Hurricane Season 2024 retrospective

by Charles Powell (University of Cambridge) and Ruth Petrie (Inigo)

Key points

• The 2024 hurricane season was hyperactive, as forecast based on record warm SSTs and development of a weak La Niña event.
• Tropical cyclone activity was unusually quiet during the climatological peak because of unfavourable conditions for development of tropical cyclone seeds in the Atlantic.
• The final part of the season was extremely active with two high impact major hurricane landfalls in Florida.
• In a changing climate, warmer SSTs may provide the fuel for more frequent intensification of tropical cyclones despite increased atmospheric stability.

Forecasts for the season

Coming into the 2024 hurricane season, almost all forecasts pointed towards an above average season. The Barcelona Supercomputing Centre collates predictions from universities, government agencies and private companies. Across all entities there was an average of 11 hurricanes forecast for the 2024 season, well above the long-term average of 7 hurricanes per season. These forecasts for increased hurricane activity were first made in March-April and remained unchanged throughout the season, despite an unusual lull in activity during the early-to-middle part of the hurricane season that included the climatological peak month of activity in August. So, why did we expect such an active season?

Figure 1: daily average sea surface temperatures in the Atlantic main development region (inset). Record warm SSTs in 2023 and 2024 highlighted.

In March 2023, global average sea surface temperature (SST) reached record warm levels. In the subsequent 15 months, global average daily SSTs set records every day. How does this affect Atlantic hurricane activity? Hurricane activity is strongly correlated with SSTs in the Atlantic main development region (MDR) where most Atlantic hurricanes form (figure 1). This metric therefore forms a key component of seasonal forecasts of hurricane activity. Daily average SSTs in the MDR also reached record levels in June 2023 and have remained exceptionally above average ever since (figure 1). At Inigo we forecast that the August-September-October (ASO) MDR SSTs would be exceptionally warm once again, with a best estimate of 28.95°C (figure 2). With SSTs at such high levels, there is an enormous amount of extra energy available to fuel tropical cyclone intensification.

Figure 2: SST forecast for Atlantic MDR in 2024.

For tropical cyclones to form and intensify into hurricanes, a favourable atmospheric environment must exist for deep convection to initiate and organise into a tropical depression which, given the right circumstances, may then evolve into a hurricane.

The atmospheric conditions in the Atlantic can be heavily influenced by ‘teleconnections’ with atmospheric and oceanic drivers around the world. Of these teleconnections, the El-Niño Southern Oscillation (ENSO) exerts the largest influence on tropical cyclone activity in the Atlantic basin. The positive or ‘warm’ phase – El Niño – tends to inhibit activity by shifting major tropical convection into the Eastern Pacific. This enhances upper-level divergent outflow, resulting in stronger westerly zonal winds in the upper troposphere over the Caribbean and tropical Atlantic. The associated vertical wind shear interferes with the organisation of deep convection and prevents tropical cyclones from intensifying. The negative or ‘cold’ phase – La Niña – has the opposite effect, tending to enhance activity by reducing vertical wind shear in the Atlantic.

Seasonal forecasts for the 2024 hurricane season suggested the El Niño event seen throughout 2023 would quickly flip to La Niña, adding to concerns for a strong 2024 season. The probability and potential strength of the forecast La Niña from April 2024 is shown in figure 3. It was forecast by the Columbia Climate School International Research Institute for Climate and Society (IRI) that in the peak of hurricane season (ASO) there was an 80% chance of La Niña conditions. However, there were indications that it would likely not be a strong event with many forecasting centres predicting only a weak La Niña event. Nonetheless, with two of the biggest predictors of Atlantic hurricane activity suggesting enhanced hurricane activity in the Atlantic, expectations were set for a very active season.

Figure 3: The IRI ENSO predictions of a transition to La Niña for 2024 in April 2024.

How did the season turn out?

The season started relatively slowly with the first named storm not forming until June 19th. This marked the latest start to a hurricane season since 2014. Activity soon picked up, with Hurricane Beryl becoming the earliest category 5 Atlantic hurricane on record at the end of June. From early July through to late September, despite a few named storms forming there was remarkably little activity. Whilst the season typically lasts from June to November, the climatological peak in activity occurs in August and September.

In 2024, after Beryl, there were no major hurricanes until the final week of September and only 4 named storms. A common measure of tropical cyclone activity is accumulated cyclone energy (ACE) which is calculated from maximum sustained winds; large ACE values indicate active periods. Figure 4 shows ACE in the North Atlantic in the 2024 season compared with climatology. After starting with a bang, ACE remained close to the climatological average through to August but then remained well below climatological levels during the peak season.

Figure 4: climatological and 2024 daily accumulated cyclone energy (ACE).

From late September, activity picked up dramatically. Four named storms formed in the space of a week, notably Hurricane Helene which formed in the Caribbean Sea and made landfall in the Big Bend region of Florida as a category 4 hurricane. Helene was an unusually fast-moving storm owing to interaction with a jet streak and a low pressure system centred over the central US. The divergence in the upper troposphere associated with the jet streak significantly enhanced moisture transport within Helene, resulting in catastrophic flooding in North Carolina.

In October, activity remained high with four named storms. All but one of these storms reached hurricane status. Notably, Hurricane Milton formed in the Gulf of Mexico and explosively intensified, undergoing the fastest 24-hour windspeed increase ever recorded in the Gulf of Mexico and becoming the most intense Atlantic hurricane (as measured by central pressure) since 2005.

Above average activity continued into the late season. In early November, Hurricane Rafael reached major hurricane status and tied the 1985 record for the strongest November hurricane in the Gulf of Mexico. In total, there were 18 named storms -less than the predicted 22. However, of these tropical storms, 11 became hurricanes and 5 went on to become major hurricanes, as was predicted. Although the 2024 season was inactive during the typical peak season, overall, 2024 has still been designated as hyperactive by NOAA, meaning total ACE exceeded 175% of the climatological median.

Why was the season like this?

This year was really a tale of two seasons. From June to August, there were 5 named storms, of which one became a major hurricane. From September to November, there were 12 named storms, of which four became major hurricanes. So why was the early-to-mid season so quiet?

Answering this question is complex and multi-faceted. There is no single explanation for the relative inactivity, but we can point to a few factors that inhibited development of tropical cyclones.

Firstly, whilst SST forecasts were accurate – even slightly underestimating peak SSTs – the expected La Niña event developed more slowly than anticipated and therefore had little influence in the early and middle parts of the season.

Moreover, the regions where tropical cyclones form changes through the season – SSTs tend to lag atmospheric temperatures by a few months, such that peak SSTs occur around August and help to fuel the climatological peak in tropical cyclone activity. This allows tropical cyclones to form in the warm Caribbean Sea and Gulf of Mexico.

However, in the early-to-mid season when SSTs are (relatively) colder, most tropical cyclone formation arises from so-called ‘seed disturbances’, or more formally ‘tropical waves’, which are formed within the intertropical convergence zone (ITCZ) – a band of convection near the equator. These tropical cyclone seeds propagate westwards and enter the Atlantic off the coast of Africa. The disturbances promote organisation of deep convection which can then intensify into a tropical storm.

Figure 5: atmospheric setup which inhibited tropical cyclone development in the early-mid season.

The latitude of the band of convection that generates tropical cyclone ‘seeds’ varies annually. This year, the ITCZ shifted further northwards than usual. As a result, seed disturbances entered the Atlantic further north than a typical year in a region where SSTs were in fact slightly colder than average. Furthermore, this coincided with a positive North Atlantic Oscillation (NAO), where high atmospheric pressure over the Azores strengthens vertical wind shear off the west coast of Africa and drags dry air from the Sahara Desert out into the Atlantic.

Together these factors resulted in a poor environment for tropical cyclone development (figure 5). Note that it is difficult to disentangle these two atmospheric events; a positive NAO favours a northward shifted subtropical high whilst a northward shifted ITCZ can also reinforce the atmospheric circulation associated with a positive NAO.

In addition to interannual and seasonal variability in the latitude of seed disturbances and the NAO phase, the tropical troposphere is gradually becoming more stable as the troposphere warms in response to climate change. In particular, the upper troposphere warms faster than the lower troposphere, reducing the atmospheric temperature gradient and therefore inhibiting tropical convection, which limits tropical storm development. Figure 6 shows the potential temperature difference between the lower and upper troposphere during August & September, the peak months of activity.

The temperature difference is decreasing over time and has been particularly low this year, which may have contributed to the lull in activity. Note that the figure shows potential temperature as it is a better indicator of atmospheric stability than temperature alone as it accounts for pressure changes as well as temperature changes.

Figure 6: potential temperature difference between 200 hPa (~12km) and 850 hPa (~1.5km) in the MDR weakening over time, increasing atmospheric stability.

In the late season, as is typical, the jet weakened, and tropical cyclogenesis moved further west as SSTs reached their peak in the Caribbean Sea and Gulf of Mexico. As in the MDR, SSTs in the Gulf of Mexico were well above average (figure 7). On sub-seasonal timescales, the , exerts significant influence on tropical cyclone (TC) activity.

As with ENSO, the MJO can strengthen or weaken vertical wind shear depending on its phase, as well as promoting uplift of air which supports tropical cyclone development. In the early season, the MJO was weak and played little role. However, from August onwards there was a higher than usual amount of MJO activity which may have contributed to the sudden uptick in activity. In summary, the favourable environment and westward shift in tropical cyclogenesis into the anomalously warm Gulf of Mexico allowed for increased TC activity later in the season.

Figure 7: daily SSTs in the Gulf of Mexico.

Role of climate in 2024 season & expectations going forward

Perhaps the most notable aspect of the 2024 season is the proportion of named storms that developed into hurricanes. Out of a total 18 named storms, 11 developed into hurricanes and 5 of these were major hurricanes. On average, there are 14 named storms each season, 7 become hurricanes and 3 go on to become major hurricanes. The 2024 season therefore saw both a higher proportion of named storms developing into hurricanes and a higher proportion of these becoming major hurricanes – should we expect this to continue in the future?

Despite modern advances in computing power, global climate models cannot explicitly resolve tropical cyclones. Nonetheless, by considering a range of models, we can develop confidence in predictions of tropical cyclone behaviour in the future climate. Assessments of a range of climate models suggest that global average TC intensity will increase, as well as the proportion of TCs that reach category 4 and above.

Owing to differences in model output, there are less confident projections that the frequency of TCs will decrease whilst the frequency of major TCs increases. In broad terms, these changes can be explained by the increase in stability in the tropical troposphere due to different warming rates with height, and an increase in energy available to fuel TCs owing to higher SSTs. In the context of these projections, the 2024 season may indicate what we see in a future climate: a greater proportion of storms reaching hurricane and even major hurricane status, amplifying the risk of large losses. The 2024 season also saw an unusually active late season.

A recent study suggests that the Atlantic hurricane season may increase by up to a month with warm northern tropical Atlantic SSTs under La Niña conditions. This can easily be understood in the context of increasingly warm SSTs: in figure 1 and 7, we see that SSTs are above the 27C temperature threshold needed for tropical cyclone development for a much longer period compared to climatology.

Conclusion

In the future, we may look back on the 2024 hurricane season as a near-miss; whilst hyperactive overall, the peak season was unusually quiet. The atmosphere and ocean conditions were primed for extreme activity throughout the season, with record warm SSTs across the Atlantic – which are likely to get warmer. However, La Niña developed later than anticipated and variability in smaller-scale processes in the atmosphere stepped in to dampen activity in the Atlantic during peak season.

The last few months of the season were perhaps a view into what hurricane seasons may look like in the future, especially when driven by La Niña conditions and with favourable conditions for TC development. The difficulties in predicting tropical cyclone activity – let alone the complexity of forecasting the potential for individual systems to develop into major hurricanes and make landfall – were laid bare in the 2024 season.

Climate science tells us that we should expect SSTs to continue to warm whilst the tropical atmosphere becomes more stable. However, the response of the ENSO and MJO teleconnections to a warming climate are not yet fully understood and therefore difficult to predict. As shown in 2024, these drivers can play a significant role in modulating TC activity on shorter timescales, adding another layer of complexity to seasonal hurricane forecasting.