Outlook for Nuclear Power: Presentations at the Technical Session of the Annual Meeting--November 1, 1979, Washington, D.C. (1980)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Need for Nuclear Power Worldwide: World Regional Energy Modeling WOLF HAFELE* SUMMARY Based on a 5-year study of the International Institute for Applied Systems Analysis, Laxenburg, Austria, this paper identifies the factual basis of today's energy situation, stressing the need for and difficulties in long-term supranational energy supply strategies. The bases are two scenarios—defined by close to observed trends of population and economic growth—that indicate a conceivable energy demand range until 2030. These scenarios are quantified for seven comprehensive world regions by way of a highly iterative model set designed at IIASA to study long-term, dynamic, and regional/global aspects of large-scale energy systems. The results illustrate the need for the world to use all available, and high-cost, energy sources. Nuclear and solar technologies replacing fossil resources will fully have to come to bear after 2030. Before, the weight will have to change from primary to secondary energy supply, e.g., to synthetic fuels production from coal. Resource allocation and trade flows will in general be restricted by production ceilings. Thus, prudent political and economic decisionmaking is in order for the world and its regions to ensure a satisfactory long-range energy supply. INTRODUCTION Nationally, there are numerous schemes and programs for planning the energy future. They provide a wealth of detail and a fairly short, or at best medium-term, planning horizon of up to about l5 years. After all, detailed measures must have a relatively short time frame in order to be feasible. Innovation, however, that is to say major changes in an existing infrastructure, involves much longer periods of time. Compare, for example, the evolutions of various primary energy carriers in the energy market. Figure l is a logistic representation showing the S *Wolf Hafele is Deputy Director and Program Leader, Energy Systems, International Institute for Applied Systems Analysis, Laxenburg, Austria.

1-F Fraction (F) 0.99 0.01 1850 1900 1950 2000 FIGURE l World: primary energy substitution. curves of energies penetrating the market (from share 0 to share l) as straight lines. Their behavior is remarkably regular, extending over more than a century. This is just to underscore the need for a long-range view once changes in the energy supply structure are at stake. If one takes into account that the life of a power plant averages 30 years, the time bracket to consider for innovation may well be 50 years. But there is also other evidence. Consider the dwindling of cheap (known) fossil reserves. Consider the impact of tapping energy sources on the world climate and on man's environment, which is not well understood. These and other aspects all suggest a long-range perspective. Looking at the energy problem from this angle, one is confronted with several difficulties: • We are not prepared, either analytically or via control of the market mechanisms, to come to grips with the interplay of short- and long-term aspects. • International interdependence will have increased considerably in 50 years from now. This will not come as a surprise, since already today 50% of the crude oil in the Federal Republic of Germany, for example, originates from one single area, the Persian Gulf. Yet we will have to do better in viewing the world as a whole. • There is a severe lack of input data, required for standard energy planning tools, with respect to the 50-year time frame and the way the world will then look geopolitically. Or what about elasticities, for example, i.e., the change in percent in the demand for a given secondary energy as a function of percentage changes in various determinants, such as prices or gross domestic product? Whereas most large-scale econometric models now in use in the United States assume availability

of elasticity inputs in one form or another, for long-range investiga- tions such elasticities cannot be obtained. For these reasons, medium- and long-term strategies of energy supply are difficult to deal with, both in terms of substance and methodology. An attempt at formulating long-term global energy supply strategies has been undertaken at the International Institute for Applied Systems Analysis (IIASA), Laxenburg, Austria, an East-West venture supported by scientific institutions of l7 nations, which is concerned with problems of civilization in a long-range and globally comprehensive fashion. Initially, IIASA1s Energy Systems Program gave much thought to a quali- tative understanding of the energy situation and a breakdown of the problem into feasibile subsets. This then enabled the program to pro- ceed to a synthesis by quantitative analysis and integration of many special studies into a global and long-term energy supply and demand picture. The tool chosen is the writing of internally consistent scenarios: the greatest possible number of necessary conditions is identified and used to narrow down the scope of subjective judgment. To support the effort, a set of computer models is relied upon. The principle guiding this approach is plausibility. Two basic scenarios are constructed, marking the range between plausible upper and lower bounds. But since there is only one reality in store for us, the scenarios are valued as guidelines toward conceivable energy futures and their outcomes as indi- cators and not as predictors. With the ordering forces of the market having lost much control, such guidelines are indispensable. This above all holds for the oil market, where prices and obtainable quantities have come to be simply vehicles of political action, irrespective of the market mechanics. They will work again properly, leading to the neces- sary investments, only when confidence and trust in the market are being restored. TWO SCENARIOS For some steps in the description of the energy problem, it is useful to consider global overall figures like the following. The world today consumes about 8 TWyr/yr of energy, i.e., commercial energy. (Note that l TWyr/yr fairly accurately corresponds to l billion tons of coal equivalent [tce] per year.) The average per capita consumption then amounts to 2 kWyr/yr (see Figure 2). About 70% of the people of the world, however, live on much less than the average, and a considerable number of them on only 0.2 kWyr/yr per capita. A conceivable addition of 0.3 kWyr/yr from burning wood and manure figures high in this context. About 22% of the world population use 2-7 kWyr/yr per person, the Euro- peans among them. The remaining 6% enjoy a per capita energy use of 7-l2 kWyr/yr. If, as in Figure 3, one assumes a doubling of the world population in the coming 50 years—a change Keyfitz considers rather conservative—and in an increase in the per capita average to 3 or 5 kWyr/yr, the world's energy demand rises to 24 or 40 TWyr/yr, respec- tively. This consideration, though so simple, helps assessing the efficiency and capacity required for future energy systems. Their

NUMBER OH w 1 COUNTRlES 70 1 60 I so • 1 % OF TOTAL POPULATlON to - 1 1 7V » 30 \ 20 V 10 - =*»-«-; 01231567 8 9 10 11 kW/CAP FIGURE 2 Per capita commercial energy consumption in the world, l975. billion people 10 + 8 6 2 study period world in transition 1800 1900 2000 2100 FIGURE 3 World population: historical and projected.

magnitudes often turn out to lie between expectation and observed reality, a thought that will come up again later. Unfortunately, it appears that such rough guidelines are not suffi- ciently detailed for real-world decisionmaking, and one is tempted to go back to the national framework. This cannot be done here, however, since it would impair the globally comprehensive vision of the problem. IIASA, in seeking a way out of this dilemma, has identified seven regions that describe the world as a whole. In this way, typical regional dif- ferences are accounted for and regional interdependencies identified (Figure 4). These regions differ above all by their states of economic development and the availability of resources and, to a lesser extent, by geographical conditions. Regions I, III, and II correspond to the so-called first and second worlds: the industrialized North. Regions IV, V, VI, and VII represent the developing third and fourth worlds, with widely differing national structures. To each world region a set of mathe- matical models is applied separately. This IIASA set of energy models is depicted in Figure 5, with the larger computer models shown in boxes. Assumptions on population and economic growth in the various regions enter the model MEDEE, which calculates final energy demand in consider- able detail, i.e., the use of energy by the final consumer. The model output, a set of secondary energy demands such as electricity, heat, gas, etc., is the input to the supply model MESSAGE. This linear pro- gramming model allocates specified quantities of primary energy, such as oil, gas, coal, uranium, etc., to the generation of secondary energy over a period of 50 years. It produces optimal discounted costs and, most important, takes into account various constraints, providing in this sense an optimal supply mix of primary energies in a region. The resulting requirements for direct and indirect investments in energy generation are accounted by the model IMPACT. The model uses an input- output approach to identify the effect these requirements may have on a given economy. The aggregated investments are then fed into the MACRO model, which helps assess the macroeconomic implications of changes in the ratio of energy consumption and investments. With a quasiformalized procedure linking international trade between the world regions, a first- order approximation of input data for MESSAGE is obtained. Another output, to be derived from the model results, is prices and elasticities. While all the models in the set in their present form have been devel- oped and applied at IIASA, their origins vary. MEDEE orginates from the University of Grenoble, MACRO is founded on work in Canada and the United States, and IMPACT and the world trade procedure come originally from the Siberian Power Institute at Irkutsk. The model MESSAGE has been completely developed at IIASA. In the light of the explications above, it is crucial to merit the iterative character of the modeling procedure. The findings described below could in no way have been obtained from one run through the modeling loop. Rather, the procedure is in steps, so that interfaces can be installed in between them for iterative modification until a consistent analysis is obtained. Thus there is indeed room for assump- tions and judgment in the light of the underlying mental model. Two scenarios have been constructed that are defined by two basic development variables, population and gross domestic product (GDP).

o o LU CC _= = >>>> D a z < oo 2< a: t- UJ 00 i < <_> 2 X < 5 I E I = Q < 2 OC LU O 2 << * —. (_} sr ^* ™" ^ LU *-- ~r» 00 LL. ^ Q 2 — — co o E LU cc o < LU tM s < 2 3 = < s ss LLJ ^ i < < 2 1 < 2 ~l-< 3 01 UJ ,_ a. < < — CS 00 ^ o 2o ^ oo .25 LU 3< ^ E i§ ^ a $£ 5 ^ LU < < oo 3 • —i ^-^ —^ OO ™^ CO c o •H k i 0) en o H frj

Scenarios, e.g. Economic Growth, by Region lnvestment and Consumption MACRO Accounting of Energy Uses MEOEE Economics lmpacts lMPACT Energy Supply and Conversion MESSAGE Population Economic lnterpretation Prices Elasticities Constraints • Resources • Buildup Rates !lnter-Regional Analyses Energy Trade FIGURE 5 IIASA's set of models for energy strategies. Both scenarios, High and Low, are rather conservative, representing moderate departures from observed trends cases. In either scenario, population is assumed to grow to 8 billion in 2030 (and would then taper off to a sustainable level). The basic difference is in GDP projections: one scenario assumes a relatively low economic growth, fairly large advances in energy end-use technology, and a rather positive attitude towards energy saving of those concerned. The other scenario assumes a modestly high growth. The less conservative assumptions made on the supply side in both scenarios include effective and timely decision- making and implementation, as well as due regard for the needs of the developing countries. In all, these assumptions are rather optimistic, marking the bounds for what may be maximally feasible, while the real world experience may turn out to be more sobering. ENERGY DEMAND Let us now look more closely at how energy demand is dealt with in the two scenarios. To this end, it is useful to consider the economic evolution of the regions in terms of percent of per capita GDP. Table l gives the per capita GDP and its yearly growth rates used in the High and Low Scenarios, l975-2030. In the Low Scenario, the per capita GDP growth for North America (Region I) goes down to 0.7%/yr, and that for Europe (or more exactly, Region III) comes to be only 0.9%/yr. Both values are meant to approximate zero economic growth.

TABLE l l975 Per Capita GDP and Growth Rates for Two Scenarios to 2030 GDP Per Growth Rate of Per Capita GDP (%/yr) Capita High Scenario Low Scenario ($) l975- 2000- l975- 2000- Region l975 2000 2030 2000 2030 I 7,046 2.9 l.8 l.7 0.7 II 3,4l6 3.6 3.2 3.l l.9 III 4,259 3.0 l.8 l.7 0.9 IV l,066 3.0 2.4 l.6 l.9 V 239 2.8 2.4 l.7 l.4 VI l,429 3.8 2.8 2.4 l.2 VII 352 2.8 2.4 l.6 l.4 The highest growth rate, by contrast, is that for Region VI (Middle East and Northern Africa) from now to the turn of the century (3.8%/yr). The Soviet Union and the Eastern European countries (Region II) have generally high values but otherwise follow the decreasing trend. Gross regional products (GRP) are obtained by multiplication of the GDP growth rates by regional population figures. (The population data here are from Keyfitz.) In the OECD countries, GRP annual growth rates, l975-2030, range between only 2% to 3%; the evolution assumed follows the general decreasing trend. More important than the economic data are the related values of energy demand, however. An adequate definition of energy flows from the source to consumption differentiates at least between primary energy and secon- dary energy, as well as energy use (see Figure 6). The latter term in fact comprises what is called energy services, resulting in a fine piece of pottery, a warm room, or adequate illumination for reading. This energy service can be consumed, other than energy itself, which follows the law of conservation. Use of energy has to do with the negative entropy or negentropy (or information) content of energy. This rather abstract quantity, equivalent to the use of capital or work or to the impact of know-how, can completely or partially be consumed or substi- tuted: the piece of earthenware may break, the room may cool down, and the light photons are absorbed. The point here is that the relationship between energy consumption— which depends on the level of a given economic activity—and the econo- mic activity itself is not unambiguous and straightforward. No wonder the issue is in the center of controversy today. It surfaces in the discussion on energy coefficients, that is the percentage of energy growth required per percent of GDP growth. And as we will see in a moment, differentiation of final energy and primary energy is very important in this context. The final energy-GDP coefficients for Regions I, II, and III center around 0.8, but those for Regions IV, V, VI, and VII are around l.5, which demonstrates the need for the developing countries still to build

up their infrastructures. By the way, more energy is needed at first to build a railway system than to ship by it computer printouts later, which may stand here for the most recent sophisticated accretion of GDP. To argue this point in a convincing and credible manner, lots of details are necessary. MEDEE, the model for assessing long-term energy demand, is meant to do so. It does in fact account for the great diver- sity of end-use categories and their interdependencies. Figure 7 summa- rizes the relevant results. The final energy-GDP coefficient is assumed to go down to as low as 0.3 for the industrialized countries, but to only slightly less than l.0 for the developing countries. The respective coefficients of primary energy and GDP, on the other hand, may differ completely, both in quantity and in quality (see Figure 8), where e for Regions I, II, and III in l975 was close to l.0 but clearly above 0.8. This is due to conversion losses at the level of secondary energy generation and to electricity generation in particular. Considerable losses arise from coal liquefaction, which plays a major role in both scenarios for Regions I and III, as will be seen later. This then may well cause the primary energy-GDP coefficient to rise to values higher than l.0. It is not possible here to treat these energy demand calculations in greater detail. One point merits special attention, however. The demand for liquid secondary energy carriers, such as heating oil or gas, is shown to be a much more severe bottleneck than is widely assumed. Therefore, it is attempted in both scenarios to limit the use of liquid fuels to practically nonsubstitutable applications, such as for trans- portation, feedstocks, and petrochemicals. Table 2 shows this use of liquids in percent. The shares vary between about 50% today and more than 90% in 2030. Therefore, it becomes more and more necessary to substitute district heat and electricity for heating oil and gas. As a consequence, there is a continuous steep increase in electrification (Figure 9). Such demand considerations lead to a per capita primary energy demand in the scenarios as in Table 3t a world average of 3 or 4.5 kWyr/yr, respectively, in 2030—similarly as was noted above—instead of 2 kWyr/ yr per person as of today. While the per capita consumption ratio of Regions I and III, compared to Regions IV and V, improves by a factor TABLE 2 Use of Liquids: Percent of Liquid Demand Used for Transportation and Feedstocks Region l975 High 2030 Low 2030 I 74 94 9l II 65 l00 l00 III 52 86 76 IV 69 90 89 V 58 9l 88 VI 74 94 9l

l0 M-l 0 c o -H --P -H 01 •d in O 0) c (4 g H

ll I in Q. '- •a o tM CM in o r» o ' 0 in in r-« 01 01 in < •H O •H <4-J M-l 0) o o a o Q) C Q) H (0 1 •H M CO D O H o CO -= 0> I p r-i C 3 •H u •H 0) o u Q O l >. CTI Vl 0) C Q) C •H

l2 30-i percent of final . energy 20- High Scenario FIGURE 9 Electrification l975-2030. 10- 1975 2000 2030 of about 2, considerable inequities between developed and developing countries remain, and the gap continues to be a problem far into the next century. The global primary energy demand is projected in Table 4, with 22 TWyr/yr in the Low Scenario and 36 TWyr/yr in the High Scenario. Mind you, it is easily possible to obtain still lower as well as higher values if the scenario assumptions are slightly varied. One way TABLE 3 Primary Energy Per Capita (kWyr/yr per capita) Regions l975 High 2030 Low 2030 I + III 6.2 l2.2 (2x) 8.2 (l.3x) IV + V 0.4 l.9 (4.8x) l.l (2.9x) WORLD 2.l 4.5 (2.2x) 2.8 (l.4x) I+ III l6.2 Ratl° 6.4 7.5

l3 TABLE 4 Primary Energy Projections (TWyr/yr) Regions l975 High 2030 Low 2030 I + II + III 6.8 20.5 (3. Ox) l3.9 (2.lx) IV + V + VI + VII l.5 l5.2 (l0.5x) 8.5 (5.8x) WORLD 8.2 35.7 (4.3x) 22.4 (2.7x) of comparison tried at IIASA is a l6 TWyr/yr scenario that retains the 2 kWyr/yr per capita average of world energy consumption. An increase in energy use in the developing countries must accordingly be offset by a negative energy consumption growth in the industrialized world. Table 5 describes this case for the seven world regions. There is at present quite some agitation to promote a negative or zero energy growth for other reasons than energy supply difficulties. Yet the impact this movement will have on our way of living cannot be grasped. At the other end of the spectrum, however, there are world energy consumption estimates of clearly more than 40 TWyr/yr. The political concept of the New Economic Order, for example, pronounced by the UN group of the 77 at UNCTAD conferences, leads to such higher energy demand values. But this is not out of focus with an observation above that the link between energy and economy is not naturally a closed one. In this light, the energy demand figures of the High and Low Scenarios fall well within the mid-range of today's projections. TABLE 5 Per Capita Primary Energy Consumption, a l6-TW Scenario, l975-2030 (kWyr/yr per capita) Base Year 2000 2030 Regions (l975) I ll.27 9.l 8.0 II 5.l0 7.2 6.2 III 4.03 3.6 3.2 IV l.06 l.8 2.8 V 0.23 0.5 0.7 VI 0.96 2.2 3.6 VII 0.5l l.0 l.2 WORLD 2.l 2.0 2.0

l4 ENERGY RESERVES AND ENERGY RESOURCES A clear distinction must be made between reserves and resources. Reserves, being resources that are explicitly known, can be mined at economic conditions. Resources then are considered to include reserves as well as a resource base: this resource base is presumed to exist with a certain probability by way of geological evidence, but its exploi- tation is not evidently economic. Both categories continuously vary in quantity, the ultimate difference between them being technology. North Sea oil is a case in point. There, with the technology of floating platforms at hand, the resources have become reserves. Consider, how- ever, that such technology has become feasible only after l973. Estimations of resources traditionally differ from each other. Geolo- gists apparently tend towards cautious estimates. Economists, on the other hand, guided by the role of price increases, proceed from a de facto unlimited resource base. Inherently different definitions apply for coal and gas, with coal being usually estimated on a geological basis and oil with a view to maintaining a certain reserve-production ratio. Resource assessment is difficult, therefore, especially if assumptions on future conditions are involved, as in the present scenario. While all this commands caution, one must come up with numbers for the scenario definition. The estimates in Table 6 of ultimately recover- able fossil resources should be looked at this way. They are grouped by three price categories. Coal, oil, and gas of the cheapest category ($25/t or $l2/boe, respectively) make up about l,000 TWyr. Simple cal- culation shows that for 40 TWyr/yr this amount would be used up in 25 years, and one comes to realize that this is where the public's concern about resource scarcity originates. Realistically, Categories II and III must also be included in the count, leading altogether to about 3,000 TWyr. Also, it seems more appropriate to assume the world's fossil energy use of the next five decades to average about l5 TWyr/yr, which makes the situation appear less tense. Of course, more details, TABLE 6 Ultimately Recoverable Resources. Coal I: $25/t, II: $25-50/t; Oil, Gas I: $l2/boe, II: $l2-20/boe, III: $20-25/boe Resource Coal (TWyr) Oil (TWyr) Gas (TWyr) Cost Category I II I II III I II III I (NA) l74 232 23 26 l25 34 40 29 II (SU/EE) l36 448 37 45 69 66 5l 3l III (WE/JANZ) 93 l5l l7 3 2l l9 5 l4 IV (LA) l0 ll l9 8l ll0 l7 l2 14 V (AF/SEA) 55 52 25 5 33 l6 l0 l4 VI (ME/NAF) <l <l l32 27 n.e. l08 l0 l4 VII (C/CPA) 92 l24 ll l3 l5 7 l3 l4 WORLD 560 l,0l9 264 200 373 267 l4l l30

l5 in particular with regard to regional differences (see Table 6), were used in the modeling, as in the case of energy demand. But, even at this summary level, large differences are obvious between different qualities of coals, oil, and gas. For example, oil recovery in Saudi Arabia or in Alaska or in the North Sea do not easily compare. It shows that the 3,000 TWyr in Table 6 are not a soft cushion to rest on. Besides fossil resources, energy supply from renewables is getting much attention, nourishing wide hopes. If one neglects the large-scale use of solar energy temporarily, one finds the potential of the remain- ing sources to be limited. Much explanation would be needed to fully prove this statement. Since this is not possible here, a summary of the estimated potential of renewables is given in Table 7. It differentiates between what is theoretically or technically possible and what may actu- ally be feasible. The realizable potential of renewables appears to be about l0 TWyr/yr, a figure much lower than the expected demand. The average energy densities of renewables, on the other hand, are 0.l-l.0 W/m (Figure l0). With an indicative energy density of 0.5 W/m , this implies that 20 million km2 of land will be needed for harvesting the realizable potential of l0 TW—an area about as big as all the agricul- tural land in the world! Another matter is the large-scale use of solar energy, the annual average density of which may be 20-40 W/m , and whose area requirement is relatively smaller. Extensive investigations5 have shown that the need for land is bound to complicate solar use in some such cases, but TABLE 7 Estimated Potential of World Renewable Energy Supply Potential Source Technical Realizable Constraint Forest and j. , j- b.U fuel farms Solar panels Soil storage 5.0 Heat pumps Hydropower 2.9 Wind 3.0 OTEC l.0 Geothermal 0.2 Organic wastes 0.l Glacier power 0.l Tidal 0.04 TOTAL 20 5.l l.0 l.5 l.0 0.5 0.6 0.l 0 0 9.7 TW ecological climatological economic technological ecological social economic ecological climatological technological economic balanced technological computational

l6 might be resolved in general. A much greater long-term problem may be capital cost and energy storage. An overriding concern is the tremen- dous demand for material needed to cover such areas. The minimum mate- rial density is generally estimated to be l0 and l00 kg of concrete and iron per square meter. The solar growth rates in Figure ll are inferred on this basis. With a hypothetical doubling of world concrete and iron production to be invested in solar power plants, between 250 and l,500 GW(th) could be installed per year. Therefore, and for other supporting reasons, the growth rate rather than the solar power potential appears to be a leading constraint on large-scale solar power application over the next 50 years. This, of course, assumes optimistically that capital cost and storage requirements can be met. Now a few words about nuclear energy. In the present context, the question is above all the magnitude of what nuclear energy can at best Demand lamie of demand densities, typical urban systems 1975 W/m' -r 10 f t 1.0 - range of demand densities, Regions l -Vll 2030 0.1" FIGURE l0 Energies densities. 0.01 -4— Supply wind (North Sea coast) wet geothermal heat pump, soil (FRG) fuelwood plantation OTEC, tropical oceans wind, continents surface layer fuelwood biogas geothermal dry hydropower

l7 G\N(lli)/yr 10000 1200 4 -/- present world total material production present iron ore or concrete production 1 0.1 1 10 100 1000 10000 net annual material requirements (106 tons) FIGURE ll Material requirements: solar conversion systems of various net densities. contribute by 2030, all institutional and societal issues put aside. There is only a vague answer to it. Detailed studies—not discussed in the present context—indicate a worldwide realistic upper limit of l0 TW of installed electric capacity. It is trivial to show that the maxi- mum could be less. To l0 TW(e) installed capacity, a (commercial) pri- mary energy consumption of l7 TWyr/yr would correspond. (Note that TWyr/yr always implies annual calorific input to produce power.) At such a growth rate, the sensitive parameter is the natural uranium re- quirement, e.g., of light water reactors. For example, the International Fuel Cycle Evaluation (INFCE) considers that 4.3 million tons of natural uranium at prices up to $l30/kg are available in the Western world today. For the world as a whole and with worldwide prospection at the present U.S. level, one could consider uranium resources of, say, 20 million tons--but this number should be taken as a working hypothesis rather than as established fact. This large amount would be used up by about 2020, however, if light water reactors (LWR's) or other nonbreeders were the only nuclear technology deployed. Therefore, breeders must play their part in time. For example, the plutonium from LWR's could be fed into fast breeder reactors (here LMFBR's, see Figure l2) and be used as breeder inventory that is not consumed but breeds more fuel. A once- through of 20 million tons of natural uranium would lead to about 24,000 tons of Pu, which means that the l7 TWyr/yr in question could be produced for a virtually unlimited period of time. However, this possi- ble reactor strategy presupposes fostering now an intensive buildup of

l8 installed 12 - • nuclear capacity inTWe 10 4- /reference trajectory LWR-U5 once-through strategy (consumptive use) classical reactor strategy (intensive use) 25 cumulative nat. uranium demand in106t 1990 2000 2010 2020 2030 FIGURE l2 The classical reactor strategy. fast breeder reactors that would become operative on a large scale by the turn of the century. There are, of course, various other possibi- lities besides this reactor strategy, but all of them require breeding. In this respect it is useful to realize that the fusion reactor of the future, based on the present design, will also be a breeder reactor. Granting the central fusion process of energy release in the plasma to be typically different from the process of nuclear fissioning, there are yet remarkable parallels between fusion and fission breeders in strategic energy planning and reactor operation: lithium in fusion corresponds to U-238 (and Th-232) in fission breeding, and tritium in fusion to plutonium (and U-233) in fission. In both cases, there are radioactive inventories and radioactive wastes. Both types of breeders today are geared to electricity generation. Lithium as well as uranium plus thorium resources are similar in size, either one yielding an energy output of about 20 kWh/g. In spite of these qualitative similar- ities, a technically mature fusion reactor could offer considerable quantitative advantages over the fission breeder. ' Technical matura- tion of fusion reactors, however, will still continue far into the twenty-first century, and no more than 2-3 TWyr/yr from fusion are to be expected in 2030. Its share will possibly increase thereafter. Table 8 attempts to summarize the world's resources, indicating what the potentials as well as the constraints are in producing and using these resources. The data are rather optimistic. Much more would have to be said if time permitted. ENERGY SUPPLY STRATEGIES Where do resources come in in IIASA's set of energy models, attempting to simulate the energy demand-supply situation? Figure 5 identifies

l9 TABLE 8 Resources, Production Potentials, and Constraints Source Production (TWyr/yr) Resource (TWyr) Constraints Wood Hydro Total 2.5 l-l.5 00 00 00 Economy — environment Economy — environment Economy — (nature) 6-(l4) Oil and Gas 8-l2(?) l,000 Economy — environment — resources Coal l0-l4(7?) 2,000(7) Society — environment-economy Nuclear Burner l2 for 2020 300 Resources Breeders Fusion <l7 by 2030 2-3 by 2030 300,000 300,000 Buildup rates — resources Technology — buildup rates Solar Soft Hard l-2 2-3 by 2030 00 Economy — land — infrastructure Buildup rates — materials oo resources in the lower right oval as an input to MESSAGE. Several exten- sive LP runs of the MESSAGE program for the seven world regions lead to —within the defined context—optimal energy supply strategies. For the purposes of this presentation, the globally aggregated supply is of interest. Figure l3 shows the evolution of the primary energy mix by 2030. It is to be taken with a grain of salt, but note the slightly reducing share of gas and the overall fairly constant share of oil together with synthetic fuels, e.g., methanol. Within this band, oil is increasingly Hydro —Renewables 1975 1985 2000 2015 2030 FIGURE l3 World: low demand, primary energy or equivalent.

20 replaced by synthetic liquids after 2000. One such source is autothermal coal liquefaction, assumed to increase the need for coal. Together, such and the traditional uses of coal lead to a rather uniform overall share of coal in the primary energy market, with the traditional share decreasing steadily. This decline is offset by a rise in nuclear energy for electricity generation. Among the various nuclear shares, that of the fast breeder increases quickly after the turn of the century. The rest of the primary energies remain fairly small until 2030, but let me repeat that solar and fusion could take on greater importance later. These results relate to the Low Scenario. In terms of primary energy market shares, the mix for the High Scenario does not differ signifi- cantly. But, of course, what we are after above all in this context is the absolute contributions of the various primary energies in the year 2030. They are listed in Table 9. Indeed, oil production does not seem to decrease at all. It appears rather high, providing about 6.83 TWyr/yr in 2030 in the High Scenario. So does gas production, contributing 4 times the value of today. All nuclear energy production, of nonbreeders (nuclear l) as well as breeders (nuclear 2), amounts to 8.l TWyr/yr, that is about 23% of the total energy supply in the High Scenario, and yet the number is far from the theoretical l7-TW potential. Most remarkably, these primary energies are topped by a coal production of almost l2 TWyr/yr, or about l3 billion tons of coal equivalent per year. Solar, which is just about 0.5 TWyr/yr in 2030, is a more or less ad hoc input to MESSAGE since the program rejected solar contributions at the estimated cost levels. Solar may at best be 2 TWyr/yr perhaps, but surely not more. Hydropower, too, may figure higher in 2030 than indi- cated by, say, 50%. In relative terms, the contributions of both solar and hydro appear rather small; absolutely speaking they are enormous, and even more so are the absolute numbers for the other primary energies. One may wish to react to these incredible quantities by reducing energy demand to values lower than in the Low Scenario. But, in doing so, one is brought to face the problems that were discussed under the heading of a l6-TW scenario. In particular, a clarification would be needed as to the regions and the manner and the extent in which the energy demand should be reduced. The present scenario approach offers the advantage of confronting us directly with the huge orders of magnitude that are required. This is unlike a national approach, which may allow one to escape into imports if futures appear too dim. When treating the world as a whole, as is done here, one must explicitly specify where imports for certain world regions could originate. As far as hopes for oil are concerned, extrac- tion must be assumed to materialize somewhere in the world. National difficulties can no more be dismissed or exported to the abstract level of the global market, given a worldwide perspective. Indeed, questions of oil import were very important for our regional calculations. Thus Region VI, largely though not fully identical with OPEC, appears to play a dominant role still by 2030. It cannot be ex- pected that this region will endeavor transferring its wealth of oil into inflationary capital, as would be the case if it simply complied with import requests as they are received. Rather, it is sensible to

2l TABLE 9 Two Supply Scenarios, Global Primary Energy: l975-2030 (TW) Primary Source l975 High Scenario 2000 2030 Low Scenario 2000 2030 Oil 3.62 5.89 6.83 4.75 5.02 Gas l.51 3.ll 5.97 2.53 3.47 Coal 2.26 4.95 ll.98 3.93 6.45 Nuclear l 0.l2 l.70 3.2l l.27 l.89 Nuclear 2 0 0.04 4.88 0.02 3.28 Hydro 0.50 0.83 l.46 0.83 l.46 Solar 0 0.l0 0.49 0.09 0.30 Other 0.2l 0.22 0.8l 0.l7 0.52 TOTAL 8.2l l6.84 35.65 l3.59 22.39 expect a limit on the region's oil production, here assumed to be 33 million barrels a day. The block diagram in Figure l4 illustrates the oil export-import situ- ation for all regions in l975 and 2030, according to the High Scenario. One GWyr/yr, the unit given, corresponds to about l4,000 barrels a day. Region IV (South America) and Region V (South East Asia and Africa) also were exporters in l975, supplying Region I (North America) and Region III l lll 1200 lV Vl 400 1975 REGlON lMPORTS 100 90 700 2030 "HlGH' EXPORTS 1400 lMPORTS EXPORTS 1500 FIGURE l4 Oil trading regions, l975 and 2030 (GWyr/yr).

22 (Western Europe, Japan, and Australia). In 2030, the High Scenario indicates for Region I oil self-supply and thus no need for import, in spite of the absolute increase in the region's energy demand; and for Region III, a reduction in imports, largely on account of autothermal coal liquefaction. The remainder should help alleviate the most severe energy needs of Region V. However, it is plain to see that Regions I and III will, on account of their purchasing power, still import more than is foreseen by the scenario, at the disadvantage of the developing Region V. One understands from the above that the main factors determining resource allocation and trade flows are first of all production ceilings and the possibility of increases in the production rate. The resources themselves are not actually exhausted by 2030. Compare the cumulative oil consumption by 2030 in Table l0: it is 68% in terms of Price Cate- gories I and II, but only l% of highest-cost oil, which is oil shales and tar sands. For natural gas the ratios are 49% and 0%, and for coal, 6l% and 0%. In other words, out of a world total of about 3,000 TWyr of resources only about 900 TWyr will have been used by 2030. This is, of course, the relatively cheap and clean resources, having less impact on the environment than others. With an annual requirement of about 40 TWyr, for example, the rest of about 2,000 TWyr would last for another 50 years. After what we know today and what we can now anticipate for the future, the main constraint of the coming five decades will be production ceilings, with resources constraints coming to bear in the following half century. By that time a transition to nuclear and solar will be inevitable. This exercise in how supply schemes affect resource allocation demon- strates the usefulness of the scenario approach, by which anticipated TABLE l0 Cumulative Uses of Fossil Fuels, l975 to 2030, High Scenario Total Resource Total Consumed Available (TWyr) TWyr % Oil Conventional 464 3l7 68 (Cat. I+II) Unconventional 373 4 l (Cat. Ill) Natural gas Conventional 408 l99 49 (Cat. I+ II) Unconventional l30 0 0 Coal Cat. I 560 34l 6l Cat. II l,0l9 0 0

23 events are put into a chronologically meaningful order. Still, one should remember that it is scenarios we are dealing with here, and not predictions. COAL IN EUROPE As was shown in Table 9, the highest absolute shares in 2030 in both scenarios are those of coal production. They are worth looking into in more detail, given the present selling difficulties in coal and one at best short-term trend toward coal use, other than burning, for power generation. But at IIASA, short-term and long-term are put into differ- ent boxes, which, in the context of coal use, means that one needs a long-term perspective to clarify matters. Although quantification is difficult, it is possible to derive to this end actual technological implications from the scenarios. Calculations have been made for all regions, and in particular for Western Europe in Region III. It appears that Western Europe does not have enough indigenous coal to meet an earmarked requirement of l4,000 million tce in 2030. Thus it would have to drive coal mining to the extreme of, say, about 500 million tce, and import the rest. This, in principle, could come from the United States, but means that they would have to mine 2,000 million tce for their own needs plus 900 million tce for Europe! In short, the insights gained from these calculations make coal a scarce resource after the turn of the century. It figures high in world trade and is likely to be processed by various new technologies, in order to substitute oil as a liquid secondary energy carrier. Autothermal liquefaction, among the various coal conversion processes, is a process by which carbon atoms are transformed into hydrocarbons, such as methanol. A greater conversion efficiency than the present 25%-29% would be desirable but requires, e.g., exogenous addition of large amounts of hydrogen (see Figure l5). Such an advanced process would require only one-third of carbon needed in present autothermal processes, and the energy content of the resulting methanol would at equal parts be derived from hydrogen and carbon. This throws new light on the possible coupling of methanol production with nuclear or/and solar, serving to produce electrolytic hydrogen, for example. The overall constraining factor for such systems would be capital cost. One cannot conclude a discussion on coal without touching on the CO2 problem. The natural CO2 content of the world's atmosphere compares to the release from burning about 500 TWyr of coal. With a fossil energy production of about 900 TWyr by 2030, suggested by the High Scenario, a doubling of the atmospheric CO2 content has to be expected, and an impact on the climate lasting longer than for centuries. Climatologists agree that an uneven warming of the earth's atmosphere would result, with a minor variation of the average, but large changes (about l0°C) in polar areas and partial melting of the polar ice caps. At the present state of the art, there is no conception of what this will mean for actual climate patterns, nor how certain the development is to occur: experts

24 CARBON SUPPLY (COAL) + 1/2 02 26 4 kcal METHANOL SYNTHESlS: j HYDROGEN SUPPLY I A) FOSSlL ROUTE (COAL) + 1/2 02 -«. CO C0 + H20 - B) NON FOSSlL ROUTE catal. 3/2 H20 — 3/4 02 21.7kcal + 26.4 kcal -9.6 kcal -102 kcal CARBON EFFlClENClES: A) FOSSlL ROUTE - n (COAL) + (n 1) 02 + H20 ii:3.5-:-4 CH3OH ^ (n -1) C02 + WASTE HEAT B) CONSERVATlON 1 (COAL) + NONFOSSlL * 3/2 H20 -» CH3OH + 1/402 + WASTE HEAT ROUTE ENERGY FIGURE l5 Methanol production routes. can only appeal to decisionmakers for highly flexible global energy strategies.8 The prudent use of the carbon atom as discussed above may serve as an example. CONCLUDING REMARKS The foregoing considerations were meant to bring light into the complex interplay of medium- and long-term facets of the energy problem. It appears that, if due regard is given to both types of aspects, it is well possible to point out ways for remedying the energy situation. The problems involved are only partially a matter of substance and can be largely overcome if political and economic measures are guided by prudence and willpower. The considerations discussed are founded on a 5-year study of IIASA's Energy Systems Program, with contributions from scientists from the USA and USSR and l5 other countries in East and West. The study is being documented in a l,000-page volume on "Energy in a Finite World—A Global Systems Analysis," which is due to appear at the beginning of l980. REFERENCES Marchetti, C., Nakicenovic, N., Peterka, V., and Fleck, F., The Dynamics of Energy Systems and the Logistic Substitution Model, vols. l and 2, AR-78-lA/B/C, Laxenburg, Austria, International Institute for Applied Systems Analysis (l978).

25 2. Keyfitz, N., Population of the World and Its Regions l975-2030, Laxenburg, Austria, International Institute for Applied Systems Analysis, forthcoming. 3. U.S. Department of Energy, Distributed Energy Systems in California's Future, interim report, vols. l and 2, HCP/P7405-0l/02, Washington, D.C., Office of Technology Impacts (l978). 4. Leontief, W., et al., The Future of the World Economy, A United Nations Study, New York, Oxford University Press (l977). 5. Hafele, W., Der Beitrag der Sonnenenergie zur Deckung des gegenwar- tigen und zukunftigen Energiebedarfs, presentation, BMWF/ASSA Symposium on Solar Energy Research on the Occasion of Austria's National Holiday, Vienna, ASSA Information Service (l978). 6. Kulcinski, G. L., Kessler, G., Holdren, J. , and Hafele, W., Energy for the Long Run: Fission or Fusion?, American Scientist, 67(1), pp. 78-89 (l979). 7. Hafele, W., Holdren, J. P., Kessler, G., and Kulcinski, G. L., Fusion and Fast Breeder Reactors, RR-88-8, Laxenburg, Austria, International Institute for Applied Systems Analysis (l976). 8. Williams, J. (ed.), Carbon Dioxide, Climate and Society, IIASA Pro- ceedings Series Environment, CP-78-5, London, Pergamon Press (l978).