A Critical Appraisal of the IIASA Energy Scenarios

NOT FOR QUOTATION WITHOUT PERMISSION OF THE AUTHOR

A CRJTICAL APPRAISAL OF THE

IIASA ENERGY ~CEWIOS

Bill Keepin

October 1983 WP-83-104

Working Papers a r e interim reports on work of t h e International Institute for Applied Systems Analysis a n d have received only limited review. Views or opinions expressed herein do not necessarily r e p r e s e n t those of t h e Institute or of i t s National Member Organiza- tions.

INTERNATIONAL INSTITUTE FOR APPLIED SYSTEMS ANALYSIS 2361 Laxenburg, Austria

PREFACE

The energy developments of t h e last decade stimulated many scientific studies of t h e global energy system and its possible future evolution. One of the most extensive of t h e s e studies was carried o u t by IIASA's Energy Systems Program between 1973 a n d 1980, culminating in t h e final report, E n e r g y in a E n i t e World. An important aspect of t h e IIASA work involved t h e development of mathematical models for t h e purpose of analyzing possible transitions from t h e present dependence on fossil fuels to future sustainable energy systems.

In 1981 I came t o IIASA t o study t h e energy models developed here, focus- ing in particular on t h e i r impressive application t o t h e global energy system published in E n e r g y in a R n i t e World. However, as t h e work progressed, I came across a number of troubling aspects t h a t eventually led m e t o t e r - minate t h e work I was doing a n d investigate further. This paper is the result of t h a t investigation, and I offer it in t h e hope t h a t it will contribute t o main- taining standards of high quality in future scientific work.

Many persons have helped m e a great deal in this work, only a few of whom can be mentioned here. I owe t h e greatest debt to Valerie Jones, who provided tremendous support, encouragement, and m u c h needed assistance.

In addition, I a m grateful t o Brian Wynne and Mike Thompson for many hours of discussion and general encouragement. 1 wish t o t h a n k Rhonda Starnes and Bonnie Riley for carefully preparing t h e manuscript, and my sis- t e r Mavis for painstakingly proofreading t h e tables.

Bill Keepin

CONTENTS

1. INTRODUCTION

2. DESCRIPTION OF THE IIASA ENERGY SCENARIOS 2.1. The Model Loop

3. ANALYSIS OF THE IIASA ENERGY MODELS 3.1. The Analytic Approach

3.2. Primary Energy

Construction of t h e scenariette for oil Comparison of scenariette with scenario Scenariettes for other energy resources 3.3. Secondary Energy

Scenariette for electricity generation Comparison of scenariette with scenario Scenariette for liquid fuel supply

3.4. Scenariettes for Other Regions 3.5. Conclusions

4. ROBUSTNESS OF THE IIASA ENERGY MODELS 4.1. Sensitivity t o Estimates of

Inexpensive Uranium

4.2. Sensitivity with Respect t o Relative Cost Structure

4.3. Documentation of Sensitivity Analysis 4.4. Conclusions

5. DISCUSSION OF THE MODELS 5.1 Iteration in MMI

5.2. Small Feasible Region 5.3. New Representation of MMI

6. CONCLUSIONS

APPENDIX A: ESTIMATED RESEARCH EFFORT

APPENDIX B: SCENARIETTES FOR PRIMARY FOSSIL FUELS APPENDIX C: SCENARIETTE FOR SECONDARY ENERGY SUPPLY SYSTEMS

APPENDIX D: SCENARIETTES FOR OTHER REGIONS APPENDIX E: SENSITIVITY ANALYSIS AND

URANIUM SCENARIETTE

APPENDIX F: EARLY SENSITIVITY STUDIES REFERENCES

LISI' OF TERMS AND ABBR!WM"I'ONS

ACT Mi DOGR

FBR GFS GT GW H-M hydro I1 ASA IMPACT lSi k t LFP

LP

LWR MEDEE-2 MESSAGE MMI MW P CT PETG region I through VII s c e n a r i e t t e scenario STEC t

TW

Advanced coal technology (fluidized bed)

Alternative scenario (subscript indicates a particular sensi- tivity t e s t )

Documentation of the Global Runs, a volume documenting t h e IlASA energy models (ESP, 1982)

E n e r g y in a Finite World, t h e final report of t h e IIASA energy study (Hafele, 1981a)

Fast breeder r e a c t o r Gas fired steam Gas t u r b i n e

Gigawatt (10' watts)

Hiifele-Manne model (forerunner of MESSAGE) Hydroelectric power

International I n s t i t u t e for Applied Systems Analysis Economic impacts model (input/output)

Original lIASA scenario (subscript indicates a particular sensi- tivity t e s t )

Kilotonnes ( t h o u s a n d m e t r i c tonnes) Liquid fuel power

Linear programming Light water r e a c t o r Energy d e m a n d model Energy supply model (LP)

The s e t of IIASA energy models (VEIIEE-2, MESSAGE, IMPACT) Megawatt ( 10' watts)

P r e s e n t coal technology (with limestone scrubber) Petroleum and gas

Partition of t h e world into seven regions (Figure 1)

Projection of f u t u r e energy system, obtained directly from assumed exogenous i n p u t s to MMI

Projection of f u t u r e energy system, obtained as final ~ u t p u t s from MMI (and published i n EIFV)

Solar t h e r r , ~ a l electric conversion Metric tonne

Terawatt (1012 watts)

Our interest in conclusions has been so great t h a t the method of reaching them has been neglected: it mattered little how much pre- judice or blind acceptance of authority was connected with them, so long as they were understood and remembered.

- ^F.M.

McMurry, 1909

This paper p r e s e n t s some disturbing findings about one aspect of a major scientific study of t h e world's energy system. The final r e p o r t of t h e seven- year study was published in 1981, entitled Energy in a f i n i t e World. Although t h e s t u d y claims t o provide a n objective, factual analysis for political decision making, some of t h e major conclusions a r e n o t scientifically justified. Princi- pal r e s u l t s include detailed projections of t h e world's e n e r g y supply s y s t e m s for t h e coming half-century. These were produced from a n apparently sophis- t i c a t e d s e t of iterative c o m p u t e r models. However, t h e models a r e found t o be largely trivial, because t h e i r final o u t p u t s a r e nearly identical t o t h e i r inputs, which a r e a r b i t r a r y , u n s u b s t a n t i a t e d assumptions. F u r t h e r m o r e , despite claims of robustness, t h e energy supply projections a r e found t o be highly sen- sitive t o m i n o r variations i n d a t a t h a t a r e well known t o be u n c e r t a i n . The sizeable contribution from t h e n u c l e a r fast breeder r e a c t o r (FBR), is d u e t o a 2% c o s t advantage t h a t is introduced 25 y e a r s from now. Since f u t u r e e n e r g y costs a r e highly u n c e r t a i n , cost-minimization l i n e a r programming models a r e unsuitable for describing robust energy supply f u t u r e s .

In addition t o t h e s e analytic findings, some a s p e c t s of t h e work a r e improperly p r e s e n t e d in t h e published documentation. In one case, t h e impor- t a n t role of t h e FBR is t r a c e d t o u n d o c u m e n t e d input d a t a . F r e q u e n t s t a t e - m e n t s t h a t t h e c o m p u t e r models formed a n iterative loop a r e contradicted elsewhere. P r e l i m i n a r y work t h a t revealed serious difficulties with r o b u s t n e s s is n o t cited, a n d s t a n d a r d sensitivity t e s t s a r e not included. Nevertheless, several "robust" conclusions have been drawn from t h e projections a n d widely publicized. One of t h e s e implies t h a t n u c l e a r power plants m u s t be built a t t h e average r a t e of o n e p l a n t every few days for t h e n e x t 50 years.

The overall conclusion in this p a p e r is t h a t t h e e n e r g y supply projections are opinion, r a t h e r t h a n credible scientific analysis, a n d t h e y therefore can- n o t be relied upon by policy m a k e r s seeking a g e n u i n e understanding of t h e energy choices for tomorrow.

A CRITICAL APPRAISAL OF THE l M A ENERGY SCENARIOS

Bill Keepin

1. INTRODUCTION

In t h e wake of t h e first oil price shock in 1973, just as "the energy prob- lem" was catapulted into t h e limelight as a major international issue, a comprehensive study of t h e global energy system was initiated a t t h e Interna- tional 'Institute for Applied Systems Analysis ('IIASA). The study lasted for more than seven years, and involved over 225 person-years of effort, with a research budget of some $6.5 million.* As described in a review of recent energy stu- dies, t h e IIASA work "is the most ambitious such study carried out thus far"

(Perry, 1982). 'In addition to the 60-odd research reports and various confer- ence proceedings t h a t were produced, t h e final report of t h e study is docu- mented in a two-volume s e t entitled Energy in a finite World (Hafele, 1981a).

The f i r s t of these (Vol. 1, 225 pages), subtitled Paths to a 3ustainable Future, is for the general reader, providing descriptions of the various aspects of t h e

See Appendix A

study a n d the associated findings. The second volume (Vol. 2, 850 pages), sub- titled A Global S y s t e m s A n a l y s i s , is t h e full technical r e p o r t which is intended for energy specialists a n d t h e i n t e r e s t e d scientific community. In addition, a 60-page E z e c u t i v e S u m m a r y (McDonald, 1981) has been widely distributed, and various magazine a r t i c l e s have been published in s u c h journals as S c i e n c e , S c i e n t i f i c A m e r i c a n , P u t u r e s , 1 k e E n e r g y J o u r n a l , etc. Although t h i s paper draws on all of t h e s e sources, t h e m o s t i m p o r t a n t r e f e r e n c e is t h e full techni- cal r e p o r t , E n e r g y in a Finite World, Volume 2. This book is hereafter abbrevi- a t e d t o EIFW.

More t h a n 140 s c i e n t i s t s c a m e t o IIASA for periods of various lengths t o participate in t h e study, including "economists, physicists, engineers, geolo- gists, m a t h e m a t i c i a n s , psychologists, a psychiatrist, a n d a n ethnologist" (EIFW, p. xvi). This multidiscipl.in.ary group c a m e from 20 different countries, encom- passing n o t only East and West, b u t developing countries a s well. As s t a t e d i n S c i e n c e , "an explicit a t t e m p t was made t o incorporate a s m a n y views and t o be as objective a s possible" (Hafele, 1980a). In addition, s o m e 34 institutions, organizations, a n d industrial firms supported or cooperated in some way with t h e project, including i n t e r n a t i o n a l organizations s u c h as t h e United Nations Environment P r o g r a m m e (UNEP) a n d t h e International Atomic Energy Agency (IAEA). Cooperating r e s e a r c h i n s t i t u t e s in t h e United States included t h e National Center for Atmospheric Research (NCAR), t h e Electric Power Research Institute (EPRI), and t h e Stanford Research Institute. F u r t h e r sup- porting and/or cooperating organizations included t h e Nuclear Research Center Karlsruhe (FRG), Volkswagen Foundation (FRG), Federal Ministry of Research and Technology (FRG), t h e Meteorological Office (UK), t h e National Coal Board. (UK), t h e Austrian National Bank, a n d t h e Siberian Power Institute (USSR).

The above information is provided to give some idea of t h e size and scope of t h e energy studies carried o u t over a period of several years a t IIASA. This is important because t h e sheer magnitude of t h e project contributes (both explicitly and implicitly) t o the authority and credibility of t h e main conclu- sions of t h e study. This paper focuses on two hypothetical "scenarios" of t h e world's energy future t h a t were developed as p a r t of t h e IIASA Energy Pro- gram. The importance of these scenarios lies in t h e fact t h a t they are t h e basis for many widely publicized conclusions drawn from t h e study.

The principal a r g u m e n t developed in this paper is t h a t t h e quantitative analysis behind t h e scenarios does not scientifically support t h e conclusions drawn from them, and t h a t these conclusions are more accurately described as opinions r a t h e r t h a n findings. There a r e two major analytical results esta- blished in this paper t h a t support this claim. First, t h e complex computer models used in t h e quantitative analysis do not play a significant role in deter- mining t h e final numerical results of t h e scenarios. Instead, t h e s e results a r e nearly duplicates of various unsubstantiated assumptions and arbitrary judg- m e n t s t h a t were supplied a s inputs t o t h e mathematical analysis. Second, t h e scenarios a r e seriously lacking in robustness with respect t o minor variations in certain input data. Although this lack of robustness was apparently recog- nized in early sensitivity studies, t h e later publications and final reports do not cite t h e early sensitivity work, nor do t h e y include standard sensitivity analyses.

This study focuses only on t h e quantitative scenarios themselves, which constitute just one aspect of t h e IIASA energy study a s a whole. Because this paper develops a strongly critical point of view with respect t o this particular aspect of t h e study, some very important caveats m u s t be clearly understood from t h e outset. First of all, many, if not most of t h e 140 scientists who parti-

cipated in the study had little or no direct involvement with t h e formulation of the scenarios o r t h e conclusions drawn from t h e m . In fact, a good n u m b e r of these participants disagree (some very strongly) with t h e methods used t o develop t h e scenarios and/or t h e conclusions drawn from t h e m . In addition, many of those who did work with t h e scenarios were involved in aspects t h a t a r e totally unrelated t o t h e r e s u l t s presented here.* Finally, m u c h of t h e IIASA energy work was unrelated to, or only distantly connected with t h e scenarios (e.g. t h e logistic substitution model).

Thus i t c a n n c t be overstated t h a t t h i s paper addresses only one aspect of t h e IIASA energy study, a n d i t is definitely n o t a general criticism of t h e e n t i r e program. In fact, the program contributed i n many i m p o r t a n t ways to a m o r e complete understanding of many aspects of t h e global energy system. I t was t h e first serious a t t e m p t to systematically account for, a n d g a t h e r consistent d a t a from all regions of t h e world with roughly equal emphasis. Given t h e mag- nitude and complexity of t h e global energy system, t h i s was n o simple task. A genuine a t t e m p t was m a d e t o properly incorporate all nations on e a r t h , which required painstaking analysis a n d aggregation of m a s s e s of detailed economic, geographical, demographic, and resource d a t a from countless sources. Furth- ermore, a g r e a t deal of effort went towards studying t h e global potential of each major source of energy. In addition, t h e program produced some very significant contributions, such a s t h e outstanding empirical results obtained by Marchetti and Nakicenovic (1979) with t h e logistic substitution model.

Finally, perhaps t h e m o s t important contribution h a s been t h e innumerable personal a n d working relationships, interactions, a n d c!ontacts t h a t developed a t IIASA, a n d as a r e s u l t of t h e many conferences and workshops t h a t were held. Indeed, t h e international setting a n d t h e many different c u l t u r e s t h a t

Examples are the U S A . work on carbon dioxide and solar energy.

were represented provided a richly stimulating and highly challenging environment in which to conduct a major scientific research program on a topic a s broad and politically charged a s t h e world's energy system.

I t is difficult to assess how much of t h e energy work a t IIASA was devoted to the development of quantitative scenarios and t h e analysis behind them.

According to EIFW, "the work took more than two years of intensive effort"

(p. 391). In any case, the scenarios are unquestionably a crucial aspect of t h e study, as revealed by the emphasis they a r e given in the published documen- tation. Of the 850 pages in EIF'IY, 300 are devoted t o the scenarios. In addition, half of the Executive S u m m a r y is focused on t h e m , and a 570-page volume is available which is entirely devoted to the mathematical models used to pro- duce t h e scenarios (ESP, 1982). Finally, t h e scenarios a r e t h e principal focus of "speech upon speech" (Hafele, 1983a) as well as magazine articles summar- izing t h e IlASA energy study published in Science, Scientific American, and F W u r e s . The scenarios a r e not intended as predictions, but r a t h e r as "indica- tors"; nevertheless, several "robust conclusions" (Hafele, 1983b) or "robust observations" (Hiifele, 1983a) are drawn from them. This suggests t h a t t h e underlying analysis is robust with respect t o uncertainties in the many found- ing assumptions, and t h a t a broad range of plausible energy futures is encom- passed by t h e scenarios.

The analysis in t h i s paper does not assess t h e realism or implications of most of the basic assurrlptions in t h e scenarios (such as t h e economic growth assumptions). In addition, this paper does n o t take a stand for or against any particular energy policy, especially a s regards controversial m a t t e r s such as the future role of nuclear or solar energy. Rather, the purpose of this work is to assess the scientific integrity of the analysis behind the IIASA energy scenarios. An earlier critique explored t h e significance of many of t h e basic

assumptions and methods (Lovins, 1981). Another critique focused on the energy models themselves (Meadows, 1981), but t h e analyst did not have access t o t h e detailed documentation t h a t is now available.

Section 2 provides ^s. brief description of t h e IIASA energy models as represented in t h e documentation. Section 3 explores t h e role of t h e s e models in generating t h e scenarios from the input assumptions, and t h e prin- cipal finding is t h a t t h e models a r e largely superfluous. This is followed in Section 4 by an investigation of t h e sensitivity of t h e scenarios to certain input data t h a t a r e known to be uncertain, and t h e finding is t h a t t h e scenarios a r e inherently unstable with respect to small variations in t h e s e data. These results a r e t h e n partly explained and clarified in a general discus- sion of t h e models presented in Section 5, which is followed by t h e conclusions in Section 6. Finally, a comprehensive s e t of appendices is included. These a r e specifically intended t o provide sufficient documentation for t h e r e a d e r t o reproduce t h e r e s u l t s presented a n d discussed in t h e text. Thus, although t h e computations a r e not difficult, some of t h e appendices a r e long and often tedi- ous, but this could not be avoided.

2. DESCRIPTION OF THE IIASA ENERGY SCENARIOS

This section provides a brief description of t h e IIASA energy scenarios and the mathematical models t h a t were used to generate them. The information presented here is drawn from several sources, including a 570-page document entitled m e IIASA Set of h e r g y Models: D o c u m e n t a t i o n of t h e Global Rum

( E S P 1982), which contains innumerable details concerning t h e models and

numerical data. This volume is hereafter abbreviated t o DOGR.

The overall purpose of the IIASA energy study was "to understand t h e fac- tual basis of t h e energy problem, t h a t is, to identify t h e facts and conditions for any energy policy" (Hafele, 1980a). This was done in a n a t t e m p t "to pro- vide decision- and policy makers with the information they need to make stra- tegic choices" (EIFW, p. 800). The principal means for doing this was via quan- titative analysis in the form of detailed scenarios describing how t h e global energy system might evolve over the next 50 years. "For our quantitative analysis, we had t o be realistic and pragmatic; otherwise we would n o t have been able to achieve t h e factual basis on which to consider possible longer t e r m solutions" (EIFW, p.xiv).

Of course the f u t u r e is uncertain, and therefore two scenarios were developed: a "high" scenario, which assumes high economic growth, corresponding to high energy consumption; and a "low" scenario, which presumes somewhat restrained economic growth, resulting in lower energy consumption. As described in EIFW, "Two scenarios ( t h e High and t h e Low) a r e constructed a s a means of spanning t h e conceivable evolutions of global energy systems over t h e next 50 years" (p.565). The scenarios a r e not intended to be forecasts or predictions, but r a t h e r to be comprehensive and interna1l.y consistent analyses from which "robust conclusions" (Hbfele, 1983a) about the world's energy future may be drawn and communicated to policy

makers and energy specialists.

The IIASA energy scenarios were generated with a set of t h r e e computer models for t h e demand, supply, a n d capital investment sectors of t h e global energy system. For this purpose t h e world was divided into seven regions (labeled I through VII) as shown in Figure 1, and scenarios were developed for each region individually. In each case, "high" and "low" scenarios were developed for each region, making a total of 14 regional scenarios. The indivi- dual results from t h e seven regions were t h e n aggregated to yield high and low scenarios for t h e entire globe. International trade of resources such as Mid- East oil, was handled on an interregional basis.

2.1. The Model h o p

The s e t of mathematical models a n d related procedures t h a t were used t o develop t h e IIASA regional scenarios a r e illustrated schematically in Figure 2.

This figure, which has been widely publicized, is a duplicate of Figure 13-1 of EIFW (p. 401). The formal mathematical models a r e designated by boxes with heavy borders in t h e figure, and t h e "assumptions, judgments, a n d manual calculations" a r e indicated by ovals with t h i n n e r borders. The major flows of information a r e indicated in t h e figure by solid arrows for direct flows, and dashed arrows for feedback Flows. Note t h a t this flow of information circulates in a clockwise fashion, which is why this is called a model loop. This is impor- t a n t , because i t m e a n s t h a t t h e t h r e e models a r e not just used in simple suc- cession, but r a t h e r they are used i t e r a t i v e l y , with the flow of information cir- culating around and around until an internally consistent scenario is obtained. This model loop is applied to each world region separately, with t h e globally unifying element being t h e manual procedure for "Interregional Energy Trade". Note t h a t t h e t e r m "scenario", as it is used here, does not sim- ply m e a n a hypothetical conjecture about t h e future. Rather, it refers t o t h e

Region I ( N A ) N o r t h America

Region II ( S U I E E ) Soviet U n i o n and Eastern E u r o p e

Region Ill ( W E I J A N Z ) Western Europe, Japan, Australia, N e w Zealand, S. Africa, a n d Israel

Region I V

...

( L A ) L a t i n America

Region V ( A f I S E A ) A f r i c a (except N o r t h e r n A f r i c a a n d S. Africa), S o u t h a n d Southeast Asia

m]

Region V I ( M E I N A f ) M i t l d l e East and N o r t h e r n A f r i c a

a

Region V I I ( C I C P A ) China and Centrally Planned Asian Economies

FIGURE 1 The ILASA world regions (reproduced from Figure 1 -3 in EIFV, p. 1 1).

final quantitative r e s u l t s of a comprehensive mathematical analysis.

Only a brief description of t h e IIASA energy models will be given h e r e - for f u r t h e r detail t h e r e a d e r is referred t o EIFW a n d DOGR. "Logically, t h e description of a loop of consistent subscerlarios could s e t o u t with a n y of i t s parts" (DOGR, p. viii). Thus t h e description of t h e model s e t can begin any- where - I begin with t h e energy consumption model, MEDEE-2. This is a s t a t i c accounting model which combines basic assumptions about population and economic growth with a large array of assumptions about lifestyles, require-

Scenarios Dtfinition

r---

(economic. popu.

I

r

lation growth)

I

^!

I ^I _I

I

^I Lifestyles.

I I Consumption

I I MEDEE - 2 Ttchnical Efficiencies

"In

!

Theorv" ^I

Secondary Fuel Mix and Substitutions

Maximum Bui1d.u~

Rates. Costs

S

Rerourcer

Production Limits

) P

for each world region - - - - - - - - - - - - -

Interregional Entrpy Trade

(-)

Assumptions, judgrne~~ts, manual calculations

1-

1

Formal mathematical models

-

Direct flow of information (only major flows shown)

----

Feedback flow of information (only major flows shown)

FIGURE 2 The widely publicized representation of t h e IIASA s e t of energy models, abbreviated MMI in t h e text (reproduced from EIFW, Figure 13-1). The capital l e t t e r s (D

= demand; S

=

secondary; P = primary; I = imports) and t h e words "In Theory" have been added as discussed i n t h e text.

m e n t s for energy services, technical efficiencies of energy-using devices. etc., t o produce profiles of final energy demand from 1980 t o 2030. In all. several thousand coefficients and p a r a m e t e r s are required for t h e full specification of t h e 14 regional scenarios. The major output is a t i m e series projection of final energy demand by s e c t o r and fuel type. Note t h a t this demand is not t h e stan- dard "demand curve" from economics, but r a t h e r a projection of future requirements for energy as a function of time. This is t h e n converted t o a demand for secondary energy (also a time series), the principal components of which a r e requirements for electricity and liquid fuels. This secondary energy demand is then furnished as a n input t o MESSAGE, t h e energy supply model.

MESSAGE is a dynamic linear programming model t h a t minimizes t h e total discounted cost of fulfilling a given secondary demand, subject t o a variety of constraints on resources a n d technologies. Thus, under several exogenous assumptions about availability of resources, costs a n d build up of technolo- gies, e t c , MESSAGE c o m p u t e s t h e optimal (i.e., least-cost) energy supply s t r a - tegy for t h e n e x t 50 years t h a t fulfills t h e energy demand specified by MEDEE-2. Notice t h a t t h i s is n o t a n economic equilibrium model; in t h e lIASA study, t h e t e r m s demand a n d supply refer t o t h e consumption and production of energy, respectively, as functions of time. Each r u n of MESSAGE requires t h e specification of s o m e 1600 constraint variables a n d 2600 activity variables (Meadows, 1981), although many of these a r e simply zero, or constant across different regions a n d scenarios (Basile, 1981). The outputs from MESSAGE include t h e marginal costs (shadow prices) of supplying secondary energy, which a r e fed back to MEDEE-2, resulting in a sub-loop iteration t h a t adjusts supply a n d demand. The major o u t p u t s from MESSAGE a r e t h e n fed into IMPACT, t h e economic model.

IMPACT is a dynamic input-output model which assesses t h e overall economic consequences of t h e energy strategy spelled out by MESSAGE. Specif- ically, t h e model c a l c u l a t e s t h e d i r e c t a n d indirect r e q u i r e m e n t s for capital investment, land, water, materials, manpower, equipment, a n d additional energy. These variables a r e t h e n fed back t o modify t h e original assumptions about t h e overall development of t h e economy: "after a first round of model r u n s , t h e built-in feedback m e c h a n i s m changed t h e original assumptions so t h e r e is n o real 'beginning' of t h e model loop" (Schrattenholzer, 1981). 'I'hus,

"the main model loop is closed with IMPACT" (DOGR, p. ix), a n d t h e resulting updated economic growth assumptions a r e supplied t o MEDEE-2, leading t o c o r r e c t e d e s t i m a t e s of final energy demand (Kononov a n d Por, 1979, Figure 1).

The flow of information h a s now r e t u r n e d t o t h e original s t a r t i n g point, com- pleting t h e description of one full iteration of t h e m a i n model loop. The e n t i r e process is now repeated several t i m e s until a n internally c o n s i s t e n t s c e n a r i o is obtained. Since t h e r e a r e t h r e e models in t h e loop, e a c h of which addresses a different facet of t h e energy system, a balanced s c e n a r i o is expected from t h i s process, as t h e o u t p u t s from e a c h model a r e adjusted a n d c o r r e c t e d by t h e o t h e r two models.

As explained in EIFW, t h i s procedure is not y e t fully s t r e a m l i n e d a n d com- puterized - most of t h e feedbacks a r e m a n u a l a n d t h e interfaces between t h e models a r e not completely formalized, leaving room for "judgmental interven- tions" a t various stages. But t h i s does not weaken t h e formalized iterative process itself (Hafele, 1980b).

As

s t a t e d in EIFW, " t h e flow of information is mechanized" (p. 400), a n d t h e streamlining i s c u r r e n t l y in t h e process of being developed (Hafele, 1982).

In s u m m a r y , "the global High a n d Low scenarios a r e t h e r e s u l t s of apply- ing t h e model loop iteratively until satisfactory consistency was achieved"

(DOGR, p. x), which in t u r n "required several iterations of t h e model set"

(McDonald, 1981).

3. ANALYSIS

OF

THE M AENERGY MODELS

Models should be designed for gaining insight a n d u n d e r s t a n h n g

...

-

E n e r g y in a finite World, Vol. 2, p. 399

In this section, a r a t h e r disturbing r e s u l t is established. Starting with t h e input assumptions to t h e IIASA energy models, a greatly over-simplified analysis of f u t u r e e n e r g y supply is c a r r i e d out (using only a h a n d calculator).

Although this paper-and-pencil analysis entails n o equations or dynamic processes, it t u r n s out t o reproduce t h e JIASA energy supply scenarios almost exactly. The unavoidable coriclusion is t h a t t h e major dynamic r e s u l t s of t h e scenarios a r e essentially prescribed in t h e input assumptions themselves, a n d t h e apparently extensive analysis performed by t h e models is equivalent t o a back-of-the-envelope calculation. In fact, in m a n y cases, t h e energy models serve a s a simple identity transformatiori from t h e inputs to t h e outputs.

3.1. The Analytic Approach

We begin t h e analysis by giving thought t o which r e s u l t s from t h e IIASA energy scenarios a r e m o s t i m p o r t a n t . Recall t h a t t h e t i m e scale for t h e IIASA study is 50 years; with a t i m e span of this length, t h e m o s t o n e c a n hope for from any model is t o discern m a j o r dynamic behavior p a t t e r n s , a n d possibly t h e i r interrelationships. For this reason we will n o t consider m o s t of t h e innu- m e r a b l e details contained in t h e scenarios. Instead, we c o n c e n t r a t e on major dynamic variables. In particular, we will r e s t r i c t o u r a t t e n t i o n t o prirnary a n d secondary energy flows ( a n d t h e i r costs), since t h e s e a r e t h e principal focus of MMI. Thus, a p a r t f r o m e n e r g y costs, m o s t economic considerations a r e excluded f r o m this analysis, a s a r e all aspects of t h e energy s y s t e m t h a t e i t h e r played m i n o r roles (e.g., solar a n d m o s t renewable resources, conserva- tion m e a s u r e s ) , or were omit.ted a l t o g e t h e r from t h e MMI analysis (e.g., social

a n d political factors, explicit environmental considerations).* In addition, "the analytic approach adopted for energy studies a t IIASA assumes an essentially surprise-free world

-

no global-scale disasters, no sweeping scientific discoveries". (EIFW, p. 395). We shall do t h e s a m e .

The particular energy forms to be considered a r e t h e following:

Primary e n e r g y (extraction of resources): oil, coal, n a t u r a l gas, and uranium.

s c o n d a q s n e r g y : electricity generation, a n d liquid fuels.

(Note t h a t n a t u r a l gas can be placed in either category).

The analysis presented h'ere is carried o u t in g r e a t e s t detail for one par- ticular world region, comprising Western Europe, Japan, Australia, New Zea- land, a n d South Africa (called region 111 in EIFW; see Figure l). This region was chosen for several reasons, one of t h e most i m p o r t a n t being t h a t i t is t h e only region for which t h e iterative process of MMI i s described in EIFW (pp. 404-407).

In addition, t h e available data for this region a r e excellent a n d voluminous.

Finally, region 111 contains t h e homelands of virtually all t h e scientists who developed t h e demand and supply components of t h e model loop (MEDEE-2 and

MESSAGE).^

The model's s t r u c t u r e and principal assumptions a r e therefore particularly suited t o this region (and most subsequent work with t h e model h a s involved applications within region 111). Thus if t h e value of t h e model is called into question for region 111, i t i s likely t o be even less useful for t h e other six world regions. In any case, a number of r e s u l t s a r e included for o t h e r

*Some of these aspects were considered by the IIASA group out& of t h e formal MMI analysis. For example, t h e global emission and concentration of carbon dioxide t h a t might result from the scenarios was analyzed in considerable detail (see EIFW).

tMEDEE was originally developed by two French scientists for application t o France, and then later adapted for use in t h e IIASA model loop. MESSAGE is a third- or fourth- generation offspring stemming from an early linear programming model conceived by W. Hiifele (FKG) and Alan Manne (USA). Subsequent versions were developed by scien- tists from Europe and Japan (both in region 111).

world regions as well, including region I (USA a n d Canada), region V (South and Southeast Asia, a n d sub-Saharan Africa excluding South Africa), and t h e oil trading regions in aggregate.

In this analysis, t h e model loop will be t r e a t e d a s just o n e model, or black box, about which nothing is known except t h e i n p u t s , t h e outputs, a n d t h e demand for secondary energy. Thus I will n o t delve into t h e m a t h e m a t i c a l details of t h e individual models themselves.* The model loop will be r e f e r r e d t o as MMI (which s t a n d s for MEDEE-2, MESSAGE, IMPACT, a n d t h e i r various i n t e r - linkages), or else just simply as t h e model. In addition, t h e t e r m s "assump- tions" and "input assumptions" refer t o various p a r a m e t e r s , t i m e s e r i e s d a t a , cost coefficients, etc., t h a t a r e supplied as exogenous i n p u t s t o MMI. These a r e indicated in Figure 2 by t h e ovals labeled P ( p r i m a r y ) , S (secondary), a n d I (imports). The secondary e n e r g y d e m a n d is indicated by t h e oval labeled D in Figure 2. In t h e p r e s e n t analysis t h e s e endogenous demand projections a r e taken a s given; therefore, this work is focused only on t h e supply side of t h e scenarios. F ~ n a l l y , t h e outputs from MMI a r e simply t h e scenarios themselves.

The numerical d a t a used i n this analysis c o m e from t h e following sources.

The i n p u t assumptions and t h e secondary e n e r g y d e m a n d a r e taken f r o m DOGR (see Appendix B) a n d t h e s c e n a r i o r e s u l t s themselves c o m e directly from t h e final c o m p u t e r p r i n t o u t s of t h e IIASA global energy scenarios, available f r o m t h e IIASA Energy Group. See Appendices B, C, a n d E for examples of t h e numerical data.

We begin t h e analysis by exploring t h e specific role t h a t t h e model (MMI) played in calculating t h e scenarios f r o m t h e assumptions. For t h i s purpose,

*Some general considerations will be discussed i n Section 5 which will help to explain why t h e models behave a s t h e y do.

we will s t u d y t h e relationship between t h e model outputs (scenarios) and t h e model inputs (assumptions). The idea is t o s t a r t with t h e assumptions and to use t h e m t o g e n e r a t e rough approzi7nations of t h e scenarios. Then, by com- paring t h e s e approximations with t h e actual scenarios, we should get some idea of t h e effect of the model's calculat,ions and i t e r a t i o n s in producing t h e scenarios. Thus, in a sense, t h e input assumptions will be distilled from t h e model in order t o expose t h e dynamic role of t h e model itself.

To t h i s end, we s t a r t with c e r t a i n input assumptions a n d proceed in a heuristic m a n n e r , combining t h e m in a simple a n d obvicus way. This will pro- duce a c r u d e zeroth-order scenario which is based purely on selected i n p u t assumptions. The criterion for selection will usually be cost minimization, meaning t h a t an unrefined form of optimization is involved. However, no equations will be solved, no dynamics will be simulated, no iteration will be performed, a n d no significant calculations or consistency checks will be car- ried out. Instead, a straightforward analysis will be performed by intuitively selecting what s e e m t o be t h e most important i n p u t assumptions a n d putting t h e m t o g e t h e r in a n a t u r a l way. In m o s t cases, t h e analysis will simply a m o u n t t o plotting a few curves on t h e s a m e g r a p h (where t h e curves t o be plotted a r e given explicitly in t h e form of input assumptions t o MMI). The resulting scenario will t h e n be compared with t h e a c t u a l scenario t h a t was produced a s o u t p u t from MMI.

Throughout, this discussion, t h e t e r m scen,arin will be understood t o denote t h e published r e s u l t s t h a t were obtained by t h e IIASA Energy Group from MMI. Meanwhile, for convenience, t h e simplistic scenario obtained from t h e i n p u t assumptions will be called t h e scenariette. Note t h a t t h i s analysis is not a n a t t e m p t t o design a riew or realistic e n e r g y model; r a t h e r , t h e aim is to understand t h e effect t h a t t h e dynamic calculations a n d iterations performed

by MMI have on t h e assumptions t h a t are fed into

MMI.

This will be done by effectively viewing t h e model's output alongside its inputs; t h u s t h e scenariette is a crude sketch compiled from certain inputs to MMI, and t h e scenario is the output from MMI.

3.2. Primary Energy

There are four primary energy sources to be considered; oil, coal, n a t u r a l gas, and uranium. Since oil is a key component of t h e global energy system, it is a natural starting point. The input data to MhlI specify t h r e e separate cost categories of this resource, which together define a kind of s t e p function for t h e cost of oil. Category I is t h e least expensive, with a unit cost of $62/kWyr,*

and includes mainly conventional domestic oil, both existing reserves and those remaining t o be discovered. Category 11 ($103/kWyr) includes some addi- tional undiscovered reserves, as well as some oil from unconventional sources.

Category 111 is t h e most expensive ($129/kWyr), consisting of oil from uncon- ventional sources such as oil shales, t a r sands, offshore a n d polar oil, and oil obtained using enhanced recovery techniques. These categories and cost assumptions a r e t h e s a m e f ~ r all world regions, and each particular region is endowed with a given (assumed) amount of oil in each category. For example, region 111 has 17.48 TWyr of oil in category I, 3.3 TWyr in category 11, and 121.36 TWyr i n category 111. These figures represent t h e overall amounts of t h e s e resources t h a t are sitting in the ground a t t h e beginning of t h e 50-year time span, available for extraction. Similar cost categories exist for t h e other pri- m a r y energy resources.

*This is equivalent to approximately $12.30 per barrel (1975 U S dollars). For categories I1 and 111 the corresponding figures are f 20.40 and $25.60 per barrel, respectively.

Construction of the scenariette for oil

We now use t h i s input s t r u c t u r e to sketch a rough portrait of oil supply for region 111. Since conventional oil is t h e least expensive, we use it first. For simplicity, we will assume in t h e scenariette t h a t t h e price of this oil will not change a s i t is depleted, i.e., we assume t h a t t h e cost of crude oil from domes- tic reserves r e m a i n s constant down to t h e last drop. This is economic sacrilege, but i t is acceptable for a rough sketch, and it makes things easy: we simply go ahead and use up all t h e conventional oil first (category I), a n d only after i t has disappeared do we move on t o t h e more expensive unconventional sources. Thus, in t h e scenariette, t h e highly simplistic s t e p function (defined by t h e input cost data) is adopted as t h e nonlinear cost function for oil supply.

Since we have decided to use up t h e cheap oil first, t h e next question is how long will i t last; i.e., how quickly will t h e oil from category I be consumed?

Looking again a t t h e inputs t o MMI, we find c e r t a i n constraints (called max- i m u m resource extraction r a t e s o r "production limits" in Figure 2) t h a t limit t h e r a t e a t which domestic oil c a n be extracted during e a c h time period.

These constraints a r e supplied t o MMI in t h e form of time series data (meaning t h a t a ceiling on annual extraction is specified for each five-year time period between 1980 a n d 2030). In t h e scenariette we e x t r a c t a s m u c h domestic oil a s possible (because it is t h e cheapest source of oil, by assumption). Thus t h e assumed constraint on domestic oil extraction is simply taken to be t h e domestic oil production curve in t h e scenariette. The only thing we have to do is keep a running tab on t h e cumulative amount of oil e x t r a c t e d -when we pass t h e 17.40 TWyr m a r k (mentioned above), we have r u n out of category I oil (domestic crude), a t which time we switch (very abruptly) to category I1 oil (unconventional); for the high scenariette* t h i s happens between 2020 and +This refers to the scenariette obtained from the assumptions of the high scenario; the

"low scenariette" is analogously defined.

2025 (see Appendix B for details). We then continue in t h e s a m e fashion:

extract oil a t t h e maximum allowable r a t e until category I1 oil is exhausted, t h e n switch to category 111, and so on.

In addition t o domestic oil, t h e r e is also imported oil t o consider, so we again consult t h e input assumptions. This time we find a constraint t h a t s e t s upper limits on t h e a m o u n t of oil t h a t can be imported as a function of t i m e , and t h i s constraint is simply adopted as t h e curve for imported oil in t h e scenariette.

This t h e n completes t h e portrait of primary oil supply, which is displayed in graphical form i n Figure 3 (see Appendix B for details). To g e n e r a t e this fig- ure, t h e individual d a t a points were plotted and then connected by straight line segments t o produce curves. Note t h a t t h e curves a r e plotted cumula- tively t o illustrate t h e composition of crude oil supply and i t s evolution over t h e 50-year time horizon from 1980 to 2030. Observe t h e r a t h e r abrupt shift from category I t o category I1 oil t h a t occurs around 2020

-

this is due t o t h e oversimplihed assumptions made i n constructing t h e scenariette. These sud- den changes a r e even more pronounced in t h e corresponding s c e n a r i e t t e s for region I (Figure D.2 in Appendix D) and in t h e global oil supply t o be discussed l a t e r (Figure 13).

Comparison of scenariette w i t h scenario

Now t h a t we have completed this first p a r t of t h e scenariette, i t is interesting to compare i t with t h e results from t h e published IIASA scenario itself. To do this, we s t a r t with a duplicate of t h e graph in Figure 3, onto which t h e final scenario results a r e superimposed by plotting individual d a t a points (which corne directly from the MMI computer output listings - see Appendix R). The result of this superposition is shown i n Figure 4. Thus Figure 4 is

Imported 011

600

-.

Category 1 1 1

200 -- Domestic Oil

Category I

1-

FIGURE 3 Scenariette for crude oil supply (region 111 high). The curves displayed h e r e a r e obtained directly from t h e exogenous input assumptions t o t h e IIASA energy models.

identical t o Figure 3 except t h a t some data points have been added; t h e s e points are t h e final scenario results, which are plotted using four different shapes (circles, triangles, squares, and crosses) t o distinguish four distinct s e t s of outputs from MMI. It is important t o understand t h e format of Figure 4, because it is used throughout this section for comparing scenariette and scenario results. The main thing t o remember is t h a t t h e curves display t h e scenariette (inputs), and t h e points display the scenario (output).

In Figure 4 we see something quite surprising. The data points from t h e scenario fall almost exactly onto t h e scenariette curves. There are some minor differences for imports, but t h e s e a r e insignificant.

A brief review is called for a t this point. We s t a r t e d with a handful of input assumptions to MMI; these were used to put together a rough sketch of t h e crude oil supply in region 111. In doing so we made some unrealistic assumptions, while a t t h e same time ignoring various considerations s u c h as

600 --

Category I l l

Domestic Oil 200 ^--

Category I

-. - . .

M M I Scenario Results:

Imported Oil Domestic Oil:

A Category I Category I I

FIGURE 4 Comparison of scenariette with the IIASA sceriario results for crude oil sup- ply (region 111 high, cf. EIFW Figure 17-1 l E , p. 560). This figure is identical to Figure 3, with the addition of the data points, which are the final outputs from the IIASA energy models. Note t h e agreement between scenariette and scenario.

price elasticities, consistency, relationships with other sectors of t h e energy system, etc. The most t h a t was expected from this was a rough qualitative correspondence with t h e scenario dynamics, and yet somehow t h e scenariette developed here agrees almost perfectly with t h e scenario itself, which is sup- posed t o be the product of a careful, detailed, iterative self-consistent optimi- zation procedure. But perhaps t h i s is just an anomaly that. holds only in this one particular case. To find out, i t is necessary to investigate some further cases.

Scenariettes for other energy resources

The development of similar scenariettes for natural gas, coal, and uranium* produces the curves shown in Figures 5, 6, and 7 respectively (see

*The uranium scenariette is obtained in a somewhat different fashion from the other primary energy scenariettes; see Appendix E.

GWyrIyr

J - Scenariette

A Scenario 800 --

Imported Gas

Domestic Gas

FIGURE 5 Comparison of scenariette and scenario results for natural gas supply - re- gion 111 high (cf. EIFW, Figure 17-12E, p.568). The curves a r e inputs t o MMI, t h e points a r e outputs from MMI.

FIGURE 6 Comparison of scenariette and scenario results for coal supply - region LII high (cf. EIFW, Figure 17-14E, p.572). The curves are inputs to MMI, t h e points a r e out- puts from MMI.

G W y r I y r -

2400 -.

-

Scenariette

"

A Scenario 2000 -.

1600 -.

1 200 --

Appendix B).

400

0

Again, for comparison, t h e scenario results are shown as data points, and once

--

Domestic Coal

I

again t h e agreement is essentially perfect. No analysis of any kind was

1980 1990 2000 2010 2020 2030

A

-

Scenariette

A Scenario

J

FTGURE 7 Comparison of scenariette and scenario results for natural uranium extrac- tion

-

^region 111 high.

involved in generating Figures 5 and 6; t h e curves a r e plotted directly from t h e exogenous input listings t o MMI, and t h e points a r e plotted directly from t h e

output

listings from MMI (see Appendix B). In some ways these plots look deceptively trivial, which obscures t h e i r importance. I t is crucial t o under- stand t h a t they are not t h e r e s u l t of some curve-fitting exercise. Rather, t h e data points are the outputs from MMI, and t h e solid curves a r e t h e input assumptions to MMI. The fact t h a t they agree perfectly means t h a t , in effect, t h e scenario results are prescribed exogenously in t h e input assumptions, and t h e model itself just reproduces these assumptions.

Perhaps these findings a r e not so surprising if we consider t h a t we have looked only a t the high scenario. I t is quite possible t h a t t h e entire energy system is operating a t maximum capacity in t h e high scenario, straining every bolt as i t were, so t h a t t h e system comes right up against t h e con- straints. If so, then it is important t o look also a t t h e low scenario, where t h e strain on t h e system should be eased considerably. This is done in Appendix B, and again, essentially perfect agreement is observed between inputs and out- puts in almost all cases.

This concludes t h e discussion of primary energy. The principal finding is t h a t both stocks and flows of primary energy sources in t h e IIASA scenarios a r e effectively prescribed in t h e form of exogenous assumptions and con- straints. In t h e schematic diagram of MMI in Figure 2, most of t h e s e assump- tions a r e contained in t h e oval labeled P (for primary). Note t h a t this oval lies entirely outside t h e iterative model loop, a n d t h a t t h e r e a r e no "major feed- backs" i n t o this oval, indicating t h a t these assumptions a r e not subject t o modification. In fact, t h e model essentially performs t h e same analysis present.ed above in developing t h e scenariette.

3.3. Secondary Energy

As discussed earlier, a principal objective of MMI is to describe a n energy supply system t h a t fulfills t h e demand a t t h e lowest cost. Therefore we shall begin t h e analysis of t h e secondary energy system by considering t h e cost assurnptions for various secondary energy supply technologies. These a r e given in Table 1. which is reproduced f r o m Table 17-4 in EIFW (p.527). The capi- tal and variable costs have t h e constant values shown, for all regions and all time periods. Furthermore, these costs are identical in both t h e high and low scenari.os, even though these scenarios a r e intended to "span a sufficiently wide range i n order t o incorporate t h e unavoidable uncertainties" (EIFW

p. 425). The assumption of fixed costs is one of t h e main reasons for t h e high degree of structural uniformity exhibited in t h e high and low scenarios for all seven world regions. The final product costs increase in some cases from t h e values shown after t h e cheapest category of t h e corresponding fuel is exhausted. Although these variations in cost are minor, t h e y a r e responsible for some curious behavior t o be discussed in Section 4. For now, we present two secondary supply scenariettes; one for electricity and one for liquid fuels.

TABLE 1 Cost assumptions for major competing energy supply and conversion techno- logies (reproduced from EIF7V, p.527, Tabie 17-4).

Capitol Vorioblr Fino1 Product

Cost cost Cost

(1975$/k W ) (19755/k wyr) (1975S/k U'yr)

Electricity Generation

Coal with scrubber 550 23 154

Conventional nuclear n c t o r (eg., LWR) 700 50 136

Advanced reactor ( e . ~ , FBR) 920 5 0 143

Coal, fluidized bed 480 36 152

Hydroelectric 620 8.5 8 5

Oil fired 350 19 2 5 6 .

Gas fired 325 16 216

Gas turbine 170 '17 24 1

Solar central station 1900 28-60 297

Synthetic Fuels

Crude oil refinery 5 0 3.7 7 5

Coal gasification ("high Btu") 480 40 125

Coal liquefaction 480 40 125

As mentioned above, t h e demand projections for secondary energy in these scenariettes a r e taken from t h e endogenous "Secondary Fuel Mix and Substitutions" procedure, labeled D (for demand) in Figure 2. Thus t h e present analysis t r e a t s these projections a s given, and focuses on t h e supply side of t h e scenarios.

Scenariette for electricity generation

Given t h e objective of cost minimization, we s t a r t by looking a t t h e rela- tive cost assumptions for electricity generation. In t h e last column of Table 1, hydroelectric power is found t o be t h e l e a ~ t , expensive technology, a t $85 per kWyr. Following this, t h e next cheapest is nuclear power, running from $136 for LWR t o $143 for FBR, then comes coal-fired power a t $152 t o $154,* and t h e remaining electricity sources become increasingly more expensive. Thus we s t a r t with the cheapest source (hydro), take as much as possible, t h e n move on t o t h e next cheapest source (LWR), again taking as much as possible; and continue in this fashion until t h e demand is met. Thus, to build t h e scenariette, the technologies a r e chosen in the order of their cost, and each one contributes an a m o u n t equal t o its supply constraint. This guarantees t h a t when we reach t h e demand level, we have specified t h e least expensive supply mix t h a t m e e t s it.

This procedure for developing t h e electricity scenariettes is described in more detail in Appendix C. The end r e s u l t is a scenariette consisting of an assemblage of constraints, stacked one on top of t h e other, defining t h e evolu- tion of t h e electricity supply system. These constraints, which a r e called

"maximum build-up rates" in Figure 2, Form another group of assumptions supplied t o MMI. In most cases they are derived from t h e following difference equation (EIFW p.530)

~ t

=

Y Y L - I +

s

⁽¹⁾

where yt represents t h e annual addition t o t h e capacity of a particular tech- nology during t h e time period

t ,

y is a constant growth parameter, and g is an initial condition t h a t s t a r t s t h e process off a t t h e "start-up" time, t o . As will be

*Since these cost about the same and are both coal burning technologies, no distinction is made in t h e scenariette. It. so happens that this distinction was unimportant in the scenario as well.

seen shortly, this very simple equation, which produces exponential growth,*

is by far t h e most i m p o r t a n t lactor in determining t h e dynamics of t h e secon- dary supply mix. The p a r a m e t e r s y, g , and t o have one fixed s e t of values for all t h e developed regions (I, 11, 111). and a fixed but different s e t of values for t h e rest of t h e world (regions I V through VII).

Figure 8 &splays t h e s c e n a r i e t t e for electricity supply in region 111. The a r e a labeled "coal & other" in this figure is due almost entirely t o coal.

"Other" refers t o a thin sliver (due t o c u r r e n t oil- and gas-fired power plants) which disappears by 2010, a n d a barely discernible contribution from solar energy a f t e r 2020. The d e m a n d projection is shown in Figure 8 by a dashed 1ine.t Since t h e d e m a n d is taken as given, a dashed curve is used t o distin- guish i t from t h e solid curves, which a r e t h e results of t h e scenariette.

Comparison of scenariette w i t h scenario

Turning now t o Figure 9, we find t h a t t h e MMI scenario i s identical t o t h e s c e n a r i e t t e u p through 2010. Notice t h a t after 2015, t h e d a t a points for LWR a n d FBR s e e m t o be deflected away from t h e demand projection as t h e y approach it. During . t h e s e final 1 5 years of t h e t i m e horizon, coal is being phased o u t very rapidly, resulting in extensive underutilization of coal-fired capacity. However, MMI imposes an economic penalty for excessive underutili- zation, s o t h a t t h e rapid decline of coal is a t t e n u a t e d somewhat, producing t h e observed deflection. This s a m e effect occurs t o a l e s s e r e x t e n t within t h e nuclear coritribution itself, a s L,WR gives way t o FBR.

'Denoting the start.-up time by t o , the initial condition is Y t

=

9 . With this condition, Equation

0

(1) has the unique solution

( t - f , + J )

~ t

=

^{9 [ ~ -} l ] / ( y - I) f o r t

r

t o ,

which is exponential in t

.

The numerical values for exceed unity in ti11 cases (see DOGR).

3 It is interestine t o note in passing that this demand projection entails a 2.7 iold increase in elec- tricity consumption per person living in region III by 2030.

GWvr/vr 1100--

- -

Demand Projection

900 ^--

800 ^{- -}

700'-

600.-

Hydro 0

1980 1990 2000 2010 2020 2030

FIGURE 8 Scenariette for e l e c t r i c i t y generation -region 111 high. The solid curves a r e obtained directly from t h e exogenous i n p u t assumptions t o t h e IIASA energy models.

The d a s h e d curve is t h e endogenous d e m a n d projection.

Notice t h a t MMI has no knowledge of t h e physical significance assigned to the particular results t h a t i t produces. For example, it might be tempting to conclude from Figure 9 t h a t t h e f a s t breeder r e a c t o r (FBR) will dominate t h e future electricity supply. However, this is an assumption supplied t o t h e model, and not really a result or conclusion derived f r o m t h e model. The curve labeled "FBR in Figure 9 is t h e immediate consequence of t h r e e parameter values [y, g , t o in Equation (1) ] supplied directly t o MMI by t h e user which reflect his or her ideas about t h e future role of FBR in the electricity supply. But t h e model itself knows nothing about t h e physical interpretation attached to t h e resulting curve, nor can it in any way assess t h e feasibility, desirability, or implications of such an option. It simply displays t h e curves

Demand Projection

1 -

Scenariette

'OoO]

]

^scenario

/ / 4'

Coal & Other

Hydro

0 I : -t---t--

1980 1990 2000 2010 2020 2030

FIGURE 9 Comparison of s c e n a r i e t t e a n d scenario r e s u l t s for electricity generation -

region 111 high. This figure is identical to Figure 8, with t h e addition of t h e d a t a points, which a r e t h e o u t p u t s f r o m t h e IIASA energy models. Note t h e close a g r e e m e n t .

t h a t r e s u l t from t h e user's inputs, and as such, i t serves a s a framework for displaying whatever free-hand sketches t h e user dreams up.

I t might still be tempting t o imagine t h a t t h e low scenario will not behave quite so predictably, since t h e energy system is under considerably less strain in this case, b u t Figure 10 reveals t h a t this i s not t h e case. Once again, t h e scenario coincides with t h e scenariette for 35 years before t h e model exerts its influence.

Scenariette for liquid fuel supply

The analysis for t h e supply of liquid fuels is essentially t h e same as for electricity, so only t h e results a r e presented h e r e (see Appendix C for details).