international journal of hydrogen energy 34 (2009) 31–39

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/he

Hydrogen’s role in an uncertain energy future

Patrick Moriartya,*, Damon Honneryb

a
Department of Mechanical and Aerospace Engineering, Monash University, P.O. Box 197, 900 Dandenong Road, Caulfield East,
3145 Victoria, Australia
b
Department of Mechanical and Aerospace Engineering, Monash University, P.O. Box 31, Clayton Campus, 3800 Victoria, Australia

article info abstract

Article history: This study explores global energy demand, and hydrogen’s role, over the 21st century. It
Received 18 September 2008 considers four illustrative cases: a high (1000 EJ) and a low (300 EJ) energy future, and for
Received in revised form each of these conditions, a high (80%) and low (20%) fossil fuel energy share. We argue that
21 October 2008 neither high energy future is probable, because of resource limitations, and rising energy,
Accepted 21 October 2008 environmental and money costs per unit of delivered energy as annual energy demand
Available online 28 November 2008 rises far beyond present levels. The low energy/low fossil case is most likely, followed by
the low energy/high fossil case, although both require large cuts in energy use, and most
Keywords: probably, lifestyle changes in high energy use countries. Hydrogen production would be
Hydrogen best favoured in the low fossil fuel options, with production both greater, and implemented
Energy consumption earlier, in the higher energy case. It is thus least likely in the low energy/high fossil fuel
Renewable energy case.
Global warming ª 2008 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights
Carbon capture and storage reserved.

1. Introduction rising share of fossil fuels. For electricity production, their

share has been rising for several decades [1,3].
Two questions are of crucial importance in discussing the How much energy will we need in the future? A more
future of energy in the 21st century: how much energy will we equitable future world would require reductions in the
consume annually, and what sources of energy will we use? present large differences in per capita primary energy
The answers to these questions are by no means certain. In consumption. Among the high-income countries, in 2004 Italy
2006, total global primary energy consumption was about had the lowest per capita energy use at 132.7 GJ [2].
493 EJ [1,2]. (EJ ¼ exajoule ¼ 1018 J.) Following the International (GJ ¼ gigajoule ¼ 109 J.) The UN median estimate for 2050
Energy Agency (IEA) convention, energy generated from global population is 9191 million [4]; if all used energy at this
hydroelectricity and other renewable primary electricity rate, global primary energy use would be 1220 EJ. This value is
sources is converted to primary energy on a one-to-one basis similar to the maximum value of 1173 EJ in 2050 in the various
[2]. Again following IEA practice, primary energy in this paper scenarios in Riahi et al. [5]. Various other researchers present
includes non-commercial fuel wood. Table 1 shows total futures with roughly 1000 EJ or more primary energy for 2050
global primary energy from 1970 to 2006, illustrating both the [6–11], with some envisaging even higher values later in the
steady growth in energy use over this period, and the recent century. While these researchers do not necessarily view their

Abbreviations: ASPO, Association for the Study of Peak Oil; CCS, carbon capture and storage; CO2, carbon dioxide; EIA, Energy Information
Administration (US); EJ, exajoule (1018 J); EWG, Energy Watch Group; FC, fuel cell; GJ, gigajoule (109 J); Gt, gigatonne (109 tonne); H2, hydrogen; He,
high energy (1000 EJ); Hf, high fossil fuel (80%); IEA, International Energy Agency; IPCC, Intergovernmental Panel on Climate Change; Le, low energy
(300 EJ); Lf, low fossil fuel (20%); Mt, megatonne (106 tonne); RE, renewable energy; WETO, World Energy Technology Outlook.
* Corresponding author. Tel.: þ61 3 9903 2584; fax: þ61 3 9903 2076.
E-mail address: [email protected] (P. Moriarty).
0360-3199/$ – see front matter ª 2008 International Association for Hydrogen Energy. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijhydene.2008.10.060
32 international journal of hydrogen energy 34 (2009) 31–39

supplydTable 1) and ‘low’, 20% fossil fuels. We impose one

Table 1 – Global primary energy use, EJ, 1970–2006.
final limit on future energy; for simplicity we assume the
Energy source 1970 1980 1990 2000 2006 primary/secondary energy ratio remains constant, while
Coal 64.2 75.7 93.7 98.2 129.4 acknowledging that it may alter with primary energy level and
Oil 94.4 124.6 136.2 148.9 162.9 mix.
Natural gas 38.1 54.9 75.0 91.8 107.8 We argue that neither high energy case is probable, because
Fossil fuels 196.7 255.1 305.0 339.0 400.1 of resource limitations, and rising energy, environmental and
money costs per unit of delivered energy as annual energy
Nuclear 0.7 6.7 19.0 24.5 26.6
demand rises far beyond present levels. The low energy/low
Renewablea 29.4 37.6 48.5 55.6 66.2
fossil case is most likely, followed by the low energy/high fossil
All energy 216.8 299.5 372.4 419.0 492.9
case, although both require large cuts in energy use, and most
Fossil % 90.7 85.2 81.9 80.9 81.2 probably, lifestyle changes in high energy use countries.
Hydrogen production would be best favoured in the low fossil
Sources: Refs. [1,3].
a Values are only approximate: fuel wood data not accurately fuel options, with production both greater, and implemented
known. earlier, in the high energy case. It is thus least likely in the low
energy/high fossil fuel case.

figures as projections, they clearly regard 1000 EJ by 2050 as at

least possible. 2. Challenges to future sustainable energy
Because of the various serious constraints facing high provision
energy use, low energy futures must also be considered. The
energy conservation needed for these can occur in two ways; As shown in Table 1, the present global energy system is
from increasing technical energy efficiency of power genera- dominated by fossil fuels, and the official forecasts discussed
tion and energy-using equipment, or from less use of energy- above see little change in this pattern before 2030. An
consuming equipment. One recent study [12] estimated that important question is for how long these business-as-usual
for their ‘2  C’ scenario, global primary energy consumption in projections can continue without running into constraints in
2050 could be held to 422 EJ, lower than today’s value. Amory the form of limited reserves of fossil fuels, or severe envi-
Lovins is a strong advocate of the potential for technical ronmental problems from their combustion, including not
energy efficiency, arguing that energy efficiency in a variety of only global climate change from CO2 and methane emissions,
applications can be increased by a factor of 10–100, and that an but also air pollution problems. In this section we examine the
overall four-fold reduction in energy use is possible [13]. An implications for 21st century fossil fuel use of resource
annual primary energy use of 300 EJ could thus be considered depletion and greenhouse gas emissions.
as illustrative of a low energy future. The Association for the Study of Peak Oil (ASPO) [14] project
We focus here on primary energy because our chief that combined global annual production of oil and natural gas
concern is with climate change impacts and fossil fuel (even including that from unconventional sources, such as oil
depletion. But what relation does primary energy have to the sands) will peak around 2010 at about 290 EJ (47.5 billion
energy available for consumption? In 1973, the ratio of global barrels of oil equivalent), before falling to around 245 EJ in 2030
primary energy to total final consumption (secondary energy) and 140 EJ in 2050. Their combined production was 271 EJ in
was 1.34, but had risen to 1.47 by 2004 [2]; primary energy rose 2006 [1]. Oil production has not risen over the past three years.
faster than final energy use. If all energy was derived from coal Simmons [15] stresses that much of the world’s present oil
fired power stations this ratio would approach 3.0; but if and gas supplies ‘come from large fields now too old, and new
derived solely from hydroelecticity, the ratio falls to 1.0. fields that are too small.’
Greater use of renewable energy does not necessarily guar- For coal, the Energy Watch Group (EWG) from Germany
antee a better secondary energy return, as it is likely that forecast that if present trends continue, global production will
increased renewable energy would see greater need for energy peak around 2025, at about 152 EJ, compared with the 2006
storage and conversion. Each additional process will act to level of 129 EJ [1,16]. By 2050, production will not be much
reduce the energy available for consumption. Increased fossil lower than 2025, but by 2100 the EWG study projects it to have
fuel use would ultimately need greater use of coal and non- fallen to about half its 2006 value. For fossil fuels overall,
conventional oil sources, again raising the primary/secondary production in 2006 was 400 EJ [1]. Combining the ASPO and
ratio. EWG projections, peak production could occur as early as 2020
Within the range of possible energy futures, what role will at about 423 EJ, and by 2050 could have fallen to 286 EJ. The
hydrogen play? In this paper we limit primary energy EWG authors stress that peak coal production in China, the
consumption in 2050 to two cases: a high energy energy future largest producer, will determine the timing of global peak
(He ¼ 1000 EJ) and a low energy future (Le ¼ 300 EJ). Other production. Their conclusions thus receive some support from
values for future energy are of course possible, but they will recent Chinese research which forecasts peak coal production
probably fall within these limits. We then explore hydrogen’s in China occurring between 2025 and 2032 [17]. Also, Rutledge
role by considering the energy supply mix for each energy [18] has applied ‘Hubbert linearisation’ to coal production
future. To do this we define two energy supply mixes, namely forecasting, and has come up with similar results to EWG [19].
a high (Hf) and a low (Lf) fossil fuel share, where ‘high’ means Both ASPO and EWG stress that accurate data on reserves for
80% fossil fuels (roughly their present share in global energy all fossil fuels are lacking for many countries [14,16].
 international journal of hydrogen energy 34 (2009) 31–39 33

Of course, many estimates forecast much higher possible delivering a unit of secondary energy to users will all rise as
fossil fuel use until well into the future (e.g. [6,20,21]). These in premium reserves are depleted.
turn are based on assessments of recoverable reserves (and Fossil fuels also produce a number of environmental
future additions to reserves) far more optimistic that those problems associated with their exploration, extraction,
discussed above [1,6,21,22]. (However, even the annual BP transport and combustion. The 2007 Intergovernmental Panel
reports caution the reader about the unaudited nature of the on Climate Change (IPCC) reports argue that avoiding a rise of
reported reserves.) They are also far more optimistic about the more than 2  C above pre-industrial values could require cuts
rate of technical progress and annual production from in global greenhouse gas emissions by the year 2050 to as low
unconventional resources, such as oil from tar sands and as 15% of the year 2000 emissions [25]. The European Union
natural gas from coal seams. Two energy authorities, the IEA has proposed such a cap of 2  C as necessary to avoid
[2] and the US Energy Information Administration (EIA) [3], dangerous anthropogenic change [26]. Reductions of this
have recently projected both total primary energy use and its magnitude would require emissions peaking in a few years
composition out to 2030. For 2050, projections are available [25] with substantial cuts in place by 2030. Hansen and
from the European Commission in their 2006 World Energy colleagues [27] go further, arguing that atmospheric CO2
Technology Outlook__2050 (WETO) [20], and a 2007 IEA study concentrations must be reduced from the 385 ppm present
[11]. The EIA report projects fossil fuel use at 638 EJ and the IEA level to 350 ppm. In brief, there are time limits for emission
563 EJ, both for their 2030 base case [2,3], similar to the WETO reductions; major cuts are needed in two decades or less, not
study reference case values of 564 EJ for 2030, rising to 660 EJ in by the end of the 21st century [26].
2050 [20].
So far, it has been implicitly assumed that if needed,
reserves will be produced at the maximum rate, limited only 3. High fossil fuel supply mix
by geological constraints, economics and good field develop-
ment practices. But fuel-exporting countries could well decide The reports discussed above all assume high energy
to limit their annual production to levels far less than the consumption in the future, with 80% or more coming from
maximum, even if reserves are much greater than the pessi- fossil fuels in their reference scenarios in 2030, and for the
mistic estimates of ASPO/EGS. Limiting output would help WETO report, still over 70% in 2050 [2,3,20]. Since fossil fuel
both maximise their total revenue and reserve some fossil use in recent years has grown strongly, forming a rising share
fuels for future generations. Output reduction could also occur of primary energy, a high energy future based largely on fossil
by damage to infrastructure, whether natural- or human- fuels (He–Hf case) deserves serious consideration. This option
caused. For example, Hurricane Katrina in 2005 did extensive also benefits from the massive past investment in the energy
damage to oil rigs in the Gulf of Mexico. supply system. But given the possible constraints discussed
Increasingly, future petroleum will need to come from both in the preceding section and by several other
unconventional sources, such as oil sands, deep water and researchers [24,28], we need to ask how likely such an energy
polar oil. The Canadian oil sands are a vast deposit covering future is, and if implemented, for how long could it be
141,000 km2 with an estimated 1.7 trillion barrels (10,400 EJ) of maintained?
bitumen in place, of which about 10% is recoverable with Clearly, if drastic greenhouse gas reductions are needed,
current technology. However, the rate of development of new high fossil fuel use requires some form of carbon sequestra-
annual capacity is limited by several factors including labour tion, as there is no other way of reducing the resulting emis-
shortages and environmental constraints [22,23]. Even opti- sions to the atmosphere. Carbon can be sequestered in soils
mistic projections only see 8.9 EJ (four million barrels/day) of and forests, but in a warming world there are doubts about the
oil from Canadian oil sands by 2020, compared to the 1.6 EJ permanence of these sinks; a number of coupled carbon cycle-
produced in 2006 [22]. Any further weakening of oil prices will climate models show annual terrestrial carbon uptake
adversely affect their economic viability. declining before 2050, and sinks becoming carbon sources
A growing proportion of natural gas will either need to be later in this century. The authors reviewing these models
shipped to market in the form of liquids such as liquefied conclude: ‘Overall, it is likely that, at least on a global scale,
natural gas or methanol, because of remoteness from terrestrial ecosystems will provide a positive, amplifying
markets, or come from unconventional sources such as ‘tight’ feedback in a warming world, albeit of uncertain magnitude’
gas, coal bed methane, or even methane hydrates. Liquefac- [29]. It thus seems unwise to rely on soil or biomass carbon
tion uses up 7–13% of the energy in the input natural gas, with sinks to offset energy/industry emission reductions.
further energy expended in carrying the liquid to markets [22]. We are left with carbon capture and storage (CCS), either by
For coal, official estimates of ultimately recoverable reserves capture of CO2 from large power stations and industrial
are vast, at around 100,000 EJ or more [6,16,22]. Only a small plants, or directly from the air. The three methods for CO2
share of reserves can be recovered by open-cut mining; the capture from large plants include: using absorbants such as
rest requires underground mining, which has higher extrac- amides to extract CO2 from the exhaust gas stream (the most
tion costs and lower labour productivity. Bockris [24] also developed method and the only one that can be retrofitted);
stresses that only a small proportion of the coal in place can be combusting the fuel in pure oxygen; and gasifying the coal fuel
economically extracted, particularly if seams are thin. Fuels into H2 and CO2, followed by separation of the CO2 [30,31]. It is
extracted will have increasingly higher sulphur content (oil even possible that coal gasification could proceed in the
and coal), or for natural gas, carbon dioxide content. The absence of CCS [30], given the energy efficiency advantages of
monetary, environmental and energy costs (and risks) of combined-cycle gas turbines.
34 international journal of hydrogen energy 34 (2009) 31–39

An advantage of air capture is its unrestricted location, drastically curtail use of energy-using equipment, which
allowing it to be done near sequestration sites [32], thereby would inevitably require significant lifestyle changes in OECD
largely eliminating transport costs. Annual CO2 sequestration countries. We have presented some ideas on how this could be
is restricted neither by annual emissions from large CO2- achieved, particularly for transport, in recent papers [26,37].
emitting plants, as is conventional CCS, nor even by total Both technical efficiency and reductions in use of energy-
global emissions. It is thus in principle possible for one using equipment need to proceed in step.
country, or one group of countries, to actively reduce global In the low energy (Le–Hf) future, only 240 EJ fossil fuel
CO2 atmospheric concentrations, rather than merely slowing energy is required, but even this level could not be maintained
their build-up. But while CO2 burial would be identical to for long if production was limited by ASPO/EWG assumptions.
conventional CCS, air capture would likely be much more CCS would still be needed, at an annual rate of 13.5 Gt, given
energy-intensive. Italian researchers [33] have analysed the the assumptions of the previous paragraph. Since the rate of
energy costs of two possible designs for air capture, and found construction of new fossil fuel power stations would be small,
that the process energy required per tonne of CO2 was larger CCS would mainly have to come from add-on technology to
in both cases than that obtained by combusting the coal. Its existing power plants, or from air capture, both energy-
high energy and monetary costs can be illustrated by the need intensive.
to process almost 12,000 Gt of air per year to remove the CO2 Fig. 1 shows historical global consumption of fossil fuels,
emitted from current annual global transport alone. It will together with likely limits to annual production in the decades
therefore only play a supporting role to centralised emission up to 2100, following the assumptions of ASPO/EWG. Also
capture. shown are limits to fossil fuel combustion for energy-related
CCS introduces what has been termed ‘moral hazard’ into CO2 emissions if (i) 15% or (ii) 50% of their year 2000 value
the energy question [34]. Already in the UK, some environ- by 2050 or thereafter, assuming 71.5 Mt CO2/EJ. These corre-
mentalists see CCS as a convenient excuse for building new spond to the 85% (upper) and 50% (lower) of year 2000 CO2
coal-burning power stations, even though CCS on the huge emissions cuts given by the IPCC for 2  C maximum rise [25].
scale requireddbillions of tonnes sequestered annuallydstill The down-sloping curve after 2025 gives the maximum
faces a variety of serious technological, legal and public possible annual fossil fuel (EJ) use. Maximum use would
acceptance challenges [35]. Air capture would compound the require most CO2 emissions to be sequestered, the amount
problem, by providing a plausible reason for even further diminishing with time. If emissions are not sequestered,
delay in tackling today’s emissions from fossil fuel combus- annual fossil fuel (EJ) use must be limited to 51 EJ (170 EJ) for
tion. The slow progress on nuclear waste disposal also 15% (50%) of year 2000 CO2 emissions.
suggests that these fears are well-founded. High fossil fuel energy futures seem unlikely, given both
A related point is the technological optimism shown in the possible resource constraints and the massive CCS effort
many discussions on energy futures. The costs and time required for 800 EJ fossil fuel use. Further, the energy costs of
frames required for implementing radically new technologies energy will continue to rise with cumulative fossil fuels
like CCS, or orders of magnitude scale-ups of existing ones, extracted, since such large annual outputs require tapping
such as those that have been proposed for RE, are systemati- progressively more unconventional (and more difficult-to-
cally under-estimated [36]. Also under-estimated is the polit- extract) sources. A litre of petrol from oil sands requires 3–4
ical opposition likely to siting CO2 storages, liquefied natural times more energy inputs than does a litre from conventional
gas terminals, or new nuclear power or reprocessing plants. oil. For shale oil, energy costs would be even higher [22]. Per
And although the difficulties facing technologies which allow unit of primary energy, less secondary energy will be delivered
continuation of large-scale energy systems are usually than is the case today, and the economic and environmental
understated, the difficulties facing social change solutions to
energy problems (e.g. energy conservation, discussed below)
are often overstated [37].
The quantities of CO2 involved are enormous. In 2004 the
fossil fuel mix averaged 71.5 Mt CO2/EJ [2]. Assuming a similar
fossil fuel mix in 2050, 800 EJ of fossil fuels would emit 57.2 Gt
CO2 annually. If emissions of CO2 from fossil fuels had to be
cut to 15% of their year 2000 value (in line with similar cuts in
CO2-equivalent emissions from all sources), the world could
only release 3.6 Gt [1], thus needing sequestration of 53.6 Gt
annually.
Clearly, the low energy future (300 EJ of global primary
energy in 2050), whether fossil- or non-fossil based, would
require massive energy reduction efforts, given present and
projected energy trends. The favoured approach is to improve
the technical efficiency of all energy-using equipment, Fig. 1 – Fossil fuel use 1950–2006 and projected ASPO/EWG
including power plants [22]. However, while the overall energy [14,16] limits, 2020–2100. Also shown are the CO2
efficiency of the world economy has risen greatly in recent emissions relative to the year 2000 emissions. The 15%
decades, as measured by Gross World Income/EJ, so has total and 50% limits correspond to the IPCC [25] limits for a 2 8C
energy use and emissions [2]. A second approach is to maximum temperature rise.
 international journal of hydrogen energy 34 (2009) 31–39 35

costs of final energy will accordingly rise. Much of this hard- environmental costs [38], major hydro electric expansion will
won delivered energy will then be needed for CCS. On the continue to face serious public acceptance issues. Recently,
other hand, resource limitations and CCS difficulties in the the serious environmental problems of the Three Gorges Dam
low energy future, although present, are far smaller, but in China have received attention, and more relocation is likely
would involve major societal change if technical efficiency because of slope instability as the watertable rises [46]. The
cannot deliver enough reductions in the very short time frame WETO report [20], even in the most favourable case, projects
available [26]. only 19 EJ hydro globally for 2050, compared with 10.7 EJ today.
Given that the economic potential is only 25–30 EJ [38], and is
nearing its limits in the OECD, this projection seems
4. Low fossil fuel energy supply mix reasonable.
In future, water availability will increasingly constrain
The He–Lf future could be based on renewable energy (RE) bioenergy production, whether it is produced from foodstuffs
sources and/or nuclear energy. Several researchers think that or purpose-grown cellulosic energy plantations. Around 40%
RE sources could readily supply the 800 EJ required [38,39]. of food production today is provided by irrigation, which
Fossil fuels would supply 200 EJ, only half today’s use, but still already requires massive overdrafts of fossil groundwaterd
unsustainable under ASPO/EWG assumptions (Fig. 1). If CO2 around 200 km3 globally. In both India and China, boreholes
emissions were held to 15% of year 2000 values, 10.7 Gt of CO2 are being sunk as far as 1 km, chasing falling watertables [47].
would need to be captured and sequestered. In a recent paper [48], we showed that for borehole depths of
We doubt that non-carbon sources could supply 800 EJ by greater than about 167 m, pumping energy alone would
2050, or even later. RE sources can be conveniently divided require all the biomass energy that could be grown from the
into two groups: intermittent and non-intermittent [35]. An extracted water. For greater depths, the energy return is
important problem facing the non-intermittent RE sources negative. In the US, much controversy exists as to whether
hydro and biomass is the effect of ongoing climate change on corn ethanol even gives a net energy benefit. But even the
their potential in 2050 and beyond. Since it is unlikely that we optimists agree that the net energy return is smalldand this
will be able to prevent a rise of 2  C or more above pre- for corn grown on prime rain-fed farmland.
industrial [5], significant climate changes will occur despite The conflict between food and bioenergy is already
even our best efforts. apparent. Corn prices have risen, as grain is increasingly
Hydropower production depends not only on average diverted to ethanol production [49]. This conflict can only
annual precipitation, but on its inter- and intra-annual vari- intensify in the future, if an expanding global population tries
ability, and on the frequency of extreme precipitation events. to increase agricultural and forestry production as well as
In general, while all general circulation models show bioenergy fuels in the face of the adverse hydrological
increased overall precipitation in a warmer world, many areas changes discussed above. Temperature increases during the
will have reduced rainfall. Decreased precipitation translates growing season can also adversely affect grain yields because
into decreased hydro potential, as has been shown for both of reductions in photosynthetic activity at high temperatures
southern Europe and tropical Africa [40,41]. [47]. In view of all these factors, Field and colleagues [50]
Climate change can influence hydro production in several recently estimated that only about 27 EJ of biomass could be
other ways. Intense rainfall events are expected to increase in harvested annually without threatening food supplies or
relative frequency, and such occurrences cause dispropor- worsening climate changedless than today’s use, which is
tionate soil erosion [42,43]. It follows that reservoir siltation in often produced unsustainably [21]. In an earlier paper [48] we
future will be even more of a problem than today, where 0.5– have presented a fuller case for our argument that global
1.0% of reservoir capacity is lost annually [43]. Increasing bioenergy potential (including waste and residues) will be of
siltation rates will both lower the output from existing hydro minor importance in an increasingly environmentally con-
schemes, and affect the economic viability of proposed new strained world.
schemes. The only other source of non-intermittent RE is geothermal
Global warming will both lower the share of precipitation energy. Conventional geothermal energy can only supply
that falls as snow on mountain ranges such as the Rockies and minor amounts of electricity [35], but considerably more direct
the Himalayas, and bring forward the discharge due to annual heat. However, a recent MIT study [51] saw enormous poten-
snowmelt, resulting in more temporally skewed streamflows tial for Enhanced Geothermal Systems. The total estimated
[44]. Unlike regional precipitation, increased regional resource base for the US alone is in the millions of EJ, far above
temperatures are predicted with high confidence [45] and estimated global reserves for fossil fuels, or even the total
higher temperatures will increase reservoir evaporation. The fossil fuel resource base [51]. Two parameters are crucially
effect of all these changes will be reduced hydro potential, not important for assessing the cost (in both money and energy
only because of projected reduced flows, but also because the terms) for developing these resources: the depth of the heat
uncertainty of flows and their variation under changing source, and its temperature. Electricity production needs
climate increase the economic risk of hydro projects. In some temperatures of at least 150–200  C. The report shows that
river basins, it may even be difficult to maintain existing there is no resource at or above 200  C at depths 4 km in the
hydro output, although for Arctic-draining rivers, flows and US, and only small amounts at 4–5 km depth. It also shows
hydro potential are anticipated to rise [44]. that the cost of wells rises roughly exponentially with depth,
With 40–80 million people displaced to make way for large so the energy costs of produced electricity will probably also
hydro schemes over the past half century and their often high show disproportionate increases with depth. Disappointingly,
36 international journal of hydrogen energy 34 (2009) 31–39

there was no discussion of energy costs in this report of over electricity rises, the ratio of primary to secondary energy will
300 pages. The IEA [11] give the global potential of geothermal fall, so that the 300 EJ of primary energy in the Le–Lf case will
electricity production as only 85 GW (about 2.4 EJ/year) over deliver more secondary energy than will the Le–Hf case. As has
the next 30 years. been noted, partly offsetting this will be the eventual need for
Another possible source of constant output energy is energy conversion and storage.
nuclear power, which in 2006 provided about 5.4% of global Even in the Le–Lf case, 0.7 Gt CO2 CCS would still be theo-
primary energy [1]. van der Zaan [10] looked at a 10-fold linear retically needed. In fact, on the assumptions used here, CCS
expansion of nuclear power between 2000 and 2075. Even this could only be avoided if fossil fuels provided 50.9 EJ or less of
optimistic assumption only gave nuclear energy providing primary energy. Such a level of fossil fuels would also be
15% of total (commercial) primary energy by 2030, and 20% by sustainable for several centuries (Fig. 1), making it an attrac-
2075. But Fieveson [52] points out that nuclear plants often tive option.
take decades to plan and build, and despite much talk, ‘there
is little evidence of a vast surge in construction before 2030,
the farthest point in time where the projections at least 5. Implications for the future of hydrogen
roughly can be based on actual plans’. Hence nuclear energy is
very unlikely to provide anywhere near 15% of primary energy Given the various possibilities for future energy, what do the
by 2030. The IPCC [25] also see negligible change in nuclear official reports mentioned in Section 1 have to say about
energy’s share by 2030. Finally, greatly expanding nuclear hydrogen production? Only the WETO study provides detailed
power worldwide inevitably means that some plants will have projections. For their hydrogen (H2) casedthe most optimistic
low security and safety standards. As Socolow and Pacala [53] for H2dthe report projects that production globally will rise
put it, ‘the world’s least well run plant can imperil the future from about 4.6 EJ in 2030 to 43.8 EJ in 2050, compared with total
of all the others’. The high energy costs of nuclear plant primary energy production in 2050 of 850 EJ [20]. The reference
construction also place limits on the rate of nuclear power case has H2 production in 2030/2050 of only 1.3/14.7 EJ. The
introduction [21]. hydrogen case assumes optimistic technology advances,
For nuclear energy, proven recoverable reserves of particularly in H2-based transport, where nearly all the H2
uranium are only 2210 EJ. Adding undiscovered conventional would be used. Even so, production would still only be 5.1% of
resources would still only give 7830 EJ [54], or roughly 3000 EJ primary energy demand in 2050. The EIA study does not give
of electricity [21]. Orders of magnitude higher values are total H2 production, but projects negligible use in 2030 in
theoretically possible if unconventional resources in phos- transport, even in the high oil price scenario [3]. The 2007 IPCC
phate rock and seawater are included, but such resources mitigation report contains little discussion on H2, but notes
would probably not deliver net energy in conventional reac- that it ‘will only begin to make an impact around 2050’ [22].
tors. Breeder technology could also extend uranium reserves, Researchers on H2 energy have far been more optimistic.
but these reactors are unlikely to make a significant contri- Momirlan and Veziroglu [8], in a paper discussing the transi-
bution before the middle of the century, since working tion from the present fossil fuel era to extensive H2 utilisation,
prototypes are not expected before 2035, and probably never, present an indicative graph showing 200 EJ H2 production in
given their many problems [8,9,35]. 2030 (about 23% of total primary energy), and 750 EJ in 2050
It seems likely that intermittent sources of RE, chiefly wind (75% of total energy). Interestingly, they forecast a decline in
and direct solar, will have to supply most non-fossil energy in primary energy production after 2050 if no H2 is produced.
2050 and beyond. The technical potential for these sources is Their scenario for 2050 is similar to our He–Lf future, having
undoubtedly very large [39], but would increasingly require 1000 EJ primary energy, with 250 EJ coming from fossil fuels
conversion (probably to hydrogen) and storage as they and the rest from RE converted to H2.
assumed steadily larger shares of total primary energy [35]. Possible routes to large-scale introduction of H2 can be
This would both raise costs and decrease net energy delivered. conveniently reduced to three. First, H2 could be introduced
A number of other problems, which may or may not prove because of various technical breakthroughs, either leading to
serious threats, include: public opposition to siting; the strong direct demand for H2, or to direct production of H2. Many
possible need to import energy from solar farms thousands of H2 researchers consider that because of oil depletion dis-
kilometres away; availability of fresh water for solar cell/ cussed above, transport represents the earliest opportunity
mirror cleaning and hydrogen production in desert areas; for large-scale introduction through the use of H2 fuel cells
dust/grit damage to solar cells and mirrors in desert areas; (FCs) (e.g. [9,20,28,57]). But H2 faces a number of competitors
deterioration of PV cell output with very high temperatures for supplying future transport energy, including non-
[55,56]. Trainer [56] has stressed the very high costs of solar conventional oils such as oil sands and liquid fuels from
energy, even in a favourable location such as northern Aus- biomass. Both are already in production, and are more readily
tralia. Wind and solar electricity in 2006 together amounted to fitted into the existing transport system. Recently, interest has
less than 0.8 EJ globally, or 1/1000 of the 800 EJ needed in the shifted to plug-in hybrid vehicles, largely powered by mains
He–Lf future. We simply do not know what difficulties will face electricity [58]. H2 even has a strong competitor for fuelling FC
such a huge scale-up. vehicles in methanol.
As RE sources that deliver primary electricity, such as Of course, these competitors also face serious obstacles to
wind, hydro and photovoltaic, increase their share of primary large-scale introduction, as already discussed for oil sands
energy, comparisons with a fossil fuel-based system become and biomass fuels. Plug-in hybrids are still not available, and
increasingly problematic. As the share of renewable primary the favoured lithium ion batteries have relatively short lives
 international journal of hydrogen energy 34 (2009) 31–39 37

[58]. Methanol is a liquid and so avoids the on-board storage 2. For the Le–Hf future, supply constraints, although far less
problems of H2dbut methanol FCs are a long way from likely than in case (1), might still be a problem (Fig. 1). If
commercial deployment. Several major technical break- emissions are kept to 15% of year 2000 levels, 13.5 Gt CO2 of
throughs are needed for H2-FC vehicles, and could occur, but CCS is still needed. As in the He–Hf future, coal gasification
such breakthroughs are just as likely in competitor vehicle might still be adopted for the small number of new plants
fuels. Nevertheless, hydrogen could well be an important fuel required, but the quantities of H2 produced would be much
for truck, bus, rail or ship propulsion, where on-board smaller.
hydrogen storage volumes are less constrained [20]. Similarly, 3. The He–Lf future would be most advantageous for
hydrogen could also be important for stationary FC power hydrogen, but like the He–Hf future, we argue is unlikely to
generation, but cost reductions are needed. An additional occur. If implemented, H2 production would mainly come
benefit here would be the potential for combined heat and from either the need for conversion and storage of large
power, giving very high overall efficiencies [9]. amounts of intermittent RE, or possibly from high-
If technology breakthroughs make direct photolysis temperature reactors. Fossil fuel use would be similar to
(water-splitting with a photocatalyst) [21], or biological H2 (but smaller than) the Le–Hf case, with similar H2 produc-
from algae or other organic substrates [59,60] economically tion possibilities.
feasible, H2 would be produced directly, and so could become 4. The Le–Lf future is in our view the most likely, and would
a preferred energy carrier. Combining the various approaches require little or no CCS. The 240 EJ from non-fossil sources
shows promise for increasing H2 production rates [61]. CCS would probably need large quantities of H2 production,
has also been seen as ‘the essential bridge to the hydrogen from conversion/storage of intermittent RE.
economy’ [62]. It is at least possible that any direct supply
could act synergistically with direct demand for H2 for In summary, we need to consider not only which energy
stationary or vehicular FCs and thus create a major opportu- pathways favour H2 production, but, just as importantly, how
nity for H2 even before 2050. Lewis [63], for example, envisages likely they are to occur. In general, the maximum future
H2 from nuclear reactors powering FC vehicles. production of hydrogen depends on total primary energy use,
Second, increasing the share of intermittent RE in elec- since as a derived fuel, production must be less than this. We
tricity grids would eventually require either dumping of have argued that high levels of energy, from whatever source,
electricity if excess to requirements, or else conversion to are unlikely by 2050, or even later, and that the Le–Lf case is the
some other energy form and storagedwith H2, a strong most probable future for energy in 2050 and later. H2 prospects
contender. Depending on the particular grid, this conversion also depend to an important extent on the shares of the
might need to occur at quite low levels of intermittent RE in various energy sources, with large amounts of intermittent RE
the energy mix, thereby hastening its introduction. most favourable. Nevertheless, breakthroughs in direct
Third, in what we have argued is the unlikely event [35] supply of H2, or H2-FCs, could allow much earlier introduction.
that non-intermittent RE sources such as hydro, biomass or
geothermal, together with nuclear power, account for most
energy production, conversion of electricity to H2 would 6. Conclusions
seem unlikely any time soon. H2, and new energy sources
generally, are late-comers and so face ‘entrenchment’ from Energy use in this century and beyond faces deep uncer-
established energy sources and energy carriers, namely tainties. There are widely conflicting opinions on the size of
fossil fuels and electricity. Only after electricity needs were ultimately recoverable fossil fuel reserves, and the extent to
fully met by CO2 emission-free sources would hydrogen which unconventional resources can be tapped. If, as expec-
production from any excess electricity generation be ted in most forecasts, fossil fuel use continues to grow, the
considered, since otherwise, available energy (and CO2 sequestration of vast amounts of CO2 would be needed if we
reductions) would always be maximised by using carbon- are to limit global warming. Large emitters such as power
free electricity directly [64]. Such might be the case for Ice- plants could probably only capture around a third of the
land, which already has 100% electricity from hydro and amount needed, requiring the deployment of air capture, an
geothermal, and no scope for directly exporting electricity. untried and energy-intensive technology.
However, export of energy-intensive aluminium is an Non-carbon sources face their own uncertainties. The
alternative to H2 production. future of nuclear energy depends heavily on the successful
We can now discuss the implications for H2 of the four case and timely development of either breeder reactors or fusion
studies: energy. Yet after nearly half a century of effort, neither are
near commercialisation, and fission technologies face deep
1. The He–Hf future is unlikely to lead to major H2 production, public opposition. Ongoing climate change will adversely
not only because of possible supply and environmental affect hydro and biomass energy expansion. Geothermal
constraints. Large CO2 emitters such as coal power plants energy could only be significant if EGS, another untried tech-
could only capture about one third of global emissions [30], nology, is deployed on a large scale. The potential for inter-
so large-scale air capture, an untried and energy-intensive mittent RE sources, wind and solar, is far greater, but is
technology, would be necessary. However, it is possible that unevenly distributed spatially, and both face orders of
coal gasification (and thus H2 production) could be adopted magnitude scale-up to be major energy suppliers. They will
for power plants, even without CCS, because of higher plant eventually also need conversion and storage, which will
efficiencies. greatly raise the costs of delivered energy. For all these
38 international journal of hydrogen energy 34 (2009) 31–39

reasons, we have argued that a low energy future is more [15] Simmons M. Shock to the system: the impending global energy
likely. supply crisis. Harvard International Review 2006;28(3):62–5.
This study considered four illustrative cases: a high [16] Energy Watch Group (EWG). Coal: resources and future
production. EWG series No. 1/2007; 2007.
(1000 EJ) and a low (300 EJ) energy future, and for each of these
[17] Tao Z, Li M. What is the limit of Chinese coal suppliesda
conditions, a high (80%) and low (20%) fossil fuel energy share. STELLA model of Hubbert peak. Energy Policy 2007;35:3145–54.
We argue that high energy futures are very unlikely, because [18] Rutledge D. Hubbert’s peak, the coal question and climate
of resource limitations, and rising money and energy costs per change, http://rutledge.caltech.edu/; 2007 [accessed 08.04.08].
unit of delivered energy as annual energy demand rises. The [19] Strahan D. The great coal hole. New Scientist 19 January
low energy cases are technically more achievable, although 2008:38–41.
both require large cuts in energy use, and most probably, [20] European Commission. World energy technology outlook-
WETO H2, EUR 22038. Directorate General for Research.
lifestyle changes in high energy use countries. H2 production
Available from: http://ec.europa.eu.research/energy/pdf/
would be best favoured in the low fossil fuel options, with weto-h2_en.pdf; 2006.
production both greater, and implemented earlier, in the high [21] Lewis NL, Nocera DG. Powering the planet: chemical
energy use case. But in contrast to the Le–Lf case, which we see challenges in solar energy utilization. Proceedings of the
as the most probable future underpinning a hydrogen National Academy of Sciences of the United States of
economy, H2 production is least likely in the Le–Hf case. America 2006;103(43):15729–35.
[22] Sims REH, Schock RN, Adegbululgbe A, Fenhann J,
Konstantinaviciute I, Moomaw W, et al. Energy supply. In:
Metz B, Davidson OR, Bosch PR, Dave R, Meyer LA, editors.
Acknowledgements
Climate change 2007: mitigation. Cambridge, UK: CUP; 2007.
p. 251–322.
Patrick Moriarty would like to acknowledge the financial support [23] Woynillowicz D. Tar sands fever. World Watch 2007;20(5):813.
of the Australasian Centre for the Governance and Management [24] John O’Mara Bockris. Will lack of energy lead to the demise of
of Urban Transport (GAMUT) in the preparation of this paper. high-technology countries in this century? Intenational
Journal of Hydrogen Energy 2007;32:153–8.
[25] Barker T, Bashmakov I, Bernstein L, Bogner JE, Bosch P,
references Dave R, et al. Technical summary. In: Metz B, Davidson OR,
Bosch PR, Dave R, Meyer LA, editors. Climate change 2007:
mitigation. Cambridge, UK: CUP; 2007. p. 25–93.
[1] BP. BP Statistical review of world energy 2008. London: BP; 2008. [26] Moriarty P, Honnery D. Mitigating greenhouse: limited time,
[2] International Energy Agency (IEA). Key world energy limited options. Energy Policy 2008;36:1251–6.
statistics 2006. Paris: IEA/OECD; 2006. [27] Hansen J, Sato M, Kharecha P, Beerling D, Masson-
[3] Energy Information Administration (EIA). International Delmotte V, Pagani M, et al. Target atmospheric CO2: where
energy outlook. US Department of Energy. Available from: should humanity aim?, www.arxiv.org/abs/0804.1126; 2008
http://www.eia.doe/gov/oiaf/ieo/pdf/0484(2007).pdf; 2007. [accessed 05.05.08].
[4] United Nations (UN). UN world population prospects, the [28] Mason JE. World energy analysis: H2 now or later? Energy
2006 revision. Available from: http://esa.un.org/unpp/; 2006 Policy 2007;35:1315–29.
[accessed 07.04.08]. [29] Heimann M, Reichstein M. Terrestrial ecosystem carbon
[5] Riahi K, Grubler A, Nakicenovic N. Scenarios of long-term dynamics and climate feedbacks. Nature 2008;451:289–92.
socio-economic and environmental development under [30] Pearce F. Cleaning up coal. New Scientist 29 March 2008:36–9.
climate stabilization. Technological Forecasting & Social [31] Kobayashi H. Climate change and future options for carbon
Change 2007;74:887–935. sequestration. Foresight: the Journal of Future Studies,
[6] Lincoln SF. Fossil fuels in the 21st century. Ambio 2005;34(8): Strategic Thinking and Policy 2004;6(3):153–62.
621–7. [32] Keith DW, Ha-Duong M, Stolaroff JK. Climate strategy with
[7] Hoffert MI, Caldiera K, Benford G, Crisswell DR, Green C, CO2 capture from the air. Climate Change 2006;74:17–45.
Herzog H, et al. Advanced technology paths to global climate [33] Baciocchi R, Storti G, Mazzotti M. Process design and energy
stability: energy for the greenhouse planet. Science 2002;298: requirements for carbon dioxide capture from air. Chemical
981–7. Engineering and Processing 2006;45:1047–58.
[8] Momirlan M, Veziroglu TN. The properties of hydrogen as fuel [34] Parson EA. Reflections on air capture: the political economy
tomorrow in sustainable energy system for a cleaner planet. of active intervention in the global environment. Climate
International Journal of Hydrogen Energy 2005;30:795–802. Change 2006;74:5–15.
[9] Marban G, Valdes-Solis T. Towards the hydrogen economy? [35] Moriarty P, Honnery D. Intermittent renewable energy: the
International Journal of Hydrogen Energy 2007;32:1625–37. only future source of hydrogen? International Journal of
[10] van der Zaan BCC. Nuclear energy: tenfold expansion or Hydrogen Energy 2007;32:1616–24.
phaseout? Technological Forecasting & Social Change 2002; [36] Hultman NE, Koomey JG. The risk of surprise in energy
69:287–307. technology costs. Environmental Research Letters 2007;2:
[11] International Energy Agency (IEA). Renewables in global 1–6.
energy supply: an IEA fact sheet. Paris: IEA/OECD; 2007. [37] Moriarty P, Honnery D. Low mobility: the future of transport.
[12] Krewitt W, Simon S, Graus W, Teske S, Zervos A, Schafer O, Futures 2008;. doi:10.1016/j.futures.2008.07.021.
et al. The 2  C scenarioda sustainable world energy [38] Moriarty P, Honnery D. Can renewable energy avert global
perspective. Energy Policy 2007;35:4969–80. climate change? Proceedings of the 17th international clean
[13] Weizsackervon E, Lovins AB, Lovins LH. Factor 4: doubling air & environment conference, Hobart (Australia); 2005.
wealthdhalving resource use. London: Allen & Unwin; 1997. [39] de Vries BJM, van Vuuren DP, Hoogwijk MM. Renewable
[14] Association for the Study of Peak Oil and Gas (ASPO). ASPO energy sources: their potential for the first-half of the 21st
Newsletter (Also earlier newsletters). Available from: www. century at a global level: an integrated approach. Energy
peakoil.net, May 2008;No. 89 [accessed 14.05.08]. Policy 2007;35:2590–610.
 international journal of hydrogen energy 34 (2009) 31–39 39

[40] Lehner B, Czisch G, Vassolo S. The impact of global change DE-AC07–051D14517; 2006. Available from: http://
on the hydropower potential of Europe: a model-based geothermal.nrel.gov.
analysis. Energy Policy 2005;33:839–55. [52] Feiveson HA. Faux renaissance: global warming, radioactive
[41] Harrison GP, Whittington HW, Wallace AR. Sensitivity waste disposal, and the nuclear future. Arms Control Today
of hydropower performance to climate change. 2007;37(4):13–7.
International Journal of Power and Energy Systems 2006; [53] Socolow RH, Pacala SW. A plan to keep carbon in check.
26(1):42–8. Scientific American September 2006:28–35.
[42] Thacker PD. As the planet heats up, will the topsoil melt [54] Price R, Blaise JR. Nuclear fuel resources: enough to last? NEA
away? Journal of Soil and Water Conservation 2004;59(1): News 2002;20(2):10–3.
12A–5A. [55] Fathi NY, Salem AA. The reliability of the photovoltaic
[43] Pritchard S. Overloaded. International Water Power & Dam utilization in southern cities of Libya. Desalination 2007;209:
Construction 2002;54(7):18–22. 86–90.
[44] Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, [56] Trainer T. Renewable energy cannot sustain a consumer
Hanson CE, editors. Climate change 2007: impacts, society. Dordrecht, NL: Springer; 2007.
adaptation and vulnerability. Cambridge, UK: CUP; 2007. [57] Kruger P. Electric power required in the world by 2050 with
[45] Barnett TP, Adam JC, Lettenmaier DP. Potential impacts of hydrogen fuel productiondrevised. International Journal of
a warming climate on water availability in snow-dominated Hydrogen Energy 2005;30:1515–22.
regions. Nature 2005;438:303–9. [58] Anderson M. Souped-up battery prepares to slay the gas-
[46] Macartney J. Three gorges dam is a disaster in the making, guzzlers. New Scientist 1 March 2008:28–9.
China admits. The Times. Available from: http://www. [59] Das D, Veziroglu TN. Advances in biological hydrogen
timesonline.co.uk/tol/news/world/article2537279.ece 27 production. International Journal of Hydrogen Energy 2008;.
September 2007. doi:10.1016/j.ijhydene.2008.07.098.
[47] Brown LR. Plan B 3.0: mobilizing to save civilization. NY: WW [60] Manish S, Banerjee R. Comparison of biohydrogen
Norton & Co.; 2008. production processes. International Journal of Hydrogen
[48] Moriarty P, Honnery D. Global bioenergy: problems and Energy 2008;33:279–86.
prospects. International Journal of Global Energy Issues 2007; [61] Melis A, Melnicki MR. Integrated biological hydrogen
27(2):231–49. production. International Journal of Hydrogen Energy 2006;
[49] Sakr Y. The fight between fuel and food. Risk Management 31:1563–73.
2007;54(4):64–8. [62] Simbeck DR. CO2 capture and storagedthe essential bridge
[50] Field CB, Campbell JE, Lobell DB. Biomass energy: the scale of to the hydrogen economy. Energy 2004;29:1633–41.
the potential resource. Trends in Ecology and Evolution 2008; [63] Lewis D. Hydrogen and its relationship with nuclear energy.
23(2):65–72. Progress in Nuclear Energy 2008;50:394–401.
[51] Massachusetts Institute of Technology. The future of [64] Moriarty P, Honnery D. Car travel in Australia: an uncertain
geothermal energy: impact of enhanced geothermal systems future. Proceedings of the 30th ATRF, Melbourne, 25–27
(EGS) on the United States in the 21st century. DOE Contract September 2007.