With the collaboration of NVIDIA and Lockheed Martin on a project for the National Oceanic and Atmospheric Administration (NOAA), the digital twin may have reached its terrestrial limit—the size of the Earth. The companies are planning to build a 3D model that shows and predicts ocean surface temperatures. NVIDIA will be using its Omniverse platform and its GPU-based servers, most likely the DGX H100, for this project.
Why Are Ocean Surface Temperatures Important?
Oceans cover 71 percent of the Earth’s surface. They interface with the atmosphere and that interaction influences the world’s weather. Normal and severe weather events—from El Niño to hurricanes—can be predicted from ocean surface temperatures. Ocean surface temperature data is also used by conservatists and biologists (sea turtle and whale tracking, coral reef monitoring, etc.) and commercial fishing operations.
How Are NVIDIA and Lockheed Martin Going to Help?
NOAA collects, stores and displays terabytes of temperature data received from satellites, buoys, ships, ocean reference stations and marine telemetry. NVIDIA and Lockheed Martin Space aim to help by providing hardware and software systems to analyze and display the vast amount of data.
As described in the press release, “Lockheed Martin’s OpenRosetta3D platform will utilize AI and ML (machine learning) to ingest, format and fuse observations from multiple sources into a gridded data product and detect anomalies. NVIDIA Omniverse Nucleus, the collaboration and database engine of its Omniverse world simulation platform, will convert data into the Universal Scene Description framework, enabling data-sharing across multiple tools and between researchers. Agatha, a Lockheed Martin-developed visualization platform, will ingest this incoming data from Omniverse Nucleus and allow users to interact with it in an Earth-centric 3D environment.”
NVIDIA and Lockheed Martin Have a History of Collaboration
This is not the first time Lockheed Martin and NVIDIA have collaborated on a project. They previously teamed up to show wildfire information using Lockheed Martin’s AI/ML platforms and NVIDIA’s Omniverse for the U.S. Department of Agriculture Forest Service and Colorado Division of Fire Prevention & Control (DFPC).
The wildfire project, announced in November 2021, used AI to predict wildfire spread. Showing a wildfire in its surroundings in 3D greatly helps in predicting its spread. Fire spread speed is a function of slope (fire travels much faster uphill), the amount and condition of fuel (vegetation of varying dryness, which can be judged by color) and wind. Adoption of 3D computer graphics to understand fire behavior is far superior to 2D maps with contour lines to show elevation that attempt to show vegetation with symbols or a legend.
The Hardware
The AI number crunching necessary to display and analyze data (including prediction) for the ocean temperature project is expected to run on NVIDIA’s DGX H100 systems. The DGX H100 is a mini supercomputer, small enough that you can put it under your arm but packed with 80 billion transistors in the form of eight NVIDIA H100 GPUs with 640 GB of total GPU memory. Built specifically for AI applications, the DGX H100 has 30 TB of SSD storage. If all that horsepower is not enough, you can stack and array DGX H100s to get the horsepower you need.
Digital Twins Taking Over the World
You may have heard the term “digital twin” used, misused and abused often enough for it to have become jaded. True, it is the latest buzz phrase. No tech conferences can avoid its mention. Bentley’s annual customer showcase, the Year in Infrastructure, was replete with digital twins. Siemens wants to have all the industries it services (railroads, automotive, energy, manufacturing) to have digital twins. For Autodesk, BIM is only a stepping stone toward the digital twin of a building. Dassault Systèmes has a digital twin of a heart and a brain. It is a deluge that started with a trickle. A bicycle equipped with sensors at a 2015 PTC conference that used the Internet of Things (IoT) to make a digital model mirrored every movement of the bicycle. Was that not a digital twin?
What Is Omniverse?
NVIDIA’s Omniverse is an extensible software platform for 3D simulation and collaboration. It’s a bit like a game engine, though the company isn’t keen on the comparison.
“Omniverse is very different than a game engine,” said NVIDIA’s larger-than-life CEO, Jensen Huang, at GTC 2022, the company’s annual conference. “Omniverse is designed to be datacenter scale, and hopefully someday, planet scale.”
The name begs a comparison to another trendy verse: the metaverse. Popularized primarily by Facebook (which went ahead and renamed itself Meta), the metaverse is often likened to a 3D evolution of the Internet, a limitless virtual world in which you can interact with others—whether to socialize, play or work. Game are already there, with titles like GTA (Grand Theft Auto), which allow players to immerse themselves in massive 3D worlds. With Omniverse, design, engineering and scientific professionals can be right behind.
Hardware and AI to Save the World
With the ocean temperature 3D model and the forest fire model, NVIDIA leads the way with planet-scale digital twins.
Having easy, visual access to oceanic surface temperatures should help scientists who are studying climate change. Ocean temperatures are good indicators of icecap melting and atmospheric warming due to greenhouse gases.
We can only hope the temperature of the ocean, including past (using historical data), present (real time view of data) and future (forecasts from analyses and modeling) will one day be as easily viewable to us as Google Earth is today. Not just by U.S. meteorologists but by scientists in all countries, and even the public.