Researchers have used deep learning to model more accurately than ever how ice crystals form in the atmosphere. Their paper, published in PNAS this week, points to the potential to significantly increase the accuracy of weather and climate forecasts.
Researchers used deep learning to predict how atoms and molecules behave. First, models were trained using small-scale simulations of 64 water molecules to predict how electrons interact in atoms. The models then replicated these interactions on a larger scale with more atoms and molecules. It is this ability to accurately simulate electron interactions that allowed the team to accurately predict the physical and chemical behavior.
“The properties of matter result from the behavior of electrons,” says Pablo Piaggi, a research associate at Princeton University and lead author of the study. “Explicitly simulating what is happening at this level is a way to capture much richer physical phenomena.”
It is the first time this method has been used to model something as complex as the formation of ice crystals, also known as ice nucleation. This is one of the first steps in the formation of clouds, from which all precipitation originates.
Xiaohong Liu, a professor of atmospheric sciences at Texas A&M University who wasn’t involved in the study, says half of all precipitation events — whether it’s snow, rain or sleet — begin as ice crystals that then grow larger and result in precipitation. If researchers could model ice nucleation more accurately, it could give a big boost to weather forecasting overall.
Ice nucleation is currently predicted based on laboratory experiments. The researchers collect data on ice formation under different laboratory conditions, and this data is fed into weather forecast models under similar real-world conditions. This method sometimes works well enough, but is often inaccurate due to the sheer number of variables involved in actual weather conditions. If even a few factors vary between the lab and the real world, the results can be vastly different.
“Your data is only valid for a specific region, temperature or laboratory environment,” says Liu.
Predicting ice nucleation based on the way electrons interact is much more precise, but also very computationally intensive. It requires researchers to model at least 4,000 to 100,000 water molecules, and even on supercomputers such a simulation can take years. Even that would only be able to model the interactions for 100 picoseconds or 10-10 Seconds – not long enough to observe the ice formation process.
However, using deep learning, the researchers were able to complete the calculations in just 10 days. The length of time was also 1,000 times longer – still a fraction of a second, but just enough to see nucleation.
Of course, more accurate models of ice nucleation alone won’t make the forecast perfect, Liu says, since it’s a small but critical component of weather modeling. Other aspects are also important – for example, understanding how water droplets and ice crystals grow and how they move and interact with each other under different conditions.
Still, the ability to more accurately model how ice crystals form in the atmosphere would greatly improve weather forecasting, particularly those related to whether and how much it is likely to rain or snow. It could also aid in climate prediction by improving the ability to model clouds, which affect the planet’s temperature in complex ways.
Piaggi says future research could model ice formation when substances like smoke are in the air, potentially improving the accuracy of the models even further. Due to deep learning techniques, it is now possible to use electron interactions to model larger systems over longer time periods.
“It essentially opened up a new field,” says Piaggi. “It already plays an even greater role in simulations in chemistry and in our material simulations and will continue to do so.”