In our column "In Defense of Non-Conventional Energy Sources" we gave an overview of the "Inhaber Report" scandal, an attempt by the Canadian Atomic Energy Control Board (abetted by various prestigious journals and scientists) to foist the results of a hopelessly flawed study upon the public. The report—entitied Risk of Energy Production (AECB1119) — written by Herbert Inhaber, purports to show that solar energy sources are nearly as dangerous to human beings as are conventional sources (such as coal) and more dangerous than is nuclear energy!
Because we know that MOTHER EARTH NEWS' readers are interested in alternative energy sources, this column will attempt to describe—in detail—how biased data selection was combined with incredible bungling to reach AECB-1119's preposterous conclusions.
A Comprehensive Comparison?
First, Inhaber claims to have tallied up the direct health risks—to workers and to the public—brought about by the construction and operation of conventional and nonconventional energy sources. He claims to count accidental deaths and injuries, disease, and all phases of the fuel cycle (such as, for example, acquisition of the raw material to build energy facilities, the manufacture of components, the construction of power plants, the mining of fuels, plant operation and maintenance, and the disposition of wastes).
This approach (while of some use if carried out successfully) omits the impact of the energy sources in question upon human well-being through disruptions of climate, of ecosystems,. and of sociopolitical affairs. Yet such "indirect" damages (including the loss of agricultural productivity due to climate disruption by fossil fuel-produced carbon dioxide, and the proliferation of nuclear weapons induced by the spread of nuclear reactors) put far more human lives in danger than do "direct" accidents and pollution-induced illness.
Even among those environmental risks that he did consider, Inhaber omitted many important phenomena. Public disease, for example, appears in the report in only two forms: as days lost from work—and premature deaths—from respiratory illness associated with emissions of sulfur oxides in the presence of particulates, and as days lost due to cancer caused by radiation from nuclear power plants.
Inhaber offers no estimate of the disease effects of the oxides of nitrogen, hydrocarbons, or trace minerals (mercury, lead, cadmium, nickel, etc.) that are emitted when fossil fuels are burned; entirely neglects to mention public disease caused by water pollution from contamination by the hydrocarbons and the trace metals that are released during extraction, processing, and transportation of fossil fuels; and he ignores all of the diseases and genetic effects that may appear in future generations as a result of nuclear-power-produced radiation or chemical mutagens from fossil fuels.
Admittedly, the data needed to make meaningful estimates of some such effects are simply not available. That's all the more reason to find fault with the repeated assertions by Inhaber—and especially by the AECB—that his analysis is a comprehensive treatment of the accident and disease risks of energy sources. Such statements, in turn, attempt to justify the report's sweeping conclusions about which power sources are more dangerous than which others.
A Consistent Comparison?
Actually, most of Inhaber's analysis of public health and occupational health consequences consists of multiplying tonnages of materials (say, the amount of steel needed to build a windmill) by the man-hours—in various occupational categories — needed to make the material, fabricate it into the needed components, and construct a finished energy facility from these parts. Next, Inhaber multiplies the man-hours in the various occupational categories by the corresponding rates of incidence of occupational risk. And if the air pollution produced per ton of material and the ton-miles of transportation are known, he adds in estimates of the public health consequences of these two processes as well.
So far, so good? The catch is that, while Inhaber has indeed applied this method to his analysis of the nonconventional energy sources and of hydropower—and while his report and various published summaries lead the casual reader to believe he used the same approach for coal, nuclear power, oil, and natural gas—he, in fact, has not counted the risks of creating these conventional facilities at all, only the risks of operating and maintaining them!
Inhaber made a number of less obvious blunders, too. His data on conventional energy technologies relate to real commercial systems, in which economic forces have acted to minimize the amounts of materials and labor used. But the same estimates for nonconventional energy technologies, on the other hand, are for preliminary commercial designs, for experimental systems, or for entirely hypothetical installations in which there has been little or no economic optimization. Moreover, from the array of data available, Inhaber has almost invariably chosen examples that are most demanding of materials and labor—and many of these designs will probably never be built commercially, since they'll be beaten out in the marketplace by the more frugal designs that the study chooses to ignore (or by still more thrifty ones not yet invented).
Conceptual Confusions and Misreadings
One of the study's most elementary errors is the assumption that one Btu (or joule or kilocalorie) is the same as any other (Heat is taken as the equivalent of electricity, electricity at the power plant is considered to be equivalent to mechanical energy delivered to the wheels of an automobile ... and so on.) This error alone is responsible for inflating Inhaher's estimate of the risks of methanol production by a factor of between three and six!
However, Inhaber's most important conceptual confusion is his complete misunderstanding of the storage and backup requirements of intermittent energy sources such as sunlight and wind.
He provides his solar power plant with enough energy storage to match or exceed the annual energy output of a coal-fired or nuclear-powered plant with the same rated capacity. Because of this particular solar plant's very large built-in energy storage (the materials and labor requirements for which Inhaber cheerfully tallies up), the installation is more than a match for the conventional baseload plant it replaces. Yet Inhaber irrationally charges the solar plant with an additional megawatt-year of coal "backup" for every megawatt-years of solar-generated electricity it produces. Such unnecessary backup systems account for 65-85% of his upper limit risk estimates for wind, photovoltaic, and solar-thermal-electric technologies.
Other conceptual errors in AECB-1119 include: adding together occupational and public risks (which should not be combined because—among other reasons—the public risks are largely involuntary, while the bearers of occupational risks often have made a conscious choice to accept the hazards in exchange for the wages and other benefits of their jobs); ascribing the same number of "days lost" to premature death from aggravated respiratory disease as to death from cancer (which is incorrect because the former disease usually takes weeks or months off its victims' lives, while the latter—on the average—cuts a lifespan by many years); and inflating the risks of coal as compared to nuclear power by assuming mining conditions and pollution control practices that are out of date and now illegal.
Inhaber's 50-fold misreading of a windplant's material requirements was discussed in our previous column. In connection with ocean thermal energy conversion (OTEC), he understates—by more than twice —the generating capacity of the unit whose material requirements he tabulates, while overstating the construction labor requirements by 10 to 50 times!
The report's author also claims to have used the highest available nuclear accident risk values, in order not to be accused of pronuclear bias. Actually, Inhaber's "upper limit" estimate of nuclear accident risk is from two to several hundred times lower than those given in three of the four references on reactor safety he cites. He also implies that he has included estimates of the risk of radioactive-waste management. He hasn't.
Inhaber's assessment of the risks involved in constructing a methanol factory are scaled from figures for an oil refinery, and are for operation and maintenance, rather than for construction. His material and labor requirements for obtaining natural gas (figures that he never actually translates into risks, as noted above) come from data for fuel-oil and coal-gasification facilities.
To obtain material requirements for energy storage at a solar-thermal-electric plant, he subtracts the weight of a hypothetical plant with no storage from the weight (as given in a later report from the same project) of a similar plant that includes storage. Since the difference in the quoted weights was caused by the correction of earlier errors and changes in assumptions—as well as by the addition of a storage unit— the procedure is useless. Yet, this same storage system size was then used for Inhaber's calculations for photovoltaic and windpower plants as well even though, as a heat storage system, it's completely irrelevant to those technologies.
Because the study considers energy sources in Canada, Inhaber scales up the size of his solar-thermal-electric power plant by 32% to account for Canada's having 32% less sunlight than the southwestern U.S., the area for which the plant was designed. In reality, only the size of the collector needs to be increased to enable such a plant to produce the same amount of energy as it could in America's Southwest, but Inhaber has increased the size of the boiler, the generator, and even the coal backup system by the same factor.
He inflates the material requirements for the production of photovoltaic cells by a bizarre procedure that depends on rejecting a number that's easily seen to be correct and cobbling up a "correction" factor which is based, in part, on a misunderstanding of the difference between the peak and average output of solar cells.
Then, to build up his estimate of risk per unit of energy, Inhaber combines some of the highest figures for the material requirements of renewable energy systems with some of the lowest system performance figures available. His risk estimates are inflated still further by the false assumption that people in the most dangerous occupations (e.g., roofing and sheet metal) perform most of the work on renewable energy installations.
In a similar scam, Inhaber uses material requirements for an unconventional power plant from a respectable reference, but ignores that reference's figure for the plant's annual output in favor of lower figures found elsewhere. In one example, he takes his material requirements—for windplants—from the Project Independence report, but ignores that study's annual load factor of 0.34 (34 megawatt-years of annual output for each 100 megawatts of rated capacity) in favor of the figure of 0.06. (This lower load factor has the effect, of course, of increasing the materials needed per megawatt-year of output.)
The study also makes the unwarranted assumption that facilities producing methanol from biomass will wear out in 20 years, while oil refineries will last three decades This sort of faulty figuring inflates the construction risk of biomass by 50%. Inhaber also assumes that such a plant will be twice as big as an oil refinery per unit of energy output, ostensibly because methanol contains about half as much energy per pound as does gasoline. This twofold increase in methanol's construction risk is unwarranted, for methanol's simpler technology probably would result in a plant that's smaller per unit of energy output than is a complex gasoline refinery.
Inhaber also assumes that all the materials used to construct unconventional energy facilities—including rock for energy storage and sand and gravel to make concrete—must be hauled 300 miles, the average transport range for coal used in power production. This assumption leads to a large risk associated with transportation accidents, when—in reality—such common materials must travel only a few dozen miles.
To estimate the maintenance risk of rooftop solar heating systems, Inhaber starts with a very generous estimate of the time he thinks professional roofers would need for the job, and computes their injury rate for this number of hours. Then he multiplies this risk by four, assuming that homeowners who maintain their own collectors will take twice as long to do the work and will fall off the roof twice as often per hour!
Covering the Trail
How could such an array of mistakes have escaped the notice of the reviewers of the articles in which Inhaber summarized his own "results" for Science, New Scientist, and Energy? Part of the answer is that none of Inhaber's detailed calculations—and few of his assumptions—were included in the articles he submitted for publication. The bodies were all buried in the report's appendices, and further concealed under layers of arithmetic mistakes and typographical errors To unravel the chain of fallacies that led to the conclusions amounted to a major undertaking even for an expert at risk assessment like our colleague, John Holdren (whose group is responsible for the critique reported here).
One of the ironies of the Inhaber fiasco is that Science magazine, whose February 1979 article by Inhaber gave his conclusions their widest circulation, had published only eight months earlier an editorial entitled "Bad Science and Social Penalties" by Cyril Comar (a nuclear advocate from the Electric Power Research Institute), which contains a description of "bad science" that fits AECB-1119 like a glove. Comar was also remarkably prescient in predicting the kind of public response Inhaber's work has received: "Bad science, being more newsworthy, will tend to be publicized and seized on by some to support their convictions."
We are in complete agreement with Comar's prescription for dealing with the general problem: "Perhaps the best solution lies in greatly increased efforts of individual scientists ... to vigorously take the initiative and responsibility, to immediately expose bad science whenever it occurs, and thereby to reestablish credibility.
Paul Ehrlich (Bing Professor of Population Studies and Professor of Biological Sciences, Stanford University) and Anne Ehrlich (Senior Research Associate, Department of Biological Sciences, Stanford) are familiar names to ecologists and environmentalists everywhere. As well they should be. Because it was Paul and Anne who—through their writing and research—gave special meaning to the words "population," "resources," and "environment" in the late 1960's. (They also coined the term coevolution, and did a lot to make ecology the household word it is today.) But while most folks are aware of the Ehrlichs' popular writing in the areas of ecology and overpopulation (most of us—for instance—have read Paul's book The Population Bomb), far too few people have any idea of how deeply the Ehrlichs are involved in ecological research (research of the type that tends to be published only in technical journals and college textbooks). That's why it pleases us to be able to present these semi-technical columns by authors/ecologists/educators Anne and Paul Ehrlich.