- If you live in the southeast, you're likely to die a decade or so earlier than most other regions.

Why?

Well, if you look at deaths from natural disasters at first it sure doesn't explain it.

I was really surprised to see that deaths from natural disasters are going down with climate change, making these disasters so much worse and so much more expensive.

I would've thought the opposite.

Well, it turns out the official numbers in this graph are hiding a much bigger story, so big that the scientists who uncovered it didn't believe the data.

- We thought we made a mistake.

We were like, this can't be true.

And we spent like five years trying to find what mistake we had made before we would decided, okay, this is probably actually true.

- And not only do these results fundamentally alter our understanding of deaths from climate disasters, they may solve this longstanding mystery in public health.

Natural disasters used to be incredibly deadly compared to today's standards, especially floods and droughts.

The 1931 China floods killed between 2 to 4 million, mostly from drowning and famine.

It was likely the worst natural disaster in recorded history.

And just before the floods, China experienced one of its worst droughts, which killed up to 3 million in the famine that followed.

These disasters were of an unimaginable scale.

For comparison, the deadliest disaster in the last 50 years in the US was Hurricane Maria with an estimated death count of more than 4,600.

We've come a long way since 1900.

Before global trade networks, it was common for famines to follow natural disasters, killing millions of people.

Today, those kinds of famine-related deaths have become incredibly rare thanks to better infrastructure, global food aid and international cooperation.

The biggest bubbles in this graph representing millions of lives lost are largely a thing of the past, and that's no small feat.

So if it's true that overall disaster deaths have fallen dramatically, then why have the number of billion-dollar disasters in the US risen over the last several decades?

I mean, in just 2024 and 2025 alone, we saw hurricanes Helene and Milton devastate the southeast, the historic LA fires and deadly flooding in Texas.

So I went to the source of the data, Debarati Guha-Sapir, who created EM-DAT, often described as the most widely used global database on natural disasters.

- I won't put my hand on the fire about the completeness of our data mortality.

Whether or not we are capturing everything is a question mark in my mind.

- As EM-DAT's creator, Debarati knows the data better than anyone and she's identified the main issue when it comes to what we know about deaths from natural disasters.

- So one issue is whether or not the data we are using, which shows the mortality is decreasing, is actually reflecting reality.

- So basically is the data complete.

Debarati's data measures total deaths, which includes what are called direct and indirect deaths.

As you'll see, this is an important distinction.

- When we talk about direct deaths, we're talking about specific traumas that have occurred as a result of the event itself.

So you can imagine people are on boats and they drown, or the winds from a cyclone cause a tree to get knocked over and they fall on a person.

- The good news is we've gotten better at preventing these direct deaths, but indirect deaths, which can happen long after the disaster has ended, like families getting sick from contaminated drinking water after a hurricane are, notoriously difficult to measure.

Which is why graphs like this can struggle to show the full picture.

But how do you measure what doesn't leave a clear fingerprint?

Epidemiologists have wrestled with this for years.

Take extreme temperature for example.

- If you look at epidemiological studies that look at year to year variations in temperature and mortality, these relations become very clear.

When it gets hot, you just see these really quick upticks in how many people are dying.

- When it gets hot, your heart has to pump a lot.

It's harder to cool you down.

And so a ton of people have heart attacks, but we don't usually write it down as like heat stroke.

We write it down as heart attack.

So in the in the data, it's like misdiagnosed.

- But Solomon and Rachel wanted to know if the same kind of hidden death toll also shows up for tropical cyclones.

And that's a tough question because to measure the real impact of hurricanes, you need to compare against a world where no hurricanes ever happened.

That's a tall order since hurricanes have always defined life in the region.

- It's like a bell and you hit it and then it's ringing for a long time.

And so if you hit it again before it stops ringing, the thing you hear is like the super position of those two hits and then you can hit it again and again and soon it's like, can I figure out what hitting the bell once sounds like?

Every state is getting hit by another hurricane before it stops ringing from the last hurricane.

And so what you're seeing in the data is this kind of like really noisy pattern of a society that's kind of just ricocheting from storm to storm to storm.

- They looked at every tropical cyclone, all 501 of them that impacted the continental United States from 1930 to 2015.

According to official government statistics, these cyclones led to a total of around 10,000 direct deaths or an average of 24 deaths per cyclone.

But they found that hidden underneath that number is an unbelievably larger one.

- When states are hit by hurricanes, people dying at an accelerated rate for 15 years, there's this like really, really long shadow to the storms where basically you know the effects are reverberating through society for a long time.

- So an average storm generates about 24 excess direct deaths.

And we're saying that if you account for all these indirect deaths, these tropical cyclones are actually causing between seven and 11,000.

- Seven to 11,000 deaths instead of 24.

That's 300 times more deaths.

- We thought we made a mistake, we were like, this can't be true.

And we spent like five years trying to find what mistake we had made before we decided, okay, this is probably actually true.

- We spent many years trying to make sure that it wasn't some kind of fluke in the data.

It wasn't sort of a missed model specification because these kinds of results could really impact how we think about disaster resilience and investment into the future.

- I was also curious if once you take into account all of these indirect deaths, are tropical cyclones getting more or less deadly over time?

- We see like no reduction in lethality over a hundred years, really like zero.

- So if the true average death toll from a single cyclone isn't 24, but more like 10,000, then how many people have died in the US from cyclones in total?

- So we did estimate the total mortality burden of all of the cyclones that have hit the United States between 1950 and 2015, and that's between three and 5 million deaths.

- More people have been killed by hurricanes than have died in battle for the United States.

That includes World War I and World War II and the Civil War.

- So if instead of thousands, millions of people have been killed by cyclones in the US, then who is dying?

- If you talk to anyone in public health, they would tell you that if you look at a map of life expectancy in the United States, there's a real bias where in the southeast US people just don't live as long and there's higher mortality rates in the Southeast.

And it's been kind of a mystery, like why is that?

- When Solomon and Rachel created their model and removed cyclone activity from the southeast, they uncovered something really unexpected.

- It turns out that if you remove the effect of hurricanes, it removes a lot of that effect for people that are under 45.

What that's basically saying is that a lot of the southeast bias in mortality for at least people under 45 could possibly be assigned to hurricanes.

It looks like hurricanes are like distorting life expectancy across the United States.

- And some age groups are particularly affected by cyclones.

- A lot of it is infants and people that are over 65 that make up a large fraction of the deaths and infants in particular have a very high risk.

- But the reasons for the particularly devastating impacts on infants aren't what you might think.

- So the storm is causing infant mortality to be higher years and years after, for infants that were not in utero, did not physically experience the storms themselves.

- So tropical cyclones are killing a lot of people who didn't even exist when the cyclone hit for more than a decade after the cyclone has passed.

So what's happening here and why haven't overall indirect deaths gone down at all?

- I think what's happening is that people are doing stuff to protect themselves from the dangers that they know about.

And when people don't realize that they're facing so much danger, they're not taking any preventive action.

And so as a society, we've essentially done nothing to protect ourselves from these indirect deaths.

- So we've gotten better at preventing the deaths that we can see from things like wind flooding and storm surge.

But we haven't gotten better at preventing the hidden deaths, which actually feels really surprising given how much we've progressed in the last century.

So what's going on here?

Well, Solomon and Rachel actually think they have a pretty good idea.

- One is the state and local governments have a reduced capacity after a storm, and that persists for a long period of time.

- There's like never money set aside to rebuild after a storm.

So then people rework the budget.

Usually what that means is you're taking it out of something that is an investment in the future because those people in the future are not around to advocate for themselves.

- The second massive factor is economic.

So if you're constantly being hit with these economic shocks and you are always happening to rebuild your garage or you know repair your roof, that's money that you could be spending on different kinds of food, different kinds of healthcare services.

- We think social networks are a big one too.

Storms can cause people to move around a lot.

- This could be particularly important, a driver for our elevator rates of mortality amongst the elderly population and also potentially for the infants.

So you might be relying on particular care resources and then people sort of relocate or they have to change jobs and they're no longer as available in a time of emergency.

- The fourth thing is kind of maybe the more obvious thing, which is changes the environment.

- Places are getting hit.

People could be exposed to kinds of toxic waste or chemicals that they wouldn't have been exposed to otherwise.

- And then the fifth thing is stress.

- We've seen more recent studies that show that stress is a really big predictor of your long-term mortality and long-term health.

And there's been studies that have shown that hurricane related stress causes elevated rates of mortality.

- Now, if you don't live in the southeast and you think this research doesn't affect, you think again.

- The direct damages and deaths from most disasters look like they're declining.

And it looks like the indirect or the hard to detect ones are not declining.

And that is like a pretty clear pattern across like a lot of things.

- And that same pattern shows up with heat.

It's easy to think that heat is not such a big deal 'cause it's far less sort of visually dramatic than something like a hurricane.

But heat is a really substantive and silent killer - For fires, it's a similar story.

The smoke just kills so many people.

- A recent study found that pollutant reductions from the Clean Air Act are for the first time actually reversing in many parts of the US due to wildfire smoke.

In many western states, wildfire smoke has eroded over 50% of the progress made, and this problem is only projected to get even worse as our climate warms.

So when we zoom out and look at these less visible long-term deaths, it becomes clear just how much more we really need to be doing to support communities in the weeks, months, and even years following a disaster.

- Our study shows that the impacts are much greater and persist for much longer than we had previously thought, suggesting that there's a lot more that we can do to help communities recover.

- And I think the important message that we've really learned from all of the research in the last couple years is that that number, which we call the direct damage, it's literally like the stuff that's destroyed during the storm.

It can be very big, but the vast bulk of the harm from hurricanes actually emerges like in the years after.