Artificial Intelligence to Predict Extreme Weather Conditions: Is It Possible?
Researchers have found a way to produce an artificial intelligence algorithm that has the ability to predict incoming storms.
A computer model was created to assist weather forecasters in determining incoming storm signals. The computer predicts faster and more accurately, thanks to the group of researchers from Penn State, AccuWeather, Inc., and the University of Almeria in Spain. The group developed an artificial intelligence that can locate cloud movements using satellite images. The AI is found on the Bridges supercomputer, located at the Pittsburgh Supercomputing Center.
Accuweather and Artificial Intelligence
According to Steve Wistar, a senior forensic meteorologist at AccuWeather, the AI will be an incredible addition to their team and it will help him serve the people better.
“When it comes to forecasting, there are countless data you need to analyze,” Wistar said. “The atmosphere drastically changes over time and so we need to record every bit of information that we can every single time to report accurate weather condition.”
With this, the developer of the AI worked hand in hand with Wistar and other meteorologist at AccuWeather to evaluate over 50,000 historical satellite images that depict the weather in the US.
In their study, the team determined a persistent cloud pattern which they call the “comma-shaped” clouds. This cloud formation is associated with thunderstorms, hail, blizzards, and high winds. Then, the researchers input this information to the computer and program the device to automatically detect comma-shaped clouds using satellite images.
“Since comma-shaped clouds are indicator of extreme weather conditions, the AI we programmed can assist meteorologists to forecast severe weather events,” stated Rachel Zheng, a doctoral student and one of the main researcher.
Is the Artificial Intelligence Reliable in Predicting Weather? According to researchers, AI can recognize comma-shaped clouds with 99% of accuracy for every 40 second prediction. “The AI can determine comma-shaped clouds just like a human being,” said Zheng. “It can also recognize pre-developed comma-shaped cloud,” added the researcher. “This advanced scheme aims to protect the people and their property against incoming weather conditions. As much as possible, we wanted to save lives by giving accurate and on time weather forecast,” Wistar stated. In addition to this, the project also allowed AccuWeather and the College of IST research group to see the struggles of climatologists and meteorologists. “This collaboration made us realize how much struggle our weather forecasters face in interpreting large quantities of data from different source including sensor networks, radars, and satellites,” stated James Wang, the leader of the research group. “The AI we have developed can give accurate and timely interpretation of the condition of the atmosphere especially during extreme weather conditions,” he said. “This research is our first-attempt in creating an artificial intelligence that can interpret weather-related visual data and report it to the research community. There will be more research after this because we wanted to help meteorologists report precise and well-timed forecast,” Wang added. Weather conditions are inevitable and they are ever-changing. That’s why researchers and weather experts count on this newly-developed AI to help them report useful weather updates.