In our state of Montana we would vote for him for anything (in ascending order) from dogcatcher to president of the United States to queen of the Helena Rodeo.
At this point in the story we need to take only a few steps backward in time to see the Mann Gulch connection appearing. Prior to his election to Congress, Mansfield had been a professor of history at the University of Montana, which starts to connect him both ways, to Japan and to Mann Gulch. At the University of Montana the future ambassador to Japan taught Far Eastern history. The connection with Mann Gulch starts appearing distinctly when we recall that the University of Montana is in Missoula, and thus the home town of Senator Mansfield was the headquarters of Region One of the United States Forest Service and the base of the Smokejumpers who were dropped into Mann Gulch.
The direct connection between Senator Mansfield and Mann Gulch must have been coinstantaneous with the fire. Why not? Surely the former boy who worked in the mines of Butte was at least as shaken as the rest of us by what happened to the boys working in Mann Gulch. They were two dangerous places for boys to work.
As to the Mann Gulch connection, the act had been almost as swift as the thought. The last victim to reach Helena before dying was dead by noon of August 6, 1949, and by October 14, little more than two months later, Mike Mansfield had rushed through Congress his amendment to the Federal Employees’ Compensation Act doubling the amount allowed to nondependent parents of children injured or killed while working for the federal government—from a pitiful two hundred to four hundred dollars. A rider attached to this amendment made it retroactive to include the Mann Gulch dead.
The part of this story that takes us from Mann Gulch to the three forest fire research laboratories ends this way:
1. The first of the three laboratories was built in Macon, Georgia, the senior senator’s home town.
2. The second laboratory was built the next year in Senator Mansfield’s home town. It is only twenty yards from the base of the Smokejumpers who flew to Mann Gulch.
3. The third was built in Riverside, California, the state that had suffered most in the last bad fire year.
So the forest fire research laboratories are where the forests and the politicians were, and it would be hard to kick at that.
MY MEETINGS WITH BRACKEBUSCH took some steam . out of my desire to find the mathematicians immediately. As I talked to him, I began to get some idea of how much more I would have to know in order to explain the Mann Gulch fire in light of the latest advances in fire science. When Laird and I were in the woods, I suppose we thought of ourselves as educated men, if there was ever an occasion to think on such a matter, but I at least knew that my education, starting with what I got from my father, had never included much math, so I began thinking that, before meeting the mathematicians, I had better retire to my cabin at Seeley Lake and do some homework. I had no trouble gathering a small pile of articles written by the two mathematicians or about them. Often it’s a lot easier to find out about important persons than to find them—especially it is not hard to find out where a man came from if his last job was on a project to determine whether an atomic-powered airplane would work. Such an experiment had been conducted near Idaho Falls, Idaho, on what significantly turned out to be the Lost River. When this experiment got lost or whatever an atomic experiment does when the government doesn’t want it around anymore, some high-powered young scientists were turned loose, and Jack Barrows, who had been one of Gisborne’s favorite students and was now the first director of the Northern Forest Fire Laboratory, moved fast and brought up five of them, one of whom was Richard C. Rothermel, who even as a student at the University of Washington had worked in aeronautical engineering. It took some seven years, however, for Rothermel to make the switch from making models of atomic-powered planes to making models of forest fires.
It took the young scientists another six or seven years of testing and correcting their new equipment, especially the wind tunnels, before they could be confident of the results of their experiments. One of the constant and crucial questions was whether the equipment really conducted controlled experiments, whether, when it reported changed results, the changes were solely the result of the one fire factor the scientists were studying or also of other factors they had not succeeded in eliminating. Almost as difficult was the problem of getting fire to burn under controlled conditions with enough consistency and continuity to allow the results to be measured accurately. So it was 1968 before Rothermel and his group were sure enough of what they were doing to move into the field of prediction and to assume responsibility for what Barrows had long wanted, an overhaul from top to bottom of the Fire Danger Rating System.
Even then there were problems, including, of course, the problem of gaining the acceptance of the old-timers, to whom forest fires are the ultimate reality and wind tunnels and computers are gadgets. Since old-time woodsmen change even slower than equipment, it took at least another seven years and well into the 1970s before mathematical models of forest fires and their predictions became sought after by agencies and businesses that live off the woods.
Administratively, Rothermel is now leader of the Fire Behavior Project at the Intermountain Fire Sciences Laboratory (as the Northern Forest Fire Laboratory has been renamed). Scientifically, his story is close to synonymous with the introduction and development of mathematical models of fires in the Forest Service. In 1981, in Washington, D.C., he received one of the highest honors granted by the Department of Agriculture, the Superior Service Honor Award. He was cited for “outstanding creativity in developing fire behavior prediction technology and training programs, enhancing the implementation of the Forest Service’s revised fire policy.” The “revised fire policy” that the developing “fire behavior technology” had made possible was a change from the policy of putting out all wildfires as soon as possible (the goal expressed in the slogan “ten o’clock fires”) to the increasingly prevailing theory, called “fire management,” of letting a selected number of fires burn themselves out. Although this new policy remained unnerving to some woodsmen, it has proved to have much practical value, provided it is wisely used, which has come to mean depending heavily upon the Fire Danger Rating System. A rough estimate of the financial benefit that might come from fighting only some wildfires might be guessed from the fact that in the mid-1970s the Forest Service’s annual expenditure had increased to three hundred million dollars and was still rising. To the economies brought about by the policy of letting some fires burn themselves out should be added a richness of ecological benefits, or the Indians long ago wouldn’t have set so many prairie fires in the autumn to enrich their pastures in the spring. To the value of fire “providing a suitable habitat for wildlife or forage for livestock” can be added the controlling of insect and plant disease. The Forest Service, moreover, has not been the sole beneficiary of letting some fires burn and even setting some. State forests and private timberlands also pay close attention to the Fire Danger Rating System and, of course, so do logging companies, especially in the autumn when they have to burn their piles of slash from the summer.
Frank Albini until 1985 was a physical scientist at the laboratory, which he joined in 1962 when one of its chief problems was to gain the confidence of the loggers. In addition to being a brilliant scientist, he turned out to have a quiet, persuasive literary style that helped to make him an effective, half-concealed salesman for the extended use of mathematical models in the woods. As a scientist, he has been from his doctoral days at California Institute of Technology a maker of mathematical models, whether for Hughes Aircraft, the Institute for Defense Analysis, or General Research Corporation. As a student, he specialized in plasma physics, and, although that field has highly specified aims and subject matter, it nevertheless reflects in its ultimate goals the Greek etymology of “plasma,” which has to do with form or mold. So a-modeling he has always been.
The term “making a mathematical model” shouldn’t slip by us so often without being stopped to i
dentify itself. “Making a model” for many of us suggests a bright boy using his Erector set to make a model of the Brooklyn Bridge as a structure of girders and then leaving it on the table for his mother to take down. Of course, that’s not completely wrong, except that a mathematical model of a fire is a structure of knowledge and, as Albini says, a “surrogate for reality”; its girders are quantitative generalizations about things that burn in the woods and in open country, therefore quantitative generalizations about different kinds of fuel and the influences on them of such powerful environmental factors as wind, slope, temperature, and humidity.
These girders of scientific knowledge are quantitative products of controlled observations of fire experiments and actual wildfires. The challenge is to pick the right analytical generalizations about things that will burn or contribute to their burning and fit them together in such a way that they will describe a fire that is predictable in its intensity, rate of spread, flame length, and other characteristics. Quantitative models of wildfire, then, have their practical as well as their aesthetic aspects. Making mathematical models of wildfire becomes a double pleasure, and Rothermel and Albini, who derive great pleasure from building these models, were placed in charge of refining the Fire Danger Rating System—one job is part of the other, and it is a good guess that the practical and aesthetic pleasures are not separate from each other.
Albini helps us in modeling a picture of modeling by pointing out that the “origin of this kind of approach to decision making and design is the ‘preliminary design’ technique used in aircraft manufacture. It is no accident that aeronautical engineers are often found in model-building jobs.” He then adds, “My undergraduate training was in aeronautical engineering.” And, as we remember, Rothermel’s was too.
Albini was once telling me about one of his projects, which had primarily to do with predicting the speed of missiles, both those missiles made by others which he had only seen and those he was recommending be built. His comment at the end was that “it’s a lot easier to predict the speed of a missile than that of a wildfire.” Generally, he said, “it’s easier to predict the behavior of objects made by man than natural objects.” Having lived long enough to absorb a considerable number of lumps and bumps from whatever hovers around outside under the name of “nature,” I said to him, “That shouldn’t have surprised you.”
“No,” he said, “it didn’t. Long ago a science teacher told me, ‘The universe, she is a bitch.’” Several times since, I have thought about this sentence. It’s probably right.
WHEN WE FIRST SAT DOWN at the long conference table in the lobby, I was puzzled about what I must have done long ago that had led me in retirement to a confrontation with two mathematicians and two wind tunnels. Some of the explanation had to be personal. As a boy I had to confront dangers of the forest that ever since have left me dreaming I am on a fire-line and the fire is about to jump the line and will if I wake up, so I try not to. My wife’s ashes, scattered on the mountain she named after herself, undoubtedly direct me back to the scene of the great tragic forest fire falling in her line of vision if she can still see me. It is probably less important that when I first saw the Mann Gulch fire it was still burning through rocks like snakes on fire. In my dream from which I cannot quite awake I still sometimes see a deer with all its hair burned off except what rims its eyelids. There is no use trying to eliminate all that is personal in order to be scientific. The long conference table at which the four of us sat was big enough to take in everything and long enough to seat eighteen or twenty. Perhaps the empty spaces had been reserved for the dead Smokejumpers and Gisborne. They belonged there but were never there, but they were never far away.
Although I have had few school courses in science, I have always tried hard to be accurate with facts. In my family we were expected to be, and, in addition, I found that being accurate with facts was a kind of game and I liked to play it. Later, when I came to know some great scientists, I found that to them science was a kind of game on a grand scale. The game Laird and I hoped to play with the mathematicians was to match our analysis of the fatal race between men and fire with their mathematical study of the same race. If nothing else, the results might tell us whether the Smokejumpers had much of a chance against this fire. Critics have always talked loosely about this or that tragedy being “inevitable,” but I seldom thought any of them were, and I also thought it would make a lot of difference to everyone involved if this tragedy was.
Mathematicians are very clear writers, as one should expect; their only prose weakness, also to be expected, is that they write for each other. So it turned out to be helpful to have figured out ahead of time what some of their main plays were going to be, because then it was not necessary to know, at least immediately, the meanings of all the words. As theoreticians, they start by finding it odd that, although men had been fighting fire long before they knew how to light one, they haven’t formed a theory of why it spreads. I thought this odd myself and oddly applicable to me, so I made an effort to learn about some of the first mathematicians to take up this problem. Evidently, W. R. Fons in 1946, basing his work on a theory that a spreading fire is in fact a series of ignitions, was the earliest to make a mathematical model of a fire. All the other mathematical modelers of fire whom I read also started by looking for a definition of fire spread never before given, and they ended with a definition which, when reduced to its main simplistic terms, says that a spreading fire is a series of little fires. Just so we wouldn’t glide by the center of this analysis with a few simplistics of mine, I asked Albini if he would write down an explanation of the process of fire spread for me and anyone who ever reads about the Mann Gulch fire:
As the fuel burns at a point just ignited, it releases the energy that the plant has gathered from the sun and stored up as plant tissue. The tissue decomposes as it is heated by the fire (called “pyrolyzing”), releasing combustible gases that burn as a free flame. This in turn heats the remaining solid matter to drive off more combustible gas…. Much of the heat is carried away as hot gas, up into the smoky buoyant plume above the fire. But much, too, escapes as radiant energy from the bright flame, returning to the form in which it was released from the sun and captured by the living plant. In the form of radiation, the energy flees with the speed of light and travels in straight rays until absorbed by matter. When it is absorbed, this energy raises the temperature of the matter which has captured it. Fine, dry plant components very near the flame are thus heated very quickly to the temperature at which they must decompose, giving off combustible gases that are in turn ignited by the nearby flame. In this way fresh fuel is added to the fire, to replace that just consumed in the flame. So the fire spreads.
Therefore, when Dodge spoke of a solid “wall of flame” behind him, 250 to 300 feet deep, he was speaking figuratively as a poet, as most of us do. What was behind him were hundreds of thousands of little fires multiplying so fast that only a computer could keep up with them.
I had to walk around this explanation of the process of the spread of wildfire several times before going on, because everything ahead comes from it. The mathematical analysis of wildfire requires a structure of thought, not just some close observations of smoke and flame, and this structure is spoken of as a “philosophy” by the mathematicians. As a philosophy it has a center from which everything flows, and the center is a definition and the definition turns out to be this explanation of fire spread. If a spreading fire is a bunch of little fires becoming many more little fires, then a lot of counting has to be done to make a study of it. Think of what a lot of counting of a lot of pine needles had to be done to come to the following conclusion:
These equations show that the rate of spread in our ponderosa pine needle fuel beds decreased by 4.23 percent for each 1-percent increase in fuel moisture. Rate of spread in white pine needle fuel beds decreased by 4.55 percent for each 1-percent increase in fuel moisture. If the effect of moisture remains linear as moisture content increases, the pond
erosa pine needles would not sustain a rate of spread at a fuel moisture of 24 percent in the still air environment. Similarly, the limit for white pine needles would be 22 percent. (Richard Rothermel and Hal E. Anderson, “Fire Spread Characteristics Determined in the Laboratory,” U.S. Forest Service Research Paper)
It also helps to remember what this definition looks like when it is in operation in the Fire Lab: a wind tunnel with a computer nearby. What you don’t see is what is installed inside the computer—the structure of thought being outlined here, including of course as its centerpiece the quantitative definition of the spread of wildfire. “Facts” Laird and I gave the mathematicians about the Mann Gulch fire that seemed worth considering would be given to the computer, which would consider them within the structure of thought it had been given. I suppose it is something like a creamery—a lot of things are churned around in it, and some are supposed to rise to the top as butterfat.
As the structure of thought has, as its centerpiece, a definition, so the definition of the spread of wildfire has fuels as its centerpiece. In the analysis of fire spread with fuels at the center, fuels are first analyzed as particles, some of the most important factors being their size (“the ratio of surface area to volume”), heat of combustion, and ash content. Nothing is more important about the arrangement of particles than their compactness, with the limitations of being so close together as to stop a fire or so far apart as not to let it get started. Once started, however, these combustibles are “whipped,” “smothered,” or “kept alive” by such environmental factors as the wind velocity, the slope of the ground, and the moisture content. Just from knowing that usually the height of the forest fire season in Montana is August, when it is hot, dry, and windy, we know something of the varying influence these forces can have on any fire or sector of it.