Computer models now with triple capacity

[cycloneye]

WASHINGTON (AP) -- The National Weather Service has tripled its computer capacity in an effort to get ahead of the weather by crunching numbers faster than ever.

Millions of weather measurements are fed into computers every day and run through programs called weather models that use complex formulas to calculate the weather and how it is likely to change.

With increased computing speed, those models can be run more quickly and can use more data to improve forecasting.

"Literally, we are going from making 450 billion calculations per second to 1.3 trillion calculations per second," David L. Johnson, director of the NOAA National Weather Service, said in a statement Thursday.

Louis W. Uccellini, director of NOAA's National Centers for Environmental Prediction, said the increase in computing power means the agency will be able to run higher resolution models with more sophisticated applied physics and use these models in the prediction of hurricanes, floods, tornadoes and winter storms.

"The advanced computers are critical to advancing NOAA's ability to make ever-increasingly accurate weather forecasts and climate outlooks," he said.

The new supercomputers are part of a $180 million, nine-year contract with IBM


Let's wait and see if the models get more better than before with this new capacity for them to compute the numerical guidance and have a much better long range forecast.

[Aquawind]

The quality will improve undoubtedly. It will take them awhile to make the proper changes for a significant improvement. Of course the quality of the data going in is key. The programmers are drueling alrighty. It simply speeds up the process to improve the input and thus get more reliable data. We'll take it. I wonder if a bunch of new/additional models will come out because of this. Maybe some of the model masters have some input?

Paul

[MGC]

Great, now the model can arrive at the wrong conclusion quicker.....MGC

[Aslkahuna]

Quality and Quantity of data is certainly an important factor as is speed of computation BUT if the Physics of the model (those complex equations) aren't refined then as the previous message said is true-we will arrive at a wrong conclusion faster.

Steve

[ALhurricane]

I posted this in the winter forum as a response, so I guess I'll put it here as well...

As my synoptic professor always preached to my class, computer models will only be as good as the data going into them. It's great we're getting faster machines, but the truth of the matter is that until we get more data into the models, then I think the improvements will not be as spectacular.

I know it's a pipe dream, but we really need more RAOB sites in the CONUS. Throw in the fact that the Pacific and Atlantic are data void and will be for some time, then I think significant modeling improvements are still a long way off.

Just my 2 cents, but of course I welcome all technological improvements.