Climate Change and Hurricane Intensity: What the Data Shows

As global sea surface temperatures rise, are hurricanes becoming more powerful? The relationship between climate change and hurricane intensity is a consequential and genuinely difficult question, and the honest answer is nuanced: some trends are now well established in the data, while others remain areas of active research and debate.

Framing the Question

The scientific community distinguishes between three interrelated aspects of hurricane behavior: frequency (how many form), size (geographic extent), and intensity (maximum wind speeds). Climate change is predicted to affect all three, but the evidence and mechanisms differ significantly for each. This article focuses on intensity, the maximum sustained wind speeds that a hurricane can achieve, because this is where the clearest observational signals emerge and where the physics is most straightforward.

One point is worth making early. The public conversation fixates on whether storms are becoming windier, which is the hardest signal of all to isolate. The clearer changes, and arguably the more consequential ones for damage, are in rainfall and in storm surge riding on higher seas. Understanding what has already changed requires accounting for several complicating factors: measurement bias over time, natural variability from ocean cycles, and the inherent difficulty of separating human-caused climate signals from background noise.

A series of vertical stripes shading from blue to deep red, showing global average temperature rising from 1850 to recent years — Each stripe is one year's global average temperature, from 1850 (left) to the present: blue for cooler, red for warmer. The unmistakable shift to deep red is the warming that, through the oceans, is reshaping hurricane behavior. Credit: Ed Hawkins, University of Reading · CC BY-SA 4.0

Why Warmer Water Matters

The fundamental physics is well understood. Hurricanes extract energy from warm ocean water, much like a heat engine. The warmth of the sea surface temperature (SST) is one of the primary thermodynamic constraints on how intense a hurricane can become, a limit called the potential intensity. This relationship is grounded in the Clausius-Clapeyron equation, which describes how atmospheric moisture content increases with temperature. For every 1 degree Celsius of warming, the atmosphere's saturation vapor pressure increases by approximately 7 percent.¹

Emanuel's (2005) seminal work on potential intensity demonstrated mathematically that hurricane strength is tied to the maximum enthalpy (heat energy) available from the ocean. As sea surface temperatures warm, the theoretical upper limit on hurricane intensity increases.² This is not controversial among atmospheric scientists; the mechanism is sound and the mathematics robust.

A world map showing 2023 temperatures running far above the long-term average across most of the globe and oceans — Global temperatures in 2023 relative to the 20th-century average, deep red where it is warmest. Because hurricanes draw their energy from warm seas, this ocean warmth raises the ceiling on how intense a storm can become. Credit: NASA Scientific Visualization Studio · Public domain

Key Concept: Potential intensity is the theoretical maximum wind speed a hurricane can achieve given sea surface temperature, mid-tropospheric air temperature, and atmospheric moisture. Warmer SSTs allow higher potential intensity values, creating a more favorable thermodynamic environment. However, potential intensity is not the same as actual intensity: storms must still develop, organize, and survive in that favorable environment.

The Observational Challenge

Here is where the picture becomes complex. The satellite era, when we have direct wind speed measurements via satellite, began around 1966, providing roughly 60 years of reliable data. This window is long enough to detect some trends, but not long enough to fully separate climate signals from natural variability cycles like the Atlantic Multidecadal Oscillation (AMO).

HURDAT2: the historical record

The most comprehensive Atlantic hurricane dataset is HURDAT2, maintained by NOAA's National Hurricane Center. Historical records before the satellite era (pre-1966) relied on ship reports, coastal observations, and barometric pressure readings. These early records have systematic biases: weaker storms were underreported, landfalling storms over sparsely populated regions were missed, and measurement techniques improved significantly over the 20th century. Any trend analysis must account for these detection biases.

What satellite-era trends show

From 1966 to present, the data reveal a detectable increase in the proportion of Category 4 and Category 5 hurricanes in the Atlantic basin.³ Studies using satellite-derived intensity estimates (which measure cloud-top temperatures rather than direct wind speed) show a shift toward more intense storms. However, the absolute number of hurricanes per season has not shown a consistent increasing trend, which is crucial. If frequency were constant but intensity were increasing, this would point to climate factors. If both were increasing in lockstep, attribution becomes harder.

Kossin et al. (2020), analyzing global satellite data from 1979–2017, found evidence that the global proportion of rapidly intensifying hurricanes has increased and that storms are reaching their peak intensity at higher latitudes.³ These are important findings, but they come with caveats: satellite measurement techniques have evolved, and global data are more limited than Atlantic data.

Detection bias and its implications

A major challenge: improvements in observational technology mean we detect weaker storms now that would have been missed in earlier decades. Kossin (2024) and others have documented that changes in observational capability can create false trends in the historical record. A storm measured today with modern Doppler radar might have gone undetected in 1970. This detection bias inflates the apparent proportion of strong storms in recent years.

The Bias Problem: It is fundamentally difficult to know whether an increase in Category 4–5 hurricanes reflects true intensification or better detection. Researchers attempt to correct for this using homogenization techniques, but these are imperfect and model-dependent.

Rapid Intensification: A Clearer Signal

One of the most robust observational findings involves rapid intensification (RI), a hurricane strengthening by at least 35 mph (55 km/h; 30 kt) in 24 hours. Rapid intensification events are sudden, dramatic, and pose major challenges for forecasting and coastal preparation.

Wehner et al. (2022), using carefully homogenized satellite data, found that the global proportion of hurricanes undergoing rapid intensification has increased over the satellite era, consistent with climate model predictions. This signal is stronger than the signal for overall intensity change, suggesting that the primary climate change effect may be on storm behavior (how quickly they can strengthen) rather than on the absolute upper limit of intensity.

A satellite image of Hurricane Otis intensifying explosively into a compact, powerful storm near Mexico's Pacific coast — Hurricane Otis (2023) exploded from a tropical storm into a Category 5 in roughly a day before striking Acapulco, a textbook case of the rapid intensification that is becoming more common over abnormally warm water. Credit: CIRA / NOAA · Public domain

The mechanism is plausible: in a warmer ocean, storms encounter a higher heat flux, allowing for faster intensification. Additionally, reduced wind shear (a difference in wind speed and direction with altitude) in certain warming scenarios can also promote rapid intensification. These are mechanistic explanations grounded in hurricane physics.

How Sea Level Rise Compounds Storm Surge

Even if future hurricanes are only marginally more intense, rising sea levels dramatically amplify damage. A hurricane with the same wind speed striking a coastline that is 1 ft (30 cm) higher due to sea level rise will experience less friction as it approaches shore, allowing the storm surge (the rise in sea level caused by the hurricane) to extend farther inland and inundate areas previously protected.

Sea level is rising at approximately 0.13 in (3.3 mm) per year globally, and faster in some regions due to subsidence and ocean dynamics.¹ By 2050, many coastal areas will experience an additional 1 ft (0.3 m) of sea level rise beyond current levels. This means that a "100-year flood" under current sea level may become a decadal or more frequent event by mid-century. The impact is equivalent to the storm surge being stronger, even if the hurricane itself is unchanged.

A line graph of global average sea level rising steadily over the past century — Global mean sea level has climbed steadily for over a century and the rate is accelerating. Every inch of that rise becomes a higher launch pad for storm surge, pushing the same hurricane's floodwaters farther inland. Credit: Robert A. Rohde / Global Warming Art · CC BY-SA 3.0

This compounding effect has received less public attention than intensity changes, but it may ultimately be more consequential for coastal risk, particularly for low-lying areas and island nations. The mechanics of storm surge explain why even a small rise in the baseline matters so much.

Rainfall: The Clearest Signal

One area carries less uncertainty: the relationship between warming and rainfall. The Clausius-Clapeyron relationship implies that storms in a warmer atmosphere will contain more moisture. Multiple studies have documented that precipitation rates in tropical cyclones have increased in recent decades, with expected increases of 3–10 percent per degree Celsius of warming.¹

A satellite rainfall map showing extreme accumulations from Hurricane Harvey centered over Houston, Texas — Hurricane Harvey's record rainfall over Houston (purple). A warmer atmosphere holds more moisture, about 7% more per degree Celsius, so storms like Harvey can wring out heavier, more concentrated rainfall than they once did. Credit: NASA (IMERG) · Public domain

Knutson and Ploshay (2016) analyzed GFDL simulations and found that while wind speed may not change dramatically, rainfall amounts are projected to increase substantially with warming. Observed rainfall from hurricanes like Harvey (2017), which produced 60 in (1,525 mm) in parts of Texas, and Eta (2020) are consistent with this projection, though attribution of any single event to climate change remains difficult.

From a practical perspective, rainfall-driven flooding is now a leading cause of hurricane mortality in the United States, surpassing storm surge in recent years. Climate change may be having a more significant impact on this hazard than on wind speeds.

What Remains Uncertain

The scientific consensus has solid ground beneath it on mechanisms and some observed trends, but significant uncertainties persist.

Frequency, or the lack of a clear trend

The global frequency of tropical cyclones is not clearly changing in observations. Some studies detect a slight decrease, others find no trend, and some regional studies find increases. Climate models do not predict a large change in frequency, and observations are consistent with this. However, there is a spatial shift: storms that form in some regions may be migrating poleward due to changing atmospheric conditions.

The attribution challenge

Attributing a specific observed trend (like the increase in strong storms) to human-caused climate change requires comparing the observed record to climate models run with and without human greenhouse gas forcing. These models have their own biases and uncertainties. While the IPCC AR6 assessment (Seneviratne et al., 2021) concludes that it is "likely" that the global proportion of very intense cyclones has increased due to human influence,¹ this conclusion rests on models more than on observations alone, because the observational record is too short and noisy to provide a definitive signal.

Natural variability: the AMO and ENSO

The Atlantic Multidecadal Oscillation (AMO), a long-term cycle in Atlantic sea surface temperatures, has a period of 60–80 years. Over a 60-year satellite record, detecting a climate-change signal separate from a cyclical AMO signal is statistically difficult. The El Niño Southern Oscillation (ENSO) modulates wind shear from year to year, affecting intensity potential. Both natural oscillations must be accounted for to isolate anthropogenic trends.

Model uncertainty and grid resolution

Climate models used to project future hurricane intensity vary in their predictions. Some models project modest increases (10–20 percent), while others suggest larger changes. Much depends on model resolution: high-resolution models (which explicitly simulate individual storms) give different answers than coarse-resolution models (which parameterize storm effects). As computing power increases, model predictions converge somewhat, but uncertainty remains.

In our own numerical-modeling work, the resolution problem is not abstract. A coarse grid smears a hurricane across cells too large to resolve an eyewall, so it systematically understates peak intensity. That limitation is a large part of why intensity projections still disagree, and why higher-resolution models tend to produce stronger storms.

The IPCC AR6 Assessment

The Sixth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR6, 2021) provides the most authoritative synthesis. Key findings from Working Group I, Chapter 11, include:

It is virtually certain that ocean heat content and upper ocean temperatures will continue to increase through 2100 under all emissions scenarios, providing a warmer environment for hurricanes.
It is likely that the global proportion of Category 4–5 hurricanes will increase with warming, based on both theory and models. The magnitude of this change is projected to be 1–10 percent per degree Celsius of warming, but the range of uncertainty is large.¹
It is very likely that precipitation rates in tropical cyclones will increase, a consequence of the Clausius-Clapeyron relationship.
Uncertainty remains regarding global tropical cyclone frequency and the spatial distribution of storms. Some models project a poleward migration; others do not.
Rapid intensification events are likely to increase, though the magnitude is uncertain.

Confidence Levels Matter: "Likely" in IPCC terminology means 66–100 percent confidence; "very likely" means 90–100 percent. These are not certainties, but they reflect the weight of evidence. The use of probabilistic language acknowledges genuine scientific uncertainty.

What Does This Mean for Preparedness and Planning?

From a practical standpoint, several implications emerge:

Intensity matters for wind damage: a 10–20 percent increase in the proportion of the strongest storms translates to meaningfully more damage from winds, even if the overall frequency stays constant.
Rainfall flooding is increasingly critical: hurricane-driven rainfall is now the leading precipitation hazard, and climate change is making this worse, independent of intensity.
Storm surge is compounded by sea level rise: for coastal planning, the combination of intensity, rapid intensification, rainfall, and rising sea levels creates a multi-factor risk that is substantially higher than any single factor alone.
Uncertainty cuts both ways: climate models project more intense storms, but the magnitude is uncertain. This does not justify inaction; it justifies prudent planning that accounts for a range of possibilities.

A Note on Individual Events

Climate scientists are often asked: "Was Hurricane X caused by climate change?" The answer is almost always: "Climate change didn't cause this specific storm, but it made certain aspects more likely or more severe." Hurricanes require specific atmospheric and oceanic conditions to form; without those, no storm occurs. What climate change does is alter the background environment, making certain intensities more probable, rainfall rates higher, and storm surge from a given wind speed more damaging due to higher baseline sea levels.

This distinction may seem semantic, but it is crucial for understanding attribution. A hurricane is not "caused" by climate change the way a car accident might be caused by a drunk driver. Rather, climate change loads the dice, making stronger storms and wetter storms statistically more likely.

Seneviratne, S. I., et al. (2021). Weather and climate extreme events in a changing climate. In Climate Change 2021: The Physical Science Basis (Working Group I, IPCC Sixth Assessment Report), Ch. 11. Cambridge University Press. https://www.ipcc.ch/report/ar6/wg1/ ↩ ↩² ↩³ ↩⁴ ↩⁵
Emanuel, K. (2005). Divine Wind: The History and Science of Hurricanes. Oxford University Press. https://doi.org/10.1093/oso/9780195149418.001.0001 ↩
Kossin, J. P., Knapp, K. R., Olander, T. L., & Velden, C. S. (2020). Global increase in major tropical cyclone exceedance probability over the past four decades. Proceedings of the National Academy of Sciences, 117(22), 11975–11980. https://doi.org/10.1073/pnas.1920849117 ↩ ↩²