UW News

November 5, 1997

The world’s most detailed weather system gives forecasters a “dramatic” view of local conditions

First it was the Chinese, then the Egyptians who more than 3,000 years ago began studying and predicting the weather. Then in the 16th and 17th centuries meteorology became a science with the invention of instruments to measure the elements. Now a supercomputer is ushering in a new era of high-precision local weather forecasting.

The University of Washington has just switched on the latest version of its MM5 weather forecasting system, the first in the world to diagnose and forecast local weather on a scale of a few thousand yards. Already, says Clifford Mass, UW atmospheric sciences professor who has spearheaded the system, the MM5’s results are proving to be “dramatic.”

In fact the system has improved the ability to diagnose and forecast local weather so greatly that the National Weather Service in Seattle now has a computer work station dedicated to the MM5 output.

This is the latest development in a forecasting sea change that began last year when Mass helped organize a consortum of eight local, state and federal agencies, including the UW and the weather service, to purchase a powerful Sun Microsystems UltraSparc multiprocessor. Using advanced weather forecasting software, Mass and UW research meteorologist Mark Albright, were able to begin forecasting the weather in exquisite detail.

The software was able to use weather service data to simulate the evolution of the atmosphere over the Northwest by solving a series of complex equations every 12 kilometers (7.5 miles). Such a computer simulation pictures the atmosphere in terms of a three-dimensional grid, with each grid point representing a location where a weather calculation is made. The closer the grid points, the more detailed the weather forecast.

But in order to reduce the grid points even further, more computing power was essential.

Recently Sun Microsystems donated new processor chips to the project, speeding up the MM5 system 1 1/2 times, and allowing calculations every four kilometers (2.4 miles). This is by far the most detailed view of Northwest weather ever, Mass says.

The souped-up MM5’s four-kilometer portrait covers an area stretching from the Columbia River on the coast of Washington to the southern part of Vancouver Island, over the crest of the Cascade Mountains to the eastern slopes, then down to the Columbia Gorge on the Oregon border. In addition there are 12-kilometer and 36-kilometer (22.4 miles) grids over much larger areas. The forecasts are issued twice a day for the following 48 hours.

For the first time, says Mass, the computer is able to picture the rises and slopes of the region’s topography so accurately that it “sees” the effects of Mount Rainier, including the rain showers on one side of the mountain and the rainshadow on the other. It is also able to calculate wind directions with considerable accuracy.

In time, the system might also include forecasting the amount of water available to a region. That’s because Kenneth Westrick, a UW graduate student, is using the MM5 data to forecast river and stream flow for the Snoqualmie Basin, an important 602-square- mile watershed stretching west from the Cascades. Westrick is taking different measurements from the MM5 forecasting system, from winds to precipitation, and coupling them to an advanced hydrological computer model created by Dennis Lettenmaier, a UW civil engineering professor, and his students.

Forecasters have lacked precipitation data for the region because of the paucity of rain gauges and other observations in the mountains. But by using the MM5 meteorological data, the effects of climate and land-use change can be explored.

Another important use of the MM5 system, possibly within a year, will be its coupling to regional air-quality monitoring systems to forecast levels of ozone and pollutants in the atmosphere.

Mass does not see the MM5 narrowing its focus beyond four kilometers in the near future. Because of the absence of accurate meterological data from the North Pacific there is debate about the value of high-resolution forecasting systems. “Some say we shouldn’t go to high resolution but instead to an ensemble of lower-resolution forecasts,” says Mass.

If, for example, northwest winds from the Pacific are forecast as coming from the southwest, he says, “all the resolution in the world is not going to help you.”

### Contact Mass at (206) 685-0910, or cliff@atmos.washington.edu

<!—at end of each paragraph insert