Burundi - Issues and options in the energy sector

RepoK No. 377&BU 3778 Burundu: ossues and Optorn Dn the Energy Sector June 1982 Report of the joint UNDPA/%rki Bank Energy Sector Assessment Program This document has a restricted distribution. Its contents may not be disclosed without authorization from the Government, the UNDP or the World Bank.  CURRENCY CONVERSION US$1.00 = 90.00 BuF (Burundi Franc) ENERGY CONVERSION FACTORS TOE = ton of oil equivalent = 10,500,000 kcal (kilocalories) = 3940 kWh (kilowatt-hours) electricity and hydro = about 9.0 m3 of fuelwood - about 1.5 tons of charcoal = about 3.5 tons of peat at 35% moisture content - 0.92 tons = 1240 liters gasoline - 0.95 tons = 1210 liters kerosene - 0.99 tons = 1150 liters gas and fuel oil This report is based on the findings of an energy assessment mission comprising Mr. D. Hughart (Energy Economist), Ms. C. Tobias (Consultant), and Mr. C. H. A. Killoran (Senior Power Engineer), who visited Burundi in April and May, 1981. Ms. S. Baile (Researcher) assisted in the report s preparation and secretarial assistance was provided by Ms. Linne Toehl.  FOR OFFICIAL USE ONLY Report No. 3778-BU BURUNDI ISSUES AND OPTIONS IN THE ENERGY SECTOR June 1982 This is one of a series of reports of the Joint UNDP/World Bank Energy Sector Assessment Program. Finance for the work has been provided, in part, by the UNDP Energy Account, and the work has been carried out by the World Bank. This report has a restricted distribution. Its contents may not be disclosed without authori- zation from the Government, the UNDP or the World Bank.   - i - TABLE OF CONTENTS CHAPTERS Page No. I SUMMARY AND RECOMMENDATIONS 1 Energy Problems ... .........1.......... 1 Energy in the Economy ....................... 1 Woodfuels .................................. 6 Peat ... 7 Petroleum Products. 7 Electricity. 8 Other Energy Sources. 8 II WOODFUELS 10 Supply and Demand .......................... 10 Institutions ........................... 11 Forestry Extension . 12 Current Forestry Projects .......................... 12 Fuelwood Conservation ....................... 13 III PEAT 16 Akanyaru River Basin .16 Highland Peat Bogs .17 Nyamuswaga Bog .18 Potential Production .19 Institution .21 IV PETROLEUM 23 Demand and Supply .23 Petroleum Stockpiles .25 Contingency Planning .29 Potential Alternatives to Petroleum Product Imports ... 31 V ELECTRICITY 32 Present Generation and Transmission Facilities .33 Projects Under Construction .34 Future Demand and Supply .35 Rusizi II .37 Rwegura .37 Other Projects .38 Tariffs and Finances .38  - iii - CHAPTERS Page No VI RENEWABLE ENERGY 39 Wind and Solar ....................... 39 Producer Gas .40 ANNEXES 1 Rural Cookstove Project ............................ 42 2 Petroleum Stockpiles .45 3 Extension of the Mombasa-Nairobi Pipeline .48 4 Petroleum Exploration .50 5 Electricity .53 MAP Electric Power System (Regional Map) .59  I. SUMMARY AND RECOMMENDATIONS Energy Problems 1.01 Burundi confronts a range of problems in the energy sector. The first and most important is the increasing scarcity of the fuelwood and other traditional fuels on which almost all of Burundi's population depends for cooking and other basic energy needs. The extraordinarily high cost of oil imports, now estimated at $100 per barrel, is another, as is the low reliability of supply by any means other than air freight. Together with its neighbors Zaire and Rwanda, Burundi faces the problem of assuring reliable supplies of electric power over the medium and long term, when the capacity of plants that now exist or are under construction in the inter-connected system serving the three countries is fully exploited. Energy in the Economy 1.02 Burundi is a small, landlocked country of 27,820 km2 in central Africa. Ninety-five percent of the population of about 4.2 million lives on isolated farms in rural areas, engaged primarily in subsistence agriculture and small-scale coffee cultivation. Burundi has traditionally had adequate rainfall for crop raising, but with increases in population density have come reductions in fallow periods and an increasing use of crop residues as fuel rather than mulch, which have combined to decrease the soil quality of agricultural lands. Population density overall is about 150 inhabitants per km2, but reaches over 350 even in some primarily agricultural areas. About 140,000 people (70% of the total urban population) live in Bujumbura, the capital city, located on Lake Tanganyika. 1.03 Coffee exports comprise 90% of total merchandise exports and Burundi's external current account balance is therefore heavily influenced by fluctuations in coffee prices and volume. In 1976 and 1977, it registered a surplus, but between 1978 and 1980 the deficit has increased each year due in part to falling coffee prices but also to increasing imports. In 1980, the external current account deficit was equivalent to about 10% of GDP, and it is expected to increase during the next few years. Petroleum product imports have accounted for about 16% of total merchandise imports (c.i.f. basis) in recent years, which is about three times the 1975 proportion but still much below their share in many other developing countries.  - 2. - 1.04 Nearly all of the energy consumed in Burundi is in the form of woodfuels or agricultural residues. Most of these fuels are gathered rather than sold, and although consumption data are not yet available, annual consumption per capita is at least 0.5 m3 fuel wood equivalent, and traditional fuel consumption accounts for at least 90% and perhaps over 95% of the total energy consumed. 1/ 1.05 Table 1.1 shows the estimated commercial energy balances for the years 1973-1980. The table shows that 75% of Burundi's commercial energy needs are met by petroleum products, while electricity accounts for about 23% and peat about 2%. The overall growth in apparent commercial energy consumption during the 1973-1980 period was about 9.0% p.a. The growth rate appears to have accelerated from 4.8% p.a. in 1973-77 to 14.9% p.a. in 1977-80, but part of this increase is probably accounted for by petroleum inventory changes. The acceleration in electricity consumption (from 5.6% p.a. to 10.6% p.a. in the same two periods) was also pronounced, however. Economic growth as measured by GDP slowed in the meanwhile from about 4.7% p.a. in 1973-77 to 4.2% p.a. in 1977-80. 1.06 Reliable data on energy consumption and output by economic sector are not available and therefore neither an analysis of the very high 1977-80 growth rates nor of whether they are likely to continue can be made. 1.07 Commercial energy consumption per capita and per unit GNP is lower in Burundi than in almost any other country, and, as Table 1.2 shows, is only about one-third the average for other contries similar in size and income level. This is owing partly to the relatively high cost of fuel imports imposed by the country's geographic position and partly to the low energy intensity of its productive sectors. 1.08 As illustrated in Figure 1.1, there are very strong associations between specific sources of energy and specific categories of energy demand in Burundi's economy, and weak links among the sources or among the users. Most of the population uses firewood, charcoal and agricultural residues for cooking and other household energy needs and, conversely, the bulk of traditional fuels consumption is accounted for by rural households. Oil imports are similarly associated with transpor- tation and construction, while electricity is obtained virtually entirely from imports and used predominately in urban areas for lighting and operating electrical machinery. While a limited amount of interfuel substitution is possible, energy issues in Burundi must really be treated in parallel sub-sectors rather than as an integrated whole. The 1/ A conservative consumption estimate of .5 m3/person/year gives an annual consumption of about 525,000 TOE for those fuels, or 92% of total energy consumption in 1980. If the amount consumed were 1 m3/person/year, the total would be 1,050,000 TOE or 96% of total energy consumption in 1980.  Table 1.1 BURUNDI - COMMERCIAL ENERGY BALANCE, 1973-80 (tons oil equivalent) 1973 1974 1975 1976 1977 1978 1979 1980 PRIMARY PRODUCTION Hydro - - - - - - - 1080 Peat - - - - 10 510 680 1020 Total 10 510 680 2100 IMPORTS Gasoline 9800 13500 10100 12700 12890 15270 13700 18050 Kerosene 3200 2500 1700 1700 1040 1310 1070 910 Diesel & Fuel Oil 6500 9900 7000 9100 9210 12290 11750 16300 (Subtotal) (19500) (25900) (18800) (23500) (23140) (28870) (26520) (35260) Electricity 5990 6170 6330 6720 7540 8630 9100 10200 f 25490 32070 25130 30220 30680 37500 35620 45460 TRANSFORMATION Petroleum Products -370 -370 -340 -340 -340 -400 -400 -470 Electricity 260 260 240 240 240 280 280 330 FINAL CONSUMPTION Petroleum Products 19130 25530 18460 23160 22800 28470 26120 34790 Electricity 6250 6430 6570 6960 7780 8910 9380 11610 Peat - - - - 10 510 680 1020 Total 25380 31960 25030 29120 30590 37890 36180 47420 Sources: Ministry of Commerce and Industry. Notes: Ton of oil equivalent taken as 10.5 Banque de la Republique du Burundi million kcal. Bank, IMF, and USAID estimates. Electricity converted at 2666 kcal/kWh. Thermal generation assumed to require 3800 kcal/kWh. Stock changes included in consumption. Traditional fuels excluded since only very rough estimates based on presumed annual population could be provided.  Table 1.2 (BURUNDI) INTERNATIONAL COMPARISONS OF COMMERCIAL ENERGY CONSUMPTION, 1978 Population GNP/Capita Energy Consumption (millions) ($) (min, kg, o.e.) (kg, o.e./cap.) kg, o.e.I$GNP (% Petroleum) BURUNDI 3.9 160 38 10 0.06 (77) Chad 4.3 150 64 15 0.10 (100) Upper Volta 5.6 160 106 19 0.12 (100) Malawi 5.7 180 239 42 0.23 (49) Rwanda 4.5 190 57 13 0.07 (73) Benin 3.3 230 137 42 0.19 (88) 1 Sierra Leone 3.3 230 259 78 0.35 (100) I (unweighted avg.) (4.5) (190) (144) (34) (18) (85) Note: Burundi's GNP per capita in 1980 is estimated at $200. Comparators are countries whose populations and per capita GNPs are between 67% and 150% of Burundi's. kg, o.e. = kg oil equivalent = 10,500 kcal. Source: Table 1.1, World Bank Atlas and Data Files.  - 5 - Figure 1.1: Energy Flows in Burundi (not to scale) Sources Uses Agricultural Residues Agricultural Rrlhueod andl R ural hou sehoId s Forestry (cooking, heating Resource Base Forests, Savannah Fuelwood Peat _ = e3 Tea factories, missions, etc. Domestic Hydro Urban households Domestic Hydra B-/ Urban industrial/commercial Electricity Imports (lighting, electrical appliances, etc.) Urban industrial/commercial, (process heat, engines, etc.) Oil Imports Transportation and construction World Bank -23571  -6- structure of this report follows this approach, with separate chapters covering the sources and uses of firewood and peat, oil products and electricity. Moreover, efforts to strengthen Burundi's energy planning capability must also be concentrated at the sub-sector level rather than by establishing an overall energy planning unit. 1.09 It is worth emphasizing that while it is more convenient to label the energy sub-sectors according to form of energy involved or the source of supply than by the set of energy demands with which they are associated, this does not imply that energy supply policy issues are more important than those related to energy demand. In the fuelwood and peat (or "rural household cooking, heating, and lighting") sub-sector in particular, efforts to improve the efficiency with which energy is used may prove to have a higher and faster return than efforts to increase supply. Woodfuels 1.10 The principal issue in the fuelwood sector is how to deal with the rapidly dwindling supply. Current or planned reforestation projects, overseen by the Director of Waters and Forests in the Ministry of Agriculture, appear to be doing as much tree planting as is possible with the available financial and manpower resources. Therefore, the mission recommends that attention be given to other issues, namely the shortage of trained Burundi foresters, the weaknesses of the forestry extension program, the lack of data on rural household cooking fuel consumption and the need to introduce conservation methods. Since Burundi currently has only one trained local forester, several students should be sent abroad for university level forestry training. A current project already provides for the training of technicians to assist the foresters. A countrywide survey of the household use of firewood and agricultural wastes is in progress. It should provide important data to allow the identification of those areas of the country where forest resources are in shortest supply, to permit the best allocation of the sub-sector's resources. Current forestry extension programs have suffered from a lack of funds which has led to poorly trained and poorly supervised extension workers. Efforts have been made to develop an audio-visual program on tree care for rural farmers which is a good beginning. Well-trained extension workers are needed to teach not only tree care and woodlot management, but also fuel conservation methods. Relatively little has been done thus far to develop and introduce better charcoal production methods or fuel conserving cookstoves. Charcoal producer licensing fee arrangements should be changed to encourage wood conservation, and improved kiln designs should be introduced. Improved cookstoves for firewood and charcoal should be developed and extension workers should assist the local population with their construction and use. Peat 1.11 Burundi's peat resources are a potential substitute for fuelwood, for small industries and for rural institutions, and perhaps eventually for urban households. ONATOUR, a parastatal under the  - 7 - Ministry of Mines and Energy, oversees peat development. Over the past two years the assessment and development of Burundi's peat reserves has become much more systematized. A current USAID project is developing production and markets for highland bog peat. The IDA-financed Musongati nickel study will test various techniques for the extraction and processing of peat from the flooded Akanyaru Basin. The principal issue in the peat sector is the uncertainty about the extent of exploitable peat reserves. If peat from the Basin can be economically extracted without causing serious environmental problems, Burundi's exploitable peat reserves will increase from hundreds of thousands to millions of dry tons. A recently discovered bog, Nyamuswaga, might contain up to four million dry tons of peat which could possibly be extracted more easily than that in the Akanyaru bogs. This amount is four times the total content of all the highland bogs combined. Pilot production is to begin in 1982. Additional studies are needed of the potential industrial market and of means of reducing transportation costs. Petroleum Products 1.12 Burundi has no domestic oil production or refining activities. Geological assessments offer only limited encouragement to further exploration for oil, and limited consumption levels make a refinery inadvisable. The Ministry of Commerce and Transportation supervises petroleum imports. Almost all petroleum products are currently imported into Burundi by truck from Nairobi. A supply interruption occurred during the Uganda war, and a principal issue in this sector is the availability and appropriate scale of strategic petroleum reserves. Another issue is how to deal with the very high cost of imported oil, now about $100 per barrel, including freight but exclusive of taxes, twice what most other countries must pay. Over a third of the cost of Burundi's oil imports is attributable to the differences between (a) the cost of the pipeline-truck transportation route via Nairobi that is now used for most imports and rail-lake transport from Dar es Salaam and (b) Kenya's export price and prices in the open market. 1.13 SEP (Societe' d' Entreposage de Petrole), which is owned jointly by five foreign oil companies, operates petroleum storage facilities near the port at Bujumbura. The Government is constructing an additional storage capacity near Gitega. These tanks will bring Burundi's storage capacity to the equivalent of over eight months of consumption, without counting stocks held by retailers and consumers. Filling this capacity may not be economically justified at this time unless oil is obtainable on concessionary terms or the likelihood of a cutoff in supplies is thought to be high. Moreover, little information is available on the needs of priority sectors, and an allocation system for occasions when there is a supply cutoff has yet to be designed. Priority should be given to developing a suitable plan for stockpiling of products. The costs of building and maintaining reserves should be passed on to the consumers who will benefit from them. 1.14 Burundi is at a disadvantage in bargaining over petroleum prices given the limited number of suppliers and the high cost of transport.  Kenya charges a preinium price for exported oil products. However, gasoline and diesel fuel purchased at spot market prices in the Persian/Arab Gulf and shipped to Mombasa would cost about $14 per barrel less than the Kenyan export price. Another alternative might be for Burundi to purchase oil in international markets, especially in the Middle East, and bring it across Tanzania. -E~Thct-ticity 1.15 Currently, 95% of the electrical power consumed in Burundi is imported from Rusizi I in Zaire to Bujumabura. Burundi has substantial hydropower potential which it would like to develop to decrease its dependence on foreign sources. Two 2 MW units of an 8 MW Chinese- financed hydroelectric plant were comnpleted in January, 1982. Several smaller plants are also under construction, and the Government has arranged funding for an 18 MW plant at Rwegura. In addition, the Governmenit has decided to participate in the Rusizi II project jointly with Rwanda and Zaire. 1.16 The principal issue in this sector is the appropriate development in indigenous power resources in view of possible investments elsewhere in the region. Electricity produced at Rwegura would cost about BuF 17 per kWh (in 1982 prices) which is on the high side. REGIDESO, under the Ministry of Mines and Energy, has primaary responsibility for the power sector, and the mission recommends that a strong planning unit be established within REGIDVSO to draw up and implement an overall power sector development plan, to set priorities and to determine when and how the available sites should be developed. In addition, the regional organization, EGL, needs assistance in formulating regional strategies and ensuring that member countries are not put in vulnerable positions. Other Energy Sources 1.17 With the possibly significant exceptions of solar water heating for the brewery and textile mill) other energy sources such as solar, biogas and wind seem to offer no substantial, practical substitute for current energy sources at the present time. The muission therefore considers that major investment in alternative energy resources, primarily solar, biogas and wind, should not be considered until there is an analysis of these industrial applications and other potential unarkets. Lionited experimentation with solar and biogas should, however, continue and Lmay provide energy for some rural institutions. Producer gas maight be able to substitute for some liquid fuels if a suitable raw rnaterial other than wood can be found. If the Nyamuswaga bog can be economically exploited, that peat could perhaps be used in gasifiers. 1.18 CRUEA (Centre de Recherche des Utilisation des Energies Alternatives) was organized at the Science Faculty at the University of Burundi to develop alternative energy resources. It has only a small regular budget supplemnented by donations fro-.n outside. So far its work has concentrated on hardware developmaent with inadequate attention given to the socio-economnic characteristics of the potential market. The  - 9 - mission recommends that CRUEA's work focus on areas which are likely to provide practical solutions to the country's energy problems.  - 10 - II. WOODFUELS Supply and Demand 2.01 Burundi's already scarce forest resources are dwindling rapidly. One estimate 1/ is that even with all of the current reforestation work, tree cover could be reduced to 40% of present levels by 1990 and wood production could decline to only 124,000 m3/year, which would be far too low to meet the expected demand. 2.02 Natural forests comprise the remnants of dense montane tropical forest along the Zaire/Nile crestline in the northwest (about 40,000 ha), small areas of savannah woodland (15,000-20,000 ha) in the south, and forest galleries along water courses and in steep gorges. These forests have not been managed for maximum yields, due in part to the absence of trained local foresters, and their increment is low, about 1 M3/ha/year. In addition, it is estimated that there are 25,000 ha of forest plantations, mainly Eucalyptus, as well as roadside trees and small woodlots. The average annual increment of plantations is about 10 m3/ha. The total annual forest increment is about 310,000 m3, including both natural and planted forests. Wood supply from roadside trees and small woodlots is not known precisely. Total increment appears to be far below the conservative consumption estimate of 1 million m3/year. Forest plantations would have to cover at least 100,000 ha if they were to meet even the most conservatively estimated demand. 2.03 Accurate data on firewood use in rural areas, where 95 percent of the population resides, are unavailable 2/. Estimates of average annual wood consumption per capita range from .25 m3/year to 1 m3/year. Given Burundi's population of about 4.2 million, this would place total wood consumption at between 1.0 million m3/year and 4.5 million m3/year. Actual consumption is likely to be closer to the low end of the range, given what is known about supply. Total marketed firewood has been estimated at 156,000 m3/year 3/, but most firewood is gathered, not marketed. Marketed firewood supplies the principal towns and rural institutions, such as missions and hospitals, and rural industries, such as tea factories. 2.04 In urban areas, per capita consumption (accounting for charcoal in terms of the wood used to produce it) is estimated to be much higher, 1/ USAID Bururi Forest Paper Project, 1981. 2/ The current IDA forestry project includes a sample survey in progress which will provide data on use of wood and agricultural wastes in rural households. 3/ "Burundi Rapport Sur le Sous-Secteur Forestier" by M. J. G. Devitt, 1977.  - 11 - due to the unavailability of alternative fuels like agricultural wastes, the inefficiency of charcoal production, and the higher cash incomes. Charcoal is the principal cooking fuel of the low and middle income groups in Bujumbura. A recent survey 1/ showed that 87% of the households used charcoal as the principal cooking fuel. Most of the rest cook with firewood. Most of those surveyed in Bujumbura purchased the fuel they used for cooking, whether they were charcoal users or firewood users. The amount spent per month per person on charcoal for cooking was about 240 BuF. This is equivalent to about 10 to 15% of the amount spent on food. Total annual charcoal consumption for Bujumbura can be estimated at roughly 20,000 metric tons/year or the amount produced from about 200-300,000 m3/year 2/ of wood. Bujumbura accounts for about 3.5% of the total population, but from 6 to 25% of total consumption of fuelwood depending on which average yearly consumption estimates is used -- .25 m3 or 1 m3- 2.05 The fuelwood scarcity has direct impact on food crop production, since the rural population is increasingly using agricultural residues to meet their fuel needs. In some cases, agricultural wastes are the primary fuel. This reduces soil fertility and crop yields which in turn increases pressure on land, and leads to shortened fallow periods, creating a cycle of soil degradation and falling production. The wood shortage has also affected the urban population, as charcoal prices increased from around 150 BuF/sack in 1970 to about 650 in 1980. This 333% increase compares with a less than 200% increase in the consumer price index for that period. Institutions 2.06 The Forestry Department is under the Ministry of Agriculture. The Director of Forestry also deals with fisheries and wildlife. Forestry development has had low priority in budget allocation planning; protection programs have until recently been small and ineffective and Burundi has only one trained forester. The parastatal National Wood Bureau operates a sawmill and imports building wood. 2.07 A survey of wood use in Burundi is currently underway. It will provide essential data for planning the development of the forestry and fuelwood sectors. The survey will measure both urban and rural wood consumption for all purposes. A rural household sample survey will collect information on consumption of both firewood and agricultural 1/ C. Tobias, "Household Cooking Practices and Fuel Use in Bujumbura", 1980. The survey included a random sample of 121 households at three income levels in Buyenzi, Ngagara, Nyakabiga and Zinama zones. 2/ Based on average monthly consumption of 12 kg per capita. This is a very rough estimate and could be too low since lower income households were disproportionately represented in the sample survey.  - 12 - wastes. Once consumption patterns are known, it should be possible to set priorities for forestry and conservation projects on a regional basis. 2.08 The extension program in the forestry sector has been very limited. Although efforts are being made to strengthen that program, additional financing will be required to make a significant impact. Training and supervision of extension workers is generally inadequate. Little effort has been made thus far to teach the rural population to plant and manage woodlots, or how to conserve fuel when cooking. An audio-visual program consisting of a series of drawings accompanied by recorded training cassettes has been developed as a teaching aid on tree care, but only one copy has been made and it has not been used very extensively. Forestry Extension 2.09 The seriousness of the fuelwood crisis in Burundi will of necessity make itself increasingly evident to the population over the coming years as more and more time must be devoted to gathering firewood and purchased fuel becomes more and more expensive. However, a well- designed extension program to teach people how to plant and care for woodlots, and to use more fuel-efficient cookstoves, could help diminish the effects of the crisis. There is a serious need for well-trained, well-supervised extension workers in Burundi. It is not essential that they be university-trained; in fact, such training may be counterproductive. Rather they must have a good basic understanding of the problems and potential solutions which are appropriate for local conditions. Students should be selected from the collines, and training should take place as near to their work sites as possible, with the understanding that their work is in the field and not in the cities. Current "moniteurs" (low level forestry employees) will need retraining. An increase in the budget of the Forestry Department will be required to allow adequate training and supervision. Well-trained and supervised extension workers could assist the rural population with energy efficient cookstove construction, woodlot preparation and care, and could help create an awareness both of the seriousness of the problem and of ways to decrease its impact on them. Current Forestry Projects 2.10 A number of forestry projects are currently in progress in Burundi. Belgium and Saudi Arabia are co-financing a 12,000 ha reforestation project in Mugamba-Bututsi, in south-central Burundi. The European Development Fund will finance the reforestation of 3200 ha at Mumirwa and 6000 ha near Gihofi. The French Fonds d'Aide et de Cooperation is financing the planting of a 500-ha protection belt around the Zaire-Nile crestline in northern Burundi. An IDA project is financing a 2000-ha plantation near Bujumbura and 5000 ha near Vyanda, as well as 30 nurseries to produce seedlings for sale to communes and individuals and studies and technical assistance to strengthen forestry planning and the Forestry Department. Three IDA Rural Development  - 13 - projects also have forestry components, and FAO-UNDP will train 35 forestry technicians. Most of these projects have a regional rather than a sectoral focus, and they generally concentrate on tree planting rather than on extension work, conservation programs and training. Fuelwood Conservation 2.11 Programs to improve the efficiency with which fuelwood is used can, in principle, conserve fuelwood much more cheaply than it can be produced by tree-planting projects. Relatively little has been done thus far in Burundi to develop and introduce better charcoal production methods or fuel conserving stoves for firewood or charcoal. Some current projects, including the IDA forestry projects, have, however, begun to address the issue. 2.12 Charcoal is produced primarily by small-scale entrepreneurs, who cut trees and make the charcoal in nearby earthen pits. Yields by this method are very low. A modern efficient kiln, which could utilize gases and liquids formed during the charcoal production process, could theoretically increase overall energy efficiency to about 80%. However, an imported steel kiln might not be appropriate in the Burundian context, and in any case it would be difficult to transport and use the by- products of charcoal production 1/. A simpler, less expensive improved kiln can provide 40 percent energy efficiency. A masonry kiln should be tested as should other models made of local materials 2/. At the present time, there is little incentive for Burundi's small-scale charcoal producers to invest in more efficient but more expensive kilns to conserve wood. Charcoal permits are issued for production of a specific number of sacks (27,000 BuF for 900 sacks currently). If permits were instead issued to produce charcoal from a specified forest area, producers would have an incentive to conserve wood by seeking more efficient methods. Once this incentive were set up, the task would then be to develop a kiln suitable for the Burundian charcoal makers. Some of them should participate in the selection and testing of various kiln models. 2.13 The typical cooking method with firewood consists of an open fire surrounded by three rocks, a very energy inefficient method. Enclosing the fire, regulating airflow into the stove and adding a chimney could increase efficiency by a factor of two. Various stove models have been developed which are low in cost and utilize available local materials. A large clay mass stove, known as the Lorena, is 1 Three metal kilns were ordered in 1981 by the IDA-financed Forestry Project and their use should provide useful information on the potential for improved kilns. 2/ One such kiln was developed by Jos Mabonga-Mwisaka (Uganda) and could be tested in Burundi.  - 14 - relatively sophisticated, yet can still be built for around $10. It is built of a clay-sand mixture and has dampers to control airflow. The chimney can be made of bamboo, bricks, metal, etc. The simplest improved cookstove is little more than a wall of clay to enclose the fire and focus heat into the cooking pot, with a small opening for a chimney. 2.14 The problems involved in any improved cookstove project are socio-economic and organizational as well as technical. Good extension work is crucial for the success of a stove dissemination project. An improved cookstove can be accepted and used if there is intensive field work to develop a model suitable for local conditions, heighten public interest and resolve day-to-day problems with stove use. Annex I outlines a pilot project designed to test the potential of the stove improvement approach. The first step is to do a household sample survey to evaluate current cooking and fuel use practices. Next this information is utilized in conjunction with technical designs of improved cookstoves to develop a model suitable for local conditions 1/. It is then field-tested. Almost daily interaction with the households testing the stoves is required for several weeks to detect problems early, and make appropriate design modifications 2/. A large scale stove dissemination project would require well-trained extension workers to carry out this necessary follow-up. 2.15 In rural areas it might be necessary for extension workers to go from colline to colline 3/ to demonstrate the stoves since there are no established villages. A-model developed for Bujumbura would not necessarily be suitable for rural areas, where people cook with open fires. The open fires provide some light in the house and the smoke acts as an insect deterrent. Rural people should be directly involved in the development of a stove model designed to meet their specific needs, and constructed of locally available materials, primarily clay, at little or no cost. 2.16 Charcoal is currently used in Bujumbura in a small open grill of thin metal, called an imbabura, and much heat is lost both by conduction and by radiation. Charcoal grills can be improved simply and at very low 1/ Prototype stores are being developed in Burundi by AIDR under control to the IDA-financed Forestry Project. 2/ This procedure proved successful on the USAID peat stove pilot project in 1980. 3/ Literally, hill. In Burundi, extended family units reside in compounds called rugos located on the hillsides. Each hill or colline is likely to have several rugos. Part of the family's farmland is located near the rugo. The collines are not like villages, and they are not officially recognized administrative units. The smallest recognized unit is the commune of which there are 79 in the country.  - 15 - cost. One way is to make a double-walled stove and fill the space between the walls with clay, sand or pebbles which serve as insulation. If the design of the current imbabura were changed only slightly to focus the heat on the pot instead of allowing it to be lost by radiation out the sides, a greater fuel efficiency could be obtained. 2.17 An adapted metal charcoal stove might be successful in Bujumbura if the price is right (the average cost of the current imbabura is about 125 BuFL/), and it can be clearly demonstrated that the stoves are convenient and use less charcoal. A variety of models should be field- tested to determine which one(s) are best liked by local cooks. It might then be possible to introduce the improved stove on a large scale using local stove makers and sellers, with a smaller group of extension workers who will identify and help resolve problems with the new stove. 1/ Tobias survey mentioned above, 1980.  - 16 - III. PEAT 3.01 Peat is a potential substitute for firewood for institutions and some small industries, and may also be developed as a household fuel. Although the proven resources are substantial, the exact extent of the exploitable peat reserves in Burundi is not presently known because not all the known bogs have been thoroughly surveyed and production problems may make the greater part of the resources unavailable or prohibitively expensive. 3.02 Burundi's major peat resources lie in an area known as the Akanyaru Basin (or Gran Marais) in the northern part of the country on the border with Rwanda. Recent estimates of the amount of peat in the basin range from 100-200 million tons to 1.5 billion tons. 1/2/ However, the peat is colloidal and the area is flooded or very swampy throughout the year, posing serious problems for extraction. In addition, there are several smaller highland peat deposits located in swampy valley along the Zaire-Nile crestline which are already in production. Estimates are that about one million tons exist in those areas. Another bog of four million tons was recently located east of Ngozi at Nyamuswaga. It will be considered separately due to its size. Akanyaru River Basin 3.03 The full extent of the Akanyaru reserves is unknown since they are underwater and floating vegetation often makes exploration difficult, but the three valleys of Buyongwe, Ndurumu and Kirundo have about 30% of the peat in the Akanyaru and are drier than most of the peat-bearing areas within the basin. The quality is similar to Finnish peat, but the average ash content is higher, about 10 percent, although still within acceptable limits 3/. Manual extraction of this peat is not feasible because most of it is below the water table, but several methods are available for "hydraulic" peat mining, or extracting peat in the form of a slurry which would have a moisture content of 94 to 97 percent, depending on the technology chosen. The moisture content can be reduced by air drying to 30 percent, and mechanically compressing the dry peat 1/ The low end is an estimate provided by the Irish survey team in 1979 and reported in Burundi Alternative Energy-Peat II. USAID, August 1980. The higher estimate is by EKONO, a Finnish consulting firm. It is difficult to determine the exact extent of the submerged Akanyaru reserves. 2/ Peat is conventionally measured by weight adjusted to a 35% moisture content basis. On this basis, its calorific value per unit weight is 25- 30% of that of oil, so a ton of peat at 35% moisture has about the calorific value of two barrels of oil. 3/ EKONO sets 15% as the upper limit.  - 17 - into pellets would reduce the moisture content to 25 percent. High temperature, high pressure processes can further upgrade the peat, but are not currently feasible for Burundi 1/. The scale of hydraulic peat mining operations requires a substantial market: the studies to date have been connected to a nickel project under consideration at Musongati, 160 km to the south. The EKONO study suggests that this project use peat from the Buyongwe bog, which can provide about 100 million tons of peat slurry containing 6% dry solids. Ndurumu bog is proposed as a potential backup and could double the amount of peat available for the project. Depending on the methods chosen for peat and nickel processing, the Buyongwe supply could last 8 to 29 years. An EKONO estimate that may be overly optimistic indicates that the cost of dredging, air drying and shipping peat in the loose, uncompacted state to Musongati would be $55.90/ton. Transport accounts for almost 30% of the price in the EKONO estimate. 3.04 Unanswered questions remain about the environmental effects of large-scale exploitation of the Akanyaru peat resources. It is not known whether the bogs can be drained without causing serious damage to the surrounding area, which includes densely populated farmland, or downstream areas, including parts of Rwanda. Highland Peat Bogs 3.05 Resources in the highland bogs, excluding Nyamuswaga, are estimated to be slightly over one million tons. Five bog sites, Kishubi, Kitanga, Nyacijima, Kashiro and Kurunyange, have proven reserves of over 718,000 tons2/. The quality of the peat varies from bog to bog and even from area to area in the same bog. Ash content ranges from 15-40%. This peat cannot be clean cut due to the variation in quality. Instead it must be macerated, a process which mixes peat found at different levels to produce a uniform quality sod. Thus far, this process, and the forming of peat into sods or bricks for drying, has all been done manually. However, this does not appear practical for large-scale 1/ A discussion of other methods for processing peat, some still in the experimental stage, can be found in the EKONO report. A report by Professor Bertel Myreen of JP-Energy Oy, Helsinki, discusses the wet carbonization process. That process, which is still in the experimental stage, would appear to produce high quality peat fuel, but at very high cost. Effluents produced by the process can cause major environmental damage if not properly treated. The energy yield of this process is predicted to be about 72%. A small amount of electricity from another source is required, but most of the electricity used can be generated at the processing plant by a back-pressure turbogenerator. This is the only process which uses anything but sun or air to dry the peat, which has been considered for Burundi. It does not appear recommendable. 2/ Burundi Alternative Energy - Peat II, USAID 1980.  - 18 - production. The USAID project will test semi-automatic machines which will macerate the peat and form it into peat sods for drying 1/. A 12 ton fully automatic machine was to be tested in 1981 or 1982. 3.06 Most of the highland bogs are located in areas where all available land is used for cultivation. Boglands are especially suitable for dry season cultivation due to the water retention properties of peat. Adequate provision must be made for the farmers who are displaced during peat extraction operations. A layer of peat about 1.5 m in depth should remain after peat extraction is completed so that the bogs can be used for cultivation again. Nyamuswaga Bog 3.07 Nyamuswaga bog may have four million tons of peat. The overall ash content is 17.1 percent -- in the high acceptable range -- but the standard deviation is large. Drainage problems at this bog will be very complex due to its size and a hydrological survey should be done in the near future to determine if and how the bog can be exploited. Draining the bog could seriously affect its capacity to store water which could lead to flood damage in case of heavy rain 2/. It appears likely that drainage should be carried out very slowly Tat 1 m/year) to avoid disturbing the equilibrium of the bog 3/. The hydrological survey should make specific drainage pattern recommendations. 3.08 Nyamuswaga peat is located at a considerable distance from Bujumbura and transport costs may make it too costly for that market. Alternative potential markets should be analyzed to determine the economic feasibility of production at this bog. The possibility of producing briquettes near the bog site should be considered as a way to reduce transport costs. 3.09 Only 10% of this bog is currently under cultivation, but this may soon increase due to the high population density in nearby areas. This makes it essential to carry out the necessary studies as soon as possible. If the bog can be economically exploited now, it should be, while it is still possible to avoid disruption of cultivation and accompanying hardship for farmers. 1 The USAID report contains a discussion of various types of machinery suitable for extraction and processing of peat from the highland bogs. 2/ One bog site, Kivogero, had to be abandoned due to damage caused by flooding which was partly due to improperly constructed drainage channels. 3/ See Justin McCarthy "Danida Prospection Project - Final Report", March 1981.  19 - Potential Production 3.10 The development of the peat industry in Burundi depends on the results of the studies of the Nyamuswaga and Akanyaru Basin reserves. If only the highland bogs can be brought into production, then annual production would peak at 59,000 tons in 1990 and decline thereafter, falling to 42,000 by the year 2000. If Nyamuswaga bog, which has four times as much peat as all the highland bogs combined, were brought into production, both annual and total production could be substantially increased. Hydrological surveys have been undertaken, and pilot production with equipment capable of producing 6,000 tons per year is slated. If the Akanyaru reserves can be used, then the situation changes drastically. An annual production in excess of one million tons could be reached. 3.11 Consumption of peat thus far has been very limited due to the small amount produced. Table 3.1 shows sales for 1977-79. The Burundi army purchases 80% of all peat sold. Ten percent is sold to small brickmaking operations and to a bakery in Bujumbura. The remainder goes to missions located near the bog sites. In 1980 about 3,000 tons of peat were produced. Production in 1981 was about 6000 tons and is expected to reach 10,000 tons in 1982. 3.12 The potential market for peat is large, but projections of consumption are highly speculative. The USAID-financed Peat II Project is oriented exclcusively toward household users. The household market will require the greatest investment of time and resources to develop because it involves a change in cooking fuel in literally thousands of households. That market should probably not be developed unless the amount of exploitable peat reserves increases significantly. It should be easier to develop the institutional and industrial markets. The institutional market is projected to expand to include schools, hospitals and prisons, as well as a larger military market. Currently the army is the largest consumer of peat, but it is obliged to buy it. The industrial market will be a milk factory (AKIRAMATA), and two tea processing factories under construction. The EDF (European Development Fund) provided assistance for developing the Kigozi bog which will supply the EDF-financed tea factories. In exchange ONATOUR agreed to supply them with 4,600 tons of peat per year for 20 years. The COTIBU textile plant has indicated an interest in obtaining 22,000 tons per year of peat, and the projected brick factory and lime kiln may each be able to use 3,000 tons annually. Some firewood users have converted to peat and liked it, but found it was not always available. Others were not trained in how to use it and found that the peat damaged their ovens. The artisanal/commercial market includes bakeries, restaurants and possibly hotels, as well as brickmaking operations and lime kilns. Brickmakers have been generally happy with results obtained using peat. One restaurant has already had constructed a large clay stove for cooking with peat but ran into technical difficulties.  - 20 Table 3 .1: Peat Sales 197-7-1979 Year Sales Receipts (Bu F) Quantityl/ 1977 122,331 37 1978 3,720,230 1,501 1979 6,756,912 2,123 Source: USAID Burundi Alternative Energy-Peat II Project Paper, August 28, 1980. l/ Unit of measure is not given in the source table. Probably metric tons.  -_ 21 - 3.13 If exploitable peat reserves expand to include Nyamuswaga then the urban household market should be developed- Peat is currently almost unknown as a household fuel in Burundi. If it is not well dried and properly stacked for burning, it smokes excessively. A low cost stove was developed for household cooking with peat in the summer of 1980 for the USAID peat project. It was a clay stove based on the Lorena model. The peat stove, which can also be used with wood, and with charcoal with minor adaptations, was tested in about 15 households and acceptance was generally good. The stoves were used continuously as long as the cooks had an available supply of peat. Since no small-scale retail marketing network currently exists for peat, and since it is too difficult to transport it from ONATOUR's storage area without a car or truck, stove- owning households wishing to purchase peat after October 1980 were unable to get it. 3.14 The successful introduction of both peat and stoves for household cooking will depend on an effective extension campaign to train both cooks and artisans in proper construction and use techniques for stoves and peat. A distribution network must be set up so that households may purchase a day's supply of peat very close to home, just as they now purchase charcoal. Although rural households may suffer more from the shortage of cooking fuel, it does not appear that they are likely peat users since they have little disposable cash with which to purchase any kind of fuel. The benefit to them will have to come via decreased use of fuelwood for charcoal for urban consumers, thus making more wood available in the countryside. 3.15 The largest potential market for peat in Burundi is the proposed Musongati nickel processing plant, which would require from 300,000 to 1,200,000 tons/year of peat depending on the processing alternatives selected. Peat supplies on this scale would have to come from the Akanyaru Basin. Institution 3.16 The institution responsible for the peat industry is ONATOUR - Office National de la Tourbe. It was created in 1977 as a planning and coordinating agency for the sector. It has responsibility for all activities relating to the exploration, exploitation and commercialization of peat in the country. It is at present a very small institution and its future development will depend on how much exploitable peat actually exists in Burundi. 3.17 The USAID project is providing expatriate experts to assist in the operation of ONATOUR and to train counterpart staff. Given the scarcity of trained local personnel, it may be difficult to adequately staff ONATOUR, and expatriate staff may actually run the organization without adequately preparing their counterparts to take over management when they depart. However, until studies on the Akanyaru peat and the Nyamuswaga bog are sufficiently advanced that the scope for future peat development in Burundi seems clearer, it may be best to maintain ONATOUR as a small and flexible institution. Burundi cannot afford to have a  - 22 - large number of professionals specialized in peat if exploitable resources turn out to be limited to the current highland bogs. In the short-term, efforts should be focused on establishing good management and accounting practices, and training Burundi staff in their use. Additional technical training should be provided at bog sites. If large- scale production at Akanyaru and Nyamuswaga bogs is found to be feasible, then a corresponding development of ONATOUR would be necessary.  - 23 3 IV. PETROLEUM 4.01 The petroleum sector in Burundi is modest ir scope. There is no domestic oil or gas production or refining activity, supply is entirely in the form of imported products. Demand is low, about 45 mn liters per year or 800 barrels per day.l/ The country is not entirely without petroleum exploration potential, but prospects in the limited prospective area are not very encouraging. The major issues in the sector are related to the very high cost of imported oil -- over $100 a barrel -- and the vulnerability of the supply routes to interruption. The government needs to consider measures to. (a) reduce insofar as possible the prices paid for imports; (b) prepare contingency plans for dealing with interruptions in supply; and (c) make more efficient use of oil and exploit opportunities for substituting lower cost domestic or imported energy sources for oil. Demand and Supply 4.02 Burundi-s per capita consumption of oil products, about ll liters per year, is low even in comparison with other countries at similar per capita income levels. While this primarily reflects unfortunate circumstances, the country's poor access to world markets and lack of a mineral or other resources base for oil-intensive development, it has moderated the effect of recent oil price increases on the national economy. Petroleum products in 1980 accounted for about 15.7% of total merchandise imports (c.i.f. basis), about three times the average for 1975 but still a much lower figure than those of many other countries. 4.03 The product mix of Burundi's oil consumption is heavily weighted toward vehicle fuels, gasoline and gas-oil account for over 90% of the total. The distribution of 1980 imports was as follows; Premium gasoline 47.3% Regular gasoline 2/ 5.5% Gas-oil 37.3% Fuel oil 7.2% Kerosene 2.6% 1/ For comparison, the Mtombasa refinery is designed for a throughput of 95,000 barrels per day. 2/ The low utilization of regular grade gasoline is explained by the high cost of transporting fuel and the lower efficiency of engines burning regular gasoline compared to those using premium or gas-oil.  - 24 - Not all of the vehicle fuels are used in vehicles; stationary engines and small boilers account for part of the demand. And an estimated 30% of the vehicle fuel consumption is accounted for by road construction contractors. 4.04 The largest single industrial consumer is the brewery, which uses about 2.4 mn liters p.a. or 5% of the nation's oil. The textile plant used about 0.46 mn liters in 1980, but is designed to use 7.2 mn liters p.a. in full production. Fuel oil is used primarily to raise steam in both the brewery and the textile plant, making it easier to replace than vehicle fuels. 4.05 'While consumption levels may be low, the unit cost of oil to Burundi is particularly high. Product imports cost about $100 per barrel including freight but exclusive of taxes, or twice what many other countries have to pay. The premium price paid by Burundi reflects its geographic position, primarily through the high cost of the transport routes it requires but also through the disadvantage at which this puts the country in bargaining with a limited number of potential suppliers. 4.06 There are only two surface routes over which large quantities of petroleum products can be imported. The southern route, used for all oil imports through 1977, crosses Tanzania by rail from Dar es Salaam, where there is a refinery, to Kigoma, where oil is transshipped to barges for the trip up Lake Tanganyika to Bujumbura. The northern route began to be used for petroleum products in 1978 when floods and tankcar shortages (possibly related to Tanzania's operations in Uganda) reduced the capacity of the southern route. The northern route has been used for the bulk of petroleum imports since 1979. It runs from the Mombasa refinery by pipeline to Nairobi, where oil is loaded into tank trucks for a 1,500 km journey to Bujumbura via Uganda and Rwanda. Occasionally, products are imported by barge from Zambia. 4.07 Both routes are high cost and make Burundi vulnerable to interruptions in supply due to problems with refineries or transport links in other countries. The cost is particularly high for the northern route, about 25qUS/liter, or $40 per barrel. In addition to the cost of transport itself over this route, Kenya charges a premium price for exported oil products. Gasoline and diesel fuel purchased at spot market prices in the Persian/Arabian Gulf and shipped to Mombasa would cost about 9c/liter, or $14 per barrel, less than the Kenyan export price. 4.08 Transport costs on the southern route are about a third of those on the northern route. An estimate of southern route costs for general cargo comes to $112/ton (9