Were the record-setting 30 named storms of the astonishingly brutal 2020 Atlantic hurricane season a one-off random event? Or a portent of a new era of increased named storms for the Atlantic?

The number of Atlantic named storms has unquestionably been on the rise in recent decades. But while hurricane scientists are confident that global warming is making the strongest storms stronger, scientific debate is continuing on how global warming may be affecting the number of tropical cyclones (the catch-all term describing hurricanes, typhoons, tropical storms, and tropical depressions). Consider this literature review on what factors may be causing the recent rise in Atlantic named storms. There are, no surprise, more questions than answers, but recent research suggests that we are now in an era of more frequent Atlantic named storms.

One fact is unequivocal: while the number of global tropical cyclones of at least tropical storm strength (one-minute average winds of 39+ mph) has remained roughly constant in recent decades (around 85 named storms per year since 1970), the number of Atlantic named storms has increased significantly. Every 10 years, NOAA updates its definition of what constitutes the “normal” climate based on the latest data.  The long-term averages from the period 1991-2020 had two more named storms and one more hurricane than in the 20 years from 1981 to 2010,  the National Hurricane Center tweeted in early April, increasing to 14 named storms and seven hurricanes. The number of major hurricanes also increased, from 2.8 to 3.2 per year, but this increase was obscured because the totals were rounded.

For the past 30 years, 1991-2020, compared to the previous 30 years, 1961-1990, the difference is even more stark. There were an average of 10 named storms, six hurricanes, and two intense hurricanes between 1961 and 1990. So the period 1991-2020 had a 44% increase in named storms, 20% increase in hurricanes, and 60% increase in major hurricanes.

Figure 1. The number of named storms in the North Atlantic, 1972-2020. The year 1972 is considered the beginning of the era of reliable satellite coverage of the Atlantic. A linear trend line is superimposed. (Background image: Spiral feeder band from a NOAA P-3 flight into category 4 Hurricane Joan in 1988, taken by Jeff Masters.)

The increase in Atlantic named storms could be driven by some or all of at least five factors, roughly in order of confidence, and each discussed in more depth below:

1) A reduction in small sulfur-containing particles over the Atlantic as a result of stronger air pollution regulations in the U.S. since the early 1970s,  allowing more sunlight to reach the surface and heat the oceans;

2) A lack of major volcanic eruptions in recent decades, allowing sea surface temperatures, SSTs, to rebound from the cooling effects of the eruptions of El Chichón in 1982 and Pinatubo in 1991;

3) Improved technology that has allowed the identification of weak, short-lived tropical cyclones that would have escaped detection in previous years;

4) Human-caused global warming, which has increased SSTs in the Atlantic, and caused atmospheric circulation changes beneficial to tropical cyclone formation; and

5) Natural variability driven by decades-long cycles.

Formation of tropical storms requires warm waters, a moist atmosphere, and low wind shear. Typically, an SST of at least 26 degrees Celsius (79°F) is required. Thus, an increased number of named storms can be expected when ocean temperatures rise, if all other factors remain constant. Indeed, SSTs over the main development region for Atlantic hurricanes have risen by about 1 degree Fahrenheit (0.6°C) over the past 50 years (Figure 2), coinciding with the observed increase in named storms.

Figure 2. Departure of SST from average (in degrees F) for the tropical Atlantic’s main development region for hurricanes, from the coast of Africa to the Caribbean, between 10-20°N. SSTs have risen by about 1.8 degrees Fahrenheit (1°C) over the past 100 years, and by about 1 degree Fahrenheit (0.6°C) over the past 50 years. (Image credit: Climate Central; manually updated up to 2020 using NOAA/NCEI data)

1. Decreased air pollution over the North Atlantic

Since the U.S. Clean Air Act was passed in 1970, fine particulate air pollution downwind over the North Atlantic from fossil fuel combustion (mostly sulfate aerosols) has decreased by nearly a factor of two (Figure 3). As a result, more sunlight has been reaching the surface in recent decades, contributing to the observed SST warming trend (Figure 2). Warmer SSTs are beneficial for tropical cyclone formation.

Figure 3. The concentration of small air pollution particles (sulfate aerosols) over the North Atlantic has dropped by nearly a factor of two since the early 1970s, primarily due to more stringent air pollution regulations in the U.S. (Image credit: Murakami et al., 2020, “Detected climatic change in global distribution of tropical cyclones”, PNAS May 19, 2020, 117:20, 10706-10714)

In a 2019 review paper by 11 hurricane scientists, “Tropical Cyclones and Climate Change Assessment: Part I. Detection and Attribution,” at least 10 papers linking a decrease in sulfate aerosol pollution to increased Atlantic hurricane activity were cited. Four of those 11 authors of the review paper gave low-to-medium or medium-to-high confidence to the theory that decreased fine particle pollution has caused a “highly unusual” increase in Atlantic tropical cyclone frequency since the 1970s. The other seven authors gave low confidence to this hypothesis.

In a more recent 2020 study on the subject, “Detected climatic change in global distribution of tropical cyclones,” Hiroyuki Murakami and coauthors used a specialized global model with a 50-km grid to study causes of the long-term changes in the number of Atlantic named storms. Their model showed a reduction in sulfate aerosol pollution to be a primary reason for the observed increase in Atlantic named storms since 1980.

Figure 4. The 1991 eruption of Mt. Pinatubo in the Philippines put enough sunlight-reflecting sulfur particles into the stratosphere to cool Earth’s climate by 0.5 degrees Celsius (0.9°F) for over a year – the largest cooling impact of any eruption of the past 138 years, since Krakatau erupted in 1883. (Image credit: Sgt. Val Gempis)

2. No large volcanic eruptions since 1991

Two major eruptions in the tropics – at Mexico’s El Chichón in 1982 and at the Philippines’ Mt. Pinatubo in 1991 – hurled climate-cooling sulfur particles into the stratosphere, which suppressed global temperatures during the 1980s and 1990s.  The biggest temperature drop after a major volcanic eruption occurs during the first several years, but it takes about 8-10 years for global surface temperatures to recover.

Authors of several modeling studies, including the 2020 paper cited above, “Detected climatic change in global distribution of tropical cyclones,” have found that levels of sunlight-blocking volcanic dust in recent decades lower than in the 1980s and 1990s have led to more Atlantic named storms since the mid-1990s.

Figure 5. The number of named storms lasting two days or less (“shorties”) in the North Atlantic, 1878-2020. The number of shorties has increased significantly since the 1940s. (Background image: Sunset thunderstorm in Hurricane Sally as seen from an Air Force hurricane hunter aircraft on September 13, 2020.)

3. Improved technology leads to identification of more named storms

Figuring out how the number of tropical cyclones has changed over the years is complicated by deficiencies in the storm database, which extends back to 1851. It is well-known that prior to the arrival of aircraft hurricane reconnaissance in 1944 and geostationary satellites in December 1966, tropical storms in the Atlantic were undercounted. Depending upon how many missed storms are assumed, one can come up with a century-scale record of Atlantic named storms showing no significant trend, or a significantly increasing trend.

Even the database for the period with high-quality satellite data starting in 1972 suffers from inconsistencies. For example, the advent of new satellite instruments, such as QuikScat and the AMSU in the early 2000s, combined with new analysis techniques such as Cyclone Phase Space analyses introduced at that time, meant that short-lived storms which previously would have been classified as extratropical were instead identified as named tropical or subtropical storms.

Landsea and colleagues, in a 2010 paper titled “Impact of Duration Thresholds on Atlantic Tropical Cyclone Counts,” identified four Atlantic named storms in 2007 and two in 2008 that likely were named  because of newly available technology and analysis techniques introduced since 2003. The authors also showed that the increasing trend in North Atlantic tropical storm frequency since the 1940s was largely the result of  an increase in short‐lived storms called “shorties” and  lasting two days or less (Figure 5). The authors theorized that a significant portion of the increase in short-lived Atlantic named storms in recent decades occurred as a result of changes in the observing system, which has allowed increased detection of shorties.

Figure 6. Tracks of all 72 Atlantic named storms lasting two days or less at tropical storm strength from 2001-2020. The majority of these “shorties” tracked through the Gulf of Mexico or subtropical Atlantic. (Image credit: NOAA)

The vast majority of the 72 shorties that occurred between 2001 and 2020 (Figure 6) took place outside the tropical Atlantic, where SSTs for tropical cyclone formation are now much more common than before. As seen in the tweet below from Richard Dixon, many areas of the Atlantic have more than two additional weeks per year with SSTs warm enough to support tropical storm development compared to 20 years ago:

However, Villarini and colleagues in 2011, in “Is the recorded increase in short-duration North Atlantic tropical storms spurious?,” attempted to correlate an increase in shorties to the increase in tropical Atlantic SSTs, but they were unable to do so. The authors wrote that:

Using statistical methods combined with the current understanding of the physical processes, we are unable to find support for the hypothesis that the century‐scale record of short‐lived tropical cyclones in the Atlantic contains a detectable real climate signal. Therefore, we interpret the long‐term increase in short‐duration North Atlantic tropical storms as likely to be substantially inflated by observing system changes over time. These results strongly suggest that studies examining the frequency of North Atlantic tropical storms over the historical era (between the 19th century and present) should focus on storms of duration greater than about two days.

Given their recommendation, one can next consider the database of tropical storms lasting more than two days (Figure 7). There is a significant increasing trend since reliable satellite data became available in 1972. However, little trend in these longer-lasting storms is apparent when looking back to 1878, after adjusting for a presumed undercount in the number of these storms of one-to-three per year prior to 1966, as done in Landsea and colleagues(2010). In a 2021 paper, Schreck and colleagues recommended in “Optimal Climate Normals for North Atlantic Hurricane Activity” using the 50-year 1971-2020 annual average of 9.2 long-lived storms to represent the annual number of long-lived Atlantic named storms, and adding 4.6 to represent the 2001-2020 average number of short-lived storms. That approach would result in a climatology of 13.8 named storms, 6.4 hurricanes, 2.6 major hurricanes, and an Accumulated Cyclone Energy (ACE) index of 102.9.

Figure 7. The number of named tropical and subtropical storms lasting more than two days at tropical storm strength in the North Atlantic, 1878-2020. Over the entire data record, there has been little trend, though a significant increase has occurred since reliable satellite data became available in 1972. *Data prior to 1966 has been adjusted for a presumed undercount of 1-3 storms per year as outlined by Landsea et al. (2010), Impact of Duration Thresholds on Atlantic Tropical Cyclone Counts. (Background image: center of Hurricane Sally as seen from hurricane hunter aircraft N43RF on September 14, 2020, taken by James Carpenter, NOAA.)

4. Human-caused global warming

Burning fossil fuels creates greenhouse gases that warm the planet and its oceans, making tropical cyclone formation more likely. However, higher levels of greenhouse gases also act to dry out middle levels of the atmosphere, discouraging tropical cyclone formation.

It could well be that human emissions of greenhouse gases have been a significant contributor to the observed increase in Atlantic named storms in recent decades, as a result of the resulting increase in SSTs. However, it is unknown to what degree the competing effect of mid-level drying of the atmosphere may cancel out or overpower the increased SST effect.

According to a 2020 review paper by 11 hurricane scientists, “Tropical Cyclones and Climate Change Assessment: Part II: Projected Response to Anthropogenic Warming“, “the vast majority of individual studies (22 out of 27 studies) project a decrease in global tropical cyclone frequency with greenhouse warming.” However, the reasons for this prediction were unclear, with the authors writing: “The physical mechanism responsible for reduced global tropical cyclone frequency in the large majority of models remains uncertain.” Seven of the 11 authors rated confidence in the projection of decreasing tropical cyclone frequency as low-to-medium.

Figure 8. Top: model predictions of the historical (upper left) and future percent change (upper right) of genesis density of tropical cyclones (the number of genesis events per 1° latitude square per year). Bottom: model predictions of the historical (lower left) and future percent change (lower right) of track density. The future predictions end in the year 2120, and are for CO2 increasing by 1% per year. Over the past 30 years, CO2 has been increasing by only about 0.5% per year, so these model runs are merely illustrative of the type of changes that will occur, and do not show the actual magnitude of the changes expected. (Image credit: Emanuel, 2020, “Response of Global Tropical Cyclone Activity to Increasing CO2: Results from Downscaling CMIP6 Models“)

One of the 11 authors, Dr. Kerry Emanuel of MIT, argued in a 2021 paper, “Response of Global Tropical Cyclone Activity to Increasing CO2: Results from Downscaling CMIP6 Models,” that because of the relatively coarse resolution of the models used in those 22 studies (and other factors), “there is little basis for confidence in the projection by most climate models that overall tropical cyclone frequency will decline.” His modeling results, using a “downscaling” technique that simulates hurricanes at a much higher resolution than used in the 22 studies, projected an increase in Atlantic named storms as a result of  increased greenhouse gas emissions. The additional storms primarily occurred in more northerly portions of the subtropical Atlantic, where SSTs are often too cool for tropical cyclone formation during hurricane season (Figure 8).

Emanuel found that higher levels of greenhouse gases increased named storm numbers primarily by making the atmosphere more unstable (by warming the ocean, among other factors). Less important influences were a decrease in wind shear, and an increase in the amount of spin storms had (because of more storms forming closer to the pole, where they can leverage the Earth’s spin to gain more spin themselves). An increase in dryness at mid-levels of the atmosphere counteracted these effects, but not enough to drive a decrease in named storms, as so many other models found.

In a more recent study on the subject, Bernett and colleagues in 2021 reported in “Tropical Cyclone Frequency Under Varying SSTs in Aquaplanet Simulations” that global tropical cyclone frequency in a warming climate is controlled by shifts in the global atmospheric circulation. In particular, as the globe-encircling band of heavy thunderstorms called the Intertropical Convergence Zone (ITCZ) shifts poleward, the number of named storms that form increases. Thus, human-caused global warming could be responsible for the observed increase in Atlantic named storms by shifting the ITCZ northwards.

5. Natural variability resulting from decades-long cycles

It is widely recognized that the natural El Niño/Southern Oscillation (ENSO) cycle modulates Atlantic hurricane activity though increases and decreases in wind shear; high wind shear acts to tear storms apart, leading to quiet hurricane seasons. Wind shear increases during an El Niño phase and decreases during neutral and La Niña phases. However, the ENSO cycle is on time scales of 2-7 years, and cannot account for multi-decadal changes in Atlantic hurricane activity. The impact on Atlantic hurricane activity of other natural cycles, such as the Pacific Decadal Oscillation (PDO), has not been established.

In the 1980s and 1990s, observations and modeling results “discovered” a natural 40- to 60-year cycle in the climate system, centered in the North Atlantic. This cycle, dubbed the Atlantic Multidecadal Oscillation (AMO), appeared to exhibit a significant influence on Atlantic hurricane activity through warming and cooling of Atlantic waters via the interactions between ocean currents and wind patterns. The late Dr. Bill Gray of Colorado State University, who pioneered the science of seasonal hurricane forecasting, frequently stated that the AMO was a major reason for the uptick in Atlantic hurricane activity since 1995.

However, a March 2021 paper, “Multidecadal climate oscillations during the past millennium driven by volcanic forcing,” threw cold water on this idea. In a commentary at realclimate.org, co-author Dr. Michael Mann of Penn State University – the scientist who originally coined the AMO phrase – wrote, “The AMO doesn’t actually exist. It’s an artifact, during the historical era, of competing anthropogenic (greenhouse warming and sulfate aerosol cooling) drivers and, during the earlier period, an artifact of the fact that volcanic forcing happens to have displayed a roughly multidecadal pacing in past centuries.”

As a result of these new findings, it’s uncertain whether natural decades-long natural cycles might play a significant role in the recent uptick in Atlantic named storms.

Figure 9. The legion of “Doom”: tracks of all 51 named storms from 2001-2020 to cause at least a billion dollars in damage or get their name retired. Only two systems that were at tropical storm strength two days or less – “shorties” – made the list: Tropical Storm Imelda of 2019, which did $5.1 billion in damage in Texas, and Tropical Storm Allison of 2001, which did $12.7 billion in damage, primarily in Texas and Louisiana. (Image credit: NOAA)

Summary: What matters most is higher frequency of severe storms

The number of Atlantic named storms has been on the rise in recent decades, and there is evidence that we are now in an era in which more frequent Atlantic named storms will be the “new normal”. The increase in numbers has mostly been for weak, short-lived storms, strongly suggesting that changes in the observing system may contribute. However, longer-lasting storms have also been increasing in number: Two possible causes are increases in sea surface temperatures as a result of reduced fine particle air pollution, and increases in SSTs since the mid-1990s given fewer climate-cooling volcanic eruptions in recent decades.

The role of human-emitted greenhouse gas emissions is controversial, since there are competing influences. Most models show fewer named storms due to increased greenhouse gases, but some higher-resolution models show significantly more named storms. Natural decades-long cycles currently do not have strong support for exerting a substantial influence on Atlantic named storm numbers.

Keep in mind the importance of primarily concerning ourselves with the frequency of the most damaging events, not the overall frequency of storms. Over the past 20 years (2001-2020), shorties comprised 27% of all Atlantic named storms. Fifty-one named storms during that span have had their names retired, or caused at least $1 billion in damage. Only two of those storms – Tropical Storm Imelda of 2019 and Tropical Storm Allison of 2001 – were shorties (Figure 9). Hurricane damage is overwhelmingly driven by long-track major hurricanes. Thus, the most concerning statistic in the new 30-year “normals” for hurricane activity is the 14% increase in major hurricanes – from 2.8 to 3.2 per year – an increase not reflected in the official rounded numbers.

Hurricane scientists are in strong agreement that human-caused global warming is increasing the intensity of these most dangerous of storms. In a 2019 review paper by 11 hurricane scientists, “Tropical Cyclones and Climate Change Assessment: Part I. Detection and Attribution,” 10 of 11 authors concluded that the balance of evidence suggests that there is a detectable increase in the worldwide average intensity of global hurricanes since the early 1980s. Eight of the 11 authors concluded that the balance of evidence suggests that human-caused climate change contributed to that increase.

Also see: A new yardstick: What is ‘normal’ in a changing climate?

Website visitors can comment on “Eye on the Storm” posts. Please read our Comments Policy prior to posting. Comments are generally open for 30 days from date posted. Twitter: @DrJeffMasters and @bhensonweather

Jeff Masters

Jeff Masters, Ph.D., worked as a hurricane scientist with the NOAA Hurricane Hunters from 1986-1990. After a near-fatal flight into category 5 Hurricane Hugo, he left the Hurricane Hunters to pursue a...