Slashdot Mirror
Supercomputers Race to Predict Storms
pillageplunder writes "CNN has an interesting article on how different supercomputers from around the world are working to predict large storms tracks. The 3 days it takes now has been cut in half. Cool read."
Actually if you read the article you will realize that it only takes about an hour of number crunching, but that the three day storm path accuracy errors have been cut in half... and that 5-day forcast is getting much more accurate.
I guess we should read articles before submitting them...
D.O.U.O.S.V.A.V.V.M.
Yeah, the Earth Simulator http://www.es.jamstec.go.jp/ does try to do this, however it does a vast number of other things as well. The other systems focus on more specific incident (such as Ivan), thus more computing power is aim'd at a short-term problem.
Earth Simulator does - Atmosphere & Ocean Simulation, Solid Earth Simulation, Multiscale Simulation, and Advanced Precipitation Simulations. (And other cooperative projects).
I am
Nope, the Earth Simulator is to predict overall climatic change, not specific weather conditions.
According to New Scientist 28/08/2004, it's a little more to do with long-term climate change, rather than predicting if you need your umbrella tomorrow in Bristol... Earthquakes and the Earth's magnetic field are also modelled too apparently...
A snip at $430 million...
No Norm, those are your safety glasses; I'll wear my own thanks...
i) because most numerical weather codes are already written in Fortran. This means that people with the right scientific knowledge tend to be Fortran programmers, and makes porting a whole lot easier.
ii) Fortran compilers are the ones where the most work has gone into optimising the hardcore mathematical routines. Thus, the compiled code has traditionally been faster. This may no longer be true.
Athletic Scholarships to universities make as much sense as academic scholarships to sports teams.
Fortran supports a very large set of highly optimized intrinsics that have been perfected for numerical computations over many many years as well as vast libraries of parallel implementations.
I saw in another post regarding my parent post that part of the reason is the tweaks for the fortran compilier being released.
I'm just trying to understand why they aren't moving to a more object oriented method of design for weather modeling. So they can drop in objects that don't require the entire code base to be recompiled.
Fortran is still the dominant language for programming high performance code. I'd still rather use C, but it's not really that different. When you're trying to optimize a piece of software for a machine architecture you need to use a language that is pretty low-level. The closer to assembly you are, the greater chance you have to best exploit the functionality of the hardware. C++/Java are right out.
You can trademark "futurecast".
funny munging
The data is already being measured (google for metar) by our tax dollars. Before you do it again look there, unless you need more location data (for micro climates and such).
Degaussing scares the bad magnetism out of the monitor and fills it with good karma.
Landfall dynamics are a VERY active point of research in hurricanes right now. Land changes a lot of variables which we can normally take for granted in a hurricane over the water... the surface has different properties, elevation changes make the air behave differently, land doesn't evaporate near as much water vapor as the ocean, etc.
u idance/at lantic/store/early_AAL06_04090300.png
So, with land, you leave the realm of an initial value problem with relatively well-understood boundary conditions that you have with a storm over the ocean to a realm that has much-less-well-understood boundary conditions. The problem becomes much harder to close, much less solve. And with a system like the hurricane which REQUIRES good knowledge of the boundary (after all, the hurricane is fueled by latent heat release by condensation of water vapor which comes from the ocean), not knowing the boundary as well as you can makes prediction much much harder.
Charley's swerve was forecast by a good number of models, but NHC played the worse case scenario card a little too long by persisting on a landfall near Tampa Bay.
Frances' stop was due to a very irregular pattern, much like a saddle point. If you are pushed any direction, you get very different behavior. You can see that on the following model ensemble plot... there's a small cluster of 48 hour predictions that are slower than the others.
http://euler.atmos.colostate.edu/~vigh/g
Ivan's bounce off Jamaica is a seriously cool research topic, since Jamaica is a mountainous island. That big elevation change could make it more "visible" to the core of the storm (unlike the plains of Florida). This will be a serious research topic for decades to come. Many of the models did not handle it well (which isn't too surprising since Jamaica is a relatively small island and the models that are used frequently are global or near-global models). And some previous storms (Gilbert, 1988) didn't even notice Jamaica as they passed over, so experience is a split decision.
So, hopefully that sheds a little insight on this issue. Land is a BIG problem for track forecasting, and we're just starting to work out the kinks.
-Jellisky