Nvidia has created a digital twin of the Earth's climate as a way to improve both weather forecasts and how to predict how the system will change over the next 100 years.

Dubbed "Earth-2," the new model combines data from the climate technology industry with AI to create a new emulation of how the climate works and how it adapts to different scenarios.

A digital twin of Earth's climate already exists, but according to Nvidia, this cloud-based solution runs on GPUs rather than CPUs and can provide predictive models in seconds rather than hours.

These forecasts are rendered in high resolution and can show cloud cover, wind speeds, or even the path of a typhoon. [Earth-2 can be used by weather forecasters, researchers, and other professionals to make predictions about the climate, how it is changing, and to visualize the weather more accurately.

CorrDiff uses the same diffusion modeling techniques as text-to-3D or text-to-image models such as MidJourney and Luma Genie to produce higher-resolution images than current models.

According to Nvidia, the images are 12.5 times higher resolution than previous models, can run models 1,000 times faster, and are 3,000 times more energy efficient because they run on a GPU.

The company claims that the technology corrects many of the inaccuracies in previous forecasts and predictions caused by coarse resolution and uses synthetic data and predictive modeling to compensate for missing information, allowing critical decisions to be made faster. [Earth-2 was built for a variety of uses, but primarily to connect to billions of dollars worth of proprietary data owned by climate change companies, governments, and weather forecasting giants like The Weather Company.

"To effectively address current and future weather and climate-related challenges, it is more important than ever to incorporate reliable, globally scaled, actual weather data and insights into a digital twin environment to better analyze, plan, and simulate weather impacts It is more important than ever to better analyze, plan and simulate weather impacts by incorporating globally scaled, real-world weather data and insights into a digital twin environment," said Sheri Buckstein, CEO of The Weather Company.

"We have long collaborated with NVIDIA on GPU acceleration of our proprietary weather modeling system, GRAF, and with the Earth-2 API, we plan to deliver higher resolution, more energy efficient simulations at a lower cost.

Jensen Huang, founder and CEO of Nvidia, said Earth-2 will help encourage human action to better prepare for and mitigate extreme weather events caused by climate change. [Historic droughts, devastating hurricanes, and generational flooding appear in the news with alarming frequency.

In addition to simple forecasts, Taiwan hopes to use them to predict the path of typhoons and the impact they might have on global supply chains.

"Taiwan is an important component of the global supply chain, and flood risk analysis and evacuation preparedness is at the core of our mission," said Chia-Ping Cheng, CWA administrator.

More than 130 typhoons have hit Taiwan since 2000, and it is hoped that Earth-2 will help authorities mitigate the risk of future typhoons by understanding when and where they will strike.