One of the main reasons for growing bio-energy plantations is a reduction of the carbon emissions whilst obtaining energy benefits. Most of the intensive silvicultural treatments needed to ensure high levels of productivity in plantations (e. g. site preparation, weeding and fertilization — see Sect. 5.5) require some energy inputs. For example, fertilizer application uses fossil fuel based energy (and is responsible for some carbon dioxide emissions) during manufacture and/or mining, transport and application. In this section we will contrast the carbon costs of fertilization and other silvicultural treatments with the potential carbon gains from increased growth in short-rotation plantations.

Energy and carbon budgets for specific scenario’s in bio-energy crop systems are usually calculated through a life cycle analysis (LCA) approach, which takes account of energy inputs and outputs, as well as carbon gains and losses associated with every step of the production process. The major steps to construct carbon and energy budgets have been proposed by Schlamadinger et al. (1997) and subsequently implemented by inter alia Matthews (2001) Heller et al. (2003) and include:

• Definition of the system boundaries to do all calculations. A reference system is normally chosen, against which different alternative management scenario’s are contrasted.

• Estimating the total energy benefits from a specific scenario (including waste products).

• Estimation of total energy inputs in the production system (including energy investments in infrastructure and energy losses along the fuel chain).

• Estimating the carbon sequestration that takes place.

• Estimating of total carbon emissions from each specific scenario.

• Estimating the net emission of other greenhouse gasses, e. g. N2O, which can be presented as CO2 equivalents for the purposes of the budget.

Most authors stress the importance of detailed carbon and energy budgets for every step in the LCA because changes in management regime of biomass production systems can lead to large variations in the carbon and energy budget. Mead and Pimentel (2006) make a good case to show that individual silvicultural operations should be evaluated to decide on the optimum energy production system, as their efficiencies may differ widely. In order to obtain meaningful results, these calculations will have to be done on a site-specific basis because recommendations on the type and intensity of soil preparation (Smith et al. 2000, 2001; Zwolinski et al. 2002), fertilization (du Toit et al. 2010; Kotze and du Toit 2012) and weed control (Little and Rolando 2008) differ strongly across site types.

The figures most commonly used to evaluate the suitability of bio-energy systems are:

• The energy ratio (energy produced per unit of energy input)

• Net energy yield (energy output minus energy input per hectare)

• The carbon emission coefficient (inclusive of the greenhouse gas emissions expressed as CO2 equivalents)

Despite the strong dependence of the carbon and energy balance on individual cultural practices when applied to specific site types, a few useful generalisations can be made: Energy ratio’s of forestry crops (in temperate climates) typically vary between 10 and 25, compared with annual crops that vary between 1 and 5 (Mead and Pimentel 2006). Energy ratio’s as high as 55 have been reported for temperate climate forest crop systems under intensive management and fertilization (Heller et al. 2003). Energy ratio’s of 42 have been calculated for warm climate eucalypt crops in semi-arid areas (Wu et al. 2007) and ratio’s in excess of 60 have been estimated for a hypothetical Pinus taeda and Eucalyptus grandis systems that included site preparation and fertilization inputs (Mead and Pimentel 2006). The three warm-climate case studies have all shown considerable room for improvement of the energy ratio if specific vegetation management, harvesting and/or transport regimes are adopted. New research should focus on the ability of warm climate tree crops across different site qualities to produce energy efficient biomass through appropriate silvicultural management strategies.

Pyrolysis is defined as the thermochemical decomposition of organic materials in the absence of oxygen or other reactants executed under specific parameter conditions (Table 7.1). Several types of pyrolysis have been developed: vacuum pyrolysis (Roy 2000), pressurized pyrolysis (Tomeczek and Stanislaw 2003), fast pyrolysis (Bridgwater and Peacocke 2000), flash pyrolysis (Demirbas and Arin

2002) , torrefaction pyrolysis (Prins et al. 2006) and slow pyrolysis (Antal and Gr0nli

2003) . The processes differ from each other according to the conditions maintained in the pyrolysis reactor (Table 7.1). Fast pyrolysis leads to a high yield of bio­oil, while vacuum and slow pyrolysis offer a good compromise for the production of char and bio-oil, providing relatively high yields of both as well as providing superior quality char products (Bridgewater 2011; Bridgwater and Peacocke 2000; Chen et al. 2003). Bio-oil represents a valuable liquid fuel for boilers, while chemicals, nutritional supplements and/or pharmaceutical products may be isolated from it, provided that the challenging separation of these compounds can be achieved efficiently. The char represents a good feedstock both for boiler fuel and for the production of activated carbon. In more recent years char has also become an option for soil improvement and carbon sequestration (Antal and Gr0nli 2003).

Fast pyrolysis thermal decomposition of a feedstock using a relatively high heating rate can yield liquids (collectively termed bio-oil) of up to70-75 % of the weight of the starting material (Bridgwater 2011; Butler et al. 2011). One of the main advantages of fast pyrolysis lies in the fact that it is an effective method for densification of voluminous biomass. Different fast pyrolysis technologies exist, namely ablative, cyclonic, rotating cone, entrained flow, bubbling fluidised bed (BFB), auger, circulating fluid bed (CFB), transported bed, screw and auger kiln and wire-mesh reactor fast pyrolysis (Hoekstra et al. 2012; Bridgwater 2011; Butler et al. 2011). Currently, BFB and screw kiln reactors can be used for commercial — scale production of biofuel. Because of its properties relating to aging, instability, corrosion and viscosity, fast pyrolysis bio-oil may be upgraded physically, chemi­cally or catalytically (Bridgwater 2011). Bio-oil can be a substitute for fuel oil used for electricity generation and biorefinery.

Slow and vacuum pyrolysis processes deliver higher char and gas yields, limiting their energy applications to electricity and heat production. Slow pyrolysis follows the conventional carbonisation process of charcoal or biochar production, with recent improvements in recycling of gaseous/liquid products providing much of the heat required by the process, increasing overall energetic efficiency. The main difference between vacuum and slow pyrolysis lies in the method of removing gaseous vapours from the reaction zone. In vacuum pyrolysis, a vacuum is created which serves the same function as purge gas used in slow pyrolysis. Because of the lower pressures applied, aerosols tend to evaporate more easily. This removes them from the reaction zone and results in a significantly reduced vapour residence time of 2-3 s for vacuum pyrolysis, relative to the 165-170 s required for slow pyrolysis. Vacuum pyrolysis is therefore a modified slow pyrolysis resulting in improved quality of both liquid and char products (Carrier et al. 2011).

Torrefaction is classified as a mild pyrolysis technique, because it takes place in an inert atmosphere at relatively low temperatures (between 200 and 300 °C (Uslu et al. 2008)). The technology is less sophisticated than fast, flash, slow and vacuum pyrolysis technologies and can be seen as something in-between the combustion of dried biomass and of pyrolysis products. From a chemical point of view, torrefaction removes oxygen from the original biomass resulting in a solid product which has a lower O/C molar ratio (Van der Stelt et al. 2011). Torrefied biomass has potential for application in various industries: as raw minerals for pellet production, as reducer for smelters in the steel industry, for the manufacturing of charcoal or activated carbon, for use in gasification and for co-firing during boiler operation.

All biomass contains Carbon (C), Oxygen (O), Hydrogen (H), Nitrogen (N) and various other trace elements. On average wood contains about 49 % C, 6 % H and 44 % O with some variation between wood species (Fengel and Wegener 2003).

A high carbon content is directly related to a high density and therefore a high calorific value and desirable for combustion and gasification.

The major elements contained in any biomass, such as C, N, H, can be determined with elemental analysis and/or x-ray fluorescence (XRF) as described by Perkel (2012).

C, N and H can be determined by combustion in a pure oxygen atmosphere at high temperatures, which converts the sample elements into CO2, H2O and N2. The amount of the product gases is proportional to the amount of C, N and H in the sample.

In the XRF technique a material is exposed to high energy x-rays, which eject electrons from the sample atoms. The transition of outer electrons to the vacant spaces results in photo energy that is characteristic for the element and can be detected by a fluorescence detector.

The biogeochemical cycling of nutrients in forest ecosystems has been well described for several forest ecosystems (Jorgenson et al. 1975; Likens and Bormann 1995; Ranger and Turpault; du Toit and Scholes 2002; Laclau et al. 2005, 2010a; Dovey 2012). Nutrients reside within forest and plantation systems in several nutrient pools that differ in size and in the form in which a particular plant nutrient is held. Figure 10.8 is a simplified representation of a number of basic nutrient pools as well the movement of nutrients (usually termed nutrient fluxes) (a) into

Inputs

Atm. deposition N — fixation Fertilization Weathering

and out of the system (as part of the so-called input-output budget), and (b) among pools within the system (Ranger and Turpault 1999; du Toit and Scholes 2002). Forests are at risk of malnutrition and subsequent decline in productivity if the biogeochemical cycles of nutrients are decoupled in time or space. For example, an ecosystem can systematically be depleted of nutrients if outputs exceed inputs by a large margin. Several short rotation commercial forestry systems where only stemwood is harvested, are not at risk of nutrient depletion (du Toit and Scholes 2002; Ackerman et al. 2013). However, in the context of this chapter, increased nutrient removals in bio-energy plantations resulting from the harvesting of tree crowns, bark (and even roots) in addition the conventional stem wood harvest is probably the biggest concern (see examples in Fig. 10.9 below). Forest productivity can also be negatively affected if nutrients are not necessarily lost from the system, but end up in pools that are (temporarily) decoupled from their usual cycle and thus unavailable to the stand during a particular phase of growth. An example would be the lock-up of nutrients in the forest floor (Morris 1986) or the precipitation of a micronutrient in a plant-unavailable form following measures to raise soil pH, such as liming. In practice, management operations that affect nutrient pool sizes usually have a direct or indirect effect on nutrient fluxes too. For example, slash removal for bio-fuels will affect (for example) the content of nitrogen on the site, but also the rate of mineralisation which transforms N in a form that can readily be taken up by trees (du Toit and Dovey 2005; Deleporte et al. 2008; Gonqalves et al. 2008b; Mendham et al. 2008). Similarly, flre will affect the quantity of N and P oxidised during burning, but it may also strongly influence the availability of soil N through changes in mineralisation rates after the flre, as well as P availability (through changes in soil pH caused by the increase in pH (the ash-bed effect).

Natural resource inventory is an expensive and time-consuming activity and full enumerations, where every tree in the population under investigation is counted and measured, are only done in exceptional cases. A discipline of science evolved around the methods to extract a representative subpopulation for measurement, as a cost effective replacement for full inventories. In essence a sub-sample of trees is selected from the total population and all of the parameters described by the inventory procedure are measured. The mean of the parameters measured for the sub-population is then accepted, with an error margin, as the mean for the population. A discussion on the selection of representative samples follows.

Combretum and Acacia woodlands are found in patches in some parts of Southern Africa. In South Africa, Combreteaceae woodlands are dominated by Terminalia sericea which is usually found in association with other species. Terminalia sericea typically occurs on deep (>1 m) sandy soils, and may be nearly monodominant following disturbance (Scholes 2004). Shallower or rocky, infertile soils are domi­nated by Combretum apiculatum in semi-arid situations, giving way to Combretum collinum in slightly wetter areas (Scholes 2004).

Semi-arid shrubland vegetation formation covers over 900,000 km2 and com­prises either microphyllous wooded grassland or shrubland in which there is a more-or-less continuous grass layer (Timberlake et al. 2010). Due to low rainfall, drought, low temperatures, exposure to wind or salinity and toxicity or extreme oligotrophy of the soil, operating singly or in various combinations, the trees are sparsely distributed and are around 5-8 m high (White 1983; Timberlake et al. 2010). The characteristic species include various Acacia species (A. erioloba, A. luederitzii, A. fleckii, A. hebeclada, A. mellifera, A. tortilis), Boscia albitrunca, Dichrostachys cinerea and Terminalia sericea (Scholes 2004; Timberlake et al. 2010). Although the characteristic species are mostly Acacias, only one or two of these species may dominate the vegetation formation (Scholes 2004) depending on the availability of moisture. For example, Acacia mellifera and A. erioloba are widespread on the arid fringe; A. karroo in the southeast; A. tortilis and A. nilotica in the northeast, and A. robusta and A. sieberiana in the moister areas (Scholes 2004).

Stumps after harvesting represent a significant potential for increased utilization of bioenergy. But the utilization of stumps for energy in countries like Sweden and Finland is mostly constrained by ecological concerns of the physical, chemical and biological impacts on the soil (Lindholmet al. 2010). In Finland some 850,000 m3 of stump wood was utilized in 2009, and the production is increasing (METLA 2011). The mass of the stump and coarse roots may be some 25 % of the utilized stem (Marklund 1988). For example, Fig. 6.6 demonstrates the close correlation between mass and stump diameter for Spruce stumps in Europe.

Using specially designed stump pulling and splitting heads (Fig. 6.7), harvesting in good conditions can produce 2-4 dry tonnes per productive machine hour, equating to roughly 100 stumps per productive machine hour (Athanassiadis et al. 2011). The split material is bulky and forwarding productivity rates of 7-9 m3 per productive machine hour on extraction distances of 50-500 m were found (Laitila et al. 2008). Productivity was lower than for any other forest product, with 27 % of the time being used on unloading alone.

Stumps are normally “seasoned” by leaving them in roadside piles for some time (e. g., one year) mainly to allow precipitation and wind to erode the worst of the soil and stone contamination, and are invariably crushed with tub-grinders before or after transportation. Studies in Sweden indicate that pre-grinding and screening of stumps at landing reduces contaminant levels and loading time, and saves 15-20 % on transport costs (Thorsen et al. 2011). Apart from the volumes and generally good quality of the fuel, a considerable benefit of stump utilization from a cost perspective is that they represent an additional resource within the same procurement area. However, stumps from pine and eucalyptus dominated industrial plantations are not as easily lifted as those of spruce, and larger and more robust lifting heads would be required, depending on soil type, rooting pattern, and stump size.

Butanol can be generated biologically by acetone-butanol-ethanol (ABE) fermenta­tion. This metabolic route is present in bacterial species in the genus Clostridium such as C. acetobutylicum and C. beijerinckii. ABE fermentation has been indus­trially applied using starchy biomass as an alternative to chemical synthesis from fossil-fuels. Similarly, the sugars present in lignocellulose hydrolysates can be fer­mented into 1.2-biobutanol, also known as 2,3-butylene glycol (2,3-BD), by mixed acid-butanediol fermentation (Menon and Rao 2012). Several microorganisms have the 2,3-BD pathway but the most commonly used is Klebsiella oxytoca, given its wide sugar spectrum and adaptive potential. Branched alcohol mixtures with high iso-butanol content can also be formed from glucose through synthetic pathways present in bacteria such as Escherichia coli (Rodriguez and Atsumi 2012).

Although these technologies are not as mature as those involved in the production of bioethanol, there is a growing interest in these kinds of biofuels, reflected by the investment in research for their commercial application by companies such as

Dupont, BP, Gevo and Green Biologics. The economic feasibility of production of these biofuels depends on the use of cheaper feedstocks to increase yields and productivities as well as on the development of more efficient recovery processes (Jin et al. 2011). In this context, the use of lignocellulose residues, selection of bacterial strains as well as process development is expected to reduce butanol production costs (Kumar et al. 2012).

Typical process yields for biochemical conversion of woody biomass available in the Southern hemisphere to bio-alcohols are presented in Table 7.4. The impact of biomass properties on biochemical conversion processes is addressed in Chap. 8.

The argument that fuelwood is not necessarily a preferred energy source but one forced upon the rural poor due to economic considerations, was mentioned briefly in the previous section. The main economic factors that drive the use of fuelwood in developing countries are the low costs of procuring fuelwood and the low income of consumers. Often the main cost associated with fuelwood collection is the opportunity cost of the time (which could be substantial) taken to collect fuelwood (Cushion et al. 2010) (Fig. 9.1). In this case, rural households can allocate scarce cash resources to other needs such as education of children, investment in agricultural tools and capital for income generation activities. Such cost savings would best be reflected by the replacement value of the energy sources that fuelwood substitutes, rather than direct cost of fuelwood (Shackleton 2004).

The commercial role of fuelwood can also be significant when community members supplement their incomes by selling fuelwood. Sometimes this activity even becomes their main source of cash. Notably, this includes also the poorest of the poor where many rural landless people are among those specialising in fuelwood production (Vedeld et al. 2004).

While it seems that when household income increases, people prefer to move up the energy ladder to other energy sources (typical scenario where fuelwood is seen as an “inferior” good), some researchers also report that fuelwood use increases with increased income for very poor households where fuelwood is seen as a “normal” good (Arnold et al. 2003; FAO 2005). Other social and environmental considerations

that might have an effect on household fuelwood consumption include climate (People might prefer to cook on fuelwood as it also provides heat for their house in winter (Chirwa et al. 2010)), access to markets and forest resources as well as health considerations (Exposure to indoor air pollution from biomass fuels is linked to many respiratory conditions and diseases (Cushion et al. 2010)).

The choice of energy source should thus be seen in the wider socio-economic developmental context and has various positive and negative implications for the user (mostly related to cost and convenience) as well as for the energy and CO2 balance of the country. Table 9.2 summarises the key implication related to each of the energy sources along the energy ladder.

When the cost of emissions reduction is brought into the equation, progress up the energy ladder becomes less clear cut. In cases where biomass is the least expensive source of fuel and is produced sustainably, traditional biomass energy use could result in much lower emissions than fossil fuel alternatives. In such cases it might not be beneficial from a national emissions point of view to promote coal derived electrification schemes but rather to focus on more efficient traditional use of fuelwood and sustainability of fuelwood supply (Cushion et al. 2010).

It is, however, also important to consider that the long term success of bioenergy programmes are linked to the socio-economic impact that it will have on rural producers of biomass. The socio-economic implication of bioenergy programmes has often been neglected in academic studies on the sustainability of biofuels. Most work focussed on the environmental implications and the interaction between the environment and economy while the social dimension remained weakly defined (Lehtonen 2011). The following sections of this chapter will consider the costs and benefits of bioenergy programmes to rural producers of biomass and users of bioenergy.

Inventory analysis involves data collection and calculation procedures to quantify the relevant inputs and outputs of a product system. The data collection can be a resource intensive process. For each unit process that is included within the systems boundaries the relevant inputs and outputs, such as use of resources and releases to air, water and land associated with the system, need to be considered. Interpretation may be drawn from these data, depending on the goal and scope of the LCA. These data also constitute the input to the LCIA (ISO 14040 1997) (Fig. 11.4).

A system model thus needs to be built according to the requirements of the goal and scope definition. The systems model is the flow model for a technical system with certain types of system boundaries (e. g. ‘cradle-to-grave’ or ‘cradle-to-gate’).

The process of conducting an inventory analysis is iterative. As data are collected and more is learned about the system, new data requirements or limitations may be identified that require a change in the data-collection procedures so that the goals of the study are still met. Sometimes, issues may be identified that require revisions to the goal or scope of the study.

The result is an incomplete mass and energy balance for the system. It is incomplete in the sense that only the environmental relevant flows are considered, which more or less include the use of scarce resources and the emissions of substances considered harmful. Environmentally indifferent flows such as water vapour emissions from combustion or industrial surplus heat are disregarded. Figure 11.4, below, is an illustration of the main steps and flows involved in an LCA.

When dealing with systems involving multiple products, allocation methods are needed. The materials and energy flows as well as associated emissions need to be allocated to the different products according to justifiable, clearly stated and well documented procedures. The allocation method can be, based on a unit of mass or energy, or in some instances on the financial value of the products.