TL: ENERGY WITHOUT OIL The Technical and Economic Feasibility of Phasing Out Global Oil Use SO: Energy Policy Unit, Greenpeace International (GP) DT: January 13, 1993 Keywords: energy policy greenpeace oil reductions research gp atmosphere facts cars forecasts alternatives / Energy Policy and Research Unit, Greenpeace International, Keizersgracht 176, 1016 DW Amsterdam, The Netherlands. INTRODUCTION Greenpeace International are shortly due to complete a major study into the technical, economic and environmental feasibility of phasing out fossil fuel use over the next century. This draws extensively on analysis carried out for Greenpeace by the Stockholm Environment Institute (SEI), Boston, and others over the past 18 months. The full report of this study is due to be completed in February 1993. Recent major oil spill disasters have brought home to the world some of the true costs of our dependence on oil. In view of the seriousness of both the current Shetland oil disaster and the insidious and sustained impact of oil use on our seas and atmosphere, the interim results of the Greenpeace study are being launched in this special report. Though the full report addresses all fossil fuel use, "Energy Without Oil", as its title suggests, focuses principally upon the feasibility and implications of phasing out global oil use. THE ENVIRONMENTAL IMPACTS OF OIL AND OTHER FOSSIL FUELS The environmental implications of the combustion of large quantities of fossil fuels, currently 6,200 million tonnes of coal, oil and gas per year, are becoming increasingly understood. The release of significant quantities of sulphur dioxide and nitrogen oxides is linked to the formation of acid rain, the destruction of ecosystems and impacts upon the human economy (WRI, 1990). Oil consumption mainly comes from the transport, industrial, electricity and petro-chemical sectors. This report deals mainly with the 80 per cent of oil used in the former three sectors. The 620 thousand barrels of oil being presently spilt in the Shetland Islands equates to 55 minutes and thirty seven seconds of oil consumption in the United States, and eight hours of oil consumption in the United Kingdom. These figures illustrate the magnitude of this country's and the world's dependence on oil. Over a third of oil consumed globally each year is burnt in the internal combustion engine. Transport's share of annual global oil consumption is over 8 billion barrels. In the twenty four countries of the OECD transport is responsible for over 50% of all the oil consumed. This is burnt in the 680 million vehicles of which 430 million are cars. Each year 50 million more vehicles are produced; a rate of over 100,000 per day or one neew car for every two new babies. The magnitude of the growth in vehicle production is illustrated by the fact that the oil spilt in the Shetlands would fuel 52,000 vehicles, the number of new vehicles produced every twelve hours of every day. There are currently 680 million vehicles on the planet, increasing at the rate of more than 30 million per year (Greenpeace, 1991). Urban smog and low-level ozone have been identified as serious pollution problems in many regions of the world, with implications for human health and ecosystems (BMJ, 1990, SCAQMD, 1989). Current emissions of carbon monoxide, hydrocarbons and NOx from the transport sector are 232, 37 and 31 million tonnes respectively. One projection suggests that under a "business-as-usual" scenario, these emissions could increase to 615, 174 and 193 million tonnes respectively by the year 2100. Each year 12 billion barrels, half the world's oil consumption, is shipped around the world. To transport this amount of oil requires the equivalent of 19,000 Braers. Oil transported around the world has increased by 11 per cent over the past ten years. The amount of oil afloat at anyone time in supertankers is 500 million barrels, the equivalent of 800 Braers. Each tanker runs the gauntlet of a Braer disaster. Between 1978 and 1992 fourteen of the world's largest spills have released over 1 million barrels into the world's oceans. The high visibility spills at sea are only a tiny proportion, some 12%, of the total amount of oil released to the oceans each year. In total some 25 million barrels are released into the marine environment each year. This is the equivalent of 40 Braers. The amount of oil spilled each year deliberately from tankers cleaning out tanks, terminal operations, dry docking, bilge and fuel oil spillage is at least 8 million barrels, the equivalent of 13 Braers In recent years, climate change has moved from the realms of a purely scientific issue, to one which has exercised the minds of politicians and entered the public domain. There have been a number of major scientific assessments of the potential for climate change over the past decade (DOE, 1985, SCOPE, 1986, US EPA, 1987, Dahlem 1988, WMO, 1988). In the past three years, a major international study under the auspices of the Inter- governmental Panel on Climate Change (IPCC, 1990a & 1992) has largely confirmed the previous findings that a continuation of current trends in fossil fuel combustion, deforestation and the release of chlorine based chemicals will likely increase global temperatures to unprecedented levels. The 23 billion barrels of oil burnt annually contribute some 31% of all carbon dioxide which humankind puts into the atmosphere each year. As of 1991 there were 1,000 billion barrels of proven global oil reserves, 44 years of reserves at current consumption rates. If this amount was burnt it would release 113 billion tonnes of carbon into the atmosphere. The climate impacts of burning these oil reserves and other fossil fuel reserves are shown below in Figures 9 and 10. FOSSIL FUELS AND THE GLOBAL ENERGY SYSTEM Fossil fuel combustion is the major source of greenhouse gas (GHG) emissions, largely in the form of carbon dioxide (CO2), accounting for 85 percent of current net anthropogenic CO2 emissions (Subak et al., 1991). Over the past 50 years, fossil fuel consumption has increased fivefold, from approximately 57 exajoules in 1937 to around 282 exajoules in 1988 (Fulkerson et al., 1990). Predominant among these fuels is the use of coal and oil, though natural gas use is projected to increase rapidly over the next few decades (IEA, 1991, Stern, 1990). Oil use has grown from approximately 16 exajoules to 130 exajoules in the same period. If fossil fuel consumption continues to grow, doubling of pre-industrial CO2 concentrations could occur as early as 2030, leading to a projected increase in global average temperature of 1.5 to 4.5xC (IPCC, 1990a & 1992). The reduction in greenhouse gas (GHG) emissions, and particularly gases such as carbon dioxide which are linked to fossil fuel combustion, provided the basis for a study commissioned by Greenpeace International (Greenpeace, 1992). THE FOSSIL FREE ENERGY STUDY The Stockholm Environmental Institute study for Greenpeace was developed with the main objective of assessing the technical, economic and policy implications of moving towards a fossil free energy system. The study considered the world in terms of ten separate regions, reflecting, to the extent possible, varying patterns of economic activity, personal consumption, energy use, and energy resources. These regions are listed in Table 1. The regional groupings were based on those used in U.S. EPA and IPCC studies. The time frame for this study extends to the year 2100. This is an end year common to several global studies. The long time frame is necessary, as the climate effects of GHG emissions are expected to lag substantially behind the emissions themselves. Given the speculative nature of such long-range scenarios, greater emphasis should be placed on the time period between now and 2030. In fact, the next 40 years present the most challenging period if we are to reverse the trend of rising emissions of carbon dioxide, and develop policies that enable humankind to meet climate stabilization targets. In addition, technical and economic estimates are far more tangible during this period. Computer Models The study utilised three computer models to assess the climate, technical, economic and policy implications of moving towards a 'fossil free' energy system over the next century. A main fossil free energy scenario (FFES) with variations was developed. The computer models were used to a) to develop climate targets which the energy modellers used as limits in order to bring climate impacts down to manageable levels; b) to assess the climate impact of the fossil free energy scenarios developed; c) to develop a disaggregated, sectoral global energy scenario which used an increasing quantity of renewable energy sources; d) to assess the pricing and overall cost implications of the scenario; and e) to inform the evolution of policies which would be needed to achieve the scenario. The three models were as follows: STUGE. The STUGE (Sea level and Temperature change Under the Greenhouse Effect) greenhouse model (Wigley, et al., 1991) was used in order to assess the implications of possible future GHG emissions, and in the development of emissions control targets. STUGE is a PC-based climate model, which is designed to simulate the future atmospheric concentrations, radiative forcing, temperature change and sea-level rise resulting from any given emissions profile over the duration of the next century. It treats the major greenhouse gases of CO2, CH4, N2O, CFC-11, CFC-12 and a halocarbon substitute HCFC-22 individually, and permits the rapid assessment of future emissions scenarios. This model was chosen because it has been used extensively by the IPCC Working Group 1 (scientific assessment) (IPCC, 1990a & 1992) for sensitivity testing of the type conducted in this study. The model employs a simple parameterised approach which is adjusted such that the results concur with IPCC figures, which are the mean responses from an extensive range of simulations utilizing highly complex general circulation models (GCMs). STUGE has a more simplified structure, and only reports global-mean values, although it should be stressed that the results are as likely to be accurate as those from any of the more complex models. The model is composed of a suite of separate algorithms. Emissions are converted to concentrations using models that agree well with the conclusions of IPCC Working Group 1 (WG1). The subsequent concentration changes are converted to radiative forcing changes. WG1 (Shine et al, 1990). The resulting radiative forcing is applied to an upwelling-diffusion climate model (Wigley and Raper, 1987 & 1990) to make transient global- mean temperature projections. The climate model gives output for the global-mean temperature changes and the thermal expansion component of the global-mean sea level change. The temperature results are input to ice-melt models for small glaciers and for the Greenland and Antarctic ice sheets, which are the same models as used in IPCC WG1. The ice- melt and thermal expansion terms are finally combined to give the total sea level rise projection. LEAP The Long-range Energy Alternative Planning (LEAP) system provided the organizing analytical framework for the energy and emission scenarios. An associated Environmental Database (EDB) was used as a source for specific emission coefficients. Together, LEAP/EDB comprise a computerized modelling system designed to explore alternative energy futures, along with their principal environmental impacts. The Greenhouse Gas Scenario System (G2S2) was used as an additional source of regionally and nationally detailed energy and emissions data. LEAP, EDB, and G2S2 have been developed by the Stockholm Environment Institute - Boston Centre at Tellus Institute, and applied in numerous countries and regions throughout the world. As a `bottom-up' energy modelling system, LEAP's principal elements are the energy and technology characteristics of end-use sectors and supply sources. LEAP has two important advantages. First, it allows very detailed specifications for key physical parameters in each end-use sector. Thus, the scenarios which were developed embody the impact of a variety of factors - including technological change, demographic variables, and structural shifts in the economy - on energy use. Second, the accounting framework in LEAP enables its results to be internally consistent; that is, assumptions made about energy use in one sector are consistent with those made in another. For example, a reduction in petroleum use in the transport sector automatically leads to a reduction in distribution losses and energy use for petroleum refining. Similarly, with its links to the Environmental Data Base, LEAP can track the pollution resulting from each stage of the fuel cycle, including the reduced emissions from extraction, processing, distribution, and combustion that would result from more efficient use of fossil fuel. While LEAP is capable of incorporating econometric equations (e.g., production functions), it is not an econometric model that determines future energy use based on historical data. A fundamental aspect of such models is the presumption that trends observed in the past will continue into the future. Because econometric models - often considered `top down' models in contrast to the `bottom-up' approach used in LEAP - are based upon historical relationships, they have difficulty in reflecting changes in the variety of technologies available, or other structural shifts that differ from historical trends. For very long term analysis, it becomes increasingly difficult to expect that historically-derived econometric relationships will continue to hold. In addition, due to their usual high level of sectoral aggregation, econometric models forego detail about changes specific to subsectors and energy end-uses such as lighting or process heat. Such changes can have major effects on overall energy use. Unlike `top-down' equilibrium models such as Edmonds-Reilly, LEAP does not simulate price and income interactions to seek a `market equilibrium' between supply and demand for each scenario [1]. As energy efficiency increases, for example, the demand for energy falls and some reduction in fuel prices might be expected. Likewise, as the cost of energy services is reduced by the use of least-cost technologies, the demand for energy can, in turn, rise somewhat. However, such interactive relationships between price and demand are notoriously difficult to accurately quantify, and the high variation among elasticities can lead to dramatically different results. Nonetheless, the basic economic concept of supply and demand cannot be ignored. A dramatic reduction in fossil fuel demand could result in decreases in fossil fuel prices, possibly limiting the cost-effective penetration of competitive non-fossil supplies. Appropriate pricing policies, such as energy or carbon taxes, can compensate for this effect, ensuring full penetration of alternatives to fossil fuels. A related concept, often referred to as the `take-back effect', may occur in cases where more efficient technologies (e.g., compact fluorescent bulbs) decrease the cost of an amenity (e.g. lighting), leading to additional use (leaving the light on longer). Whether such a `take-back' actually takes place in reality depends on the specific end use and is subject to debate in the literature. A few of the efficiency targets embody adjustments for such effects. For these reasons, the end-use `bottom-up' approach, embodied in LEAP, was chosen specifically to enable us to incorporate and simulate several important effects, including technological improvements and transitions, the limits imposed by saturation of several energy-intensive activities, and structural shifts among economic sectors and subsectors. We were interested in selecting the most economic resources we can to create a climate stabilized world. The end-use approach allows us to consider numerous detailed potential steps, such as efficiency improvements and fuel switching opportunities, which have been identified in other studies. ASF The ASF, or Atmospheric Stabilisation Framework, (ICF, 1990) is a suite of models designed to explore the relationships between human economic activity, the emission of greenhouse gases and global warming. Simulation of the energy sector is performed using an adaptation of the IEA/ORAU Long-Term Global Energy-CO2 model, more commonly known after it's authors as the Edmonds- Reilly model. Both the ASF and the Edmonds-Reilly models have been extensively utilised for long-range energy assessments (US EPA, 1990, Edmonds & Reilly, 1985, Mintzer, 1988, IPCC, 1990a & 1992). In broad terms, the Edmonds-Reilly model may be described as a global macro-economic energy model which balances estimates of energy supply and demand through adjustment of energy prices. In this respect it constitutes an alternative modelling framework to the LEAP end-use model (or 'bottom-up' approach) employed in the construction of the FFES. Macro-economic formulations of supply and demand are constructed upon projected levels of population, wealth generation, energy resources, technological change and energy production costs. Within each time frame and region the initial projections of energy supply and demand are adjusted until they balance, using a partial equilibrium routine to modify the energy prices. The supply component of the model permits estimates of future supply from both conventional and unconventional fossil fuels, nuclear, hydro and solar electricity, synthetic fuels and biomass. These supply options compete with one another for market share based upon price forecasts and ability to satisfy demand. Inter-regional trade is permitted for fossil fuels such that a region may be a net exporter or importer of energy. The model tracks emissions of radiatively significant gases from the energy sector, and in particular CO2, to give a picture of the overall release of greenhouse gases from the energy sector. The ASF adaptation of the Edmonds-Reilly model differs primarily in its use of exogenously derived end-use model demand data up to 2025. This feature fixes the energy demand to match the exogenously determined demand for secondary fuels and electricity for the first 40 years, during which time the supply and demand are matched through supply-side adjustment only. After 2025 the model operates in pure macro-economic mode with adjustment of both supply and demand being possible for equilibrium to be reached. As the principal aim of this study was to investigate the economic implications of the FFES which was formulated using an end-use energy model, the ASF variant was favoured in order that demand might be maintained at the FFES level while supply could be optimised as a function of cost. Assumptions for the study. A number of assumptions were made to guide the modelling exercise by SEI - Boston. It should be noted that these are not necessarily Greenpeace policy, but were used in order to make the study comparable with other studies: 1) Fossil fuel combustion was to be eliminated by the year 2100. This outcome was a scenario 'constraint', and did not result from an economic cost-benefit or modelling analysis of the value of substituting for every energy use of fossil fuels. 2) Carbon removal technologies were not considered. Though technologies exist for the capture of the CO2 emitted from fossil fuel combustion, issues such as the integrity of the long-term storage and management of large quantities of carbon dioxide, are as yet unanswered. 3) Nuclear power was eliminated by 2010. Nuclear power is not regarded by Greenpeace as an attractive substitute for fossil fuels, because of the significant risk factors from catastrophic accidents and proliferation, the high costs of generating nuclear electricity in many countries such as the U.K., U.S., and developing countries, plus the environmental concerns of solid and liquid radioactive wastes. Nearly 70 per cent of current nuclear reactors will reach the end of their planned lifetimes by 2010 (Nuclear Engineering International, 1991). The assumption is thus achieved through a moratorium on new reactor construction, and accelerated phase-out of existing reactors. 4) New renewable and other resources were subject to environmental restrictions. Concerns about the construction of new, large hydro facilities, such as siltation, erosion, submergence, human settlement impacts, etc, were reflected in a downgrading of the global technical/economic potential of hydropower by 35 percent. No municipal waste incineration was considered, given an emphasis on materials re-use and reduction policies, and concerns about incinerator emissions. It was also assumed that biomass for energy would only be produced in a sustainable manner, with no net carbon emissions to the atmosphere. 5) Conventional assumptions for GDP and Population were made, with one exception, that relating to equity. This is not to signify acceptance by Greenpeace of such assumptions, but to allow cross-comparison of the FFES with other policy scenarios. Most global energy scenarios, regardless of their source, share a common attribute: they are strongly driven by the projected growth in population and income, the latter usually represented by Gross Domestic Product, (GDP). In general, GDP and population estimates have not received scrutiny and evaluation commensurate with their critical importance in most energy use and emissions projections. This disparity is most likely due to their highly political nature and the difficulty in constructing credible alternatives to standard sources (e.g., United Nations and World Bank). Consider, for example, the population and GDP growth rates assumed by the IPCC in its most recent projections (Swart et al., 1991a). World population grows to over 11 billion by the year 2100, with over fivefold growth in Africa from 560 million to almost 3 billion. South and East Asia are projected to support 3.6 billion people in 2100, up from 1.4 billion in 1988 (see Figure 1). Income (GDP) per capita in the U.S. grows to approximately $83,000 (1985 U.S. $) per capita in 2100, and total global GDP grows more than 14 times. Meanwhile, the income of the average African rises to only $6000 110 years from now, less than half the current OECD average, and over ten times lower than the richest countries. In the FFES analysis an approach to regional income equity was proposed by the SEI-Boston wherein the ratio of highest to lowest average regional income drops to 2:1 by 2100, compared with the current ratio of over 14:1. This is not true equity, of which GDP is only a very crude and inadequate measure anyway, but the gap does continue to narrow after the year 2100. In addition, the rate of GDP growth for regions in the South is at a level regarded as high but credible by SEI-Boston. Any higher and such growth woulkd be likely to have serious socio-economic and environmental consequences. SEI achieved this by maintaining the IPCC projected regional growth rates over the next 20 years, and then gradually adjust them over the 2010-2100 period. GDP is redistributed among regions to achieve the same total world GDP as the IPCC forecast does in 2100: $213 trillion or $19,000 per capita (1985 U.S. $). Figure 2 compares the resulting GDP per capita for these two cases. GDP per capita in all regions is higher than the 1985 OECD average of $12,200 (1985 U.S. $). All 110 year economic forecasts are normative; no objective or scientific methods can be applied to such predictions. Rather than emphasize a `business-as-usual' approach, SEI presented an equity one, which one could argue is just as feasible. Since, in the long run, equity is likely to be a prerequisite for true sustainability, it is important to establish climate stabilization within such a context. Equity is a desirable objective not only among regions, but within them. The FFES and variations are based upon a notion that income distributions will improve, enabling near universal access to household amenities (lighting, refrigeration), commercial and transportation services, and consumer products. Population projections were drawn by SEI-Boston mainly from the World Bank (Bulatao et al., 1990), the same source used by the IPCC. We utilised the IPCC (1990a) assumptions regarding total global GDP growth, in order to maintain the ability to compare our results with the range of reference case studies (Swart et al., 1991a). Figure 3 illustrates that recent projections of GDP growth by IPCC and EPA (average of EPA's RCW and SCW) yield roughly similar GDP results to our own for the year 2100; all fall within a range of $213 to $255 trillion (1985 U.S. $). 6) Structural Change. The concept of structural change in economies is fundamental to the FFES. With the rapid GDP growth rates for the South embodied in our scenarios, SEI anticipated the general transition among sectors that has accompanied the industrialization process in the North: from agricultural and other primary production to a period of greater industrial activity, and finally to the ascendancy of the service sector. The specific path for future economic development in the industrializing countries of the South is impossible to predict; we use the model of the currently industrialized countries as one possible option. Once again this is not Greenpeace policy. SEI therefore constructed a scenario in which there is an initial shift towards the industrial sector during a period of infrastructure building, then a later transition to a more service-oriented economy. In the North, SEI project some additional shifting of economic activity from industry to services, but assume that much of this transition to service activity has already taken place. Further economic growth in OECD countries may be heavily influenced by communications, the so-called fourth sector, and avocational and leisure activities. A balance between industry and service sector activity may remain, while the character of manufacturing and services may change dramatically. Lacking a strong basis for defining residual regional differences in economic structure toward the year 2100, we assume a convergence by 2100 to a service-oriented economic structure in each region, with the industrial sector accounting for 25 percent of future GDP. These macro economic assumptions, while simple, help to capture sector-level structural change that would be expected as economies undergo transitions through industrialization in a world approaching inter-regional equity. 7) Economic Criteria. SEI sought, where possible, to ensure that measures undertaken over the near and medium term (to 2030) yielded net economic benefits or were unlikely to incur significant costs. The emphasis here is on proven or near-market technologies that have been shown to be either cost-effective or cost-competitive with other options. The dynamics of regional price changes in a high-efficiency, high-renewables, low fossil scenario are difficult to predict, particularly in light of various policies that might be implemented (e.g., carbon taxes, efficiency rebates). The economic analysis utilising the ASF model addresses this issue in greater detail. Any cost estimates for the period beyond 2030 are inherently speculative; for this period, our scenarios reflect what currently appears credible and achievable. It should be noted that a number of conservatisms were included in the analysis. These included assumptions of no major breakthroughs in solar and biomass technology, limited efficiency improvements after 2030, likely over-estimates in the Services sector energy demand, high levels of energy-intensive (eg. steel and cement) production, and not taking any improved efficiency credit for electrification in the commercial and industrial sectors. BUSINESS AS USUAL SCENARIOS Since the mid-1970s, a series of global energy projections have been published. As concerns about CO2 emissions and global warming increased in the early 1980s, several of these projections were also translated into carbon dioxide forecasts. Many suggested very high levels of growth in energy use and CO2 emissions (Keepin, 1986). By the mid 1980s, it became clear that energy patterns were undergoing some profound changes in most industrialized countries. Previously exponential growth in energy use greatly diminished, with energy use actually declining in some countries. In many industrialized countries, demand for energy-intensive basic materials (e.g., steel and cement production) was reaching a plateau, while consumer demand for basic energy-intensive equipment and amenities, such as refrigerators or central heating, was levelling off. Emerging concepts of structural change and saturation changed our understanding of energy systems and the nature of energy forecasting. The notion that economic growth must be accompanied by corresponding growth in energy consumption could no longer be accepted as inevitable. Major global forecasts of the late 1980s and early 1990s (World Energy Conference, 1989, US EPA, 1990, IPCC, 1990a & 1992, and Manne and Richels, 1990), projected somewhat lower growth in primary energy than had earlier studies. The range of estimates of primary energy requirements for 2025 dropped from a range of 600-1000 EJ to a range of 450-800 EJ, (see Figure 4). Many researchers have constructed forecasts into the early or mid 21st century. The far smaller number that have attempted to forecast to the year 2100, have done so primarily for the purpose of CO2 projections and climate studies. Figure 5 presents selected primary energy forecasts to the year 2100, with associated CO2 projections shown in Figure 6. Two reference case projections for comparison with the results of the FFES, the IPCC projections and an average of the EPA's Rapidly and Slowly Changing World cases [2]. This range of reference cases is shown in Figures 7 and 8 for primary energy and CO2. There are reasons to believe that these reference cases may be high, in particular, underlying population and economic growth rates could be unrealistically high and unsustainable. Nonetheless, they reflect two of the most widely cited and internationally reviewed long-term projections to date. Oil use in both scenarios increases over the next forty years by up to 34 per cent. In both US EPA scenarios, synthetic oil from coal and oil shales increasingly makes a contribution after 2025, providing between 103 and 267 exajoules in the year 2100 (US EPA, 1990). Synthetic fuels from fossil fuel sources are among the most polluting fuels known to man. The climate implications of a Business-as-usual Scenario. The climatic consequences of these scenarios have been explored using STUGE and are illustrated in Figures 9 and 10. By 2100 the global-mean induced radiative forcing exceeds 10 W/m2, while in the event of a higher climate sensitivity (4.5oC) the global-mean temperature is forecast to exceed 6oC above pre-industrial times (4oC under a 2.5oC sensitivity). The impetus for global warming is not diminished by 2100. Even if emissions were held static or cut, the thermal lag of the oceans and long atmospheric residency time of GHGs would ensure that global-mean temperatures continue to rise beyond 2100. Sea level is projected to increase from 35 to 115 cms. (depending on climate sensitivity), while rates of temperature increase of between 0.2 and 0.6oC per decade. In terms of environmental impacts, the BAU scenarios may be extremely severe. Sea-level rise would threaten all coastal habitats and human infrastructures, including the major ports, resulting in increased flooding, soil salinity changes, land erosion and a host of other effects. In addition, certain marine-based ecosystems may have trouble adapting to the rise in sea-level height, such as coral reefs. Climate change has the potential to disrupt existing ocean circulation patterns, and change the global supply of nutrients and heat which determine conditions for all marine ecosystems (IPCC, 1990b). On land, the rising temperatures would change the climatic conditions for ecosystems at an unprecedented rate. There is a very strong correlation between climatic regions and biotic types, such that shifts in climate force ecosystem migration or extinction. In the past climatic changes have usually occurred slowly over long time spans with little or no human intervention in migratory processes. Under the reference scenario conditions, it is very doubtful whether major land-based ecosystems can migrate successfully (given human obstacles such as farmed land, urban settlements, road networks, etc.) or fast enough to avoid mass extinction. RESULTS OF THE FOSSIL FREE ENERGY SCENARIO (FFES) The findings of the Fossil Free Energy Scenario (FFES) indicate that a combination of efficiency improvements, renewable energy technologies, and fuel switching, could achieve significant long- term reductions in CO2 emissions. As shown in Figure 11, annual CO2 emissions peak around the year 2000, and decline to 48 percent and 29 percent of current global levels by 2030 and 2075, respectively, before reaching the fossil-fuel target of zero net CO2 emissions by 2100 [3]. The initial rise in CO2 emissions reflects the momentum of current energy use patterns, the embedded stock of energy- inefficient equipment, and the time required to effect of large shifts in fuel and technology choices throughout the world. The scenario roughly achieves the 20 percent reductions by 2005 among industrial countries called for by the 1988 Toronto World Conference on the Changing Atmosphere. These reductions do not reflect Greenpeace policy as higher individual national targets would be expected. Nearly all industrialized countries have already pledged to either stabilize or lower emission levels by that time. In the scenario reported here, CO2 emissions from 1988 to 2000 decline by 3-12 percent in industrialized regions, while increases in developing regions which range from 21 percent in Latin America to 55 percent in South and East Asia, lead to a 6 percent overall increase in CO2 emissions. Between 2000 and 2010, the effects of technology improvements begin to outweigh the underlying forces of economic and population growth, and global CO2 levels begin to decline. Reductions in all industrialized regions offset continued increases in all developing regions. Beyond 2010, emission levels decline in all regions, as the current stock of energy consumption and production equipment turns over, and high efficiency end-use and electricity generation technologies are widely implemented. Reduced dependence on coal and modest levels of renewable fuels and electricity further contribute to CO2 savings. The contribution of fuels to total delivered energy is shown in Figure 12. Oil use falls slightly by the year 2000, before falling by 20 per cent in the year 2010 and nearly 50 per cent in the year 2030. Oil use is phased out by the year 2100. By 2030, a new generation of lower-cost renewable supply technologies provide a cleaner, low-CO2 mix of fuels and electricity. Over the longer-term, the transition to a solar and biomass-based energy system, as illustrated in terms of primary energy supply in Figure 13, leads to the complete reduction of fossil fuel carbon dioxide emissions by 2100. Estimated cumulative global CO2 emissions from 1988 to 2100 amount to 314 (Pg C), as shown in Table 2. Emissions from the 5 industrialized regions account for 53 percent (165 Pg C) of these cumulative emission, over 70 percent (118 Pg C) of which occur by 2030. Thus, despite high levels of economic growth in the South, the North remains the major contributor to CO2 emissions overall, and particularly in the near and medium term. Accelerating the adoption of emission reduction measures in industrialized countries would appear the most important area for additional improvements, even under the improved equity assumptions of this scenario. Improved end-use energy efficiency in residential, industrial, transport and service sectors, accounts for the major source of emissions reductions between now and 2030. For the period beyond 2030, smaller improvements in end-use efficiencies are assumed. The dip in energy use from 2010 to 2030, followed by a subsequent rise for the remainder of the scenario, largely results from strong efficiency improvements prior to 2030 that outweigh the dual forces of growing economies and population for a brief period of time. Beyond 2030, much slower rates of efficiency improvement were assumed. As a result, continued economic and population growth in the South leads to growing energy requirements. The levels of end-use efficiency improvements included in this scenario are based upon current assessments of economic and technical potential: levels based on market or near-market technologies that can be implemented within 40 years. Beyond 2030, the more conservative estimates (0.5 percent annual efficiency improvement for most end uses) could well be overly pessimistic. If so, the doubling in delivered energy from 2030 to 2100 could be avoided. Improved efficiency on the supply side, including the more efficient use of fossil fuels for electricity production (e.g., combined cycle and fuel cell systems), refinery improvements, and reduced transmission and distribution losses, provide important contributions to reducing emissions over the next 40 years. This also includes the overall efficiency gains offered by on-site and centralized combined heat and power generation (cogeneration). After 2030, approximately 20 percent of global electricity demand is supplied from centralized and on-site cogeneration. Fuel switching from coal and oil to natural gas, also plays an important role in reducing emissions in the near and medium term (to 2030). Estimates were made for each region regarding the ability to switch to lower carbon fuels, based on the availability of fossil fuel supplies, particularly natural gas, and end-use considerations. The resulting primary energy shares are shown in Table 3. The cumulative global consumption of natural gas and oil in the FFES is 6200 EJ and 6600 EJ, respectively. According to one recent analysis, global natural gas reserves are sufficient to meet 1990 production levels (75 EJ/yr) for another 60 years, or 4500 EJ, somewhat less than the cumulative natural gas use in this scenario.{1} However, such levels of gas are required partially as a function of the high levels of GDP and population assumed. In the lower variants simulated in Section 5.2.2, overall energy use is substantially reduced by around 10 percent by the year 2030, and 22 to 35 percent by the year 2100. In this situation it would be possible to adhere to Greenpeace policy of no new oil and gas exploration. Greenpeace has a policy which is opposed to new gas use unless available energy efficiency and renewable energy opportunities have been exhausted. Any additional use of gas in the FFES should be viewed within the framework of a comprehensive range of CO2 reduction measures, and only where methane leakage (CH4) can be kept to an absolute minimum. The cumulative oil and natural gas supply estimates through 2100 in the FFES are both less than half those projected by U.S. EPA in its Rapidly Changing World scenario. For regions heavily dependent on coal, with limited supplies of natural gas and other fuels, coal continues to account for an important but declining share of primary supply. For example, in Centrally Planned Asia, coal use drops from 72 percent of primary energy in 1988, to 51 percent in 2010, and to 22 percent in 2030. Globally, coal and oil use declines from 27 and 34 percent of total primary energy supply in 1988, to 21 and 26 percent by 2010, and 7 percent and 15 percent, respectively, by 2030. Meanwhile, the share of natural gas rises from 19 to 26 percent from 1988 to 2010, before dropping to 15 percent in 2030. A major transition to solar and biomass sources of energy, as illustrated in Figure 13, accounts for most emission reductions after 2030. Over the next 40 years, several renewable sources play an important role, including biofuels for transport, wind energy for electricity, and various cost-effective applications of solar technologies. After 2010, more rapid penetration of low-carbon systems begins as the current stock of technologies turns over, emerging technologies mature, and appropriate infrastructures are established. By the mid 21st century, solar and wind energy systems begin to account for the largest share of primary supply. The scenario encompasses a diversity of renewable supply sources including high-efficiency biomass gasification, agricultural residue-based cogeneration systems, hydrogen and biofuels as fossil fuel substitutes for vehicular and other uses, and direct solar thermal applications. Land use considerations will impose limits to the maximum penetration of biomass energy sources. Biomass energy -- delivered in the form of biofuels and electricity -- could require as much as 120-380 million hectares worldwide to provide 91 EJ or 24 percent of 2030 primary supply. This compares with 8.7 billion hectares in the current combined crop, pasture, forest and wood land area and 13 billion hectares of global land area. In comparison, solar and wind collection, which accounts for 118 EJ (fossil equivalent) or 31 percent of 2030 primary supply, would likely require less than 15 million hectares from land surface areas (arid areas, rooftops, etc.) that may pose fewer conflicts. By 2100, the biomass energy supply and land use requirements roughly double their 2030 levels to 181 EJ, implying land use of 290-720 million hectares, and could be of greater concern than the estimated 50-110 million hectares required for solar and wind. The high estimates assume no breakthroughs in biomass yields or solar capture efficiency (e.g., 15 percent efficient photovoltaics), and thus are conservatively high (see Sensitivity Testing Section below). Because of the potentially high water and fertilizer requirements of very high-yielding biomass species, we assumed lower biomass yields than are assumed by other studies. Land availability for biomass energy will depend on the ability of improved agricultural productivity and recycling and reduction of wood and paper products (thereby reducing land use for commercial wood), to reduce competition for suitable land. It is important to note, however, that a coal and oil shale based future would also present high land use requirements, and more destructive environmental impacts. Technology Options Switching from Higher to Lower Carbon Content Fossil Fuels. For appropriate end-uses, switching from higher to lower carbon content fuels, such as from oil to natural gas, can offer an important transition strategy. The FFES embodies a continuous shift from coal and oil to natural gas where available, with the condition that natural gas leakage can be effectively controlled. In some cases, fuel substitution can be done with little new investment, such as with dual-fuel boilers, now fairly common in industrial applications. In general, the potential for short-term fuel switching is limited by the requirements of retrofitting or replacing energy using equipment; over a 10-30 year time period, however, equipment stock turnover in all sectors can be expected to enable far greater levels of fuel switching. Switching from Fuels to Electricity. Depending on how it is generated and consumed, electricity can be either more or less efficient than fuel for end-use applications, when the full fuel cycle is considered. A commonly cited example of inefficient electricity use is the resistance-heated building. Delivered electricity from fossil fuel combustion can impose losses of 60 to 70 percent in steam cycles, compared with losses of 20 to 30 percent if the same fuel is used to heat the building directly. In contrast, some applications of electricity offer significant efficiency improvements that can help to offset electric generation losses. Examples include high COP (coefficient of performance) heat pumps for water and space heating, electric arc furnaces for steel-making, and many process applications. With advanced highly efficient electricity generation technologies (e.g., fuel cells), switching from fuels to electricity for appropriate end-uses can provide significant overall savings. In addition, electricity offers an important carrier for low- carbon renewable energy sources, such as geothermal, wind, and hydro. When moving to a renewable-based energy economy, electricity must be weighed together with other convenient energy carriers such as biogas, alcohol, and hydrogen, all produced through conversion process that also impose energy losses and usually additional non-CO2 emissions. Switching from Fossil Fuels to Renewable Fuels. In order to fully replace fossil fuels, acceptable low-carbon solid, liquid, and gas fuels will be required, preferably derived from renewable sources. The other option is an all-electric-energy world, an option that, while conceivable, we do not consider here. Biomass offers the most abundant, renewable combined feedstock and energy source, and the possibility of producing a wide range of biofuels, including methanol, ethanol, methane-rich biogas, producer gas (CO/H2), and diesel substitutes. The other major renewable fuel that we consider is hydrogen, which can be produced using either biomass or solar energy. A range of alternative fuels are used in the transport sector to replace oil use. These include biofuels, hydrogen and renewable- based electricity. These are discussed in greater detail in the Biofuels and Transport sections later. Combined Heat and Power (Cogeneration). When electricity is produced from the combustion of fuels, from 60 to 70 percent of the total energy content of the fuel is typically lost as waste heat, discharged to the local environment. Combined heat and power, or cogeneration, systems take advantage of the heat generated, for industrial process applications or for space and water heating in buildings. In these systems only 20 to 40 percent of the total energy content of the fuel is lost. Cogeneration systems can be used in the service, residential, and industrial sectors. For commercial and residential buildings, both centralized and on-site cogeneration options are possible, while large, centralized station cogeneration units can feed district heating networks, many of which currently exist in Europe, the U.S., USSR, and China. Smaller, on-site cogeneration units can be implemented in numerous industrial, as well as many commercial and residential situations. Modular, high-efficiency fuel cells offer the potential to greatly increase the cogeneration market in the medium term. For instance, in Japan, a 40-fold expansion of cogeneration and district heat using fuel cells has been suggested (Yamamura, 1991). Sufficient heat and hot water loads are needed to justify district heating systems; we thus restricted district heat penetration to colder climate areas, and areas where multi-family dwellings and denser cities predominate (i.e. former USSR, China, Northern and Eastern Europe, and part of North America). Electricity Generation. Oil is used to provide around 10 per cent of electricity generated in the OECD. In the South, the figure is much higher, frequently up to 50 per cent. A wide range of options for phasing such use out are available. The cost of generating electricity from several renewable sources is already cost-competitive with fossil resources under certain conditions; at the best solar thermal and wind sites, some remote photovoltaic installations and at numerous hydroelectric and geothermal sites (see Figure 14). Over the next 40 years, decreasing costs could make wind and solar electricity cost competitive at more abundant sites, enabling their installation in all regions, particularly at or near the site of energy use (e.g., rooftop photovoltaics) (SERI, 1990). In fact, recent research indicates that within the 1990s, significant technical improvements in large-scale electricity generation from solar thermal technology appear feasible. Preliminary economic estimates suggest that electricity from solar thermal stations could be produced at an average cost of 4 to 6 cents per kWh by the end of the decade, making it cost competitive in many regions of the world (Mills and Keepin, 1992). A largely renewable, low- emission electricity supply system can thus be envisioned for the remainder of the next century. Near-term renewable potentials (to 2020): Region-specific potentials for several near-term renewable electricity options up to 2020 -- hydro, wind, solar, and biomass residues -- were derived from three existing projections (Dessus et al., 1991, Odgen et al., 1990; Hall et al., 1992). The analysis by Dessus et al. is based on several geographical, social, and economic considerations, and provides accessible renewable reserves for solar, wind, and hydro during the 1990s and by 2020. We adopted these reserve levels, but delayed full implementation for the 1990 reserves until 2010, to reflect the time scale of the retirement of existing fossil and nuclear capacity. a) Biomass Waste Cogeneration: The total energy resource from current recoverable biomass residues has been estimated at 65 EJ (55 per cent of current oil use), (Hall, et al, 1992). This equivalent to 17,300 Braer's'. The FFES considered only a small fraction of this potential, mainly due to ecological constraints. The use of pulp and paper wastes, another major source of biomass residues, has been considered within the analysis of the industrial sector above. A discussion of the large, economic potential of biomass waste use with biomass integrated gasifier steam-injected gas (BIG-STIG) turbine cogeneration systems can be found in several recent reports by researchers at Princeton University (Ogden, et al., 1990, Williams and Larson, 1991). BIG-STIG cogeneration systems offer a relatively low-cost versatile option for utilizing biomass resources. Average overall (combined thermal and electric) efficiency is estimated at 62 percent, with a ratio of electricity to steam production of 48 percent to 52 percent. Future sugar residues and their generation potentials by region were derived from the same source (Odgen et al., 1990). A more conservative estimate of 25 percent of estimated sugar residue cogeneration potentials was adopted for 2030, with a gradual increase from now until then, and constant levels thereafter. b) Fuel cell technology (`2nd generation'). Fuel cells (and magnetohydrodynamics, MHD) provide a second generation of efficiency improvements in electricity production, starting in the year 2000, and representing all new fuel-based capacity (biofuel and hydrogen) after 2030 (EPRI, 1989, Fisher, 1990). The efficiency of these units starts at 50 percent and grows to 57 percent by 2030. For coal, the efficiencies are 10 percent lower, representing the losses from coal gasification. For coal, MHD could provide similar efficiency improvements at similar costs to fuel cells. c) High-efficiency biomass generation. Gasification of biomass at 85 percent energy efficiency provides the fuel input to fuel cells. If fuel cell technology does not prove to be the optimal technology for use of biomass resources for electricity production, then other efficient technologies such as gasified combined cycle or STIGs could be used. The biomass resources utilized for both electricity and biofuel production are discussed below. The only region in which biomass does not displace up to 40 percent of fossil fuel inputs by 2030 is the Middle East, where no biomass is assumed due to plentiful fossil and scarce biomass resources. d) Long-term solar and wind supply. The near-term renewable potentials for solar and wind noted above, are limited to the use of highest resource sites and remote locations (2010 and 2030). Wide-scale application of solar PV, solar-thermal electric, and wind technologies at more abundant, lower resource sites is not expected to be cost-competitive until sometime around the 2010-2030 period, as illustrated in Figure 14. The long-term solar and wind resource is phased into the energy mix, starting slowly in 2010. The FFES assumes that solar and wind resources could provide up to 40 percent of direct electric loads, i.e. without storage requirements, by 2100. Grubb provides a detailed discussion of grid penetration and integration issues for renewables, wherein he points out that, a few caveats regarding PVs notwithstanding, concerns over the integration of renewable sources have been grossly exaggerated. "Contributions of perhaps 20 percent of the demand could be obtained from one type of variable [e.g., renewable] source with only a modest reduction in the value of the energy, and contributions of 30-40 percent would seem to be feasible before the penalties become severe, even neglecting storage and possible power exchanges with other systems" (Grubb, 1991). In addition to Grubb's analysis, several additional factors suggest that solar and wind resources could provide even greater share of direct electric load than assumed here. First, the demand for electricity in many areas, particularly those with good solar resources, correlates well with solar resource availability. Second, the electric load from a large fleet of storage-equipped electric vehicles could be shifted to periods of excess supply (20 to 40 percent of vehicles are electric from 2030 to 2100 in this scenario). Third, the load-following capability of other base and intermediate load generation technologies will improve compared to the poorer load-following capability of current boiler technology as the supply mix switches to a more flexible mix of sources including fuel cells, combined cycle units, and small cogeneration units (internal combustion engines, gas turbines, and fuel cells). Beyond the 40 percent of load met directly, the remainder of solar and wind resources are assumed to require storage. For modelling purposes we assume hydrogen will play this role, although other storage (compressed air, battery, hot rock beds, etc.) and transmission modes are also possible, and could prove more efficient and cost-effective. Solar thermal, PV, and wind capacity must be sufficient to produce hydrogen for use in high efficiency fuel cells during hours when solar insolation or wind power is not available. Fuel cells using hydrogen could operate at 70 percent efficiency, as assumed here. An 84 percent conversion (electrolysis) efficiency from electricity to hydrogen is assumed (Ogden and Williams, 1989). Long-distance transcontinental electricity transmission offers the potential for significantly reducing storage requirements beyond what is envisioned here, particularly with successful development of superconductors for low-loss transmission. In particular, east-west transmission would increase the diversity of wind and solar resources available to follow electric system loads. The future balance between long-distance transmission and storage will depend not only on the evolution of technologies and their costs, but on other considerations such as desires for local resource self-sufficiency. Solar and wind generation could develop in both central station (e.g., large solar thermal stations and wind farms) and decentralized, on-site (e.g., integrated PV cell roofing materials) modes. We do not attempt to establish here which mode would dominate. Reduced transmission and distribution losses. By the year 2030, electricity losses are assumed to have fallen to 6 percent in each region, a level already observed in the OECD Pacific region (Japan, Australia, and New Zealand). Electricity losses are left unchanged in 2030 and beyond. This estimated improvement is highly conservative; losses could be even lower given the potential development of superconductor technology, the reduced transmission requirements due to increased on-site cogeneration, and greater reliance on hydrogen storage, if it proves to be a more efficient long-distance transmission carrier than electricity. e)Renewable Fuels (Non-Electric). In this scenario, hydrogen produced from solar and wind resources, and various fuels derived from biomass, comprise the renewable fuels that eventually displace fossil sources for gas, liquid, and solid fuel applications. Hydrogen production from solar and wind resources is described above. For simplicity of analysis, we have assumed that hydrogen satisfies most gas fuel demands, and that biofuels are largely used for liquid fuel applications. However, other assumed fuel patterns would be roughly compatible with our overall results, such as heavier use of gaseous biofuels, hydrogen from biomass, or `blended' gas supply that might consist of a mix of biogas, hydrogen, and a decreasing fraction of natural gas over time. f) Biofuels. Liquid biofuels include ethanol, methanol, and synthetic gasoline. With several conversion technologies under development, such as enzymatic hydrolysis for ethanol production, it is too early to say which fuels and processes will prove most practical and economic. Cost estimates for each, as shown in Table 4 and Appendix A, indicate that they all could be cost- competitive with gasoline within the next 20 years. This excludes the possible introduction of pollution taxes on fossil fuels. With further development in conversion technologies, and increase in petroleum product costs, costs should drop sufficiently for renewable fuels to effectively compete with other petroleum products in other applications. For example, based on distillate oil and natural gas price projections by the U.S. Department of Energy, biofuels could readily compete with petroleum products for residential, service, and industrial sector applications by early in the next century. A conversion efficiency from biomass feedstock to all biofuels of 50 percent (energy basis) was assumed, based upon equal or higher estimates from other sources (SERI et al., 1990; Cook et al., 1991). Conversion efficiency is assumed to rise to 60 percent by 2010, a lower assumption than cited in some other sources. For instance, the recent EPA Policy Options study assumes an increase to 75 percent by 2010, based on U.S. DOE analysis (US EPA, 1990). For this analysis, we assumed that most biomass feedstocks for fuel conversion are produced from herbaceous and short-rotation woody biomass plantations. Since feedstocks from dedicated woody biomass crops and plantations would likely be unavailable in large quantities before 2010, most biomass feedstocks resources until then are assumed to come from biomass residues and existing forest resources. The FFES conservatively estimates long-run biomass productivity of 10 tonnes of woody biomass per hectare in temperate regions; 20 tonnes per hectare in moist or irrigated tropical regions, and 4 tonnes per hectare in semi-arid tropical regions (all tonnes given here are dry tonnes). The assumptions are based on review of biomass productivity estimates from various sources (Hall et al, 1992, US EPA, 1990, Dessus et al, 1991). These productivity levels are considerably lower than those projected to be achievable by U.S. EPA, and are well within the range of productivities currently achieved on managed lands. Relatively conservative estimates were chosen, since very high yielding biomass plantations might be unsustainable (and unrealistic) due to high nutrient and water requirements, and potential lack of ecological diversity. Table 5 shows the decreased land use requirements that would result from a steady improvement of productivity to double the base levels by 2030. Such higher levels have already been achieved in practice -- 18-20 tonnes per hectare in temperate regions (high end of range for U.S. and Europe). The biomass plantations could occupy over 700 million hectares, or about 8 percent of current crop, pasture, forest and woodland land area by 2100. Assuming biomass productivity could be improved two-fold by then, and that a significant percentage (25 per cent) of recoverable residues can be captured, land requirements drop to 3 percent of the combined land areas. In either case, it is important to emphasize that these high land use requirements are not simply a result of the choice of biomass as a supply source, but of the underlying assumptions of this analysis: maintaining current industrialized patterns of consumption in a world of over 11 billion inhabitants. It is important to stress that the underlying assumptions used in this modelling exercise are not inevitable, they are dependent on policy and political choices. A number of variations in population, GDP and other assumptions were made (see below). SECTORAL RESULTS In 1988, industry was the largest energy consuming sector at 38 percent of total delivered energy, as shown in Table 6. Transportation accounted for 26 percent, while residential and services accounted for 23 and 13 percent respectively. The major change in sectoral distribution over the course of the FFES is the increase in Services sector and decrease in Residential sector delivered energy. Saturation of appliance, heating, cooling, and cooking needs effectively limits overall growth in residential energy consumption. In addition, the potential for residential efficiency improvements is very high; there is no overall growth in total delivered energy from 1988 to 2100. With structural economic shifts from industry accompanying economic growth in developing regions, service sector activity grows faster than any of the other three sectors, averaging 1.1 percent per year from 1988 to 2100. However, the industrial sector remains the dominant energy sector throughout the study period, largely due to the increase in basic materials consumption (steel, cement, etc.). Transport Sector. Four separate scenarios were developed for this sector (Walsh, 1992). The last of the four -- alternative fuels -- was incorporated into the FFES analysis. These scenarios were intended to test the effects on energy use of four technology and policy issues: improvements in fuel efficiency, improvement in clean air technology, the introduction of non-fossil fuels into the transport sector and the effects of controlling the expected growth in number of vehicles. Assumptions for the sector: i) Fuel efficiency for cars and light trucks improves at 3 per cent per anum from 1993 levels until 2030, after which no further improvements are made. After 2005 it was assumed that 2 per cent of all new light duty vehicles, subsequently increasing by 2 per cent per year, would use advanced technology (ie. fuel cells) that would be 50 per cent more efficient than conventional vehicles. Overall fuel efficiency improves from 24.5 miles per gallon (20.4 per US gallon) to more than 96 mpg (80 US mpg) in 2030 for cars, and from 20.5 mpg (17.1 US mpg to 80 mpg (67 US mpg) for light trucks. Fuel efficiency for heavy duty trucks and buses is assumed to improve by approximately 10 per cent between 1995 and 2000, 10 per cent per decade between 2000 and 2030, and then 10 per cent from 2030 and 2100. Motorcycle efficiency remains constant throughout. ii) Clean air legislation. Emissions of CO, HC and NOx are based on current levels in every country. Most industrialized countries, and many rapidly industrializing countries have introduced clean air legislation programmes covering exhaust and evaporative emissions standards, fuel quality standards, refuelling controls, and inspection and maintenance programmes. These are incorporated into the baseline scenario. Countries with very few or no pollution control strategies at present are assumed to adopt similar programmes after approximately a ten year delay. iii) Constraints on vehicle use and growth. A range of policies are assumed, particularly in urban areas where the majority of vehicle trips take place. These include the gradual development of 'car-free' zones, a rapid increase in public transport, and a move towards better urban planning and zoning, thus reducing the need to drive to work and travel around for shopping, education, etc. The current vehicle stock of 680 million (430 million cars, 110 million light duty vehicles, 110 million motorcycles and 30 million heavy duty vehicles), grows considerably to 4.93 billion by 2100 under a 'business-as-usual' situation. This reduces to 1,150 million and 1,600 million in 2030 and 2100 respectively in the 'low growth/alternative fuels' scenario. iv) Alternative Fuels. If significant reductions in CO2 emissions are to occur, non-fossil fuels will be needed. There are a number of technological options already available today, which could substantially lower emissions in the future. It is likely that others will emerge should the proposed policy tools be adapted to encourage their development and introductions. Stringent limitations on CO2 emissions using either standards or economic instruments were assumed to be a high priority. The use of appropriate policy instruments should, at a minimum, facilitate the expanded use of biofuels and solar-electric or solar- hydrogen. Biofuels are close to economic viability now (SERI et al, 1990). An enhanced R & D programme coupled with increases in the cost of gasoline and diesel due to pollution externalities, are assumed to make costs comparable in the period 2000 to 2010. The penetration of biofuels is assumed to reach 10 per cent of total fuel use in 2010, 20 per cent in 2020 and 30 per cent in 2030. With renewable (solar, wind, hydro etc) electricity source becoming cost-competitive with fossil fuel sources early in the 21st century, the use of electricity and hydrogen (via electrolysis) becomes an increasingly attractive option. The costs of photovoltaic cells are assumed to reach the costs of conventional fuels in the period 2015 to 2020. The cost of wind generated electricity is already competitive with fossil fuels in some situations. Penetration rates of solar-electric/ solar- hydrogen are assumed to increase from 10 per cent in 2020, to 30 per cent in 2030 and 80 per cent in 2100. The split between straight electric vehicles and solar-hydrogen vehicles is assumed to be 2:1 in the year 2030 and 1:1 in 2100. The overall results for the transport sector show a fall in CO2 emissions of nearly 40 per cent by the year 2030, and 100 per cent by the year 2100. Emissions of other transport emissions such as CO, HC and NOx fall by 86 per cent, 73 per cent, and 57 per cent respectively from current levels by 2030, and 92 per cent, 85 per cent and 74 per cent by 2100. Industrial Sector. The industrial sector, as defined here, comprises a range of economic activities from primary production to the manufacture of consumer goods. In 1988, the industrial sector accounted for 35 percent of global final energy consumption, more than any other sector. We disaggregated the industrial sector into the five most energy intensive sub-sectors -- iron and steel, non-ferrous metals, non- metallic minerals, paper and pulp, and chemicals -- and a sixth category for all other industry. This category includes food processing, textiles, machinery, mining, and other productive sub-sectors. Energy intensiveness here is given as energy use relative to economic activity. Two sets of factors will influence future energy use and CO2 emissions from industry. The first set relates to the characteristics of future industrial sector -- overall economic activity levels, shifts among industrial sub-sectors, and the changing set of products within each sub-sector. The second set relates to how energy is used for production processes -- energy efficiency, process technology and fuel mix. At an aggregate level, we represent these factors using two variables -- economic activity (GDP) and energy intensity (GJ/GDP) for each sub-sector. Changes in sub-sectoral economic activity and energy intensity in the scenarios are governed by three considerations: increasing equity, technological improvements, and what we will refer to as standardized industrialization. Increased equity considerations suggest that by the year 2100 all regions will have reached or exceeded current OECD per capita GDP with corresponding levels of materials consumption. Over the next century, assuming improved inter-regional equity, technological improvements and transfers can enable all regions to reach similar levels of high industrial energy efficiency. A principle of standardized industrialization implies a model of industrial development in the South that follows the basic pattern observed in the North, the transition from a reliance on primary materials to a more diversified industrial economy. This transition encompasses not only the shifts from industry to services, described earlier, but also among industrial sub- sectors, and within each sub-sector itself. In the scenario, global industrial energy use grows from 90 EJ in 1988 to 196 EJ in 2100, a total increase of 118 percent or an average of 0.7 percent per year (see Table 7). Most of this increase occurs after 2030, after most of the savings from efficiency improvements have been captured. As a result of industrialization in the South, the production of basic energy- intensive materials increases two and one-half to ten times during the same time period. Global production of steel increases 3.8 times; aluminum, 8 times; cement, 2.4 times; and paper and pulp, 10 times. In the short-term (to 2030), the currently industrialized countries continue to dominate global industrial energy use. By 2100, however, industrialization and population growth in the South lead to a major change in the relative regional energy shares. Industrial energy use in the countries of the former USSR and Eastern Europe declines 35 per cent from 27 EJ to 17 EJ between 1988 to 2030 and an additional 28 per cent to 12 EJ in 2100. The OECD declines less dramatically from 35 EJ in 1988 to 24 EJ in 2100. In the South, the situation is reversed. Population and production growth drive energy use up 1.3 times from 28 EJ to 38 EJ by 2030, and another 4.3 times to 160 EJ (82 per cent of total global industrial energy use) in 2100. The types of fuels used to meet this demand also alter (see Figure 15). In 1988, about 71 percent of industrial energy came directly from fossil fuels, another 17 percent from electricity, currently supplied predominantly by fossil fuels, and only 12 percent from renewable fuels or heat. By 2030, direct fossil fuel use declines to 24 percent, renewable fuels and cogeneration rise correspondingly to 32 percent and greater electrification of industry brings electricity's share to 45 percent. In 2100, the proportion of electricity use remains almost constant at 45 percent, but renewable fuels and cogeneration account for the remainder of industrial energy use. Advanced cogeneration units provide 90 percent of all industrial steam requirements, and account for 40 percent of total industrial fuel use by 2030 and 2100. While not observable from the results for final demand, the widespread adoption of cogeneration systems significantly reduces primary energy requirements. Residential Sector Projected residential sector energy consumption by region are illustrated in Table 8. The results reflect a convergence by the year 2100 on a standard set of highly efficient electric appliances, increased levels of hot water usage in most regions, and region-specific targets for cooking, space heating and cooling, based on notions of climate and cultural differences. They encompass major improvements in building shell efficiency, reduced heating and cooling energy requirements, increased district heating in temperate climates, solar water heating, electric heat pumps, and towards the end of the study period the dominance of solar and biomass fuels. Between 1988 and 2030, global residential energy use declines, amid a rapid increase in access to household amenities, such as refrigeration and hot water, a 60 percent increase in global population, and a significant decline in persons per household. Such a surprising result reflects the massive technical potential of improving household energy efficiency. Combined with fuel switching from coal to natural gas to electricity and renewable fuels, CO2 emissions decline even further substantially. Household Energy Transition in Developing Countries The major differences between our household analysis of developing and industrialized regions lies in the treatment of energy transition issues in developing countries. Developing regions currently use a significant amount of traditional biomass fuels (firewood, charcoal, animal and crop wastes) for cooking, water heating, and small-scale enterprises, such as beer brewing, food curing, and pottery. For this scenario, a full transition from traditional biomass fuels to more convenient fuels was postulated. Services Sector. This sector includes all demand-side activities not included in the industry, transport, and residential sectors. While it includes agriculture, this category is referred to as the Services sector throughout most of the report, since with a few notable exceptions, agriculture accounts for only a small fraction of total `modern' energy use. It was assumed that energy use in this category is governed primarily by trends for service activities and building stock. The FFES shows service sector energy demand increasing faster than any other sector (1.1 percent per year) from 1988-2100. Total service sector energy consumption grows over 3-fold, from 30 EJ in 1988 to almost 100 EJ by 2100, due in large part to structural economic shifts to the provision of services from primary production and manufacturing (see Table 9). In the North, this shift is already underway. In the South, rapid growth in services can be expected to accompany industrialization and the sector dominates overall energy demand thereafter. As in other sectors, service sector energy use is characterized by a shift from the North to the South. OECD service sector energy use declines throughout the scenario, from 16 EJ in 1988 to 13 EJ in 2100. In the countries of the former USSR and Eastern Europe, service sector energy use initially increases about 24 per cent from 7 EJ to 10 EJ reflecting structural shifts into the sector before declining to 6 EJ as efficiency measures dominate. In the South, service sector energy use rises throughout the scenario. Energy use more than doubles between 1988 and 2030 from 6 EJ to 17 EJ. From 2030 to 2100, energy use increases an additional 4.7 times to 80 EJ due to the combined effects of economic growth and structural shifts. Future service sector activity and hence energy use will depend on the nature of services utilized: business services (accounting, finance, etc.), restaurants, hotels, retail stores, and so on. In the scenario, we have not attempted to distinguish future changes in the mix of services. As noted above, the variation in energy intensity among sub-sectors is far less than that found within the industrial sector. In an emissions control scenario for the U.S. Congress, the Office of Technology Assessment has estimated that approximately 40 percent of total building (combined residential and service sector) energy use could be reduced by the year 2015. A combination of efficiency improvements and service sector cogeneration could also yield net benefits to the economy of up to 28 billion dollars (OTA, 1991). For the service sector, OTA assumed that new building shells could be 75 percent more efficient than 1987 U.S. average, while retrofits could achieve 40 percent improvements in building shell efficiency by 2000. A combination of high efficiency bulbs, ballasts, reflectors, and daylighting can reduce energy use by 50 percent in existing and 60 percent in new buildings (OTA, 1991). Other estimates have indicated lighting savings for U.S. buildings from 70 to 90 percent, using improved fixtures and dimming technologies (Fickett et al., 1990). In warmer regions, service sector energy use tends to favour electricity, as cooling and lighting tend to be the big energy users. In Sao Paulo, Brazil's largest city, approximately 44 percent of service sector electricity use provides lighting, while air conditioning (20 percent), refrigeration (17 percent), and cooking (8 percent) account for most of the remainder (Geller, 1991). In Thailand, lighting accounts for 31 percent of total service sector energy use (Busch, 1990). More efficient lighting generates less heat, and can result in significant additional benefits in terms of reduced cooling loads. Not surprisingly, this phenomenon is most marked in hotter climates. In Thai office, hotel, and retail buildings, lighting improvements can reduce lighting energy use by 70 percent, while at the same time reducing energy needs for cooling (12 to 33 percent, depending on building type), and ventilation (13 to 26 percent) (Busch et al., 1991). The estimated costs of saved energy for these measures are all at or below $.04/kWh (1991 U.S.$). COMPARISON OF FFES RESULTS WITH OTHER SCENARIOS Table 10 and Figure 16 compares the results of the Fossil Free Energy Scenario (FFES) with one of the US EPA reference case scenarios, and the EPAs lowest policy CO2 emission scenario, Rapidly Changing World with Rapid Reductions (RCWR). Comparison with the reference case indicates significant reductions in delivered energy and CO2 emissions throughout the time horizon considered. The contrast in oil usage between the FFES, the "business-as- usual" scenarios and other policy scenarios is stark. In the year 2030, oil usage in the FFES is between 58 and 63 per cent lower than in the BAU scenarios. In one of the more radical policy scenarios produced by the US EPA (RCWR), though oil use falls rapidly to a similar level as the FFES in 2030, oil from polluting synthetic fuels increases rapidly. Overall oil usage increases over the period. This is double the level of the FFES in 2030, and increases throughout the remainder of the century. Overall oil use is nearly double that of 1990. Although the RCWR scenario reduces CO2 emissions more rapidly than the FFES scenario through 2030, there are three important distinctions to be made. First, the RCWR scenario relies far more heavily on biomass resources and includes biomass carbon sinks in its calculations of total scenario emissions. The RCWR scenario depends on very high biomass yields to achieve the 370 EJ and 610 EJ supplied in 2030 and 2100. It scenario embodies assumptions of biomass yield improvements of 150 to 500 percent over current biomass plantations. The ecological implications of such high yield plantations are uncertain; fertilizer and water use requirements could be very high. If the FFES biomass yield assumptions were used, the land use requirement for the RCWR scenario would consume up to approximately 30 percent of global current crop, pasture, forest and woodland area by 2100. The FFES emphasizes a far greater penetration of solar and wind supply. Second, the RCWR scenario assumes a growing contribution from nuclear energy throughout the study period (12 percent of primary supply in 2100). Finally, in RCWR, CO2 emissions begin to rise after 2050, and coal and oil use are increasing as the scenario heads into the 22nd century [4]. OTHER GREENHOUSE GASES The dominant GHG is CO2 , accounting for 61 percent of the global warming impact of all GHG emissions in 1990, followed by methane (CH4) at 15 percent, halocarbons (CFC's and HCFC's) with 11 percent, and nitrous oxide (N2O) at 4 percent (Shine et al, 1990). To complete the climate analysis, assumptions were made for the other GHGs. Considering the halocarbons first, CFC-11 and CFC-12 are already due to be phased out under the terms of the Montreal Protocol agreement because of their destructive impact upon stratospheric ozone. However, the mooted substitute gases are among the most potent greenhouse gases and constitute a new and potentially substantial addition to radiative forcing. They are used in a limited number of inessential or substitutable products and are likely to be the simplest greenhouse gases to control because of their narrow production base. For these reasons Greenpeace propose the complete phase out of radiatively significant halocarbons and substitutes. The sources and sinks of CH4 and N2O are not so well understood and accounted for. While confidence in the estimates for sources and sinks of methane is growing (Lashof, 1991), there is still a high level of uncertainty. Similarly, the sources and sinks of N2O are subject to wide margins of error such that it is very difficult to confidently quantify the impact upon atmospheric concentrations of any given control policy. In consequence, Greenpeace propose that, where practicable, steps are enacted to reduce concentrations of these gases. The result of these measures could be quite substantial, but for the purposes of the target setting exercise we adopt more conservative CH4 and N2O emissions assumptions taking the average of the Rapidly Changing World with Control Policies (RCWP) and Slowly Changing World with Control Policies (SCWP) scenarios (U.S. EPA, 1990). This essentially stabilises emissions of these gases. Other gases from the transport sector which have an impact on the atmospheric system were previously discussed in the sectoral results. THE CLIMATE BENEFITS OF THE FFES Modelling the results of FFES using STUGE demonstrates that it significantly reduces the risks of climate change. Based on the FFES total emission of 314 Gt carbon, carbon dioxide concentrations in the atmosphere are kept to below 400 ppm., in contrast to well over 750 ppm. in the Business-As-Usual (BAU) scenario. Under a climate sensitivity of 2.5oC for a doubling of pre-industrial carbon concentrations, global average temperature increases are kept from increasing above 1.5oC within the time period, in contrast to more than 4oC under BAU (see Figures 17 and 18). Temperatures are actually falling from the year 2050 onwards in the scenario. Even at a climate sensitivity of 4.5oC, global temperature are kept below 2.2oC. Rates of temperature increase are brought to below 0.1oC per decade around the year 2020. Sea level increases are kept between 10 and 35 cms, in contrast to 65cms. in the BAU. Additional modelling on the implications of delaying the FFES by between 10 and 30 years showed that for each 10 year delay in taking policy action, the planet would be committed to a further 0.4oC temperature increase. THE ECONOMIC IMPLICATIONS OF THE FFES The costs of a scenario developed over a time period of more than century are necessarily speculative. The projected costs of future energy systems crucially depends on assumed costs for the various fuels, the costs of implementing policies for energy efficiency and renewable energy, the type of computer model used, and assumptions about the current energy market. A number of authors have commented on the wide range of cost estimates for carbon dioxide abatement (Grubb, 1991, Boyle, 1992). The cost estimates for reducing USA carbon emissions by 20 per cent range from a net negative cost to $450 per tonne of carbon saved (Reilly, 1992, Boyle, 1992). A major difference in cost estimates emerges between those developed using 'top down' macro-economic models, and those using 'bottom-up' end-use models. The FFES utilised both types of model to give some indication of the costs. Considerable evidence suggests that implementation of the measures in this scenario could be achieved at modest cost, or even at a net economic benefit relative to continuation with a business-as-usual world. Projections of renewable energy supply costs indicate that solar, wind, and biomass technologies could be close enough to those of fossil fuels to enable a transition to occur without major economic penalties. Major economic benefits in avoided electric capacity requirements could result from investments in efficient end-use technologies. A full benefit analysis would include the avoided costs and capital requirements (e.g., dikes to stem coastal flooding) that would otherwise be needed to mitigate the impacts of global warming, if the world were to continue its current dependence on fossil fuels. The FFES was based on data supplied from a wide range of sources. It also utilised more than 100 national, regional and global reports on carbon dioxide reductions which have been produced in the past two years. For example, analysis of energy end uses in Thai commercial buildings found that efficiency measures could cut electricity and on-peak usage, and resulting electricity bills, in half. All of these measures were found to be cost- effective, based on a tariff of $.05/kWh, $9.16/kW (Busch, 1990). Using a detailed electric utility financial model, investments in conservation were shown to be 75 percent less capital intensive than avoided electric capacity investments (Busch, 1990). Several provided significant input to the development of the scenario. These include: a) A recent global study for the U.S. Working Group on Global Energy Efficiency (Levine et al., 1991). The study compared an Efficiency scenario with a Reference scenario over the 1985-2025 period, the latter based upon EPA's Rapidly Changing World scenario. The Efficiency case achieves 29 percent and 28 percent reductions in energy use and fossil fuel CO2, respectively. At the same time, the Efficiency scenario reduces cumulative 1985-2025 capital requirements, from $7785 billion to $4111 billion (1990 U.S.$). In capital-constrained developing and Eastern Europe countries, efficiency investments could cut capital requirements in half (projected decrease from $4,657 to $2,320 billion), thereby freeing up important resources for further development. b) Perhaps closest to reflecting the types of options -- renewable energy and energy efficiency -- and their levels of penetration shown in the FFES, is the America's Energy Choices study. (UCS et al., 1991). This study found that a 70 percent reduction in CO2 emissions could be achieved at a net cumulative savings of $2.3 trillion from 1990 to 2030. c) A study of the five largest and most developed Western European countries (France, FRG, Italy, Netherlands, and U.K.) found that CO2 emissions could be reduced by 18-41 percent below 1985 levels at a net energy cost savings of 13 to 27 percent relative to a European Community business-as-usual projection. (Krause et al., 1992) The annual savings by 2020 would be sufficient to provide each European resident with approximately $600 to $900 in savings each year (1989 U.S.$). A deeper CO2 reduction, `Minimum Risk', scenario indicates that emissions could be reduced 58 percent below 1985 levels by 2020, with energy cost savings of 2 to 12 percent. A criticism of end-use models such as LEAP is that fail to take account of energy pricing factors, the price equilibrium effect as demand falls for certain fuels, and that they underestimate the difficulties and costs of implementing energy efficiency programmes (Boyle, 1992). To partially counter this criticism, some of the assumed energy efficiency improvements were decreased in order to act as a proxy for achieving less than the technical and economic potential for energy efficiency, and to account for a 'take-back' effect, whereby some of the benefits in reduced energy consumption may be taken back by consumers through leaving lights on longer, or spending money on other energy consuming activities. The ASF model was used to gain additional information on the pricing implications of the FFES. This is a partial equilibrium model, which is commonly used throughout the world [5]. By feeding in a range of assumed fuel costs for renewable and fossil fuels, using US EPA figures in the main, the relative costs of the assumed 'business-as-usual' scenario and the FFES could be developed and compared. The results are shown in Figures 19 and 20. They indicate that, accepting the assumed costs for fossil fuels and renewable energy, the secondary costs of energy consumption for the FFES are considerably lower than the BAU. What the results do not show are the additional costs of the end- use efficiency equipment needed to achieve the FFES demand profile. This is the subject of current analysis by Greenpeace International. Based on a preliminary analysis of the costs of efficient and solar-hydrogen vehicles, super-insulated buildings, highly efficient industrial processes, lighting, appliances and boilers, it was concluded that the costs of the FFES are likely to be equal to or less than a 'business-as-usual' scenario. This is based on a future where consumers are paying more than at present for each unit of energy used, but using far less fuel for each unit of energy service derived. This concurs with a recent report for the United Nations which carries out a detailed economic evaluation of a range of renewable technologies which provide 50 per cent of global supplies by the year 2050 (UNSEGED, 1992). The report concludes that "the results of the scenarios can be taken as a kind of 'existence proof' showing that renewable energy supplies could provide by the middle of the next century of the order of half of total energy at costs that are roughly comparable to those for conventional energy supplies...". SENSITIVITY TESTING A wide range of sensitivity tests were carried out on the FFES. These included the following, relative to the central assumptions of the FFES: a) Lower levels of economic growth (measured by GDP), to assess the impact of 'lifestyle' changes and alternative development models; b) Lower levels of population; c) Less intensive use of materials and resources; d) The impact of higher and lower efficiencies of solar and biomass technologies; e) The impact of lower rates of improvement in energy efficiency; f) Variations in the absolute and relative costs assumed for fossil fuels and solar technology. Details of the sensitivities were as follows: a) 35 Percent Lower GDP by 2100. Under this sensitivity, energy use and CO2 emissions decline, as the total global economy grows more slowly that in the FFES. Projected global GDP is 25 percent lower by 2030 and 35 percent lower by 2100. No other parameters, including population, are varied. The FFES scenario's 2:1 richest to poorest region equity ratio for 2100 is maintained here, as is the relationship between the industrial, service, and agriculture sectors. The lower economic growth assumptions lead to a drop in primary energy requirements of 10 percent by 2030 and 20 percent by 2100. Cumulative global CO2 emissions decrease by 10 percent, from 314 to 284 Pg C. b) Intermediate Low Population, Same Total Regional GDP. This sensitivity varies population while assuming the same regional and global total GDP: a world with fewer but wealthier people. We use an intermediate low population forecast which projects a world population of 8.0 billion in 2100, 29 percent lower that the mid forecast used in the FFES scenario (Lovins et al., 1981, Meadows et al., 1992). Since we assume no resulting change in the scale of the world economy, average per capita income increases 42 percent, and the reduction in energy use (2100 consumption is down 5.6 percent) and CO2 emissions (cumulative emissions are down 2.2 percent) are relatively small. c) Low Population, Same Total Regional GDP. This sensitivity parallels the previous one, with an even lower population forecast. A world population of 6.4 billion in 2100 is projected, 43 percent lower that the mid forecast used in the FFES. (Bulatao et al., 1990). As in the previous sensitivity, since there is no change in the scale of the world economy, the reductions in energy use (2100 consumption is down 7.1 percent) and CO2 emissions (cumulative emissions are down 2.5 percent) remain small. d) Intermediate Low Population, Same Regional GDP per Capita. Rather than assuming the same total GDP by region, unaffected by population changes, as in the preceding sensitivities, here we assume that average per capita GDP remains the same as in the FFES. As a result, total GDP declines in direct proportion with population. The two sensitivity approaches -- same total GDP and same GDP per capita -- bracket a more likely outcome; i.e. that lower population growth would accompany increasing per capita income, but that total GDP would be lower than that of a more populous world. In this sensitivity, cumulative CO2 emissions decline 7.0 percent from the FFES to 292 Pg carbon over the 1988-2100 period. The 2100 energy supply requirements decrease by 22 percent, and land use requirements for solar, wind, and biomass would decrease by approximately this level as well. e) Low Population, Same Regional GDP per Capita. This sensitivity is identical to the previous one through 2030. The 2100 projected population, and - as a result of the constant GDP per capita assumption - total GDP, are about 43 percent lower than the FFES, compared with a 29 percent decrease in the previous scenario. Cumulative CO2 emissions are 9.2 percent lower than the FFES scenario. This decrease is surprisingly small, because the effects of slower population growth take place largely over the longer term, when energy supply systems are largely based on renewables. f) Slower Penetration of Improved Energy Efficiency. This sensitivity underscores the importance of achieving the technical and economic potential for improved efficiency. In this scenario, energy efficiency improvements are scaled back by one- third through 2030. From 2030 to 2100, a 0.5 percent per year efficiency improvement rate was assumed, as in the FFES. The result of this sensitivity is a substantial increase in CO2 emissions over the next 40 years relative to the FFES scenario. Projected annual emissions increase from 5.3 to 6.1 Pg C in 2010 and from 2.6 to 3.8 Pg C by 2030. Cumulative emissions rise to nearly 400 Pg C, a 27 percent increase. The results of sensitivities a) to f) are shown in Table 11 and Figure 21. A somewhat surprising result is that annual carbon dioxide emissions are little altered from the main FFES, with the exception of slower energy efficiency improvement rates. This does not lead to a conclusion that population and GDP levels have no impact on the results; it mainly reflects the already low levels of carbon emissions achieved in the main scenario. What the sensitivity tests show is the need to make an early start on measures to improve energy efficiency throughout society. The major impact is in reducing the global energy demand. This is reduced by between 20 and 35 per cent in the lower population and lower economic growth sensitivities. The main benefit here would be in reducing the intensity and ecological impact of renewable energy technologies. For example, the land area needed for solar and biomass technologies would be reduced from 400 to 260 million hectares in 2030, and from 826 to 540 million hectares in the year 2100 respectively. Under assumptions of improved efficiencies for solar and biomass technology, and greater use of biomass residues, the land area needed could be reduced to as low as 225 million hectares (2.5 per cent of the total land area). The implications of variations in cost assumptions for both fossil fuels and solar energy were tested by assuming 25 to 50 per cent lower production costs for fossil fuels, and 30 to 50 per cent higher costs for solar respectively. The major impacts are that fossil fuels are not phased out, and a residual 20 per cent of fossil fuel related carbon emissions remain. This would imply a much higher carbon tax than the maximum of $150/tonne C assumed in the FFES. The overall cost increases of the FFES under the higher solar costs sensitivity are relatively minor, a few per cent higher over the time period. It should be noted that this in part reflects the control of fuel shares through implied regulations such as carbon dioxide emission standards. POLICY IMPLICATIONS OF THE FFES Historically, it has taken new fuels some 50 years to capture 10 per cent of the global energy market (Marchetti, 1989). In the FFES, renewable energy sources increase from their current 14 per cent of total energy supplies to 60 per cent within the next 40 years. Strong policies will be needed to reach this target. A detailed consideration of the policy implications of achieving a fossil free energy future is beyond the focus of this paper. However, a number of key policy points emerge. Pricing Policies A perfect energy market requires full knowledge of all market elements and alternatives, equal access to capital, and free competition. The perfect energy market does not exist, as government and industry interfere with the market in numerous ways. In the last fifty years, legislation, pricing, and institutions have developed to favour fossil fuels and nuclear power. These now form market barriers, preventing producers and consumers from using new technologies to save money and utilise cleaner and sustainable energy systems. Realistic energy pricing will not on its own solve global warming. As part of a wider strategy, however, it is important in sending the correct signals for investment choices and removing subsidies. Policies to encourage this include: a) the introduction of energy taxes to reflect the full environmental costs of fossil fuels and nuclear power. A phased increase of energy costs to a level double the current oil price equivalent, is justified (Hohmeyer, 1988, PACE, 1989, NRDC, 1993); b) the introduction of tax credits for renewable energy developers (already occurring in Germany, UK, Italy, the Netherlands, Denmark, several states in the USA); c) ensuring that utilities buy clean renewable energy at a reasonable price (as occurs in parts of the US, Italy and Germany); d) changing the regulations under which gas and electricity institutions operate; specifically removing the financial incentives to sell increasing quantities of gas or electricity. e) removing the subsidies from the fossil fuel and nuclear industries - $44 billion per year in the USA alone (1984 prices), and $3.4 billion per year for company car tax breaks in the UK; f) ending oil and gas exploration tax breaks, as well as a range of other subsidies - In the UK, the projected costs of decommissioning North Sea oil drilling platforms were effectively subsidised to the tune of around 9 billion. Oil companies were absolved from full decommissioning of tthe platforms in the late 1980s, and the remaining costs san be claimed against tax. In most countries the oil exploration tax regime is very attractive in order to attract oil company interest. Intervening in the Market In addition to realistic energy pricing, regulation is needed to prevent cartels forming and the market from being manipulated. Regulation has already been shown to work in the US, Japan, and most Western European countries, in areas such as building standards, appliance efficiency, and safety. Policies should include: a) new mandatory efficiency standards for appliances, vehicles, buildings, industrial motors and other technologies, should be set at ambitious levels. Efficiency standards have already worked in several countries, including Germany, Japan, and the US. The technical potential for improving vehicle fuel efficiency is very large. Though not without its flaws, the Corporate Average Fuel Efficiency (CAFE) legislation in the US did achieve a doubling of new vehicle fuel efficiency between 1978 and 1986; b) government support for public transport through subsidies, tax incentives for companies wishing to move freight by rail rather than road, and disincentives to use private vehicles in urban areas; c) planning regulations which discourage major new road-building and encourage public transport; d) integrated resource planning (IRP), in which gas and electricity utilities are required to compare the cost of supplying energy with the cost of minimising demand before building new power stations, should be introduced; e) Demand Side Management (DSM) programmes could be encouraged, in which utilities meet peak demand by helping customers use less energy, rather than by building new power stations. DSM spend is doubling from $1 billion per year in the US to some $2 billion by 1995. $10 billion per year could be justified (Pringle, 1992); f) national and local government purchasing programmes for efficiency and solar equipment would help establish initial markets; g) in the former Comecon countries, plus many countries in the South, basic data on energy use in buildings, energy prices and equipment performance are almost non-existent. Efficiency centres for the collection of such data, and for testing equipment, should be set up. Research and Development There are a number of reasons why certain technologies achieve success and capture a large portion of the energy market. Improving efficiencies and reducing costs are important objectives for accelerating the impact of renewable energy technologies, and enhanced research, development and deployment (R,D & D) can assist in this. One study has estimated that an expenditure of some $3 billion over the next few decades could double the projected contribution of renewable energy over the period (SERI et al., 1990). This and other studies have suggested that the costs of alternative fuels for transport could be comparable with oil based fuels in the period 2000-2010 (UCS et al, 1991, DeLuchi et al, 1991). A new approach towards energy Research and Development is needed. International Energy Agency governments' energy R & D budgets are currently heavily skewed towards fossil fuels and nuclear power. Only 12.5 per cent of the total budget of $7,675 million is allocated to renewables and energy conservation: over 70 per cent is allocated to fossil fuels and nuclear power (see Figure 22). Eighty-nine per cent of the energy research budget of the European Community between 1987 and 1991 was devoted to nuclear fusion and fission technologies (CAN, 1991). As a minimum energy R&D budgets from current levels within ten years. Of this support for nuclear power and fossil fuels should be phased out. Changing Institutions None of the current energy institutions are guided by environmental concerns. At the international level, organisations exist to promote and develop oil use (OPEC), coal (the International Energy Agency), and nuclear power (the International Atomic Energy Agency). Transnational corporations promote and lobby for oil, coal, gas, and nuclear power. No international organisations exist for energy efficiency and renewable energy. In the past ten years, energy loans from multilateral development banks such as the World Bank have totalled more than $50 billion. Less than 1 per cent of this energy lending went to improving end-use energy efficiency, despite the better rates of return such investments give, compared to paying for new energy supplies. A major reassessment of energy institutions is needed if global energy policies are to change. New lending criteria for power sector loans, which encourage energy efficiency investments, would be a start. The creation of a new international agency for the development and promotion of technologies for renewables and energy efficiency (TREEs) has been proposed by Greenpeace and others. A TREEs agency could provide a focus for energy funding, R&D collaboration, technology transfer, education and information supply. It could also ensure that the United Nations, development banks, and other organisations develop appropriate policies and actions for a low-CO2 future. A tax equivalent to $1 per barrel on all non-renewable energy sources, applied in industrialised countries, could raise more than $30 billion per year to fund the agency (Goldemberg, 1990). International Climate Protection Policies One hundred and fifty four nations signed the International Climate Convention at the Earth Summit in June, 1992. Article 2 of the Convention commits signatories to the stabilisation of concentrations of greenhouse gases not controlled by the Montreal protocol, at a level which will prevent dangerous anthropogenic changes, and within a time-frame which will allow ecosystems to adapt naturally. Achieving this will require a range of strong Protocols, assisted with tough national and regional CO2 reduction targets. The discussion of Protocols should not be postponed until the Convention enters into force. A priority follow-up Protocol to the Convention would be an Energy Efficiency Protocol committing signatory countries to achieving annual improvements in energy efficiency over the next few decades. A challenging target would be an average of 2.5 per cent per year over the next forty years. A Renewable Energy Protocol should also be enacted committing signatory countries to expanding all environmentally sound sources of renewable energy. A challenging target would be 5 per cent per year. REFERENCES Bevington, R., Rosenfeld, A. (1990). Energy for Buildings and Homes, Scientific American. September. BMJ, (Read, R.C. and Green, M.) 1990. Internal Combustion and Health. British Medical Journal Vol 300, 24 March. Bulatao, R. A., Bos, E., Stephens, P. W., Vu, M. T. (1989). Population Projections. (WP5328-331) World Bank, Washington, D.C. November. Busch, J., DuPont, P., Chirarattananon, S. (1991). Conserving Electricity for Lighting in Thai Commercial Buildings: A Review of Current Status, Potential, and Policies, Lawrence Berkeley Laboratory, Berkeley, CA. Draft. Busch, J. 1990. From Comfort to Kilowatts: An Integrated Assessment of Electricity Conservation in Thailand's Commercial Sector, Lawrence Berkeley Laboratory, Berkeley, CA. August. CAN (Climate Action Network). 1991. Global Warming and the EC Budget. May. Commission of the European Communities. 1991. Cost Effectiveness Analysis of CO2 Reduction Options: Country Reports. Report for the Commission of the European Communities Directorate General XII, Germany. May. Cook, J. H., Beyea, J., Keelas, K. H. 1991. Potential Impacts of Biomass Production in the United States on Biological Diversity, Annual Review of Energy Environment. 16:401-31. Dessus, B., Devin, B., Pharabed, F. 1991. World Potential of Renewable Energies: Actually Accessible in the Nineties and Environmental Impacts Analysis. CNRS-PIRSEM. Paris. DeLuchi, M.A., Larson, E.D., Williams, R.H. 1991. Hydrogen and Methanol: Production from Biomass Use in Fuel Cell Internal Combustion Engine Vehicles, Princeton University. August. DOE (US Department of Energy). 1985. Projecting the Climatic Effects of Increasing Carbon Dioxide. DOE/ER-0237. December. Electric Power Research Institute. 1989. Technical Assessment Guide, Palo Alto, CA. September. Fickett, A. P., Gellings, C. W., Lovins, A. B. 1990. Efficient Use of Electricity, Scientific American. September. Fisher, D. 1990. Options for Reducing Greenhouse Gas Emissions. Stockholm Environment Institute, Stockholm. Fulkerson, W., Judkins, R., Sanghvi, M. 1990. Energy from Fossil Fuels, Scientific American. September. Geller, H. 1991. Efficient Electricity Use: A Development Strategy For Brazil, ACEEE, Washington, DC. Goldemberg, J. 1990. How to stop global warming. Technology Review. November/December. Greenpeace. 1992. Towards a Fossil Free Energy Future. Greenpeace International, Amsterdam (forthcoming Septmeber 1992). Greenpeace. 1991. An Environmental Audit of the Car. Greenpeace International, Amsterdam. Grubb, M.J. 1991. The Integration of Renewable Electricity Sources, Energy Policy. Butterworth-Henemann, Ltd. September. Grubb, M. 1990. Analytical and Economic Challenges in Assessing CO2 Abatement Costs and Policies. Proceedings of the Workshop on Economic/Energy/Environmental Modeling for Climate Policy Analysis, Washington, D.C. Wood, D. O and Kaya, T., eds. p. 549. October. Hall, D. O., Rosillo-Calle, F., Senelwa, A. K., Woods, J. 1992. Biomass for Energy: Future Supply Prospects, in Fuels and Electricity from Renewable Sources of Energy, Island Press, Washington, D.C. Hohmeyer, O. 1988. The Social Costs of Energy Consumption. Foor the European Community (DG XII). Springer-Verlag. ICF, 1990. Atmospheric Stabilization Framework: User's Guide and Software Description. ICF, Inc., Fairfax, VA. Intergovernmental Panel on Climate Change (IPCCa). 1990 and 1992 update supplement. Scientific Assessment of Climate Change, Report from Working Group I. August. Intergovernmental Panel on Climate Change (IPCCb). 1990. Potential Impacts of Climate Change, Report from Working Group II. August. International Energy Agency. 1991a. Greenhouse Gas Emissions: The Energy Dimension. Paris, France. International Energy Agency. 1991. Energy Policies and Programmes of IEA countries: 1990 Review. OECD. Paris. Kayes, Roger J. 1991. A Review of the Current and Potential Future Significance of Trees in the Global Crabon Cycle, Greenpeace International, Oxford, UK. October 30. Keepin, B. 1986. Review of Global Energy and Carbon Dioxide Projections, Ann. Rev. Energy, Vol. 11, pp. 357-392. Krause, F., Koomey, J., Bleviss, D., Olivier, D., Onufrio, G., et al. 1992. Energy Policy in the Greenhouse, Vol 2. International Project for Sustainable Energy Paths, El Cerrito, CA, Vol. II, preliminary draft. Lashof, D.A., 1991. Discussion paper for the Climate Action Network, "Methane Controls in the Climate Convention", Natural Resources Defence Council, September. Levine, M., Gadgil, A., Meyers, S. et al. 1991. Energy Efficiency, Developing Nations, and Eastern Europe, International Institute for Energy Conservation, Washington, D.C. June. Lovins, A., Lovins, L., Krause, F., Bach, W. 1981. Least-Cost Energy: Solving the CO2 Problem, Rocky Mountain Institute, Snowmass, Colorado. Manne, A. S., Richels, R. G. 1990. Global CO2 Emission Reductions - The Impacts of Rising Energy Costs, Electric Power Research Institute, Palo Alto, CA. February. Marchetti, C. 1989. How to solve the CO2 problem without tears. In proceedings of International Energy Agency Seminar on Energy Technologies for Reducing Emissions of Greenhouse Gases. pp161- 191. OECD 12-14 April. Meadows, D. H. et al., 1992. Beyond the LImits: Global Collapse or a Sustainable Future. Earthscan Publications Ltd. London. Mintzer, I. 1988. Projecting Future Energy Demand in Industrialized Countries: An End-Use Oriented Approach. MIT Center for Energy Policy Research, Cambridge, MA. October. Nadel. S, 1990. Lessons Learned: A Review of Utility Experience with Conservation and Good Management Programs for Commercial and Industrial Customers. ACEEE. Washington. Nuclear Engineering International. 1991. World Nuclear Industry Handbook, Sutton, England. Office of Techology Assessment. 1991. Changing By Degrees: Steps to Reduce Greenhouse Gases, Washington, D.C. July. Ogden, J. M., Williams, R. H., Fulmer, M. E. 1990. Cogeneration Applications of Biomass Gasifier/Gas Turbine Technologies in Cane Sugar and Alcohol Industries, Princeton University. March. Ogden, J. M., Williams, R. H. 1989. Solar Hydrogen: Moving Beyond Fossil Fuels, World Resources Institute, Washington, D.C. PACE (Pace University Centre for Environmental Studies), 1990. Environmental Costs of Electricity. Oceana Publications, New York. Prindle, W.R. 1990. Demand-Side Management in the 1990s: Time to Come of Age. The Alliance to Save Energy. Washington. Reddy, A. K. N. 1991. Environmentally Sound Energy Development: A Case Study of Electricity for Karnataka State. Presented at Regional Conference on Environmental Challenges for Asia-Pacific Energy Systems in the 1990s, Kuala Lumpur (Malaysia). January 14-18. Rowland, F.S. and Isaksen, I.S.A.. 1988. The Changing Atmosphere. Dahlem Workshop Reports No. 7. SCOPE. 1986. The Greenhouse Effect, Climate Change and Ecosystems. Bolin et al.,(eds.). Wiley Books. Shine, K.P., Derwent, R.G., Wuebbles, D.J. and Morcrette, J-J., 1990. Radiative Forcing of Climate. In, Climate Change: The IPCC Scientific Assessment (Eds. J.T. Houghton, G.J. Jenkins and J.J. Ephraums), Cambridge University Press, p41-69. Solar Energy Research Institute, et al. 1990. The Potential for Renewable Energy: An Interlaboratory White Paper, Golden, CO. March. SCAQMD (Southern California Air Quality Management District), 1989. Air Quality Management Plan and Policy Proposals. Stern, J. 1991. Prospects for Natural Gas in Europe. Gower Books. Subak, S., Raskin, P., Von Hippel, D. 1991. National Greenhouse Gas Accounts: Current Anthropogenic Sources and Sinks. Stockholm Environment Institute- Boston, Boston, MA, December. Swart, R. J., Pepper, W. J., Ebert, C., et al. 1991a. Emissions Scenarios for the Intergovernment Panel on Climate Change, An Update, Background Paper and Work Plan, Bilthoven, Netherlands. April. Swart, R. J., Pepper, W. J., Leggett, J., et al. 1991b. Emissions Scenarios for the IPCC, An Update, Bilthoven, Netherlands. October. Union of Concerned Scientists, Natural Resources Defense Council, Alliance to Save Energy, American Council for an Energy Efficient Economy. 1991. America's Energy Choices. Union of Concerned Scientists, Cambridge, MA. UNSEGED (United Nations Solar Ebnergy Group for Environment and Development). 1992. Renewables for Fuels and Electricity. Johannsson, T. et al., (eds.). June. U.S. Environmental Protection Agency. 1987. Assessing the risks of Trace Gases that can modify the atmosphere. Vol. 1. Executive Summary. EPA 400/1-S7/001A U.S. Environmental Protection Agency. 1990a. Country Study Base Case: Summary Paper and Comparison. February. U.S. Environmental Protection Agency. 1990b. Policy Options for Stabilizing Global Climate, Report to Congress, Washington, D.C. December. Waide, P. 1992. Greenhouse Modelling and Emission Targets. Technical Report for Greenpeace International study Towards a Fossil Free Energy Future. March. Walsh, M. 1992. Global Transport Scenarios. Input to Greenpeace International study Towards a Fossil Free Energy Future. March. Wigley, T.M.L. and Raper, S.C.B., 1987. Thermal Expansion of Sea Water Associated With Global Warming, Nature 330, p127-131. Wigley, T.M.L. and Raper, S.C.B., 1990. Natural Variability of the Climate System and Detection of the Greenhouse Effect, Nature 344, p324-327. Wigley, T.M.L., Holt, T. and Raper, S.C.B., 1991. STUGE (an interactive Greenhouse Model): User's Manual, Climate Research Unit, University of East Anglia, Norwich, U.K., 44pp. Williams, R. H., Larson, E. D. 1991. Advance Gasification-Based Biomass Power Generation and Cogeneration. Presented at ESETT'91 Conference, Milan, Italy. October 21-25. Williams, R. H. 1990. Will Constraining Fossil Fuel Carbon Dioxide Emissions Really Cost So Much?, Princeton University. April. World Energy Conference, Conservation and Study Committee. 1989. Global Energy Perspectives 2000-2020, Montreal, Canada. World Meteorological Organization/United Nations Environment Programme (Jaeger,J. et al.,). 1988. Developing Policies for Responding to Climatic Change. (WMO/TD-No.225). April. World Resources Institute. 1990. World Resources 1990-91, Oxford University Press, New York/Oxford. Yamamura, S. 1991. Paper for the Regional Conference on Environmental Challenges for the Asia-Pacific Energy Systems in the 1990s (unfinished manuscript). APPENDIX A FOOTNOTES******************************** {1} Reserves estimates are derived from an analysis by Jonathan Stern, August 1991, in a report to Greenpeace International. ----------