Two hurricanes battered the Southeast this season in as little as two weeks. Each is historic—Hurricane Helene for its size, which at 400 miles across allowed it to wreak havoc from Florida to Appalachia, and Milton for its strength. With wind speeds up to 180 mph, Hurricane Milton was the fifth most intense tropical cyclone ever in the Atlantic and the strongest worldwide of this hurricane season.
Climate scientists who examine how tropical cyclones will change under a warming climate are looking to improve how well climate models can simulate these rare but destructive events. Emily Bercos Hickey is a researcher with the Climate and Ecosystem Sciences Division who examines extreme weather events, such as tropical cyclones, and the complex meteorology over North Africa, specifically the African easterly waves that are strongly linked with Atlantic tropical cyclones.
We spoke with Emily to learn more about the importance of climate modeling for hurricane prediction, and how some of the latest modeling techniques may someday soon make it possible to predict the ferocity of “outliers” like Hurricanes Helene and Milton.
Why is it important for climate scientists to model tropical cyclones?
There are a few really great ways that we can leverage climate modeling to improve our understanding of tropical cyclones. Traditional climate models can be used to simulate historically impactful tropical cyclones, which we call “hindcast” simulations. We can then recreate, or simulate, these previous tropical cyclone events, but with certain environmental conditions of the simulations modified to incorporate the effects of climate change. For example, we could simulate Hurricane Helene as if this hurricane had occurred in a pre-industrial world or at the end of the 21st century but with warmer sea surface temperatures due to a warming climate by modifying sea surface temperature. This approach allows us to better understand how the intensity of historic extreme weather events, like Milton and Helene, has changed due to our warmer climate, and will respond to future warming.
More recently, climate scientists have been using another approach to help address one of the biggest challenges we have when hindcasting previous events: Small sample size. Milton was one of an estimated 42 Category 5 Atlantic storms since 1924. And because there are very few storms similar to Milton, there is a limited number of events in the historical record for comparison. Many researchers, including the CASCADE team studying extreme weather events at Berkeley Lab, are now utilizing deep-learning based weather models that can be used to generate huge samples, something that was previously computationally prohibitive with traditional climate models. This approach is very exciting because large sample sizes allow us to perform statistical analyses that would have previously been challenging given the rarity of extreme events like Helene and Milton.
Why are hurricanes becoming more intense, and why does it seem like this year was very unpredictable?
Multiple factors are required for a tropical cyclone to form, including warm sea surface temperatures of at least 27 degrees Celsius (80 degrees Fahrenheit) and a moist atmosphere. As Milton was forming, sea surface temperatures in the Gulf of Mexico reached roughly 31 degrees Celsius (88 degrees Fahrenheit), which is more than 4 degrees Fahrenheit above the long-term average. In the current warmer climate, and with future warming, sea surface temperatures and atmospheric moisture increase, which fuels tropical cyclones and leads to an increase in storm intensity.
The 2024 Hurricane Season, which officially began on June 1 and lasts until November 30, was expected to be highly active due to above-average sea surface temperatures and hurricane-favorable wind conditions. However, the North Atlantic went unusually quiet without a new named storm at any point between mid-August and early September. There were multiple reasons for this lull, including a northward shift of the African easterly waves that often develop into tropical cyclones and an abundance of dry, dusty air from the Sahara Desert. All the ingredients can be in place for an active hurricane season, such as warm sea surface temperatures, but intraseasonal variability, like dust storms, can strongly affect what happens and can be very challenging to predict.
After these tumultuous storms, what comes next for extreme event modeling?
Something that seemed to take many people by surprise was the devastating impact that Hurricane Helene had on communities in the Great Smoky, Blue Ridge, and Appalachian Mountain regions. One of the tools that will help us better understand how a tropical cyclone might behave in regions with similar topography is high-resolution modeling. To resolve something in a climate model means to represent a specific feature accurately within a model’s grid cells. Many of the models that have been used to examine tropical cyclones are run at horizontal resolutions of 25-100km. That’s not great for simulating the extreme wind and rain associated with these events, and for resolving mountainous regions–such as those impacted by Hurricane Helene.
The newly DOE-developed Simple Cloud-Resolving E3SM Atmosphere Model (SCREAM), which is capable of using Graphics Processing Unit (GPU) architectures to run efficiently at a global resolution of 3km, allows for finer resolution modeling in combination with faster computing. With finer resolution, and faster computing, climate models can help us make more useful predictions, such as those at a geographic scale that will enable us to warn communities at risk.
Another aspect of these storms that may seem extraordinary is the severe weather that can often accompany them. On the day that Milton arrived in Florida, the National Weather Service issued 126 tornado warnings, and South Florida experienced its first EF3, on the Enhanced Fujita Scale, tornado on record. While it is not unusual for tornadoes to accompany tropical cyclones due to large changes in wind speed and direction and instability in the atmosphere, there are typically not this many associated with one tropical cyclone nor are they nearly this strong. Tornado-producing tropical cyclones are particularly challenging to study, as tornadoes are such small-scale events that they are not resolved even in fine resolution climate models like SCREAM. This demonstrates the need for continued advances in climate modeling.
The amount of rainfall with Helene, and the strength of the wind and rapid intensification of Milton, gives climate scientists a lot to consider when taking these extreme events into consideration while trying to predict the intensity of future hurricanes. Following the tragedies caused by these storms, we can continue to work towards improving climate models to be faster and more computationally efficient, with the ability to resolve characteristics at ever finer scales. This will help us to better understand how tropical cyclones and their drivers will continue to respond to a warming climate.
