The news comes as the 27th Conference of the Parties (COP27) puts an international spotlight on climate change and as effects of the changing climate are increasingly felt.
NOAA has contracted two firms to create a digital twin simulation of the Earth’s weather and climate patterns. The digital twin — called Earth Observation Digital Twin (EODT) — is currently a prototype.
It is intended to pull together and synthesize terabytes of data from NOAA’s satellites and sensors to present researchers with a fuller, easier-to-interpret picture of weather and climate, according to a press release from technology company NVIDIA. NVIDIA is collaborating on the project with aerospace company Lockheed Martin.
The digital twin is also intended to update frequently with new data, to stay current.
“Such a system has the potential to reduce the output of complex weather visualizations from hours to minutes,” the release states.
The project relies on both artificial intelligence — including machine learning — and digital twin technology.
NOAA’s devices collect details about the atmosphere, cryosphere, land, ocean and space. This generates a vast amount of information, in different data formats, which can be difficult for analysts to interpret without technology, said Matt Ross, senior manager of Strategic Programs at Lockheed Martin, during a press briefing.
Machine learning helps sift through all this data and flag changes in the situations being observed that researchers may want to analyze further.
“Through artificial intelligence and machine learning, we’re able to highlight anomalies in this data, which can reduce that sea of data down to a set of interesting points the experts can make use of, in order to provide insight and make decisions,” Ross said.
Some digital twins are also used to model how different interventions might create different outcomes and to make predictions. But the EODT, at least at present, will focus solely on capturing and accurately reflecting existing situations, NVIDIA lead product manager of accelerated computing, Dion Harris, told GovTech during a press briefing.
The data the EDOT compiles could potentially lay the foundation for other projects that do assess potential climate actions, however, such as AI surrogate models of climate mitigations, Harris said.
“This is really focused [on] the observation part of it,” Harris said. “… But that’s sort of the first step to being able to do some of those other more advanced things that you would expect to do in terms of building mitigative solutions.”
UP FIRST: SEA SURFACE TEMPERATURE
As the EODT project kicks off, its first focus will be a data visualization of the sea’s surface temperature. That phase — slated for September 2023 — will pull NOAA sensor data about details like solar wind, sea ice concentrations, and temperature and moisture profiles.
Rising sea surface temperatures — driven by climate change — can have significant impact on states and local jurisdictions.
Shifts in these temperatures can change the “circulation patterns that bring nutrients from the deep sea” to feed fish living in surface waters, the Environmental Protection Agency (EPA) notes. That can challenge communities that depend on strong fish populations for food or livelihoods.
Warmer sea surface temperatures also lead to more water vapor in the air over oceans, which can in turn result in greater risks of heavy rain and snowfalls, per the EPA. At the same time, changing sea temperatures can “shift storm tracks, potentially contributing to drought in some areas.”
Drought and precipitation continue to be pressing issues for some areas. This October, some areas along the Mississippi River saw water levels fall to the lowest seen in a decade, NOAA found. This tangled water-based transportation during what is typically a significant period for shipping crops, leaving ships backed up or running aground. Other states saw unusually heavy precipitation; for example, that same month, New Jersey experienced its “10th-wettest October on record.”
