The event

Hurricane Milton began as a tropical depression in the Gulf of Mexico on Saturday 5th October. It very rapidly intensified to tropical storm status, before undergoing explosive intensification to a high category 5 between Sunday 6th and Monday 7th, as it moved southeast towards the Yucatan Peninsula (NOAA, 2024), where the central pressure fell to below 900 mbar (NHC, 2024). This was driven and sustained by the very high sea surface temperatures in the Gulf, which previous analysis has shown to have been made 400-800 times more likely by climate change (Climate Central, 2024). Milton then turned northeast, closing in on central Florida, in a path very close to both Tampa and Orlando. Its intensity decreased slightly to Category 3 by the time of landfall on the evening of the 9th, bringing high winds, extreme rainfall and storm surges of 5-10  feet (Miami NBC, 2024) to the western coast of Florida, before moving directly across the low-lying peninsula. 

While still working on the impacts of hurricane Helene that hit only two weeks earlier (WWA, 2024), the authorities put over 1,000 first responders and more than 1,400 search and rescue personnel on the ground to support people in the path of Milton (FEMA, 2024). In addition, government authorities urged people to evacuate the area and provided free shuttles for people living in the predicted path of Milton, and issued a range of statements against misinformation (FEMA, 2024b). Over 7 million people live in the area with mandatory evacuation orders, though some did not or were unable to comply (Axios, 2024). The hurricane spawned dozens of tornadoes that resulted in many of the deaths (AP News, 2024). The full impacts of the hurricane will only be known in the days and weeks to come. At the time of writing, 16 people are confirmed to have lost their lives while more than three million are without electricity (Tampa Bay Times, 2024; New York Times, 2024; Independent, 2024).

Key Messages

• Milton formed in the Gulf of Mexico and intensified in the course of only two days into a Category 5 hurricane. It followed in the wake of Hurricane Helene that made landfall less than two weeks before Milton. Both the rapid intensification and the fact that emergency personnel were still continuing with the recovery from Helene made preparations difficult. Despite this, many people in the immediate path evacuated and the losses and damages from Milton are much smaller than from Helene. However, the full extent of the compounding impacts from the hurricanes will only be known in the weeks to come.
• At the time of writing, the day after landfall, the observation-based datasets are not all updated to include the event. We can thus not reliably estimate how rare the heavy rainfall in the path of Milton was, a usual step in attribution analysis. Instead, we assessed trends in observations for an estimated 1 in 10 year and a 1 in 100 year event in the region indicated in figure 1. In both cases the results are comparable and thus not very sensitive to the exact event definition.
• To estimate if human-induced climate change influenced the heavy rainfall, we determine if there is a trend in the observations. In three out of the four analysed datasets we find that heavy 1-day rainfall events such as the one associated with Milton are 20-30% more intense and about twice as likely in today’s climate, that is 1.3°C warmer than it would have been without human-induced climate change. The fourth dataset shows much larger changes. 
• These results are based on observational data and do not include climate models and are thus higher than the overarching attribution statement given for Hurricane Helene, where we combined observations and climate models. Nevertheless the results are compatible with those obtained for other hurricanes in the area that have been studied in the scientific literature. Despite using different temporal and geographical event definitions, as well as different observational datasets and climate models, all these studies show a similar increase in intensity of between 10 and 50% and about a doubling in likelihood. We are therefore confident that such changes in heavy rainfall are attributable to human-caused climate change. 
• The IRIS model was used to investigate Milton’s strong winds by analysing storms making landfall within 2 degrees of Milton. By statistically modelling storms in a 1.3°C cooler climate, this model showed that climate change was responsible for an increase of about 40% in the number of storms of this intensity, and equivalently that the maximum wind speeds of similar storms are now about 5 m/s (around 10%) stronger than in a world without climate change. In other words, without climate change Milton would have made landfall as a Category 2 instead of a Category 3 storm. 
• Hurricane-prone Florida has measures in place to reduce impacts, though most adaptation measures are reactive, rather than preventative. The Resilient Florida Grant Program, established in 2021, strengthens local government’s capacity to address the impacts of climate change, through vulnerability assessments, adaptation plans, strengthening coastal defences, and community outreach. Since Hurricane Michael in 2018, the Florida Building Code has been updated twice, and is amongst the toughest in Gulf and Atlantic Coast states. The state has opportunities to further integrate nature-based solutions, such as wetland restoration, and to enhance efforts to address social vulnerabilities to better protect low-income and minority communities at greater risk of flood impacts. 

Analysis of trends in extremes

In this short report we examine trends in rainfall extremes like that from Hurricane Milton, and attribute changes in wind speeds from similar events using the IRIS storm model. We compare this to results from the recent rapid attribution study for Helene and other work attributing similar storms in the literature.

Observed rainfall

Changes in the rainfall from Milton are studied by analysing changes in the wettest day each year during June-November, averaged over land areas within an area covering central Florida bounded by the coordinates 26-30 °N and 79.5-83.5 °W (shown in Figure 1). Four observational and reanalysis datasets are used: MSWEP, ERA5, CHIRPS and CPC (see appendix for details). The methods used to analyse rainfall trends follow the standard WWA protocol using non-stationary extreme value theory, as described in Philip et al, 2020 and all full study reports (e.g. WWA, 2024).

At the time of writing, the day after landfall, the observation-based datasets have not all been updated to include the event. We can thus not reliably estimate how rare the heavy rainfall in the path of Milton was. Instead, we estimated changes in likelihood and magnitude for a 1 in 10 year and a 1 in 100 year event. The results of this analysis are shown in table 1. 

Dataset	1 in 10 year event		1 in 100 year event		Magnitude change (%)
	Magnitude (mm)	Probability ratio	Magnitude (mm)	Probability ratio
ERA5	66.97	2.29  (0.68 – 8.10)	143.97	2.50  (0.65 – 13.09)	33.15  (-11.33 – 101.88)
CHIRPS	74.38	32.69 (0.33 – inf)	104.91	287.70 (0.35 – inf)	60.46 (-9.80 – 187.0)
MSWEP	72.81	1.95 (0.30 – 8380)	144.13	2.18 (0.12 – inf)	23.46 (-34.57 – 171.71)
CPC	78.06	1.92 (0.15 – 4.58)	162.69	2.08 (0.001 – 6.24)	24.51 (-21.65 – 84.97)
Synthesis		4.09 (0.09 – 3240)		7.56 (0.009 – 80000)	34.6 (-24.2 – 145)

Table 1: Magnitudes of 1 in 10 and 1 in 100 year rainfall events in 2024 in the region over central Florida shown in figure 1. The probability ratio and change in intensity associated with a 1.3C increase in global mean surface temperature (GMST) are estimated for each event, with bootstrapped uncertainties Due to the scaling assumption built into the statistical model, the change in magnitude is the same for events of all return periods. Finally, the synthesised results for all datasets are shown in the final row.

In summary, over the past ~75 years, rainfall extremes in central Florida have increased significantly with global warming. While a full attribution would require an assessment of trends in climate models as well as observations, the results in Table 1 are in line with those of Hurricane Helene and other literature on hurricanes in the region. The results for individual datasets shown in Table 1 show changes in response to GMST that are very similar to the observational results for Hurricane Helene (see Table A.2). We would therefore expect the overall attribution result to be comparable, with climate models showing weaker (but still positive) trends in response to climate change and the synthesised results indicating with high confidence that the rainfall was increased in both likelihood and intensity to a similar level as the coastal region for Hurricane Helene (WWA, 2024). 

Wind intensity

The IRIS synthetic storm model (Sparks and Toumi, 2024), developed at the Grantham Institute and previously set out here, is used to study the change in likelihood and intensity of the high wind speeds associated with storms similar to Milton. This is defined by changes in those storms making landfall at Milton’s intensity, within Category 3, and geographically within 2 degrees of Milton’s landfall on the west coast of Florida (figure 2).

This analysis finds that storms with Milton’s wind speeds have become approximately 40% more frequent; equivalently, winds associated with storms of similar rarity have become nearly 5 m/s more intense, due to 1.3 °C of global warming (figure 2). In practice, this means that without climate change Hurricane Milton would have been a Category 2 rather than a Category 3 hurricane when it made landfall. 

Sea Surface Temperatures 

Climate Central’s Climate Shift Index: Ocean (Ocean CSI) tool was recently used to rapidly compute the influence of human-caused climate change on Sea Surface Temperatures along Hurricane Milton track. Here we summarise their methods and main findings.

The methodology underpinning the CSI Oceans tool is based on peer-reviewed research (Giguere et. al, 2024). It uses a combination of an empirically-driven attribution method usually using the observed SSTs in OISST data (Huang et. al, 2021), but due to the impacts of Helene on the daily updates from NOAA’s National Centers for Environmental Information (NCEI) based in Asheville, N.C., observations of this event remain unavailable at the time of writing. Thus the European product OSTIA is used instead as well as model simulations using an ensemble of 13 debiased CMIP6 models. A detailed description can be found in the assessment of Hurricane Helene (WWA, 2024).