Machine Learning to Predict Severe Weather

Meteorologists use a number of models and data sources to track shapes and movements of clouds in order to predict weather conditions. However, with expanding weather data set and looming deadlines, it is becoming nearly impossible to monitor all storm formations.

To overcome these huge data sets, a team of scientists at Penn State, AccuWeather, Inc., and the University of Almería in Spain has developed a computer model which is based on Machine Learning. This model can help weather forecasters to perceive the potential of severe storms rapidly and precisely by using linear classifiers that identify rotational movements in clouds from satellite images. This AI solution ran on the Bridges supercomputer at the Pittsburgh Supercomputing Center.

During the study, these scientists worked with Wistar and other AccuWeather meteorologists in order to analyze more than 50,000 historical U.S. weather satellite images. Experts identified and labeled the shape and motion of ‘comma-shaped’ clouds in these images. The cloud patterns thus identified are strongly associated with cyclone formations, which can lead to severe weather conditions. Then by using machine learning algorithms, scientists taught computers to recognize and detect these comma-shaped clouds automatically in the satellite images. The computers can hence assist experts by pointing out in a real-time where to focus in order to detect the onset of severe weather.

The researchers found out that their method can detect comma-shaped clouds with 99% accuracy, at an average of 40 seconds per prediction. Also, it was able to predict 64% of severe weather events. This kind of accuracy can surely outperform other existing severe-weather detection methods.

Rachel Zheng, a doctoral student in the College of Information Sciences and Technology at Penn State and the main researcher on the project, stated that “Our method can capture most human-labeled, comma-shaped clouds and it can also detect some comma-shaped clouds before they are fully formed. Our detections are sometimes earlier than human eye recognition.”

To this, Wistar added, “The calling of our business is to save lives and protect property. The more advanced notice to people that would be affected by a storm, the better we’re providing that service. We’re trying to get the best information out as early as possible.”

This project enhances earlier work between AccuWeather and a College of IST research group which was led by professor James Wang, who is the dissertation adviser of Zheng.

Wand said, “We recognized when our collaboration began that a significant challenge facing meteorologists and climatologists was in making sense of the vast and continually increasing amount of data generated by Earth observation satellites, radars, and sensor networks. It is essential to have computerized systems analyze and learn from the data so we can provide a timely and proper interpretation of the data in time-sensitive applications such as severe-weather forecasting. This research is an early attempt to show the feasibility of AI-based interpretation of weather-related visual information to the research community. More research to integrate this approach with existing numerical weather-prediction models and other simulation models will likely make the weather forecast more accurate and useful to people.”

Concluded Wistar, “The benefit of this research is calling the attention of a very busy forecaster to something that may have otherwise been overlooked.”

This article does not necessarily reflect the opinions of the editors or management of EconoTimes.