AI can speed analysis and the assembly of forecasts, and is being employed at the weather scale, which is beyond nowcasting, extending forecasts into the next couple of days.
Thanks to simulations and modeling, we’ve come a long way in our ability to predict – at least on a short-range basis – extreme weather events such as hurricanes and droughts, as well as day-to-day weather. When a hurricane approaches, residents and municipalities within its path may have up to a week’s pre-warning that trouble is coming. Now, the power of AI is helping to amplify the power of accuracy of these simulations even further.
“Weather forecasting has generally used physics-based models that are mostly very computationally intensive,” according to Amy McGovern, leader of the National Science Foundation’s AI Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography at the University of Oklahoma, quoted in an article released by the National Academies. “They take a long time to run, and they’re always behind if you’re trying to do real-time forecasting. And they’re not as accurate as you go out in time, to the 10-day or 14-day scale. There are a variety of ways that AI can be used to try to improve that.”
The speed at which AI can analyze and assemble forecasts is a factor. “AI is being used to improve ‘nowcasting,’ which is forecasting on a super-short time horizon – usually zero to 60 minutes,” said McGovern. “We’ve done that with hail nowcasting, which tries to provide a forecast for the probability of severe hail over the next hour.” In addition, AI is being employed at the “weather scale,” which is beyond nowcasting and into the next couple of days.
“Scientists are using AI in lots of other ways to improve the models themselves – like putting AI pieces into those models so that they can be a lot faster and more accurate. We can take in larger amounts of data in the same amount of time, for example. They are working on making pure AI forecasts, but those are not ready for primetime yet. They will be, but they’re not yet.” For example, there are biases between various weather models that researchers hope AI will help aggregate.
See also: AI’s Role in Weather Forecasting Still Cloudy
Weather Forecasting Gets an AI Assist from Technology Providers
To advance AI’s capabilities further, NVIDIA, the AI processor provider, has announced two microservices that can boost climate-change modeling by a factor of 500. This capability is being added to NVIDIA’s Earth-2, a digital twin platform for simulating and visualizing weather and climate conditions.
The new “NIMs” (or Neural Inference Microservices) offer climate technology application providers advanced generative AI-driven capabilities to assist in forecasting extreme weather events.
NVIDIA NIM is the company’s set of inference microservices designed to accelerate the deployment of foundation models on any cloud or data center. The CorrDiff NIM and FourCastNet NIM microservices are intended to help weather technology companies more quickly develop higher-resolution and more accurate predictions.
“High-resolution forecasts capable of visualizing within the fewest kilometers are essential to meteorologists and industries,” according to a statement from NVIDIA. “The insurance and reinsurance industries rely on detailed weather data for assessing risk profiles. But achieving this level of detail using traditional numerical weather prediction models like WRF or High-Resolution Rapid Refresh is often too costly and time-consuming to be practical.”
The CorrDiff is 500x faster than traditional microprocessors, and is intended to increase the resolution of lower-resolution images or videos — for the entire United States and predicting precipitation events, including snow, ice and hail, with visibility in the kilometers.
NVIDIA Earth-2 combines the power of AI, GPU acceleration, physical simulations, and computer graphics to simulate and visualize weather and climate predictions at a global scale. The platform consists of microservices and reference implementations for AI, visualization, and simulation. NVIDIA NIMs for Earth-2 allow users to leverage AI-accelerated models to optimize and simulate real-world outcomes for climate and weather.
FourCastNet NIM is a service for global weather forecasting, enabling enterprises to develop solutions that use large datasets by up to 20X larger to capture extreme weather events. CorrDiff NIM is designed to help researchers generate more datasets to get better probabilistic distributions for weather events.
