Finnveden et al. (2009) distinguish between two types of LCAs: attributional and consequential LCAs. The attributional LCA is defined by its focus on describing the

environmentally relevant physical flows to and from a life cycle and its subsystems. The consequential LCA is defined by its aim of describing how environmentally relevant flows will change in response to possible decisions (Curran et al. 2005). Similar distinctions have been made in several other publications, but often using other terms to denote the two types of LCA, and sometimes including further distinctions of subcategories within the two main types of LCA (Guinee et al. 2002). Baumann and Tillman (2004), for instance, distinguish at least three types of LCAs, namely LCAs of the accounting type, LCAs of the change-oriented type and standalone LCAs.

LCA studies of the accounting type are comparative and retrospective. This type of LCA is well suited to different types of eco-labelling and can be used in purchasing or procurement situations, since these applications involve a comparison of existing products. LCA studies of the change-oriented type are comparative and prospective. This makes them useful in product development, building design and process choices, since decision-making involves a comparison of options that may be implemented or produced in the future. A standalone LCA is used to describe a single product, often in an exploratory way in order to get acquainted with some important environmental characteristics of that product, identifying the ‘hot spots’ in the life-cycle, i. e. which activities cause the greatest environmental impact (Baumann and Tillman 2004).

In general, the attributional method is the most used in LCA, but in LCA of bioenergy systems the consequential methods appears as the most broadly applied. Almost 75 % of relevant studies reviewed by Cherubini and Str0mman (2011) compare the environmental impacts with those of a fossil reference system, as they are aimed at addressing the needs of policy makers, since consequential LCA is more relevant for decision-making.

A non-linear artificial neural network (ANN) was the final statistical approach evaluated. Jordan and Bishop (1996) describe neural networks as a graph with

patterns, represented by numerical values attached to the nodes of the graph, and transformation between patterns achieved via message passing algorithms. The power of an ANN is rooted in the fact that it is designed to replicate a biological neural system. An advantage of using this methodology is that an ANN has the ability to learn and adapt to the underlying structure of the data set being analysed. The ANN approach is also capable of handling non-normality, non-linearity, and collinearity within a data set of interest (Haykin 1994). Typically, an ANN is trained using a sample or training set that consists of both dependent and independent variables. In this study the training data set was extracted from the 225 sample plots and included maximum LiDAR heights as dependent variable and mean IKONOS reflectance and compartment age as independent variables.

The structure of an ANN is made up of several layers (Fig. 2.8), most notably an input layer (containing the training data), a number of hidden layers, and an output layer. Each layer is in turn made up of a number of neurons, which represent the fundamental processing unit of any ANN. A neuron consists of three parts (Fig. 2.9), namely a set of synapses or connecting links, an adder or transfer function, and an activation function. For this study we first normalised (subtract mean and divide by the standard deviation) spectral reflectance and age before introducing these independent variables to the neurons via the synapses or connection links.

Inputs were standardised (0-1) to facilitate inclusion of the age variable and weights were initially set using a random seed value. Optimisation of the network was undertaken using bootstrapping methods, with 50 % of the training data removed from the training set and used to assess the accuracy of the network weights. Fifty bootstrapped models were calculated for each epoch, with a total of 500 epochs specified. The bootstrapping approach allowed for the calculation of mean effects of each input variable (confidence interval), similar to the standard

output of a multiple regression analysis. Goodness of fit statistics, similar to those reported in regression models, was used to assess the model. Finally, the neural network was applied directly to the population data to create spatially explicit maps of LiDAR height after model training was completed.

There can be little doubt that the selection and genetic improvement of fast­growing tree taxa (in this context referring to species, provenances, families within provenances, hybrids or clones) have strongly boosted productivity on intensively managed plantation forests in the tropics (Zobel and Talbert 1984; Verryn 2002, 2008; Pallett and Sale 2002; Kanzler et al. 2003; Wu et al. 2007; Boreham and Pallett 2007). Particularly impressive tree breeding successes in short-rotation pulpwood plantation forestry also include improvements in properties that enhance processing (e. g. wood properties, stem form and ease of debarking — Malan and Verryn 1996; Dvorak et al. 2008) or properties that allow for better survival and productivity of a specific taxon under adverse circumstances (e. g. disease resistance, cold/drought/frost tolerance, improved water use efficiency and herbicide resistance — Hodge and Dvorak 2007; Herbert 2012; du Toit 2012; Mitchell et al. 2013). It is therefore imperative for any intensive bio-energy production system to invest in a focussed tree improvement programme that can conserve a broad genetic base of fast-growing tree families, constantly breed for resistance to newly emerging pests and diseases, and constantly improve quality of the biomass and its suitability for the particular production process. The word-wide trend in highly productive short-rotation plantation forests is to move increasingly towards planting a variety of genetically improved, vegetatively propagated hybrid clones (rather than raising seedlings from half-sib or full sib families within species), which are deployed in a mosaic of small blocks to minimise risk. Some of the more important reasons for this trend revolve around the following facts: (a) genetic gains are large and guaranteed, (b) hybrid vigour can be obtained, (c) disease resistance can be obtained through hybridisation, (d) large improvements in stand uniformity, with which comes ease of mechanisation and an increase in the partitioning to above-ground biomass (Stape etal. 2010).

Another advantage of planting clones rather than pure species of fast-growing exotics has to do with the invasive potential of some taxa when planted in a non­native environment. There is an increasing body of evidence showing that several hybrid plants are sterile or do not produce large quantities of viable seedlings (Eldridge et al. 1994; Lopez et al. 2000; Owens and Miller 2009; Chen 2010). It follows that highly bred hybrid tree clones that are less fertile or sterile can potentially be selected for planting biomass crops, so that they do not pose an invasive threat. This aspect needs further testing and experimentation, but holds promise for the establishment of “greener” bio-energy crops.

Equally important to the genetic tree improvement process that may improve stand productivity is intensive experimentation with site-taxon matching. Matching the planted taxa to site conditions is obviously important in biophysically complex landscape where climatic and edaphic conditions differ markedly in space and time. There are several examples all over the world where species/provenances or families that were not well adapted to specific climate conditions have been devastated by a single risk factor, e. g. the infection of pine species that have little resistance to Diplodia pinea when planted on sites that experience hailstorms (Swart et al. 1988), and the stem breakage caused in Acacia mearnsii and Eucalyptus grandis stands in the KwaZulu-Natal Midlands of South Africa following episodic, heavy snowfalls (Gardner and Swain 1996). However, new pests or diseases can easily be introduced to regions of exotic plantations, e. g. Phytophthora pinifolia in Chile or Sirex noctilio in South Africa (Tribe and Cillie 2004; Duran et al. 2008; Hurley et al. 2012). Furthermore, there are documented evidence of pests and diseases of indigenous trees that infect or attack distantly-related exotics, e. g. Crysoporthe austroafricana stem canker that recently spread to infect exotic stands of Eucalyptus grandis stands in sub-tropical parts of South Africa (Wingfield et al. 2008). It follows that regional enterprises with fairly uniform climates (and therefore chiefly rely on just one or perhaps a few species) are at risk. In such cases, it pays to invest in research on several taxa that could potentially be suited to a site: (a) to allow for the planting a mosaic of different families/clones and so to minimise risk, (b) to have alternative taxa that can be deployed rapidly and effectively in case of the introduction of a significant pest or disease to the region, and (c) to improve site-taxon matching for lesser known taxa.

Several authors have been successful in matching species or provenances to a broader region using climatic similarities, often with a computer-aided approach, e. g. bioclimatic parameters. However, finer scale matching of taxa to specific site types require local knowledge of tree response to risk factors, as well as stand productivity and quality of biomass that can be obtained under the range of pre­vailing site conditions. This involves planting and testing all promising taxa across a wide range of site types in the region where it could be grown. It also presupposes that a sophisticated yet simple site classification and site evaluation system exists (e. g. Louw and Scholes 2002; Smith et al. 2005; Louw et al. 2011; Louw and Smith 2012) — sophisticated enough to take both risk factors and the drivers of resource availability to stands into account, yet simple enough to implement in practice. Information gathered from studies such as the aforementioned examples is the only reliable foundation upon which detailed site-taxon matching can be based. Accurate site-taxon matching becomes even more important when (a) the genetic base of the planting material becomes increasingly narrow, e. g. in the sequence provenance! family! clone, and (b) when the silvicultural regime tends toward short rotations of unthinned crops. A stand consisting of a several families of half-sibs that are only moderately suited to a specific site will still contain some individuals that are well suited to the prevailing conditions, and those individuals could be the final crop trees remaining after thinning in a medium to long rotation system. This is often the case in plantation forests grown for sawtimber on 20­35 year rotations in Southern Hemisphere countries. However, if a single clone (that is not particularly well suited to the site conditions) is planted in a short, unthinned rotation, the productivity of that stand is guaranteed to be sub-optimal, and the productivity loss will be compounded over successive rotations if a coppice system is used. Furthermore, in the event of climatic extremes, an entire crop could be affected by pests or diseases, brought about by stress. It is therefore essential to have a risk profile as well as a “response surface” for the productivity potential of each important or potentially important taxon across a the broadest possible range of site types, to aid the silviculturalists in their decision making.

Many forest owners are already members of cooperative organizations that provide services (forest management, harvesting and logistics) and access to market chan­nels. Also, some bioenergy conversion plants create specialized cooperative units which manage the longer term procurement of biomass on their behalf (or a group of forest owners invest in a bioenergy plant). Irrespective of the structure or origin, this business process model is characterized by the undertaking to maximize the interests of their members.

In this case, the forest owner can either supply directly (as in case above) or can choose to leave the procurement process to the cooperative. In the latter case, the cooperative carries the costs to the contractors in the supply chain. Advantages of cooperative supply are: (1) that the interests of the forest owner are seen to; (2) that the cooperative can incubate good contractors and streamline business processes; (3) economies of scale and benefits (e. g., speed of payment) of being a large and consistent supplier to an energy plant; and (4) that dividends, bonuses, or revenue adjustments can be made retrospectively.

The MC of partially dried wood can be measured with a resistance meter. In this method two pins are inserted into the wood and the measured resistance can be converted to a MC value. This technique works, however, only for MC values below fibre saturation point (around 30 % MC), on large enough samples and is generally not very accurate.

For bioenergy purposes the MC is typically determined with the ovendry method (ASTM D1762). For this a small sample is cut, weighed when wet, dried at 103 ± 2 °C for at least 24 h and weighed again when dry. MC can then be calculated according to Eqs. 8.1 or 8.2.

Tree water-use also changes relative to the age of the tree, typically following a sigmoid curve, with a low initial water use followed by rapid increases flattening out to a plateau once canopy cover is achieved (Scott and Smith 1997; Dye and Bosch 2000). In certain long-rotation sawtimber stands (>30 years for pines, and >15 years for eucalypts), tree water-use (streamflow reduction) has even been observed to tail off with age (Scott and Prinsloo 2008). However, with the advancement of tree improvement programmes and genetic selection for faster growth rates, rotation lengths have generally decreased, particularly in eucalyptus plantations (Verryn 2000). Rotation lengths are predicted to shorten even further with the advent of intensive biomass production plantations, typically reducing from 6 to 5 years on good quality pulpwood sites, and from 12 to 8 years on sawtimber sites. Combined with higher stand densities, these plantations will produce trees of smaller individual volumes at harvesting, but greater biomass production per land unit overall. Due to the shape of the tree water-use curve relative to tree age, reductions in rotation length are likely to decrease the overall water-use of the plantation, resulting in lower streamflow reductions (Fig. 10.2).

A challenge in evaluating the wider scale hydrological impacts of rotation length changes associated with biomass plantations is that individual compartments within a commercial forest plantation go through a growth cycle from planting to clearfelling and back to planting again, with constantly changing water use impacts. However, at a landscape scale it may be helpful to consider the planted area as

Fig. 10.3 Hypothetical changes in tree growth and water-use relative to stand age, before canopy closure within a pulpwood stand compared to an intensive biomass production stand

a mosaic of compartments representing all ages from seedlings to mature trees, growing simultaneously, and cycling through the various stages of growth and water use. For areas growing just one species under one rotation length, the net water use impact will be that of plantations at the “water-use mid point” of their rotation. However for areas growing multiple species at different rotation lengths water use impacts may need to be weighted relative to species and rotation length predominance.

Energy development in many industrialised countries shows a reverse trend, turning away from fossil fuels and nuclear power towards regenerative energy sources. International commitments towards clean development mechanisms, such as the Kyoto protocol, increasing fossil prices due to a dwindling supply, and the general aversion to nuclear energy in many developed countries, lead to a dynamic development of alternative energy sources in the past two decades. The European Union, which may serve here as an example for the industrialised nations, has defined the goal to increase renewable energy in their energy portfolio from the current value for the year 2011 of 13 % (Eurostat 2013) to 20 % in the year 2020 and is planning to move beyond that mark (European Commission 2010). About 70 % of the current European renewable energy production was from biomass in 2011 and about 70 % of that biomass was based on wood and wood residues (Eurostat 2013). In the last decade the amount of wood based biomass used for energy has increased by more than 50 % (Fig. 1.2). In the same time frame (2002-2011), net imports of biofuels have increased by a factor of ten, indicating that the import of solid and liquid biofuel is part of the strategy to transform the energy portfolio in the European Union. Due to restricted access to land and high production costs, industrialised

nations increasingly try to source their biofuels from tropical countries, preferably in the form of easily transportable fuels such as oils, fatty acid distillates, bioethanol or solid pellets, all of which can facilitate co-firing in power plants or be blended in fuel for motor vehicles.

However, to embark on commercial bioenergy projects for the mitigation of climate change might be a two-edged sword. Non sustainable practice, in particular with palm oil production, have raised concerns about the suitability of biofuel production for mitigating climate change since it may degrade existing natural resources and may further increase climate change due to deforestation and the deterioration of natural ecosystems, especially in tropical countries (Wicke et al. 2008; Butler et al. 2009). Many of these concerns are valid, particularly in the tropics, where conversion from one land-use to another is quite common, where it is often loosely regulated and controlled and land tenure is frequently unsolved. Conversely, if bioenergy is produced sustainably, it offers the potential to provide an energy resource, sequester carbon and at the same time alleviate poverty in many developing countries in the tropics. In this context it is important to see both aspects of current bioenergy use. The traditional low-tech fuelwood aspect for everyday cooking and heating that is prevalent in most developing countries, and the high-tech approach of many industrialised countries, which will also be realised mainly in the developing countries of the tropical and sub-tropical areas. Some tropical and sub-tropical countries have embarked on commercial bioenergy production, driven partly by local demands and by the increasing resource needs of the industrial countries of the Northern Hemisphere (Mathews et al. 2000; Wright 2006).

In this section an example of tree biomass modelling is presented, introducing several key techniques. All models were parameterised using the freely available statistical software package R (R Development Core Team 2012).

The starting point for this modelling section is a set of data obtained from 99 Eucalyptus grandis trees from the Karkloof experiment at different ages, ranging

from 0 to 11 years. The data is described in du Toit (2008). The biomass has been scaled up to tree level and four biomass fractions that should be modelled (stemwood, stem bark, branches including bark, and foliage). The objective is to create models for the different biomass fractions as well as for the total biomass under the constraint of additivity.

A method for forcing additivity was introduced, based on a separated estimation of total aboveground biomass of the tree and the compositional estimation of proportions for the biomass fractions. The total biomass was estimated with a traditional ln-transformed linear model to account for heteroskedasticity (Eq. 3.6). As independent variables DBH, height and the compound variable D2H were tested. The best model was selected based on the Akaike Information Criterion, which has theoretic foundations in information theory (Burnham and Anderson 2004). The AIC penalises the inclusion of additional variables and thus helps to keep the models parsimonious. The best model fit in this case was achieved by Eq. 3.6.

ln (ABMtotal) = a + b ln (D2H) + c ln(H) (3.6)

Here ABMtotal is the total above ground biomass (kg) of the tree, D is the DBH (cm) and H is the tree height (m). For back-transformation the variance of the residuals о2 was calculated, multiplied by 0.5 and added to Eq. 3.6 for back- transformation bias correction (Eq. 3.7) as proposed by Baskerville (1972).

ABMtotal = e(a+b ln(D2H)Cc ln(H)+^20.5) (3.7)

An almost unbiased model was achieved (Fig. 3.6).

The next step was to model the proportional distribution of the biomass fractions in a simultaneous approach based on the Aitchison-Simplex and an isometric log-ratio transformed (ILR) model (Aitchison 1982; van den Boogaart
and Tolosana-Delgado 2008). The model foundations for this approach have been laid down by Aitchison (1982, 1986) and the method was successfully applied for the estimation of compositions in Geosciences (van den Boogaart and Tolosana — Delgado 2008). The data are modelled as a closed composition with a relative geometry. This means that the individual components are forced to add up to 1 and their relative proportions are of interest rather than their absolute values. Similar to the logarithmic transformation the ILR-transformation takes care of the effect of heteroskedasticity. Expressed in simple words the compositional relative proportions are transformed into a euclidian orthogonal coordinate system. The dimensionality (D) is hereby reduced by one, because the compositions have to add up to 1. This means, that the last component in the euclidian space is not directly predicted but automatically derived in the back-transformation step. Classical multivariate analysis is applied in the euclidian space and then the results are back — transformed into the original coordinates to provide meaningful composition parts.

The ILR-transformation is based on a CLR-tansformation, and a multiplication with a triangular Helmert matrix (van den Boogaart and Tolosana-Delgado 2008). The matrix multiplication does the dimensional reduction from D to D-1. The triangular (D, D-1)-Helmert matrix can be derived from a normalised Helmert contrast matrix as shown in van den Boogaart et al. (2013). Further details on the theoretical framework are provided by Aitchison et al. (2002) and van den Boogaart and Tolosana-Delgado (2008). One convenient side-effect of this method is that the solution in the transformed space can be found by applying well known linear statistics. The “compositions” package of R was used (van den Boogaart and Tolosana-Delgado 2007; R Development Core Team 2012). The code for modelling compositions in R is available from the web page of this book.[2]

After ILR-transformation of the data combinations of the following independent variables were tested: DBH, H, D[3]H, ABMtotai (see Eq. 3.8).

The best model in the transformed space was achieved by a linear combination of DBH, height and total aboveground biomass (Eq. 3.8)

ilr (compositionABM/ = bBHD + cH + dABMtotal (3.8)

The regression parameter output is provided in Table 3.2.

These are the parameters in the ILR-transformed space. Backtransformation is conveniently available in the statistical software package compositions. The model can be evaluated, based on the two different model parts separately and on the total model, where the compositions are multiplied with the total biomass estimation.

The extraction of loose material that is not chipped and which weighs less per load than solid wood is less damaging to the site than infield chipping and extraction of heavier loads on the same site (Stupak et al. 2008). This is due to the lower mass and bulk density of the material and reduced impact of the total mass on the soil surface. It also results in inefficiency; however, as less tonnage is extracted due to the loads being volume-restricted. Residual biomass harvesting system selection decisions therefore need to match biomass type with specific machines to result in optimally productive and cost efficient harvesting systems (Stupak et al. 2008). Conventional harvesting systems, which are presently in use form the basis of residual biomass harvesting systems selection, and are dependent on the location of comminution:

• Terrain chipping

• Chipping at roadside or landing

• Terminal chipping

• Chipping at processing plant

However, conventional forwarders or agricultural tractor/trailer units fitted with forestry trailers (i. e., with crane and bogie axle), encompassing CTL, TL and FT systems remain the preferred means of extracting small trees, bundles, tree parts or harvesting residues. This is due to their design and availability to the sector.

These come in various adaptations, all of which aim to maximize the payload of a bulky, low mass product and selection remains specific to particular situations. A FT system incorporating cable and/or grapple skidders is an option as an integrated roundwood/biomass harvesting system provided sufficient space is available on the landing to cater for storage of the resultant biomass. From this point decisions can be made to either comminute at roadside or transport the loose biomass further towards the final consumption point.

Lignocellulosic biomass can be biochemically transformed into different bioen­ergy products which are either liquid (bioethanol, biobutanol, 1,2-biobutanol, branched liquid alcohol mixtures) or gaseous (biogas, biohydrogen) and can be used as replacement for conventional combustible fuels used for electricity genera­tion, transport and various heating applications. Table 7.3 lists the lignocellulose source (energy crops or residues generated by the forestry industry and pulp mills), the technology for biological transformation and associated products of the transformation.

Lignocellulose conversion into biofuels through biochemical transformation involves the following steps:

• Pretreatment of biomass for enhanced accessibility and digestibility by enzymes

• Hydrolysis of cellulose and hemicellulose to sugars and/or organic acids

• Microbial transformation of sugars/acids into designated biofuels

• Product recovery

From the abovementioned bioenergy products, bioethanol and biogas have been the most extensively studied and are nearest to commercialization, while butanol is receiving increased attention. The conversion technologies involved as well as biomass specifications are described in more detail in the following sections.