IBM is putting its supercomputing power behind an effort to launch the world's most sophisticated weather forecasting models.
The Weather Company, an IBM subsidiary, is teaming up with the University Corporation for Atmospheric Research (UCAR) and the National Center for Atmospheric Research (NCAR) to develop these models using IBM Power Systems and OpenPOWER technologies. They'll be optimized to run on IBM's next generation POWER9-based systems, which are slated for commercial release at the end of this year.
Sophisticated regional weather forecasting has been around for more than a decade, and global forecasting models exist as well. This collaboration, however, is promising an advanced, global system that wasn't previously feasible.
"Weather forecasting is one of the most challenging things in computer architecture today," Mary Glackin, head of weather science and operations for The Weather Company, told ZDNet. Because of the advanced physics and interdependencies involved, "meteorologists and weather modelists can use all of the computing power you've got," she said.
The first part of this new effort involves adapting NCAR's Model for Prediction Across Scales (MPAS) community model to run more efficiently on next-generation computers. Even small improvements in efficiency will translate into improved accuracy and resolution, Glackin said. While the model will be global, it will also offer increased granularity, pinpointing where storms and other weather systems are developing and forecasting how they will evolve. The model could, hours in advance, forecast weather events at county and sub-county levels.
Improved weather forecasting has clear benefits for individuals and communities, making it easier to keep people safe, as well as for industries such as agriculture vulnerable to weather conditions.
"As we're able to make finer-scale forecasts more accurately you'll see better decisions being made, and that should translate to benefits around the world," Glackin said.
Meanwhile, improved modeling is just one element of The Weather Company's and IBM's efforts to develop cutting-edge weather forecasting. Simply observing weather will become more sophisticated, Glackin said, as IBM leverages IoT sensors and data pulled directly from endpoint devices. Even connected cars should eventually offer useful data on factors such as precipitation and road conditions.
Additionally, there's room for IBM's cognitive capabilities to come into play, leveraging data streams and forecasting models to enable "tailored" decision making, Glackin said, for both business and consumer needs.