麻豆淫院

A new machine learning tool can reduce errors in national flood prediction programming, resulting in more accurate predictions of where floods will occur. In a new published in AGU Advances, scientists found that when the AI was used in combination with the National Water Model, developed by the U.S. National Oceanic and Atmospheric Administration, the resulting hybrid model was four to six times more accurate.

The AI was trained on NOAA data for the United States, but the system can be specialized for any country.

This AI is a neural network, or a deep learning model, and is trained to find errors. The network was trained on historical observational and National Water Model simulated data on rainfall and flooding. Researchers created a hybrid program by combining the AI system with the National Water Model, which forecasts streamflow for the entire United States.

"So especially for floods, the performance of the pure AI model is quite poor," said Vinh Ngoc Tran, hydrologist at the University of Michigan and head researcher on the study. "The advantage of AI models is that they are very simple. You only need to use the data to train the model and provide the forecast, but the most important thing we need to be concerned about is ensuring prediction accuracy for flood events that can cause significant damage."

Across the United States, NOAA has nearly 11,000 operational water gauges that collect data on previous floods and water levels, but NOAA also keeps track of other data than just water. The agency gathers detailed information on variables such as vegetation, urbanization, and drainage networks for the different gauges.

The sheer amount of available information is helpful, but it also makes it harder to narrow down the research to figure out where things went wrong or to simply account for everything when creating a flood model. This will cause errors in the forecasting system, and that is where AI can step in.

Aptly named Errorcastnet, the AI-based system looks for errors in the national model. It looked at old floods and what the model forecasted for the floods. For times when NOAA's model did not correctly forecast the flood, the AI would categorize the errors into two groups: the errors it could reduce and those that could simply not be fixed.

The AI learns the problems within the model and works to correct them. The errors that cannot be fixed, like limitations within the model itself or incomplete data, are still important to track. It helps the AI continue to train and improve the forecasting by focusing only on the errors that it can fix.

"You can't throw away physics," said Valeriy Ivanov, physical hydrologist at the University of Michigan and an author of the study. "It's just by definition you can't. You have to understand that systems are different. The landscapes are different. You have to account for dominant physical processes in your predictive model."

Researchers found that when only using Google's AI flood forecasting program, which uses historical data to make predictions but doesn't consider details such as elevation, vegetation, and reservoirs that National Water Model incorporates, the model would generally underpredict flood flows.

"We understand the power of AI," Ivanov said. "No one denies it. It's definitely there. But it should not negate decades of research. It should not negate the understanding of physics and understanding of complexity of physical processes in watersheds."

By bettering NOAA's forecasting model, the researchers said they think it could also improve the potential economic impacts of floods. More accurate flood predictions could mean businesses can better prepare for the coming floods. Tran, Ivanov and their team hope that as the program grows, potential floods could be predicted in detail up to several days or more before they happen.