One of the things that one notices when living in Korea, are the sometimes way off weather forecasts in the local news.
I don’t know how this compares to to other countries, but the official KMA figure for forecasting accuracy is 86.2% (down from 86.8%. Not bad if it’s true). But not satisfied with this figure the KMA is fighting back. How? By buying a new supercomputer and a new building to house them.
The nation will introduce a new supercomputer by 2009 for weather forecasts and meteorological studies, a state weather agency said yesterday.
The KMA has drafted the 25.3-billion-won ($27 million) construction plan. It will request the National Assembly this month to approve a budget of 10 billion won in first-stage costs, to be implemented next year.
The provisionally named National Meteorological Supercomputer Center will be built on an area of 6,627 square meters and includes an emergency electricity supply system, air-conditioning and heating equipment, and an ultra-high voltage power system.
The third supercomputer is expected to significantly improve accuracy of weather forecasts and help the nation to better tackle climate change and disasters.
When I ask people regarding inaccurate weather forecasts, they always comment that the ability for the KMA has degraded ever since they bought the supercomputers.
The state-run forecaster currently runs two supercomputers which were introduced in 1999 and 2004. But it has faced criticism due to misleading weather forecasts.
Critics and the public have been denouncing the KMA for failing to properly operate the expensive computers. They allege the computers were put under the control of inexperienced forecasters.
I think the KMA got this whole thing wrong. Yes, they should be trying to improve their forecasting accuracy, but there are better and cost effective ways to doing it instead of buying new expensive supercomputers. And from the looks of it it looks like the KMA is more interested in deflecting criticism and trying to show the public that its supercomputers work instead of taking steps that matter the most. Well, what can I say, more taxpayer money down the drain.
This begs the question, if the supercomputer and the brand new center doesn’t improve the forecasting accuracy rate, then will the KMA blame the U.S. for the inaccurate weather forecasts, considering that the supercomputers are U.S. made Crays? Not that I believe that Cray is responsible.
9 Comments
I’ll say they flushed that money down the drain. An array of Apple Xserves would be able to match the computing power of the Cray for a cost of less than 1/20th the W50 billion quoted in this article. Good work for the Cray salespeople, but what a stupid, shortsighted move by the KMA.
What about the dust that I’ve been seeing over the horizon in the last few days? I really think that we aren’t told the whole truth about the dust storms from China. Last year, for example, you could see the orange dust on the horizon in early January, but the Korean government claimed that they started in March. Are my sinuses really more sensitive than the computerized equipment that is at the disposal of the Korean government? Certainly not.
The atmospheric sciences (especially short term, 3-7 day forecasts) based on “grid” models used to provide hopefully, accurate predictions are only as good as the individual who interprets the models. Experience is a wonderful thing. Perhaps the KMA is still dealing with the learning curve on an individual basis.
The earth’s atmosphere is often referred to as “an ocean of air”. Unfortunately, the state of the art atmospheric analysis models can only look at and extrapolate small slices and/or parcels of the atmosphere. Imagine the difference in scale when comparing the ocean’s Gulf stream current to the Jet stream at 28,000 to 35,000 feet. There are many multi-dimensional eddies and currents we do not have the ability to monitor at all depths or heights. The “state of the art” computer models will definitely point to possibilities but the experienced forecaster still has to provide the human element of thought and informed decision making.
It’s cold in January and hot in August.
I’m curious… How do you figure any percentage, right or wrong with forecasts like “Partly sunny with chance of clouds?”
I mean if they say 20% chance of rain, is it wrong if it doesn’t rain? Do I get a check block for saying 50% chance of sunshine and it doesn’t show? 86.8% horse-shit if you ask me.
I think most of us can look at a pressure map or a satellite image and take a good guess of what’s to come the next day.
Also, I think you’re right, they don’t need another supercomputer, they need experience and they need to quit using the “Indian rock” technique to predict weather.
I really never understood weather fanatics. Refreshing weather.com every other second to see what the barometer is going to do next.
Wasting $27 mil for marginal return… sounds like something a government would do. Why not spend that money to improve metro air quality?
Based on my purely anecdotal evidence, the quality of weather forecasting in Korea has gone downhill big time since the late 90s. When I first arrived here I was amazed by the accuracy of the weather forecasts (almost to the hour of when it would start/stop raining). But these days I can do a better job of predicting the weather with my bag of chicken bones and pigeon blood!
I really never understood weather fanatics.
Having survived a flash flood once on a camping trip in upstate NY and sat through a hail and lightening storm in the woods on a different occasion, I think I understand why one might be a “weather fanatic”.
Correction: Last spring, not ‘Last year’.
In any case. Lost Nomad has posted that the Korean government has finally admitted that a dust storm in China has been dropping dust over Korea…since Monday (I first spotted the dust nearly 2 weeks ago over the horizon and my allergies have been acting up for at least 14 days). What’s with that? Waiting for the dust to be so severe that they can’t deny it’s presence?