Climate Change and Hurricanes

Hurricanes Helene and Milton recently brought attention to the question of how climate change (or other human activities) might be affecting hurricanes/tropical cyclones (TCs). This topic can be controversial, because we have a relatively limited record (as for most extreme events) and because TCs are difficult for climate models to simulate, being right on the edge of what state-of-the-art models can resolve. However, the lack of conclusive modeling and observational evidence can make the impacts of climate change on TCs seem more uncertain than they really are. We know enough about TCs to have confidence in some aspects of their response to warming, and to know why we are unsure of other aspects. To illustrate these points, I want to present here an analogy I like to use for thinking about climate change and TCs.

Let’s start with the most recent IPCC report, which gives a summary of the state of the science. The end of the TC section (11.7.1) states that “average peak TC wind speeds and the proportion of Category 4–5 TCs will very likely increase globally with warming” and “It is very likely that average TC rain rates will increase with warming”, but that it is only “likely” that “the global frequency of TCs over all categories will decrease or remain unchanged.“

So we’re confident that TCs will get stronger with warming, but we’re less sure how their frequency will change. The analogy I have for this is a field of crops. As MTG correctly stated, TCs start from “seeds” – clusters of small convective storms – which organize to form larger systems that eventually develop into TCs. During this process, the seeds and developing TCs can be disrupted by interactions with the large-scale meteorological environment (wind shear, dry air intrusions, etc.), so not all seeds grow into TCs. In the same way, a farmer sows seeds which will hopefully grow into mature crops, but this process can be disrupted by weather events (droughts, storms), pests and other factors outside the farmer’s control.

                                                           (Illustration made with Gemini)

Continuing the analogy, global warming makes the tropical oceans “more fertile” for TC growth. TCs extract energy from the temperature difference between the warm ocean surface and the cool upper troposphere., so they can extract more energy and become more powerful as surface waters warm [the upper troposphere is also warming, but it’s better to think of this as stretching the depth of the atmosphere, so TCs deepen and extend up to roughly the same temperature, rather than as affecting the relevant temperature difference.] Similarly, if the soil in a field becomes more fertile, plants can grow larger, giving bigger yields.

Our understanding of what drives TCs gives us confidence in this response to warming, even if climate models improperly simulate TCs or if we don’t see clear trends in the limited historical data. What we’re less sure about is how the number of TC seeds, and their chance of developing into TCs, will change under warming. This is as if we were unsure how many seeds the farmer would sow each year or what fraction of those seeds would grow into mature plants.

TC seeds and their transition are difficult to constrain because they depend on relatively small-scale meteorological factors, which are influenced in turn by things like the ENSO cycle and Saharan dust emissions. This complex set of interactions, spanning meteorological “noise” to major modes of climate variability, determines how favorable a given environment is for TC growth, obscuring trends and leading to spread across climate models in their projections. Nevertheless, there are some things we can say here. For example, we might expect a longer TC “growing season” as the ocean waters stay warm for longer 

There’s also been some intriguing recent work by Tsung-lin Hsieh and co-authors on TC seeds (see here). Zooming out, TC seeds are favored in regions of large-scale ascent (where the tropical atmosphere is rising on average) and where vortex formation is favored over wave-like behavior; i.e., vortex stretching is favored over vorticity advection. Pushing this further, Tsung-Lin recently found that changes in TC seeds are anti-correlated with cloud feedbacks in climate models. In other words, models which warm more have larger decreases in TC seeding. The common link is reductions in tropical convection, in the area of tropical ascent, which lead to less cloud cover and fewer seeds developing, though this still needs some filling in. 

The probability of seeds transitioning to TCs can be similarly quantified with a “ventilation index”, which measures wind shear and the dryness of the mid-troposphere (see Tang and Emanuel, 2010). Another reason TC frequency is expected to decline with warming is an increase in wind shear, decreasing the probability of a seed transitioning to a TC.

Two other aspects of tropical cyclones the IPCC report discusses are where they form and how fast they move. With regards to formation regions, we are confident that TCs will form at higher latitudes over time because the atmosphere’s circulation generally expands poleward under warming. So the regions that are favorable for TC growth will also expand polewards. 

Changes in how fast TCs move – in their “translation velocity” – are more unclear. Helene had a devastating impact on North Carolina because it moved very quickly (it had a fast translation speed) and managed to get quite far over land before decaying. Helene’s speed also meant that it didn’t do much mixing of the waters in the Gulf of Mexico, leaving them warm and ready for another TC (Milton) to form.

There isn’t much consensus on how TC translation speed will change under warming, but most studies seem to suggest TCs will actually slow down under global warming (example). This would be bad news for coastal regions, as slower moving TCs will dump more rain on landfall. But as with TC seeds, TC translation speeds are the result of complex interactions with environmental conditions, so it’s harder for us to be confident in how they will respond to climate change. Both problems seem ripe for AI/ML to solve.

Previous
Previous

California Wildfires and Climate Change

Next
Next

Paper Summary: Impact of Atmospheric Cloud Radiative Effects on Annular Mode Persistence in Idealized Simulations