Artificial intelligence (IA) becomes a strong ally to improve climate forecasts and anticipate environmental risks. Nvidia's latest investment is the Earth-2 system, which would function as a “digital twin” of planet Earth to predict climate phenomena that put the population at risk. The platform seeks to support tornado prevention, cyclones, floods and other catastrophes, in addition to offering an in-depth overview of the long-term effects of climate change.
Nvidia's system works through machine learning. Training with global data allows you to simulate terrestrial climate variables, atmospheric and oceanic. In this way, the platform quickly analyzes data on greenhouse gas emissions, temperature, water cycles, soil changes, ocean chemistry and other factors, building more accurate predictive simulations.
Platform creates high-resolution climate simulations
The platform combines machine learning technologies running on supercomputers to create a virtual replica of physical systems, such as those that operate focused on Earth's climate. Second Dion Harris, Head of Data Center Product Marketing at Nvidia, the model allows us to more accurately simulate how the climate may evolve over time.
The system will make it possible to synthesize hundreds of observations and reconstruct Earth's weather and climate over the last few years. 50 years or more. This historical data will enable advanced analytics, with weather and climate simulations in different locations at a detailed scale of up to 2 kilometers. The system is expected to be initially tested in Taiwan, region frequently hit by typhoons, supporting the planning of measures to reduce the impacts of these phenomena.
