A monster-sized, weather-predicting computer could soon save lives and property.
"We can forecast that there might be heavy rain on one side of the mountain and not so heavy on the other side to help forecast things like the recent mud slides near Hope," said Professor Roland Stull, director of the new Geophysical Disaster Computational Fluid Dynamics Centre at the University of British Columbia.
Stull and his associates have created a new super-computer which is so powerful and so speedy that it can solve equations used to predict the weather fast enough to actually warn people of danger.
The mega-computer, made up of 264 networked processor computers, can do 170 billion computations per second. It is the fourth fastest computer in Canada and is ranked No.255 in the world.
"Meteorologists, when they forecast the weather, use all kinds of tools like satellites, radar and such, but all of those other tools only tell them the current weather," explained Stull.
"The only tool meteorologists have to tell them about future weather is computerized codes that solve the rules of how the atmosphere works.
"Well these are complicated equations that take a long time to solve. But we don’t have the luxury of taking a long time to solve them because we have to solve them before the weather actually happens.
"So we have the need for speed."
Stull has divided the whole of B.C. into 3.3 square-km grids. The super computer, fondly known as "monster," can accurately predict weather in each grid.
Stull hopes to reduce the grid size to just 1 km in future.
"We can make a different forecast for Stanley Park, North Vancouver, Victoria, or even Whistler," said Stull.
Having this type of high resolution information allows Stull and his associates to predict specific weather-related phenomena.
For example they could tell whether an avalanche is likely to occur on one mountain as opposed to offering a general avalanche warning for an entire region.
"Until now avalanche forecast was done by taking the observed weather from wherever the weather station was, which may be many kilometres away from the avalanche area" said Stull
"Instead, we can take a weather forecast, for two days from now, right over that mountain, and forecast how much snow will fall on that avalanche track.
"Now we have two days to run the avalanche model and see whether the avalanche will be triggered."
"Monster" can also provide life-saving information for forest firefighters. By predicting wind and weather changes it could guide firefighters to hot spots or offer enough warning to remove firefighters from areas likely to become dangerous due to changing weather.
"We could say ‘get these firefighters out from this region because the fire is going to explode’ and save lives," said Stull.
The computer is fed weather information every night from weather reporting agencies world-wide. It takes half a day to analyze the data.
The information is shared with a number of partners in both industry and government on an experimental basis at the moment.
But, said Stull, the hope is that in the future it could be used to warn B.C. Ferries of wind storms, forestry companies of blow down threats, ski areas of avalanche risk, and highways workers of landslide, rockslide or avalanche risk.
