Last week, I wrote that the National Weather Service is not keeping up with its European counterparts in investments in supercomputing and, consequently, is not positioned to be the world leader in global weather prediction. I stand by what I wrote, but it may have unintentionally led to a misconception that the NWS is not working diligently to improve its wide-ranging array of weather forecasting models and tools. It most certainly is — but I argue that’s still not enough.
To its credit, since 2013, the NWS has doubled its operational computing capacity, substantially upgraded its HWRF hurricane model (leading to immediate improvements in forecasting), and put into operation a high resolution model for forecasting thunderstorms (known as the HRRR).
Some additional upgrades are right around the corner. In December, the Global Forecast System (GFS) model will be upgraded to a higher resolution and, in January 2015, NWS is slated to triple its current overall computing capacity.
The procurement of a new “game-changing” supercomputer is still in the works, however.
“In late January 2014, NWS was literally weeks away from executing a task order under its current contract with IBM that would have significantly increased its computing capacity and allowed the NWS to meet the goals communicated publicly in 2013,” says a NWS white paper on the status of its supercomputing. “However, IBM’s decision to leave the x86 server business by selling the line to Lenovo caught NOAA, and everyone else, by surprise and prevented us from moving forward with this procurement.”
There is no question that the NWS is making a legitimate effort to advance its supercomputing and forecasting abilities, and that progress is being made. But two important realities emerge from the current state of affairs: 1) Bureaucratic and policy obstacles have slowed the NWS effort to obtain more supercomputing power 2) Europe is — as of this moment — investing vastly more resources in supercomputing technology (per my post last week: Congress authorized approximately $24 million in forecasting equipment and supercomputer infrastructure following Superstorm Sandy, whereas the NWS’ European counterparts the ECMWF and UKMet Office, recently invested $64 million and $128 million, respectively).
NWS admits more funding is required to keep pace with the Europeans in its white paper: “Despite the ongoing and planned increases to the US operational weather compute capacity, due to these more recent announcements from our UK and European partners, the US computer capacity will not surpass the European infrastructure without additional investment.”
In email correspondence, Cliff Mass, a professor of atmospheric science at the University of Washington, described the situation at the NWS as a glass half full (or empty, depending on ones perspective). On the one hand, it is making strides in its forecasting abilities. On the other hand, to keep up with the Europeans, a pitcher of water, i.e. a large investment in computer capacity, is needed to fill the glass.
In its supercomputing white paper, the NWS equivocates some on urgency of obtaining more computing resources. It speaks of having a broader mission and more responsibilities than the Europeans – suggesting it cannot as easily focus its investments in supercomputing. “[U]nlike the European Center, we must invest in the people – the forecasters – so essential to the U.S. weather prediction capability, who work hand in hand with local emergency managers, communities and state decision makers to keep people safe in advance of and during extreme weather,” the white paper says.
The white paper also suggests that if the Europeans have the computing power and the information is being shared, perhaps the need for the U.S. to keep up isn’t as pressing. “The American public benefits from the weather prediction capabilities of all three major players – the US, the UK and the European Center,” it says.
My view is that this equivocation is a bit of a cop out and that NWS should settle for nothing less than a full glass. It should make the case to Congress that it needs the investments to have the best possible computing and weather forecasting systems to protect life and property.
Historically, despite its status as governmental organization extremely popular in Congress given its mission — not to mention talented, effective workforce — the NWS has been shy to advocate for itself according to Richard Hirn, counsel for the NWS Employees Organization, its labor union. “I can’t tell you how many times appropriators have told me -‘sure- we will give the NWS more money, but they have to ask for it!'”, Hirn said.
So will NWS leadership have the courage to do so?