Author: Tronserve admin
Tuesday 27th July 2021 12:41 AM
U.S. Will Triple Weather Computing Capacity - Is It Enough To Surpass The European Model?
There is so much talking and opining about the status of the United States and its weather modeling capabilities these days. Candidly, it has gotten to the point that I sort of “eye roll” when I see a Tweet or Blog about it. Yes, the European Center for Medium-Range Weather Forecasts (ECMWF) or Euro model leads the pack in jargony metrics utilized in the meteorological community to establish superiority. Debates on Twitter and sidebar conversations at the Waffle House about “Euro vs American” models have confused the public into thinking the U.S. model is some version of a rotary phone compared to a smartphone. That’s simply not the case. The Euro and American GFS global modeling system, which I have written about numerous times, are both advance smartphones (to carry the analogy forward). The best forecasters use both of them along with an array of other modeling tools. Meteorologist Mike Lyons provides an excellent overview at this link about the models and why computing capabilities have historically given the Europeans a slight advantage. However, the United States government made a major announcement this week that may signal that its model is making a play for weather model front-runner.
Supercomputing is at the heart of better weather prediction capabilities. NOAA announced a tripling ... [+] NOAA COMMUNICATIONS WEBSITE
Is this enough to surpass the Europeans? I don’t know - maybe. To me, it really doesn’t matter whether the U.S. is leading the pack. What is more important to me is that all of the models are advancing to give us the best possible weather prediction capability to save lives. If this “so-called” weather modeling arms race pushes innovation and enhanced prediction capabilities, I am all for it. If the U.S. ends up slightly ahead, so be it. I should also mention that the United Kingdom weather modelers recently announced massive computing upgrades that rival (or perhaps surpasses) European and U.S. efforts.
At higher levels, administrators and meteorologists laugh at the notion that this is some type of competition. As president of the American Meteorological Society a few years back, I literally watched a high-level ECMWF official chuckle at dinner with me in England when I mentioned bickering about the models. Another point that goes overlooked is that the U.S. federal weather enterprise has to leverage its budget in many directions: new satellites, radars, forecast offices, and so forth. The National Weather Service budget is roughly $1 billion and is one of the best investments that our government makes on behalf of its citizenry. Frankly, I would like to see that budget increased given the importance of weather to our economy, transportation network, agriculture, national security, energy generation and more.
ECMWF has a different paradigm and budget strategy so can sink resources into the biggest, fastest supercomputers. They also use a sophisticated technique for ingesting satellite and other datasets into their model. By the way, much of that satellite data is from U.S. and other international satellites so the success of the Euro model is not independent from other partnerships.
Forecast for Hurricane Sandy (2012). The Euro model predicted the left turn into the U.S. several ... [+] NOAA
The U.S. has big computers too. However, the National Oceanic and Atmospheric Administration (NOAA) announced this week that it is tripling supercomputing capacity for weather and climate modeling over the next few years. According to a NOAA press release, two new Cray supercomputers will come online by 2022 and extend supercomputing capacity to 40 petaflops. Since the average person doesn’t speak technical computer language, it is useful to just understand that more petaflops mean faster computing. Faster computing capacity enables solving equations that help predict weather with more information, higher spatial resolution, and better representation of rapidly-evolving meteorological processes.
This increase in high performance computing will triple the capacity and double the storage and interconnect speed, allowing NOAA to unlock possibilities for better forecast model guidance through higher-resolution and more comprehensive Earth-system models, using larger ensembles, advanced physics, and improved data assimilation. The new computers will provide operational capacity to implement research and development advancements made under NOAA’s emerging Earth Prediction Innovation Center (EPIC) to make the U.S. Global Forecast System the best model in the world.
--NOAA Press Release--
Neil Jacobs is the acting Administrator for NOAA. He said in the press release, “NOAA is excited for the incredible opportunity ahead to partner with university and industry scientists and engineers (on EPIC) to advance U.S. numerical weather prediction, and this supercomputer upgrade lays the foundation for that to happen.” National Weather Service Director Louis Uccellini added, “The National Weather Service ran a competitive acquisition to ensure we have the supercomputing power needed to implement all the great modeling advancements we anticipate over the next several years.” Through a competitive process, NOAA awarded the 8-year contract to CSRA LLC, a General Dynamics Information Technology company.
Meteorologists are bullish on the announcement. Ryan Maue is a meteorologist and one of the top experts on weather model applications. His Tweet on Thursday sums up the feeling of many of us, “Huge! NOAA has awarded $1/2 billion contract in order to triple supercomputing capacity for weather forecasting.”