Letter: 'Nuclear Winter' Hysterics

In reference to your article on "Nuclear Winter" (The Scientist, May 1, 1989, page 1) and the commentaries that continue to appear in your Letters section: Most discussants of the current spate of computer-based atmospheric catastrophes assume that computer models and outputs are more reliable than the reality of observation. Although large-fire research in a war and wildland context has been going on for many years, a popularized study of the supposed effects of large fires, directed toward support of the antinuclear weapons movement, was written and first published in Europe as "The atmosphere after a nuclear war: Twilight at noon," by Crutzen and Birks (Ambio, 1982, pages 115-25). The title did not have much glamour, but the concept did. The idea was reworked, a new computer model was developed, and the results were published in the U.S. in 1983 as "Nuclear winter: Global consequences of multiple nuclear explosions," by R. Turco, O. Toon, T. Ackerman, J. Pollack, and C. Sagan (Science, 1983, pages 1293-1302). There was a list of to-be-published references that would founder a whale. These initial papers prompted a deluge of related material, reviews by learned authorities, congressional hearings, and excited meetings. Most, with few exceptions, accepted the computer model results without question. The complexity of the "Nuclear Winter" computer programs precludes an unbiased rigorous examination of the computations without essentially repeating the development of the computer model. But this is unnecessary. There have been large fires, both experimental and natural, in the past, and it is possible to rationally compare computer output with reality. Based on more than 20 years of observation of wildfires and research experience in large experimental fires, I have identified serious errors in the "Nuclear Winter" computer model and its output. The Project Flambeau experiments (1960-66) demonstrated that conventional atmospheric convection models were not applicable to large fires. Even now, a satisfactory theory of how cities and wildlands burn has not been developed. Unfortunately, neither the computer modelers nor the commentators understood this problem, and a simple, large heat source suitable for the computer model was used instead of a realistic fire simulation. This inevitably led to erroneous results. The most fatal consequence was the basic "Nuclear Winter" requirement that smoke from nuclear weapon-initiated fires somehow moves from the troposphere into the stratosphere. It is clear from observations that smoke absorption of solar energy has heated the atmosphere downwind from large fires. The modelers assumed that the smoke from the large fires moved into the stratosphere. The resultant heating, together with obscuration of the sun and various dynamical processes, produced a simulated drastic cooling at ground levels. But there have been no observations of smoke in the stratosphere above even the largest natural fires. No stratospheric smoke, no "Nuclear Winter." When this became apparent, the modelers struggled mightily in theoretical efforts to place smoke in the stratosphere. Their approach has been to use various atmospheric cloud convection computer models to demonstrate that superconvection columns form above large nuclear burst-initiated fires. These theoretical models form convection columns tens of kilometers in diameter that reach well into the stratosphere and theoretically carry smoke to this region. This effect has never been observed, and for good reasons. First and simplest - it has been shown that burnable, contiguous cities and forests do not exist. At best (or worse), simultaneous ignition of a city or forest will produce spotty fire "islands" no larger than about one to two square kilometers. There is observational evidence that the fires will be even smaller. And experimentally, convection columns above fires with even the most modest separation (a few tens of meters) simply do not merge until they near the top of their maximum height of rise. It follows that the master convection column is a computer model artifact. It remains to be demonstrated by observation that any smoke from large fires has ever been deposited in the stratosphere. No stratospheric smoke, no "Nuclear Winter." Second, computer models of large fires have failed to reproduce the observed winds inside a fire from the first one in 1967 to the most recent. The models (even the more recent three-dimensional ones) ignore the complexities of fuels burning in the atmospheric boundary layer where the basic convection process starts. They treat the fire as a simple, albeit large, heat source and ignore the air flow into the fuels that is necessary for combustion. This is clearly in error. Analysis of convection columns above large natural and experimental fires shows that much of the air in the convection column must pass through the flaming zone (otherwise the fire goes out for lack of oxygen), yet there is no provision for this inflow in any of the computer models. This is understandable. It is a very difficult problem. Third, observation of real, large fires shows that most dry convection columns reach only about five kilometers in height. Even the most intense known fires, such as the World War II firestorms in Japanese and German cities and the Tillamook fire in Oregon, produced columns reaching only seven to eight kilometers in altitude. This is far from stratospheric levels. Occasionally, the rising smoke column will be sufficiently moist for a cumulus cloud to form. Recently, moist convection has been incorporated into the smoke column models, to provide sufficient energy for the monster computer model cloud to reach into the stratosphere. Condensation and rain will remove smoke from the convection column. The "Black Rain" following the atom bombing of Nagasaki illustrates this effect. Again, little or no stratospheric smoke. And no "Nuclear Winter." Computers are at best but an extension of man's intellect, and when used blindly will result in bad results and decisions. Computer results must face the ultimate test of comparison with reality, and the observing scientist, not politically motivated hysterics, must be the tester. THOMAS Y. PALMER Large Fire Consultant Southwest Environmental Technology Labs Fallbrook, Calif.

Letter: 'Nuclear Winter' Hysterics

Interested in reading more?

Become a Member of