That’s indeed good news—and it’s backed by robust long-term data. Global death rates from extreme weather events (storms, floods, droughts, heatwaves, etc.) have declined dramatically over the past century, even as the planet has warmed and population has grown.

According to data from the International Disaster Database (EM-DAT), compiled by the Centre for Research on the Epidemiology of Disasters (CRED) and visualized by Our World in Data:

• In the early 20th century, weather-related disasters often killed hundreds per million people annually.
• By the 2020s, this has fallen to fewer than 10 deaths per million people per year—a drop of over 95%.

---

From Watts Up With That?

Anthony Watts

If you rely on today’s headlines, you would be forgiven for believing that extreme weather is becoming ever more deadly and that humanity is steadily losing its ability to cope.

Floods, storms, and heat waves are routinely framed as evidence of a growing “climate emergency,” often accompanied by claims that climate change is “juicing” weather extremes and driving escalating loss of life. Yet when we step back from the headlines and examine the long-term data, a very different and far more encouraging picture emerges with a new peer-reviewed study just published Empirical evidence of declining global vulnerability to climate-related hazards by Giuseppe Formetta and Luc Feyen.

They write:

“The most striking feature of modern climate history is not worsening vulnerability, but a dramatic collapse in weather-related mortality.”

Over the past century—and especially since the mid-20th century—the risk of dying from extreme weather has fallen dramatically. This is not a matter of interpretation or projection. It is documented in global disaster databases and confirmed by peer-reviewed research. This new research bolsters what the “Or World in Data” graph shows, seen below.

Data compiled by the Centre for Research on the Epidemiology of Disasters (CRED) and summarized by Our World in Data show that global death rates from weather-related disasters have dropped by roughly two orders of magnitude since the 1960s. In 1960, extreme weather events killed more than 300 people per 100,000 worldwide. By 1970, that figure had already fallen below 100. By 1990, it was closer to one or two per 100,000. In recent years—particularly since about 2014—global death rates from storms, floods, and wind-related disasters have routinely fallen below one per 100,000 people.

“Extreme weather still happens. What has changed is its ability to kill.”

This decline has occurred during a period when the world’s population more than doubled, when coastal development expanded rapidly, and when the total value of infrastructure and assets exposed to weather hazards increased many times over. If extreme weather were becoming more dangerous in the way commonly suggested, mortality risk should have increased alongside that exposure. Instead, it collapsed.

A common response is to dismiss these figures as artifacts of population growth or improved reporting. That objection fails once vulnerability is analyzed properly. Rather than looking at raw death counts, which are easily distorted, researchers increasingly examine mortality rates relative to the number of people actually exposed to hazardous events. When that is done, the downward trend becomes even clearer.

The comprehensive peer-reviewed study published in Global Environmental Change by Giuseppe Formetta and Luc Feyen examined global vulnerability to climate-related hazards from 1980 to 2016 using one of the most complete disaster loss databases available . The authors calculated mortality rates as a function of exposed population and economic loss rates as a function of exposed GDP, allowing them to separate changes in vulnerability from changes in population, wealth, and reporting practices.

“When deaths are normalized to exposure, vulnerability to extreme weather has fallen sharply across the globe.”

Their results were striking. Across seven major weather-related hazard types, global mortality rates fell by more than a factor of six between the 1980s and the most recent decade analyzed. Economic loss rates, when normalized to exposed GDP, declined by nearly a factor of five. These results were robust across different assumptions about the size of affected areas and were observed for floods, wind storms, drought-related hazards, and other major categories.

This matters because it directly challenges one of the core assumptions embedded in many climate-impact narratives: that increasing hazard intensity automatically translates into increasing human harm. The real-world data show that this assumption is wrong. Vulnerability is not fixed. It changes, often rapidly, in response to development, technology, and institutional capacity.

Meanwhile, the count of weather-related incidents increased:

One of the clearest findings in the Formetta and Feyen study is the strong negative relationship between wealth and vulnerability. As countries become wealthier, mortality rates fall. The steepest declines occur at low income levels, where even modest improvements—basic infrastructure, early warning systems, communications, and emergency response—produce large reductions in loss of life.

This dynamic has led to a convergence in vulnerability between poorer and richer countries over time. While significant gaps remain for certain hazards, particularly coastal flooding and wind-related events, poorer countries today are far less vulnerable than they were just a few decades ago, and in many cases they have reduced vulnerability faster than wealthier nations.

“Development, not decarbonization, has been the dominant life-saving force.”

This reality is rarely reflected in how disaster losses are discussed publicly. Rising economic damages are often cited as evidence that extreme weather is becoming more destructive. What is almost never mentioned is that these figures are reported in absolute terms. As societies grow wealthier, they place more valuable assets in harm’s way. A flood that once damaged crops may now inundate industrial parks, ports, and dense urban infrastructure. Dollar losses rise even if risk declines.

When losses are normalized to the amount of wealth exposed, the trend reverses. Proportional losses are falling, not rising. This distinction is central to understanding risk, yet it is almost entirely absent from mainstream reporting.

“Higher losses do not mean higher danger. They usually mean more wealth in the path of nature.”

The focus on absolute losses and isolated disasters feeds directly into worst-case climate narratives. Many climate-impact models implicitly assume fixed vulnerability or slow adaptation, treating societies as passive victims rather than adaptive systems. The historical record shows otherwise. Adaptation is not speculative; it is measurable, ongoing, and already responsible for enormous reductions in mortality.

Formetta and Feyen explicitly note that improved protection measures, early warning systems, and disaster risk management have counterbalanced increases in exposure, resulting in net reductions in both human and economic vulnerability . This is not a footnote. It is a central conclusion of their analysis.

“Adaptation has outpaced hazard—not the other way around.”

Yet this evidence sits uneasily alongside policy narratives that frame emissions reductions as the primary means of protecting lives from extreme weather. Emissions policy may be debated on many grounds, but the historical record is clear: reduced vulnerability has been driven overwhelmingly by wealth, technology, and governance, not by climate mitigation.

This matters because policies that restrict economic growth or limit access to affordable energy risk undermining the very factors that have proven most effective at saving lives. Making societies poorer in the name of climate protection is more likely to increase disaster risk than reduce it.

None of this implies that extreme weather is harmless or that preparation is unnecessary. Large loss-of-life events can still occur, particularly where infrastructure is weak or governance fails. The data show variability, and tragic outliers remain possible.

What the data do not show is an escalating global mortality crisis driven by climate change. The dominant long-term trend is one of extraordinary progress.

“The real story of extreme weather is not one of failure, but of resilience.”

There are limitations, of course. Disaster databases are imperfect, particularly in earlier decades, and some hazards—especially extreme temperature events—are difficult to track consistently over time. Formetta and Feyen are explicit about these uncertainties and deliberately avoid mixing incompatible datasets. That methodological caution stands in sharp contrast to the confidence often placed in long-range model projections.

Taken together, the evidence leads to a clear conclusion. Extreme weather remains a fact of life, but its capacity to kill has been dramatically reduced because human societies have become better prepared, better informed, and more resilient.

That is the real good news. It deserves far more attention than it receives. If you are a journalist covering extreme weather, there is a basic question worth asking before publishing the next “climate-fueled disaster” headline: Are people actually becoming more likely to die from these events?

The long-term data say no.

---

Meanwhile, media coverage often emphasizes the increasing frequency and intensity of events (which is true for some, like heatwaves and heavy rainfall, per IPCC and World Weather Attribution reports) and dramatic individual disasters, creating an impression of rising deadliness.

Economic damages are rising due to more people and assets in harm’s way—but lives saved show humanity is adapting effectively.