UW professor calls out National Weather Service for inferior predictions
We’re still trying to understand why weather forecasting is so hit and miss in the U.S., especially after the storm over the weekend turned out to be much weaker than the National Weather Service said it would be.
Though we did have wind gusts of up to approximately 40 miles per hour around Puget Sound, the main storm — initially being compared to the historic Columbus Day Storm — ended up moving farther offshore. It moved enough to make winds much weaker than expected, University of Washington Professor of Meteorology Cliff Mass says.
Mass argues that the forecasts weren’t necessarily wrong. The biggest failure was communication and helping people understand the uncertainties of the incoming storm.
Still, there is something to say about the accuracy of the forecast that originally called for widespread power outages and led people to scramble for supplies. It’s a problem that Mass has been vocal about on his blog.
In a recent interview with New York Times Magazine, Mass points out that — at one point — the National Hurricane Center was off by more than 325 nautical miles in its forecasting model for Hurricane Matthew’s course. That is just one example of a government agency that has what Mass calls an underutilized supercomputer system that is responsible for keeping the public informed.
Mass says that while the U.S. has the potential to lead the way in weather prediction, that prediction right now is inferior compared to agencies such as the European Centre for Medium-Range Weather Forecasts.
“Weather prediction has become immensely better, but they are not fulfilling their potential,” Mass said of NOAA. “They could be far better.”
NOAA received a supercomputer system after European weather models predicted the path and strength of Hurricane Sandy in 2012, while U.S. models struggled to do so. The computer processes billions of observations in order to improve forecasts. Mass says that computer isn’t being used to its full potential.
According to Mass, if the computer was being used more efficiently it could run through dozens of forecast models. That would help determine the likelihood of, say, a storm brewing off the West Coast moving over the mainland. That would help save lives and even help the economy, Mass says.
“It’s really foolish not to make the investments into weather prediction,” he said.
In January, NOAA announced the supercomputer system was “running at record speed.” It has the capacity to process and analyze earth observations at quadrillions of calculations per second to support weather, water, and climate forecast models.
KIRO 7 Meteorologist Morgan Palmer says the failure to properly predict the course of Saturday’s storm “will be something meteorologists will study to determine what went wrong in the multi-million dollar supercomputers.”
The National Weather Service offered up its own explanation.
3500+ miles of open ocean + a half dozen global forecast models with differing solutions in time and space = difficult forecast. #wawx
— NWS Seattle (@NWSSeattle) October 16, 2016
At least until forecast models are improved, the grocery industry will thrive as people scramble to fill their pantries.