Cyclones meet their mathematical match

Cyclones, hurricanes, typhoons and tornadoes are some of the most extreme and damaging natural phenomena on Earth. Driven by a complex interplay of physical processes and interactions between the air, water and land, cyclonic events have proven exceptionally difficult to model.

Now, KAUST researchers have developed a comprehensive physically based mathematical approach that allows cyclones and tornadoes to be modelled in their entirety for the first time, based on fundamental physical principles^[1].

“We’ve developed a way to simulate the formation and development of atmospheric cyclones and tornadoes in a detailed and comprehensive manner: this allows us to study and explore these weather phenomena within interactive virtual environments,” says lead researcher, Alejandro Amador Herrera.

Cyclones, which are alternatively referred to as hurricanes or typhoons depending on their ocean region of origin, form as a result of a complex interplay of heat and water dynamics in the atmosphere, including the effects of water vapor, cloud water, cloud ice, rain, snow and precipitated ice (graupel).

Previous modelling approaches either rely on computer graphics-based visual simulations using tuned equations to reproduce observed visual effects, or computationally heavy numerical grid-based calculations methods using empirical mechanistic cell-to-cell energy transfers. Neither approach attempts to model the fundamental microphysics of cyclogenesis.

“Current models often require supercomputers and only focus on specific parts of cyclones, lacking a unified approach,” says Herrera. “Our physically based simulation captures the multi-scale turbulent heat and water transport in the atmosphere, explicitly modelling vorticity and the fluid and heat dynamics that generate vortices at the microphysical level.”

Herrera, with Dominik Michels and colleagues from KAUST, in collaboration with researchers from Germany and Poland, constructed a system of interdependent equations modelling different cyclonic processes. This made it possible to simulate cyclones and tornadoes at various scales without intensive computations.

“Our approach captures the turbulent flow needed to accurately reproduce these phenomena and avoid the visual artifacts found in other models, such as hurricanes missing the ‘eye’ structure or tornadoes that appear abruptly,” says Herrera. “Our results also closely match real storm data and specialized simulation methods, demonstrating both accuracy and efficiency.”

The researchers tested their approach in a range of different simulation settings, showing remarkable accuracy compared with observations, including a highly accurate simulation of the genesis and evolution of Hurricane Katrina, which devastated the city of New Orleans in 2005.

“Now, we aim to extend our framework to include additional atmospheric effects, such as different vegetation types, seasonal changes and lightning, which will further increase realism and applicability,” Herrera says. “Beyond visual realism, however, our model can also generate high-quality synthetic data for training artificial intelligence models, which will improve predictive capabilities and applications in weather forecasting, disaster preparedness, and climate research.”