricultural outputs or other types of products to the northern part of, of Madagascar because the infrastructure was inaccessible. And this is really, this is a strong message for for attribute or the impacts of climate change and impact attribution in general. And if we know these results from attribution studies, we can prepare populations. We know who's affected and how they're affected, and we can prepare those populations vulnerable to the impacts of climate change ahead of the next extreme weather event, which we know will occur.

[00:21:15.13] 

I want to talk more about that, the kind of, the predictive and kind of preparatory aspects of this. I mean, then say more about that. What can you do? How does we think of attribution as retrospective? Often, but this is also prospective or like looking into the future. How does that work? How can we prepare for the future using this kind of science?

[00:21:34.19] 

So part of the toolset attribution science provides through statistical methods is also estimating how many times in the future such an extreme weather event of the same magnitude would occur, let's say in 2050. And this is also another strong message. Again, these are models obviously, so we don't really know that this is like point to point happening in 2050, but it gives us a good estimate to prepare ahead of time and really positively jump the gun and prepare populations, prepare also infrastructure, the economy, or if you think about agriculture, plant the right crops. If we know for instance in six months time there's a drought coming, you'll probably plan different crops and seeds compared to a drought not happening. Or to a lesser degree, very interesting stuff.

[00:22:30.06] 

I think we're just about out of time, so we probably have to wrap it up there. But Lisa, thank you so much for coming on. This was super interesting. Thanks for your time.

[00:22:37.15] 

Thank you. Love talking to you. Thank you.

[00:22:39.19] 

Lisa Thalheimer is a research scholar at the Migration and Sustainable Development Research Group at the International Institute of Applied Systems Analysis, also known as IEASA. Thank you for listening to Changing Climate, Changing Migration. Make sure to catch all of our new episodes by subscribing to the podcast on Apple Podcasts, Spotify or wherever it is you get your podcasts. And please leave us a review which makes it easier for other people to find us. You can find every episode of this podcast on our website at migrationpolicy.org/podcasts. On our site, you can also dive into the full catalog of MPI research on climate migration, including a special series of articles in our Migration Information Source online magazine. That's at migrationpolicy.org/climate. Drop me a line at [email protected]. I'm always happy to hear from listeners. This episode of the podcast was produced by Elizabeth Navarro. Michelle Mittelstadt provided editorial oversight and. And additional assistance came from Lisa Dixon. The theme music you're hearing is Touch by Patrick Patrikios. I'm Julian Hattem. Thank you again for tuning in to Changing Climate, Changing Migration. See you next time.