Energy 64 (2014) 506e512
Contents lists available at ScienceDirect
Energy
journal homepage: www.elsevier.com/locate/energy
A top-down approach to assess physical and ecological
limits of biofuels
Carlos de Castro a, *, Óscar Carpintero b, Fernando Frechoso c, Margarita Mediavilla d,
Luis J. de Miguel d
a
Department of Applied Physics, Campus Miguel Delibes, University of Valladolid, Spain
Department of Applied Economics, Av. Valle Esgueva 6, University of Valladolid, Spain
Department of Electric Engineering, Fco. Mendizabal, University of Valladolid, Spain
d
Department of System Engineering and Automatic Control, Paseo del Cauce, University of Valladolid, Spain
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 4 May 2013
Received in revised form
11 September 2013
Accepted 17 October 2013
Available online 9 November 2013
The aim of this article is to analyse the physical and ecological limits of biofuels (in particular, ethanol). To
this end, three aspects are discussed. First of all, the territorial element, i.e., the real productivity of
energy crops (in particular, ethanol), to show that, in general, the productivity is much lower than most
of the literature on the subject suggests. Similarly, we then stress the high total territorial impact (the
Ecological Footprint) associated with this kind of energy production. Thirdly, we offer data concerning
the very modest nature of the energy balance of biofuels (EROEI) and their real power density,
considering the productivity of the crops grown for biofuel. As a comparison, these three aspects are
estimated for other forms of energy, in particular, photovoltaic energy. We conclude that when the set of
estimated parameters has been analysed, there exist reasonable doubts concerning the use of biofuels on
a regional and global scale, so they should not, in principle, be promoted as a renewable energy source,
nor is it desirable on such a scale.
Ó 2013 Elsevier Ltd. All rights reserved.
Keywords:
Biofuels potential
Energy return
Ecological footprint
Biofuels productivity
1. Introduction
There is currently an important social, economic, political,
ecological and technological controversy surrounding the use of
biomass as a biofuel to produce energy. There are many opposing
studies and discussions in the scientific literature concerning the
production and use of these biofuels in such matters as: the
emissions of gases associated with the greenhouse effect [1e6], the
energy return (EROEI) [1,7e10] and other energy related issues
such as the exergetic evaluation [11e14], the real yields productivity [15e18], and the environmental, economic and social costs
and benefits [17,19e29].
Given that the use of biomass would, in principle, be included
within the scope of renewable energies, we should take into account not only the usually stressed advantages [1,18,23], but also
the evaluation of the physical and ecological limits associated with
its use.
* Corresponding author.
E-mail address: ccastro@termo.uva.es (C. de Castro).
0360-5442/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.energy.2013.10.049
Following a holistic approximation, top-down methodologies
have been proposed and explained for wind potential [30,31] and
solar power on a global scale [32]. Prieto and Hall [33] use also a
top-down approach based on the whole Spanish photovoltaic
installed base instead of the classic individual plants of a given
technology of a given latitude (bottom-up methodology) and use
this approach to analyze the extended boundaries of the energy
inputs required to the whole photovoltaic system.
Bottom-up methodologies are based on available concrete cases
and estimations of the land productivity, power density, EROEI,
exergy, etc. of a particular area of a region or country, which are
generalized and extrapolated to all this region or country and
sometimes to the World as a whole (see Table 1). Against this
methodology, we propose [30,32] a top-down methodological
approach to assess the very same issues, considering global and/or
regional statistics, for instance, for biofuel production: satellite data
of the total land occupation of plantations, total biofuel production,
etc. (see next sections); always with a holistic approach that takes
into account related issues like: eroded land, use of water and
materials, global greenhouse emissions, etc.
In this work we aim to use this same methodology for the
biomass case, making a global analysis of some of the most
C. de Castro et al. / Energy 64 (2014) 506e512
Table 1
Reported or inferred land yields for sugarcane and biofuel yields from Brazilian
sugarcane and the World. See text and Appendix.
Authors
Land yield
(Tn/Ha)
Biofuel yield
(L/Ha)
Connor & Hernández [34]
73.5 (Brazil)
60.7 (India)
82.4 (Brazil)
5475 (Brazil)
4522 (India)
6500 (Brazil)
5300 (India)
5500 (Brazil)
6712 (Brazil)
6314 (Brazil)
6250 (Brazil)
6641 (Brazil)
5227 (World) 5684
(Brazil)
4737e5559
(Brazil) (harvested)
Worldwatch Institute [15],a
Pimentel & Patzek [35],b
Macedo et al. [36],b
Ramírez-Triana [7],b
Giampietro [37]
Balat & Balat [18],c
Johnston [16],d
This work
e
77
77
77
75
(Brazil)
(Brazil)
(Brazil)
(Brazil)
64.5(World), 70.2
(Brazil)
60.2e69.9
(Brazil)
a
Notes: These are the values given as “typical yields” in Table 3.1 of the book
(bottom-up methodology). However, from Table 23.1 of the same book, from the
data given there (Brazil total land occupation of sugarcane, % dedicated to ethanol
production, Brazil ethanol production), the productivity for Brazil can be calculated
as 5213 L/Ha (top-down methodology).
b
The data of Pimentel & Patzek [35] and Macedo et al. [36] have been taken from
the revised work of Ramírez-Triana [7] dedicated to the calculation of the EROEI of
four authors. Pimentel & Patzek [35] and Macedo el al [36]. give the extreme values.
c
Balat & Balat [18] give the productivity for the ethanol produced from sugarcane
or other crops from different regions and countries; from their data, the ethanol
average for Brazil, China, the EU27 and North America is 4223 L/Ha. The great difference is due to the fact that the productivity of the sugarcane of Brazil is much
higher than that of other countries and that of other crops (mainly corn).
d
Johnston [16] provides data concerning the bioethanol produced by the sugarcane from around the world, while only the average is shown here.
e
To be able to compare data with other authors, we give here our estimated value
of the productivity of real harvested land. However, as we explain in the text, the
total real production per hectare dedicated to crops (what is grown every year is not
always harvested) is lower. We believe that the latter should be the real “productivity” which should be used in the global calculations. See the text for more detailed
explanations.
important aspects to be taken into account in its generalized use
for biofuels.
2. Theoretical calculations and results
2.1. Yield productivity of land for biofuels
Let us commence by remembering the stated productivity from
different authors (or at least what is inferred from their work) for
the bioethanol from sugarcane in Brazil or other countries and regions (Table 1). We also compare these data with those provided by
Johnston et al. [16] for the world average and with what we
calculate in this work for Brazil, a country where the productivity is
above the average.
Johnston et al. [16] show that the productivity of many crops has
been overestimated (and this still happens) by 100% or more using
a methodology based on actual, reported yields derived by
spatially-explicit agricultural census data. In addition, the updating
of the studies cited in this work concerning the average values
found by Johnston et al. [16], in general, not only lower the productivity of the land, but also the calculations which depend on this
average, or typical, productivity. This occurs, for instance, in the
case of the EROEI or the balance of greenhouse gases with respect
to the consumption of fossil fuels. In fact, the work by Johnston et al.
[16] can even be considered to be optimistic since its world average
is close to the performance of Brazil, while the value it gives for
Brazil is higher than the highest we have found.
In effect, if we take Brazil as a reference, as the world’s principal
producer of biofuels together with the USA and a big region, the
507
data are revealing. According to the Brazilian Ministry of Agriculture, Livestock and Food Supply [38], and taking five year periods to
avoid the influence of anomalous years, the area of sugarcane
harvested in Brazil went from an average of 4.9 million hectares in
the period 1997e2001 to 5.61 million hectares in the period 2002e
2006, to finish in 8.18 million hectares in the period 2007e2011.
Meanwhile, the production of crushed sugarcane for the same periods went from 295.04 million tonnes to 373.19 million tonnes, to
finish in 568.12 million tonnes. In terms of crop productivity, this
supposes an increase from the 60.24 Tonnes/hectare in the first of
the abovementioned periods, to 69.86 Tonnes/hectare in the last
period.
However, not all the sugarcane crop is dedicated to the production of ethanol: an important part is also dedicated to the
production of sugar. If we concentrate on the TRS (Total Recoverable Sugar), that is, the part of the sugarcane crop for industrial use
(ethanol or refined sugar), the progressive move towards ethanol
has become evident over the last few years (Fig. 1).
In the case of ethanol, the same reference data allows us to
obtain its productivity for Brazil (see Appendix) and also to lower
the performance figures obtained in the previous works. Thus, if we
consider the three abovementioned five year periods, the average
production of ethanol in Brazil grew from 12.8$109 L to 25.1$109 L.
With these figures, the country’s productivity was between 4737
and 5559 L/Ha, similar to the conservative estimates of Connor &
Hernández [34] and of Pimentel & Patzek [35], and some way below
the rest of the estimates included in Table 1.
The statistics used by the authors of Table 1, and those that we
ourselves have previously used, are based on the harvested land
dedicated to sugarcane. Although the productivity calculations are
made on the basis of the harvested surface (what the Brazilians call
“disponivel para colheita”), it is necessary to look closely at the total
territorial requirements (both direct and indirect) that the cultivation of sugarcane has through the cultivated area (which, besides
the harvested surface in any year, also includes what is called “em
reforma”, or not yet available for harvest). The total area occupied
by the sugarcane crop increases on average when this last type of
land is included. In the case of Brazil, there are no complete statistics of the differences between the real territory occupied by the
sugarcane crop and the harvested crop. However, for the state of
Sao Paulo (with over 50% of Brazil’s production), there are recent
measurements taken using advanced satellite data tools [39],
which have already been incorporated into open-access databases
(UNICA [40], 2012). A comparison between the cultivated areas and
Fig. 1. Brazilian destination of the total recoverable sugar of sugarcane.
Source: Brazilian Ministry of Agriculture, Livestock and Food Supply [38]. Own
calculations.
508
C. de Castro et al. / Energy 64 (2014) 506e512
the harvested areas in the state of Sao Paulo between 2003 and
2011 show that the average harvested area is 91% of that of the
cultivated area. If this proportion is maintained for the whole of
Brazil, the litres of ethanol obtained per cultivated hectare would be
in the range of 4300e5100.
On the other hand, even the productivity of sugarcane on a
regional, national and world scale are often conditioned by negative factors that reduce the crop’s productivity (e.g., such climatic
problems as unexpected and excessive rainfall during the harvest
season in some areas, unexpected flowering or frost, etc.). Such
circumstances have always existed and will continue to do so, and
such events must be taken into account. For instance, climate
change itself may make unexpected climatic and ecological conditions more probable and more frequent.
2.2. Biofuels energy return (EROEI)
The considerations of the above section tend to decrease the
EROEI calculated by other authors, as they are usually based on a
higher productivity of the cultivated land than the real one. Of
course, if the productivity per hectare in the calculations of the
EROEI is reduced, both the outputs and energy inputs in the tables and the LCA (life cycle assessments) used for such calculations are also reduced. Nevertheless, they will not be reduced in
the same proportion and thus the EROEI estimated by the
different authors will not be conserved. This is so because, for
instance, if it is estimated that a typical mill needs 22,000 ha,
supposing a productivity of 80 tonnes/ha [15], then it is probable
that with a smaller productivity, the average mill will really need
more hectares and the kilometres covered by trucks from the
processing plant to the fields where the crop grows, cover, on
average, greater distances. Other energy inputs such as those due
to the use of fertilizers, insecticides, etc., which have been used
on crops that, in the end, turn out to have a low productivity, or
none at all, must also be taken into account for the regional EROEI
calculations.
Thus, for instance, Ramírez-Triana [7] aims to compare and
standardise the results of four studies of the EROEI of ethanol from
Brazil [35,36,41,41]. To do so, he supposes a productivity of 77
tonnes/Ha (6314 L/Ha). With this theoretical productivity he reduces or increases the outputs calculated by the different authors,
while leaving the calculated inputs unchanged. He thus reaches
EROEIs of 2.63 (Pimentel & Patzek [35]), 3.19 (Oliveira et al. [41]),
7.52 (Macedo et al. [36]) and 8.84 (Boddey et al. [42]).
Following the same methodology and standardisation as in
Ramírez-Triana [7], but with the productivity we have calculated
here, we would obtain lower values for all of them. The dispersion
in the EROEI calculations is so high (see Table 2), it is clear that, at
least in part, the calculations are biased.
There are frequently mutual accusations concerning the overestimation or underestimation of some inputs. Thus, for example,
Giampietro et al. [49] argue that the boundaries of the inputs
should be widened to include the energy due to the human work
effort, while others (Macedo et al. [36]) suggest that this should
not be considered because such persons would necessarily
consume the same energy in another place. At the same time, the
latter authors argue that all the by-products of the manufacture of
the biofuels should be considered as outputs (for instance the
glycerine manufactured), whether they are used to produce energy or not. However, the former authors have the exact opposite
belief, reasoning that energy is not obtained with these byproducts and that, in fact, their manufacture, use and disposal
consumes energy.
As things stand, if the energy of the by-products is included, the
energy associated with the human work effort in the case of biofuels should also be included, since it is probable that this work
effort is much higher than that required by other sources of energy.
Another by-product not usually considered in the studies of the
EROEI and which, in our judgement, is correctly estimated in Dong
et al. [46], consists of quantifying the soil loss of the crops and,
from that, the energy content of the organic material which
“disappears” with it. For the case of wheat in China, the calculation
of Dong et al. [46] gives 7.83 GJ/Ha$yr. A similar approximate
calculation for the bioethanol of Brazil can be estimated by taking
into account the fact that about 12 tonnes of soil per hectare/year
are lost [50] (chapter 7); that, with an average organic material
content of 5% (Bianchi [51] gives values of 0.72e18.9% as extremes,
and an average of 7.7% for soils of Parana State) and taking an
energy content of 20MJ/kgC (Dong [46] takes 22.6 MJ/kg of organic
matter), this would suppose about 12GJ/Ha-yr of losses, which
would be around 10% of the energy content of the ethanol produced. Pereira & Ortega [43] give 1.82 kg of topsoil eroded per litre
of ethanol, which would give 1.82 GJ/L or a loss of 8% of the energy
content of the ethanol itself.
If this energy loss were included in the studies of the EROEI cited
by Ramírez-Triana [7] and with his standardisation, the highest
EROEI (that of Boddey et al. [42] (8.84)) would come out at 5.1.
Further, if we take into account the fact that the real productivity is
lower than that taken by Ramírez-Triana [7]; the result would be
even smaller.
It would seem reasonable to believe that, with these criteria, the
highest estimates of the EROEIs for the ethanol of Brazil could be
overestimated. Furthermore, if we also take into account the fact
that the Brazilian EROEI for ethanol is probably higher than that of
other countries such as India (with a lower productivity), our estimate in Table 2 could also seem optimistic.
In this estimation, we should also take into account the
differences due to calculations of the EROEI based on bottom-up
Table 2
Some reported EROEIs and the EROEIs used in this work for biofuels.
Source
Reported EROEIs or output energy/Input energy
This work
(optimistica)
Sugarcane
ethanol Brazil
Corn ethanol U.S.
Giampietro & Mayumi [10]: 1.5, Oliveira et al. [41]: 3.7, Pereira & Ortega [43]: 7.2, Macedo et al. [36]: 8.9, Triana [7]: 2.63e8.84
5
Pimentel & Patzek [35]: 1.12, Pimentel et al. [44]: 0.69, Oliveira et al. [41]: 1.1, Hill et al. [23]: 1.25, Worldwatch Institute [15]:
1.34e1.38, Wang [45]: 0.7e1.5
Dong et al. [46]: 0.18e1.48 (China), Giampietro & Mayumi [10]: 1.1 (Italy), Dong et al. [46]: 1.19 (Italy)
1.25
Pimentel et al. [44]: 0.92, Pleanjai & Gheewala [47]: 2.42e3.58, Yee et al. [48]: 3.54
Yee et al. [48]: 1.44
Hill et al. [23]: 1.93, Worldwatch Institute [15]: 1.43e3.4
Dong et al. [46]: 1.09
3
1.5
Corn ethanol
(other countries)
Palm oil
Rapeseed oil
Soybean
Wheat China
a
The EROEI shown here is not a calculation but a rough estimate based on the analysis of the reported EROEIs that are considered optimistic (see text) and which will be
used as global world average values for four sources: sugarcane, corn, palm oil and other sources.
509
C. de Castro et al. / Energy 64 (2014) 506e512
methodologies with respect to the results that could be given by
top-down methodologies. Normally, in bottom-up EROEI calculations, statistics are gathered from studies of concrete cases which
are extrapolated to the country or region and type of crop being
dealt with. The methodology proposed here would try to make the
estimation on the basis of global or regional data. In this way, if we
look, for instance, at the “Brazilian Energy Balance” provided by
Brazil’s own Ministry of Mines and Energy [52] (page 46, 2011), we
can see that the production of crushed sugarcane in 2010 was
160.33 Mtm, and whose power content, taking that obtained per
metric tonne according to Beharry [53], would be around 39 million
metric tonnes oil equivalent or 1.64 EJ. However, the official data
show that the energy used to produce that crushed sugarcane is
equivalent to 13.2 million metric tonnes oil equivalent (page 68)
(0.55 EJ). This gives an EROEI for the crushed sugarcane of around 3.
In 2010, 50.5% of the production of crushed sugarcane was dedicated to ethanol production and the rest to sugar. In Brazil, it is
normal for the production of sugar and ethanol to be carried out in
the same factories, so it is, in practice, very difficult to calculate how
much energy is used in each process. If we suppose that the energy
needed to produce ethanol and sugar is similar, then the EROEI
factor of the crushed sugarcane for ethanol of 3 would be
conserved. This means that the EROEI of a derivative of the crushed
sugarcane (whether ethanol or sugar) must be lower, even though a
part of the crushed sugarcane is used as an energy source to produce ethanol.
From these figures, it can be concluded that, either the Brazilian
ethanol has an EROEI <3 or that the data provided by the said
report are not correct. In the first case, it would point in the direction mentioned in this work, that our calculations and estimations are still optimistic.
On the other hand, it is important to take into account the fact
that the productivity of Brazil or the USA for ethanol depends
mainly on the importation of fossil fuels from around the world
associated with the growing quantities of fertilizers per hectare
[54]. Thus, it would be difficult to achieve such productivity for
the world as a whole, given the nature of the Earth’s closed
system. So this increasing fertilization, which depends on the rest
of the world, tries to compensate precisely the ever increasing
loss of nutrients that occurs with the working of the land and the
harvesting. Neither could this be compensated for by the majority
of the world’s arable land being given over to the production of
ethanol.
2.3. Power density of biofuels
We shall now look at the real power density of biofuels. A good
top-down approach taking socio-ecological and real biomass energy yields is what Campbell et al. [55] did. These authors argued
that “the area with the greatest potential. that reduces net
warming and avoids competition with food production is abandoned/degraded lands”.
From this premise, Campbell et al. [55] calculate that the
extractable gross heat content in 386 MHa would be 27 EJ/yr
(0.86 TW) or 0.222 Wheat/m2 (Wheat is calorific power). Although
they suggest that there are less than 386 MHa available for cultivating biofuels, they do not take into account the land that is now
used as pasture and which the expected increase in degradation,
population and consumption will involve their use for other ends.
Obviously, if this energy were used to produce electricity with 10e
30% efficiency in the conversion [53] or biofuels with an efficiency
of approximately 15% (see below), there would be less than 10 EJ of
useful energy in the form of electricity, or less than 5 EJ in the form
of biofuels (the latter with a very low power density of <0.033 W/
m2 of occupied land). Of course, highly productive land is currently
being used to produce biofuels, which is why the power density
could be greater (but also unsustainable, according to Campbell
et al. [55] own criteria).
In this sense, Table 3 calculates the power densities that can be
inferred from the work of other authors and they are compared
with what is offered in this article. For the case of the highly
developed bioethanol of Brazil, and taking into account the fact that
767 GJ/Ha$yr (2.43 Wheat/m2) is the power content of 75 Tn/Ha of
sugarcane [53], the result is that the power content of the ethanol
with respect to the power content of the sugarcane is approximately 15% (calculations from Table 1), and even less if we take into
account the energy losses during manufacture.
Thus, a top-down calculation of the power density for biofuels
should begin with an estimation of the real area used for the
world production of biofuels. However, the available estimates of
the power density are based on a hypothetical mean productivity
of a particular region or country; in general, they are based on
concrete cases that use the bottom-up methodology. As argued
above, there are doubts about this methodology and, therefore,
the estimates of global land occupation for biofuels are normally
underestimated.
For instance, UNEP [56] estimates that 36 MHa were dedicated
to fuel crops in 2008 and points out that 64.5$109 L of ethanol and
11.8$109 L of biodiesel were produced. Multiplying by their
respective energy contents (see note on Table 3), a gross energy of
1.75 EJ or 0.155 W/m2 was obtained for that year. In reality, the
power density is less as more than 36 MHa were surely used to
produce that much, as indicated by our top-down calculations for
Brazil. The methodology used by the UNEP [56] is bottom-up and
has been shown to be conservative, as the UNEP itself recognises in
the light of the work by Johnston et al. [16]: “some of the land
requirement data. may represent rather conservative estimates”
(UNEP [56], page 63). In addition, the net energy is inferior to the
gross energy due to the EROEI ratios. Taking into account the
optimistic EROEIs estimated in Table 2, the biofuel production of
Table 3
Power densities calculated from the productivities reported or inferred by different authors.
Source of biofuel
Inferred gross power densities (W/m2)
Gross power densities:
This work (W/m2)
Sugarcane
0.47 (Macedo et al. [50], Brazil), 0.37 (Worldwatch Institute [15], India), 0.35
(Johnston et al. [16], World mean)
0.264 (Balat and Balat [18], US) 0.141 (Balat and Balat [18], China), 0.12
(Johnston et al. [16], World mean)
0.627 (Worldwatch [15], Malasya), 0.29 (Johnston [16], World mean)
0.06 (Hill [23]), 0.073 (Worldwatch [15], Europe), 0.043 (Johnston et al. [16], World mean)
0.155 (UNEP [56])
0.30e0.36 (Brazil)
Corn ethanol
Palm oil
Soybean
Global biofuelsa
Note: the light heat content is: 22.1 MJ/L for ethanol and 33 MJ/L for biodiesel.
For our Brazil calculations, we take the range 4300e5100 L/Ha estimated before.
a
See text for the calculation.
<0.15
510
C. de Castro et al. / Energy 64 (2014) 506e512
the different world regions provided by the UNEP [47] (reflected in
Table 4) and the net energy production of an occupied surface area
greater than 36 MHa; then we reach a global net power density of
biofuels lower than 0.073 W/m2, which could be bigger than the
estimation for abandoned land using Campbell’s [55] criteria
(0.033 W/m2), but which confirms roughly the estimations of
Giampietro and Mayumi [10] (<0.05 W/m2).
Thus, the net 0.073 W/m2 is an optimistic calculation, because
we have used an optimistic value for the gross power density (see
Table 3) and because we have used the high spectrum for the
calculated EROEIs. In comparison, from Prieto and Hall [33] data we
could calculate the electric power density produced at approximately 4.8 We/m2 in the Spanish photovoltaic parks, that with a
EROEI of 2.41, gives a net energy density of 2.8 W/m2, roughly 40
times bigger.
2.4. Ecological Footprint of biofuels
The inverse of the power density is related to the land requirements for delivering this power. This idea allows us to directly
approach the concept of the Ecological Footprint [57], which has a
great political relevance as it is often used by policymakers [58].
In the case of fossil fuels, the Ecological Footprint (EF) calculates
the associated emissions of CO2, subtracts a percentage which is
absorbed by the oceans, and estimates the woodland surface area
needed to absorb this CO2 from the atmosphere. The result is given
in “global” hectares (gHa) per capita, which, for 2008, was 1.47 gHa/
per capita [59].
In the case of biofuels, Oliveira et al. [41], compare the footprint
of gasoline with other biofuels in concrete cases. In this sense, and
for what interests us now, we can calculate this Ecological Footprint
(EF) per unit of power potential (the inverse of the calculated power
density) on a global scale. The approximate result is shown in
Table 5. In the case of biofuels, it is also necessary to add to the
calculated EF both the emissions associated with the fossil fuels
used in their manufacture and those associated with the change of
use in the land and its erosion, so our result is clearly conservative.
For the renewable energies, the EF per watt generated can also
be calculated in this way, considering the inverse of its power
density and applying the factor that, according to the methodology
of the EF, is assigned to each type of land or surface area in order to
convert it to gHa (see Table 5).
In the light of this impact indicator, it can be seen that the biofuels
would be surprisingly much worse than even the fossil fuels.
3. Discussion and conclusions
The following table (Table 6) shows a summary of the results
found in this work:
It is clear from these figures that, in general, biofuels should not
be considered by policymakers as a source of renewable energy,
since the energy flow and materials needed for their production
include not only the land but also soil loss and a considerable
Table 5
Land needs per watt produced and Ecological Footprint per generated watt and per
capita to replace 12 TW of fossil fuels.
Source (w12 TW)
m2/W
gm2/W
gHa/Per capita
Fossils
Biofuels
Solar electricity
Wind
6.5 (forest)
>6.7 (cropland)
0.21 (cropland)
0e2 (cropland)
8.2
>16.3
<0.49
0e5
1.47
>2.87
<0.09
0e0.9
Note: In the case of the fossil fuels, we use the necessary woodlands and “global” m2
(gm2) to absorb the emissions associated with its use in accordance with the
methodology of Wackernagel and Rees [57]. For the biofuels, the land needs are
calculated as the inverse of the power density calculated in this work (<0.15 W/m2).
For photovoltaic power, we use 4.8 We/m2 (see text); the table shows the worst of
the cases: installing parks in areas of cropland, and for wind power, we use 0.5 W/
m2 with calculations based on data from de Castro et al. [30] and Smil [60]. In the
case of wind power, the occupation of the necessary infrastructures is much smaller
if we consider that the wind farms allow a double use and that they are compatible,
for instance, with cropland and pastureland. In the case of photovoltaic power, if we
use the net power (2.8 W/m2) we will arrive to 0.156 gHa/per capita, but at present,
the associated fossil emissions required from the photovoltaic industry [42] could
reach the Ecological Footprint to roughly 1/3 of fossil fuels instead of roughly 1/10 in
croplands.
Ecological Footprint in a finite, non-renewable planet, in addition to
the use of fossil fuels (especially when the EROEI is low).
The use of biofuels on a global scale over the last decade makes
neither ecological nor sustainable sense. Among the arguments
that lead us to this conclusion are the following: i) their reduced net
power density; ii) the corresponding high Ecological Footprint; iii)
the percentage of GHG (Greenhouse gases) in comparison with
petroleum, which is possibly higher than the emissions of this fossil
fuel in terms of net energy (low EROEI) when the emissions of N2O
are also taken into account [4]; and iv) the negative balance of use
and land conversion in the real world [2,3,24,61].
In addition to these disadvantages which have been described in
the paper, it is important to mention other problems that have been
pointed out by other research works. These reasons are supported
by the data given in this work. However, in addition to these disadvantages, it may also be necessary to add other, equally important ones which arise with the large-scale proliferation of biofuels.
Among these are, for instance: i) the change in use of agricultural
land dedicated to feeding humans or animals and the directly [24]
or indirectly associated deforestation (the displaced crops are
transferred to forested areas [2]) in a world with severe problems of
hunger, deforestation and loss of biodiversity [22,28,29,62e64]; ii)
the high water use needed for its cultivation and industrial processing in a world with ever greater water shortages [25,65]; and
iii) the loss of organic material from the soils due to the erosive
processes associated with the energy crops in a context where
reversing the erosion process is urgent [22,43,46]. The above considerations, unfortunately, convert the generalisation of biofuels
into a worrying example of an unsustainable energy source with
negative socioeconomic implications [20].
When considering the question of energy it is, in general, better
to be prudent and not confuse the concrete technological potential
Table 4
Production of biofuels by region in 2008.
Region/fuel
Gross production
(litres)
EROEI
Net energy
(EJ)
US ethanol
Brazil and others ethanol
UE biodiesel
Others biodiesel
38$109
26.5$109
5.9$109
5.9$109
1.25
5
1.5
3
0.168
0.469
0.065
0.130
Note: For production see UNEP [56]. For the EROEI see Table 2. Net energy is
calculated as: net energy ¼ gross energy$(EROEI-1)/EROEI. The assigned energy is
22.1 MJ/L for ethanol and 33 MJ/L for biodiesel. The global EROEI gives 1.85.
Table 6
Brazil ethanol and World biofuels yield, EROEI, density power and EF.
Brazil ethanol
World biofuels
Real yield
(L/Ha)
EROEI
Gross density
(W/m2)
EF (gHa/MW)
4300e5100
<2120
<5
<1.85
0.3e0.36
<0.15
>680e815
>1630
Note: Real yield is the litres obtained per hectare really used to produce the biofuel.
The Ecological Footprint is given in global hectares needed to produce a continuous
power of one MW (compare this with the 820 gHa/MW calculated for fossil fuels).
C. de Castro et al. / Energy 64 (2014) 506e512
with its translation to reality on a global scale. Normally, it is not
enough simply to visualise our dilemma or trilemma [26], imagining that the technology could be viable if it were used in the best
cases and discarded in the worst. The calculations and estimations
carried out here are very far from the best cases because the real
world is much more complex and subject to internal (e.g. average
practices) and external (e.g. climate change) inefficiencies, as well
as to scientific, economic and political interests [10] which, unfortunately, are capable of promoting solutions which are far from
being good practices.
To summarise, the above pages have clearly shown the “technical” limitations of the potential production of biofuels. It is a
question of limitations that also affect other renewable energy
sources [30,33,32]. As biofuels on a large scale are highly dependent
on fossil fuels, the increase in their production would have an effect
(for example, in CO2 emissions) at most equivalent to a more efficient use of petroleum. In this latter case, it would mean a lower
impact on land use. The scenario considered in this article is BAU
(business as usual), and therefore, modifications concerning
changes in the social organisation of consumption and production
have not been incorporated. Nevertheless, although changes in
dietary patterns, for instance, have not been considered here, such
changes have been shown to have a very relevant influence on the
energy potentials of biomass [61], perhaps even more so than the
expansion of the yields due to technological improvements.
Although these questions also depend on the technical side, in the
light of the latter’s limitations, the effectiveness of the socioeconomic changes and the incentives for advancing towards a more
environmentally sustainable and socially just model should not be
ignored.
Given the current circumstances, and taking into account
criteria of social equality and ecological sustainability, it would be
convenient if, on a global scale, the technology associated with
biofuels were subject, at the very least, to a moratoria, putting a
stop to its expansion and even, perhaps, tending to a progressive
decrease in its use. It is to be hoped that this conclusion concerning
environmental policy, reached through the information and analysis given here and the arguments set out with their referenced
works, should not result in misinterpretations of the said
conclusion.
Author contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the
manuscript.
Acknowledgement
This work has been developed within the project CGL200914268 funded by Spanish Ministry of Science and Innovation
(MICINN). Óscar Carpintero would also like to thank financial
support from the Spanish Ministry of Science and Innovation
(Project ENE2010-19834, Project CSO2010-21979, and Project
HAR2010-18544).
Annex. Ethanol productivity calculations for Brazil
To calculate the performance of the ethanol, we carried out a
reversible process starting from the contents of the TRS of the
sugarcane. We did so in the following way:
1. Using the MAPA [38] data and the proportion of mean TRS
contained in the annual harvest of crushed sugarcane, and the
needs of the TRS per litre of ethanol (anhydrous and hydrated)
511
provided by MAPA (1 L of anhydrous ¼ 1.812 kg of TRS; 1 L of
hydrated ¼ 1.7412 kg of TRS), we have calculated the quantity of
TRS needed to obtain the ethanol produced in that year.
2. We then calculated the quantity of crushed sugarcane that
would have to be produced to achieve that volume of TRS.
3. Given the yield for crushed sugarcane each year, we have
obtained the area in hectares that would have to be dedicated
to growing sugarcane exclusively for the production of
ethanol.
4. Dividing the production of ethanol by the surface area in hectares calculated in the previous step, we obtain the yield per
hectare (litres/Ha) of ethanol.
References
[1] Farrell AE, Plevin RJ, Turner BT, Jones AD, O’Hare M, Kammen DM. Ethanol can
contribute to energy and environment goals. Science 2006;311:506e8.
[2] Fargione J, Hill J, Tilman D, Polasky S, Hawthorne P. Land clearing and the
biofuel carbon debt. Science 2008;319:1235e8.
[3] Searchinger T, Heimlich R, Houghton RA, Dong F, Elobeid A, Fabiosa J, et al. Use
of US croplands for biofuels increases greenhouse gases through emissions
from land use change. Science 2008;319:1238e40.
[4] Crutzen PJ, Mosier AR, Smith KA, Winiwarter W. N2O release from agrobiofuel production negates global warming reduction by replacing fossil
fuels. Atmos Chem Phys 2007;8:389e95.
[5] Haberl H, Sprinz D, Bonazountas M, Cocco P, Desaubies Y, Henze M, et al.
Correcting a fundamental error in greenhouse gas accounting related to bioenergy. Energy Policy 2012;45:18e23.
[6] Streets DG, Waldhoff ST. Greenhouse emissions from biofuel combustion in
Asia. Energy 1999;24:841e55.
[7] Ramírez-Triana CA. Powers of Brazilian ethanol: comparison between
assessment approaches. Energy Policy 2011;39:4605e13.
[8] Hammershlag R. Ethanol’s energy return on Investment: a survey of the
literature 1990-present. Environ Sci Technol 2006;40:1744e50.
[9] Science letters. Letters-energy returns on ethanol production. Science
2006;312:1746e8.
[10] Giampietro M, Mayumi K. The biofuel delusion: the fallacy of large-scale
agrobiofuel production. London: Earthscan; 2009.
[11] Berthiaume R, Bouchard C, Roden MA. Exergetic evaluation of the renewability of a biofuel. Exergy Int J 2001;1:256e68.
[12] Talens Peiró L, Villalba Méndez G, Sciubba E, Gabarrell y Durany X. Extended
exergy accounting applied to biodiesel production. Energy 2010;35:2861e9.
[13] Velásquez HI, De Oliveira S, Benjumea P, Pellegrini LF. Exergo-environmental
evaluation of liquid biofuel production processes. Energy 2013;54:97e103.
[14] Piekarczyk W, Czarnowska L, Ptasinski K, Stanek W. Thermodynamic
evaluation of biomass-to-biofuels production systems. Energy 2013;62:
95e104.
[15] Worldwach Institute. Biofuels for transport. London: Earthscan; 2007.
[16] Johnston M, Foley JA, Holloway, Kucharik C, Monfreda. Resetting global expectations from agricultural biofuels. Environ ResLett 2009;4:014004. http://
dx.doi.org/10.1088/1748-9326/4/1/014004. 9 pp.
[17] Howarth RW, Bringezu S. Biofuels: environment consequences and interactions with changing land use. In: Proceedings of the scientific Committee
on problems of the environment (SCOPE) International biofuels project Rapid
assessment. Ithaca NY, USA: Cornell University; 22e25 September 2008.
Gummersbach Germany, http://cip.cornell.edu/biofuels/.
[18] Balat M, Balat H. Recent trends in global production and utilization of bioethanol fuel. Appl Energy 2009;86:2273e82.
[19] Danielsen F, Beukema H, Burgess ND, Parish F, BrÜhl CA, Donald PF, et al.
Biofuel plantations on forested lands: double jeopardy for biodiversity and
climate. Conservation Biol 2008;23:348e58.
[20] Hecht L. Smell of gigantic hoax in government ethanol promotion. 21st century science and technology. Spring-summer; 2007. p. 62e7.
[21] Pimentel D, Patzek T, Cecil G. Ethanol production: energy, economic and
environmental losses. Rev Environ Contam Toxicol 2007;189:25e41.
[22] Pimentel D, Marklein A, Toth MA, Karpoff M, Paul GS, McCormack R, et al.
Biofuels impacts on world food supply: use of fossil fuel, land and water resources. Energies 2008;1:41e78.
[23] Hill J, Nelson E, Tilman D, Polasky S, Tijany D. Environmental, economic and
power costs and benefits of biodiesel and ethanol biofuels. Proc Nat Acad Sci
2006;103:11206e10.
[24] Koh LP, Miettinen J, Liew SC, Ghazoul J. Remotely sensed evidence of tropical
peatland conversion to oil palm. Proc Nat Acad Sci 2011;108:5127e32.
[25] Dominguez-Faus R, Powers SE, Burken JG, Alvarez PJ. The water footprint of
biofuels: a drink or drive issue? Environ Sci Technol 2009;43:3005e10.
[26] Tilman D, Socolow R, Foley JA, Hill J, Larson E, Lynd L, et al. Beneficial biofuels.
The food, energy and environment trilemma. Science 2009;325:270e1.
[27] Timilsina G, Shrestha A. How much hope should we have for biofuels? Energy
2011;36:2055e66.
512
C. de Castro et al. / Energy 64 (2014) 506e512
[28] Ajanovic A. Biofuels versus food production: does biofuels production increase
food prices? Energy 2011;36:2070e6.
[29] Nonhebel S. Global food supply and the impacts of increased use of biofuels.
Energy 2012;37:115e21.
[30] de Castro C, Mediavilla M, Miguel JL, Frechoso F. Global wind power potential:
physical and technological limits. Energy Policy 2011;39:6677e82.
[31] Miller LM, Gans F, Kleidon A. Estimating maximum global land surface wind power
extractabilityandassociatedclimaticconsequences.EarthSystDynam2011;2:1e12.
[32] De Castro C, Mediavilla M, Miguel LJ, Frechoso F. Global solar electric potential: a review of their technical and sustainable limits. Renew Sust Energy
Reviews 2013;28:824e35.
[33] Prieto PA, Hall CA. Spain’s photovoltaic revolution: the energy return on investment. Spring Briefs Energy/Energy Anal 2013.
[34] Connor DJ, Hernández CG. Crops for biofuel: current status and prospects for
the future. Chap. 4. In: Howarth RW, Bringezu S, editors. Biofuels: environment consequences and interactions with changing land use. Proceedings of
the scientific Committee on problems of the Environment (SCOPE) International biofuels project Rapid assessment. Ithaca NY, USA: Cornell University;
22-25. September 2008. p. 65e85. Gummersbach Germany.
[35] Pimentel D, Patzek TW. Ethanol production: energy and economic issues
related to US and Brazilian sugarcane. Biofuels, solar and wind as renewable
energy Systems. Nat Resour Res 2007;16:3235e42.
[36] Macedo IC, Seabra JE, Silva JE. Greenhouse gases emissions in the production
and use of ethanol from sugarcane in Brazil: the 2005/2006 averages and a
prediction for 2020. Biomass Bioenergy 2008;32:582e95.
[37] Giampietro M, Mayumi K. Complex systems thinking and renewable energy
systems. In: Pimentel D, editor. Biofuels, solar and wind as renewable energy
systems. Benefits and risk. Springer; 2008. p. 173e213.
[38] Ministry of Agriculture, Livestock and food Supply (various issues): AgriEnergy statistical Yearbook, [Brasilia, DF].
[39] Rudorff BF, de Aguilar DA, da Silva WF, Sugawara LM, Adami M, Moreira MA.
Studies on the rapid expansion of sugarcane for ethanol production in São
Paulo state (Brazil) Using Landsat Data. Remote Sensing 2010;2:1057e76.
[40] UNICA 2012. www.unica.com.br.
[41] de Oliveira ME, Vaughan BW, Rykiel Jr EJ. Ethanol as fuel: energy, carbon
dioxide balances and ecological footprint. Bioscience 2005;55(7):593e602.
[42] Boddey RM, Soares LB, Alves BJ, Urquiaga S. Bio-ethanol production in Brazil.
In: Pimentel D, editor. Biofuels, solar and wind as renewable energy systems.
Netherlands: Springer; 2008.
[43] Pereira CLF, Ortega E. Sustainability assessment of ethanol production from
sugarcane. 1st international workshop advances in cleaner production; 2010.
[44] Pimentel D, Marklein A, Toth MA, Karpoff MN, Paul GS, McCormack R, et al. Food
versus biofuels: environmental and economic costs. Hum Ecol 2009;37:1e12.
[45] Wang M. Updated energy and greenhouse gas emission results of fuel ethanol.
The 15th International symposium on alcohol fuels. Sepp. 26e28. San Diego,
CA, USA; 2005.
[46] Dong X, Ulgiati S, Yan M, Zhang X, Gao W. Energy and eMergy evaluation of
bioethanol production from wheat in Henan Province, China. Energy Policy
2008;36:3882e92.
[47] Pleanjai S, Gheewala SH. Full chain energy analysis of biodiesel production
from palm oil in Thailand. Appl Energy 2009;86:S209e14.
[48] Yee KF, Tan KT, Abdullah AZ, Lee KT. Life cycle assessment of palm biodiesel:
revealing facts and benefits for sustainability. Appl Energy 2009;86:S189e96.
[49] Giampietro M, Ulgiati S, Pimentel D. Feasibility of large-scale biofuel production. Bioscience 1997;47(9):587e600.
[50] Macedo IC. (Organizer) Sugar Cane’s Energy e Twelve studies on Brazilian
sugar cane agribusiness and its sustainability. UNICA -Uniao da Agroindústria
Canavieira do Estado de Sao Paulo; 2005.
[51] Bianchi SR, Miyazawa M, de Oliveira EL, Pavan MA. Relationship between the
mass of organic matter and the carbon in soil. Braz Arch Biol Technol
2008;51(2):263e9.
[52] Ministeriode Minas e Energia. Brazilianenergy balance, 2010. RiodeJaneiro;2011.
[53] Beeharry RP. Extended sugarcane utilisation for exportable electricity production in Mauritius. Biomass Bioenergy 1996;11(6):441e9.
[54] FAO. FAO statistical database. http://www.fao.org/corp/statistics/en/; 2011.
[55] Campbell JE, Lobell DB, Genova RC, Field CB. The global potential of bioenergy
on abandoned agriculture lands. Environ Sci Technol 2008;42(15):5791e4.
[56] UNEP. Assessing biofuels. United Nations Environment Programme; 2009.
www.unep.org.
[57] Wackernagel M, Rees W. Our ecological footprint: reducing human impact on
the Earth. Gabriola Island, BC: New Society Publishers; 1996.
[58] UNEP. Search “ecological footprint” in www.unep.org.
[59] WWF. Living planet report. http://wwf.panda.org/about_our_earth/all_
publications/living_planet_report/; 2012.
[60] Smil V. http://www.vaclavsmil.com/wp-content/uploads/docs/smil-articlepower-density-primer.pdf (retrieved: October 5, 2010).
[61] Haberl H, Erb KH, Krausmann F, Bondeau A, Lauk C, Müller C, et al. Global
bioenergy potentials from agricultural land in 2050: sensitivity to climate
change, diets and yields. Biomass Bioenergy 2011;35:4753e69.
[62] Koh LP, Ghazoul J. Biofuels, biodiversity and people: understanding the conflicts and finding opportunities. Biol Conser 2008;141(10):2450e60.
[63] Scharlemann JP, Laurance WF. How green are biofuels? Science 2008;319:43e4.
[64] Pimentel D. Ethanol fuels: energy balance, economics and environmental
impacts are negative. Nat Resour Res 2003;12(2):127e34.
[65] UNDP. Human development report 2006: Beyond Scarcity: power, poverty
and the global water crisis. www.undp.org.