Daily Weather Forecasters Rise to the Task With Balloon Launches

Today’s San Diego weather forecast for partly cloudy, cool weather with a slight chance of showers began to take shape 10 days ago.

To be exact, it began Oct. 2 at precisely 4 p.m., Pacific Daylight Time, with the simultaneous release around the world of 732 weather balloons.

From San Diego to Bermuda to Madrid to Karachi to Macao and hundreds of other points in between, the specially equipped, synthetic rubber balloons sent a constant stream of radio messages on temperature, air pressure, water vapor and upper-air winds to ground stations during roughly 90-minute flights that soared to nearly 100,000 feet.

The radioed information was fed by high-speed telecommunication lines from various nations and island dependencies into regional weather centers in Tokyo, Nairobi, Melbourne, Moscow, and Bracknell, England, and then into the main U.S. National Meteorological Center outside Washington.

Scientists there entered the raw figures into highly sophisticated computer weather models developed by center researchers and crunched them through one of the world’s fastest supercomputers. The result was 10-day forecasts for the Northern Hemisphere, for selected overseas nations, and for the United States.

That forecast showed that upper-air wind patterns--the critical element in long-range forecasting--across Asia and the Pacific Ocean were going to continue the unusually cool weather for the West Coast, including San Diego and Southern California.

As the hydrogen-filled gray balloons continued to be launched daily at midnight Greenwich Mean Time (GMT)--with a slightly smaller number at noon GMT--weather personnel refined their predictions, with 120-, 108-, 72-, 48- and 24-hour forecasts disseminated nationwide.

Like his colleagues throughout the United States, Wilbur Shigehara, chief meteorologist in San Diego, reconciled the computer model printouts with satellite data and other information to provide a local forecast for today.

“Upper-air data is the backbone, the fundamental basis for most forecasts that are made not only over the United States but worldwide,” said Harlan Saylor, deputy director for the National Weather Service forecasting center in Silver Spring, Md. In addition, aviation worldwide could not operate far-flung networks of flights without the data.

The information is obtained through the worldwide weather balloon network coordinated by the World Meteorological Organization (WMO), a United Nations agency based in Geneva.

San Diego is among three California sites--the others are Oakland and Vandenberg Air Force Base--and 94 U.S. locations where the balloons are launched twice daily. The balloons--those from San Diego are sent aloft from Montgomery Field at 4 a.m. and 4 p.m. (noon and midnight GMT)--contain a lightweight box with several sensors to measure the various weather data and a transmitter to relay it back by radio. A radio dish at the airport tracks the balloon to determine wind speed and direction.

“It’s truly an international program,” said James Neilon, director of the International Activities Division of the National Weather Service. “We like to say that the atmosphere knows no boundaries.”

And, as Shigehara said, only half in jest, “The flapping of a butterfly’s wings over the Sahara Desert can have an impact on San Diego’s weather.”

The information is shared instantaneously by the 159 countries that belong to the WMO.

“We literally get a snapshot of the world’s upper-air pattern at a single time,” Neilon said.

“It’s part of what is called the World Weather Watch.” Regional forecasts for Europe, Asia and the Southern Hemisphere are made by several national centers according to geographical location. Those centers include ones in the Soviet Union, West Germany, Japan and Australia.

Although there are more than 700 balloon stations, Neilon said that researchers would like more. The balloon releases have traditionally been limited to land areas, restricting information from the vast areas of the earth covered by oceans. The ideal network needed to obtain a complete picture of atmospheric conditions calls for balloon release points every 300 miles or so around the globe. But whereas there are almost 100 stations in the continental United States, an area of the southwest Pacific of similar size has only 12 release points.

The data returned by the balloons paints what researchers call an “initial state,” the state of atmospheric conditions, which is then analyzed by the computer models to turn out the long-range forecasts. In essence, the computer takes the information measured by all the balloons and runs out, based on the models, what will happen according to the laws of physics and the general west-to-east movement of wind patterns.

“We’ve really come a long way in the evolution of developing weather systems over a period of 10 days or so,” Saylor of the National Meteorological Center said. However, important limitations still do exist. “We can’t specify the initial state with complete accuracy because in some areas of the world, we don’t have sufficient numbers of observations,” he said. “And the models themselves still lack some of the things we need, such as how to (put into mathematical form) the cooling that results from rain, etc.”

The 1-to-10-day analyses from the Maryland weather center come in throughout the day to National Weather Service offices nationwide, in the form of maps. The maps include polar projections of winds around the world in the northern atmosphere and for the Western Hemisphere, and detailed wind and temperature projections for the U.S. mainland.

“What does it all mean? That’s the bottom line,” said San Diego’s Shigehara. “Like last week, I could look at the readouts and see high-level winds off Oregon in excess of 100 miles an hour coming in from Canada and know that they would be bringing cold air into our area.

“Beyond two days’ forecasting, you have to know what is happening with the upper winds 5,000 to 10,000 miles away. The upper-air pattern here is related to winds over Japan, and Asia, and even Europe.”

A map from the WMO for northwest Europe shows that wind data from every point on the compass is required to put together a forecast with any accuracy beyond two days.

Shigehara has a 90% confidence factor in the model’s forecasts out to three to five days. But he acknowledges the 10% of the time when the wind data misleads the models or sudden upper-air changes fail to be noted by balloon measurements.

“Last week, on Friday, there were a couple of weather fronts from the Bering Sea that were not quite detected by the model and therefore, while we advertised sunny and warm based on computer analysis, the weather stayed cool and cloudy for the weekend,” Shigehara said.

Such problems occur when weather changes are taking place in the geographic gaps between balloon measurements. Sometimes a key station will not report, forcing the models to estimate based on the remaining stations.

On Wednesday, Shigehara received updated 72-hour forecasts that showed a high-pressure system developing over the Pacific Northwest to bring drier, warmer air into Southern California this weekend. But it clashed with the previous longe-range predictions that he had been receiving, which called for cooler temperatures and clouds.

On Thursday, the newer computer forecasts reverted back to the cool and cloudy predictions.

“It’s a classic case of what we call yo-yo forecasting,” Shigehara said. “One of the runs, for whatever reason, might have missed some key upper-air wind data and the model got thrown off.

“But most of the time, we get it right, from balloon to the models and into the computer and into our forecast.”