|The Global Greenhouse Regime. Who Pays? (UNU, 1993, 382 p.)|
|Part III National greenhouse gas reduction cost curves|
|10 Abatement of carbon dioxide emissions in Brazil|
Brazil energy economy
Energy subsector analyses
Changing land-use trends
Jose Moreira and Alan Poole
This chapter concentrates on energy-related carbon dioxide (CO2) emissions, the principal greenhouse gas, of which Brazil is an important contributor at a global level. The clearing of forest land for non-energy uses is its largest source of CO2 emissions. Quantitative estimates of this source are controversial, however. It is not possible to be precise as to the amount and costs of carbon conservation associated with land use changes in Brazil. For this reason, this chapter will focus on the abatement of fossil CO2 emissions released from the energy sector. But we will also discuss policies to decelerate deforestation and outline some crucial interrelationships between land use, energy policy, and carbon conservation strategies.
Contemporary energy related emissions of CO2
In 1990, Brazil's total energy consumption reached 183.7 million tonnes of oil equivalent (MTOE). In that year, 37 per cent of Brazil's primary energy came from fossil fuels, 26 per cent from biomass fuels and 37 per cent from hydroelectricity (see Table 10.1). Important features of Brazil's energy balance are: the preponderant role of hydropower in electricity generation (94.4 per cent); the relatively large amount of biomass that is used at an industrial scale (alcohol, charcoal and wood - see Table 10.2); and the small penetration of natural gas.
The energy system resulted in 1990 in the emission of approximately 51.6 million tonnes of carbon (TC) as CO2 from the combustion of fossil fuels (Table 10.3).
Table 10.1 Energy supply and consumption, Brazil 1990 (M TOE)
|Coal||Natural gas||Petroleum||Subtotal fossil||Biomass||Hydro||Nuclear||Other primaryb||Electricity|
|Gross internal supply||9.21||3.75||55.06||68.02||47.33||67.75||0.58||183.68||n.a.|
|Other energy sector||0.82||1.16||1.93||3.91||13.78||-||-||17.69||10.88|
|Final energy demand||7.18||2.00||43.04||52.55||32.50||-||-||85.05||60.86|
Source :Boletim do Balanco Energetico Nacional,1991.
Data for fuel include derivatives of the primary energy source (for example, coke from coal, alcohol from sugarcane). Electricity (last column) is calculated assuming that 1 MWhe = 0.29 TOE.
a Includes transformation and other losses and energy use (for example, in
b For gross internal supply and electricity generation includes all energy consumption, thereafter electricity is excluded. The actual inputs for electricity generation are included in this column, since these are different than the coefficient of 0.29 TOE used in last column. There is a slight discrepancy.
Lessons from climate change in Brazil
Each year, much human suffering has been caused by climate hazards in Brazil and billions of dollars have been lost ... The Brazil study concludes that human actions have inadvertently led to increased societal vulnerability to climate variations. Deforestation in the Northeast has made the semiarid region more vulnerable to droughts. Inadequate urban planning in Rio de Janeiro has made the city much more vulnerable to floods. Deforestation in the Upper Paraguay River Basin may be altering the pattern of floods and droughts in the Pantanal (Great Swamp) region, thus contributing to ecological imbalance.
On the other hand, some action has been taken to increase resilience to climate variations. The modernization of the salt industry in the state of Rio Grande has made that industry more resistant to heavy rains. Agricultural research has led to the development of several new crop varieties more resistant to climate variability. Relief action in Northeast Brazil has [reduced] the very heavy impacts of droughts on the poor rural population.
It is clear that the same climatic event may have different impacts according to local socio-economic and environmental characteristics. Rainfall that brings terrible floods to Rio de Janeiro is a beneficial event in the Pantanal area. Drought may cause huge losses to agriculture in the Northeast but can also bring economic benefits for the salt industry in the same region. Moreover, while the majority of the poor population will suffer from droughts, a small group of large land owners and businessmen may indeed profit from them. Climatic variations can thus have differing effects on different regions, ecosystems, economies and social classes.
Government policy needs to integrate short-term relief actions during extreme climatic events with long-term actions aimed at increasing societal resilience to climate variability and change. It also needs to pursue a goal of sustainable development, by seeking to increase the technological capacity of people to face climatic extremes, reduce the social impacts of the weather, reduce poverty (since the poor are the most vulnerable) and improve our knowledge and thus increase our capacity to accommodate adverse variations of climate. A policy of sustainable development would pursue both a reduction in greenhouse gases emissions and an improvement of the capacity of the environment to adapt to possible future climate changes. To achieve this will require international cooperation to enable the transfer of resources and technology that will allow developing countries to use the most environmentally appropriate available technologies to their development process.
Excerpt from M Parry, A Magalhaes, Nguyen H Ninh, The Potential Socio-Economic Effects of Climate Change, A Summary of Three Regional Assessments, UN Environment Programme, Nairobi, Kenya, 1991, p. 12.
Table 10.2 Biomass supply and consumption, 1990 (MTOE)
|Wood for charcoal||Other wood||Sugarcane||Otherb biomass||Total|
|Gross internal supply||12.31||15.14||18.14||1.74||47.33|
|Other energy sectora||6.43||0||7.41c||0.06||13.78|
|Final energy demand||5.88e||15.14||10.17||1.31||32.50|
Source: Boletim do Balanco Energetico Nacional, 1991 .
a includes transformation losses and fuel use; specifically losses in producing charcoal from wood and alcohol from cane;
b Basically pulp mill liquor;
c Sugarcane bagasse;
To simplify our analysis, we assumed that hydroelectricity has zero CO2 emissions. Similarly, we assumed that sugarcane supply (alcohol and bagasse) and 'other renewables' (pulp industry liquors) also produce zero CO2 emissions. In the case of alcohol, however, we included fossil fuel inputs to alcohol production.
Estimates of carbon emissions from wood uses are problematic and vary widely according to end use sector. Most wood used in the residential and agricultural sectors does not appear to contribute to deforestation. The relatively dispersed use in rural areas does not exceed natural regeneration. In this regard, Brazil's fuelwood problem is unlike that in many other countries. Little information is available to quantify deforestation in the areas that it occurs. In this study, we simply assumed that 20 per cent of the fuelwood used in the residential and agricultural sectors results in net CO2 emissions.
Conversely, fuelwood used in the industrial sector is more geographically concentrated. Demand from this subsector of fuelwood consumption often destroys natural forests in Brazil. Again, precise information as to rates and magnitudes of this phenomenon are scarce. We have assumed that 60 per cent of industrial fuelwood contributes to CO2 emissions. Charcoal, however, presents a special case requiring the analyst to adopt additional assumptions. The impact on the forests of converting wood to charcoal (the largest use of fuelwood) is highly controversial. An estimated one third only of Brazil's charcoal is supplied from planted forest. The remainder is made with fuelwood obtained by clearing the natural forest, mostly the drier forests of the cerrado. Most of this land is cleared for grazing and/or agriculture of which charcoal is a mere by-product. We assume, therefore, that 75 per cent of the charcoal from natural forest results in deforestation. Thus, we estimate that roughly 50 per cent of fuelwood used to make charcoal results in net CO2 emissions.
Table 10.3 CO2 emissions from energy consumption (MTC)
|Gross internal supply||51.64 (59.77)e||11.44||63.08 (71.21)e|
|Other energy sector||3.29||3.31c||6.60|
|Final energy demand||46.11||8.13||54.24|
a For conversion factors see .
b Sugarcane and 'other biomass' assumed to have no CO2 emissions. For wood it is assumed that deforestation results from 20% of residential, commercial and agricultural use; from 60% of industrial fuelwood use; and from 50% of charcoal use. These shores are somewhat arbitrary as discussed in the text.
c All due to conversion of fuelwood to charcoal.
d CO2 equivalent of fossil fuel used, if they were burnt. See text.
e Includes non-energy CO2 equivalent in parenthesis. (See note d above).
On this basis, about 40 per cent of total fuelwood use generates annual CO: emissions of 11.4 millions of tonnes of carbon as carbon dioxide (TC) or 22 per cent of that resulting from fossil fuel use. Evidently, fuelwood is not yet a renewable resource in Brazil. In spite of its importance, fuelwood-related carbon emissions constitute only 3-4 per cent of a total deforestation-related emission each year of about 300 million tonnes of carbon (MTC).
Total (fossil fuel plus biefuel) energy-related emissions of CO2 in 1990 are estimated to be about 63 MTC. This figure increases to about 71 MTC if the non-energy use of fossil fuels is valued at its combustion equivalent. This figure is roughly 470 kgC/capita per year.
Existing energy scenarios and their impacts
An official study was recently performed in Brazil as a basis for policy recommendations (SNE 1991a). This study analysed two basic policy postures. The first, a reference scenario called 'tendencies', assumed that current political and economic constraints in the energy sector (including those on pricing policy) will persist. The second scenario, called the 'alternative', postulated substantial changes in the status quo. This latter scenario sought to actively pursue greater efficiency in the overall use of energy resources; to promote competition and private sector investment in energy supply; and to stimulate the use of certain energy forms in some applications - principally biomass, but also coal and natural gas.
Relative to 'tendencies', the 'alternative' scenario contains moderate reductions in energy use. In both scenarios, lower and higher economic growth cases were studied. In the alternative scenario, primary energy use falls approximately 9 per cent by 2000 and approximately 18 per cent by 2010 in both the lower and the higher economic growth cases (see Tables 10.4 and 10.5). The fact that the reduction is proportionately the same in both the lower and higher growth cases is somewhat odd. One might expect a proportionately greater reduction in energy use in the higher economic growth case, since faster growth permits the more rapid penetration of new technology in the marketplace. Other key differences between the alternative and tendencies scenarios are that in the former, hydroelectricity and petroleum products are projected to grow less, and biomass use to grow more in the latter.
Total fossil fuel emissions are 12 per cent less in the low growth alternative case relative to the 'tendencies' in 2000 and 21 per cent less in 2010. However, the fossil fuel emissions in the lower growth version of the 'alternative' scenario are still 42 per cent and 103 per cent higher than in 1990. Transport offers the biggest abatement in fossil fuel CO2 emissions in the National Energy Secretariat (SNE) scenarios. Even so, this sector remains the biggest single source of carbon (see Table 10.6).
Fossil carbon emissions from electricity generation do not fall in the alternative scenario, despite substantial reduction of total electricity consumption. Indeed, thermal generation increases substantially, though this growth is mostly the result of increased use of biomass (mainly from sugarcane). Thus, the 'alternative' scenario retains the significant increase in fossil fuel emissions found in the 'tendencies' scenario. From 9 TC/GWhe today (Moreira 1991a), the carbon intensity of electricity increases to 23 TC/ GWhe in 2000. Even so, this ratio is still very low by international standards.
Table 10.4 Official energy scenarios - basic characteristics
|Low economic growth||High economic growth|
|Gross domestic product (1990=100)||37.3||85.3||100||146.3||238.3||146.3||238.3||163.5||292.5||163.5||292.8|
|GDP per capita (1990=100)||58.5||105.7||100||122.6||172.7||122.6||172.7||137.0||212.2||137.0||212.2|
|Primary energy consumption (MTOE)||74.7||137.1||183.7||266.5||407.3||241.8||330.3||288.1||473.9||262.9||386.6|
|Fossilfuel CO2 emissions(MTC)a||24.3||52.0||59.8||95.4||153.9||84.6||121.5||104.1||181.8||92.6||143.9|
|Fossil fuel emissions (1990=100)||40.6||87.0||100||159.5||257.4||141.5||203.2||174.1||304.1||154.8||240.6|
|Fossil fuel emissions per capita (kg C/cap.)||253||429||397||532||742||471||586||580||876||516||693|
|Fossil fuel CO2 emissions per unit |
of GDP (1990=100)
Source projection from Reexame de Matriz Energtica Nacional, Ministerio
da Infraestrutura, Brasilia, 1991.
a Based on gross internal supply of fossil fuels and thus includes a non-energy component.
Table 10.5 Sources of energy supply-official scenarios(16)
|Gross internal supply(MTOE)||139.2||183.7||288.1||473.9||262.9||386.6|
Sources: Reexame da Matriz Energca Nacional; Balanco Energca
High economic growth case.
The 'alternative' scenario projects large economic benefits. Energy sector investments, for example, fall by US$26 billion between 1991 and 2000 and $58 billion between 2001 and 2010 (in the high economic growth case) (SNE 1991a). Moreover, almost all of the carbon abatement in the alternative scenario result from 'no regrets' policy measures. That is, these steps are justified on non-greenhouse grounds (Ayres 1991), and accounting for CO: emissions only makes already economically viable changes even more desirable. Thus, these steps should be taken regardless of the scientific and other uncertainties surrounding greenhouse phenomena.
We have briefly described these scenarios to delineate the CO2 emissions implied by official energy planning. The SNE study is also the only recent example of an integrated energy analysis in Brazil. This planning did not emphasize CO2 abatement, however. New scenarios are needed badly to evaluate the costs of carbon abatement. The following end-use analysis is a first step in that direction.
Table 10.6 FossiI CO2 emissions by activity
|Million tonnes of carbon||Percentage|
|Thermal electricity generation||2.24||7.73||7.72||4.3||9.2||10.4|
|Other energy sector use and losses||3.97||7.39||6.51||7.6||8.8||8.8|
|Subtotal energy sector||6.21||15.12||14.23||11.9||18.0||19.2|
|Subtotal final fuel demand||46.11||68.81||59.65||88.1||82.0||80.8|
|Total fossil fuel||52.32||83.93||73.88||100||100||100|
Source: based on Secretaria Nacional de Energia, Reexame da Matriz Energca Nacional, 1991. Low economic growth scenario.
Methodology of subsector analysis
An important input for CO2 emissions abatement policy formulation is the analysis of sets of energy end-uses and transformations that constitute the various fuel cycles. This approach is not widely used in Brazil, and is most advanced in the electrical sector. In sectors where quantitative analysis is rudimentary, we will make qualitative observations. The time horizon for quantitative analysis is the year 2000. While politically realistic, this horizon is quite short for energy planning or carbon abatement analysis.
To construct a CO2 emission abatement cost curve entails that we address a series of methodological issues in the Brazilian context. These include: the reference scenarios; and the parameters and scope of cost-benefit analysis. Choosing a reference scenario is not trivial - especially in Brazil where key economic and political factors are highly uncertain. The reference scenario greatly influences the estimate of abatement potential. Here, we take the socio-economic assumptions and the energy projections of the low growth, 'tendencies' scenario (described above) as our reference scenario.
Cost-benefit parameters used in the analysis must be comparable to those used in other countries, or subject to sensitivity analysis. A key parameter is the cost of capital, which is higher in Brazil than in the industrialized countries. We assume a 12 per cent discount rate. This discount rate is quite high but it reflects Brazil's high debt burden and the scarcity of internal investment capital.
The scope of the cost-benefit analysis can greatly affect the evaluation of net costs and benefits. This issue is relevant to quite specific technologies (such as vehicles) as well as to broader system changes (such as shifting transport modes). Analysis of the latter type of changes, however, is particularly susceptible to changes in analytical scope. The narrowest definition of benefits is the energy saved or substituted. We always incorporate such benefits in this study. In cases such as electrical equipment, this definition of benefits is adequate. But in others, it gives an erroneous estimate of net costs. Governments, for example, undertake many energy-related investments to achieve multiple objectives, including stimulating productivity, improving the quality of products, responding to environmental constraints, even improving social welfare (as with much rural electrification or public transport). In such cases, much broader definitions of benefits enter the picture which may outweigh the values attributed to energy savings or carbon abatement.
Finally, we used current 'frozen' costs in 2000. That is, our analysis of energy savings (and associated carbon reduction) in 2000 is based on current new technology costs and current energy prices. We selected only more efficient technologies that are already economic compared with existing technologies. We also assumed no increase in the real price of oil, oil derivatives, and electricity in 2000. This approach is conservative as the cost of new technologies often falls as it enters service; and the price of fossil fuels may be expected to increase in real terms as it becomes increasingly scarce.
Energy sector use and transformation
Electricity generation in Brazil is dominated by hydropower (94 per cent). As already noted, we assume that hydropower generates zero net CO2 emissions. We expect that this share will fall by 2000. In the official SNE scenarios, the share falls from 94 to 85-88 per cent (ELETROBRAS 1990). Here we assume a moderate shift to thermal generation such that hydropower falls to 90 per cent of total generation by the year 2000. By then, 20 per cent of newly available generating capacity coming on-line would be thermal (of which 10 per cent would be coal-fired; 40 per cent would be natural-gas-fired; 30 per cent would be fuel-oil- and refinery residues-fired; and 20 per cent would be biomass-fired). We estimate that generating 1 GWhe would result in 37.2 TC emissions in Brazil in 2000, an increase from today's 9 TC/GWhe. This coefficient would be very low by international standards. The analysis also assumes that 16 per cent of total generated electricity is lost in transmission and distribution losses incurred in delivering power to the residential, commercial and public illumination sectors.
The shift to greater thermal generation is a response to a series of problems faced by the power sector. These include: a profound financial crisis (hydropower is capital intensive); the large market uncertainty confronting supply planning (larger hydropower has long lead times); and the need to attract more private sector investment in generation, which tends to favour thermal power plants. A more hydropower-intensive scenario is conceivable but we have not reviewed this possibility as it would require extensive, system-level analysis beyond the scope of this chapter. It is noteworthy, however, that Brazil's thermal electrical sector would not be very carbon-intensive.
Other energy transformation
This end use includes all energy inputs to extract and process fuel oil rigs, refineries, sugar plantations, etc. Half of the sectoral carbon emissions are due to the conversion of fuelwood to charcoal. As is generally the case with biomass fuels used in industry, the production of charcoal is inefficient. This inefficiency can be greatly reduced; this must be done in any case to ensure that fuelwood from sustainably managed forests for charcoal manufacture is economically viable. Reducing conversion waste in the charcoal fuel cycle will deliver the biggest abatement of carbon emission in the subsector. Although less dramatic, incremental improvements in efficiency are also available in most industries. These small but pervasive gains would accumulate into significant reductions of fossil fuel usage and carbon emissions.
Electricity final demand
Electricity generation will remain a relatively small source of CO2 emissions for many years in Brazil. Yet it is still desirable to reduce these emissions by demand side management. Many of these opportunities to increase efficiency have a large 'negative cost.' exploiting these opportunities will also relieve the financial pressures felt by the electrical sector in Brazil.
The residential sector consumes about 22 per cent of total electricity and contributes 30-37 per cent of the electricity evening peak. We have considered a range of technical options that could be implemented in this sector up to 2000. For each of the following final uses, we estimated the carbon abatement that can be obtained by improving electricity end-use efficiency: lighting, water heating, refrigeration and air conditioning. In Table 10.7, we show the technical and economic specifications for the new technologies and those being replaced; the amount of carbon abated per year per unit of equipment; the levellized annual cost of the new technology; and, in the last column, the net costs when the annual value of the electricity saved is deducted from the levelized costs.) A negative figure means that the new technology brings economic savings, even before considering carbon credits.
Table 10.7 Residential sector technologies to limit CO2 emissions implementable by 2000
|Old technology (1)||New technology (2)||Avoided C (kg/y/unit) (3)||Levelized annual cost ($/kg C) (4)||Annual cost of saved elec. ($/kg C )(5)||Net annual cost ($/kg C) (d) = (4-5)||Total C avoided (MTC/y)||Total net cost (million $/y)|
|Incandescent bulb (standard)||Efficient incandescent||0.260||0.192||1.989||-1.797||0.0385||-69.26|
|Incandescent bulb (standard)||Fluarescent (standard)||1.560||0.485||1.989||-1.504||0.1386||-208.40|
|Incandescent bulb (standard)||Compact fluorescent||1.910||1.338||1.989||-0.651||0.1129||-224.50|
|Electric shower||Solar shower||17.350||4.146||1.989||2 157||0.2563||553.00|
|Refrigerator (standard)||Efficient refrigerator||5.210||2.255||1.989||0.266||0.0857||22.80|
|Air conditioner (standard)||Efficient air conditioner||7.460||1.297||1.989||-0.692||0.0170||-11.80|
Based on the present stock of residential lamps (220 million incandescent lamps and eight million fluorescents), we estimate the stock in 2000. We assume that 50 per cent of old, inefficient incandescents will be replaced by efficient incandescents; 30 per cent by fluorescents; and 20 per cent by compact fluorescents. Such a programme would reduce carbon emissions by 0.29/MTC in 2000, providing a total net annual levelized benefit of US$502 million.
In 1987,75 per cent of Brazil's 26.3 million electrified households had electric showers. Assuming that the number of showers increases by 2.1 per cent per year up to 2000 and further, that solar water heaters achieve a 50 per cent penetration level in that year, we estimate that another 0.256 MTC can be conserved at an annual net cost of US$553 million.
Refrigeration accounts for about 33 per cent of total residential electricity use and is very wasteful (Jannuzzi and Schipper 1991; Geller 1991). We projected the stock of refrigerators in the year 2000 based on today's stock of 27 million. Assuming that 50 per cent of the stock in 2000 is efficient, we estimated that 0.086 MTC can be abated at a total net cost in that year of US$22.8 million.
Table 10.8 Commercial and public sector technologies to limit CO2 emission* implementable by 2000
|Old technology(1)||New technology(2)||Avoided C (kg/y/unit)(3)||Levelized annual cost ($/kg C) (4)||Annual cost of saved elec. ($/kg C) (5)||Net annual cost($/kgC) (6) = (4-5)||Total C avoided (MTC/y)||Total net cost (million $/y)|
|Incandescent bulb (standard)||Efficient incandescent||0.830||0.229||2.690||-2.461||0.0053||-12.95|
|Incandescent bulb (standard)||Fluorescent (standard)||4.330||1.309||2.690||-1.381||0.0618||-111.16|
|Incandescent bulb (standard)||Compact fluorescent||2.290||1.345||2.690||-1.345||0.0485||-96.26|
|Fluorescent (standard)||Efficient fluorescent||1.000||0.990||2.690||-1.700||0.0645||-109.68|
|Incandescent bulg (standard)||Mercury/sodium||15.050||1.928||1.548||0.380||0.015||5.74|
a Includes lamp (8000h), ballast (12000h) and light fixture (20000h).
b Average cost of mercury and sodium package with 80% mercury lamps.
We assumed that 2.85 million air conditioners will be in use in residences in the year 2000. Of this total, 80 per cent will be installed after 1990. We further assumed that all will be energy saving units. On this basis, 0.017 MTC can be conserved in 2000 at a total net benefit of US$11.8 million per year.
Commercial, service sector and public lighting
Total energy consumption of the commercial and services sectors has increased threefold during the past fifteen years, and electricity accounts for more than 90 per cent of the increase. We considered only opportunities to increase the efficiency of lighting. In Table 10.8. we present the technical and economic data for the new and old technologies, and quantities and costs of carbon abatement. We assumed the current tariff for commercial electricity of US$0.10/kWhe. (In the future, we would prefer to use an estimate of the economic cost of electricity). We also assumed that 26 million incandescent and 48 million fluorescent lamps were used in these two sectors in 1990 (Jannuzzi et al 1991). We postulated a 2 per cent annual growth rate for incandescent sales and 3 per cent for fluorescents (taken as 40W lamps). We estimate that all the incandescent lamps could be replaced in the year 2000 as follows: 20 per cent by krypton-filled bulbs; 45 per cent by standard fluorescents; and 35 per cent by compact fluorescents. This substitution would conserve 0.116 MTC of emissions and increase total net benefits that year by US$220 million. Replacing all fluorescents of 40W with more efficient 32W units with electronic ballasts would conserve another 0.064 MTC, and increase total net benefits by US$110 million.
In Table 10.8, we also present results for improvements in public illumination. If 800,000 incandescent lamps are replaced by mercury lamps and 200,000 by sodium lamps, then 0.015 MTC could be conserved at an annual cost of US$5.7 million.
To calculate industrial electricity demand in 2000, we assumed an average industrial economic growth of 3.5 per cent per year (as suggested by the low growth 'tendencies' scenario) and the same energy intensity (total energy/ GNP) as in 1988. We analysed electricity consumption for each industrial subsector and factored in technological improvements believed to be currently economic. We included the following measures: housekeeping; efficient lighting; and more efficient electric motors, variable speed electric motors, electric ovens, and electrolytic processes. We priced industrial electricity at an average US$0.0581kWhe. In Table 10.9, we list technical and economic data for these improvements.
Table 10.9 Industrial sector technologies to limit CO, emissions, implementable by 2000
|Annual cost of |
|Net annual cost |
|High efficiency motor||0.115||383||1560||-11 77|
|Variable speed drivers||0.197||357||1560||-1203|
|Electric ovens and boilers||0.152||197||1560||-1363|
From many evaluations performed by electric utilities (CEMIG 1989; CESP 1990) it is clear that electricity demand can be reduced by 10 per cent with better end-use management. These measures should include:
a better choice of electric motor size, especially by avoiding oversized motors which are common (Latone e' al 1990);
appropriate design of the factory's internal electric distribution grid;
installation of small size transformers in parallel with the main one, to be used during idle factory periods;
correction of the load factor;
avoidance of short term peak demand (through the use of demand controllers); and
better mechanical coupling between electric motors and the equipment driven by them.
These actions would conserve 0.32 MTC for a total annual net benefit of US$368 million.
The industrial sector contains about 14 million fluorescent 40 watt lamps. We assumed that this stock grows by 3.5 per cent per year and that all new lights will be the 32 watt efficient type (with new ballast). These steps would conserve 0.020 MTC/year yielding a total annual net benefit of US$11.3 million.
High efficiency electric motors
These devices are available in Brazil and offer an improved average efficiency of 7 per cent. We assumed that 60 per cent of motors will be replaced by high efficiency models up to the year 2000. This step would save 3.1 TWhe in that year (Geller 1991), saving Brazil US$135 million and avoiding 0.115 MTC emissions.
Variable speed electric motors
Variable speed motors can be run partly loaded without decreasing energy efficiency. More important, in applications such as refrigeration and air circulation, such motors can operate partly loaded with less energy consumption than do fixed frequency motors that provide only on/off cycles. About 30 per cent of the total motor market (measured in kWhe/year) is used to drive variable loads. If half of this load is met by variable speed motors, Brazil would reap a total annual net benefit of US$237 millions and conserve 0.197 MTC.
Electric ovens and boilers
At least 10 per cent of the electricity used in electric ovens and electric boilers can be avoided by recycling the exhaust heat or by installing more efficient equipment (Geller 1991). We expect that the retrofit rate will be low. Nonetheless, one third of the potential saving could be achieved by the year 2000. We estimate that the total annual net savings that year would be US$207 million, thereby conserving 0.152 MTC.
Studies on electricity intensive industries (CNE 1989) have shown that improvements in electrolytic processes can reduce electricity consumption by 7 per cent in metallurgical industries and by 10 per cent in chemical industries. Brazil could avoid generating 2 TWhe by this means in 2000, saving a total annual net benefit of US$21.2 million and reducing carbon emissions by 0.075 MTC.
Final electricity demand summary
Modern technologies can improve end-use electricity efficiency, although their cumulative impact on carbon emissions is small in a country like Brazil where most of the electricity is provided by hydroelectric power plants. We estimate that the potential to conserve carbon in the residential, commercial, service, public illumination and industrial sectors in 2000 is about 1.7 MTC - or only about 3.2 per cent of the total fossil fuel carbon emissions of 1990. This result follows from the predominance of hydroelectricity supply in Brazil. Nonetheless, these technologies should be promoted in any case because most bring net economic benefits. The few which involve net costs today should become cheaper as the technology is accepted more widely and economies of scale are achieved. The only high cost technology where this trend may not hold is solar heating to replace electric showers.
Final fuel demand
The final consumption of fuels is the predominant source of CO2 emissions in Brazil (see Table 10.3). We focus on the transport sector in this section because it is the largest source of fossil CO2 emissions.
Transport is central to the achievement of carbon emission abatement. In Brazil, this sector is responsible for 44 per cent of fossil carbon dioxide emissions. Most (82 per cent) of these emissions come from road transport (SNE 1991b). Transport-related carbon emissions result from three broad factors: the carbon emission coefficient of the fuel used; the efficiency with which different energy forms are used in different modes and markets; and the demand for transport services in different markets and different modes that serve these markets.
In this section, we review the measures which influence these three determining factors in the transport sector. Fuel substitution can modify the first factor, the carbon emission coefficient. Vehicle efficiency affects the second. 'Structural charges' influence both the demand for different modes of transport; and the performance of the vehicles operating within them, that is, the second and the third factors listed above. These first two classes of measures are dominated by energy sector objectives and priorities. The economics of the third, the category of 'structural charges,' is determined by non-energy societal benefits. Consequently, economic analysis of the third factor is more complex than for the first two.
The emergence of CO2 emissions as an issue will have a profound effect on the development of substitutes for petroleum derivatives in the long run. The only large scale commercial substitutes in the world which reduce rather than increase emissions are ethanol from sugarcane (the production of which is concentrated in Brazil) and compressed natural gas.
Brazil's well-known alcohol programme (PROALCOOL) now fuels nearly five million Brazilian cars with pure (hydrated) ethanol. The rest use a gasoline-alcohol mixture. PROALCOOL is now in a difficult situation due to the collapse of oil prices in 1986. Some have suggested that alcohol output should be reduced gradually (World Bank 1990). We have not calculated the societal cost of maintaining current output, nor the cost of increasing the output level. Such an estimate should include the impact on sugar prices of reducing alcohol output (Borrel 1991) and on employment - especially in the Northeast where the sugarcane industry is in crisis. However, the net cost of merely maintaining the output of alcohol should be lower than for expanding it (see Table 10.10).
The 'alternative' SNE scenario projected moderate growth of 2.1 million of tonnes of oil equivalent (MTOE) or 37 per cent by 2000, thereby increasing the market share of ethanol from 17 per cent to 19 per cent of total transport fuel. We adopted this estimate in our own scenario. This substitution would conserve about 1.5 MTC (see Table 10.10), even allowing for fossil fuel inputs for alcohol production. The net cost of this expansion is heavily influenced by assumptions as to gasoline prices. At a gasoline price of US$25 per barrel, we estimated the net cost of expansion to be US$2301TC. Changes in alcohol production technology may also substantially reduce costs by improving the utilization of the residues of sugarcane processing in the cogeneration of electricity. In Brazil, conventional steam turbine technology does not offer much hope of reducing alcohol production costs. But the new gasification/gas turbine (BIG/GT) technology could reduce costs significantly (Ogden et al 1990). Table 10.10 illustrates this possibility. Such major reductions are likely to be commercially proven only by the end of the decade.
Table 10. 10 Transport sector opportunities to limit CO, emissions by 2000, preliminary
|Avoided CO2 (MTC)a||Net cost/TC (US$)a|
|Alcohol: maintain existing output||4.2||not availableb|
|Alcohol: expand output|
|Current technology at $25/b gasoline||1.5e||230c|
|Current technology at $35/b gasoline||1.5e||115c|
|New technology at $25/b gasoline||very small||75d|
|New technology at $35/b gasoline||very small||-35d|
|Natural gas||0.2f||near zero (+,-)|
|Improvement in automobiles||1.9-2.5k||-135i|
|Highway system recovery||2.4g||800g|
|Improved urban transportation||0.4-0.8l||near zero (+,-)|
a Includes rough estimates of fossil fuel inputs for alcohol production (15%
of alcohol output) and of refinery efficiency for gasoline (95%). Elsewhere in
column a 95% refinery efficiency is assumed.
b Estimated to be lower than increasing alcohol output. Baseline scenario assumes maintenance of existing output.
c Assumes a litre of hydrated alcohol is equivalent to 0.7 litres of gasoline (small efficiency credit).
Cost of good exiting distillery is $0.20 per litre. Allows for refinery efficiency of 95% in gasoline production and fossil fuel inputs equivalent to 15% of alcohol production.
d Assumes alcohol production cost of $0.14 litre, only available by end of the decode.
e Based on 'alternative' scenario relative to 1990 (see text).
f Based on 'alternative' relative to 'tendencies' scenario.
g Assumes baseline of 14 MTOE diesel consumption and 3.7 MTOE gasoline in 2000 compatible
with 'tendencies' scenario assuming same proportion of total diesel and gasoline transport use
as today. Assumes an average 15% improvement for all vehicles. While up to 40% improvement
is possible from worst to best conditions, not all of the roods needing improvement, about 50%
of the roads are in the 'worst' category.
h Assumes 25% of market of 21.4 MTOE in 'tendencies' scenario shifted to this engine type.
i Assumes conservatively average vehicle use at 45,000 km per year with 3-year engine lifetime,
engine 25% mare expensive and average efficiency improved by 15%.
j Assumes existing gasoline price ($0.26 per litre). Average cost of measures is $0.17 per litre,
adjusting cast estimates of Ledbetter and Ross (1990), for 12% discount rate.
k Ledbetter and Ross (1990) estimate US average fleet fuel economy could increase 25% by
2000. Assume here that in 'tendencies' baseline average Brazilian fuel economy increases to existing US level (estimated at 9.2 km/litre).
l See text. This is the least defined case. It helps to illustrate the impact of a relatively small effort to improve urban transportation.
Natural gas is promoted as a substitute for diesel, primarily in public transport. The main goal is to reduce atmospheric pollution (NOx, SOx, particulates) in metropolitan areas. Several cities aim to replace all diesel in vehicle fleets by around 2000. The 'alternative' scenario estimates that 0.9 MTOE of natural gas may be used in this fashion which would conserve O2 MTC. Some of this gas may displace gasoline instead of diesel, since this is more lucrative at current relative prices. The cost of the measure (excluding environmental benefits) is near zero. That is, the natural gas option roughly breaks even with diesel at today's prices.
Road transport vehicle efficiency
Significant improvements in vehicular fuel economy are possible. The rate of improvement will be determined primarily by technological innovation in the automotive industry in the industrialized countries and secondarily, by the pace that these changes penetrate the Brazilian market.
Light vehicle efficiency
The current Brazilian automobile averages about 7.5 km/l (gasoline equivalent). The SNE 'tendencies' scenario projected an increase of 30 per cent in light vehicle (basically automobile) fuel consumption to 17 MTOE, incorporating modest improvements in fuel economy. We extrapolated from US data on trends in automobile fuel efficiency (Ledbetter and Ross 1990), and considered only measures that cost less when operated at the consumer retail price of gasoline today. On this basis, we estimated that the average fuel use of the automobile fleet could be decreased by 20-25 per cent relative to the 'tendencies' scenario, saving 3.4-4.3 MTOE of fuel per year.) The associated carbon abatement depends on whether these savings cut gasoline rather than alcohol usage. If we assume that two-thirds of the fuel saving is gasoline, then carbon emissions would fall by 1.9-2.5 MTC. Based on US costs, the average cost of these measures would be negative (see Table 10.10) - even if we ignore the likely benefits of reducing the emissions of other local pollutants. Our projected large fuel saving contrasts strikingly with that of the SNE official 'alternative' scenario, which projected that only 0.4 MTOE could be saved by increasing automotive fuel efficiency.
Heavy vehicle efficiency
We estimated heavy vehicular fuel use from figures available for diesel consumed by Brazilian road vehicles. On this basis, this end use in 1990 was 15.5 MTOE. The SNE 'tendencies' scenario projected that it would increase to 21.5 MTOE. More efficient diesel engines offer substantial fuel savings at a zero or slightly negative net cost (Cummins 1991). We assumed that 15 per cent of the fuel can be saved in this sector. As the initial cost of the engine is about 25 per cent greater than less efficient motors, the net cost would be US$30-401TC conserved. If the more efficient motors achieved a 25 per cent additional market share than in the SNE 'tendencies' scenario, then another 0.9 MTC would be conserved. Other measures can improve heavy vehicle fuel economy, including improved maintenance and motor regulation and more appropriate sizing of vehicles to their tasks.
In this section, we outline structural changes that foster energy efficiency. These are: the re-paving and rehabilitation of the existing highway system; and the reform of urban transport to improve public transport and the overall productivity of urban transport infrastructure. Both approaches entail large public sector investments in which non-energy costs and benefits usually determine policy decisions. Broad rather than narrow economic analysis must be used to estimate the economic feasibility of these changes.
Highway systems recovery
Brazil has an extensive, but badly deteriorated, intercity highway system. Roughly half of its 130,000 km system requires intense reconstruction. The economic cost of this decayed infrastructure is high in terms of trip time and reliability, vehicle maintenance and lifetime, and hazard. The poor system also reduces energy efficiency. For inter-urban trucks and buses the loss may be as high as 40 per cent on poor quality paved roads relative to well-maintained roads (GEIPOT 1989). Assuming an average fuel economy gain of 15 per cent and a baseline interurban vehicle fuel consumption consistent with the 'tendencies' scenario, CO2 emissions in 2000 could be reduced by about 2.4 MTC. The investments required are large and the lifetime of the assets is as short as five years (Lee 1991). The cost (if fully charged to carbon abatement) could exceed US$8001TC conserved (including only fuel savings). Much - perhaps most - of this reconstruction is economically justified without reference to carbon emissions. In that case, the energy savings and carbon abatement can be treated as a by-product obtained at zero marginal cost.
Improved urban transportation
Approximately one third of transport fuel is consumed in the capital cities and larger metropolitan areas. Changes in the structure and operation of the urban transport systems can influence the evolution of fuel demand, though the evaluation of this potential is still in its infancy. Measures are diverse and include land use control, disciplining the automobile's use of road space, coordinating traffic flow and strengthening collective transport (Poole et al 992).
The city of Curitiba has already addressed the problem of urban transport in a comprehensive manner. The city has coordinated urban land use, roadspace and collective transport policy for more than fifteen years, and has improved bus systems and traffic controls. Fuel consumption per car in Curitiba is about 30 per cent less than the average for other cities of its size in Brazil (Lerner 1989). This differential may indicate the potential impact of such measures if adopted widely in Brazil.
An aggressive programme to address urban transport imperatives might result in CO2 emissions savings of 5-10 per cent relative to the SNE 'tendencies' scenario for 2000. This potential could be realized in spite of the inertia of a decentralized system involving thousands, even millions of actors. If lower cost solutions are emphasized in the next decade, then this energy and carbon savings could be achieved at zero net cost. Such measures would include creating and integrating public transport systems, and controls on traffic flow and parking.
By far the largest source of anthropogenic carbon dioxide emission in Brazil is deforestation, principally in Amazonia. The carbon stock of the seasonal and rainforest vegetation in Amazonia is estimated to range from 140 to 200 TC/ha; that of pasture is 10 TC/ha; and of cropland, 5 TC/ha. The forest carbon stock may be adjusted as new information becomes available on subsurface biomass of the vegetation. Changing land use also reduces soil carbon content. In pasture soil, for example, the carbon content may be about 10 per cent of the approximately 100 TC/hectare of forest soil, or about 90 TC/ha. less than in forests. (Houghton et al 1991).
Thus, assuming a deforestation rate in Amazonia of 1.8 million hectares per year, gross CO2 emissions would be 250-360 MTC (though not all appears immediately in the atmosphere). To this figure should be added emissions from deforestation in other regions of Brazil. Unfortunately, we have no estimates for this source. Although substantially smaller, these are not insignificant.
Biological processes also continually accumulate carbon from the atmosphere, as is the case with regrowth of natural vegetation on deforested areas, abandoned land, or forest plantations. The rate of natural regrowth can vary by a factor of twenty in humid tropical areas depending on the local land-use situation (Nepstad et al 1990). The scale of this countervailing sequestration is poorly understood.
Despite these uncertainties, it is clear that Brazil's annual emissions from deforestation (250-360 MTC) dwarf those of fossil fuel use (60 MTC) as well as from biomass use for energy (about 11 MTC). This fact is consistent with the observation that fuelwood use is not a major factor in overall deforestation, though it may be significant in some regions (for example, charcoal from cerrado and mangroves). The primary direct causes of deforestation are clearing for pasture and cropland, with logging often opening up the occupation process.
Focusing on Amazonia, any substantial decrease in the rate of deforestation is likely to be associated with decreased economic growth. Macroeconomic modelling suggests that for every 1 per cent reduction in deforestation regional GDP would have to fall by roughly 1.7 per cent (Reds 1991). While pessimistic, the model suggests a first approximation of the cost of CO2 abatement by halting deforestation, roughly US$4/TC according to the model's author. This low cost (equivalent to a tax of $0.50/barrel of oil) is probably an upper limit, since the model assumes historical relationships. A strategy to change these relationships should be both cheaper and allow a less drastic trade-off between economic growth and deforestation. Such a strategy must go beyond police enforcement or reducing/eliminating legal and financial incentives to deforestation, though these are important (for example, Binswanger 1991). New or modified economic activities must be developed or strengthened both in forested and deforested areas (Sawyer 1990) based on land-use zoning. Settlement and economic activity, for example, should be stabilized, consolidated, and (in many areas) intensified in the largely deforested areas along the frontier and the 'pre-frontier'. While complex, restructuring Amazonia's economy is likely to be a large, 'no regress' source of CO2 abatement.
The relationship between land-use trends and energy policy has been little explored in Brazil. The most important such interaction is fuelwood for industry and charcoal. This nexus is the most important direct energy-related source of deforestation. A key issue is whether a decisive move to put these uses on a sustainable basis is justified or whether they should be phased out.
Another important land-use issue in relation to energy arises from hydroelectricity development in Amazonia. The relative priority, rate of development, and ultimate potential may all be influenced by a strategy to minimize deforestation. The infrastructure and migrations occasioned by hydro are the key concern. Some projects may provoke deforestation. Others help to decrease it as, for example, on the Tocantins river (Moreira et al 1990). This indirect effect on carbon emissions is likely to be larger than differences in direct electricity CO2 emissions resulting from alternative scenarios of hydropower/thermal generations (as discussed above).
Two subsidiary issues also connect land-use and energy policy issues. The unavailability of electrical power to isolated communities (most especially in Amazonia) constrains economic development. Poverty, in turn, fosters more carbon-emitting and intensive resource exploitation (Poole et al 1990). Relatedly, fuels such as diesel sold for use in Amazonia are subsidized (Reds 1991). The common denominator of CO2 emissions reinforces the need to consider energy, land use and regional development together.
The analysis of CO2 evolution and abatement measures is still incipient in Brazil both in regard to energy and to land-use change. The creation of credible, systematic and internationally comparable 'abasement cost' curves for Brazil is still not possible. As a consequence there is as yet little basis on which to agree on specific CO2 limitation targets. The problem is exacerbated by the wide range of uncertainty surrounding the prospects for economic growth. The work briefly described here is part of an effort to better understand the potential and economics of CO; abatement. In Figure 10.1, we summarize the tentative results of this study for the costs and abatement potential of measures to reduce fossil energy emissions. We identified about 13.7 MTC of technologically feasible and economically justified carbon abatement, relative to the SNE 'tendencies' scenario in the year 2000. This reduction potential amounts to about 16 per cent of the SNE reference projection for carbon emissions in that year. We did not allow for the more pervasive energy and carbon reducing effects of technological innovation in all end-using sectors; nor did we include the impact of shifts in the sectoral composition of the economy on our estimate of energy and carbon conservation. Moreover, as our bottom-up calculation is extended to other sectors, we expect to increase substantially the size of the carbon reduction potential above 13.7 MTC.
Figure 10.1 Cumulative annual carbon emissions avoided by 2000 for technology improvements in Brazil
Although the quantitative analysis is incomplete and preliminary, it permits some observations which are relevant for policy. It appears that substantial savings in CO2 emissions can be achieved at 'negative cost' or very low cost (say, less than US$10/TC), both in energy and land-use change. These savings should be substantially cheaper than those of many measures being considered by the industrialized countries. However, the fact that these are 'no regress' sayings does not mean that they are easy to achieve. This fact is relevant for a possible policy of international resource transfers, which should be very attentive to 'no regrets' opportunities in developing countries.
In Brazil's energy sector, the major 'no regress' sayings involve increasing energy efficiency in all consuming sectors. As explained earlier, conditions and policies favouring greater energy efficiency are also likely to favour, and are associated with, higher medium term economic growth (low inflation, correct price signals, investment in modernizing processes and products, competition, etc). Energy efficiency itself should directly contribute to improved overall productivity and thus to economic growth.
With regard to land-use change, we argue that the trade-off between Amazonia's economic growth and reducing deforestation is not so acute as some suggest. However, a strategy is needed urgently to change the economic dynamics of Amazonia's frontier regions. The case of Amazonia also highlights the need for cost-benefit analysis to explicitly consider who pays the costs and who receives the benefits of policy changes aimed at conserving carbon.
The medium term potential for relatively low cost CO2 abatement is probably much larger for land-use change than in the energy sector. This judgement does not mean that energy efficiency should be ignored, however. Many measures to reduce energy-related emissions are as cost effective, or even more so, than those to reduce emissions from land-use change. Moreover, energy is fast becoming relatively more important. Finally, as we have observed, energy policy can influence future land-use, especially in the Amazon region.
Using the technologies discussed in this chapter, the total amount of carbon abatement is approximately 13.5 million TC. In Figure 10.1, we plotted a total amount of 9.3 million TC, since we excluded the 4.2 million TC due to the use of alcohol at the existing output (as shown in Table 10.10) because we did not know the net cost associated with this technology. The opportunities identified in this chapter are not the complete picture. Other technologies exist, and with their full inclusion on the demand side of the Brazilian energy matrix, probably more CO2 abatement than is forecasted in the official alternative scenario can be achieved.
Ayres, R U (1991) 'No regrets options for greenhouse gas abatement,' in A Comprehensive Approach to Climate Change, ed. T Hanisch, Center for International Climate and Energy Research, Oslo
Binswanger, H (1991) 'Brazilian Policies that Encourage Deforestation in the Amazon', World Development, 19(7)
Borrel, P (1991) 'How a Change in Brazil Sugar Policies Would Affect the World Sugar Market,' PRE, Working Paper, World Bank, Washington DC
CEMIG (1989) Companhia Energca de Minas Gerais, 'Estudo de Economia de Energia', Belo Horizonte-MG, Brazil
CESP (1990) Companhia Energca de SPaulo, Personal communication, SPaulo, Brazil
CNE (1989) ComissNacional de Energia, 'Estudo Sobre Energo Intensivos, Assessoria Tecnica', April, 1989, Brasilia, Brazil
Cummins Newsletter (1991) published by Cummins Engine Co. Inc., Columbus, USA
ELETROBRAS (1990) Programa Decenal de Gera 1991/2000, Sistemas Interligados SE/CO/S, N/NE. Planning of Grupo Coordenador do Planejamento dos Sistemas Eletricos, Rio de janeiro, Brazil
GEIPOT (1989) Empresa Brasileira de Planejamento de Transportes. Manual para Culo de Custos Operacionais de Velos Rodovios, Ministo dos Transportes, Brasilia, Brazil
Geller, H (1991) Efficient Electricity Uses: a Development Strategy for Brazil, American Council for an Energy Efficient Economy, Washington, DC, USA
Houghton, R A, Skole, D L and Letkowitz, D S (1991) 'Changes in the landscape of Latin America between 1850 and 1985: II Net release of CO2 to the atmosphere,' Forest Ecology and Management, 38, pp 173-199
Jannuzzi, G M, Gadgil, A, Geller, H. Sastry, A (1991) 'Energy efficient lighting in Brazil and India', Proceedings of the 1st European Conference on Energy-Efficient Lighting, Stockholm, 28-30 May
Jannuzzi, G M and Schipper, L (1991) 'The structure of electricity demand in the Brazilian household sector', Energy Policy, 19(11)
Latone, COF. Nobre, E C and Burgoa, J A ( 1990) 'Diagnco do Potencial de Conserva de Energia na Industria' Proceeding of the 5th Congresso Brasileiro de Energia 'desafio dos anos 90', November 5-9,1990. Rio de Janeiro, Brazil
Ledbetter, M, and Ross, M (1990) 'Supply Curves of Conserved Energy for Automobiles'. Lawrence Berkeley Laboratory, Berkeley, USA
Lee, S H (1991) Planning Manager, Secretaria de Transportes e Obras, Santa Catarina, Brazil, Oral Communication
Lerner, J (1989) 'Curitiba Mass Transit Systems', Workshop on New Energy Technologies, Transportation and Development, International Institute for Energy Conservation, Washington DC, September
Moreira, J R and Poole, A D (June 1991) 'Energia e Melo Ambiente', Secretaria de Ciencia e Tecnologia, Federal Government, Brasilia, Brazil
Moreira, J R and Poole, A D (1992) 'Hydroelectricity and its Constraints' draft for publication, ed. Thomas Johansson, 'Fuels and Electricity from Renewable Sources of Energy', Island Press.
Moreira, l R. Poole, A D and Serrasqueiro, M A (1990) 'Alternatives Energcas da Amazonia'. Part IV of Projeto/Anse da Implanta de Grandes Projetos Energcos - O Caso do Setor Elico do Brazil. Institute of Electrotechnology and Energy. University of Sao Paulo
Nepstad, D, Uhl, C and Serro, E A (1990) 'Surmounting Barriers to Forest Regeneration in Abandoned Highly Degraded Pastures, A Case Study from Paragominas, Para, Brazil', in Alternatives to Deforestation, ed. Anthony Anderson, Columbia University Press, New York
Ogden, J M, Williams, R H and Fulmer, M E (1990) 'Cogeneration Applications of Biomass Gasifier/Gas Turbined Technologies in the Cane Sugar and Alcohol Industries', Center for Energy and Environmental Studies, Princeton University, Princeton
Poole, A D, Ortega, O and Moreira, J R (1990) 'Energia pare o Desenvolvimento da Amaznia ', Instituto de Eletrotica e Energia for UNDP/SUDAM, Macrocenos Amazonia 2010, Bel Brazil
Poole, A D, Pacheco, R S and Melo, M A C (1992) Collective Transport Policy in Developing Countries: A Case Study of Brazil, International Development Research Institute, Ottawa
Reis, E (1991) 'A Amazonia e o efeito Estufa, Perspectivas da Economia Brasileira 1992, IPEA - Instituto de Pesquisas e Economia Aplicada, Brasilia, Brazil
Sawyer. D (1990) 'The Future of Deforestation in Amazonia: A Socioeconomic and
Political Analysis', in Alternatives to Deforestation, ed. Anderson, Columbia University Press, New York
SNE (1991a) 'Reexame da Matriz Energetica Nacional', Ministerio da Infraestrutura, Federal Government, Brasilia, Brazil
SNE (1991b) 'Balanco Energetico Nacional', Secretaria Nacional de Energia, Ministerio da Infraestrutura, Federal Government, Brasilia, Brazil
USDOE (1990) The Potential of Renewable Energy, Interlaboratory White Paper, Office of Policy, Planning and Analysis, Washington DC
World Bank (1990) Brazil Energy Strategies and Issues Study Pricing and Investment Policy, Washington DC