Supercomputing Upgrade Produces High-Resolution Storm Forecasts

dcblogs writes 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. This improvement in modeling detail is a result of a supercomputer upgrade. 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.

Comments

Have the solutions converged?

[RichMan]

I was at a supercomputing conference back in the 90's. There were wonderful reports on doubling the resolution of the grid and so on. Advances in the scale are all good.

The questions are
a) with the increase in detail of the simulations have we converged on a solution. That is do solutions at scale N and 10N match. If they do then the resolution and model are aligned for accuracy in the solution.
b) do the simulations agree with reality.

If a) and not b) then there is something wrong with the model that is not related to compute power or problem resolution, and no amount of compute power will fix it.

Re:Have the solutions converged?

[pushing-robot]

TFA ironically begins with the quote '"I don’t think we will ever have enough [computing power] to satisfy us,” says researcher.'

The summary is vague, and the article not much better, and neither say anything about whether the 'new model' is matching observations any better than the old.

It would be nice if they could at least clarify if the sole pair of comparison images are even the same forecast, because the new model shows not only more detail but a completely different prediction.

Come on kids, this isn't a network news sound bite. This is the Internet, and you're a tech news site. Would it kill you to go past the press release?

Maybe I'm just bitter about this because I live in mountains where a coin is a more accurate forecasting tool than the weather service.

Re:Have the solutions converged?

[EETech1]

TFL in TFA goes over it.

www.computerworld.com/article/2484337/computer-hardware/noaa-goes--live--with-new-weather-supercomputers.html

It's been a complicated process to get to this point. The NWS has had to ensure that the software running on the new system is producing scientifically correct results. It had been running the old and new systems in parallel for months, and comparing the output.

This comparative testing involved examining output data to determine whether it is numerically reproducible out to five decimal places. There is also a statistical analysis of weather predictions on the new system against the actual weather conditions.

The process wasn't just an examination of numerical data. NWS scientists also studied the weather products and examined them for subtle differences. "There is a lot of human, highly experienced, subjective evaluation," said Kyger.

There are computational differences involved in switching to new chips and a new operating system. They are subtle, and appear in decimal places six through 12.

As you go further out in a forecast, the differences compound. The changes may appear in the fifth day of an extended, five-day forecast as a difference of one degree.

And as the resolution increases ...

[Thorfinn.au]

Better resolution is good, but with each improvement in the system, the input data also needs to be improved and remeasured.
Ultimately the ground features need to be modelled in greater detail to match the increased resolution of the grid.
Which comes own to knowing where each tree/building and similar sized static feature is and how this affects the model.
However, as the grid increases it should not need to know where the butterflies are .

WRF has gotten pretty good, actually

[erikscott]

I'm a computer engineer, not a meteorologist, but I've worked with them off and on for about eight years now. One of the most common models for research use is "Weather Research and Forecasting Model" (WRF, pronounced like the dude from ST:TNG). There are several versions in use, so caveats are in order, but in general WRF can produce really good results at a 1.6KM grid for 48 hours in the future. I was given the impression that coarser grids are the route to happiness for longer period forecasts.

WRF will accept about as much or as little of an initializer as you want to give it. Between NEXRAD radar observations, ground met stations all over the place, two hundred or so balloon launches per day, satellite water vapor estimates, and a cooperative agreement with airlines to download in-flight met conditions (after landing, natch), there's gobs of data available.

The National Weather Service wants to run new models side-by-side with older models and then back check the daylights out of them, so we can expect the regular forecast products to improve dramatically over the next (very) few years.