Addressing the climate crisis with a digital twin of Earth

The increasing adoption of big data analytics (BDA) has acted as proof of the opportunity and value presented by predictive modelling. Organizations are able to forecast widely divergent factors such as future demand for products and services, the possible effects of supply chain disruptions, business growth modelling, and more. And as we grow more adept at managing data and connecting the dots across different categories of data, the opportunities presented by predictive analytics will continue to grow.

Climate change is one of those issues that BDA seems to fit perfectly: there is the opportunity for a small change today to have a large effect in the future; but convincing people to make those small changes, some of which are costly, can be a challenge. Often, the investments in the environment that are made today will not have measurable effects until many years, even decades, into the future.

Until recently, it has been difficult to create a model of climate change that can show the effects and progression of extreme weather across decades, across the entire globe. Predicting the weather is largely based on physics: the interaction of hot and cold air and water in the atmosphere. A climate model, on the other hand, must incorporate variables across atmospheric physics, but also chemistry (like carbon emissions), biology, water and ice and land, and human activities.

In essence, to create a comprehensive climate change model, a complete digital replica of the Earth is required: a digital twin. To that end, NVIDIA has released their plans to build a supercomputer dedicated to predictive climate change modelling. The system will be built on Omniverse, the company’s open simulation platform and will be known as Earth-2 (E-2).

To date, climate simulations have been created at a resolution of 10 km to 100 km. But these models do not provide the level of detail that is needed to model changes to the water cycle across the globe: those changes that will cause longer, more severe storm systems and water shortages. Additionally, these models cannot simulate atmospheric details, like the reflective properties of clouds; and any variation in these models can throw results off tremendously, particularly in the case of a multi-decade simulation.

E-2 is intended to be the ultra-high resolution model that is needed to accurately demonstrate the effects of climate change years into the future: bringing the resolution down to a single metre.

A comparable system was recently unveiled to conduct healthcare research. Cambridge-1 was created to advance medical research and cooperation across a number of healthcare and related organizations: from the NHS, to King’s College, to AstraZeneca. Dedicating this computing power (and an estimated $100 million investment) to improving medical research was elevated to a top priority during the global pandemic. Now, NVIDIA is dedicating similar resources to predicting climate change.

In the announcement of E-2, the company said, “So, we will dedicate ourselves and our significant resources to direct NVIDIA’s scale and expertise in computational sciences, to join with the world’s climate science community.”

“All the technologies we’ve invented up to this moment are needed to make Earth-2 possible. I can’t imagine a greater or more important use.”

Even in the face of a unified scientific community, climate change deniers can be found worldwide.

https://cdn.statcdn.com/Infographic/images/normal/19449.jpeg

But creating a comprehensive, meaningful model of climate change – and the potential costs and benefits of strategies to combat it – could have an enormous effect on individual, organization, and government commitment to reaching critical climate change goals.