AI-powered weather and climate models are set to change the future of forecasting, researchers say

A new weather forecasting and climate prediction system uses artificial intelligence to achieve results comparable to the best existing models while using much less computing power, its creators say.

On paper Published in natureA team of researchers from Google, MIT, Harvard University and the European Centre for Medium-Range Weather Forecasts say their model offers enormous “computational savings” and could “enhance the large-scale physical simulations that are essential for understanding and forecasting the Earth system.”

The NeuralGCM model is the latest in a steady stream of research models that use advances in machine learning to make weather Climate predictions are faster and cheaper.

What is NeuralGCM?

The NeuralGCM model aims to combine the best features of traditional models with Machine learning approach.

At its core, NeuralGCM is a so-called “general circulation model.” It contains a mathematical description of the physical state of Earth’s atmosphere, and solves complex equations to predict what will happen in the future.

However, NeuralGCM also uses machine learning—the process of finding patterns and regularities in massive amounts of data—for some less understood physical processes, such as cloud formation. The hybrid approach ensures that the output of the machine learning modules is consistent with the laws of physics.

The resulting model can then be used to forecast weather for days and weeks ahead, as well as look months and years ahead for climate predictions.

The researchers compared NeuralGCM with other models using a standardized set of prediction tests called Weather Bench 2For three- and five-day forecasts, NeuralGCM performed as well as other machine learning-based weather models such as royal And GraphiccastFor longer-range forecasts, over ten and fifteen days, the accuracy of the NeuralGCM model was roughly comparable to the best existing conventional models.

The NeuralGCM model has also been successful in predicting less common weather events, such as tropical cyclones and atmospheric rivers.

Why Machine Learning?

Machine learning models rely on algorithms that learn patterns in the data they are fed, and then use that learning to make predictions. Because climate and weather systems are so complex, machine learning models require massive amounts of historical observations and satellite data to train.

The training process is very expensive and requires a lot of computer power. However, once the model is trained, it is fast and cheap to use for forecasting. This is a big part of its appeal for weather forecasting.

The high cost of training and low cost of use are similar to other types of machine learning models. For example, GPT-4, It said It took months of training at a cost of over $100 million, but he is able to answer any query in moments.

One of the weaknesses of machine learning models is that they often struggle in unfamiliar situations – or in this case, extreme or unprecedented weather conditions. To do this, model You must be able to generalize or extrapolate beyond the data it was trained on.

NeuralGCM seems to be better at this than other machine learning models, because its physics-based core provides some basis for reality. As the Earth’s climate changes, extreme weather events will become more common, and we don’t know how well machine learning models will be able to keep up.

No one is actually using machine learning-based weather models for daily forecasting yet. However, it is a very active area of ​​research – and one way or another, we can be confident that future forecasts will involve machine learning.

This article was republished from Conversation Under a Creative Commons license. Read Original Article.