Electronic flood-prediction tool gains wider use in West

Poking out of the silent white snow of Deadman's Hill, a solitary antenna beams in an instant the depth, weight, and water content of the five-foot snowpack here at 10,000 feet in the Rockies.

Hydrologists can collect the same data by hand. But it takes two hours by car and another hour by bumpy snowmobile to get here from Denver. And they must tramp around in hip-deep drifts, sample the snowpack with a hollow aluminum rod, weigh it, measure its depth, and resample the snow at several points along a 100 -yard course.

In either case the data are essential in figuring how much water will run off the mountain ranges. In this pristine setting the snow looks harmless. But melted snow from a thousand such locations can mean flooding. With the automation at Deadman's Hill, part of the Snotel system of 500 electronic data-collection points in the West, snowpack data can be translated instantly into stream-flow forecasts.

Snotel, a US Soil Conservation Service project, has meant the difference between this month's forecasted flooding in Utah and southern Idaho, and last year's flooding in the same areas, which took hydrologists, farmers, and residents by surprise.

It's obvious from the snowy peaks in many areas of the west this year that there is more snow later in the year than usual. But that alone doesn't mean flooding will occur, explains Robert Clark, chief hydrologist of the National Weather Service.

Any number of variables can come into play, he says - temperature, precipitation, the amount of snow and its water content. All of those variables change daily. And while weather data can be collected nearly immediately, snowpack information traditionally has been days and weeks old. With Snotel, snowpack data can be correlated more closely with weather data so critical to how and when snow will melt, he explains.

''Snotel is so new (its data haven't) been incorporated in objective forecast. But now there's a realization in the hydrologic community of the value of this,'' says Bernie Shafer, data analysis leader of the Soil Conservation Service's Portland, Ore., water supply forecast staff.

Snotel, he says, offers the third dimension in the hydrologic picture scientists are trying to develop. Satellite photos reveal the extent of the snowpack, Snotel gives the snowpack's composition, and meteorologic detail helps determine how the snowpack will be affected.

''If we can put all this together we have an extremely powerful tool,'' he says. ''Predictions will be more accurate and forecasts will be more timely. Our best shot before was to do it (a forecast) once a month.''

Last year after the floods that took hydrologists by surprise, analysis of Snotel data data showed that ''it wasn't that the snowpack had a tremendous volume of water, but that the runoff was compressed into half the normal time,'' he says, because of rain and unseasonably high temperatures.

Scientists found that feeding Snotel data and other information into a computer model can give an improved forecast, even though Snotel data goes back only five years. (Because Snotel history is so short, long-term averages are harder to determine for computer modeling.) The computer models, as they are refined, will be used to give ''reasonable maximum and minimum stream flows,'' says Mr. Shafer.

Still, forecasting is different than preventing flooding. But it is the next best thing, scientists say. And there are 25,000 organizations who pay to get the Soil Conservation Service's forecasts.

Agriculture consumes 90 percent of the total water supply in the West and farmers rely heavily on forecasting for cropping patterns, says Shafer. For example, if farmers were advised of a low water supply they might plant fewer crops, or different kinds of plants, he says.

Similarly, municipal water utilities and reservoir operators depend on forecasting for water and electric (hydropower) supplies and flood control. Last year's unexpectedly heavy stream flows, says Dr. Clark of the National Weather Service, put twice the normal 6 million acre-feet of water into Lake Powell in Arizona when hydrologists had forecasted only 8 million to 10 million acre-feet.

''This year (using Snotel) we were able to anticipate months ago the course of things because we could update forecasts every day,'' he says.

Though use of Snotel is still relatively new, he says, ''five years from now we won't be able to do without it.''

This year's forecasts of flooding in Idaho, Utah, western Colorado, northern Nevada, and southeastern Oregon have allowed people to be prepared, Dr. Clark says. He says there has been time to raise levees, make thousands of sandbags, dig ditches around homes, and lay out plastic sheeting in anticipation of the floods.

He says estimates are that $23 billion in damage was averted in the United States last year by successful stream-flow forecasting. ''If you were able to move only the television sets out of 500 homes in a flooded community because of a flood warning, you've still saved a lot of money.''