A limited nuclear attack on the United States would trigger vast hurricanes of fire and cause millions of more deaths than standard government studies estimate, according to scientific research released yesterday.

Although the devastating effects a nuclear war would have on the Earth's climate have gotten considerable attention and been named "nuclear winter," much less attention has been paid to the medical implications of nuclear war. However, reports made public at a weekend symposium of the National Academy of Science's Institute of Medicine conveyed a uniform warning of destruction and suffering on a far wider scale than foreseen by government experts, even if the Soviet Union unleashed only 1 percent of its nuclear arsenal.

Such an attack, consisting of 100 one-megaton bombs on or near the nation's 100 largest urban areas, could result in the immediate deaths of 36 million to 56 million people, Theodore A. Postol of Stanford University's Center for International Security and Arms Control concluded in a 50-page study.

By contrast, he said, standard government estimating procedures for predicting the consequences of nuclear attacks, which are based on blast effects alone, would project 14 million to 15 million dead.

The enormous increase in projected fatalities, Postol said, comes from a "very large expected range of superfires" and the highly toxic environment within the burning region. The huge fires, he said, "could kill everyone within a six- to eight-mile radius of ground zero."

"Winds at the ground could reach hurricane force," Postol said, envisioning a mammoth replay of the firestorms of World War II, "and air temperatures within the zone of fire could exceed that of boiling water."

Three Princeton University researchers -- Frank von Hippel, Barbara Levi and William Daugherty -- added in a separate study that tens of millions of people would be killed and injured regardless of whether the weapons were aimed at civilian or military targets.

One megaton bombs exploded over the 100 most important U.S. military-industrial targets, they estimated, would cause 38 million deaths and some 15 million injuries because most of those targets are in major urban areas.

Casualties would be lower in an attack on 99 key strategic nuclear targets, but this could still kill 19 million and injure 8 million others, they concluded. Chicago's O'Hare International Airport, for instance, is on this list because it houses tanker planes for refueling strategic nuclear bombers. Staten Island, N.Y., and Long Beach, Calif., are targets because they will soon be home bases for battleships equipped with long-range, nuclear-armed cruise missiles. And Washington, D.C., qualifies thanks to the Pentagon and the White House.

In short, said Princeton physicist Von Hippel, the distinction nuclear strategists often make between attacks on population centers, industrial targets and military targets is not as significant as it has sometimes seemed. It "may be comforting to the targeteers," he said, but "it should not be comforting to the rest of us."

"Our finding of tens of millions of deaths from just 100 warheads," the Princeton study said, "also shows that any strategic defense system would have to reduce an attack to well below this level before the results would be less than catastrophic."

Other studies presented at the symposium, funded by the Carnegie Corp. of New York and the National Research Council, painted a bleak picture for survivors. The immune systems of many would be severely damaged, much as if they had AIDS (acquired immune deficiency syndrome). The weakened immune systems would open the door for prolonged epidemics of diseases such as tuberculosis, leprosy, pneumonia, Legionnaire's disease and probably cancer, according to Brown University researchers.

All of the major injurious effects associated with nuclear warfare -- ionizing radiation, ultraviolet radiation, trauma, burns, depression and stress -- attack the same component of the immune system (the T-cell) as AIDS does, said Dr. David S. Greer, dean of medicine at Brown University, and coauthor Lawrence S. Rifkin. "A striking similarity exists between AIDS and the anticipated immuno-suppressed condition of survivors of a nuclear war," they reported.

The nation's medical resources, meanwhile, would be overwhelmed. According to Dr. Herbert L. Abrams of the Stanford and Harvard departments of radiology, a one-megaton explosion over Detroit would produce injuries requiring 42,000 hospital beds in burn treatment centers. Only 1,333 such beds exist in the United States; 41 are in Michigan.

The survivors in Detroit would also need 134,000 beds in intensive-care units, according to Abrams. There are 61,000 nationwide. There would be 57,286 whole blood units available in the country; Detroit would need 1,280,000.

A massive (6,559-megaton) nuclear attack on the entire nation, Abrams added, would kill most physicians, leaving 48,000 doctors to treat the 32 million people that the Federal Emergency Management Agency has estimated would be injured, Abrams said. There would be only one surgeon for every 410 patients with serious blast injuries and one plastic surgeon for every 15,364 with severe burns, he added.

The standard way government agencies estimate casualties of a nuclear war is to extrapolate from the blast effects of the World War II attack on Hiroshima. Modern nuclear weapons are not only much larger, Postol said, but they can produce much greater fires.

A one-megaton bomb, for instance, has the explosive power of 1 million tons of TNT, capable of producing temperatures of about 100 million degrees at its center -- four to five times the temperature at the center of the sun. A one-megaton burst over Baltimore, Postol said, would be seen in Washington "as a ball of fire many times brighter than the noonday sun."

"This intense source of heat and light is capable of setting many simultaneous fires over vast areas of surrounding terrain," he said. Once started, he said, the fires could heat large volumes of air near the Earth's surface. And as that hot air rose, cool air from regions outside the vast burning area would rush in to replace it, producing ground winds of hurricane force and boiling temperatures.

"The ferocious hurricane of fire would also be accompanied by the release of large amounts of potentially lethal toxic smoke and combustion gases, creating an environment of extreme heat, high winds and toxic agents in the target areas," the study added.

Postol said the climatic changes that could result from the smoke has drawn considerable attention, but little heed has been paid to what would happen to the people caught in the fires. He estimated that one-megaton or "near-megaton" air bursts "in or near urban areas" would produce two to four times as many deaths as projected by the government. By the same token, there would be fewer injuries.

There is uncertainty about the potential radius and size of the huge fires, but Postol said the government's death projections clearly are seriously understated. He pointed out that saturation incendiary bombing at Hamburg, Germany, in World War II produced one firestorm that covered five to six square miles and had a power of 1 million or 2 million megawatts, killing 50,000 to 60,000 people.

"By comparison," he said, "a mass fire initiated by a one-megaton airburst over a typical urban area in the United States could involve areas considerably larger than 100 square miles and have peak intensities of perhaps 15 to 50 million megawatts."

The Princeton researchers adopted the new "conflagration model" in making estimates on their computers, said to be the first outside of government capable of such calculations.

In addition to a one-megaton attack on 100 varying targets, they studied what they said was a commonly projected "major attack" of 3,000 warheads aimed at the nation's missile silos, missile launch-control centers and key strategic nuclear installations, a total of 1,215 targets. This, they estimated, would kill 16 million to 35 million people and seriously injure 9 million to 41 million others.

The aftermath of such attacks, Greer said, would include "a very large increase in AIDS in the surviving population."