UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An economic analysis of fossil-fuel substitution for climate change mitigation Graham, Peter John 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2001-0391.pdf [ 4.39MB ]
JSON: 831-1.0075338.json
JSON-LD: 831-1.0075338-ld.json
RDF/XML (Pretty): 831-1.0075338-rdf.xml
RDF/JSON: 831-1.0075338-rdf.json
Turtle: 831-1.0075338-turtle.txt
N-Triples: 831-1.0075338-rdf-ntriples.txt
Original Record: 831-1.0075338-source.json
Full Text

Full Text

An Economic Analysis of Fossil-Fuel Substitution for Climate Change Mitigation by PETER JOHN GRAHAM, RPF B . S c . F . , University of New Brunswick, 1994 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F F O R E S T R Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S Department of Forest Resources Management Faculty of Forestry W e accept this thesis as conforminq to the inquired standard/ THE UNIVERSITY OF BRITISH COLUMBIA August 2001 © Peter John Graham, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of jor&i The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT In 1997, the Kyoto Protocol was adopted to limit greenhouse gas emissions in an attempt to mitigate climate change. The impetus for this thesis is Canada's commitment under this international agreement to reduce national greenhouse gas emissions to 6% below its 1990 levels by 2008-2012 as well as reducing our dependency on fossil fuels. The question posed here is: can using biomass from afforested lands and industrial wood waste as a fuel for energy production be an economically viable tool to reduce greenhouse gas levels in the atmosphere? To answer this, I first examine the two stages of afforestation's role in reducing greenhouse gas levels: its initial use as a carbon sink, and then its use as a renewable energy source that substitutes for fossil fuels. Next I examine the potential supply of biomass from afforested lands as well as from industrial wood waste. The production of ethanol from wood-biomass is then considered. Ethanol offers an excellent opportunity for greenhouse gas mitigation due to market potential, an ability to offset significant emissions from the transportation sector, and reduce emissions from COyintensive waste-management systems. I follow with a case study of the economics of a hypothetical ethanol production facility using mathematical modeling. The results indicate that a facility capable of producing 122 million litres of ethanol annually would have a net present value of $245 million over a planning horizon of 36 years. This facility would require a supply of up to 960 oven-dry tonnes of wood-biomass per day and would result in net annual reductions of greenhouse gas emissions of approximately 349,000 tonnes of COyequivalent (non-discounted). This includes the carbon sequestered through the afforestation of 66,000 hectares over 24 years as well as avoided emissions from fossil fuel substitution. ii In conclusion, then, I am able to answer the question with which I began: using biomass from afforested lands and industrial wood waste as a fuel for energy production can be an economically viable tool for reducing greenhouse gas levels in the atmosphere. By doing so, Canada can take a step towards meeting its Kyoto Protocol commitment, and would be taking a leading role in the vital move toward mitigating climate change. This will also reduce reliance on fossil fuels and reduce the sensitivity of transportation fuel prices to changes in gasoline prices. iii TABLE OF CONTENTS A B S T R A C T T A B L E O F C O N T E N T S iv LIST O F T A B L E S vii LIST O F F I G U R E S viii A C K N O W L E D G E M E N T S x C H A P T E R 1: I N T R O D U C T I O N 1 1.1 CLIMATE CHANGE 2 1.2 DEFINITIONS 3 C H A P T E R 2: OPTIONS T O M I T I G A T E G R E E N H O U S E G A S EMISSIONS 6 2.1 FORESTRY MEASURES: SEQUESTERING CARBON 7 2.1.1 Land-Use Change 8 2.1.2 Afforestation / / 2.2 ENERGY SECTOR MEASURES: REDUCING EMISSIONS 12 2.2.1 Renewable Energy 13 2.2.2 Fossil fuel Displacement : / 7 2.3 CONCLUSIONS 19 C H A P T E R 3: W O O D - B I O M A S S S U P P L Y 20 3.1 SUPPLY FROM AFFORESTATION 20 3.1.1 Availability of Land 20 3.1.2 Cost of Land. 22 i v 3.1.3 Productivity of Afforested Land 27 3.2 SUPPLY OF INDUSTRIAL WOOD WASTE 31 3.2.1 Availability of Wood waste 32 3.2.2 Cost of Wood waste 34 3.3 CONCLUSIONS 34 CHAPTER 4: THE MITIGATIVE POTENTIAL OF WOOD-ETHANOL 36 4.1 WOOD-ETHANOL TECHNOLOGIES 36 4.2 MARKETS AND INCENTIVES FOR ETHANOL PRODUCTION 38 4.2.1 Markets 38 4.2.2 Incentives 40 4.3 CONCLUSIONS 43 CHAPTER 5: A CASE STUDY - AN ECONOMIC ANALYSIS OF A HYPOTHETICAL ETHANOL PRODUCTION FACILITY 45 5.1 ACCOUNTING AND ECONOMIC ASSESSMENT PROCEDURES 45 5.1.1 Carbon Accounting 46 5.1.2 Cost Accounting 46 5.1.3 The Significance of Time in Carbon Mitigation Studies 50 5.2 METHODOLOGY 51 5.2.1 Objective Function 52 5.2.2 Constraints 62 5.3 RESULTS 65 5.3.1 Model Descriptions 66 5.3.2 Optimal Allocation of Resources 67 5.4 SENSITIVITY ANALYSIS 71 5.4.1 Sensitivity to Elements of Objective Function 71 5.4.2 Sensitivity to Constraints 72 V 5.5 ANALYSIS OF RESULTS IN TERMS OF GHG EMISSIONS REDUCTION 79 5.6 CONCLUSIONS 84 CHAPTER 6: DISCUSSION 87 6.1 RECOMMENDATIONS FOR FURTHER STUDY 92 BIBLIOGRAPHY 94 VI LIST OF TABLES T A B L E 1. CHANGES IN COMPONENTS OF TERRESTRIAL CARBON STOCKS UNDER DIFFERENT LAND-USE CHANGES. . . 9 T A B L E 2 . R A T E OF SOIL CARBON INCREASE ( O D T / H A ) COMPARED TO ADJACENT AGRICULTURAL CROPS FOR A NUMBER OF SHORT-ROTATION PLANTATIONS AT DIFFERENT DEVELOPMENTAL STAGES IN BOTH NORTH AMERICA AND EUROPE 1 0 T A B L E 3 . C 0 2 - E Q U I V A L E N T GREENHOUSE GAS EMISSIONS BY SECTOR, ACTUAL 1 9 9 0 , PROJECTED 2 0 1 0 AND K Y O T O TARGET 13 T A B L E 4 . A L B E R T A ELECTRICITY GENERATION CAPACITY 1 7 T A B L E 5 . CARBON EMISSION FACTORS FOR SELECTED ENERGY SOURCES 18 T A B L E 6 . PERCENTAGE OF FARMS BY T Y P E 2 1 T A B L E 7 . ALBERTA'S WOOD RESIDUE ESTIMATES ( 1 9 9 8 ) - A N N U A L QUANTITY, O D T S 3 2 T A B L E 8. BY-PRODUCT YIELD FACTORS - GREEN TONNES / M F B M (MILLION METRIC BOARD FEET) LUMBER 3 3 T A B L E 9 . SUMMARY OF CANADIAN INCENTIVE PROGRAMS 4 2 T A B L E 1 0 . V A L U E S OF YA, BA, AND EA USED IN MODELED SCENARIOS 5 4 T A B L E 11. T O T A L MILL RESIDUES BY DISTANCE ZONE (RD) 6 4 T A B L E 1 2 . SHADOW PRICES OF WOOD WASTE ASSOCIATED WITH SCENARIO-A IN $ / O D T 6 8 T A B L E 1 3 . SUMMARY OF RESULTS OF SCENARIO-A BY PLANNING PERIOD 6 8 T A B L E 14 . SHADOW PRICES OF WOOD WASTE ASSOCIATED WITH SCENARIO-D IN $ / O D T 6 9 T A B L E 15 . SUMMARY OF RESULTS OF SCENARIO-D BY PLANNING PERIOD 7 0 T A B L E 1 6 . SUMMARY STATISTICS FROM THREE SCENARIOS FOR BOTH SCENARIO-A AND SCENARIO-D 7 3 T A B L E 17 . CONVERSION RATES USED IN THE CASE STUDY OF A HYPOTHETICAL WOOD-ETHANOL PRODUCTION FACILITY 8 0 T A B L E 18. SUMMARY OF G H G EMISSIONS STATISTICS FROM THREE SCENARIOS OF SCENARIO-A AND OF SCENARIO-D. 81 vu LIST OF FIGURES FIGURE 1. SURVEY RESPONSE TO STATEMENT: C A N A D A NEEDS TO INVEST IN REDUCING EMISSIONS OF GREENHOUSE GASES : 2 3 FIGURE 2 . SURVEY RESPONSE TO STATEMENT: PLANTING TREES WILL YIELD BENEFITS TO MY FARM (E.G., REDUCE WIND, IMPROVE WATER QUALITY) 2 4 FIGURE 3 . SURVEY RESPONSE TO QUESTION: W H A T TYPE OF TREE PLANTING PROGRAM WOULD YOU PARTICIPATE IN TODAY IF YOU WERE ADEQUATELY COMPENSATED FOR LAND AND PRODUCTION LOSSES? RESPONDENTS CAN CHOSE MULTIPLE PROGRAMS 2 5 FIGURE 4 . SURVEY RESPONSE TO QUESTION FOR THOSE RESPONDENTS WHO WOULD VOLUNTARILY PLANT TREES ON THEIR LAND IF IT DID NOT HAVE A NEGATIVE EFFECT ON THEIR ELIGIBILITY FOR GOVERNMENT AGRICULTURAL PROGRAMS OR ANY TAX BENEFITS: WHAT TYPE OF PLANTING WOULD YOU ENGAGE IN? 2 5 FIGURE 5 . SURVEY RESPONSE TO QUESTION: W H A T WOULD YOU SAY WAS YOUR MAIN REASON(S) FOR NOT CONSIDERING PLANTING TREES ON YOUR LAND? 2 6 FIGURE 6 . SURVEY RESPONSE TO QUESTION: A T THE END OF THE CONTRACT, IF THERE WAS NO POSSIBILITY TO EXTEND IT, HOW LIKELY ARE YOU TO TAKE THE FOLLOWING ACTIONS? 2 6 FIGURE 7 . T O T A L ABOVEGROUND GROWTH FUNCTION FOR HYBRID POPLAR 3 0 FIGURE 8. A N N U A L CARBON ACCUMULATION CURVES FOR POPLAR 3 1 FIGURE 9 . REPRESENTATION OF STUDY AREA DIVIDED INTO ZONES 5 2 FIGURE 1 0 . CHART OF LINEAR CAPITAL COST FUNCTIONS AS DESCRIBED IN EQUATIONS ( 5 .4A) AND ( 5 .4B ) 5 6 FIGURE 1 1 . CHART OF LINEAR OPERATING COST FUNCTIONS AS DESCRIBED IN EQUATIONS ( 5 . 5 A ) AND ( 5 . 5 B ) 5 7 FIGURE 1 2 . GRAPHICAL REPRESENTATION OF THE QUADRATIC L A N D RENTAL R A T E FUNCTION, EQUATIONS 5 . 7 A AND 5.7B 5 9 FIGURE 13 . EFFECT OF EVEN-FLOW CONSTRAINTS ON PLANTING DECISIONS OVER THE PLANNING HORIZON OF SCENARIO-A (THE SMALL-SCALE FACILITY) 7 5 viii s FIGURE 1 4 . EFFECT OF EVEN-FLOW CONSTRAINTS ON DISCOUNTED COSTS ( N P C ) AND REVENUES ( N P R ) OVER THE PLANNING HORIZON OF SCENARIO-A , 7 5 FIGURE 1 5 . EFFECT OF EVEN-FLOW CONSTRAINTS ON PLANTING DECISIONS OVER THE PLANNING HORIZON OF SCENARJO-D (THE LARGE-SCALE FACILITY) 7 6 FIGURE 1 6 . EFFECT OF EVEN-FLOW CONSTRAINTS ON DISCOUNTED COSTS ( N P C ) AND REVENUES ( N P R ) OVER THE PLANNING HORIZON OF SCENARJO-D 7 6 FIGURE 1 7A. GRAPH REPRESENTING THE EFFECT OF TIME AND TRANSPORT DISTANCE ON THE SHADOW PRICE OF WOOD WASTE IN SCENARIO-A, THE SMALL-SCALE FACILITY WITH 2 5 % EVEN-FLOW CONSTRAINTS 7 7 FIGURE 18. EFFECT OF EVEN-FLOW CONSTRAINTS ON G H G BALANCE OVER THE PLANNING HORIZON OF SCENARIO-A.. 8 2 FIGURE 19 . EFFECT OF EVEN-FLOW CONSTRAINTS ON G H G BALANCE OVER THE PLANNING HORIZON OF SCENAR/O-D. 8 3 FIGURE 2 0 . EFFECT OF EVEN-FLOW CONSTRAINTS ON NET G H G BALANCE (DISCOUNTED AND NON-DISCOUNTED PHYSICAL CARBON EQUIVALENT) AT THE END OF THE PLANNING HORIZON OF SCENARIO-A AND SCENARIO-D 8 3 FIGURE 2 1 . EFFECT OF EVEN-FLOW CONSTRAINTS ON MAXIMUM ABOVEGROUND C O , SEQUESTRATION (NON-DISCOUNTED) OF SCENARIO-A AND SCENARJO-D 8 4 ix ACKNOWLEDGEMENTS The research for this thesis has been partially funded by the Sustainable Forestry Management Network. I would like to thank my advisory committee: Dr. G . C . van Kooten; Dr. E . Krcmar-Nozic; Dr. J. Saddler; and Dr. G . Bull. I would also like to thank Dr. A l i Esteghlalian, Pavel Suchanek, Dr. Bryan Bogdanski, and Diane Park for their valuable contributions at various stages of my research. x C H A P T E R 1: I N T R O D U C T I O N In this thesis, I examine the costs and benefits of substituting wood-biomass for fossil fuels in conjunction with a policy of afforestation. If afforestation is adopted as part of Canada's strategy to reduce greenhouse gas emissions, what are the costs and benefits of using the trees as a substitute for fossil fuels? The impetus for this thesis comes from two related issues: climate change and the international commitment to its mitigation; and the importance of the reduction in our dependency on fossil fuels for energy. Reducing our dependency on fossil fuels, of which there is an essentially finite supply, would reduce the effects of sudden and major fluctuations in oil and gas prices on consumers. The incentive for mitigating climate change through the reduction of greenhouse gas emissions is to avoid the resulting economic, social and environmental impacts of the potential damages. In this thesis I do not enter into arguments about whether the current period of global warming is entirely (or in part) due to human activity or a phase in a natural cycle or natural phenomena (e.g., solar flares). Nor will I argue one way or another on the relationship between climate change and greenhouse gases. The scope of this paper is limited to specific strategies and recommendations identified by the Canadian Government (Forest Sector Table, 1999) as helping Canada meet its commitment to reduce greenhouse gas emissions. 1 1.1 CLIMATE CHANGE In December, 1997, at the Third Conference of the Parties in Kyoto, Japan, the Parties to the 1992 United Nations Framework Convention on Climate Change signed a document known as the Kyoto Protocol, agreeing to limit emissions of six greenhouse gases (GHG's). Canada committed to reducing G H G emissions by 6% below 1990 levels by the first commitment period, 2008 to 2012. Unfortunately this target does not appear to have been based on an adequate amount of scientific study or economic analysis; when economic growth is taken into account, the 6% figure jumps dramatically. Under business-as-usual scenarios, emissions are projected to increase to 764 megatonnes (Mt) of carbon annually by 2010. Canada's Kyoto commitment amounts to approximately 565 Mt of carbon. Therefore, meeting our target requires an effective decrease in emissions not of 6%, but of approximately 26% (van Kooten and Hauer, 2000). In April , 1998, a process was initiated by Canada's federal and provincial/territorial Ministers of Energy and Environment to develop a national implementation strategy on climate change and help determine the costs and impacts of reaching our Kyoto Protocol target. To do this, sixteen committees or Tables were established covering various economic sectors, including the forest sector. In November, 1999, the Forest Sector Table published its Options Report that evaluated options "in terms of their costs and mitigation potential as well as a number of other considerations including their implications for competitiveness, environmental and health impacts and employment" (Forest Sector Table, 1999). As stated above, this thesis examines the costs and mitigation potential of the production of ethanol from biomass supplied from industrial wood waste as well as from trees harvested 2 from afforested land. Ethanol's contributions to meeting Canada's Kyoto target are its use as a substitute for fossil fuels, its substitution for other octane-boosting gasoline additives, and its potential to be used in electricity generation. Also, the wood-biomass to be used as feedstock for ethanol production can itself be useful in reducing net emissions of G H G ' s : afforestation increases the size of the terrestrial carbon sink; using industrial wood waste replaces other carbon-dioxide intensive management methods such as landfill and incineration. Clearly, the production of ethanol from wood-biomass merits serious consideration. 1.2 DEFINITIONS The Parties to the United Nations Framework Convention on Climate-Change ( U N F C C C ) have agreed to a definition of 'forest' as it applies to land-use, land-use change and forestry activities. Therefore, the definitions of afforestation, reforestation, and deforestation are linked to the definition of a forest. These U N F C C C definitions will be followed throughout this thesis. Forest is a minimum area of land of 0.05 - 1.0 hectares with tree crown cover (or equivalent stocking level) of more than 10-30 per cent with trees with the potential to reach a minimum height of 2 - 5 metres at maturity in situ. A forest may consist either of closed forest formations where trees of various storeys and undergrowth cover a high proportion of the ground or open forest. Young natural stands and all plantations which have yet to reach a crown density of 10 - 30 per cent or tree height of 2 - 5 metres are included under forest, as are areas normally forming part of the forest area which are temporarily unstocked as a result of human intervention such as harvesting or natural causes but which are expected to revert to forest. Afforestation is the direct human-induced conversion of land that has not been forested for a period of at least 50 years to forested land through planting, seeding and/or the human-induced promotion of natural seed sources. Reforestation is the direct human-induced conversion of non-forested land to 3 forested land through planting, seeding and/or the human-induced promotion of natural seed sources, on land that was forested but that has been converted to non-forested land. For the first commitment period, reforestation activities will be limited to reforestation occurring on those lands that did not contain forest on 31 December 1989. Deforestation is the direct human-induced conversion of forested land to non-forested land. Internationally, these terms have been a matter of contention. The general definition of afforestation is the most agreed upon, and was therefore the only action for which assessed options were closely analyzed by the Forest Sector Table in its 1999 report. However, Canada's definition of afforestation overlaps with Europe's definition of reforestation. Reforestation is defined by European countries as planting trees on land that was once forested (implying a long period of time since the land was denuded), while the Canadian definition includes the re-establishment of trees after harvesting. Despite the fact that the definition of reforestation proposed by Canada is similar to the one used by the United Nations Food and Agriculture Organization (FAO), it was not accepted internationally. Also, under the Protocol, deforestation occurring during the commitment period is considered a debit (source of C 0 2 emissions); however, deforestation is not included in the baseline year (1990) measurement. Therefore, the full level of deforestation in the commitment period is a liability (Forest Sector Table, 1999; MacLaren, 1999). In effect, this methodology did not directly acknowledge any change in deforestation rates between the baseline year and the commitment period. In the recent negotiations at the 6th Conference of the Parties in Bonn, Germany (COP6-2, June, 2001), a general consensus was reached regarding the definitional framework. It was agreed that the definitions should pertain to the human-induced change in land-use, and not the exact condition of the land prior to, and following, land-use change. This avoids problems 4 related to the wide range of definitions for 'forest' and 'agriculture' for example. As a result, although the final definitions have not been set in the Protocol, the international community, including governments (except the United States) and industry, is now more confident in what activities will be included under the Protocol. First I briefly examine the various international and federal agreements and commitments related to the potential policy options. In Chapter 2, options to mitigate greenhouse gas emissions from both the forestry and the energy sector are discussed. The potential supply of wood-biomass from industrial wood waste and from the harvesting of afforested lands is discussed in Chapter 3. In Chapter 4,1 look specifically at the production and benefits of ethanol as an energy product derived from wood-biomass. The formulation and results of a mathematical model of a hypothetical wood-ethanol facility are discussed in Chapter 5, with conclusions and recommendations presented in Chapter 6. 5 CHAPTER 2: OPTIONS TO MITIGATE GREENHOUSE GAS EMISSIONS Carbon dioxide (C0 2 ) mitigation measures can be divided into two categories: source-oriented measures and sink-enhancement measures. Source-oriented measures try to reduce emissions of carbon into the atmosphere. They include energy sector activities such as energy conservation and efficiency improvement, fossil fuel switching, renewable energy and nuclear energy. Because the transportation and energy sectors are the largest source of carbon, they have been the focus of most of the mitigation work globally. However, the forestry industry's sink-enhancement measures are also important for mitigating atmospheric C 0 2 levels. A terrestrial carbon sink such as a forest effectively captures and disposes of atmospheric C 0 2 , storing the carbon in solid form (e.g. wood). Thus, enhancing forest sinks is a crucial part of mitigating climate change (Jepma et al., 1996). This chapter begins by examining the options for sequestering carbon in the forestry sector through land-use change, and specifically afforestation. The second half of the chapter looks at the options for reducing emissions in the energy sector through the use of renewable energy and fossil fuel substitution. This review reveals significant opportunities that, if pursued, would move Canada in the right direction and help to meet its Kyoto commitment. 6 2.1 FORESTRY MEASURES: SEQUESTERING CARBON In forests, C 0 2 is taken from the atmosphere and converted, through photosynthesis to carbon and stored as biomass in four carbon pools: aboveground biomass; dead organic matter; belowground biomass; and soil organic carbon. Aboveground biomass consists of tree stems, branches and foliage. Belowground biomass refers to root biomass, and soil organic carbon (SOC) consists of microbiotic organisms in the soil. The success of forestry measures to sequester carbon must examine their effects on each of these carbon pools. The Kyoto Protocol identifies a number of carbon sequestration options with potential for reducing greenhouse gas emissions. The options related to the terrestrial carbon sink, for which credits or debits are assessed (in Article 3.3 of the Kyoto Protocol), are reforestation, afforestation and deforestation (see Chapter 1). The Intergovernmental Panel on Climate Change G H G inventory guidelines assume that the forest product pool is in equilibrium, and that biomass energy is neutral (i.e. zero emission credits or debits). These assertions assume that the forests are managed sustainably. Jepma et al. (1996, p.246) identify the following seven subclasses of forestry measures to mitigate C 0 2 emissions: 1) Halting or slowing deforestation; 2) Reforestation and afforestation; 3) Adoption of agroforestry practices; 4) Establishment of short-rotation woody biomass plantations; 5) Lengthening forest rotation cycles; 7 6) Adoption of low-impact harvesting methods and other management methods that maintain and increase carbon stored in forest lands; 7) Sustainable exploitation of forests, sequestering carbon in long-lived forest products. These options are viewed as intermediate responses as they are all ultimately limited (Marland et al., 1997; Wright et al., 1993), the first six by area availability and the seventh through the effect of leakage on market demand for timber and the eventual decay of wood (Jepma et al., 1996). However, as this thesis will show, combining forestry measures with energy measures offers the potential for continuous C 0 2 mitigation through sequestration in combination with fossil fuel switching and renewable energy. Thus, these forestry measures are certainly worth further consideration. 2.1.1 Land-Use Change Converting non-forest land to forests will typically increase the size of the terrestrial carbon sink due to increases in carbon in aboveground, belowground and soil organic pools (Table 1). The diversity of flora and fauna will also increase, except in situations where biologically diverse non-forest ecosystems are replaced by forests that consist of single or a few species (e.g., plantations of monocultures and especially exotic species) (IPCC, 2000). Soil carbon plays a significant role in the effects of land-use change (including afforestation) on the carbon balance, and is recognised in the accounting approaches regarding land-use change in the Kyoto Protocol. The amount of carbon stored in the soil organic carbon pool is about 3 times that found in aboveground and belowground pools, and twice the amount 8 found in the atmosphere (Eswaran et al., 1993). Whether the soil organic carbon levels increase or decrease with afforestation depends on the initial conditions of the soil and its location. In most regions, conversion of natural to agricultural land-use results in a rapid depletion of soil organic carbon (SOC) content. Only in those regions with low inherent fertility (nutrient deficiency or toxicity), that have been cultivated for an extended period using best management practices (e.g., fertilisation), is it possible for afforestation to lead to S O C depletion (Lai and Bruce, 1999). S O C depletion is also possible with afforestation if the length of rotation is too short, thereby not allowing sufficient carbon transfers from litterfall and root growth into the soil. Table 1. Changes in components of terrestrial carbon stocks under different land-use changes Source: IPCC (2000). (Direction of arrows indicate either and increase or decrease in carbon stocks. Double arrows indicate a faster relative rate of change.) Biomass Litter/Woody Debris Land-use Change Above ground Below ground Short-term Long-term Soil Organic Matter Wood Products and Landfills Cultivated land to forest ft ft ft ft ftb Non-cultivated land to forest ft fi fi ? fib Forest to cultivated land u u u fi -Forest to grazing land uu u fia u ? -a It is assumed here that upon conversion of forest to grazing land, woody debris is not, or is only partly, removed. Dead roots, in particular, would not normally be removed. If woody debris is removed or burned, only dead roots would add to the short-term increase of woody litter. b Assuming that the forested land is subsequently harvested and used for wood production. 9 Lai and Bruce (1999) suggest that growing specific species for biofuel production is an important strategy for restoring degraded soils. "Assuming C sequestration rate of 0.25 Mg/yr [0.25 tonnes/yr] . . . the C sequestration potential for restoring severely damaged cropland soils is 0.025 Pg/yr [25 tonnes/yr]." (Lai and Bruce, 1999, p. 179.) In a review of the limited studies in this area, Samson et al. (1999) found that estimations of soil carbon increases compared to adjacent agricultural crops ranged from -3.5 to +13 M g ha"1 yr"', depending on plantation species, site, and age (Table 2). The characteristics of tree species considered suitable for short-rotation plantations provide an excellent opportunity for carbon sequestration in soils. The combination of biomass inputs through litterfall, lignified organic matter and deep root systems, along with increased shading, permits significant soil carbon increases under short-rotation plantations (Samson et al., 1999). Table 2. Rate of soil carbon increase (ODt/ha) compared to adjacent agricultural crops for a number of short-rotation plantations at different developmental stages in both North America and Europe. Source: Samson et al. (1999). Source Location Species Depth sampled (cm) Plantation Age (years) Carbon (Mg ha'yr 1) Zan, 1998 Quebec Willow 0-60 4 5 - 13 Mehdi et al., 1999 Quebec Willow 0-60 6 0.8-2.2 Hansen, 1993 North-central U.S. Poplar 0-100 12 - 18 30 1.6 3.2 Grigal & Berguson, 1998 Minnesota Poplar 0- 100 7 - 8 0 Estimated Data Dewar & United Willow - 8 5.9 Cannell, 1992 Kingdom Poplar - 26 7.3 Hansen, 1993 North-central U.S. Poplar 0- 100 4-6 0-30 Loss (not quantified) "C = (-43.5 + 4.48age) / age 1 -3.5 Grigal & Berguson, 1998 Minnesota Poplar 0- 100 5 7-8 15 -2.6 -1 1.2 a C= soil carbon accumulation; age = age of plantation; Equation r = 0.76 10 Land-use change also changes the nature of economic activity. If we look at converting land from agriculture to forestry, there are benefits in the form of new socio-economic opportunities related to the forestry activity. However, there are also costs in the form of forgone agricultural benefits. Fearnside (1997) discusses the social impacts of land-use change including population displacement and loss by some (often disadvantaged) section of society of the use of common property (IPCC, 2000). 2.1.2 Afforestation Due to its large areas of marginal agricultural land, Canada can reduce net C 0 2 emissions through afforestation. From their studies in New Zealand, Ford-Robertson et al., (1999) conclude that "pastoral farming [e.g. raising livestock] is a considerable source of greenhouse gases, and afforestation would rapidly reverse the situation and provide a substantial carbon sink" (p. 143). There are three general options for lands afforested for carbon sequestration: leave trees standing as a carbon sink, harvest the timber for wood products, or harvest the trees as a biomass-fuel. Leaving wood standing does not take advantage of potential economic opportunities and is not an option likely to appeal to landowners. There are also considerable problems with the second option. First, although sequestering carbon in wood products allows for a relatively slow rate of C 0 2 emissions due to the life span of the products, only about 40% of the log volume is used in creating those wood products. That leaves the remaining 60% of the log volume as waste (Row and Phelps, 1995). The second problem is "leakage." In the context of changing land-use, leakage can occur when, due to afforestation, a significant increase in supply of wood products 11 results in a drop in prices that lowers the marginal value of forest production.1 Where that shift in the marginal value is significant enough to result in a switch to agriculture production, the result is deforestation; a reduction in the size of the global carbon sink thereby negates the positive effect of afforestation (Sohngen and Sedjo, 1999). Afforestation itself can result in leakage by reducing the non-timber value of existing trees and forests in other areas. Therefore we look to the third option. Harvesting fast-growing trees such as hybrid poplar from afforested land offers a relatively inexpensive source of wood-biomass for energy conversion. The use of such plantations as a source of energy feedstock can result in emission savings by offsetting fossil fuel use. Also, if managed sustainably (e.g. trees are planted to replace those harvested), such plantations can result in zero net C 0 2 emissions depending on the discount rate (MacLaren, 1999; van Kooten and Hauer, 2000). In conclusion, afforestation can increase the size of the terrestrial carbon sink which in turn contributes to net reductions in greenhouse gas emissions. This benefit is relatively short-lived as there is a finite amount of land available. However, the opportunity for long-term reductions lies in the use of the afforested biomass as feedstock for energy production, thereby offsetting emissions from the production and combustion of fossil fuels. 2.2 ENERGY SECTOR MEASURES: REDUCING EMISSIONS Terrestrial sinks keep carbon locked up and out of our atmosphere. However, in order to reduce the level of atmospheric greenhouse gases in the long term, we must reduce our emissions ' This assumes no increase in demand. A concurrent increase in demand reduces the effects of leakage: demand for wood products can increase along with the increasing size of the population 12 of them in the first place. The transportation sector is the largest source of anthropogenic greenhouse gas emissions in Canada, followed by the industry and energy sectors (Table 3). Emissions from these sectors can be reduced through a number of actions including the use of renewable energy, the displacement of fossil fuel use, and by improving the efficiency of energy consumption. The first two actions are discussed below while the third is discussed specifically in regards to ethanol in Chapter 4. Table 3. C0 2-equivalent greenhouse gas emissions by sector, actual 1990, projected 2010 and Kyoto target, Mt. Source: van Kooten and Hauer (2000). Sector Actual 1990 Projected 2010 Kyoto Target Difference11 Residential 49 48 46 2 Commercial 26 34 24 10 Industrial 125 138 118 20 Transportation 147 197 138 59 Fossil fuel 75 123 71 52 Production Electricity 95 119 89 30 Agriculture 61 72 57 15 Other 23 34 22 12 T O T A L 601 764 565 199 a Projected 2010 minus Kyoto target. Column entries may not sum to total due to rounding. 2.2.1 Renewable Energy In the Kyoto Protocol, the emissions from renewable energy production such as the combustion of biofuels are not included in the accounting of a country's G H G inventory. Therefore, in terms of the carbon balance, converting wood-biomass into energy production includes the following benefits: the maintenance of an emission-sequestration equilibrium; the one-time gain in carbon uptake from initial afforestation; and off-setting emissions from reduced 13 use of fossil fuels. In order to realise these benefits, the obstacles to achieving an increase in the use of renewable energy must be overcome. Before climate change issues came to the forefront, the doubling of world oil prices in 1979 led to considerable interest in renewable energy alternatives. Studies by Helliwell and Margolick (1980) looked at the economics of electricity generation from wood waste and found that, even before 1979, the use of wood wastes to replace fossil fuels was highly profitable from the point of view of society. The Canadian forest industry is a large consumer of purchased electricity despite the fact that it self-generates approximately half of its energy requirements. The forest industry is constrained in many cases from achieving economies of size in power generation by either an inability to sell excess power into the provincial grid or by a lack of fibre to use as fuel (CPPA 2000). Changes in the structure of the electricity market are required to permit an increase in wood waste utilisation. "The forest products industry is unique among major industrial energy consumers in that its production processes and by-products create the potential for the industry to generate renewable energy and virtually eliminate direct fossil fuel C 0 2 emissions. Maximizing the energy generated from biofuels ... creates a significant opportunity for the forest sector to contribute to reducing Canada's G H G emissions" (Forest Sector Table, 1999, p. 13). Currently, much of Canada's industrial wood waste is going into landfills. However, as demand for industrial wood waste increases beyond supply, the value of wood fibre from fast growing energy plantations will increase. In British Columbia, the main reasons that large industrial consumers of electricity, such as pulp and paper firms, did not switch to wood waste was because of the low price of natural gas offered by the provincial utility. In their analyses, Helliwell and Margolick (1980) found that the 14 optimal scale of electricity generation was bound by the introduction of a spatially determined cost function for wood waste. (This relates directly to transportation costs, which are examined in Chapter 5). They also found that energy policies in British Columbia were foreclosing or delaying a substantial amount of investment in the energy use of wood wastes. This assessment is confirmed by the following comparison of British Columbia and Alberta's policies. The energy policies of British Columbia have not changed significantly in the past 20 years and neither have the levels of investment in renewable energies, except where necessary due to environmental regulations.2 B C Hydro, a public utility, supplies electricity to the province and export market principally from hydroelectric generating stations (about 90% of its total capacity). The remaining capacity is generated from one conventional thermal station and from two combustion turbine stations. B C Hydro estimates that their current generating capacity can meet demand until 2007 based on an increase in consumption of 2% per year. As new, large-scale hydro developments are not politically acceptable, B C Hydro is now looking to wind, small hydro plants, hydrogen technologies and biomass to meet future demand. Currently B C Hydro is assisting in the construction of one biomass-fuelled pilot plant scheduled to produce 500 kilowatts (kW) of electricity by 2001 (BC Hydro, 2001) and has agreed to buy power from a proposed 25 megawatt (MW) wood waste powered generating station operated by Lytton Power Ltd. Studies are also underway to investigate potential locations for wind power generation. Small and micro-scaled hydroelectric generators appear to be the main focus of the utility, and some progress has been made with the help of the provincial 2 Due to air quality concerns, beehive burners, commonly used by sawmills to dispose of wood waste, are being phased out. Some cogeneration plants have been built in the areas where disposal costs and waste volumes are highest but most wood waste is now put into landfills. 15 government's restructuring of water rental rates. Unfortunately, based on their progress to date, B C Hydro's commitment to meet 10% of its load growth through green energy sources by 2010 is unlikely to be met. The province of Alberta began restructuring its electric utility industry in 1995 to introduce competitive market forces, thereby increasing incentives for efficiency, eliminating regulatory burden, promoting customer-oriented service, and putting downward pressure on electricity prices (Alberta Energy, 2001). The results of this deregulation in Alberta have included an increase in Power Pool 3 participants from 38 in 1997 to 59 in 2000, and a total of 1,395 M W of new generating capacity between 1998 and 2000.' The new generation for 2001 and proposed generation for 2002 to 2006 is shown in Table 4. Although coal has made the greatest gain, renewable energy sources such as wind and biomass are increasing in usage despite the lack of a specific "green energy" policy in Alberta. The paucity of current, regionally specific research in this area is likely due to the relatively low oil and gas prices in North America. However, the recent substantial increases in North American oil and gas prices as well as concerns over climate change have created renewed interest in renewable energy strategies. In Europe, particularly Scandinavia, where fossil fuel prices are already high, there has been greater development of full-scale biomass energy operations. 16 Table 4. Alberta Electricity Generation Capacity. Source: Alberta Energy (2001). Generation Installed Capacity Type (MW) 1998-2000 2001 2002-2006 Coal 0 50 1860 Natural Gas 6 369 250 Cogeneration'1 1190 106 1280 Gas Turbine 133 0 0 Waste Heat 6.5 0 0 Wind 15.2 103 100 Hydro 12.8 0 112 Flare Gas b 15 0 0 Gas 0 40 525 Combined Cycle 0 0 0 170 Biomass & 17 0 20 Waste Wood T O T A L 1395.5 668 4317 a Cogeneration is the combined production of electrical power and useful heat. b Flare Gas is waste gas captured in oil wells before they are flared. The gas is then used to generate electricity. c Combined Cycle is a variation on the cogeneration process, in which the useful heat is used to operate a steam-driven generating process to produce additional electrical power. 2.2.2 Fossil Fuel Displacement The substitution of renewable energy products (such as ethanol produced from sustainably managed plantations) for fossil fuels results in a reduction in the emission of C 0 2 and other greenhouse gases into the atmosphere. The few studies that have looked at the net reduction in carbon emissions show a range in estimated emission savings of 1.7 to 9.0 tonnes of carbon per hectare per year depending on forest type, discount rates, energy conversion efficiency, and the particular fossil fuel being displaced (Wright et al., 1993; van Kooten et al., 1999b). Technological innovations resulting in increases in conversion efficiency for biomass-3 The Power Pool, created under the Electric Utilities Amendment Act, 1998, is an independently governed body through which all electricity generation is bought and sold in the province of Alberta. 17 fuels will result in increased benefits in net reductions of atmospheric carbon through the displacement of fossil fuels. The amount of emission savings are primarily a function of the carbon content of the feedstock and the substitution ratio of biomass to fossil fuel required to produce an equivalent amount of energy (Table 5). Van Kooten et al. (1999b) estimate the cost of substituting wood-biomass for coal in electricity production ranges from $27.60 to $48.80 per tonne of carbon. This is based on a value of $7.50 per m 3 for hybrid poplar on energy plantations, a substitution ratio of 2.6 - 4.6 m 3 of wood per tonne of coal 4 to generate an equivalent amount of energy. Table 5. Carbon Emission Factors for Selected Energy Sources. Source: van Kooten et al. (1999b). Fuel Higher Heating Value Carbon Content Carbon Coefficient (MJ per kg) (kg C per kg fuel) (kg C per GJ) Wood 15.5-19.7 0.500 25.6 Coal 29.31 0.707 24.12 Natural Gas 0.0317 (m"3) 0.482(iri3) 13.78 Crude Oi l 42.82 0.850 19.94 Kerosene (jet fuel) 46.5 0.858 18.45 Gasoline 47.2 0.869 18.41 Diesel fuel 45.7 0.865 18.93 Liquid petroleum gas 50.0 0.818 16.36 Black Liquor 3 14.36 0.372 25.91 "Source: Levelton (1999). In Table 5 we see that more wood, and therefore more carbon, is required to produce an amount of energy equivalent to fossil fuels. However, wood is a renewable resource and the carbon emitted through energy consumption is assumed to be fully sequestered by the next 4 Based on the range in Higher Heating Value for wood shown in Table 5. 18 rotation of trees, resulting in zero net emissions. Fossil fuels on the other hand are non-renewable and therefore the amount combusted is fully accounted as a debit on a country's G H G emissions inventory, as specified in the Kyoto Protocol. 2.3 CONCLUSIONS Afforestation will result in an increase in the size of the terrestrial carbon sink. As there is a finite amount of land available for afforestation, and trees do not live forever, we must look for a use for the mature trees to provide longer-term carbon sequestration. We encounter the problem of leakage when wood products are considered; therefore, the alternative of using the afforested wood-biomass for energy is a preferred option. Renewable energy production has the advantage of being emissions-neutral under the Kyoto Protocol. Net emissions can be further reduced through the use of wood waste that would have been incinerated or put into landfills, and through the displacement of fossil fuels that would have been burned to provide the energy previously. There are also opportunities for industry and communities to reduce their electricity costs through biomass-fuelled power generators scaled to their particular requirements. The establishment of plantations and the operation of biomass systems will also result in an increase in employment, with most of the jobs created in rural areas. In conclusion, greenhouse gas mitigation would be realised most effectively through the combination of sink-enhancement and source-oriented measures. A system of renewable energy production that would use biomass from afforested land as feedstock would displace the use of fossil fuels in energy production while increasing our terrestrial carbon sink. This certainly would assist Canada in meeting its Kyoto commitment. 19 CHAPTER 3: WOOD-BIOMASS SUPPLY The supply of wood-biomass for bioenergy production relies on two main sources from the forest sector: afforested land, and industrial wood waste. (Another potential source not covered in this thesis is forestry residue left over at logging sites and log sorting areas.) This chapter will examine the availability and costs of biomass supply from afforested land and wood waste, as well as productivity from afforestation, as these factors limit the scale and economic potential of wood-based bioenergy and its contribution to greenhouse gas reduction. 3.1 SUPPLY FROM AFFORESTATION The supply of biomass from afforested lands is a function of the amount of land available, the productivity of that land, the growth and yield of the species planted, and the costs associated with all of these factors. It is important to note that, although many previous studies on the potential of afforestation to contribute to Canada's Kyoto commitment have assessed the physical availability of suitable land, few have assessed its economic availability. These and other factors will be discussed herein. 3.1.1 Availability of L a n d The physical availability is fairly easy to estimate given government statistics on land and agriculture. However, the choice of statistics makes a considerable difference in the final results, as is clear from the following two examples. 20 One method of estimating physical availability is to determine how much of the agricultural land is not being actively managed for agricultural production. Agricultural statistics from a 1996 Alberta census (Government of Alberta, 2001) show a total farm area of approximately 52 million acres (21 million hectares) and an average farm size of 881 acres (357 hectares). The latest farm production figures available from the Alberta Government (2001) show that 21.6 million acres of agriculture crops will be harvested in 2001 (including 3.2 million acres in summerfallow). In Alberta, about 45% of farmland (based on the average farm size) is used for cattle farming and therefore does not produce crops (Table 6). Assuming cattle farms are not included in the measurement of crop acreage, and an average farm size of 881 acres for both crop and cattle operations, the current area of cropland not in managed crop production would be 7 million acres (= 52 million acres x (1 - 45%) - 21.6 million acres), or 2.8 million hectares.5 That 2.8 million hectares could be considered as potentially available and suitable for afforestation. Additional land may be available if some of the area occupied by cattle farms could accommodate trees, and if some of the current crop acreage is only marginally productive for agriculture and would be better suited to tree growth. Table 6. Percentage of Farms by Type. Government of Alberta (2001). Farm Type % of farms Cattle (Beef) 45.2 Grain and Oilseed (except wheat) 18.9 Wheat 9.6 Miscellaneous Specialty 8.8 Field Crop (except grain and oilseed) 7.0 Other Types 10.5 Total 100.0 5 45% of farms = 45% of farm area assuming equal farm size (881 acres). Separate averages for the size of crop and cattle operations would improve the accuracy of the estimate. 21 Another method, using Statistics Canada data, was used by van Kooten et al. (1999a), estimating that 7.25 million hectares of marginal agricultural land are physically suitable for planting trees in the boreal Peace River region of north-eastern British Columbia and north-western Alberta. However, the range in estimates of land availability is not the primary concern, because it is the cost of land that restricts the potential scale of afforestation, not its physical availability. 3.1.2 Cost of Land There may be several million acres of land that would be suitable for afforestation, but the area available ultimately depends upon the landowner: what is the value of that land; how much would they have to be paid to plant trees on it? Very few studies have addressed this aspect of land availability in the economics of afforestation. Therefore, as part of the research for this paper and related projects, a survey of landowners in western Canada (Suchanek et al., 2001) was conducted to determine their willingness to accept tree planting and the significance of a variety of factors upon their decisions. It is unrealistic to assume simply that substitution will occur if the value of tree production is greater than the value of the current agricultural regime, as there are many factors other than crop value that affect land rent or the landowners' willingness to accept afforestation (Jepma et al., 1996). Factors range from relatively quantifiable values such as efficiencies of scale and transaction costs, to non-market values such as visual quality (aesthetics) and resistance to change. Thus, in addition to personal and farm-business data, the Suchanek et al. survey elicited detailed information about a farmer's attitudes and preferences regarding climate change, tree planting contracts, type of planting, species, and many others (see Suchanek et al., 2001). 22 A selection of trends obtained directly from the survey results are interesting as they reflect the opinions of the respondents with respect to afforestation. The majority of landowners surveyed are aware of the climate change issue and most of the respondents agreed that some investments are needed to reduce greenhouse gas emissions (Figure 1). Most respondents felt that planting trees would yield benefits to their farm (Figure 2), such as prevention of soil erosion, providing shade, making use of idle land, and diversifying production. Most respondents also indicated that they would voluntarily plant trees if it did not have a negative effect on their eligibility for government agriculture programs or tax benefits. no strongly disagree neutral agree strongly agree opinion/don't disagree know Figure 1. Survey response to statement: Canada needs to invest in reducing emissions of greenhouse gases. 23 no opinion/don't strongly disagree disagree neutral agree strongly agree know Figure 2. Survey response to statement: Planting trees will yield benefits to my farm (e.g., reduce wind, improve water quality). The landowners' acceptance of different types of plantations was another consideration, as the type of plantation used can have significant impacts on biomass yield, carbon sequestration rates, and financial costs. Types of plantations include block, shelterbelt, strip planting, and planting individual trees. Block planting would generally be the least expensive option per tree and would yield the highest volume and carbon sequestration levels per hectare. The planting and harvesting of individual trees would be the most expensive per tree and would yield lower volumes per hectare. The cost of planting and harvesting shelterbelts, on a cost-per-tree basis, is less than individual trees but more than block planting. When adequate compensation was offered for planting trees, the survey respondents did not indicate any preference regarding the type of plantation except a slightly greater preference for shelterbelts (Figure 3). However, for those respondents who would voluntary plant trees, their preference was for shelterbelts (Figure 4). This implies that they would require more compensation for block planting, which would take up more of their land. 24 Figure 3. Survey response to question: What type of tree planting program would you participate in today if you were adequately compensated for land and production losses? Respondents can chose multiple programs. 6 9 % shelterbelt individual experimental block forest alley cropping Christmas trees planting planting trees Figure 4. Survey response to question for those respondents who would voluntarily plant trees on their land if it did not have a negative effect on their eligibility for government agricultural programs or any tax benefits: What type of planting would you engage in? Those respondents who chose not to plant, even when offered adequate compensation, did so primarily because of their resistance to change (Figure 5). When those who said that they would plant trees if adequately compensated were asked what they would do with the trees at the end of the contract period, their response shows an appreciation for the economic value of the 25 trees (Figure 6): most chose to wait until the trees reached maturity, or when their rate of growth falls below the discount rate (the financial rotation age). Figure 5. Survey response to question: What would you say was your main reason(s) for not considering planting trees on your land? 67% not likely likely very l ikely 0 Immediately Harvest T rees B De lay Harvest ing T rees Until They R e a c h Maturity Q N e v e r Harves t T rees 1 m Harvest T rees On ly If Agricultural Pr ices A r e H igher T h a n N o w Figure 6. Survey response to question: At the end of the contract, if there was no possibility to extend it, how likely are you to take the following actions? 26 Suchanek et al. (2001) asked respondents if they would accept $x per acre to plant trees (with x varying between respondents). They then applied a probit model to the random utility framework in order to estimate the probability of the acceptance of a tree-planting program. The results indicate that the main, significant factors influencing the landowners' willingness to accept tree planting were: i) compensation offered less the opportunity cost; ii) number of acres of farmland already covered with trees; and iii) visual appeal of landscape. The landowners' main reason for keeping land in agriculture was that the forestry options presented appeared financially unattractive. The results of the analysis confirm the presence of other costs, perhaps non-market, in addition to the opportunity costs of agricultural production, that need to be compensated for (see van Kooten et al., 2001). But, as long as the compensation offered was at least as much as the opportunity cost of not producing, the landowner would accept the bid and allow afforestation. The importance of the landowners' willingness to accept tree planting is seen in the case study described in Chapter 5, where willingness to accept is converted to a land rental rate function and included in the cost of afforestation. If the objective of an afforestation policy is to maximise the area planted given a limited budget, then attention must be paid to the owners of the land and how their preferences influence the compensation they will require. 3.1.3 Productivity of Afforested L a n d The productivity of afforested land, in terms of how much wood-biomass can be produced, depends on site productivity, species planted, and the management regime used. In an 27 afforestation program, site productivity depend on the quality of the land that the landowner is willing to lease for tree planting. The species planted and management regime depends on the management objectives. For example, if the objective is to sequester carbon and increase wildlife abundance and diversity, the management strategy may be to plant long-living species and leave them to grow, decay and regenerate naturally in perpetuity. This strategy may be preferred from a socio-economic standpoint if carbon storage plus biodiversity, recreational and other benefits exceed opportunity costs. Or, if the objective is to plant trees as an energy crop, which is the focus of this thesis, the rotations must be short, but with high enough yields to minimise costs. The productivity of land initially available for afforestation is not likely to be very high. The agricultural land generally assumed to be available for afforestation falls under Classes 4, 5 and 6 of the Canadian Land Inventory classification system of agricultural land suitability (Samson et al., 1999). The descriptions of these classes are: • Class 4: land having severe limitations restricting crop range or requiring special conservation practices; • Class 5: land having very severe limitations restricting the capability to produce perennial forage crops; • Class 6: land capable of producing only perennial forage crops. But the land in these Classes may be highly suitable, and productive, for growing trees. The site productivity, in this situation, is a measure of the potential of the site to support tree growth. The growth rate of trees is a function of nutrient and moisture availability, as well as the physiology of the tree itself. Site productivity can be enhanced through a number of 28 means including fertilisation, site preparation, or irrigation. The least expensive method of managing a poor site is to plant species that are more tolerant of productivity-limiting factors. The species preferred for short-rotation forestry programs are hybrid poplar and willow. Most of the information on hybrid poplar production comes from North America while information on willow production has been developed in Europe. Plantations of hybrid poplar and willow can be economically viable on sites of soil Class 4 and 5, but that production on poorer, Class 6, sites is not economically viable (Samson et al., 1999). Willow is typically grown at densities of 12,000 to 15,000 cuttings per hectare.6 They are managed using a coppice system on a 3-4 year rotation, which can be sustained for 20-25 years (Samson et al., 1999). Harvesting costs for willow plantations represent a large percentage of the total costs of production. Harvesting technology is constantly changing but the current systems involve a single pass with a machine that cuts and chips the willow stems. Harvesting losses (wood-biomass left on site) and damage to stools (stumps from which coppice or vegetative growth originates) are the main factors limiting harvesting productivity. Hybrid poplar is typically grown at densities of 1,100 to 1,400 cuttings per hectare. Most planting is still done by hand although mechanised planting systems have been developed. Rotations are generally around 12 years in length, and harvesting can be carried out by conventional forestry equipment. The longer rotations of hybrid poplar are less nutrient demanding and therefore require minimal fertiliser additions compared to the shorter rotations of willow. Thus, hybrid poplar is a better candidate for afforestation purposes due to these site 6 Cuttings are lengths of stem cut from existing willow trees. The cuttings sprout once planted in a suitable medium such as mineral soil. 29 productivity limitations. In the case study presented in Chapter 5,1 employ a yield function based on the Chapman-Richards model (Figure 7) (van Kooten et al., 1999a). Kort and Turnock (1999) developed biomass equations for shelterbelt species, including hybrid poplar, based on stem diameter measurements. These biomass equations were then transformed to estimate annual carbon accumulation. Figure 8 shows the results for poplar from the Kort and Turnock study in comparison to the accumulation rates used in the case study in Chapter 5. Initial growth rates are similar but after about 15 years the Kort and Turnock rates maintain their slopes, while the rates of accumulation used in the case study decline after the trees reach maturity and their growth begins to slow. The Kort and Turnock results do not take into account the trees changing growth rate with age and suggest, incorrecty, a potentially infinite accumulation of carbon. « 5 0 0 T o m o m o m o m o L D o T - T - C N C N I O O M - ' S I - U I I O C O Age (years) Figure 7. Total aboveground growth function for hybrid poplar. 30 35 ., c o J3 < -K 50 0 10 20 30 40 Age (years) Figure 8. Annual carbon accumulation curves for poplar, including estimates by Kort and Turnock (1999) and the function used in the case study presented in Chapter 5 of this thesis, derived from van Kooten et al. (1999a). The supply of biomass from afforested land, then, will be limited by the quality of land available and the growth characteristics of the species planted. If site-specific management regimes are applied, hybrid poplar plantations on marginal farmland may be economically viable sources of wood-biomass for energy production. 3.2 SUPPLY OF INDUSTRIAL WOOD WASTE The potential supply of industrial wood waste as a source of biomass depends on availability and cost. As mentioned previously, this thesis does not include woodland residues (e.g., branches, tops, stumps, roots) associated with regular harvesting activities because of two main factors: obtaining reasonable estimates throughout the study area is very difficult; and the availability of woodland residues will likely be limited to roadside accumulations due to 31 economic and ecological constraints. Sources in this study are thus limited to wood processing residues, which mainly consist of sawmill and pulpmill residues. 3.2.1 Availability of Wood Waste In 1998, Alberta's forest sector produced approximately 1.4 million oven-dry tonnes (ODt) of surplus wood waste. This includes 445 thousand tonnes of shavings and sawdust (whitewood waste) (Table 7). Based on the distribution of surplus whitewood waste in Alberta, McCloy and O'Connor (1999, p.23) conclude that: " A logical location for a wood-ethanol facility would be Grande Prairie where there are a total of 194,545 surplus BDts 7 within a 200-km radius. This feedstock supply can be augmented by primary clarifier sludge from pulp mills in Grande Prairie, Peace River and Taylor B C estimated at 16,000 BDts as well as unknown quantity of chip fines. Grande Prairie is connected by rail to Edmonton where a number of oil refineries are located." Table 7. Alberta's wood residue estimates (1998) - Annual Quantity, ODts. Source: M c C l o y and O'Connor (1999), p.22. Residue Type Production Utilisation Available Residue Surplus % Bark 714,332 270,771 443,561 31.8 Sawdust/Pins 399,966 179,527 220,439 15.8 Shavings 414,178 189,121 225,057 16.1 Trim Blocks 37,117 29,784 7,333 0.5 Log Yard Debris 416,003 115,404 300,599 21.5 Other 1,321,316 1,230,109 91,207 6.5 Dry Waste 186,234 180,289 5,945 0.4 Wet Waste 178,847 135,100 43,747 3.1 Sludge 158,922 101,529 57,393 4.1 Total 3,826,915 2,431,634 1,395,281 100.0 7 BDt (Bone-Dry tonne) = ODt (Oven-Dry tonne) 32 The availability of such residues relates explicitly to provincial harvest and allowable cut levels. M i l l residues are simply the by-products derived from conventional forest product processing, but waste/input ratios vary by tree species, log condition, type of processing and end product. A report by Intergroup from 1982 derived mill residue estimates separately for bark, sawdust, shavings and chips, based on each study region's allocation of hardwood and softwood harvest to sawmills and pulpmills. Sawmill chips were not included in available supply, as it is assumed that all would be used in pulp production (although Intergroup found that the supply was greater than demand in some cases). More recently, Forintek Canada Corporation estimated surplus sawmill residues in Canada (except British Columbia and Quebec) in terms of by-product yield (Table 8). Table 8. By-product yield factors - Green tonnes / Mfbm (million metric board feet) lumber. Source: M c C l o y and O'Connor (1999), p.13. Pulp Chips Sawdust Shavings Bark 1.73 0.446 0.217 0.496 In Intergroup's (1982) analysis, mill residues in British Columbia were based on provincial surveys while in Alberta (as well as the other provinces) mill residue factors were applied to volumes of roundwood received at the mill to calculate oven-dried (OD) weights of various residues (e.g. chips: 0.108 ODt/m 3 ; bark: 0.048 ODt/m 3 ; sawdust & shavings: 0.044 ODt/m 3 ). Forecasting residue supply from mills can be complicated by uncertainty in annual harvest levels (especially in B . C . where land-use planning is highly politicised and can lead to 33 sudden changes in regional harvest levels). Residue supply is expected to decrease due to developments in technology and process efficiencies that tend to reduce the amount of waste from the various processing stages. In addition to decreased residue production, an increase in competition from other secondary processing industries, such as medium-density fibreboard, is expected and will continue to reduce the availability of wood residues (CFS, 1999; Bronson Consulting Group, 1999; Skog et al., 1995; Rinebolt, 1995). 3.2.2 Cost of Wood Waste Currently, industrial wood waste is often given free-of-charge to those willing to bear the expense of transporting it. However, with the development of secondary industries that use wood residues in their production processes, these residues are taking on a value greater than the cost, to the mill, of disposal. The cost of mill residues depends on the type of waste (e.g., bark, chips or sawdust) and the distance of the residues from the secondary industries. Whitewood chips and sawdust are the most valuable due to demand from Medium-Density Fibreboard (MDF) and other engineered wood product industries. A decrease in the availability of wood waste, combined with an increase in the number of secondary wood-processing industries and bio-energy systems, will result in an increase in the value of the wood waste. 3.3 CONCLUSIONS There are currently millions of tonnes of industrial wood wastes that are being incinerated or dumped into landfills. In most cases this waste is available for the cost of transportation. However, due to increasing wood processing efficiency as well as competition for the waste, the 34 supply of inexpensive industrial wood waste is expected to decrease. As a result, the demand for biomass from fast-growing plantations will increase. Using the best-suited species with site-specific management regimes on marginal farmland, afforestation may be economically viable. Thus, although the economic balance may change to favour one over the other, both afforestation and wood waste are important sources of the supply of wood-biomass for bioenergy production. 3 5 CHAPTER 4: THE MITIGATIVE POTENTIAL OF WOOD-ETHANOL In Chapter 2,1 discussed options for reducing greenhouse gas emissions within the energy sector. In this chapter, I look specifically at the ethanol industry, its technical challenges and its potential to contribute to net reductions of greenhouse gas emissions in Canada. It should be noted that, although the Kyoto Protocol's accounting system is not yet finalised, it is important to avoid double counting of emissions credits from reduced fossil fuel consumption with credits for producing ethanol that will offset the same fossil fuel consumption. The production of ethanol from cellulosic materials is increasing internationally, with production plants being established in the United States, Canada and Sweden. Production and use of ethanol, as an alternative transportation fuel or as an octane booster, help reduce greenhouse gas emissions from road vehicles and can promote sustainable development and management of forests and forest industries through waste minimisation. This presents a two-fold environmental advantage as it has been realised that stabilisation of atmospheric greenhouse gas concentrations requires both reduction of fossil fuel consumption and preservation and enhancement of carbon sinks and reservoirs, such as forests (Lashof and Hare, 1999). 4.1 WOOD-ETHANOL TECHNOLOGIES The technology for converting grain to ethanol has been around for some time. The impetus for developing wood-ethanol conversion processes stems from the availability of low-cost wood residues and, more recently, the desire to reduce greenhouse gas emissions. Ethanol is 36 derived from sugars present in lingo-cellulosic materials such as agricultural, hardwood and softwood residues. However, due to the different structural and chemical compositions of these three materials, different processes have been developed to deal with the different feedstocks. Softwoods have greater lignin content than hardwoods or agricultural residues, and it is the lignin that initially hinders the processing of the sugars into ethanol. There are currently five types of wood-ethanol processes, each at a different stage of development. The M c C l o y and O'Connor report prepared for the Forest Sector Table (1999) examined the current technologies with a key requirement being their ability to deal with softwood residues, the dominant feedstock in Canada. The conversion of wood-biomass to ethanol involves the following steps: pre-treatment of feedstock, hydrolysis, fermentation, and ethanol recovery. Iogen Corporation is a leader in the field of wood-ethanol production and, with the assistance of the Canadian federal government and a partnership with Petro-Canada Corporation, has constructed a demonstration plant in Ottawa. Iogen uses an enzymatic process designed initially to process agricultural and hardwood wastes. A co-product of the process is lignin, which can then be used as starting material in other processes or as a fuel to produce steam or electricity. Due to its current stage of development and the association of U B C ' s Forest Products Biotechnology group with this work, the enzymatic hydrolysis process is used in the case study presented in Chapter 5. Ac id hydrolysis is a process that has been known for over 100 years, but was abandoned due to the lower production costs of petrochemical-derived ethanol. Currently, high capital and operating costs limit the development of industrial scale facilities. Two other processes that 37 involve acid hydrolysis were assessed in M c C l o y and O'Connor's report, one involving the solution of feedstocks with acetone and acid, the other involving the gasification of wood. Both of these processes are at early stages of development, but the gasification process may prove to be an effective and efficient technology. The gasification process has the advantage that many of its technological components are currently used at industrial scales in other energy applications. Also, in theory, gasification is able to handle a wide range of feedstock types. 4.2 MARKETS AND INCENTIVES FOR ETHANOL PRODUCTION M y main source of information on markets and' incentives for ethanol production is the M c C l o y and O'Connor (1999) report prepared for the Forest Sector Table. In it, the current and projected ethanol markets are discussed along with a number of incentive mechanisms that may assist the industry's growth in Canada. Comparisons to the United States are made due to the current level of development of the ethanol industry in that country. 4.2.1 Markets Currently in Canada, about 175 million litres of ethanol are being produced from grain. No industrial wood-ethanol plants are presently operational. Total ethanol production represents only 0.5% of the 35 billion litres of gasoline sold in Canada each year, and only 5% of the potential market if all gasoline was blended with 10% ethanol. The wholesale price of ethanol is estimated at 40 cents per litre in Alberta and 43 cents in British Columbia, including Federal and Provincial tax incentives (M c Cloy and O'Connor, 1999). The wholesale price of gasoline is approximately 21 cents per litre. 38 Ethanol is used by gasoline companies in six provinces and the Yukon Territory. Some companies, such as Mohawk Oil , use it for its clean-burning properties to address a particular market niche while others, such as Petro-Canada, use it for its octane-boosting properties. However, ethanol must compete with other octane-boosting agents such as M M T 8 and M T B E 9 that are currently used by refiners. In light of the current health and environmental concerns related to the use of M M T and M T B E (discussed in the following section), ethanol presents an attractive and effective alternative to these additives, as well as contributing to the reduction of harmful emissions from gasoline combustion. These benefits, combined with relatively low costs, indicate the opportunity for the ethanol industry to take a larger share of the fuel additive market and increase its share of the transportation fuels market as well. A n examination of U.S. ethanol markets provides an indication of the potential for growth in the Canadian industry given appropriate incentives. The U.S. currently produces about 6.5 billion litres of ethanol per year from 42 production facilities. Almost all U.S. gasoline retailers use ethanol for at least part of the year. Domestic U.S. consumption accounts for approximately 5.2 billion litres. With annual gasoline consumption of 450 billion litres, and most of the ethanol used in 10% blends, ethanol is used in approximately 12% of the gasoline sold in the U.S. The potential short-term growth in the U.S. ethanol market will depend mostly on changes in state legislation to favour ethanol use. Also, an increase in production of E85 8 MMT (Methylcyclopentadienyl Manganese Tricarbonyl) is an additive used to boost octane in unleaded gasoline. The Canadian government imposed a ban on the inter-provincial trade of MMT, but this was later rescinded due to the threat of legal challenge by manufacturers and provincial concerns. 0 MTBE (Methy Tertiary Butyl Ether) is a by-product of the natural gas industry used as an additive to increase oxygen content of gasoline. 39 vehicles (vehicles that run on 85% ethanol blends) will increase demand. Studies looking at the role of ethanol in climate change mitigation scenarios project annual demand of 36 billions litres by 2015 (Sheehan, 1998), or 145 billion litres by 2020 (Lynd, 1996). These projections assume the success of research and development programs in reducing the cost of ethanol production, and a substantial increase in the sales of E85 vehicles. The U.S. Department of Energy (1998) projects the production of cellulose-based ethanol to be between 20 and 28 billion litres by 2010. 4.2.2 Incentives Health and environmental concerns are leading incentives to reduce G H G and other noxious gas emissions. Ethanol can play an important part in these reductions. As previously mentioned, M M T and M T B E are currently widely used instead of ethanol as octane-boosting agents in gasoline. The manganese component of M M T is a known neurotoxin at high exposure levels but current research on the effects of low-level exposure is inconclusive. M T B E is a more expensive octane booster than M M T but has the added benefits of lowering carbon monoxide (CO) and volatile organic compound emissions (VOCs). However, M T B E has been linked to groundwater contamination in the US and is a significant source of G H G emissions, as it is a by-product of the natural gas industry. The combustion of gasoline releases V O C s , nitrogen oxides (NO x ) , C O , and particulate matter (PM). The combination of V O C s and N O x with sunlight results in the formation of low-level ozone, the main component of smog. C O is a deadly poison and the inhalation of fine particulate matter (PM) is a serious health concern. 1 0 A blend of 10% ethanol in gasoline has the 1 0 In 1990, Health Canada and Environment Canada estimated that 13% of all P M 2 5 emissions in Canada were from the transportation sector. 40 potential to reduce V O C s by up to 10%, and reduce C O emissions by 8 - 30%. The use of ethanol-blended gasoline has the potential to reduce significantly particulate emissions although this claim has yet to be fully substantiated with data (M c Cloy and O'Connor, 1999). If blended at the refinery, as opposed to 'splash blending' outside the refinery, ethanol-blended gasoline can reduce N O x emissions, thus further reducing the potential for smog. A number of incentive mechanisms exist in both Canada and the United States that should lead to an increase in the production of ethanol. First, the U.S. will be discussed; there, these mechanisms include legislation to stimulate demand, tax reduction on ethanol-blended gasoline to stimulate production, and the provision of guaranteed loans for capital costs to reduce risk. Legislation requiring minimum oxygen content levels in gasoline has been passed in a number of states to deal with air quality problems, specifically to reduce C O levels in the winter and ground-level ozone in the summer. Minnesota has been very effective at increasing ethanol production in that state by requiring that almost all gasoline sold in the state contain ethanol. The U.S. federal government has offered reduced taxes on ethanol-blended gasoline since the 1970's (prompted by the oil crises), and the incentive is currently at 21.6 cents per litre (Cdn). Individual states have provided additional incentives (up to 3.25 cents/litre (Cdn) in Alaska) directly to ethanol producers to encourage production in their state. Additional incentives in the form of special loans, property tax assessments, income tax credits, and preferential purchasing policies can assist in the financing of individual projects. In Canada, there is currently a Federal Excise Tax exemption for ethanol produced from biomass (wood or grain). In addition, a National Biomass Ethanol Program administered by the 41 Farm Credit Corporation provides a line of credit to qualified ethanol manufacturers as a means of rescheduling their long-term debt, thereby reducing risk related to the potential for future increases in taxes and feedstock prices or a decrease in oil prices. However, this program is limited in terms of the ethanol production volumes it can cover (M c Cloy and O'Connor, 1999). Five of the provinces have incentive programs for ethanol production as summarised in Table 9. Table 9. Summary of Canadian Incentive Programs. Source: M c C l o y and O'Connor (1999). Province Incentive Conditions Quebec 19.76 cents per litre ethanol Not yet proclaimed (as of 1999) Ontario 14.7 cents per litre ethanol Manitoba 2.5 cents per litre of fuel containing Ethanol produced in a minimum of 10% ethanol Manitoba from Canadian biomass (grain or wood) Alberta 9.0 cents per litre ethanol British Columbia 11.0 - 15.0 cents per litre fuel Applies to blends containing a minimum 85% ethanol Two large grain-ethanol plants have been financed in Ontario and Quebec through the provision of agreements that guarantee the level of tax incentives for a number of years, thereby reducing the level of uncertainty regarding future changes to government policies and programs. The difference between the level of federal incentives in Canada and the U.S. appears to be the main reason why Canada's ethanol industry has not kept pace with the U.S. As M c C l o y and O'Conner (1999) point out, "the Canadian fuel industry is much more national in scope than the US industry and prefers single marketing and product programs across the country" (p.48). For this reason, provincial incentives are not sufficiently effective in developing the ethanol industry. A doubling of federal incentives, combined with financing assistance at the Provincial 42 level, would likely result in a rapid increase in investment (most likely from the oil companies themselves) in ethanol production. 4.3 CONCLUSIONS Given that one litre of ethanol contains 65% of the energy of one litre of pure gasoline, and accounting for relative efficiencies, one litre of 10% ethanol-blended gasoline replaces 0.8 litres of gasoline. The lignin co-product from the wood-ethanol conversion process can be burned to generate power, thereby replacing power generated from natural gas or used as a source material for other (perhaps chemical production ) processes. This combined reduction in the use of gasoline and natural gas translates to about 0.0027 t - C 0 2 per litre of avoided emissions through the production of wood-ethanol. By 2020, ethanol production is estimated to reach 525 million litres per year, resulting in 1.4 million tonnes of avoided C 0 2 emissions per year (M c Cloy and O'Connor, 1999). However, the potential G H G reductions from the increased production of ethanol will depend greatly on the development of additional tax incentives and the development of the technology. The adoption of stronger incentive programs, particularly at the federal government level, would provide the industry with the initial push it requires to justify the high capital cost and risks associated with wood-ethanol production. The potential market growth for ethanol described in this chapter is likely to be maintained due to the general population's level of concern for the environment and due to efforts to meet our Kyoto commitment through reductions in G H G emissions from the transportation sector. Whereas the potential damage resulting from climate change may be perceived by many as not critical enough to require 43 immediate action, the effects of decreased air quality in rapidly-growing cities are pushing policy and law makers to demand cleaner sources of transportation fuel. 44 CHAPTER 5: A CASE STUDY - AN ECONOMIC ANALYSIS OF A HYPOTHETICAL ETHANOL PRODUCTION FACILITY This case study examines the economics of a hypothetical wood-ethanol production facility located in Grande Prairie, Alberta. Mathematical programming is used to determine the optimal allocation of resources (afforested land and wood residues) from the point of view of a profit-maximizing ethanol producer. The model is solved using MicroSoft Excel with the Premium Solver Platform developed by Frontline Systems Inc. The allocation of land-use and resulting production of ethanol is then examined in terms of the impacts on net C 0 2 emissions. This chapter begins with a discussion of alternative accounting and economic assessment procedures and follows with a detailed description of the model used in this case study and an analysis of results. 5.1 ACCOUNTING AND ECONOMIC ASSESSMENT PROCEDURES There are a number of approaches to assessing the potential of greenhouse gas mitigation activities. The design of the accounting and assessment procedures can greatly influence the accuracy of the results. Some activities warrant different procedures than others, but for comparative purposes these procedures must be clearly described. Before describing the methodology used in the case study, I review some of the approaches that have been used by others. 45 5.1.1 Carbon Accounting MacLaren (1999) identifies a conflict between three general carbon accounting approaches for harvested wood products: 1. Scientific purists account for every carbon source, sink and reservoir; 2. Pragmatic accountants concentrate on critical and easily quantifiable data, acknowledging that data-collection resources are limited; 3. Other accountants are concerned with international equity, so that no country or grouping is greatly disadvantaged. Conflict arises because estimates of C 0 2 absorption by forests vary significantly depending on the techniques employed to value them. L i m et al. (1999) evaluated three carbon accounting approaches: atmospheric-flow, stock-change, and production. The atmospheric-flow approach calculates the flow of carbon to and from the atmosphere, while the stock-change and production approaches calculate the net change in stocks in the forest and product pools. Under this framework, the approach taken in this thesis is pragmatic and follows a production-based methodology. 5.1.2 Cost Accounting Studies relating to energy plantations, fossil fuel substitution, or energy production establish varying boundaries on the system that is assessed. For example, some papers dealing with afforestation ignore the opportunity cost of changing land-use, or land rental rates. The various elements and activities associated with bioenergy systems described above all have associated dollar values, although some are easier to determine than others. Hourcade et al. (1996) emphasize the importance of clearly identifying the types of costs (e.g., direct, sectoral, 46 macroeconomic, or welfare) that are included in the analysis. Clearly formulated models make comparison of results much easier and more meaningful. Those responsible for preparing International Panel on Climate Change (IPCC) or national climate change recommendations often find it difficult to assess the benefits of one technology over another due to varying model formulations and unclear assumptions of the individual analyses. From the point of view of policy, the total costs of emissions reduction are not as important as the net costs (e.g. the total costs minus any positive side effects of mitigation). In a review of projects assessed under the Land-Use, Land-Use Change and Forestry ( L U L U C F ) section, the IPCC found that methods of financial analysis among the projects were not comparable. In most cases significant costs were not included in the calculations, including costs for infrastructure, monitoring, data collection and interpretation, opportunity costs of land and maintenance, or other recurring costs. A simplified comparison of the studies produces a range of $0.1 to $100 per ton of carbon (US) (IPCC, 2000; Stennes, 2000). A detailed and standardized (or more transparent) accounting system would provide more useful estimates of the costs and benefits of greenhouse gas mitigation projects. In their review of economic studies of greenhouse gas mitigation, Hourcade et al. (1996) found that the higher the underlying economic growth assumed in the baseline scenario, the greater the estimated costs of mitigation. By employing the 'with-without' principle of cost benefit analysis, the costs of mitigation can be shown relative to a given baseline. When assessing the technical potential of financial costs and savings, the preferred models are those with more detailed representations of technology, and those following a 'bottom-up' approach. Hourcade et al. (1996) found that 'bottom-up' approaches, which assume that there exist 47 substantial correctable market imperfections, show significant no-regrets potential; a 'top-down' approach, which assumes that existing markets are relatively efficient, show little no-regrets potential. A 'no-regrets' action is a preventative action that is cost-effective, resource efficient and worth doing for reasons other than G H G mitigation. Jepma et al. (1996) present two approaches to the economic analysis of C 0 2 emission reduction/absorption. The "engineering efficiency" approach estimates the financial costs of the various technologies. The "welfare economic" approach determines the costs and benefits of the application of any particular technology including an assessment of the opportunity costs of the resource allocation. In the case of an afforestation program, the engineering efficiency approach would involve the determination of the discounted value of the costs directly associated with the program (land acquisition, tree planting, maintenance, etc.) and the discounted returns from future (sustainable) harvests. The net levelized costs could then be determined in dollars, and "on the basis of this information, and by comparing with other options' cost-efficiencies, one could then decide whether or not to proceed" (Jepma et al., 1996, p.235). The welfare economic approach to analyzing the cost-efficiency of an afforestation program would have to consider the opportunity cost of afforestation in terms of lost agricultural potential. For this approach, Jepma et al. (1996) present the argument that if an afforestation program is applied on a large scale there are additional impacts that should be included in project appraisals. There may be impacts, either positive or negative, on factors such as local climate, soil fertility, social and cultural life, infrastructure, tourism, water and hydrology. 48 Jepma et al. (1996) suggest that government measures, such as subsidies and taxes, be considered in assessments as they can have distorting impacts on the efficiency of forest sink enhancement options. It is highly unlikely that all direct and indirect welfare consequences of a proposed afforestation program can be quantified or monetized; therefore an extensive and complicated social cost-benefit type of analysis would be required under the welfare economic approach (Jepma et al., 1996). The outcomes of the engineering efficiency and welfare economic assessments of the same project would not be expected to coincide. "The costs of land in monetary terms may not fully reflect the land-use opportunity costs in welfare terms, because in the former no full account is taken of indirect effects, nonmaterial consequences, distributional impacts, and externalities" (Jepma et al., 1996, p.235). A s Jepma et al. point out, the welfare economic approach implies an assessment based on a general equilibrium model, a highly complex and rarely pursued exercise. To make the exercise more manageable, response options can be evaluated on the basis of 'important' opportunity costs and externalities. The determination of what is important can become rather subjective, however. Those opportunity costs and externalities that are measurable, and that are most sensitive to the results of the analysis, would be of greater importance. Jepma et al. (1996) found that the engineering efficiency approach was the most common used in studies that estimate the costs of afforestation or halting or slowing down of deforestation. This approach is also subject to uncertainties such as the availability of land area, carbon uptake rates, and costs of establishment and maintenance of forestry measures. 49 The approach to estimating the cost of establishing carbon-sequestering tree plantations in this study is similar to that described by Jepma et al. (1996). The approach first involves the estimation of cost functions. In this study, the cost of leasing land is represented by a function derived from a survey of landowners. The next step is to refine the point estimates for plantation establishment and maintenance costs by taking into account local or regional site considerations (e.g., climate, soil zone, transportation distance). For example, Marland and Marland (1992) showed that the best afforestation scenario, based only on carbon flows, will depend on site-specific characteristics such as the expected growth-rate and the accessibility for and efficiency of harvest (MacLaren, 1999). Finally, a discounting procedure is built into the methodology to account for the time value of money and reduced G H G emissions. 5.1.3 The Significance of Time in Carbon Mitigation Studies Due to the growth characteristics of trees, carbon gain is low in the initial years of afforestation. This means that afforestation and reforestation programs will not contribute a great deal to carbon-dioxide reductions during the first commitment period, 2008 - 2012 (IPCC, 2000). However, over time and with successively more land being afforested each year, the benefits will continue to increase for decades (Nilsson and Schopfhauser, 1995). The results of a study by van Kooten (2000) highlight the influence of discount rates on the time value of carbon (C) and the marginal value of afforestation. The time value of carbon depends on the path of marginal damages (Richards, 1997), in other words, the value of reducing G H G emissions today depends on when the damages being avoided are expected to occur and how much they will cost. O f course, this inherently includes the discounting of financial costs. 50 "Given uncertainty over the relationship between atmospheric C 0 2 concentrations and global climate change, and between climate change and economic damages, we have no a priori reason not to discount future C fluxes. When physical C is not discounted, it does not matter when (and thus if) C is sequestered" (van Kooten et al., 1999b, p.3). Whereas the financial discount rate reflects the opportunity cost of money, the carbon discount rate is a function of the marginal rate of substitution between damages from emissions now versus damages from emissions later (Marland et al., 1997). "The carbon discount factor will depend on the time path of stock accumulation and decay of C in the atmosphere, the relation between C stocks and subsequent climate changes, and the marginal damages (in lost utility) of climate change given changing income and adaptation technology. The carbon discount rate is the time derivative of the carbon discount factor" (Marland et al., 1997, p.223). The carbon discount rate used in the following case study is relatively arbitrary due to the level of scientific and economic uncertainty related to climate change (as described above). The rates used in the literature range from 0 to 6%. 5.2 METHODOLOGY The study area is assumed to be a circle, with a radius of 200 km from the ethanol facility at its centre (Figure 9). It is divided into four zones (d) of 50 km (d = 1,2,3,4). Planning periods (t) are four years in length, with decisions made at the beginning of each period. The planning horizon is 36 years (t = 1,2,3,.. .,9), allowing for a reasonable life span for the facility and the potential for multiple tree-crop rotations. The age (a) of an area of afforested plantation is measured from the year of its establishment and measured in planning periods (a = 1,2,3,.. .,9). 51 Figure 9. Representation of study area divided into zones. The decision variables are: Pdt = the area (in hectares) of marginal farmland in zone d afforested (i.e. planted) in period t; Hailt = the area (in hectares) of afforested plantation of age a, in zone d, harvested in period t; WA, = oven-dry weight (in metric tonnes) of wood waste acquired in period t from mills in zone d. 5.2.1 Objective Function The objective is to maximize net present value (NPV) to the owner of the ethanol production facility by allocating the area of marginal farmland for afforestation in each zone in 52 each period, and allocating the area of harvest in each zone in each period. The objective function, NPV, is equal to the difference between the discounted value of revenues {PVR) and the discounted value of costs (PVC) resulting from the land-use decisions over the planning horizon: (5.1) NPV = PVR-PVC Revenue The total present value of revenue (PVR) from sales of ethanol over the planning horizon is a function of the yield from harvesting and wood waste acquisitions. In each period, the yield from harvest is a product of the area (ha) of plantation in each age class harvested in each zone (// a ( l t) and the volume of biomass (m3) per hectare produced in each age class (Ya). The merchantable yield (m3/ha) for hybrid poplar is calculated using the Chapman-Richards function: yield = 329 x (1 - e - ° l 5 6 ( a x | , ) ) 3 (van Kooten et al., 1999a). The length of the planning period (p) is equal to 4 years. The estimates of total aboveground biomass (i.e. including tree-tops and branches) are derived from the merchantable yield function described above and an expansion factor of 1.454 (Nawitka, 1999; van Kooten et al., 1999b). It is assumed that 100% of the aboveground biomass is removed during harvesting. This assumption is made due to the significant variation in harvesting efficiency depending on harvesting systems. The quantity per hectare of aboveground biomass, Ba (Table 10), is estimated by first converting the volume measure to weight (0.361 ODt/m 3 for western hardwoods such as hybrid poplar). Then a conversion factor of 350 litres of ethanol per ODt ( M c C l o y and O'Connor, 1999), and the market value of ethanol per litre ($/l) are applied to obtain the value of E(l. Estimates of the wood-ethanol conversion rate (litres of ethanol per ODt biomass) range from 242 1/ODt in a study of 53 Oregon cellulose-ethanol potential (Graf and Koehler, 2000) to 450 1/ODt (M c Cloy and O'Connor, 1999). The market value of the ethanol used in this scenario is $0.40/litre (Cdn) based on a study by M c C l o y and O'Connor (1999). Wood waste acquisitions (Wilt) are converted to their ethanol value equivalent by multiplying by $140.00/ODt (= 350 1/ODt x $0.40/1). The total revenue at the end of each period is discounted at a rate (r) equal to 6% per year from the middle of each period (dropping units of measurement for convenience):" 9 4 9 (//„.„,, x £ „ ) + K , x HO) (5.2) P r a = Z Z Z n . . V ^ « ) 1=1 </=! a=\ (1 + r ) ' Table 10. Values of Ya, 5 a , and Ea used in modeled scenarios. a Ya Ba Ea (period) (m3/ha) (ODt/ha) ($/ha)a 1 47.85 17.27 6,445.06 2 173.33 62.57 23,346.84 3 289.84 104.63 39,038.91 4 369.57 133.42 49,778.66 5 417.75 150.81 56,268.11 6 445.21 160.72 59,966.27 7 460.40 166.21 62,012.78 8 468.68 169.20 63,128.20 9 473.16 170.81 63,731.30 a Based on $0.40/litre o f ethanol x 350 I o f ethanol/ODt per hectare Costs The total present value of costs (PVC) associated with the afforestation and ethanol production over the planning horizon is calculated as the sum of the following present-value M c C l o y and O'Connor (1999) use the discount rate o f 6% in a study of the wood-ethanol industry. The discount rate in an analysis such as this should reflect the producers value of future returns, which includes risk. 54 costs: capital cost (CC), operating cost (OC); transportation cost (TC); plantation cost (PC); land rental cost (LC); and harvesting cost (HC): (5.3) PVC = CC + OC+ TC + PC + LC + HC The cost of acquiring wood waste is assumed to be zero. Linear representations of each of the capital, operating, and land rental cost functions require two separate equations to describe their behaviour adequately. Although non-linear equations may better describe the behaviour of the systems, the functions have been estimated by piece-wise linear functions in order to permit the use of linear or quadratic programming. Capital cost ( C Q is assumed to be incurred in the middle of the first period (t = 1). This assumes that it will take two years to. build the facility and have it capable of operating at full capacity. C C is a function of the average annual feedstock supply (in oven-dry tonnes, ODt) to the facility. 1 (5.4a) C C , = h x a=\ ,1=1 i = \ px9 (\ + r)2 £££KUXB>^, i (5.4b) C C 2 = [b2 x a=\ d=\ l=\ px9 X (1 + r) 2 55 The coefficients b„ b2 and c are set at 4437.5/ODt, $175.0/ODt and $87,500,000, respectively, to estimate the scaleable capital costs of a ligno-cellulosic ethanol production facility. The two linear functions (5.4a and 5.4b) are derived from the combination of a non-linear function developed by Van Dyne et al. (1998) and linear estimations from IVTCloy and O'Connor (1999) (Figure 10). Ba represents the oven-dry weight per hectare (ODt/ha) of plantation biomass of age a (Table 10). The denominators of functions (5.4a) and (5.4b) convert the total supply over the planning horizon to an annual average (4 years/period x 9 periods). Capacity (ODt/yr) Figure 10. Chart of linear capital cost functions as described in (5.4a) and (5.4b). The source of the operating cost ( O Q functions is the same as that of the capital cost functions. A non-linear form for OC was estimated by two linear functions (5.5a and 5.5b), employing parameters derived from Van Dyne et al. (1998) and M c C l o y and O'Connor (1999) 56 (Figure 11). The coefficients (f„ f2 and g) in 5.5a and 5.5b are set at $94.5/ODt, $37.8/ODt, and $18,900,000, respectively. The operating cost is discounted from the end of each period. 9 4 (5.5a) OC. =± /, (=1 365x p (1 + r r 9 4 (5.5b) oc2=± /2 « = 1 </=! ( = 1 365 x p 180,000,000 160,000,000 . Fq 5 fia 140,000,000 . Eq.5.5b 120,000,000 -nnn-linpar 100,000,000 -tf) o 80,000,000 . • o 60,000,000 . 40,000,000 - " " " " 20,000,000 . 0 •< c 500,000 1,000,000 1,500,000 2,000,000 Capacity (ODt/yr) Figure 11. Chart of linear operating cost functions as described in Equations (5.5a) and (5.5b). 57 The cost of renting the land required for afforestation (LC) is a function of A a t l t , the area (ha) available for harvest (planted and not yet harvested) of age a, in zone d, in period t: (5.6) A^, =Pa^-Y£H^ i i where P a d t is the toal area of age a in zone d planted until period t. This measure keeps track of the total amount of land afforested in any given period and is therefore useful in ensuring that land availability constraints are not exceeded. The land rental cost ($/ha) is based on results from the survey of western Canadian landowners (Suchanek et al., 2001) discussed in Chapter 3. It is a measure of the landowners' willingness to accept tree planting on their land. The economic analysis of the survey results incorporates opportunity costs of agricultural production and the landowners' preferences that affect the amount of compensation they would require in order to allow trees to be planted on their land for a period of time. Suchanek et al. (2001) estimate that the minimum an individual landowner would accept is $40 per acre ($98.84/ha). At this annual rate the average landowner was willing to offer 143 acres (57.87 ha). A n assumption made in this study is that any compensation less than $98.84/ha ($342.49/ha per 4-year period) would return zero hectares. When restated for the purposes of this model, the cost of establishing plantations on any area less than 57.87 hectares per landowner will be at least $342.49/ha per period. If the area required for afforestation is greater than 57.87 hectares per landowner, the cost per hectare in each period follows the landowners' willingness-to-accept function described in Equation (5.7b). In order to avoid non-linearity in the program, the land rental rate has been divided into two segments, one constant and the other linearly increasing, thereby necessitating at least two 58 scenarios of the model. The split occurs at 57.87 hectares per landowner; below this point the rate is constant at $342.49 per hectare (5.7a), and above this point the rate follows the function L C 2 (5.7b) (Figure 12). The coefficient 9.76 in (5.7b) is expressed in units of $ per farmer per hectare2. These forms of the L C , and L C 2 functions imply the combination of a linear and quadratic form of the land rental function. (5.7a) I C , = £ 4 9 Z 2 X „ , , x 395.36 1 (177) < 342.49 (5.7b) f ( 9 4 A 4 9 I x 9.76 x a= 1 tl = \ -170 d=\ « = l Id ) J 1 (i77) > 342.49 2 0 0 0 ] | 1 8 0 0 -</> 1 6 0 0 -a> 1 4 0 0 -5 1 2 0 0 -« 1 0 0 0 -g 8 0 0 -^ 6 0 0 -5 4 0 0 .„ c 2 0 0 -< 0 , , , , , , , , 0 2 5 5 0 7 5 1 0 0 1 2 5 1 5 0 1 7 5 2 0 0 Area per L a n d o w n e r (ha) Figure 12. Graphical representation of the Land Rental Rate function. In order to determine the number of landowners within each zone, an estimate of the average farm size is used in conjunction with the total area of farmland in each zone. Within 59 each zone, the area (in hectares) of farmland physically available for afforestation e {424500, 459750, 561000, 135000}) is based on planimetric measurements of improved land (Alberta Provincial Base Map, 1984).12 The total area of improved land measured in all four zones (2,107,000 ha) is very similar to the estimate (2,126,485 ha) obtained from Statistics Canada (personal communication, 2001). The Statistics Canada estimate was not used in this study as the planimetric map measurements provided more accurate estimates of area by zone. Statistics Canada estimates that there are 5,650 farms within the study area. The number of landowners per zone (qd e {1138, 1233, 1504, 362}) is estimated based on the average size of farms in the study area (= total area / total number of farms) and the available area in each zone. The area of land afforested in each zone is assumed to be equally distributed amongst the total number of eligible landowners in the respective zone. As the survey (Suchanek et al., 2001) indicated that only 75% of landowners would consider planting trees, the measured area was reduced by 25% per zone (given a fixed, average farm size) to give the available area per zone (Z,,). The total land rental costs per period are discounted from the beginning of each period. Transportation cost (TC) is a function of the weight of harvested biomass (ODt) delivered from each zone to the facility in each period (Equation 5.8). The transportation rates (TFA, d = 1,2,3,4) in $/ODt for each zone are 4.2, 12.6, 21.0, 29.4, respectively. The cost per tonne of biomass (chipped trees and industrial wood waste) delivered to the facility is a function of an average distance to each zone. For example, the transportation rate is TF, = $4.20 per tonne to transport material from 25km away. The cost includes the round-trip distance, assuming that the 1 2 In the Peace River region of Alberta, "improved land" is predominantly comprised of agricultural land. 60 outward-bound trucks are empty. The figures used in this model are based on 1999 Saskatchewan Wheat Pool rates, and are similar to transportation costs used in related studies (Intergroup, 1982; Lindenbach, 2000). Transportation costs are discounted from the middle of each period. (5.8) 7T = X 4 f ff £ TFdx YH ,,xB + w. j l (l + r ) /x(,,/2) Plantation cost (PC) is made up of establishment and maintenance costs including site preparation, planting, brush and weed control, and operating overhead. The per hectare cost (m) is estimated at $1575.50/ha in this scenario.'3 As most of this cost is incurred at the beginning of the period of establishment, the plantation cost in each period is discounted from the middle of each period: (5.9) P C = £ S P., xm lx(p/2) Harvesting cost (HC) is a product of the area harvested (HlM), the yield (m3/ha) in each age class (Ya), and the harvesting cost per cubic metre (v). It is given by Equation 5.10. The values of Ya used in this scenario are presented in Table 10. The value of v is set at $12.00/m3 in 1 3 The components o f this cost include site preparation estimated at $180/ha, planting at $950/ha, herbicide application at $240/ha (includes multiple applications), and a 15% operating overhead. 61 this scenario (van Kooten et al., 1999b; Samson et al., 1999; Girouard et al., 1995; Intergroup, 1982). Harvesting costs are discounted from the middle of each period. (5.10) M 7 = £ 5.2.2 Constraints The model includes a number of constraints. The physical constraints include the area of land and wood waste available. Additional constraints are required to ensure that harvested areas do no exceed planted area and that very young plantations are not harvested. A n even-flow requirement is also included to provide a stable supply of ethanol and a measure of employment stability for the local community. The sum of areas under plantation in any period, in each zone, must not exceed the total area available in each zone (Lu): (5.11) Y^Au^,<Llt As mentioned previously in reference to land rental cost, the rental rate for afforested land is a function of the area required by the ethanol producer and the number of eligible landowners (qrf). A different constraint on the rental rate function is required for the two linear models. The 9 4 a=l tl=\ 1 (1 + r) /x(/,/2) 62 constraints corresponding to land rental cost functions, Equations 5.7a and 5.7b, are presented in Equations 5.12a and 5.12b, respectively. (5.12a) 0 < ^ 4 , A , < 57 .87x^ (5.12b) 2 X , , , > 57.87x9,, The right-hand side of the constraints, Equations 5.13a and 5.13b, required for the linear capital cost ( C Q and operating cost ( O Q functions, is estimated from the point where the two functions intersect, at 350,000 ODt/year (see Figure 10 and Figure 11). 9 9 4 (5.13a) '"' " = l t / = l < 350,000 9x p 9 9 4 (5.13b) " = I J = I > 350,000 9x p Harvesting constraints are set up to ensure that the harvested area in each period is not greater than the area of plantations physically available (Equation 5.14). Harvesting is also restricted to plantations older than 4 years (a > 1) (Equation 5.15). 63 (5.14) Ha^< A a 4 t (5.15) Ha__]d^0 The amount of waste produced by a mill is a function of the amount and size distribution of roundwood entering the mill and the process efficiency of the mill. In this model, the maximum available levels are assumed to be fixed over the planning horizon and are based on current sawmill operations in the study area (Table 11). It is also assumed that there are no competing uses for the wood waste and it has zero value to the mill. O f course, if the wood waste constraint is effective, wood waste becomes valuable to the forestry operation or mill owner. It is then possible to determine its shadow prices. The limited availability of wood waste constrains the amount acquired in each zone in each period: (5.16) WtlJ<Rd Table 11. Total mill residues by distance zone (Rd). Source: AFPA(2001) . Distance Zone (d) Number of mills Total Quantity of Residue per Year (tonnes) Produced Available (R„) 1 4 598536 197644 2 0 0 0 3 7 243592 224678 4 2 3975083 87023a "Includes 65,000 ODt of wood waste from mills in British Columbia within a 2-hour drive (M c Cloy & O'Connor, 1999). 64 As the size of the facility is determined by the average feedstock supply over its lifetime, a relatively stable flow of biomass must be maintained in order to avoid significant feedstock excess or deficit. The periodic supply (in ODt/period) is constrained within a limit of 25% above and below the average periodic supply over the planning horizon: (5.17a) Z Z ( ^ M , x 5 f l ) + ^ r f ^ 0 . 7 5 x =1 ti=\ 9 4 a=\ d=\ l=\ (5.17b) ^ ( h ^ X B ^ + W , , , <\.25x «=1 d=\ 9 4 Y.TL(H^xBa)+WdJ a=\ d=\ / = ! Finally, all decision variables are constrained to be non-negative: (5.18) P„-Htld/,Wtl> 0 5.3 RESULTS The model was constructed in Microsoft Excel and solved using the Premium Solver Platform. To solve the model, four scenarios are run using different combinations of the two land cost functions (Equations 5.7a and 5.7b) and the two sets of capital and operating cost functions (Equations 5.4a & 5.5a and 5.4b & 5.5b) in the objective function. The behaviour of 65 each scenario, and the sensitivities of the solutions, are examined before drawing conclusions from the results. Only two of the four scenarios need to be examined further as the other two resulted in negative net present values. The two scenarios that used the quadratic land rental function resulted in net present losses due to a combination of a minimum level of afforestation and a rapidly increasing cost function (Figure 12). The model that represented the lower capital cost and operating cost functions and associated constraint (Scenario-B) resulted in an optimal solution that represents a net present loss of $433 million. Scenario-C represents the large-scale facility and has a lower bound on the biomass supply. When an attempt is made to solve this quadratic model given these conditions no feasible solution is found. This is mainly due to the combination of the steep slope of the land rental function and the constraints which prevent a positive solution. As a result, Scenario-A (a small-scale facility) and Scenario-D (a large-scale facility) are the two scenarios that represent operations with positive economic potential and therefore warrant further investigation. 5.3.1 Mode l Descriptions A linear programming model under Scenario-A considers a small-scale ethanol production facility capable of processing a maximum of 350,000 ODt of biomass per year (see Equation 5.13a). Based on current grain-fed plants, this size of facility is about average and would be near the upper limit for the industrial application of a new technology such as enzymatic hydrolysis. The supply of biomass from afforested land is limited to less than 57.87 hectares per farmer (see Equation 5.12a). This portion of the land rental rate is constant at $342.49/ha (see Equation 5.7a) and the corresponding land rental function is linear. 66 Scenario-D involves a large-scale ethanol production facility supplied with biomass from wood waste and from a limited area of afforested farmland. The scale of the ethanol production facility is constrained to have a capacity of at least 350,000 ODt per year. The capital and operating cost functions for a facility of this size follow a shallower slope (Equation 5.13b) than for the smaller facility represented by Scenario-A. The land rental cost function and associated constraints used in a model under Scenario-A are employed under Scenario-D. 5.3.2 Optimal Allocation of Resources Small-Scale Facility The optimum allocation of resources for the small-scale facility (Scenario-A) results in a discounted net present value of $244.8 million. The scale of the production facility in this case requires a capital investment of $136.3 million for the capacity to produce 122 million litres of ethanol annually. This scale of operation requires all of the available wood waste from all zones in every period. It also requires that a total of 65,785 hectares be planted over 24 years. Planting begins in the first period and harvesting begins in the second period (Table 13). The first plantation is not completely harvested until the fourth period, when the marginal yield is highest. Harvesting in younger age classes is required to offset the initial investment and maintain an even-flow of biomass later in the planning horizon. Planting continues at a constant rate (9,299 ha per period) until period 7 when no further planting is carried out because the time required to grow sufficiently valuable volume exceeds the planning horizon. A l l trees are harvested by the end of the last period. A l l available wood waste in the study area is utilised by the ethanol production facility. The demand for wood waste is higher than for afforested biomass simply because the only cost 67 directly associated with wood waste in this model is transportation; therefore it is obviously a less costly feedstock. The shadow price of wood waste depends on the zone and period in which it is available. The shadow prices of wood waste show that the ethanol producer would be willing to pay between $7 and $50 per additional oven-dry tonne (ODt) of wood waste, depending on zone and period (Table 12). For example, the ethanol producer could pay up to $33.03 for one additional ODt from zone 1 in period 1 without reducing the net present value of the business. Table 12. Shadow prices of wood waste associated with Scenario-A in $/ODt Zone d 1 2 3 4 Period t 5 6 7 8 9 1 33.03 2 25.55 3 18.08 4 10.60 50.23 43.58 36.92 30.27 37.35 31.43 25.51 19.59 32.26 26.99 21.72 16.45 27.98 23.29 18.60 13.91 24.32 20.15 15.97 11.80 21.19 17.47 13.76 10.04 18.49 15.19 11.88 8.57 16.17 13.23 10.28 7.34 Table 13. Summary of results of Scenario-A by pi anning period. Period 1 2 3 4 5 6 7 8 9 TOTALS NPV ($x 106) (10.8) 25.6 29.5 38.7 59.1 58.2 63.4 60.5 56.9 381.1 NPC ($x 106) 79.0 104.1 85.9 83.3 95.2 95.2 58.8 48.3 39.9 673.8 NPR ($x 106) 68.2 129.7 115.4 122.0 154.3 154.3 122.3 108.8 96.8 1054.9 Total Area harvested (ha) 0 8,640 5,167 5,481 9,299 9,299 9,299 9,299 9,299 65,785 Total Area planted 19,289 9,299 9,299 9,299 9,299 9,299 0 0 0 65,785 Total biomass delivered (ODkt) 509 1,050 1,050 1,241 1,750 1,750 1,750 1,750 1,750 12,600 Feedstock Cost ($/ODt) 80 40 35 30 25 22 15 13 11 68 Large-scale Facility The optimum allocation of resources for the large-scale facility (Scenario-D) results in a discounted net present value of $2,150 million, the optimal solution of the model. The scale of the production facility in this case requires a capital investment of $401 million for the capacity to produce 726 million litres of ethanol annually. This scale of operation requires that a total of 722,782 hectares be planted over 32 years, and requires all of the available wood waste from all zones in every period. The shadow price for wood waste ranges from $19 to $102 depending on zone and period (Table 14). Table 14. Shadow prices of wood waste associated with Scenario-D in $/ODt Zone d 1 2 3 4 Period t 5 6 7 8 9 1 98.64 102.19 82.63 75.28 68.28 61.74 55.36 41.07 28.03 2 91.16 95.54 76.71 70.01 63.59 57.57 51.65 37.76 25.08 3 83.68 88.89 70.78 64.74 58.90 53.39 47.93 34.45 22.14 4 76.21 82.23 64.86 59.47 54.21 49.22 44.22 31.15 19.20 A l l available land, as constrained by the land rental function (Equation 5.12a), is planted in the first period. The plantations in zone 1 occupy the maximum area over the whole planning horizon. The plantations in zones 2 and 3 maintain their coverage until period 9, and in zone 4 no re-planting is done in periods 8 or 9 (Table 15). Harvesting begins in period 2 in zone 3, and continues until the end of the planning horizon. The majority of the harvesting occurs in 12 to 16 69 year-old plantations when the marginal yield is highest. A l l plantations are completely harvested by the end of the planning horizon. Table 15. Summary of results of Scenario-D by planning period. Period 1 2 3 4 5 6 7 8 9 TOTALS NPV ($ x 106) (409.8) 308.2 515.9 313.5 400.8 299.6 367.2 320.0 383.6 2498.9 NPC ($x 10*6) 478.0 481.2 657.5 537.9 500.9 332.4 369.0 335.3 199.5 3891.6 NPR ($x 106) 68.2 789.3 1173.4 851.4 901.7 631.9 736.2 655.2 583.1 6390.5 Total Area harvested (ha) 0 91,308 94,255 59,608 91,308 70,934 91,141 101,756 122,471 722,782 Total Area planted 245,172 0 91,308 94,255 59,608 91,308 70,934 70,196 0 722,782 Total biomass delivered (ODkt) 509 6,223 10,371 8,462 10,063 7,931 10,313 10,371 10,371 74,674 Feedstock Cost ($/ODt) 854 50 43 47 37 31 28 26 14 In comparison to the small-scale facility, the greater demand for feedstock for the large-scale facility of Scenario-D is reflected in the increase in afforestation, as described above, as well as the shadow price of wood waste. As in the small-scale facility, the shadow price of wood waste depends on the zone and period in which it is available. In the large-scale facility, demand is greater and the ethanol producer would pay between $19 and $102 per additional oven-dry tonne (ODt) of wood waste, depending on zone and period. For example, the ethanol producer could pay up to $98.63 for one additional ODt from zone 1 in period 1 without reducing the net present value of the business. 70 5.4 SENSITIVITY ANAL YSIS 5.4.1 Sensitivity to Elements of Objective Function As discussed previously, the land rental cost is key factor in the NPV of the model. The land rental cost is a result of planting and harvesting decisions and therefore is integrated into their respective coefficients in the objective function. In Scenario-A, the objective function value (NPV) is sensitive to the costs related to planting and harvesting in zone 1. A decrease in planting cost of $27.15/ha in the first planning period, or an increase of $21.29 in the second period, can result in a change in the amount of land afforested. The model is most sensitive to costs and revenues associated with harvesting 12 to 16 year-old plantations due to the growth function. For example, the harvesting schedule could change if harvesting costs in the 4 t h planning period were to decrease by only $2.16/ha, from $l,369.13/ha (= $12/ha x 289.84m7ha x (1+r)"'6) for 12 year-old plantations. A relatively small change in plantation yield or transport cost could cause a similar change in the solution. The sensitivity oi Scenario-A to these variables declines with increasing distance from the facility and over time. Scenario-D is less sensitive to planting costs due to the high demand for biomass. The sensitivity of the model to planting costs increases significantly after period 2; however the effect would be seen mostly in the timing of planting rather than a change in the area planted. A minimum change of approximately $2/ha could change the planting decisions over the planning horizon. Transportation cost has a significant effect on the area chosen for afforestation. In Scenario-A no planting occurs in zones 2, 3 or 4 due to the increase in transportation cost compared to zone 1. The harvesting decision directly affects the total transportation cost; 71 therefore the value of the objective coefficient of harvesting shows the effect of transportation distance on the revenue per hectare harvested. For example, in Scenario-A, the revenue from harvesting a 12 year-old plantation in period 2, zone 2 is $2423.69/ha, compared to $3119.86/ha in zone 1. In Scenario-A, the revenue per hectare harvested decreases the further from the facility until in zone 4, depending on the yield, revenue can be negative and could reduce the optimal value of the objective function (NPV) by over $1600 per additional hectare harvested. In Scenario-D, transportation costs affect the harvesting costs in the same manner as in Scenario-A. However, because of the lower capital and operating costs (per delivered ODt), the objective coefficients of harvesting are substantially larger. In comparison to the example used above for Scenario-A, the revenue from harvesting a 12 year-old plantation in period 2, zone 2 is $6824.89/ha, compared to $7521.06/ha in zone 1. In contrast to Scenario-A, the revenue from harvesting in Scenario-D is always positive, regardless of yield, zone or period. 5.4.2 Sensitivity to Constraints The requirement for an even-flow of biomass to the facility has a strong influence on the optimal solution. In Scenario-A, the even-flow constraint is binding at the lower limit early in the planning horizon, and then it is binding at its upper limit at the end of the planning horizon. In other words, the profitability of the operation would likely increase if less planting were permitted in the second period, and if more planting were permitted in the last five periods. In Scenario-D, the lower bound of the even-flow constraint has practically no effect on the solution due to the level of demand for feedstock. The upper bound of the constraint does limit the NPV in periods 3, 7, 8 and 9. The maximum NPV of Scenario-D would improve if a greater amount of biomass were available in these periods. 72 I have analyzed each of the original scenarios under three different even-flow assumptions. The original Scenarios [Scenarios-A & D) are constrained to a tolerance of 25% variance from the average biomass supply as set in Equations 5.17a and 5.17b. The even-flow constraints are then relaxed to a tolerance of 50%, and then with the even-flow constraints omitted entirely. Table 16. Summary statistics for both Scenario-A and Scenario-D. A-25% and D-25% represent a 25% even-flow tolerance, A-50% and D-50% have a 50% tolerance, and A -100% and D-100% have no even-flow constraints. Scenario A-25% A-50% A-100% D-25% D-50% D-100% NPV ($x 106) 244.8 254.7 263.9 2,149.9 2,200.3 2,350.0 CC ($x 106) 136.3 136.3 136.3 400.9 407.0 430.7 Biomass Feedrate (ODt/day) 959 959 959 5,683 5,790 6,206 Average Feedstock Cost ($/ODt) 24 23 24 39 38 38 Annual Ethanol Production (Ml) 122.5 122.5 122.5 726.3 739.6 792.8 Ethanol Sale Price ($/l) 0.40 0.40 0.40 0.40 0.40 0.40 Ethanol Yield (l/ODt feed) 350 350 350 350 350 350 The discounted NPV increases when the even-flow constraint is relaxed, even if the capacity of the ethanol facility does not change (Table 16). The cause of this trend in Scenario-A is the decrease in average feedstock costs as the even-flow constraint is relaxed. Planting and harvesting are rescheduled to maximize net returns. In Scenario-D the cause of the trend is the 73 increase in ethanol production. Figure 13 and Figure 14 illustrate the effect of the constraint on planting decisions and the associated effect on NPC and NPR of Scenario-A. Most of the planting and harvesting decisions made in Scenario-A are directly affected by the even-flow constraint. In Scenario-A, with a 25% even-flow constraint, while it would be profitable to plant more in later periods, the even-flow constraint on the amount of biomass delivered to the facility results in a compromise solution. Some of the area planted early in the planning horizon is harvested at a relatively small loss in order to take advantage of increased revenues from later, larger harvests. Without the even-flow constraints, all planting in Scenario-A would occur in period 5 and be harvested in period 8. With Scenario-D, the planted area is almost always maximized so there is no room for increase, only a re-scheduling of planting (Figure 15 and Figure 16). Without the even-flow constraints in Scenario-D, planting and harvesting activities would occur every 12 years, corresponding to the optimal economic rotation length for hybrid poplar in this case. While the relaxation of the even-flow constraint increases N P V over the planning horizon, the periodic fluctuations become quite significant, resulting in periodic losses during planting years. Despite a lower NPV, the ethanol producer may prefer to have an even-flow constraint to maintain cash-flows or consider acquiring feedstock from somewhere else. 74 ro 2 30,000 2 3 4 5 6 Planning Period (4 years) Figure 13. Effect of even-flow constraints on planting decisions over the planning horizon of Scenario-A (the small-scale facility). A l represents a 25% even-flow tolerance, A2 has a 50% tolerance, and A3 has no even-flow constraints. 6.00E+08 5.00E+08 "? 4.00E+08 *-> c O 3.00E+08 o jo 2.00E+08 1.00E+08 0.00E+00 — • — -A1 •NPC . A1 -NPR —A— A2 -NPC - - - A- -. A2 -NPR — • — A3 -NPC ------ . A3 -NPR 3 4 5 6 7 Planning Period (4 years) 8 9 Figure 14. Effect of even-flow constraints on discounted costs (NPC) and revenues (NPR) over the planning horizon of Scenario-A. A l represents a 25% even-flow tolerance, A2 has a 50%o tolerance, and A3 has no even-flow constraints. 75 CO 300,000 250,000 200,000 150,000 4 ra ^ 100,000 50,000 3 4 5 6 Planning Period (4 years) Figure 15. Effect of even-flow constraints on planting decisions over the planning horizon of Scenario-D (the large-scale facility). DI represents a 25% even-flow tolerance, D2 has a 50% tolerance, and D3 has no even-flow constraints. 0) -*—' C 3 O u m 3.00E+09 2.50E+09 2.00E+09 1.50E-+O9 1.00E+09 5.00E+08 0.00E+00 . D1-NPC D1-NPR . D2-NPC D2-NPR . D3-NPC D3-NPR 3 4 5 6 7 Planning Period (4 years) Figure 16. Effect of even-flow constraints on discounted costs (NPC) and revenues (NPR) over the planning horizon of Scenario-D. DI represents a 25% even-flow tolerance, D2 has a 50% tolerance, and D3 has no even-flow constraints. 76 The even-flow constraint also affects the shadow price of wood waste. The shadow price of the wood waste in Scenario-A is higher in period 2 than in period 1 due to the influence of the even-flow constraints (Equations 5.17a and 5.17b) that do not directly constrain the amount of biomass acquired in period 1. In Scenario-D the demand for biomass to feed the larger facility exceeds the lower limit of the even-flow constraint; therefore the shadow price is highest in period 1. For both scenarios, AJ and DI, the shadow price of wood waste decrease over time and with increasing transport distance (Figure 17a and 17b). Q O 60.00 50.00 40.00 0) •= 30.00 Q. % 20.00 ra C O 1 0 . 0 0 -r o.oo Zone 1 Zone 2 Zone 3 Zone 4 4 5 6 7 Planning Period 8 Figure 17a. Graph representing the effect of time and transport distance on the shadow price of wood waste in Scenario-A, the small-scale facility with 25% even-flow constraints. 77 120.00 100.00 § 80.00 3 60.00 Q. | 40.00 re .c OT 20.00 0.00 .Zone 1 Zone 2 Zone 3 Zone 4 4 5 6 Planning Period 8 Figure 17b. Graph representing the effect of time and transport distance on the shadow price of wood waste in Scenario-D, the larger-scale facility with 25% even-flow constraints. In Scenario-A, the constraint associated with the smaller scale capital and operating cost functions (Equations 5.4a and 5.5a respectively), limiting periodic capacity to 350,000 ODt, is binding on the objective function. Increasing the capacity of the facility would increase NPVby $134 per ODt (up to approximately 540,000 ODt/period). The upper limit of afforestable area in Scenario-A, due to land availability (L d) and the land rental cost function (LC,) , is not binding on the solution. The maximum area under plantation at any time is 37,196 hectares in zone 1. The upper limit on the area in zone 1 is set by the land rental cost constraint (Equation 5.12a) at 65,860 hectares. As no planting occurs in the other zones, in this scenario there remains a total of 208,025 hectares available for afforestation in the study area. 78 In Scenario-D, the upper limit of afforestable area, as defined by the land rental cost function (LC,) , limits the value of the objective function (NPV) in every zone. The shadow prices associated with this land constraint vary with zone and period, generally decreasing over time and with increasing distance from the facility. For example, the shadow price for one additional hectare of afforested land in zone 1, period 2, is $1,965.56. 5.5 ANALYSIS OF RESULTS IN TERMS OF GHG EMISSIONS REDUCTION The results obtained from the maximization of the objective function, for both Scenario-A and Scenario-D, are now examined in terms of greenhouse gas (GHG) emissions. A l l G H G emissions are presented in terms of C 0 2 or C0 2-equivalents, and a physical carbon discount rate of 2.5% is used (see Stennes, 2000; van Kooten et al., 1999b; Price, 1997; Marland et al., 1997).14 The G H G pools examined in this case study include soil organic carbon (SOC), belowground carbon, aboveground carbon, emissions from plantation operations, and emissions avoided through the use of ethanol in gasoline. The conversion coefficients used in the analysis are presented in Table 17. 1 4 The discounting of GHG emissions remains a contentious issue. The view taken in this thesis is that given regulatory requirements (i.e. a ratified Kyoto Protocol) and the emergence of carbon markets, a discount rate (although arguably artificial) is appropriate. However, a precise rate applicable to this situation is unknown and therefore the rate of 2.5% is chosen arbitrarily, but based on previous studies. 79 Table 17. Conversion rates used in the case study of a hypothetical wood-ethanol production facility. Inside bole (merchantable) volume to aboveground biomass expansion factor: Wood volume to dry-matter biomass (ODt / m3): Ethanol per unit biomass (litres / ODt): Avoided emission from ethanol use (tC02-equivalent / litre): Soil organic carbon gain relative to baseline agricultural soil (tC ha'yf'): C0 2-equivalent emitted in planting operation (tC0 2/ha): Electricity/Steam from lignin co-product (kWh / litre): Physical carbon to C0 2-equivalent conversion factor ( tC0 2 / tC): 0.00269 0.2667 3.6667 0.361 350.0 1.454 0.2 1.1 The amount of soil organic carbon and belowground carbon sequestered depends on the amount of land afforested and the length of time that trees continue growing. The condition and use of the land prior to afforestation is also relevant and in this case the land available is assumed to be improved pasture and "other" land (meaning other than crops, fallow, or natural pasture). The size of the aboveground carbon pool (e.g. hybrid poplar trees) in one period, takes into account the amount present in the previous period. Therefore, if physical carbon is not discounted and all of the trees that are planted are harvested before the end of the planning horizon, there is no net gain of aboveground carbon. The emissions from planting operations are solely a function of the area planted in each period. The emissions avoided through the use of ethanol in gasoline are measured in terms of C 0 2 equivalents and are based on the periodic ethanol production from the facility. Scenario-A results in a net discounted emissions reduction of 8.5 Mt-C0 2-equivalents over 36 years. The costs incurred amount to $98/t-C0 2. The larger scale of operation 80 represented in Scenario-D results in a net discounted emissions reduction of 51.2 M t - C O equivalents, at a rate of $84/t-C0 2 (Table 18). Table 18. Summary of G H G emissions statistics of Scenario-A and of Scenario-D. A-25% and D-25%) represent a 25% even-flow tolerance, A-50% and D-50% have a 50% tolerance, and A-100% and D-100% have no even-flow constraints. Scenario A-25% A-50% A-100% D-25% D-50% D-100% NPV ($x 106) 244.8 254.7 263.9 2149.9 2200.3 2350.0 Annual Ethanol Production (L x 106) 122.5 122.5 122.5 726.3 739.6 792.8 Total Disc. Cost / Disc. tC0 2 77 75 74 83 81 82 Total Disc. Cost/Non-Disc. tC0 2 52 49 50 56 55 54 Initial Aboveground C 0 2 max.(kt-C0 2 ) a • 905 2,013 4,990 9,887 13,546 20,362 Net disc, balance (kt-C02)b 8,467 8,258 8,567 51,137 52,080 54,409 Net non-disc, balance (kt-C0 2)° 12,551 12,552 12,552 76,164 77,581 83,048 " The maximum amount of aboveground C0 2 sequestered prior to harvesting. If the planning horizon was infinite, and the plantations were managed on a sustainable basis, this would be the equilibrium level of sequestered aboveground C0 2 . b Net quantity of C0 2 sequestered and avoided over the planning horizon, and discounted at a rate of 2.5% per annum. c Net quantity of C0 2 sequestered and avoided, over the planning horizon, but not including discounting of carbon. When the periodic C 0 2 balances are examined, a distinct trend of declining emissions reductions over time is observed. However, the even-flow constraint has a significant effect on the pattern of periodic emissions reductions, as can be seen in Figures 18 and 19, but little effect on net discounted reductions (Figure 20). Within the G H G calculations, the amount of 81 aboveground biomass is very sensitive to the even-flow constraint (Figure 21), which explains most of the variation in periodic emissions reductions.15 O o OJ u c ra ra n fN o o Planning Period (4 years) Figure 18. Effect of even-flow constraints on G H G balance over the planning horizon of Scenario-A. A l represents a 25% even-flow tolerance, A2 has a 50% tolerance, and A3 has no even-flow constraints. 1 5 The maximum amount of aboveground biomass in plantation is a function of the maximum amount of land afforested over the planning horizon. This measure is relevant if one assumes that if demand from the ethanol facility declines, as it does given the limited life-span, a new facility or market will re-create the demand and therefore allow for sustainable production from hybrid poplar plantations. 82 0) u c re O O 20,000 15,000 10,000 5,000 -5,000 -•—D1 - • -_D2 . . . . . D3 Planning Period (4 years) Figure 19. Effect of even-flow constraints on G H G balance over the planning horizon of Scenario-D. DI represents a 25% even-flow tolerance, D2 has a 50% tolerance, and D3 has no even-flow constraints. c re > cr d> (S o 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 • Net disc, balance m Net non-disc, balance A1 A2 A3 D1 D2 Model-Scenario D3 Figure 20. Effect of even-flow constraints on net G H G balance (discounted and non-discounted physical carbon equivalent) at the end of the planning horizon of Scenario-A and Scenario-D. A l and DI represent a 25% even-flow tolerance, A2 and D2 have a 50% tolerance, and A3 and D3 have no even-flow constraints. 83 25,000 20,000 c 0) « 15,000 '5 CT CJ g 10,000 5,000 20,362 13,546 9,887 4,990 905 2,013 A1 A2 A3 D1 Model D2 D3 Figure 21. Effect of even-flow constraints on maximum aboveground C 0 2 sequestration (non-discounted) of Scenario-A and Scenario-D. A l and DI represent a 25% even-flow tolerance, A2 and D2 have a 50% tolerance, and A3 and D3 have no even-flow constraints. If the emissions reductions associated with the ethanol production program modeled here are to be included in the accounting of Canada's G H G emissions, some variation in conversion factors and accounting methods may be required depending on the agreed-upon accounting system in the Kyoto Protocol. However, it is clear that there would be significant net emissions reductions from ethanol production programs as modeled in Scenario-A and Scenario-D. 5.6 CONCLUSIONS The results of the mathematical programs presented here indicate that the land rental rate significantly restricts the economic success of afforestation as a source of biomass for ethanol 84 production. The results also show that a large-scale facility has a higher net present value and contributes substantially more to G H G emissions reductions than a smaller-scale facility. However, there are considerations not accounted for in the model that would make the smaller-scale facility more attractive at this time. Although Scenario-D is more economically efficient than Scenario-A due to economies of scale, it is highly unlikely that the ethanol production system represented by Scenario-D would be built at this time. The enzymatic-hydrolysis technology applied in these Scenarios is presently untested at an industrial scale and this represents the greatest risk for the ethanol producer. In addition, the land rental rate function used is based on a sample of landowners from all western provinces and may not accurately reflect the willingness-to-accept compensation of those in the study area. Another consideration not included in the Scenarios is the limitation of the current transportation infrastructure. A supply of 5,685 tonnes of feedstock per day to the larger-scale facility may not be possible given current road and rail networks. A smaller facility could more easily meet its feedstock requirements using existing transportation networks. The results of the Scenarios are also sensitive to planting and harvesting costs. Increases or decreases in these costs by as little as $2/ha can change the amount of area afforested or the timing of planting and harvesting. Transportation costs also affect these activities, but less so; they are constant within each zone and the difference between zones is significant enough to be unaffected by small changes in cost. The shadow prices of wood waste in Scenario-A indicate that the mills could charge at least $11 per ODt, depending on distance from the ethanol facility, and the ethanol producer would continue to purchase the wood waste. O f significant consideration not included in the 85 model are the characteristics of the available wood waste. Whitewood sawmill residue is much preferred over bark due to their relative ethanol conversion rates. Operating and capital costs are a function of the biomass delivered and therefore are independent of distance. It is not the cost itself but the constraints related to the capital cost, such as capacity and even-flow, which influence the model results. With some government incentives designed to reduce, or compensate for, the risk to the producer, a wood-ethanol production facility located in Grande Prairie, Alberta, of the scale represented in Scenario-A would be an economically viable project in the short term. In the long term, Scenario-D may have a greater chance of success given some maturing of the technology. A greater willingness of landowners to plant trees will likely be necessary in the future to make up for reductions in wood waste supplies due to increased efficiency of mills and increased competition from other users. 86 CHAPTER 6: DISCUSSION The use of wood-biomass from afforested lands and industrial wood waste as a fuel for energy production can be an economically viable tool to reduce greenhouse gas levels in the atmosphere. Afforestation as a supply of wood-biomass for energy production (such as ethanol) has the combined benefit of increasing the terrestrial carbon sink, and offsetting the production and combustion of fossil fuels, thereby reducing net greenhouse gas emissions. As a forestry measure for mitigating carbon-dioxide ( C 0 2 ) emissions through a change in land-use, afforestation has a number of potential benefits. The primary benefit is an increase in the size of the terrestrial carbon sink. Other benefits include the potential increase in wildlife diversity and abundance, reduced soil erosion and improved water quality. Additional benefits, in terms of net greenhouse gas reductions, include the opportunity either to extend the storage life of the carbon by converting the mature trees into wood-products, or to use the wood-biomass as a source of renewable energy and thereby offset the use of fossil fuels. The former option has distinct advantages including established processing facilities and markets; however, a potential market effect of leakage may result in negligible benefits in terms of net greenhouse gas reductions. The use of afforested biomass for renewable energy has the advantage of being emissions-neutral under the Kyoto Protocol, and further reductions are realised through the displacement of fossil fuels. O f the carbon pools associated with the forest carbon sink, soil organic carbon is thought to be the largest pool compared to aboveground and root biomass. Unfortunately, soil carbon is also the most difficult to measure efficiently. It is likely that the measuring system agreed upon 87 for accounting purposes under the Kyoto Protocol will be a compromise between precision and the cost and time required to undertake the measurements. The potential for renewable energy to contribute to mitigating greenhouse gas emissions in Canada will depend largely on changes to the structure of the electricity markets. Pulp mills and sawmills have the capability of generating electricity from their waste products in excess of their own requirements, but in many cases are prevented from doing so due to lack of access to the power grid. In British Columbia, for example, the provincial power providers (BCHydro and BCGas) are essentially monopolies, with profits going directly into the government's general revenue fund. Therefore, from a financial point of view, it would not be in the government's best short-term interest to deregulate the industry. In the long term, however, most consumers would likely benefit from a deregulated industry through cheaper basic rates and a greater choice of providers (including renewable energy providers). This is being demonstrated in Alberta where, despite an economy based on fossil fuel production, a number of renewable energy companies are entering the newly deregulated electricity market. There are currently many millions of tonnes of industrial wood waste being disposed of by incineration or in landfills. The conversion of this waste wood into energy products provides an opportunity to avoid these current C0 2-intensive waste management practices and offset emissions from the production and consumption of fossil fuels. The level of wood waste production is a function of the volume of timber harvested and the efficiencies of the primary processes. Projections of trends in future harvest levels are highly varied depending on province or region. There is generally a downward influence due to reductions in landbase but also an upward influence from favourable yield projections of managed second-growth timber. At the 88 processing end it is reasonable to assume that efficiency will increase, thereby reducing the amount of waste. Also, competition for that waste is likely to increase from the engineered wood products industry, thereby increasing the price. Future reductions in the availability of low-cost wood waste will increase the demand for wood-biomass from fast-growing plantations. The potential for afforestation to meet this demand is limited mainly by the cost and productivity of marginal farmland and its suitability for growing trees. Farmers do not only consider the opportunity cost of agricultural production when considering the option to plant trees. Other factors such as the current condition of their land and the visual appeal of the landscape influence their willingness to plant trees. The productivity of the land initially available for afforestation will be marginal in terms of its potential for agricultural crop production. Trees will grow on this land but their yield will be restricted by nutrient and moisture availability. Management tools such as fertilization and irrigation can increase yields but the value of that increase must be weighed against the extra costs. Increasing demand combined with reductions in the costs associated with plantation establishment and harvesting will result in increased productivity (due to better land and more intensive management) and therefore greater potential to compete with wood waste. Considering the conversion of wood-biomass to energy products, ethanol stands out for three main reasons. First, its use as a transportation fuel derived from renewable wood-biomass reduces the net greenhouse gases emissions through the reduced consumption of gasoline, the increase in the terrestrial carbon sink from bio-energy plantations, and the emissions avoided from incineration and landfilling of wood waste. Second, ethanol has a number of health and environmental benefits, including improving air quality, in addition to mitigating greenhouse gas 89 emissions. The third benefit of ethanol production is that, as it gains a share of the transportation fuel market, the consumer becomes less vulnerable to drastic changes in oil and gas prices (as experienced in the winter of 2000-2001). As discussed in Chapter 4, there are technological obstacles to overcome before wood-ethanol production can begin to realise its market potential. A n increase in federal incentives in the form of tax concessions on ethanol-blended fuel, combined with provincial incentives designed to help overcome the large capital costs and associated risks, will provide the industry with the necessary boost to get it up and running. Canada can gain from a successful ethanol industry not only reduced greenhouse gas emissions, but also an increase in economic activity, particularly in rural areas. Through a quantitative analysis of a hypothetical wood-ethanol production facility, I found that, given the assumptions made in the model formulation described in Chapter 5, positive net present value could be achieved from the point of view of the ethanol producer. The allocation of resources which optimized net present value led to the maximum utilization of available industrial wood residues in the study area due to the significantly lower acquisition costs compared to afforestation and harvesting. However, to take advantage of some economies of scale while satisfying even-flow requirements, some afforestation is economically viable. Land rental and transportation costs are the most significant costs limiting afforestation potential. Due to the large amount of suitable land available, to satisfy their demand for feedstock the ethanol producer need only rent the most marginal agricultural land and therefore pay the minimum rental rate. A relatively small change in rental cost at the margin would result in a change in the area afforested. The cost of transportation affects where afforestation occurs; 90 in the case study involving a moderately-sized ethanol facility (Scenario-A), it is uneconomical to plant and harvest trees over 50 km from the facility. It must be noted that the carbon accounting system to measure compliance with the Kyoto Protocol is not yet finalized and therefore some assumptions regarding changes in carbon stocks in terrestrial carbon pools may change prior to the first commitment period (2008-2010). One point that is agreed upon by the Parties to the Protocol is that double counting of emissions credits is not allowed. In other words, the avoided emissions due to renewable energy production can not be counted in addition to the reduced consumption of fossil fuels. As a "no-regrets" option, the production of ethanol from wood waste and afforested wood-biomass may be economically viable and contribute to the mitigation of greenhouse gas emissions. The development of markets for the trading of carbon credits increases the incentive for such a process, especially when credits for displacing fossil fuel use are attributed to the ethanol producer. The science of greenhouse gas behaviour and its relation to climate change and global warming is not proven; therefore an effective and internationally accepted accounting system for carbon credits and debits is still under negotiation. Consensus on which carbon pools and sinks would be included was reached at the end of the 6th Conference of the Parties to the United Nations Framework Convention on Climate Change but negotiations regarding how these pools and sinks will be measured is on-going. For the wood-ethanol industry, there remains considerable risk for investors in this new technology due to high capital costs, technological obstacles and market barriers. In order for the industry to develop, the Federal and Provincial governments will have to create incentives beyond those that currently exist. Government subsidies aimed at overcoming the risks 91 associated with this new technology can be justified on the basis of carbon gains. The government can also help promote the industry through the development of carbon markets where ethanol producers would be allowed to sell carbon offset credits. In the United States, incentives such as tax benefits, loan guarantees, and regulations have led to the development of a strong ethanol industry. The market potential in the U.S. and Canada is very large and could contribute significantly to economic growth over the next few decades. To the consumer, the debates over the future availability of fossil fuels is not as much of an issue as the effects of sudden, large jumps in oil and gas prices. Buffering the demand for gasoline through the addition of ethanol in the transportation fuel market will reduce the sensitivity of prices at the pump. 6.1 RECOMMENDATIONS FOR FURTHER STUDY To make afforestation more affordable there are opportunities to derive multiple economic benefits from each rotation of trees. For example, suitable logs can be sold to pulp and paper mills, while the remaining wood-biomass in the form of treetops and branches (and possibly roots) could be used as a feedstock for ethanol or other forms of energy production. If markets for the lignin co-product of the bio-ethanol conversion process can be found (within the chemical industry for example), the value of the lignin is likely to be higher than when it is burned for process steam and electricity in the bioconversion process. The use of lignin-based chemicals could replace current petroleum-based chemicals, thereby further reducing the use of fossil fuels. 92 As mentioned previously, the accounting system for measuring compliance with the Kyoto Protocol is not yet completed and will likely change from one commitment period to the next. As a result, estimates of greenhouse gas reduction potential will change and should be updated when possible, and especially when policy options are being assessed. 93 BIBLIOGRAPHY A F P A (2001). Wood Residue Inventory. Alberta Forest Products Association website: www.abforestprod.org/data/indes.htm. Alberta Energy (2001). Website: www.energy.gov.ab.ca, accessed July 2001. Alberta Provincial Base Map (1984). Scale 1:1,000,000. Apps, M . J . , W . A . Kurz, S.J. Beukema, and J.S. Bhatti (1999). Carbon Budget of the Canadian Forest Product Sector. Environmental Science & Policy 2(1): 25-41 Atkins, R.S. (1995). Generating Power with Waste Wood, Power Engineering 99(2): 38. B C Hydro (2001). Website: www.bchydro.com, accessed July 2001. Bronson Consulting Group (1999). Demand for wood residue for non energy products. Report prepared for National Climate Change Process, Forest Sector Table. Brown, S., A . E . Lugo, and L . R . Iverson (1992). Processes and Lands for Sequestering Carbon in the Tropical Forest Landscape. Water, Air, and Soil Pollution 64(1/2): 139-55. Bunce, H . W . F . , K . J . Blakeney, and A . E . Richmond (1991). Economic feasibility of utilizing logging slash and fire-killed timber for energy: Saskatchewan and Manitoba, Canada, 1991. Reid, Collins and Associates. E N F O R Project R-379, Forestry Canada, 1991. C F S (1999). Canada's wood residues: A profile of current surplus and regional concentrations. Canadian Forest Service, Industry, Economics and Programs Branch. Draft report prepared for National Climate Change Process, Forest Sector Table. C P P A (2000). CPPA's discussion paper on climate change. Canadian Pulp and Paper Association. Cline, W.R. (1992). The Economics of Global Warming. Institute for International Economics, Washington, D . C . Eswaran, H . , E . V . D . Berg, and P. Reich (1993). Organic carbon in soils of the world. Soil Science Society of America Journal 57: 192-194. Environment Canada (1997). Trends in Canada's Greenhouse Gas Emissions 1990 - 1995. Fearnside, P . M . (1997). Transmigration in Indonesia: lessons from its environmental and social impacts. Environmental Management 21(4): 553-570. 94 Ford-Robertson, J., K . Robertson, and P. Maclaren (1999). Modelling the effect of land-use practices on greenhouse gas emissions and sinks in New Zealand. Environmental Science and Policy 2: 101-109. Forest Sector Table, National Climate Change Process (1999). Options Report: Options for the Forest Sector to Contribute to Canada's National Implementation Strategy for the Kyoto Protocol. Canadian Forest Service, Natural Resources Canada November, 1999. Forsberg, G . (2000). Biomass energy transport: Analysis of bioenergy transport chains using life cycle inventory method. Biomass and Bioenergy 19: 17-30. Government of Alberta (2001). Agricultural Statistics obtained from website: www.agric.gov.ab.ca, accessed July 2001. Graf, A . , and T. Koehler (2000). Oregon cellulose-ethanol study: A n evaluation of the potential for ethanol production in Oregon using cellulose-based feedstocks. Prepared for the Oregon Office of Energy. Gregg, D. , and J.N. Saddler (1996). A techno-economic assessment of the pretreatment and fractionation steps of a biomass-to-ethanol process. Applied Biochemistry and Biotechnology 57/58: 711-727. Gregg, D.J . , A . Boussaid, and J.N. Saddler (1998). Techno-economic evaluations of a generic wood-to-ethanol process: effect of increased cellulose yields and enzyme recycle. Biosource Technology 63: 7-12. Hair, D. , R .N . Sampson, and T . E . Hamilton (1997). Summary: Forest management opportunities for increasing carbon storage. In: Hair, Dwight and R. Neil Sampson, eds. (1997), Forests and Global Climate Change, Chapter 14: 237-254. Hall, D . O . (1999). Biomass Energy Versus Carbon Sinks: Trees and the Kyoto Protocol. Environment 41(1): p.5. Hamilton, C , and L . Vellen (1999). Land-use change in Australia and the Kyoto Protocol. Environmental Science and Policy 2: 145-152. Hansen, U . (1998). Technological Options for Power Generation. Energy Journal 19(2): p.63. Helliwell, J.F. , and M . Margolick (1980). Some Hard Economics of Soft Energy: Optimal Electricity Generation in Pulp and Paper Mills with Seasonal Steam Flow Requirements. University of British Columbia, Department of Economics resource paper No.56: 35pp. 95 Hourcade, J .C. , R. Richels, and J. Robinson (1996). Estimating the Costs of Mitigating Greenhouse Gases. Climate Change 1995: Economic and Social Dimensions of Climate Change, Contributions of Working Group III, IPCC, Cambridge University Press. Intergroup Consulting Economists Ltd. (1982). Availability and cost of forest biomass in Canada. E N F O R Project P-224(l) IPCC (2000). Land-Use, Land-Use Change and Forestry: IPCC Special Report. Jepma, C.J . , M . Asaduzzaman, I. Mintzer, R.S. Maya, and M . Al-Moneef (1996). A Generic Assessment of Response Options. Climate Change 1995: Economic and Social Dimensions of Climate Change, Contributions of Working Group III, IPCC, Cambridge University Press. Karjalainen, T. , A . Pussinen, S. Kellomaki, and R. Makipaa (1999). Scenarios for the carbon balance of Finnish forests and wood products. Environmental Science and Policy 2: 165-175. Kort, J. and R. Turnock (1999). Annual carbon accumulations in agroforestry plantations. Agriculture and AgriFood Canada, P F R A Shelterbelt Centre, Saskatchewan. Lai, R. and J.P. Bruce (1999). The potential of world cropland soils to sequester C and mitigate the greenhouse effect. Environmental Science and Policy 2: 177-185. Lashof, D . and B. Hare (1999). The Role of Biotic Carbon Stocks in Stabilizing Greenhouse Gas Concentrations at Safe Levels, Environmental Science and Policy 2: 101-109. LeBlanc, A . (1999). Issues related to including forestry-based offsets in a G H G emissions trading system. Environmental Science and Policy 2: 199-206. Levelton Engineering Ltd. (1999). Forest Sector Table: Assessment of Gasification Technologies and prospects for their commercial application. Prepared for the Forest Sector Table, National Climate Change Process. L i m , B . , S. Brown, and B. Schlamadinger (1999). Carbon accounting for forest harvesting and wood products: review and evaluation of different approaches. Environmental Science and Policy 2: 207-216. Lindenbach, R .N . (2000). A n economic application of agroforestry in Saskatchewan: A comparative analysis. Masters of Science Thesis. University of Saskatchewan, Department of Agricultural Economics. Lofstedt, R. (1998). Sweden's Biomass Controversy: a case study of communicating policy issues. Environment 40(4): p. 16. 96 Lugo, A . E . and S. Brown (1993). Management of Tropical Soils as Sinks or Sources of Atmospheric Carbon. Plant and Soil 149: 27-41. Lynd, L . R . (1996). Overview and evaluation of fuel ethanol for cellulosic biomass: Technology, economics, the environment, and policy. Annual Review Energy Environment. Maclaren, P. (1999). Carbon accounting and forestry: a review of the subsequent papers, Environmental Science and Policy 2: 93-99. Marland, G . , and S. Marland (1992). Should we store carbon in trees? Water, Air and Soil Pollution 64: 181-195. Marland, G . , B. Schlamadinger, and P. Leiby (1997). Forest/biomass based mitigation strategies: Does the timing of carbon reductions matter? Critical Reviews in Environmental Science and Technology, 27(Special): S213-S226. Marland, G . , and B. Schlamadinger (1999). The Kyoto Protocol could make a difference for the optimal forest-based C 0 2 mitigation strategy: some results from G O R C A M , Environmental Science and Policy 2: 111 -124. McCloy , B . W . and D . V . O'Connor (1999). Wood-ethanol opportunities and barriers. Prepared for the Forest Sector Table, National Climate Change Process. Matsui, K . (1998). Global Demand Growth of Power Generation, Input Choices and Supply Security, Energy Journal 19(2): p.93. Ministry of Energy Mines and Petroleum Resources ( M E M P R ) and Ministry of Environment, Lands and Parks (MELP) , British Columbia (1995). Website: www.elp.gov.be.ca/epd/epdpa/ar/climate, accessed March, 2000. Nawitka Renewable Resource Consultants Ltd. (1999). Carbon sequestration aspects of an afforestation program in Canada's prairie provinces. Prepared for the Joint Forest Sector Table/Sinks Table of the National Climate Change Process. Nienow, S., K . T . McNamara, and A.R. Gillespie (2000). Assessing plantation biomass for co-firing with coal in northern Indiana: A linear programming approach. Biomass and Bioenergy 18: 125-135. Nilsson, S., and W. Schopfhauser (1995). The carbon-sequestration potential of a global afforestation program. Climatic Change 30: 267-293. Paustian, K . , O. Andren, H . H . Janzen, R. Lai , P. Smith, G. Tian, H . Thiessen, M . van Noordwijk, and P.L. Woomer (1997). Agricultural Soils as a Sink to Mitigate C 0 2 Emissions, Soil Use and Management 13: 203-44. 97 Pimentel, D. and G . Rodrigues (1994). Renewable Energy: Economic and Environmental Issues, Bioscience 44(8): p.536. Pingoud, K . , A . Lehtila, and I. Savolainen (1999). Bioenergy and the forest industry in Finland after the adoption of the Kyoto protocol. Environmental Science and Policy 2: 153-163. Plantinga, A . J . , T. Mauldin, and D.J. Miller (1999). A n Econometric Analysis of the Costs of Sequestering Carbon in Forests, American Journal of Agricultural Economics 81: 812-824. Price, C . (1997). Analysis of Time Profiles of Climate Change. Climate-change Mitigation and European Land-use Policies. CAB INTERNATIONAL, Wallingford, U K : 71-87. Richards, K . , (1997). The time value of carbon in bottom-up strategies. Critical Reviews in Environmental Science and Technology 27: 279-307. Rinebolt, D . C . (1993). The Potential for Using Wood for Energy and the Implications for Climate Change. In: Hair, Dwight and R. Neil Sampson, eds. (1997), Forests and Global Climate Change, Chapter 6: 117-129. Rittenhouse, R . C . (1995). Waste Wood, Biomass Crops Approaching Viability as Fuels, Power Engineering 99(5): p. 12. Rosenqvist, H . , Roos, A . , Ling, E . , and B. Hektor (2000). Willow growers in Sweden. Biomass and Bioenergy 18: 137-145. Row, C and R.B. Phelps (1995). Wood Carbon Flows and Storage After Timber Harvest. In: Hair, Dwight and R. Neil Sampson, eds. (1997), Forests and Global Climate Change, Chapter 2: 27-58. Samson, R., and Y . Chen (1995). Short-rotation forestry and the water problem. Proceedings of the Canadian Energy Plantation Workshop, Gananoque, Ontario, Canada, May 2-4, 1995. Samson, R., Girouard, P., Zan, C , Mehdi, B. , Martin, R., and J. Henning (1999). The implications of growing short-rotation tree species for carbon sequestration in Canada: Final report. Prepared for the Joint Forest Sector Table/Sinks Table: Afforestation #5. National Climate Change Process Solicitation No: 23103-8-0253/N. Schelling, T . C . (1998). The Environmental Challenges of Power Generation, Energy Journal 19(2): p.115. Sheehan, J. (1998). The role of bioethanol in global climate change. Presented at the 1998 National Conference on Ethanol Policy and Marketing, Albuquerque, New Mexico. 98 Skog, K . E . , T . C . Marcin, and L.S. Heath (1995). Opportunities To Reduce Carbon Emissions and Increase Storage by Wood Substitution, Recycling, and Improved Utilization. In: Hair, Dwight and R. Neil Sampson, eds. (1997), Forests and Global Climate Change. Chapter 12: 209-215. Sohngen, B. and R. Sedjo (1999). Estimating Potential Leakage from Regional Forest Carbon Sequestration Programs. R F F Working Paper, Washington, D . C . August, 1999: 26pp. Subak, S.P., P. Raskin, and D. von Hoppel (1993). National Greenhouse Gas Accounts: Current Anthropogenic Sources and Sinks, Climatic Change 25: 15-58. Stennes, B . (2000). Carbon uptake strategies in the western boreal forest region of Canada: Economic considerations. Doctor of Philosophy Thesis, Department of Forest Resource Management, University of British Columbia. Suchanek, P., S.L. Shaikh, and G . C . van Kooten (2001). Carbon Incentive Mechanisms and Land-Use Implications for Canadian Agriculture. Sustainable Forest Management Network Working Paper 2001 (6). Tol , R.S.J. (1996). The Damage Costs of Climate Change: Towards a Dynamic Representation, Ecological Economics 19: 67-90. U.S. Department of Energy (1998). Scenarios of US carbon reductions: Potential impacts of energy-efficiency and low-carbon technologies by 2010 and beyond. Van Dyne, D . L . , Kaylen, M . G . Blase (1998). The economic feasibility of converting renewable lingo-cellulosic feedstocks into ethanol and higher value chemicals. Presented at the A S A E Annual International Meeting, Florida, 1998. van Kooten, G . C , E . Krcmar-Nozic, B. Stennes, and R. van Gorkom (1999a). Economics of Afforestation for Carbon Sequestration in Western Canada, The Forestry Chronicle 76(1): 165-172. van Kooten, G . C , B . Stennes, E . Krcmar-Nozic, and R. van Gorkom (1999b). Economics of Fossil fuel Substitution and Wood Product Sinks when Trees are Planted to Sequester Carbon on Agricultural Lands in Western Canada, Canadian Journal of Forestry Research 29(11): 1669-78. van Kooten, G . C . (2000). Economic Dynamics of Tree Planting for Carbon Uptake on Marginal Agricultural Lands. Canadian Journal of Agriculture Economics 48: 1, 51-65. van Kooten, G . C , and G . Hauer (2001). Global Climate Change: Canadian Policy and the Role of Terrestrial Ecosystems. Canadian Public Policy. In press. 99 van Kooten, G . C . , S.L. Shaikh, and P. Suchanek (2001). Mitigating Climate Change by Planting Trees: The Transaction Costs Trap. Working paper, University of Nevada-Reno: 31pp. West, T . O . , and G . Marland (2001). A synthesis of carbon sequestration, carbon emissions, and net carbon flux in agriculture: Comparing tillage practices in the United States. Oak Ridge National Laboratory, Environmental Sciences Division, in press. Woodrising Consulting Inc., (1999). A n estimation of the impact on net carbon sequestration of forest management including wood products storage. Prepared for the Joint Forest Sector Table/Sinks Table. Woodrising Consulting Inc., Ontario, Canada. Woomer, P .L. , C A . Palm, J. Alegre, C . Castilla, D. Cordeiro, K. Hairiah, J. Kotto-Same, A . Moukam, A . Risce, V . Rodrigues, and M . Noordwijk (1999). Carbon dynamics in slash-and-burn systems and land-use alternatives: Findings of the alternative to slash-and-burn programme, Advances in Soil Science In press. Wright, L . L . , J .H. Cushman, and S.A. Martin (1993). Growing Short-Rotation Woody Crops for Energy Production. In: Hair, Dwight and R. Neil Sampson, eds. (1997), Forests and Global Climate Change, Chapter 8. 100 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items