Theoretical optimum productivity.
Factors limiting growth
Potential global productivities.
Present land use and availability.
Land availability.
Carbon balances and fossil-fuel substitution.
The fundamental determinant of biomass productivity is the amount of sunlight falling on the leaves of the plant. The ability of the plant to utilise this resource is mediated by temperature, water and nutrient availability, the plant type and species, and the plant's ability to deal with pests and diseases. Plants absorb photosynthetically active radiation (PAR) in the wavelengths from 400-700 nm. PAR represents roughly 50% of the energy of the total incoming radiative energy. Of this energy, further losses occur through reflectance by the leaves and transmission through them, interception by non-photosynthetic components both within the leaves and by the branches, twigs etc, and efficiency losses with which the energy in absorbed photons is converted into chemical energy as fixed carbon bonds. These losses dictate a maximum theoretical photosynthetic efficiency of 6.7%, if total PAR is utilised throughout the year. {Bolton & Hall, 1992}
Besides water availability, temperature is of central importance since it governs the length of the growing season i.e. C3 plants grow optimally between 20 and 30°C (and not below 0-5°C), C4 plants between 30°C and 40°C (and not below 10-15°C). Thus away from the tropics temperature constraints can severely limit the length of growing season.
To grow, plants must absorb carbon dioxide for which the stomata must be open; however, due to the thermodynamics of diffusion, water can escape to the atmosphere at a faster rate than carbon dioxide can enter the leaf. Therefore, photosynthesis results in the loss of the water. This water loss is turned to the plants advantage as it is essential for transporting nutrients through the plant and also for structural and biochemical requirements.
A major source of carbon loss in C3 plants occurs as a result of photorespiration which causes the loss of about 30% of the carbon already fixed through photosynthesis. C4 plants (which have made structural and enzymatic alterations to minimise this loss) suffer negligible losses from photorespiration. Attempts to select C3 species which have lower photorespiratory levels have been largely unsuccessful.
Theoretical calculations of the maximum potential yields for trees, all of which are from the C3 group of plants, under conditions which are neither nutrient or pest/disease limited, show the absolute potential productivities possible. Thus for Plymouth (50°N), with an annual average daily insolation of 11.1 MJ m-2, temperature constraints reduce the theoretical maximum yield (without pest and disease losses) from 156 t/ha/yr. to 50 (oven dry) t/ha/yr. {Hall et al. 1992}
Temperature. As can be seen, temperature plays a central role in the productivities a plant can achieve and not surprisingly different plant species being adapted to different temperature regimes. Thus, C3 species such as barley, willow and alder are adapted to temperate climates to maximise their growth under the prevailing conditions. Many deciduous temperate species lose their leaves when the temperature falls below levels at which effective photosynthesis can occur. Such plants thus minimise metabolic losses from the now redundant leaves and also avoid damage due to severe drops in temperature. However, the canopy must be quickly re-established at the beginning of the next growth season when high CO2-fixation rates can be achieved. Perennial species can take advantage of their existing branch structure to redeploy their leaves at a faster rate than annuals; again ensuring a longer growing season. Leaves of evergreens can remain functional for two or more years, and so avoid the costs of annual leaf production, but such a strategy has a cost in terms of lower carbon dioxide fixation rates (due to factors such as thicker leaf cuticles and leaf shape needed to survive frost and snow) than either annuals or perennials. {Ledig, 1989}
Nutrition. Large areas of the world's soils are nutrient deficient. Nitrogen is one of the most important nutrient requirements for plant growth, and its uptake from the soil is required by all plants which are unable to fix atmospheric nitrogen. Inter-cropping trials with N-fixing species in tropical plantations have shown that it is possible to maintain high yields without the use of nitrogen fertilisers. For example, trials in Hawaii which mixed Albizia with Eucalyptus achieved yields of about 25 t/ha, slightly more than pure, well fertilised Eucalyptus stands. {Debell, 1989}
Nitrogen fertilisers are applied to the world's crops in ever increasing amounts and have been linked to many environmental problems (section 5). Other nutrients (mainly K and P) and trace-elements are required for healthy growth, but these are only required in relatively small quantities which need to be determined for individual sites.
There are many management strategies which can be adopted to minimise the use of fertilisers. These may include intercropping with nitrogen fixing species, and/or the returning of a portion of the crop's residues to the fields. For example, in the case of electricity and alcohol production from sugarcane, many of the non-organic nutrients removed from the fields at harvest (especially K) can be restored by irrigation with "stillage" (the liquid residue from alcohol distillation.) Further potential for nutrient recycling exists via the redistribution of the ashes from the combustion of the bagasse onto the fields. This has many potential environmental benefits (section 5.)
Plants also show considerable variations in nutrient-use efficiency (NUE) and, when conditions are not water or nutrient limited, also show differences in the efficiency with which they convert intercepted PAR into fixed carbon. Important gains in productivity may thus be made through the selection and genetic manipulation of species which are more efficient in their utilisation of resources, or more tolerant to the lack of them. This would allow increased productivities on present cropland and reasonable productivities on land previously considered as nutrient-stressed wastelands, large areas of which are in need of rehabilitation. (table 5)
Pests and Diseases. In general attacks on crops by pests are all too obvious and in common with fungal and bacterial diseases can be highly destructive if preventative measures are not taken early. In these situations it is common practice to spray with the appropriate prescribed pesticide if the farmer can afford them. Casual browsing by deer, rabbits, etc, can be more difficult to control and often needs physical restraints if the crop is not to be lost, especially during the early stages of growth.
Integrated pest management (IPM) strategies which incorporate biological control practices may allow "energy farmers" to minimise pesticide applications with concurrent reductions in pest levels and energy inputs. Such practices rely on integrating many risk abatement and management strategies {Raske & Wickman, 1991}. For example, a reserve area may be maintained in order to ensure that a stable population of predators is present in close proximity to the crop. Any pest outbreak may then theoretically be matched by an increase in the predator population thus minimising pest damage. (section 5.)
For forestry plantations it is now recognised that of equal importance to growth management strategies are strategies designed to make plantations more robust to diseases, pests and drought. Ledig and Kitzmiller {1992} suggest that in the face of environmental uncertainties "reforestation strategies should emphasise conservation, diversification, and broader deployment of species, seed sources and families." This approach is widely followed in Brazil's commercial plantations with a great deal of success. (see Chapter 3.)
Physical (soil & land). Soils are a key determinant of plant growth as they are the medium from which they gain their nutrients, water and physical support. Soils vary in texture, mineral content, pH and the ability to retain water and nutrients which can be, and often are, modified by the vegetation and bacteria growing on and within them.
Buringh {1987} has estimated that after eliminating land areas which are unsuitable for the growth of cereal crops even with the addition of fertilisers and pesticides, about 22% of the earth's surface is capable of sustaining cereal production (present agricultural land comprises about 11% of terrestrial area.) About 1/8 of this potential cereal land is qualitatively estimated to be of low productivity. The remaining (non-potential cropland) 78% of global land area may be capable of supporting crop production of other varieties, or cereal production under different management practices and is of considerable interest for biomass growth.
Harvesting and storage. These factors are important since losses from harvesting and storage can be the same order of magnitude as losses from pests and diseases. It is estimated that about 25% of above ground tree biomass is lost during harvest and transport. Pre-harvest food losses are estimated to be around 35% of the total production, and a further 10-20% is lost after harvest during transport, storage and processing. {Hall, 1984}. Whilst much can be done to increase the efficiencies at which harvesting is carried out through improvements in both management practices and equipment, there are environmental consequences if too much of the vegetative cover is removed from above the soil surface at the wrong time. For sugarcane production it is estimated that about 25% of the tops and leaves should remain on the fields after harvesting to protect the soils from rain and wind erosion and also to maintain organic matter levels of the soil {Carpentieri, 1992}.
Present attempts to predict biomass productivities by both climate-driven and mechanistic models may be useful in estimating future biomass-for-energy scenarios.
These types of models predict net primary productivity (NPP) patterns based on a range of limiting factors of which the ratio of Precipitation (rainfall) to Potential Evapotranspiration (i.e. P/PET) is generally dominant. PET is defined as the potential total amount of water which could be evaporated from the soil plus that transpired from leaves (pores fully open) at given conditions of irradiance, temperature, air movement and air humidity (units: mm H2O/yr. or mmol H2O m-2 s-1.) A P/PET ratio of < 1 over a season or defined growth period implies that H2O is limiting growth, while a ratio > 1 implies that there is excess water which plants and soil cannot absorb and that run-off may occur. PET can be estimated from meteorological and soil data and by using energy-balance measurements which emulate plant canopy dynamics. However, due to differences between plant species, in both rates of transpiration and evaporation; P/PET can only be a rough guide.
The P/PET ratio incorporates both temperature and precipitation patterns to give plant moisture availability profiles; however, these are probably only consistent at country to regional level resolutions and not at the more global levels. For example, a Terrestrial Ecosystem Model (TEM) has recently been devised to predict global C and N fluxes and pool sizes. It has been applied to South America and has shown that on an annual basis availability of moisture was the factor which correlates most strongly with annual productivity (NPP). TEM estimates the mean NPP of the tropical evergreen forest region (natural growth) at 11.7 odt/ha/yr. (oven dry tonnes) which can be compared with recorded local productivity values of forest plantations (unmanaged and managed) of 20-40 odt/ha/yr. {Raich et al., 1991; Brown et al., 1991}. The TEM-derived productivity value was directly compared with the Miami model {Box & Meentemyer, 1991} (climate driven) which was then recalculated using the same parameters used to calibrate the TEM- this gave TEM values about 10% lower than the Miami model with the same spatial distribution of predicted NPP. Thus, despite fundamentally differing in their methodologies both of these approaches to vegetation modelling give similar results.
Future models will need to have much higher resolutions if they are to give useful optimum harvestable biomass predictions at a much smaller scale of thousands of hectares. Resolutions presently used in ecosystem productivity modelling are at the hundreds of thousands of hectares scale, or larger (0.5° grid scales, 50x50 km). {Esser, 1991; Raich et al, 1991} The use of data aggregated from relatively few sites (and for natural vegetation only) in large scale models is a problem. Thus, extrapolating model conclusions to predict NPP values for managed biomass-energy ignores the effect of management practices on increasing biomass yields. Such model predictions are usually too low in their estimates of potential biomass productivities since they cannot yet factor in site-specific or regional determinants of yield limitation- these can only be derived from much more detailed empirical knowledge. Good management practices can overcome limiting factors of nutrient availability, pests, harvesting problems and even moisture availability, and raise NPP substantially on a long-term basis if the practices are sustainable. The challenge is to identify areas where biomass production appears most promising and to adapt the natural ecosystem models for use with biomass energy production when management is applied. Obviously the inputs required to improve site productivity at any given point, will need to be related to the expected returns.{Barros & Novais, 1990}.
A large scale model should also incorporate the effects of human interventions and changing land use systems. None yet exists (to our knowledge) which allows the theoretical limitations of P/PET and other factors to be partially mitigated. Practices such as the use of more water use efficient (WUE) species, intercropping, soil management techniques, both above and below ground, irrigation, etc, can be highly effective and will need to be accounted for in such models.
Plant WUE (measured as g carbon fixed per kg of water transpired by the plant) is highly-variable between plant types and species (Fig. 2), but usually ranges from 3 to 7 g CO2 fixed by photosynthesis (before metabolic losses) per kg H2O, or expressed as t H2O per t final dry matter, usually 500-1000 t per t. Biomass production can be increased without irrigation by selecting species better adapted to water limitations and by management techniques; improvements of 3-5 times in NPP have been recorded at a given moisture regime. An important consideration as atmospheric CO2 levels increase during the next century, is whether this will improve the WUE of plant growth generally or only for certain types of plants e.g. C3 plants. Even though experiments indicate that many plants increase their WUE at high CO2 we do not yet know if this will be applicable to ecosystems or whether this occurs at the field (agronomic) scale.
The growing of biomass over large areas is believed to ameliorate the climate through the recycling of water and nutrients, through water-shed protection, and by providing a more stable microclimate. At present these self perpetuating mechanisms are largely ignored by vegetation models which do not allow for feedback between pixels (remote sensing units). It is these feedback mechanisms which are most affected by land use changes and need to be incorporated into future vegetation models at various scales.
Wastelands & Potential land for Forests.
Land Reclamation Case Studies.
The total land surface on the Earth is just over 13 billion hectares of which about 1/3 is under forest and woodlands, 1/3 under grassland + arable, and the final 1/3 ("other") includes deserts, stony, steep (mountains) and ice-covered land. Increases in cropland have come mainly at the expense of forests and woodlands, with arable land estimated to have occupied 860 Mha in 1882 and 1,477 Mha today (1989) (11.3 % of the world's surface, table 4). At the same time forested land has decreased from 5,200 Mha to 4,087 Mha (31.2% of the world's land area). Buringh {1987} has estimated total potential cropland using productivity constraints for the 10 most commonly grown crops at just over 3 billion ha. Simplistically this means that, at present crop productivities, global food production could be doubled by utilising all this potential arable land. However, such a conclusion is a tenuous extrapolation since past trends in cropland expansion have often been at the expense of woodland, and generally onto less suitable soils. This implies that future yields will decrease as a result of the falling quality of land being brought into production. Furthermore, a substantial proportion of good quality cropland is expected to be lost to non-food producing uses, such as cities and towns. Nevertheless, if present trends in rising productivity were to continue (fig 3) and be applicable to new croplands, then far more than double present food production might be expected. For example, wheat yields in the UK have increased from about 2.5 t/ha/yr. in 1945 to about 7.5 t/ha/yr. in 1987 and are still increasing.
The success of the strong agricultural development programmes in both Europe and North America has led to large agricultural surpluses. The production of these surpluses has, however, proved economically expensive. For example, Wright et al {1992} estimate that payments from modern farm programmes in the US are costing "one and half times net farm income," whilst global agricultural subsidies were estimated to be about US$ 260 billion in 1990. If present policies continue as usual this will climb to US$ 300 billion by 2000. {Economist, 1992}
Significant areas of land presently used for intensive agriculture are not capable of sustaining modern intensive farming techniques, and are thus targeted for removal from arable production. It is estimated in the US that about 30 Mha had been removed from crop production by 1988. In the EC (12 countries) surplus agricultural land resulting from rising yields and changing agricultural subsidies may reach 15 to 20 Mha by the year 2000, and at least 40 Mha into the next century as crop productivities continue to increase.{NSCGP, 1992} These cropland areas are already being removed from intensive farming under EC and US "set-aside" schemes and the US "Cropland Reduction Scheme." {CAST, 1990; Brown L.R., 1992; Hall, 1992}. For practical and social reasons related to the rural economy and environment, this land represents a significant opportunity to initiate biomass energy production schemes, especially if coupled with the use of agricultural and forestry residues.
3 The definition of "wastelands" is highly subjective. Table 5 gives a summary of the definitions given by each author. In general, such estimates try to define areas of land which have been used previously, but are now incapable of sustaining humans.
We consider wastelands to be land presently incapable of sustaining food production. This land has generally been degraded through changing methods of management, often towards more intensive and unsuitable forms of land use. A prime example is fallow land in shifting agriculture, where fallow periods previously lasted 30 or more years and may now be as little as 1 to 5 years.
Estimates of degraded and abandoned land generally lie between 700 and 1,000 Mha which is equivalent to about half the world's present arable land. The extent of this "available" land has led scientists to highlight its potential for use in mitigating the greenhouse effect by managing it to become a carbon sink. Wastelands are regarded as having a good potential for storing carbon in trees due to the relatively low levels of carbon in their soils and vegetation.
At good productivities (6 t Carbon/ha/yr. equivalent to about 12 odt/ha/yr. biomass) the reforestation of all this land could theoretically remove about 5 GtC from the atmosphere per year over the next 40 years, after which the rate of net absorption would decrease due to increasing tree maturity. Present emissions of CO2 to the atmosphere are around 7 to 8 GtC/yr. of which about 3 to 4 GtC appears as an atmospheric build up in the levels of CO2. Sequestration programmes could therefore only be regarded as a temporary measure, buying time until other sustainable forms of energy or permanent CO2 removal systems can be developed. However, productivities on degraded land are at present between 0.1 and 3 t/ha/yr. {ETC, 1992}, (productivities in US commercial forests lie between 1 and 3 t/ha/yr.). Thus, it would require large quantities of inputs in the form of management, fertilisers, pesticides and labour in order to raise the productivity significantly.
It is now increasingly realised that attempts to afforest land areas on the scales required (400-1000 Mha) for reasons aimed purely at absorbing anthropogenically produced CO2 may be misdirected. {Hall et al., 1991; Hall, 1993; NaKicenovic et al, 1993} Social, political and practical limitations to achieving high rates of reforestation are more likely to be overcome if there are concrete social and economic reasons stimulating revegetation at the local level. The development of an indigenous, modern biomass energy infrastructure, and the removal of obstacles such as subsidies to competitive fuels, may provide such a stimulus. (see chapter 6)
Estimates of the Mean Annual Increment (MAI)4 of wood at the global scale are becoming available with some accuracy, however, global estimates should be based on more dissagregated data at the regional and country level. Whilst average potential productivities of about 6 to 12 t/ha/yr. are possible on rehabilitated lands it is questionable whether carbon sequestration by itself would provide sufficient incentive for the wide-scale forest establishment required to reverse atmospheric CO2 increases. In order to achieve high productivities a variety of strategies will need to be employed. In most cases this will require a detailed local knowledge, extension services and continuing monitoring and research.
4 MAI is defined as the net accumulation of above ground biomass through annual plant growth.
Much of the degraded land may be salt-affected, for example, Alpert et al., (1992) estimate that about 950 Mha of saline land exists (table 5). 125 Mha of such land could feasibly be rehabilitated and is not presently used for agriculture or settlements. Massoud {1979} has also estimated that there is about 1,000 Mha of salt-affected lands.
Houghton estimates that there are 850 Mha of degraded lands available for rehabilitation, 350 Mha of which could come from land presently in the fallow cycle of shifting cultivation. Houghton only considered land which had previously been forests & woodlands and is now unused. Thus ignoring the fallow cycle land 500 Mha is theoretically available for immediate use since it is presently "unused." Another estimate from Myers (1989) that 200 Mha needs to be reforested mainly for watershed protection, and a further 100 Mha of wastelands are available, strongly support strategies for rehabilitating degraded lands. Myers states such strategies could have far reaching effects and need to be "carried out for reasons other than the greenhouse effect." {Myers, 1989}. (Table 5)
The estimates of the extent of degraded lands are in the same order of magnitude as the salt-affected lands; it therefore seems likely that some of these lands overlap and that a considerable proportion of the degraded land was abandoned due to rising salinity. High salt levels reduce the levels of nitrogen which is available to the plant, but nitrogen-fixing species are often tolerant to saline soils, and may achieve acceptable productivities on such land.
The potential for the use of saline land is considerable, for example, salt-tolerant plants can attain 3 to 7 tC/ha/yr. with saline water irrigation. However, the use of saline irrigation may only economically realistic up to about 100 m above water level, due to increased costs for irrigation at higher altitudes. It is therefore estimated that only about 125 Mha of the salt affected lands will be of use {Alpert et al, 1992}.
In other areas, in South-West Australia for example, where a saline water-table is close to or at the surface over large areas other remedies have been used to great effect. Due to intensive agricultural management practices which led to the removal of deep rooted vegetation in favour of cereal crops the water-table in this region of Australia rose so inundating the topsoil with saline water. A two pronged approach was used to rehabilitate this land requiring the planting of the non-salt affected watersheds with low water-use-efficient trees and the salt affected regions with saline tolerant tree species. The water table has now been lowered enough (through increases in transpiration rates of these trees) in some areas to treat the topsoil and resume carefully managed crop production.
Recent estimates of land availability have suggested that less land may be available in practice, than previously calculated. Bekkering {1992} has calculated that 385 Mha only of land may be available in a total of eleven tropical countries after allowing for future land requirements for food production to 2025. Whilst this estimate uses the "carrying capacity" model for estimating future land requirements (which does not allow for improvements in productivity,) its predictions may be more accurate than previous global and regional level calculations. NaKicenovic et al. {1993} has attempted to distinguish between "suitable" land for reforestation and land which will actually be "available" for reforestation. He calculates that about 265 Mha is available for global reforestation programmes and a further 85 Mha for agroforestry.
We consider that reasons other than pure Carbon sequestration are necessary if land resources are to become available for revegetation programmes. Thus producing biomass as a substitute for fossil fuels will generate income and at the same time achieve a certain amount of carbon sequestration. Carbon-sequestration only programmes would also be costly. For example, NaKicenovic et al. {1993} estimates that the cost of a global plantation programme to sequester 120 GtC over the period 1995-2095 would be about US$ 520 billion (average cost = US$ 4.4/tC). However, these cost estimates could be an underestimate "of the real costs by a factor of 2 to 3." {NaKicenovic N. et al., 1993}
Severely degraded lands may require more intensive management if they are ultimately to be restored to their former productivity and provide useful outputs besides carbon benefits. The dominant factor affecting the success or failure of land rehabilitation schemes is the intimate involvement of and acceptance by the local inhabitants. The Baringo Fuel and Fodder Project in semi-arid Kenya, {de Groot et al., 1992} is an example of the progress, albeit imperfect, that can be achieved by this approach and the lessons which need to be learnt if significant land rehabilitation on the required scale is possible. Projects such as these must directly involve the local people in the planning and implementation phases. In the BFFP, previously fertile land had been devegetated through over-grazing, leading to severe erosion and desertification over a period of 50 years or more.
Baringo. This project which is based around the Lake Baringo in central Kenya has been running for over 10 years and relies on the use of solar powered electric fences to exclude grazing animals from the fields until they are well established and managed. Over 1,000 ha of fields have been planted with a variety of different tree and grass species which can provide both fuel and fodder. By allowing the vegetation in these fields to regenerate fully, a sustainable supply of fuelwood, grass for fodder (sustaining the livestock at the end of the dry season) and thatching is provided. An ancillary benefit is that these fields also play a role in carbon sequestration and soil stabilisation for the region as a whole.
The Baringo fuel and fodder project is not without its problems but has been successful in halting soil erosion within the fields. It has gained the support of the local population who continue to donate large areas of their land to be included in the project which is returned once revegetated.
Other projects. The KEITA project in Niger which is on the border line of the Sahara is also an example of the gains which are possible once the causes of deforestation are addressed and effective management practices are demonstrated to work. {FAO, 1992} Although the primary aim of KEITA is to halt desertification it recognises that there is little point in stopping desertification if it does not help the interests of the local inhabitants. It reinforces the conclusion that projects which do not involve the needs and aspirations of the inhabitants are probably predestined to failure. There is now so much scepticism about the chances of success of such aid projects that they are often abandoned as soon as the aid agency ceases to oversee the project (see below). In KEITA's case this meant that visible results had to be demonstrated quickly. This superficial requirement for speed initially necessitated the use of heavy machinery to demonstrate the effectiveness of bund building and tree planting in a similar manner to Baringo. Future projects should not need such machinery. Other key factors in the success of the project are the infrastructure which was put in place, thus insuring that any farm produce could get to the market, the agronomic practices which can be achieved without the use of heavy machinery and the recognition of the important role women play in the structure of the community. Despite these "successes" questions still remain about the cost-benefit performance of this project.
In Kerkhof's {1990} study of 19 agroforestry projects in 11 different countries in Africa, he identifies several factors which mediate in the success and failure of these projects. Primarily, the needs and aspirations of the local people must be sought and not assumed, and donor agencies must be willing to adopt long-term and flexible aims. Above all there is the need for local inhabitants to be involved at all levels of decision making, planning and extension, if large amounts of money are not to be wasted.
There are two key questions which need to be addressed:
i) is there sufficient cropland available to produce food for the world's expanding population? and
ii) can biomass energy help enhance development and food production?
As seen above there are significant reserves of potential cropland available, but it appears that these resources are not distributed where they will be needed most if present predictions about the rate of population growth and areas for food production are realised. The IPCC's Response Strategies Working Group (II) {IPCC, 1990) has estimated that the need for cropland will increase in proportion to the World's rising population. Such an increase might require about 50% more cropland to be in production by the year 2025. We have analyzed data from the FAO's "Agriculture Towards 2010" project which assesses the potential cropland resource in over 90 developing countries. Data for China is not yet available, but will be essential for realistic global and regional assessments. We have estimated the potential global land resource based on this sub-set of countries, but without the Chinese data this extrapolation is limited to comparative purposes only. (see table 2b)
The FAO study (AT2010) took into account factors such as water availability, status of soils and the use of inputs such as fertilisers and pesticides. From this data we have calculated likely future cropland areas needed for food production in 2025. By subtracting these cropland requirements from the estimated total potential agricultural land resource for the three major developing regions, Africa, Latin America and Asia (excluding China), the theoretical remaining productive land in 2025 can be estimated (see Table 2). The potential energy production on this "remaining" land is then calculated assuming that such land is capable of yielding 10 air dry t/ha/yr. of biomass (i.e. 150 GJ/ha/yr.).
The area of land under cultivation is predicted to rise from the 706 Mha used in these 91 developing countries to 1059 Mha's or 40% of their potential cropland by 2025. These regional level figures disguise the local level problems which may occur when all the available cropland is already in use. For example, Asia (minus China) is already using 348 Mha and this is forecast to rise to 517 Mha in 2025; however, total potential agricultural land is estimated at only 470 Mha, and thus under these assumptions a deficit of -47 Mha is calculated by 2025. Africa which at present uses only a fifth of its potential cropland would still have 75% of its potential cropland remaining by 2025. Latin America is in an even more favourable situation, presently using only 15% of the cropland resource and 23% by 2025 (see Table 2a).
Asia therefore appears to be most at risk from population increases, being increasingly unable to meet its food requirements at present productivities. Many areas of Asia are densely populated and there seems little room for expansion into an over-utilised cropland. Previous attempts to reconcile this potential shortfall in Asia have centred on the gains it is possible to make through increased irrigation; in fact the area under irrigation has been rising steadily. However, water resources are increasingly limited with severe environmental problems resulting if this resource is overexploited. Withdrawals of water are nearing 20% of total run-off for both Asia and Europe. In 1986,17% of the world's cropland was irrigated, and this is increasing by 0.9% annually. {Hall and House, 1992}
Significantly, at the global level, continued increases in the gross quantities of food production during the 1980's have not been achieved as a result of increases in cropland areas. (fig.9a) The gains in per hectare yields which have made this possible are borne out at the country level. In India, for example, the net sown area has remained virtually constant since the mid-1970's but at the same time total cereal production has risen from about 120,000 t to 200.000 t. (fig. 9b) Despite these apparent improvements there still appears to be a significant potential to raise these yields. (fig.3)
In fact there has been a steady improvement in both the quantity and quality of food produced if inequalities in food distribution and production are ignored. For example, globally, the average per hectare yield of cereals has increased by 20% since 1978-80 to 1990, and is up by 11% in Africa; however, the average yields for roots & tubers has fallen 5% globally, while it has risen by 16% in Africa over the same time period. {WRI, 1992}
Increases in productivity resulting from the selection of crops with enhanced water-use-efficient (WUE) and nutrient-use-efficient species offer the most promise in a resource limited environment. In the case of water, WUE's for C3 plants are in the range of 2-6 mg CO2/g H2O which represents 300-1,000 t water per t biomass or an annual rainfall of 750"-2,500 mm. C4 plants have higher WUE's than most C3 plants, for example, maize requires only about 300 t H2O/t biomass produced.
Significant areas of land are available for rehabilitation. If such land is to make a long term contribution to the reduction in atmospheric CO2-levels the terrestrial carbon inventory5 will need to be raised permanently. Fossil-fuel use increases CO2 levels in the atmosphere and is thus a carbon "source"6. Sustainably grown biomass for energy is nearly CO2-neutral7, depending on production and conversion methods. The production of fuels from annual crops results in lower emissions of net CO2 compared to fossil fuels. {Turhollow & Perlack, 1991} The energy output:input ratios for both annual and perennial biomass energy "crops" are positive - the ratio can vary from just above 1:1 up to 20:1 depending on the system. The amount of CO2 released by the fossil fuels used to power the machinery for growing, harvesting and processing the biofuel is fully accounted for in these ratios. Hence a positive biofuel output fossil fuel input ratio (i.e. > 1) infers that when the biofuel is used as a substitute for fossil fuels it will result in reduced net CO2 emissions.
5 The term carbon inventory is used to describe the total mass of non-atmospheric carbon stored per hectare of land. In general it is practical to consider this carbon as the organically stored carbon held in living and decaying biomass of all types, both above and below ground.A more rigorous definition would include inorganic forms of carbon held in the soils and rocks. Since this form of carbon normally cycles between the soil and the atmosphere very slowly (measured in millennia) it is often ignored in carbon balances. However, if exposed to atmospheric weathering from the rain, wind and sun it may be released far more rapidly than normal. We make no attempt to measure its significance.
6 Any change in land-use which leads to net emissions of carbon dioxide to the atmosphere can be regarded as a "C source."
7 A "Carbon sink" is defined as a process which leads to the net removal of CO2 from the atmosphere. "Carbon neutral" is a process which has no overall effect on the levels of atmospheric CO2. It is essential that the time scales within which they are used are also defined.
If biofuel production is to be regarded as a sink, the total amount of carbon stored per hectare under the bioenergy crop must be greater than the level of stored-C in the vegetation previously on that land e.g. where annual crops or degraded lands are replaced by sustainably grown forestry plantations. Also, where biofuels are used as substitutes for fossil-fuels, the biofuel can be regarded as a "sink" in terms of the avoided CO2 emissions which would have arisen if that energy was derived from fossil fuels.
The size of the carbon sink is dependant on the level of vegetation already present on the land and the time perspective of the newly planted vegetation e.g. the rotation length of a plantation and the likely length of time the land will be used for energy production. {Marland and Marland, 1992.} Increasing both the length of rotation and the productivities will result in higher average levels of standing carbon. {Schroeder, 1992} (also fig. 4) In pure C-sequestration terms the longer the rotation length the higher the average standing stock and therefore the greater the carbon stored. However, for energy production purposes the optimum rotation length may be shorter since the rate of tree growth falls off after a certain age, so reducing the annual productivity and economic return.
The benefits of C-sequestration may thus have to be balanced against those of C-substitution in the selection of the optimum rotation length. We consider that economically greater benefits will be gained from C-substitution since the plantation will produce a valuable commodity; practical rotation lengths will thus be shorter than the optimum for C-sequestration strategies. The costs of C-sequestration only strategies can be extremely high, with projected costs varying from US$ 2 to 56 per tC sequestered. {Moulton & Richards, 1990; NaKicenovic N, et al., 1993; Hall, 1993}. The implementation of these strategies will also require large areas of land, of between 300 to 800 Mha, which may not be available unless useful products are provided to the local populations. (see section 2)
Crops with more than a one year rotation period in effect represent a reserve of standing carbon being grown to replace an annual harvest. A simple example of a 10 year rotation (on 10 ha), assuming a linear growth rate and a yield of 6 t C/ha/yr., shows that the amount of wood permanently being grown to replace 6 harvested tC/yr. = 6 tC (1 year old) + 12 tC (2 years old) + 18 tC. = 330 tC on the 10 ha. Thus, for one hectare to be harvested every year, a plantation with a 10 year rotation requires a minimum total area of 10 hectares. Once established, a plantation achieving a growth rate of 12 tC/ha/yr. would have an average standing stock of 66 tC/ha immediately prior to harvest. (fig 4)
The most clear benefit, in C-sequestration terms, would be if the plantation was established in a desert with an existing standing stock of almost 0 tC/ha. Both the above and below ground levels of carbon would be raised significantly compared to the existing level. However, if the standing stock of the previous vegetation was greater than that of the new plantation then a net reduction in the standing-C levels would result. If, the plantation biomass were to be used as a fossil fuel substitute (or long-lived product), a net reduction to atmospheric CO2 emissions would occur after a given time period. At higher levels of vegetation, prior to the plantation, the longer it would take for fuel-substitution benefits to redress the amount of CO2 initially released. This results from the relatively large amounts of CO2 which are released during the initial clearance through harvesting and transport losses. {Marland & Marland, 1992}.
The time perspective is also important. Longer rotation periods allow greater average plantation standing stocks. Of equal importance to the rotation length are the productivities gained. Higher yields result in higher average standing stocks, and also reduced unit costs. Figure 5 shows the average cost of plantation-derived fuelwood from the 5 bioclimatic regions of Northeast Brazil. The cost of the wood is clearly related to the yield, with yields below about 8 t/ha/yr. proving relatively more costly. The Brazilian study is discussed in more detail in section 3.
Marland and Marland have explored the use of plantations for a fossil-fuel substitution plantation-based model (1992). They conclude that the three most important factors in assessing whether biomass energy plantations are effective C sinks are: i) the C inventory of the natural vegetation, ii) the productivity of the plantation and iii) the time perspective adopted. In areas with high standing stock carbon (e.g. old natural growth forests) on low productivity land, the most favourable solution is simply to leave alone and protect the existing forest. The most effective plantations would be those which are established on sparsely vegetated land capable of high productivities, i.e. degraded lands and present good quality arable cropland (see above). The Marland and Marland model is mainly designed for US forestry conditions i.e. 40 + year rotations, regarding coal as the primary fuel-substitution feedstock. Assumptions about the efficiency with which biomass-based fuels can substitute for fossil fuels influence the rate at which plantations can recover the carbon released to the atmosphere during initial tree harvest, haulage and storage. In their model, these parameters are set at 0.75 tC (coal-derived) substituted by 1 tC from biomass, and 0.375 tC for liquid fuels per tonne biomass C. Their coal substitution parameter assumes that biomass would be converted to useful energy at 60% the thermal efficiency of coal (the present average).
However, if the biomass is used in efficient domestic appliances or with advanced conversion processes then the ratio of fossil-derived C substituted by biomass derived C would improve (see section 4.). Using the more advanced technologies, the amount of useful energy obtained per kg of biomass-C (which is relatively more thermochemically reactive) would be equal to or greater than the amount of energy per kg of coal-C. Effectively higher yields could be gained, and the substitution ratio improved, if more of the forestry residues arising at harvest could be used. However, it is unlikely for environmental reasons that the proportion of residues left on the field should be reduced drastically. In the field, these residues decompose releasing CO2 without providing any useful energy. Thus, even at relatively higher efficiencies of biomass conversion, one unit of biomass-C may still substitute for less than 1 unit of fossil-fuel carbon when the decomposition of residues is accounted for.
Extrapolation from this model can provide a large variation in results. It does however, show that at timescales of less than 30 years, only sparsely vegetated areas capable of high yields should be considered on pure carbon-sequestration terms. These assumptions are based predominantly on the lifetime of the plantation, and not on the rotation length, decreases in which might effectively increase the yield (for certain species), but decrease the average standing stock.
Therefore the most important points for the optimisation of fossil fuel substitution by biomass are: i) increasing the energy output:input ratio i.e. raising the ability of biomass-derived fuels to substitute for fossil fuels; and, ii) raising the level of Carbon held in the biotic pool in order to absorb some of the carbon already emitted by fossil fuel use i.e. increasing the terrestrial carbon sink.
Accounting for soil carbon would make bioenergy plantations even more favourable as carbon sinks, but detailed data is not yet available on the interaction between changing uses of land and soil-C levels.
Energy ratios allow planners to determine whether the agricultural production of a crop is a net energy producer or consumer. Balance sheets detailing the energy inputs must take the whole life-cycle of the energy crops into account since investments in machinery and infrastructure will last more than one crop growth cycle. Detailed assessments of energy inputs would therefore include not only the fossil fuel inputs for fertiliser production and the machinery used for ploughing, planting, harvesting, storage and transport, but also the energy required for the manufacture of the machinery and infrastructure required for modern agriculture. Likewise outputs should include not only the energy return from the crop itself, but also the energy content of the residue production, e.g. straw, husks, shells, stalks, etc. With intensive annual crops, lower output to input ratios can be expected due to the high level of inputs. However, such crops fit well into modern energy and agricultural markets and generally require less land per tonne of produce as a result of higher yields than perennial or woody crops.
Forestry plantation biomass production has been shown to have positive energy output to input ratio's of about 10 or more. In fact, Ledig {1981} has pointed out that for biomass plantations, increases in energy inputs are rewarded with net increases in energy outputs. The use of primary forest land would give a positive energy ratio. However, at higher levels of initial standing stock (e.g. tropical forest) the loss in carbon to the atmosphere during clearance will probably not be recovered in terms of fossil fuel substitution benefits at timescales of less than 100 + years. Thus primary forests should not be considered as sources of biomass fuels.
For liquid fuels from biomass much smaller and sometimes negative energy input: output ratios are cited. For example, a study for the European Bureau for the Environment {Taschner, 1991} gives negative ratios for the production of ethanol from potato, wheat and sugarbeet, showing positive balances only when residues are accounted for. Integrated management approaches are essential if the production of such biofuels are to help ameliorate environmental problems and not increase them. It is therefore essential to adopt management strategies which maximise energy efficiency without compromising soil fertility.
To date very little research has been directed at maximising the production of the whole plant i.e. residue production has previously been regarded as a waste-disposal problem and so much of the breeding work has been directed at raising the crop harvest index. Alexander {1985} states that if sugar breeders were to concentrate on optimising overall productivity in sugarcane instead of maximising the sugar concentration in the stem, then large overall gains are possible in both sugar production, and gross yield per hectare. Sugarcane alcohol plantations presently provide a positive energy ratio of about 4 for the Triangle programme in Zimbabwe and an average of 5.9 for the Brazilian Proalcool programme, which rises to 8.2 under the best conditions. {Scurlock et al., 1991; Goldemberg et al., 1992}
