HPC upgrade produces high-resolution storm forecasts

A supercomputer upgrade is paying off for the U.S. National Weather Service, with new high-resolution models that will offer better insight into severe weather.

The improvement is easy to see in the depiction (below) showing the difference between the older and newer model. In a six-hour forecast on a moving storm system, thunderstorms — illustrated in red and yellow — are now made visible in the newer model, called “High-Resolution Rapid Refresh” (Hrrr). This gives forecasters better information to work with when warning about a threat.

This improvement in modeling detail is a result of a supercomputer upgrade in the summer of 2013. The National Oceanic and Atmospheric Administration, which runs the weather service, put into production two new IBM supercomputers, each 213 teraflops, running Linux on Intel processors. These systems replaced 74-teraflop, four-year old systems.

More computing power means systems can run more mathematics, and increase the resolution or detail on the maps from 8 miles to 2 miles. This is analogous to increasing the pixels or detail in a standard Web image.

“You can see so much more detail in the fields” that just wasn’t captured before, said Geoff DiMego, chief of the Mesoscale Modeling Branch at the National Centers for Environmental Protection.

The Hrrr model produces output from the model every 15 minutes versus the previous hourly rate, and forecasts up to 15 hours, although there are plans to raise that to 18 hours. As a result, when the output from the model is animated it produces “a much smoother display,” DiMego said, and forecasters can detect rapid changes happening in severe conditions, such as thunderstorms, weather fronts, and derechos (a type of wind storm), information that would not be available with the standard outputs.

The long-range goal is to move to “ensemble” of predictions, said DiMego. Right now, the weather services use individual model runs, but the ensemble approach involves multiple model runs with different formulations. With that, “you can do a much better job of capturing all the events and the variables” in the atmosphere and come up with a range of possibilities, he said.

In terms of the compute power needed to run ensembles, the more the better, said DiMego. “I don’t think we will ever have enough to satisfy us,” he said.