Minnesotans For Sustainability©
Sustainable Society: A society that balances the environment, other life forms, and human interactions over an indefinite time period.
Can Renewable Energy Sources Sustain Affluent Society?
F. E. Trainer*
University of New South Wales
Figures commonly quoted on costs of generating energy from renewable sources can give the impression that it will be possible to switch to renewables as the foundation for the continuation of industrial societies with high material living standards. Although renewable energy must be the sole source in a sustainable society, major difficulties become evident when conversions, storage and supply for high latitudes are considered. It is concluded that renewable energy sources will not be able to sustain present rich world levels of energy use and that a sustainable world order must be based on acceptance of much lower per capita levels of energy use, much lower living standards and a zero growth economy.
Commonly stated figures for the cost of energy produced from renewable sources are quite impressive. For example it is claimed that electricity can be generated from the wind at 6¢ per kilowatt hour (Diesendorf, 1992, p 7). Given that the cost of coal fired electricity in the late 1980s was in the region of 3.5¢ per kilowatt hour (ELCOM, 1991, p 35) such figures appear to show that renewable energy sources will soon be able to replace fossil fuel sources. However there are a number of factors casting a quite different light on the potential of renewable energy sources. This discussion is mainly concerned with solar energy although some reference to other renewable sources is included, and it is confined to the two energy forms most crucial for industrial conserver societies, electricity and liquid fuel.
The commonly quoted figures obviously derive from trials carried out at the sites most likely to yield optimum performance. The best wind sites lie at latitudes around 45° from the equator. There are few desirable sites in the state of NSW. Modelling of solar thermal systems in Australia (Mills, 1992) has been based on Longreach and Wagga where mean daily solar radiation (global measure). is 6.03 kWh and 5.03 kWh per square metre [m2] per day respectively. The White Cliffs solar thermal electric system was built in a region where the average solar radiation is 6.5 kWh/m2/day, where intensity can exceed 1000 W/m2 for parts of the day and can actually exceed 700 W/f2 for 10 hours on a summer day (Kaneff, 1991, Figure 3). However these three sites are 25°, 35° and 30° from the equator respectively. Most of the people who live in the rich countries are 45° to 50° from the equator and average about half the Australian annual solar radiation, with winter levels far below annual averages (see below). The prospects for such countries basing high material living standards on renewable energy sources are quite different from impressions that might be gained from ideal experimental sites 25-30° closer to the equator. Energy would either have to be collected from much lower intensities of solar radiation, or collected far away and transported at a high energy loss.
The two crucial factors for this discussion are first the present and probable costs of the most likely renewable energy sources, and second their capacity to supply electricity and liquid fuels to rich countries in high latitudes, after heavy losses due to conversion, storage and transport have been taken into account.
The White Cliffs and similar solar thermal systems have been able to convert about 10-12% of the solar energy they receive into electricity. The Solar 1 central receiver system achieved 4.1% to 5.8% in three years of operation (Radosevich and Skinrod, 1989, p 146). The Tennant Creek dish under construction is anticipated to achieve 16.7% (Kanef, 1989).
A major problem for solar thermal power generation (as distinct from process heat) is that the intensity of the solar radiation is critical. A large amount of energy might be received per day but it must be above a certain value in watts per metre before electrical power can be generated. At White Cliffs the high summer level 1000 W/m2, produced 28 kW, but at 300 W/W/m2 electricity output would have been virtually zero. Tokyo, only 36° from the equator, and almost the whole of France, average less than 300 W/m2 for most of the four months of winter. For the White Cliffs system the critical level seemed to be in the region of 700 W/W/m2. Even at this level output was only around 50% of the maximum operating output (Kaneff, 1991, Figure 82). At this favourable site only about 80-85% of summer solar energy incidence was at a level greater than 700 W/m2 and in winter the figure fell to 60-75% (Kaneff, 1991, Table XV-A).
The output for the White Cliffs plant at 400 W/W/m2 was only 16% of peak capacity (Kaneff, 1991, p 178). Systems presently under development are projected to be capable of commencing operations at 400 W/W/m2 (Morrison, 1993; Luzzi, 1993). In Sydney in winter solar radiation exceeds 400 W/W/m2 for only two hours a day, and in Brisbane, 27° from the equator, for only four hours.
When data on solar radiation received at different latitudes throughout the world (University of Lowell Photovoltaic Program, 1991) is examined it appears that at most latitudes higher than 30° average midwinter levels of radiation received are below 3 kWb/m2/day. These figures suggest that in general solar thermal electrical systems are not likely to function very effectively further than 25° from the equator, except for some regions such as Central Australia. They could be useful further from the equator for much of the time if considerable quantities of gas or other fossil fuels were used as back up.
Even at Tennant Creek where annual insulation levels average 6.5 kWh/m2/day back up fuel is expected to have to provide half the electrical energy delivered from the solar thermal plant. (Northern Territory Power and Water Authority 1994, p 5). The Luz systems in favourable sites have been quite dependent on gas for back up (De Laquil, 1993, p 3225). However in the longer term, probably well before 2050, little use of gas or other fossil fuels can be assumed.
The cost per metre of the reflectors and their supporting structures required to collect solar energy for central receiver solar thermal electricity generation at present appears to be in the region of at least US$ 250/m2 and total plant costs including the receiver, heat transfer equipment, engine and generator etc, approximately double the collection cost (ELCOM, 1991, p 37, Luzzi, 1993; De Laquil, 1993, p 284). The collector cost for the White Cliffs project in the late 1980s was US$500/m2 (US$1986) (Kaneff, 1992, p 26) and total plant cost was three times the cost of the collector (Kaneff, 1991, p 197). Kaneff estimates that for the Tennant Creek dish under construction plant costs will be 2.5 times collector cost (Kaneff, 1989, pp. 16-5). The estimated total system construction cost for this project is US$ 758/m2. Of this, 84% is due to the cost of materials, which is not likely to undergo dramatic reduction in future. Kaneff estimates that for commercial models the total cost will be US$400/m2. The mid-1980s trials of the Luz solar thermal system involved total plant costs of US$ 540/m2 (Mills 1992). Luz 3 system costs are US$ 600/m2 excluding storage (Luzzi, 1993). The estimated collector cost for the proposed Sydney University solar thermal trough system is US$ 250/m2 (Mills, 1993b). A planned US 100 MW solar thermal system will have a total plant cost of US$ 750/m2 (Pacific Power, 1993). Luzzi puts current heliostat collector costs at US$ 268/m2 (1993).
The foregoing costs refer to steel and/or glass reflectors and their support. More recently reflectors made from stretched plastic membranes curved by vacuum have been introduced. Their present cost is in the region of US$ 138/m2 (Hagen and Kaneff, 1991, pp. 4-12; De Laquil, 1993, p 247). However, they set problems regarding storm damage and cleaning and their lifetime is likely to be only five years, which means that over the lifetime of a plant they would be unlikely to cost less than steel and glass reflectors.
The energy efficiency of conversion of solar radiation to-electricity in the White Cliffs project was 9.1%. The efficiency of the 80 MW Luz 2 system is approximately 12.5% (derived from Mills, 1992, p 27). However, Kaneff claims that double this efficiency might eventually be achieved in very large systems. Kaneff expects the efficiency of the Tennant Creek dish to be 16.7%. Johansson et al conclude that solar thermal systems are around 13% efficient but that 17% is a future possibility (1993, p 21; see also p 254).
An important factor related to the efficiency of solar thermal systems is the start up time. For Solar 1 normal generation was reached only two to three hours after sunrise due to the need for the system to heat up, and after the passage of a cloud it could take 45 minutes to return to normal operation (De Laquil, 1993, p 244).
This means that figures on the kilowatt-hours of solar energy at a site per metre per day will be misleadingly high indicators of the electricity that can be delivered.
In the discussion that follows US$ 500/m2 has been assumed as the plant cost and a relatively high figure of 20% has been assumed for the efficiency of solar thermal, electricity generation.
There is considerable confusion and ambiguity involved in claims of this kind. The most spectacular claims refer only to the possible cost of the material which converts sunlight into electricity and does not include the many other essential elements in a complete system. A common claim is that eventually the cost of PV technology will only be limited by the cost of the glass, because the expected technical advance will result in negligible cost for the active materials and the production of the cells. Even if this assumption is accepted the real cost of a complete PV electricity supply system would be considerable because of the many components and costs in addition to the cells.
The cost of glass assumed in some estimates is as low as US$ 5.6/m2 (Kelly, 1993, p 311). However the present wholesale cost of glass to large volume Australian solar manufacturers is US$ 36/m2 (Solarex, 1994) and it is not obvious that this and other materials costs could be expected to fall markedly in future.
The present cost of the `modules' which house the active cells and adds the cost of the frame, wiring and labour is approximately US$ 200/m2 (Kelly, 1993, p 304). Other estimates put finished module costs from three to seven times the cost of the actual cell material at present (Green, 1993, p 348). Ogden and Williams (1989) estimate that in large-scale future production, module cost would be 2.5 times the cost of the materials in the cells and equal to US$ 225 per metre. Kelly says module costs are about twice the present cost of the cells they house (1993, p 300). Profit margins to producers would have to be added at this point. Retail costs of PV modules add 50-75% of factory costs (Kelly, 1993, p 322).
The most important costs are not to do with the cells but with their assembly into modules and the 'balance of the system', i.e. setting up the modules at the power generation site. For systems now being completed the balance of system cost seems to be more than the cost of the modules (Kelly, 1993, p 300). For a small system recently installed by the Australian Northern Territory Water and Power Authority balance of system costs were approximately A$ 720/m2 (US$ 526) (Presnell, 1994). For two systems installed by the Australian Snowy Mountains Authority in the early 1990s the balance of system cost was A$ 1260/m2 (although the unusual site would indicate that this might be an unrepresentative figure) (Green, 1994). These costs are A$ 6 and A$ 10.5 per watt respectively. Chianimba et al (1991) estimate support systems alone at £100/m2 (US$ 160/m2). According to De Laquil (1993, p 312) the balance of system cost will probably fall to US$ 400-500/m2. These figures are for fixed orientation mounting. Tracking systems would involve much more expensive systems such as those used for power-tower heliostats. For large-scale generation involving millions of square metres the cost of land would have to be added.
It is not obvious that these balance of system costs can be expected to fall dramatically in future, since the processes involved are relatively simple and not likely to undergo rapid technical advance similar to that being made in cell production. However, the development of PV roofing materials could make a significant difference. Little support is given for the assumption that balance of system costs will fall to the region of US$ 50/m2 (reported by Ogden and Williams, 1989, p 34, and Kelly, 1993, p 325). Some components of the balance of system cost have been quite resistant to efforts to reduce them (Kelly, 1993, p 313).
Adding these elements loosely indicates that the installed cost of PV systems (excluding storage) is now over US$ 1000/m2. Even if the cells became costless total system costs would still probably be in the region of US$ 600-700/m2. Assuming 15% efficiency and cells in ideal conditions this comes to US$ 4-4.6/W of peak capacity. Discussions of possible photovoltaic electricity costs often fail to make clear that balance of system costs have not been taken into account and that the reference is only to the cost of producing cells.
Another factor easily overlooked when comparing PV costs with those for coal-fired electricity is that the latter often include distribution costs and the former usually do not. Transmission and distribution costs add to a little less than generation costs (Electricity Supply Association, 1995).
The efficiency of PV cells assumed most of the following examples will be 20%. At present most cells in general use probably average around 12% efficiency (Kelly, 1993). Figures achieved in laboratory tests can be misleading as these usually do not involve glass Protecting covers, accumulated dust or losses due to heating. These factors typically result in performance in the field that is some 1.0-30% below rated capacity (Kelly, 1993, p 324).
At the favourable Sydney latitude of 340 South a domestic 12 volt lighting system using three panels each of 54 watts and a 320 ah lead-acid battery delivers electricity at approximately 90 cent/kWh (assuming a lifetime of fifteen years for the panels and three years for the battery (batteries with much longer lifetimes are available, but at higher cost). These panels are rated at 162 watts, which might lead the uninitiated to expect an output of 13.5 amps in full sun. However in winter their output averages only 5.5 amps over approximately 7 hours, even on a completely cloudless day, equivalent to a 1.72 amp flow for 24 hours, i.e. only 12% of rated capacity. Taking cloud into account would probably reduce this figure by 30%.
These figures refer to a system which tracks the sun, i.e. faces it continually at right angles, whereas most photovoltaic systems are fixed, which would reduce total energy collected over a day by approximately 30%. Therefore on a typical winter day in Sydney a house with a normal arrangement of fixed panels would probably collect electrical energy at a rate averaged over 24 hours of only 6% of its peak capacity. If the energy loss from battery storage is assumed at 30% (40% is more likely) the overall energy efficiency figure falls to 4.2%.
Therefore in order to deliver energy at a rate of 1 kW in Sydney in winter, panels with a peak collection capacity of 24 kW would be needed. In other words a capacity of 1 kW will provide an average of 1 kWh/day. The 1990 NSW household average electricity consumption was 2.7 kW, i.e. 40.8 kWh/day (Office of Energy; 1992, p 17). To provide this via PV panels and batteries would therefore require panels of 40.8 kW peak capacity. At the present retail price, US$ 10/W, this capacity would cost US$ 408,000. If 20% efficiency and US$ 5/W are assumed, the cost would be US$ 112,000.
If a 15-year life for the panels is assumed the annual cost of electricity per household would be in the region of US$ 27,000 ignoring the cost of batteries. These figures make no provision for a succession of cloudy days. Whenever one sunny day is followed by a cloudy day collection and storage capacity on the sunny day would have to be doubled if energy is to be available on the following day. To cope with several cloudy days in a row either capacity would have to be multiplied accordingly or resort made to external supply.
Solar powered family vehicles charged from panels on the house roof would approximately double the magnitude of the problem. If the, car were to travel 50 km per day [31 miles] it would require approximately the equivalent of 4 litres of petrol [ca 1.1 gallon], at 34 MJ/l, i.e. 38 kWh equivalent in electrical energy (ignoring probable improvements in automobile energy efficiency and the probable 40% energy loss in battery storage). From above this would require a peak collection capacity of 38 kW, i.e. 350 metres of collection area (at 108 W/m2, i.e. 11% efficiency).
In other words 'fuel' for the car would require purchase of 700 solar panels at a cost of approximately US$ 350,000 or US$ 23,300/year, assuming a 15-year lifetime. This is in the order of 25 times the present annual outlay on fuel for a car (ignoring battery costs). Note that if the cars were to be battery powered two sets of batteries would be needed, one in the car and one at home being charged during the day. Because of either battery weight or heavy containers for hydrogen, caution should be exercised regarding assumed potential energy efficiency of cars powered by either energy source. (It is not being assumed that private panels plus lead-acid batteries is the best way to power vehicles by renewable energy.)
The cost of photovoltaic cells is likely to fall considerably in future years
and the energy efficiency of cars is likely to increase, conceivably trebling,
but the magnitudes of the above figures indicate how far technical advance would
have to proceed before this source became capable of replacing conventional
energy sources for common purposes at a reasonable cost, even in the highest
It is evident therefore that in most of the countries where energy use is high solar renewable sources cannot make a major contribution unless provision is made for large-scale storage and/or transport of energy. To use renewable sources without provision for storage would mean duplicating plant, so that non-renewable sources can be turned on when the renewables are not available.
The first problematic implication of the need to store energy is the fact that much more generating capacity is needed than might have at first been thought. If the sun is shining for only 8 of the 24 hours in a day then in those 8 hours a solar system must not only collect enough energy to meet demand during that time but must be large enough to collect sufficient energy to meet demand for the other 16 hours as well. This would require construction of a peak generating capacity 3 times as large as the coal fired capacity necessary to meet the 24-hour demand.
. The problem is made worse by the fact that in winter in high latitudes energy demand can be considerably higher than summer demand. In England a typical winter daily pattern of demand for power is twice as high as in summer (Vincent, 1984, p 13; Palz, 1978, p 241 comes to much the same conclusion for the USA in general).
This has merely been the problem of storing energy collected in the daytime for use at night. 'Interseasonal' storage is a far more difficult problem in those many regions of high population where there is little solar radiation, large quantities of solar energy would have to be stored from summer to winter, or transported long distances daily from lower latitudes.
The second and much more problematic implication of storage concerns the large losses of energy involved in converting energy into storable or transportable form and in the regeneration of electricity from these forms. This in effect greatly increases the collection area required to achieve a continual supply of electrical energy. The consequence of various storage options will now be considered.
For the White Cliffs system the average energy efficiencies for the main steps in the process were as follows (Kaneff; 1991, Figure 78):
(1) heat supply to engine from collectors = 0.62
The overall effect of these combined efficiencies was a lead-acid storage and conversion system with an energy efficiency of around 6%.
This figure means that a system capable of supplying 800 MW for 16 hours, i.e. 12.8 million kWh, after storage in lead-acid batteries would need to collect about 17 times as much solar energy in the first place, i.e. approximately 220 million kWh. If we assume solar radiation at 5 kWh/m2/day for a common but less than ideal site, the collection area would have to be 44 million m2. The additional area necessary to supply 1000 MW without storage over the 8 daylight hours would be 17 million m2. Total plant area would therefore be approximately 70 million m2 and its construction cost assuming US$ 500/m2 would be US$ 35 billion, 15 times the cost of constructing and providing, the fuel for equivalent coal fired electricity supply capacity (see below). In other words the low energy efficiency of storage results in the need for large collection areas and costs.
Assuming an approximate cost of 9¢ per Wh of lead-acid battery storage (the early 1990 retail cost of solar home lighting batteries), it would cost US$ 1.152 billion for the capacity to store 12.8 million kWh. However the lifetime of lead-acid batteries can be as little as three years. If five years is assumed the storage cost over a 20-year plant lifetime would add US$ 4.6 billion to the cost of the system.
Estimates of potentially recoverable lead resources indicate that it would not be possible to provide large-scale storage capacity using lead-acid batteries. The US Geological Survey (Erickson, 1973) estimates that the amount of lead potentially economically recoverable from the top 1 km of the continental crust is 550 million tonnes. Gordon et al (1987) estimate that all ore deposits in the top 4.6km of the continental crust contain 2.5 billion tonnes of lead. A very high proportion of this material is likely to remain undiscovered or lie in locations that are too difficult to mine (e.g. under cities or Antarctica) or in deposits too small to mine economically. (Most ore deposits are quite small.)
To provide domestic lighting, around 4% of household energy use (and less than 0.4% of all energy use) might require batteries containing 50 kg of lead per household. By 2060 there will probably be approximately 2 billion households in the world, given the UN median world population projection. The total amount of lead required would be 100 million tonnes.
If all the people expected to inhabit the earth late next century, 11 billion, were to rise to the present rich world per capita lead consumption this would total 2.5 billion tonnes in 40 years, even without any provision for large-scale electrical storage (Trainer, 1985, p 59).
A world of 11 billion people living at current US 'living standards' with
approximately one car per two people, each powered by 1 tonne of lead batteries,
would require around 5 billion tonnes of lead.
Modelling studies have concluded that operation even in highly favourable regions would be optimal when 7.5-18% of the energy came from the use of gas (Mills, 1992). The Luz systems were similarly dependent on gas (Johansson et al, 1993, p.225). These back-up energy costs are not negligible. If delivering 1 kW of electricity involves 0.075-0.18 kW supplied via gas turbines, then gas equivalent to at least 0.15-0.36 kW would have to be used, because the efficiency of converting gas to electricity is only likely to be in the region of 50%. In other words for each unit of electrical energy produced, up to one-third of a unit of energy in the form of gas would be needed
In the longer term a sustainable energy system cannot rely very heavily on fossil fuels such as natural gas. Australia's estimated economic and subeconomic natural gas resources would last 100 years at present rates of use. However, the rate is increasing rapidly, indeed by 9.6% in one year, 1989 (ABARE, 1991, p 6). Gas contributes to the greenhouse effect and potentially recoverable resources are almost as limited as petroleum. Both are likely to become quite scarce in the middle decades of next century.
The above mentioned theoretical study was based on modelling, not empirical observation and assumed a 38% efficiency of conversion of stored heat to electrical energy (Mills, 1992, p 15). This would seem to be implausibly high, given that the typical efficiency achieved in conventional coal-fired power stations is around 35% despite there being no heat losses from rock masses in storage as there would be in hot rock storage, and the fact that it is difficult to supply heat at above 3500C, compared with the 500-6000C in coal-fired power generation. As has been noted above, for the White Cliffs system the efficiency of the engine run by (high temperature) solar heat was only 22%, and of the generator 92% indicating a 20% overall efficiency of conversion of heat to electricity, there being no storage of heat and, therefore no losses from storage. Losses in the transport of heat from rock mass to boiler would also be involved but have been ignored here.
Hot rock storage systems are likely to make an important contribution to electricity supply in regions of relatively high solar radiation. Energy for space heating can be stored effectively from summer to winter in crushed rock or water masses. However, there is no realistic prospect of interseasonal storage in rock masses for electricity in view of the huge volume of storage capacity that would be required. One tonne of crushed rock will store 7 kWh in a retrievable form (Mills, 1993). To store present rich world per-capita energy use, 7.3 kW, for three months would involve storing 15,763 kWh, requiring 2,252 tonnes of rock or approximately 10,000 tonnes per household.
The necessary space would be approximately six times that of a normal house. These figures do not take into account the fact that a high proportion of energy used is in the forms of electricity and liquid fuel, and conversion of the stored heat to these forms would involve heavy energy losses, in effect trebling or quadrupling the volume of storage needed for these purposes.
Even daily storage of large quantities of energy is quite problematic. If a plant averaging 1000 MW were to store half its daily output to meet demand when the sun was not shining strongly, i.e. 12 million kWh, and the efficiency of generation is 35%, then 343 million kWh of heat would have to be stored (ignoring heat losses in storage and transport). At 7 kWh per tonne, 5 million tonnes of rock would be needed, probably requiring a structure 1 km long, 100 m high and 100 m wide when provision is made for air flow channels. At least formidable costs would be involved.
The construction of artificial ponds on a significant scale would involve very large earth works, given the relatively low efficiency of this method of delivering heat. Only about 15% of the incoming energy is stored in the salt water at the bottom of the pond. In one Dead Sea project a 250,000 m2 pond was able to deliver 5000 kW of electricity, only 1/50 of the energy falling on the pond surface (Rabl, 1990a, p 35; Rabl 1990b, p 90).
Applying this fraction to the approximately 3 kWh/m2/ day of solar energy received in Sydney in winter indicates that to meet the average per capita electricity demand for all purposes, approximately equivalent to 30 kWh per person per day (Office of Energy; 1992, p 17), a pond area of 500 m2 would be needed. For a family the associated pond area would be 25 times the area of a typical house.
A further problem concerns the need to prevent salt from getting into surrounding soils and water tables. Artificial systems must have costly and fallible plastic liners.
Cost figures deriving from trials in small ponds can also be deceptive because such ponds might not require extensive piping to withdraw and return hot water. In other than small ponds a network of pipes is necessary to ensure that strong flows of incoming and outgoing water do not set up vertical currents that will destroy the salt gradient.
A major drawback with solar ponds is their high consumption of water, which De Laquil estimates at 35 times the quantity used by a coal fired power station of the same output (1993, p 273). This is partly due to evaporation and partly to the need to draw off water in order to restore the salt gradient. According to De Laquil, for this reason alone solar ponds "are unlikely to generate large amounts of electricity" (1993, p 289).
Again it is evident that impressive figures on costs from specific trials might be far from sufficient to establish the viability of a renewable system. The estimated cost of electricity from the pond Rabl discusses was 10-15 cents/kWh (in the late 1970s). The cost of heat has more recently been estimated at only 2 cents/kWh (Hull and Nielsen, 1990, p 503). When considerations of space, let alone availability of suitable high radiation sites, are taken into account solar ponds are unlikely to provide more than quite small proportions of the total energy demand of an affluent world
In some situations energy to be stored might be transported to existing hydroelectric sites and used to pump water back up into dams. This would rule out use of the sea as the low dam and it would require construction of large volume low dams to hold the water until it could be pumped. Relatively fiat countries such as the UK would be unlikely to have sufficient dam capacity for this strategy to make a significant contribution.
The cost of this form of storage has been estimated at US$ 200/kWh, approximately 2.2 times that of lead-acid battery storage assumed above (Skylas-Kazakis et al, 1991, p 6). These figures indicate that if the batteries lasted five years the storage capacity needed for a 1000 MW power station would add US$ 18 billion to its cost.
Although vanadium is about 10 times as abundant in the earth's crust as lead it would seem to be too limited to enable this form of storage to provide current rich world living standards to large numbers of people. Consider the following very approximate estimate for daily storage. (Storage in summer for winter use would be a far bigger problem.)
Rich world per capita energy consumption is around 7.3 kW, or 175 kWh per day. If one-third of this must be available after storage at 87% efficiency then 67 kWh must be stored each day (ignoring losses in reproduction of AC electricity). Approximately 2 kg of vanadium are needed to store 1 kWh (Skylas-Kazakis, 1991). For a world of 11 billion people living as rich world people do now 3.7 billion tonnes of vanadium would have to be in use at any one time.
The US Bureau of Mines states that demonstrated world recoverable resources of vanadium total only 69 million tonnes (USBM, 1985, p 898). The approach taken by Gordon and Skinner (1987) to the estimation of potentially recoverable mineral resources indicates that only 1.8 billion tonnes of vanadium exist in all ore deposits within the top 4.6 km of the continental crust. (This figure is derived from their analysis for copper, indicating 900 million tonnes in all deposits, and the fact that the crustal abundance of vanadium is twice that of copper. The amount in deposits is proportional to crustal abundance. (see, McLaren and Skinner, 1987.)
As Gordon and Skinner discuss, it is possible to process minerals from common rock after deposits have been exhausted, but only at an extreme energy cost. There are a number of reasons why only a small fraction of potentially recoverable resources, probably less than 10%, is likely to be discovered, to be in accessible sites and to be in deposits large enough to warrant the construction of a mine.
Another significant problem concerns the low energy to weight ratio of the liquid; 5 kg of vanadium pentoxide are needed to store 1 kWh. But the energy content of 1 kg of petrol is 13 kWh. The energy per kilogram in the vanadium battery would be about 1/65 that in an equivalent volume of petrol. This would have major cost implications for the construction and performance of vehicles, and for tank capacities required for interseasonal storage. For example the winter energy use in one rich world household would require storage capacity of 355 tonnes. The equivalent energy in the form of coal would weigh approximately 4 tonnes. For a city of 3.5 million a three-month winter storage would require approximately the storage volume of 1000 large oil refineries, each with 60 tanks 15 m high and 20 m in diameter.
If the energy efficiency of generating electricity from photovoltaic cells is 15%, and of using the electricity to produce hydrogen gas from the electrolysis of water is 70% (Ogden and Nitsch, 1993, p 935), the energy efficiency of the process will be around 10%. The stored hydrogen gas can be used to produce DC electricity via a fuel cell at an efficiency of 40%. Inversion to AC would then be necessary. This means that for each kilowatt-hour of solar energy collected only 3.6% of 1 kWh will be available again as electricity after storage. In other words to have 1 kWh available as electricity after energy storage 28 times as much energy must be collected, and 28 times as much generating capacity must be constructed compared with the amount of coal-fired capacity that would meet the need. Storage in the form of liquid hydrogen would involve a further approximately 30% energy loss (20% efficient PV cells are assumed below, indicating an overall energy efficiency of 4.8% for the above system based on hydrogen gas).
A potential concern with the use of hydrogen as a major energy source arises from the fact that when hydrogen is burned it forms water vapour which enters the atmosphere. Water vapour is an important greenhouse gas. The volumes likely to issue from a hydrogen based economy might not be large compared with natural evaporation but this does not mean that they could not make a critical difference to the balance of the global climatic system.
A major problem would be where to store the large quantities of gas required. Given that 1 m3 of gas can store 1.54 kWh, to store the daily energy needed when the sun is not shining (e.g. 800 MW x 17 hours) would require storage of 15 million m3 of gas, i.e. a mine shaft 1500 km long (taking into account the 60% energy efficiency of storage). Compression of the gas would reduce the necessary volume but the proximity of sufficient storage capacity to areas of generation or demand would be problematic:
Given the assumptions and conditions arrived at in the foregoing discussion it is possible to derive approximate implications for the supply of electricity in different locations, especially concerning costs after losses due to energy conversion, storage and transport. First let us consider electricity supply for Sydney, 340 South.
Sydney has unusually high solar radiation for its latitude. However for the three winter months daily radiation is under 3 kWh/m2 and in June it is only 2.42 kWh/m2. It will be assumed, that a power station is to supply 1000 MW for 7 daylight hours and to store and supply 800 MW for the other 17 hours of the day.
A central receiver solar thermal plant would not be effective given that
solar radiation exceeds 400 W/W/m2 for only 2 hours a day in June. If
the solar energy were collected in the form of photovoltaic electricity at 20%
efficiency and 93% efficiency of inversion to AC, then 0.45 kWh/m2/day
would be collected in June, and 15.6 million m2 of collectors would
be needed to supply 7 million kWh in daylight.
Note that these figures refer only to the cost of the collection of sufficient solar energy and converting it into electricity. They do not include the cost of the plant to produce, store and pump hydrogen, operation and management costs, insurance, profits of the power station to produce electricity from the stored hydrogen, and the payment of interest on the money borrowed to build the system. The power station to produce electricity from the stored hydrogen would cost about as much as a coal-fired power station capable of providing the entire 24-hour supply. The interest payments would probably at least double the sum of all the other cost items (see below).
The cost of building a 1000 MW coal-fired plant will be taken as US$ 800 million. This is the cost of the Australian Mount Piper plant soon to commence operations (Pacific Power, 1993a, p 104). The cost of the coal to fuel the power station over an assumed 20-year lifetime, at US$ 30 per tonne would be approximately US$ 1.56 billion, giving a plant plus fuel cost of US$ 2.36 billion. The cost of the solar thermal plant would therefore be approximately 56 times that of a coal-fired plant. This example seems to indicate fairly decisively that it would not be feasible to provide electricity at this latitude in winter from a power source operating solely on solar energy.
The most serious problems for renewable energy are set by Europe. A commonly argued solution (e.g. Johansson et al, 1993; Ogden and Nitsch, 1993) is to locate solar plant in Africa and transport hydrogen to Europe. Again winter is the main difficulty. North Africa is about as far from the equator as Sydney. If the plant was located further south, for example within 200 of the equator, then the energy would have to be transported some 4000 km to northern Europe. In the winter insolation would average only about 4-4.5 kWh/m2. Half the power stations might be located 200 North of the equator and half located 200 South, to ensure that half always received summer insolation, but then most of the energy would have to be transported about 6000 km in the northern winter. The available options, pumping hydrogen gas, tanking liquid hydrogen, taking and transmitting high voltage electricity, would all involve large energy losses.
Let us first consider supplying solar electricity to Europe transported in the form of liquid hydrogen. The overall energy efficiency of delivering solar electricity via a liquid hydrogen system would be around 3.4% (without taking into account the energy cost of shipping liquid hydrogen). Therefore for a 1000 MW plant to provide 24 million kWh of electrical energy per day in Europe would require collection of 705 million kWh in the form of solar energy. In a region averaging 4.5 kWh/m2/day this would require 157 million m2 of collectors and constructing the collection plant alone would cost US$78.5 billion for a solar thermal system, at current wholesale prices, or US$ 157 billion for a PV system at US$ 1000/m2.
If the energy were to be transported to northern Europe in the form of hydrogen gas, energy losses due to pumping from plants 200 North of the equator would probably be around 15% (derived from Ogden and Nitsch, 1993; p 980). Therefore to deliver the equivalent of 1000 MW, a hydrogen flow equal to 1176 MW would have to be generated, i.e. 28.2 million kWh per day. Given the above conclusion that supplying electricity via hydrogen gas is about 4.8% energy efficient, solar energy equal to 587 million kWh would have to be collected each day and at 4.5 kWh/m2 the collection area would have to be 130 million m2, indicating a cost of US$ 65 billion per solar thermal collection plant.
More promising would seem to be the generation of electricity at these low latitudes and transmitting it 4000 km via high voltage electricity lines, requiring the storage of only the night time demand. In view of the distance involved the transmission would have to be via direct current. The main efficiencies for a thermochemical storage system would be approximately 60% for heat collection, 60% for storage of heat, 35% for electricity generation, 84% for transmission (UNSW Mechanical Engineering, 1993), and 93% for inversion of high voltage direct current. The assumption of only 16% loss from such long-distance direct current transmission is disputable. In fact it is possible that the losses would prohibit this strategy (Pacific Power, 1993b). However, combining these figures yields an efficiency of 10% for delivery at night following storage, and of 17% for daytime delivery without storage.
To deliver 7 hours x 1000 MW in the day time would require collection of 41 million kWh and this would require 9 million m2 of collector, assuming 4.5 kWh/m2/day. To deliver 17 hours x 800 MW per day during the night time would require collection of 136 million kWh and this in turn would require 30 million m2 of collection. The total area would have to be 39 million m2, indicating a total construction cost for a solar thermal collection plant of US$ 19.56 billion.
Because the construction costs of high voltage-lines could be in the region of US$ 500,000 per km (Pacific Power, 1993b), the cost of each line would be approximately US$ 2 billion. The maximum likely load for each line would be 1000-2000 MW meaning that there would have to be one line for each one to two power stations (UNSW Mechanical Engineering, 1993).
Even ignoring land availability for the hundreds of lines necessary to supply Europe or the cost of crossing the Mediterranean Sea, line costs would raise the construction cost per plant by more than the cost of building a coal-fired power station. The major problem, however would be how to store the vast quantities of gas that would be involved. In thermochemical storage 1 m3 of gas can store 1.54 kWh. Therefore to store the 136 million kWh would require 88 million m3 of gas, i.e. the approximate volume of a mine tunnel 8830 km in length, assuming 100-fold compression of the gas, just for the daily stored output for one plant.
The foregoing figures indicate that the cost included in only the plant to collect the required amount of solar energy to supply high latitude countries with electricity would be around eight times the cost of the coal-fired plant plus fuel capable of meeting the demand.
Again this overall cost figure excludes many items to be discussed below. It should also be noted that these examples have not taken into account the amount of time power plants are out of action for repairs (capacity factors of 0.7 are typical for coal-fired power stations).
Let us assume a PV plant capable of meeting 1000 MW daytime and 800 MW nighttime demand via hydrogen storage in summer when Sydney's insolation averages 6.18 kWh/m3. In winter insolation would-be less than half as high, so half the output would have to be replaced by fossil fuels.
Averaged over the whole year fossil fuel equal to one-quarter that used in a coal-fired plant would be needed. The approximate electricity consumption in Sydney is 1+kW per person requiring about 3.6 tonnes of coal per year. The use of solar plants of the type assumed would cut this to 0.88 tonnes. However if the IPCC's two-thirds reduction is accepted and the remaining energy is shared equally among all people expected by 2060 the per capita amount available for all purposes would only be 0.39 tonnes pa. In other words the fossil fuel budget would not be anywhere near sufficient to provide the necessary back up for a system delivering present developed nation per capita electricity use to all in winter.
There would be 6.5 million tonnes of steel in the 130 million m2 plant constructed to supply hydrogen gas to Europe from low latitudes, discussed above. It takes about 8000 kWh to produce 1 tonne of steel, so the energy needed to produce the steel would be about 70 billion kWh. The plant would generate around 0.7 x 100M W x 24 hrs x 365 days per year (assuming it operated on average at 70% capacity); i.e. 6123 million kWh. This means it would take 8.5 years energy output from the plant to repay the energy needed to produce the steel to build it. If the plant had a 20-year lifetime, just over one-half of its operating time would be available to supply energy for other purposes. This in effect means we would need almost twice as much solar generating capacity as we thought we needed, and that the total cost of electricity consumed would be almost twice as high as we might have expected before considering the energy costs of materials.
This factor narrows the gap with the hydrogen strategy for the supply of electricity to Europe from Africa. A PV system would require a very large collection area due to the low energy efficiency of the process. The solar thermal plus thermochemical storage strategy would be about 2.8 times, as energy efficient and therefore would require less collection capacity, but it would require far more energy intensive materials.
In the foregoing discussion only items (1) and (2) have been quantified and in the best options they alone indicate renewable energy electricity system construction costs of 10 or more times those for coal-fired electricity.
A number of these costs would probably be more or less proportional to the collection plus construction cost, such as maintenance and repair, land costs, insurance and interest.
No attention has been given to land costs in this discussion. Large areas would be required if solar electricity were to be the main source, although the best sites for generation tend to be in arid areas. Land for solar thermal plants is about four times the area of the actual collectors. For photovoltaic systems the multiple is about two. For a solar thermal power station in Africa exporting hydrogen at a rate equal to 1000 MW, almost 700 km2 of land would be required. A solar electrical supply of all present US electricity from solar thermal sources would require some 180,000 km2 of land.
The operations and maintenance cost for solar thermal systems have been estimated at approximately 40% of the cost of constructing the, plant, assuming a 20-year lifetime (Clark, 1990, p 287; De Laquil et al, 1993, p 250; Hagen and Kaneff, 1991, pp. 4-14). For photovoltaic systems lifetime operations and management costs have been estimated at equal to 5% of plant construction cost per year (Ogden and Nitsch, 1993, p 984). (Note that mirrors and photovoltaic cells might have to be cleaned each 9-14 days to optimize performance; Radosevich, 1989; Hagen and Kaneff, 1991.)
Interest charges on the capital borrowed to construct the renewable energy power station would probably be the largest single cost item. These can more than double the figure for the construction of plant (Johansson et al, 1993, pp. 19, 65). Moreira and Poole (1993, p 112) show that with a 6% discount rate on a plant which cost US$ 1200/kW to construct and a 50-year lifespan, annual capital costs would be 2 - 2.2 cents/KW. Assuming a 0.7 capacity factor, such a plant would produce 306.6 billion kWh in its lifetime and the capital cost would therefore be US$ 6 billion per kW.
In other words in this case the interest charges would come to approximately four times the construction cost. 'Financial charges' equal 1.4 cents/kWh generated in the NSW system (Pacific Power, 1993a, p 110). If a 20-year lifetime is assumed per power station this adds to almost twice the construction cost.
A one million kW plant operating for 20 years with a capacity factor of 0.7 will produce 122,640 million kWh. The early 1995 Australian selling price of electricity was approximately 8 cents/kWh. Around 45% of this was for transmission and distribution (Electricity Supply Association, 1995), leaving 4.4¢ for generation. Approximate costs contributing to generation per kWh would be, construction of the US$ 800 million plant 0.6¢, operations and management 0.24¢, coal 1.3¢, company profit 0.44¢, apparently leaving an interest component approximately three times plant construction cost.
The basic cost referred to in this paper has been for plant construction. This might have to be multiplied by two to four in order to arrive at the real outlays necessary to cover construction plus repayment of interest. It would seem therefore reasonable to conclude that the final all inclusive cost of supplying renewable electricity after storage, without use of fossil fuels for backup, might be in the region of 20 times the present cost via coal-fired plant, and possibly much higher. This is far beyond levels that could be tolerated.
An important limiting factor for wind energy concerns the availability of suitable sites. As has been noted, presently quoted performance figures refer to the best sites and much depends on the overall distribution of sites. As for most renewable resources, global wind energy far exceeds human energy demand, but most of the exploitable wind resources are at a relatively few sites far from settlement (Grubb and Meyer, 1993, p 199). Most of Africa and Asia seem to have poor wind resources. Some countries, notably Britain, Canada, New Zealand and Crete might derive a large proportion of their energy needs from their unusually favourable wind resources (Grubb and Meyer, 1993). Windmills presently require winds around 7 m/s [18 - 20 mph] before they can operate efficiently. Very few parts of Australia have wind speeds averaging above 5 m/s, "a gloomy outlook for wind power generation" (Bell, 1982, p 24), although there are some excellent sites.
Will technical advance make it possible to extract energy from areas with lower average wind speeds? This is likely: but will probably not make a large difference to wind energy potential because of the relatively low amounts of energy available in winds of lower speeds. Energy varies with the cube of the wind speed and therefore falls rapidly as speed falls. Winds of 4 m/s have only 19% of the energy in 7 m/s winds. Typical windmill power curves show that although generation can commence at under 5 m/s, at 7 m/s output is less than one-third of the maximum operating output which is not reached until 13 m/s. Most of the energy available at a site over a period is in the much higher than average winds. Thus it would not make much difference if mills were developed capable of starting to generate in areas with average speeds under 7 m/s, because there is relatively little energy available in low speed winds. Such mills would not deliver much more energy than existing mills would deliver at those sites. Caution must therefore be exercised regarding the fact that there are large areas with average wind speeds below those which present mills are capable of operating on.
Because of the variability of the wind there is a considerable probability that most or all of the units in a wind farm would not be functioning at a given point in time, and that at other times much more energy would be being generated than the system could use. Even in the UK, one of the best wind energy regions, in summer there is a 27% probability that a wind system would be generating at less than 10% of its capacity (Grubb and Meyer, 1993, p 166). In general this 'integration' problem limits windmills to contributing only 10-20% of the total electrical demand, unless there is provision for storage. In some situations the fraction might be only 5% (Grubb and Meyer, 1993, p 205). In northern Britain, where there are remarkably good wind resources, the fraction could rise to 25-45% (Grubb, 1991). However, Elliot regards 20% as the most frequently estimated figure for the UK and says that siting and other constraints could cut this to 10% (Elliot, 1994, p 8). Even for Denmark, another highly favourable region, the wind potential has been put at only 25% of present electrical generating capacity (Grubb and Meyer, 1993, p 176). Siting mills in shallow offshore waters can make a difference but according to Grubb and Meyer, Europe is not likely to derive more than 10% of present electrical demand from wind resources (ibid., p 189).
US potential is much greater than for Europe (Duxbury, 1992, p 6). Elliott estimates that if all areas in the US with adequate winds (Class 5, i.e. 7m/s average, were used, allowing for areas that would have to be excluded for urban and agricultural reasons, only 20% of US electrical demand could be met. This estimate is based on the common assumption that the energy delivered from a windmill will be 25% of its peak capacity. Elliot's own figures for Californian windmills presently in use show that electricity delivered is only 18.6% of peak rating on average. (Elliot claims that if class 3 winds could be used economically, wind could generate 10 times present US electricity demand, although he does not discuss whether this is feasible in view of the problem set by low Weibull distributions noted above. Nor does he consider the integration limit.)
The analysis given by Elliott, has been criticized for making implausibly high estimates of the amount of rangeland (90%) that could be used (Grubb and Meyer, 1993, p 198). Europe has far less open land than the USA and because of its much denser settlement most of its good wind sites have been taken for other purposes.
A great deal depends on the distribution of wind speeds at a site. In general little can be concluded from annual average figures (Berrill and Jones, 1989, pp. 22-24). An impressive average could be made up of winds that are mostly too light and too strong for efficient windmill operation. This factor will eliminate many sites apparently promising on existing wind maps, which at this stage usually only give average speeds. Seasonal variations are also important. Fortunately in Europe the winds are strongest in winter when most energy is needed, but this is not the case in most of Australia (Hutchinson et al, 1984, p 319).
Because wind energy is mostly harvested in the form of electricity its storage involves the problems and losses discussed above for solar energy. If storage is via heat or hydrogen production, losses of the order of 75-80% would occur in the storage and regeneration of electricity, implying corresponding increases in the initial generating capacity needed. The associated five-fold increase in generating cost again refers only to constructing the collection capacity and does not include costs to do with storage, transmission and conversion of the energy, etc
Although the situation regarding potential is less clear for wind than for solar energy it would seem to be unlikely that except at special locations, a high proportion of present rich world electricity demand, let alone future world demand, can come from the wind.
All the energy stored each year in the planet's plant mass is equal to approximately five times present world energy use. All agricultural land would have to be used to meet the present demand for transport fuel (de Montalembert, 1983, p 40; Grathwol, 1982, p 249, and Cook, 1991). Pimentel et al estimate that the total biomass growth per year in the USA is 3.2 billion tonnes, and that only one-fifth of this could be harvested for energy production. However, if all of it was converted (assuming 15 GJ/t and petrol at 40 GJ/t) it would have a gross energy content equal to 850 million tonnes of petroleum i.e., only about 27% of total US energy use. Pimentel et al indicate that growing, collecting and processing would use about one-third to one-half as much energy as exists in the biomass (1984, p 89; 1994, p 9).
If 11 billion people were to have the present US diet approximately 22 billion ha would be needed to produce food, i.e. 1.7 times all the land area of the planet and 16 times its cropland. Movement in this direction will accompany Third World development and this must significantly reduce land that will be available for biomass energy production. According to Pimentel et al, to meet US electricity demand via biomass would require 2.2 ha per person (1994, p 4). For a world of 11 billion living as Americans do now some 24 billion ha would have to be harvested, approaching twice the total land area of the planet.
Agricultural and forestry wastes could be used but in the long-term all these nutrients should be returned to the soils from which they were derived. Shea (1988, p 20) estimates that the potential biomass energy retrievable from all forest, crop and animal wastes in the USA, West Germany and Japan would yield only 2 - 5% of present energy demand. According to Pimentel et al "crop residues should not be removed from the land for a fuel source" (1994, p 6).
Figures from Brazil where biomass is used on a significant scale to fuel
motor transport can be somewhat misleading. Brazil has access to large areas of
land capable of growing sugar cane, at yields between 50 and 60 tonnes
per hectare per year. If Australia could double its sugar cane production this
would only meet 10% of its present transport fuel demand. Annual biomass
production per ha from the world's forests averages only about 2.2 t/ha/year
(Hall et al, 1993, p 615).
Pimentel et al calculate that even ignoring the energy cost of producing ethanol from corn (this takes more energy than there is in the resulting ethanol) and taking into account only the energy needed to produce the corn, a net yield of only 37 gallons of ethanol would be produced per acre per year. To fuel an average US car for one year would require no less than 14 acres of high quality cropland, about nine times the amount needed to feed one American (Pimentel et al, 1994, p 5; Brown, 1980, p 25).
The prospects for large-volume liquid fuel from biomass depend mostly on the potential of tree plantations. Unfortunately there is considerable uncertainty surrounding the yields that would be sustainable in the long run. Influential optimistic claims have been based on the assumption that high yields can be taken year after year from large areas of degraded land. For in stance Johansson et al (1993) and Hall et al (1993) assume that such lands can continually sustain a yield of 15-30 t/ha/year. This would seem to be highly unlikely given that even with very high inputs of energy, fertilizer and irrigation, US corn plant growth averages about 14 tonnes per ha per year, on high quality soils and in ecologically unsustainable ways. Lazarus et al estimate mature forest growth in-wet tropical areas at 8 t1ha/year, but in temperate forests only 3 t/ha/year (1993, p 155).
The US Office of Technology Assessments estimates US forest growth at 2-4 t/ha/year (Cook, 1991). Yields of 4 t/ha/year (for one 'crop') are achieved in Australian pine plantations, but there is no clear evidence as to what yields would be sustainable year after year. There is evidence of significant falls in yields with second plantings (Keeves, 1966). The world average forest growth might be as low as 2.2 t/ha (Hall et al, 1993, p 615). Pimentel, et al (1994) regard 3 tonnes/year as a probable limit for general forest yields even assuming fertilizer use (as distinct from using degraded land).
It is debatable whether any significant yield should be taken from forested land given that long-term sustainability must ensure there is no net nutrient loss. Australia's wheat production averages less than 3 t/ha and depletes the soil of nutrients equal to around 60 kg of fertilizer per ha each year in excess of the amount of fertilizer applied each year (Lipsett and Dann, 1983). These considerations seem to indicate that in the long-term plantations on unused land will not enable very large-scale production of liquid fuels.
The implications for fertilizer demand are also problematic. The minerals taken from the soil within a crop constitute about 3% of its weight (Lipsett and Dann, 1983). If 3 tonnes of biomass are harvested per hectare then 90 kg/ha of minerals would have to be replaced by fertilizer. This is about half the average annual application of fertilizer on the croplands of the rich world, ignoring the fact that a considerable fraction of applied fertilizer washes or blows off the surface. To ensure replacement of 90 kg of nutrients per hectare, application would have to be at a much higher rate. To equal present world petroleum consumption, the gross energy content of annual forest growth on 2.17 billion ha would have to be harvested at 3 t/ha and 20 MJ/t for wood. Assuming 90 kg of fertilizer per hectare, total fertilizer requirement would be 195 million tonnes, around 30% more than total present world fertilizer consumption.
There would also be various ecological problems, including loss of species diversity, the generation of 13 litres of high biological oxygen demand waste for each litre of fuel produced, and high erosion rates after harvest. Pimentel et al report 15-17.6 tonnes of soil lost per ha per year from newly harvested forest (1984, p 91, 1994).
There are considerable areas of unused cropland and of degraded forest land that could be used for biomass production. However, in view of what appear to be rapidly worsening trends in a number of indices to do with world biological and ecological production which bear directly on future food production prospects, it is likely that in coming decades there will be little or no increase in the area of agricultural land in use, and a reduction is probably due to natural limits on the supply of good land, the cost of introducing marginal land and especially the loss of land to erosion, waterlogging and other factors.
Gains in world food production have come predominantly from yield increases in recent decades. However, in the last few years it has become evident that these can no longer be relied on. After three decades of generally steeply rising yields, for a number of the most important items, including wheat, rice, meat and fish, yields now appear either to be tapering towards limits or to have actually fallen (Brown, 1993). These trends are related to over harvesting and deterioration in underlying ecological conditions.
These disturbing indices mean that there is likely to be renewed pressure to increase food harvests by increasing areas under cultivation. The pressures are likely to be very great since at present only one billion people are being provided with affluent diets and fibre (etc) consumption and world population could exceed 10 billion within decades. These pressures will probably have dramatic impacts on the areas available for biomass production.
Another factor which must be borne in mind is the reduction in forest land potentially available for biomass that would occur if all the world's people were to continue moving towards present rich world per capita consumption of forest products. If the projected 11 billion were to each consume forest products at the North American average rate world timber production would have to be six times its present level and world forest area would have to be about the same as all the planet's land area (Rees, 1992; Pimentel et al 1984). Although these are impossible multiples the development of the Third World is taking us in these directions and as a result the future demand for timber can be expected to significantly reduce land available for the biomass production of energy.
Biomass sources are therefore unlikely to meet present liquid fuel needs. If a world population of 11 billion were to use cars at the present rich world per capita rate more than 10 times the present world transport fleet would have to be fuelled. Even if a four-fold improvement in the energy efficiency of vehicles is assumed, total fuel demand would still be a multiple of present demand.
Geothermal energy is best thought of as a stock of heat that would be used up over time rather than a continually available flow. Although useful quantities are retrievable in some limited high temperature locations, the average flow of heat reaching the earth's surface is only 0.4 kW/W/m2/year. The flow is at a very low temperature, unlikely to be transformed into useful temperatures and forms at better than 1 % efficiency. The World Energy Conference concluded that all US sites might only provide 2% of US electricity for 100 years (Krenz, 1980, p 215; see also Ken, 1982; Palmerini, 1993, p 558).
Tidal power could be effective at a limited number of locations but various analyses conclude that if all potential sites were developed energy equal to less than 0.1 % of present world energy could be collected (Krenz, 1980, p 213; Kiely, 1978, p 120).
Wave power potential is more difficult to estimate given that the resource is huge but collection is especially problematic due to the destructive power of storms. It is likely to make a significant contribution at some sites. The problems of storage and integration are again involved. Wave power has been estimated as unlikely to make a major contribution to world energy needs (Foley, 1976, p 219).
Could some of the gaps be filled by a system combining sources, for example, using biomass to augment solar power in winter? This would make a significant difference only if some sources were at times available in greater quantity than was needed to meet their main use. The analysis above concludes that there would be insufficient biomass to meet liquid fuel demand, so it would not be available to contribute to electricity generation when the intermittent sources were insufficient. The potential within the other major possibility for augmenting intermittent sources, hydropower, is too limited to make a major difference.
Two important considerations must now be taken up. The first is the possibility of extending present rich world living standards to all people, and second the energy supply implications of the fundamental commitment built into our economy, i.e. limitless growth.
Could adoption of technologies enabling C02 to be extracted from
fossil fuel sources in the production of hydrogen or electricity solve the
problem? The Intergovernmental Panel on Climate Change has indicated that fossil
fuel consumption should be cut by 60-80% from its present amount. The resulting
quantity would require total carbon emission per person to be cut to 6% of the
present rich world average, in an equitable world of 11 billion people. It
remains for advocates of carbon extraction to show that a reduction of this
order is technically and economically feasible.
If the rich countries were only to average 3% growth while the Third World rose to the present per capita level of the rich countries then by 2060 world economic output would be 20 times what it is today.
The analysis above indicates that renewable energy sources will not be capable of sustaining the present level of world energy production and use, let alone any multiple of it.
What about the possibility that world economic growth will yield levels of wealth that make the costs estimated above acceptable? The magnitude of the multiples makes this implausible. It was concluded that electricity from renewable sources would probably be 20 times as expensive as it is now, and possibly much more so. If the average rate of world economic growth per capita over the past two decades were to continue to 2060 World Gross Product would only be about 4 times its present level (World Bank, 1993, p 219). Large-scale conversion to renewables will probably become highly desirable well before 2060. Rather than increased world economic wealth enabling the high costs to be borne, it is more likely that the costs will seriously hinder economic activity.
Estimates of the potential reductions achievable by conservation measures appear to lie generally in the region of 50%. Lovins is notable for claims in the region of 80% but it is not clear that these claims are supported by his own analysis. In Soft Energy Paths (1977), and Least Cost Energy, Lovins and Lovins (1981) figures are given for a number of elements and processes but these fall far short of comprising the whole economy, and they appear to average 50-60% saving rather than 80%. Several other estimates fall in the region of 50-60% (e.g. Business Council of Australia, 1991; International Energy Authority, 1987, p 29; Morris, 1982, p 142; Brouwer, 1988; Colombo et al, 1991; McLaren and Skinner, 1987, p 228; Worldwatch Institute, 1991, p 26; see Trainer 1985, p 83). Gellings et al (1991) state lower figures for electricity as does Schipper regarding cars (1991, p 128).
A 50% reduction would halve the overall magnitude of the problem but it would still be enormous. A world of 11 billion living as people in rich countries do now but using the energy conservation measures envisaged would need four times as much energy each year as the world uses now, rather than eight times as much.
In addition our society is committed to constant and limitless growth in
output and this factor must soon overwhelm the contribution any realistic
assumptions about conservation could make. If the energy needed to achieve a
given level of output was reduced by one-third, but output increased at 3% per
year, then in 14 years the energy required would be back up to the original
level, and in another 23 years it would be twice as great.
Is transition to an economy concerned with services and in information likely to head off these problems? The service sector already makes up about 70% of more of the economies of rich countries. If 3% per year growth took place while the (most energy intensive) non-service sector did not increase, by 2060 the service sector would have to make up 96% of the whole economy. It is not plausible that there can be a significant increase in the service sector without considerable increase in energy use, especially given that many important services are energy intensive (e.g. travel, tourism, transport), many deal with energy intensive processes or goods (insurance, retailing), and growth in services means growth in construction, equipment and offices.
The primary concern of this discussion has been to reinforce the case for taking the second option. This is to move to a 'radical conserver society' in which it is possible to live well on far lower levels of per capita energy consumption and, after a period of marked "de-development", without growth in economic output or energy use.
Conventional economic thought has great difficulty coming to terms with this option but there is now a considerable and rapidly growing literature arguing that the high material 'living standards' characteristic of the 'over developed' countries are unsustainable for ecological, resource, security and social reasons, and that a sustainable world order must be based on materially simple lifestyles, a high level of local economic self-sufficiency and a steady-state or zero-growth economy (e.g., Korten, 1990; Clark, 1989; Gordon and Suzuki, 1990; Douthwaite, 1992; Mollison, 1988; Trainer, 1985, 1989, 1995; Goldsmith, 1988). In addition there is now a substantial and rapidly growing 'eco-village' movement exploring the development of sustainable settlements (e.g. In Context Institute, 1991).
Unfortunately policy analysis is still overwhelmingly dominated by the unquestioned assumption that the ceaseless pursuit of more economic growth and higher material living standards is the only conceivable path. My, The Conserves Society: Alternatives for Sustainability (1995) summarizes the now considerable literature on the more simple and self-sufficient alternative path and attempts to show that it is viable and attractive.
Please send mail to
firstname.lastname@example.org with questions or comments about this web site. Minnesotans For Sustainability
(MFS) is not
affiliated with any government body, private, or corporate entity.
Copyright © 2002, 2003, 2004
Minnesotans For Sustainability