Improving the predictability of weather forecasts

Those satellite weather pictures on the nightly news give the watcher, and the meteorologist, the Big Picture - atmospherically speaking. But they tell much less about the weather conditions the man in the street will experience.

A group of atmospheric scientists with the National Oceanic and Atmospheric Administration (NOAA) here is trying to change that. They are pioneering methods to bring the art and science of weather forecasting down to human scale. They hope even to provide the public with a nearly ''real time'' picture of local weather on a cable-TV channel.

This is one of several objectives of an ambitious program called the Prototype Regional Observing and Forecasting Service (PROFS). In this program federal weather researchers are employing recent advances in computer technology to forge the meteorological tools of the future. Their aim is to increase the quality and accuracy of weather forecasting, particularly of predicting such severe conditions as thunderstorms, tornadoes, and flooding.

Weather forecasting is particularly difficult because the atmosphere is ''explosive'' and weather is a ''nonlinear'' system, explains Robert Bunting, leader of the PROFS forecast techniques development group. ''In a nonlinear system, you can drop a pebble into a pool and get a tidal wave. Or drop a boulder and get a ripple,'' he explains. Similarly, a devastating storm can form with explosive suddenness out of apparently benign atmospheric conditions. Or, every indicator can be ripe for the formation of a tornado and nothing will happen.

The first step PROFS researchers and engineers have taken to help the weather forecaster cope with his subject's legendary unpredictability is to take all of a forecaster's presently used sources of information and integrate and automate them. Currently, forecasters look at satellite weather pictures and local radar images - such as those briefly seen on TV weather briefings - and combine them mentally. This is a difficult process because the images are different in scale and time. Also, these meteorological tools provide a tremendous amount of information, much of which forecasters cannot currently use.

''The human mind simply cannot completely comprehend satellite or radar pictures, let alone integrate them, so we are using a computer to assist the man ,'' explains Dr. Donald Beran, head of the PROFS project.

Mr. Bunting demonstrates the power of this approach at a PROFS terminal.

On the computer screen flashes a satellite picture of the US last summer. It has been retrieved from a data bank. With a few keystrokes, the angel's hair pattern of cloud is set in motion above the familiar outline of the states. Suddenly, ominous spots of red appear. One is in northeast Colorado. Bunting explains that the computer has been set to display with red highlighting certain conditions indicative of thunderstorm formation.

With a few more keystrokes the picture zooms down on the storm. It also makes an effortless transition from the satellite image to that of the local weather radar. Now the screen is filled with the color image of a summer thunderstorm, which appears as ragged, concentric rings of various colors. ''The different colors correspond to different amounts of rainfall,'' he explains. With the touch of another key, the storm begins to move across the screen and its shape changes.

To date, NOAA has spent $10 million on PROFS. One of its first users will be Murray Pautz, a hydrologist in the Denver office of the National Weather Service. The weather service office will get a terminal to test this spring. ''I'm really quite impressed with the progress they've made in the last year. We're really looking forward to getting the equipment,'' he says, adding, ''It will make Denver the most exciting weather service office to work in for the next few years.''

It's hard to say what effect this new technology will have on the quality of forecasts, Mr. Pautz says. But he believes it will improve them. NOAA intends to compare forecasts made with and without PROFS to determine definitively what effect these added capabilities have on the art of weathercasting, Dr. Beran says.

The initial area to look for benefits will be in severe storm warnings, Mr. Pautz believes. Also, he thinks the animation capabilities will help forecasters , particularly in compiling morning forecasts. It relieves the forecaster of a great deal of drudgery. ''Instead of spending all our time putting out fires, it should give us time to anticipate events,'' the National Weather Service hydrologist says.

So far, PROFS has concentrated mainly on improving the meteorologist's picture of what is happening. But the NOAA scientists are also attempting to build in automated predictive capability as well. Utilizing standard programs, the researchers are experimenting with two--to six-hour weather ''projections'' that will estimate the future course of storm tracks. They are also working on a program that will predict the amount of snowfall in different areas in the Denver region.

In the second phase of this program, the federal atmospheric scientists hope to put new sources of weather information at forecasters' fingertips. For years, weather researchers have been frustrated by how little of their efF rts are incorporated into operational forecasting. In part, PROFS was originally conceived as a way to smooth this translation of research into practice.

One essential aspect of this is a much denser network of weather stations. ''You can't catch a minnow with a whale's net,'' is how Mr. Bunting puts it. Today, forecasters are trying to construct their pictures of weather patterns from observing stations often hundreds of miles apart, so spotting severe storms , which typically are 10 miles across, is usually a matter of happenstance.

Enter the concept of the ''mesoscale'' net. This is a dense regional array of automated radar and ground weather stations. Currently, most forecasting is done on the ''synoptic'' or global scale. In many conditions, say ''scattered showers ,'' an accurate synoptic scale prediction can appear wrong at any given location. Hence the importance of a regional weather network that allows forecasters to better tailor forecasts for a specific locale and to aid in issuing more accurate storm warnings.

An Achilles' heel in the current weather system is the difficulty and expense in gathering vertical temperature profiles of the atmosphere. This is one of the most valuable of atmospheric measurements. Yet it is only taken twice a day at widely spaced locations by weather balloons. NOAA researchers are in the process of perfecting what they hope will be a more cost-effective way of gathering this valuable weather information. This is a device called a profiler. It sends a beam of laser light vertically into the air. By the way it is reflected back analysts can determine the air temperature at different altitudes. Totally automated, these profilers would transmit their measurements back to the main PROFS system where the data would be integrated with other sources of information about the atmosphere.

Another instrument the researchers are eager to incorporate into the system is Doppler radar. This is a radar that measures the motion of rain drops and other particles in the air. In this fashion, it can directly measure wind patterns: another crucial and often missing link in weather reconstructions. A major problem with this type of radar has been interpreting the tremendous amount of information it generates. The NOAA scientists believe the PROFS computers can overcome this hurdle.

''By 1983 we expect to have a full-blown demonstration of the advanced system ,'' Dr. Beran reports. Then, if all goes as planned, the PROFS equipment will gradually be introduced into National Weather Service offices around the nation.

The effect this will have on the accuracy of forecasts and weather warnings is a matter of speculation at this time. But one private forecaster who has tried PROFS says, ''It's like the difference between alchemy and chemistry!''