Scientists Model a Digital Earth to Simulate Climate Change

Global network connection and data connections concept.

Scientists have warned global leaders for decades about the increasing damage climate change is doing to our planet and the dire consequences if actions are not accelerated. Now, an artificial intelligence company, NVIDIA, has developed three technologies that could bring scientists closer to simulating (and literally demonstrating) regional climate change for policy-makers and citizens in ways they can visualize. GPU-accelerated computations, physics-informed AI models, and AI supercomputers are the technologies NVIDIA is using to produce the digital twin of Earth. Appropriately named Earth-2, the model would allow climate scientists to replicate and explore effective mitigation and adaptation strategies.

“Economists, biologists, scientists, companies in countries all over the world will be able to … simulate whether the dams that they’re building in Venice are going to make a difference,” said Jensen Huang, CEO of NVIDIA.

Why This Matters

Time is running out to solve climate change -- or merely survive it. Many mitigation and adaptation strategies have been proposed in regions around the world. Still, it is difficult to foresee, let alone model, what effects or implications climate policies will make in the long run. A digital replica of our planet’s physics, biological cycles, feedback loops, and weather patterns could give policymakers and scientists a glimpse of how specific strategies could play out over time to determine which ones are effective.

Literature vs. AI

There are several climate simulators widely available online, and they are accessible to people of all ages. Many of them create visual predictions of climate change by considering factors such as policy systems and current greenhouse gas emission rates.

En-ROADS is one such model created by MIT scientists that calculates roughly 14,000 equations to develop an energy-economy integrated assessment. Although these simulations are effective and well-informed, they are literature-based. However, Earth-2 uses AI, which constantly synchronizes and calibrates to the real world through precise measurements. Therefore, NVIDIA’s simulation can predict climate change and policy implications more accurately, giving scientists an edge in effectively tackling climate problems.