SINCE 1983, scientists have been bitterly divided over whether a nuclear war is likely to result in a catastrophic global chilling. But the five scientists who introduced the term ''nuclear winter'' now acknowledge that they overestimated its severity, and their concession appears to have moderated the longstanding debate. 
Scientists say the issues involved are as pertinent to human survival as ever, despite the new friendliness of Soviet-American relations. The strategic nuclear arsensals of both nations remain intact, they note, and could come into play if the current peaceful climate gives way to war. 
The techniques developed to predict the effects of nuclear war on climate are also applicable to other climatic predictions, including the possibility that increased carbon dioxide in the air is leading to global warming, theorists say. The nuclear winter scenario is also closely related to the theory that dinosaurs became extinct when a giant meteor hit the earth and threw up a global dust cloud that caused catastrophic cooling. 
The views of atmospheric scientists studying the nuclear winter theory still vary widely, although most of those interviewed said they believe a nuclear war could have some effect on climate. But most discounted the extreme view that global chilling of the atmosphere would be severe enough to be described as ''winter.'' Scientists specializing in such studies also generally reject the suggestion that a ''nuclear winter,'' in itself, could bring about the extinction of the human race. Even Dr. Richard P. Turco, the physicist who coined the phrase ''nuclear winter,'' discounts the idea. 
Dr. Turco, a professor of atmospheric science at the University of California at Los Angeles, said in an interview that he had never believed that nuclear winter alone could wipe out humanity. ''That was a speculation of others, including Carl Sagan,'' he said. ''My personal opinion is that the human race wouldn't become extinct, but civilization as we know it certainly would.'' 
Dr. Sagan, a professor of astrophysics at Cornell University, was one of the scientists who collaborated with Dr. Turco in the article that ignited the nuclear winter dispute. The article, ''Nuclear Winter: Global Consequences of Multiple Nuclear Explosions,'' was published by the journal Science in 1983, and spawned a host of movies, plays and books predicated on the nuclear winter hypothesis. 
The other authors of the article were Dr. Owen Brian Toon and Dr. James B. Pollack, both of the National Aeronautics and Space Administration's Ames Research Center, and Dr. Thomas P. Ackerman of Pennsylvania State University. Their paper became so famous and so frequently cited that other scientists have since referred to it by an acronym of the contributors' initials: TTAPS, pronounced ''tee-taps.'' 
That paper was not the first suggestion that global cooling might follow a nuclear exchange. A 1982 article by Dr. Paul J. Crutzen of the Max Planck Institute in West Germany and Dr. John W. Birks of the University of Colorado proposed the possibility of such an effect. 
In a new paper published in Science, ''Climate and Smoke: An Appraisal of Nuclear Winter,'' the five TTAPS scientists review research conducted during the five years after their first joint paper. Drop in Temperatures While asserting that their general conclusions have been sustained, they say that a full-scale nuclear exchange in midsummer could reduce temperatures by an average of only 10 to 20 degrees centigrade (18 to 36 degrees Fahrenheit) in northern mid-latitudes. Compared with the reduction of 15 to 25 degrees centigrade (27 to 45 degrees Fahrenheit) predicted by their first paper, this chilling would be relatively mild, and in the view of some scientists, it deprives the phrase ''nuclear winter'' of realistic meaning. 
Dr. Turco said his nuclear winter forecast had changed somewhat because he and his colleagues had been able to reduce the uncertainty inherent in some of the climatic effects involved. New experimental data and analyses from other groups also helped to refine his predictions, he said. 
Dr. Stephen H. Schneider of the National Center for Atmospheric Research in Boulder, Colo., a long-standing critic of the extreme nuclear winter hypothesis, believes that a cooling of 10 to 20 degrees centigrade (18 to 36 degrees Fahrenheit) would not constitute the arrival of ''winter.'' 
''I would call it nuclear fall, not winter,'' Dr. Schneider said in an interview. ''But in any case, the TTAPS numbers have now more or less converged with ours, so I don't have a major problem with them anymore.'' 
In their latest paper, Dr. Turco and his associates say they have summarized and synthesized important experimental evidence and mathematical predictions made by other groups related to nuclear winter, thereby reducing the uncertainties inherent in their theory. 
''Essentially,'' Dr. Turco said, ''what we say is that the basic physics we proposed turned out to be correct, although the magnitude of the effects has been moderated somewhat.'' Blocking of Sunlight The theory underlying nuclear winter is that if the Soviet Union and the United States were to wage an unlimited nuclear war, much of the resulting dust and smoke from fires, especially those of burning cities, would be spewed into the upper atmosphere, where it might remain for weeks or months. This would block sunlight, resulting in a sudden drop in atmospheric temperature. 
The latest paper suggests some additional atmospheric effects of an all-out nuclear war, notably a severe depletion of the ozone layer in the Northern Hemisphere, which protects human beings from dangerous solar ultraviolet radiation. 
But although most critics of the nuclear winter theory have expressed only muted disagreement with the latest TTAPS paper, major discrepancies remain between its estimates and those of some other leading investigators. 
An important factor in such estimates, all agree, is the quantity of combustible material that would contribute to the global pall of smoke. Based on estimates by various research groups, Dr. Turco assumes the total mass of material burned, including wood, plastics, petroleum and vegetation, would be 5,075 ''teragrams'' (trillions of grams), or about 6.8 billion tons. 
But another leading investigator, Dr. Richard D. Small, a thermal science expert at Pacific-Sierra Research Corporation, a Los Angeles research organization, says he disagrees strongly with this estimate, which he believes is much too high. 
Dr. Small estimates that a maximum of 1,475 teragrams of material would be burned in the United States, provided all the weapons in the Soviet arsenal were successfully launched and detonated, and that all combustible material was actually ignited. Comparable figures for burned material in Europe and the Soviet Union would be proportionately less, ''because those regions simply have less combustible material in homes, businesses and industries,'' he said. 
''Our estimate is based on rigorous analysis of blueprints and other records of real homes and commercial and industrial structures,'' Dr. Small said. ''We add up every possible ingredient available for burning to estimate a weighted total.'' 
The greatest uncertainty in the article's assumptions, he said, is in the amount of smoke that would be injected into the atmosphere, remaining aloft long enough to reduce global temperatures. 
Other uncertainties include the amount of smoke that would be removed from the atmosphere by rain, and the height to which smoke would be lifted by fires ignited in a nuclear exchange. These and other factors could radically change the atmospheric effects. Biggest Uncertainty ''Perhaps the largest uncertainty is timing,'' Dr. Schneider said. A nuclear exchange in late spring or summer might have a significant effect on temperature, while a war in late fall or winter would have no appreciable effect, because sunlight is already reduced and temperatures are already low, he said. The growing season, moreover, would be over, and therefore unaffected by cooling. 
Dr. George Rathjens, a professor of political science at the Massachusetts Institute of Technology, charged in an interview that ''all the hype about a lot of freezing following a nuclear exchange is hyperbole.'' He said his own calculations, using the latest of Dr. Small's numerical smoke estimates in a standard mathematical model, result in a temperature drop of only about 9 degrees centigrade (16 degrees Fahrenheit) after a full-scale nuclear war in the Northern Hemisphere in midsummer. 
''I remain unconvinced that the TTAPS authors have a robust basis for their conclusions,'' he said, ''although in their latest paper they have made a fair number of concessions consistent with the work of others.'' 
Russell Seitz, whose critical analysis of the nuclear winter hypothesis attracted wide attention while he was a visiting scholar at Harvard University, is also unconvinced by the latest TTAPS paper. He said it is ''blatant semantic aggression'' to describe the available scientific data as pointing to a nuclear winter. 
In reply to Dr. Rathjens and Mr. Seitz, Dr. Turco characterized them as ''nonspecialists with political axes to grind and who have no real technical background. 
Dr. Turco said he and Dr. Sagan are about to publish a book, ''Nuclear Winter and an End to the Arms Race,'' in which they advocate reducing Soviet and American arsenals of nuclear warheads to a few hundred each. Link to Dinosaur Theory The scientific issues involved in the nuclear winter debate stem directly from a parallel debate concerning the reasons dinosaurs became extinct. In 1979, the late Dr. Luis Alvarez of the Lawrence Berkeley Laboratory in California; his son, Dr. Walter Alvarez of the University of California at Berkeley, and their colleagues proposed a new explanation for the extinction of dinosaurs and many other forms of life 65 million years ago at the close of the Cretaceous period. 
They theorized that the earth was struck by an asteroid or comet, which threw up an immense dust cloud that quickly blanketed the planet. They believe the cloud blocked sunlight, chilled the atmosphere and interfered with the photosynthetic process on which plants depend. The result, they hypothesized, was a mass extinction of the dinosaurs. That theory led directly to the nuclear winter hypothesis. The impact hypothesis has been challenged by paleontologists who believe that the extinction of dinosaurs had more complicated causes. Some paleontologists believe that if changing climate was a reason, it was probably caused by dust and gases from a continent-sized volcanic eruption in what is now India. Others believe that falling sea levels, disease and other factors caused the extinctions. 
Now that the nuclear winter debate appears to center more on technical issues than personalities, participants hope it will cool. ''It's nice to see these guys acting like scientists again,'' a researcher said. 
Nuclear Winter: Debated Points 
The dispute over whether a nuclear war would cause a ''nuclear winter,'' fatal to human life on earth, hinges on these variables: 
CONFLAGRATION 
How much flammable material is available to burn? Would fires be in urban areas, with presumably more combustible material? 
ATMOSPHERIC DYNAMICS 
Would dust and smoke injected into the upper atmosphere remain there to reduce global temperatures, or would much of it precipitate as ''black rain?'' 
TIMING 
Would the fires be in spring or summer, or in fall or winter, when temperatures are already low? 
DARKNESS 
How much light-blocking effect would the soot from the fires have? 
Photos: Dr. Richard P. Turco (NYT/Marty Katz); Dr. Stephen H. Schneider (pg. C8)