Applied Energy 77 (2004) 119–130
www.elsevier.com/locate/apenergy
Wind turbines in Brazil and Germany:
an example of geographical variability in
life-cycle assessment
Manfred Lenzena,*, Ulrike Wachsmannb
aSchool of Physics, A28, The University of Sydney NSW 2006, Australia
bCENERGIA, Programa de Planejamento Energe´tico, COPPE-CT, Universidade Federal do Rio de Janeiro,
C-211, 21945-970 Rio de Janeiro, RJ, Brazil
Received 4 March 2003; received in revised form 25 March 2003; accepted 29 March 2003
Abstract
A life-cycle assessment of a single product can produce substantially varying results that
depend on the location of production. It is the aim of this study to provide an example of this
geographical variability by examining the energy and CO2 embodied in a particular wind-
turbine manufactured in Brazil and in Germany. Our results demonstrate the importance of
adequately considering the background system of the local economy.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: Life-cycle assessment; Wind turbines; Brazil; Germany; Energy; CO2
1. Introduction
A life-cycle assessment (LCA) of a product or process aims at capturing a range of
environmental liabilities or impacts that accumulate over the entire cradle-to-grave
period. Generally speaking, these impacts occur either directly during the manu-
facture of the product, or during the process (on site), or are caused indirectly dur-
ing the provision of inputs into the manufacture, or process (off site). While direct
impacts are a unique characteristic of the very product or process, indirect impacts
can be expected to vary with the structure and performance of the supplying back-
ground system, that is the economy of the location of production (see Ref. [1] for a
* Corresponding author. Tel.: +61-2-9351-5985; fax: +61-2-9351-7725.
E-mail addresses: m.lenzen@physics.usyd.edu.au (M. Lenzen), ulrike@ppe.ufrj.br (U. Wachsmann).
0306-2619/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0306-2619(03)00105-3

---

120
M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
demonstration of the variability of energy inputs and CO2 emissions in the electricity
sectors across a wide range of countries).
It is the aim of this study to obtain an idea about the variability of the indirect
contribution to a life-cycle inventory (LCI) for nominally-identical products, caused
by geographical variations of the background supply system. We examine the life
cycle of a specific wind-turbine model produced in Brazil and Germany. These
countries were chosen partly because of data availability, and partly because of
striking differences in economic structure and energy-supply characteristics. The
analysis is carried out predominantly in energy and CO2 terms, because energy is a
proxy quantity for a wide range of other impacts (for example SO2, NOx), and in
order to enable comparisons with previous studies to be achieved.
Cumulative energy requirements for wind turbines in different countries have been
calculated before for Germany and India by Gu¨rzenich et al. [2], including the transport
between the two countries. However, these authors use non-specific process-analysis-
based ‘‘energy factors’’, that are ‘‘suitably modified for Indian conditions, wherever
found necessary’’. In contrast to their approach, we use generic data sets for both
countries, reflecting both the domestic economic and energy-use structures.
This paper is structured as follows. The following section introduces the E-40
wind turbine and describes the structure of the Brazilian and German economies
with a focus on the energy-supply systems. Following, we explain the hybrid
(process/input–output) LCI approach taken in this work, including estimates of
uncertainty and data sources. Results of all calculations are then presented in a
comparative manner, discussed, and the paper concluded.
2. The E-40 in Brazil and Germany
The wind turbine model E-40 manufactured by the German company Enercon
features a three-blade, pitch-controlled rotor with a nominal power of 500 or
600 kW (depending on the wind class). The rotor diameter and the height of the
hub are variable so that it can be efficiently adjusted to the prevailing wind
conditions of any location. The turbine does not have a gearbox and the rotor
is directly connected to the generator. The rotor blades (diameter 40 or 44 m) are
made of fibreglass-reinforced epoxy. The tower can be either of tubular steel or
steel–concrete.
In 1996, Enercon founded a subsidiary in Brazil. The E-40 model has since been
assembled, initially from locally produced blades, foundations and towers, and
imported nacelles and generators. The manufacture of completely Brazilian-made
turbines commenced in 2000. At the beginning of 2001, the subsidiary had installed
35 wind turbines in the states of Ceara´ and Parana´ with a total nominal power of
17.5 MW (81% of Brazil’s total wind-power capacity) [3].
Since the emphasis of this work is on the background supply systems, we provide
the reader with an overview of the economic and energy structure of Brazil and
Germany (Figs. 1 and 2). All sectors, except for services, are more important for
generating Gross Domestic Product in Brazil. Germany, as an industrialised

---

M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
121
Fig. 1. Breakdown of Gross Domestic Product of Brazil (left pie chart) and Germany (right pie chart) by
broad industry class [26,28].
Fig. 2. Breakdown of primary-energy use in Brazil (left pie chart) and Germany (right pie chart) by broad
industry class [25,29]. Note: Secondary transport fuels are represented by their primary equivalent (crude
oil or sugar-cane alcohol) within manufacturing (refining).
country, naturally relies to a much larger extent on the tertiary sector. The per-
capita Gross Domestic Product in 1995 was equivalent to 5928 U$PPP1 in Brazil
and 20,370 U$PPP in Germany.
Both countries feature a large consumption of oil-based liquid fuels. While natural
gas and nuclear energy are only important for German industries, hydraulic energy,
bagasse and firewood, and sugar-cane-based alcohol are unique to Brazil. Since the
latter are all renewable energy sources, the CO2 balance for Brazilian products can
1 The World Bank (http://www.worldbank.org/depweb/english/modules/glossary.htm#ppp) defines
Puchasingh Power Parties (PPP) as ‘‘a method of measuring the relative purchasing power of different
countries’ currencies over the same types of goods and services. Because goods and services may cost more
in one country than in another, PPP allows one to make more accurate comparisons of standards of living
across countries.’’

---

122
M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
be expected to be considerably lower. Note that the utilities (mainly electricity
generation) sector consumes considerably less energy in Brazil, which is due to the
conversion efficiency of hydraulic energy being much higher than that of coal.
3. Methodology: hybrid IO-LCA
Triggered by the oil crises of the 1970s, energy analysis emerged as a discipline to
investigate the total energy required to perform a given task. Initially, this discipline
employed process analysis, where the energy requirements of the main production
processes and some important contributions from suppliers of inputs into the main
processes are assessed in detail (e.g. by auditing or using disparate data sources), and
the system boundary is usually chosen with the understanding that the addition of suc-
cessive upstream production stages has a small effect on the total energy embodiment [4].
Process-based energy analyses of wind-energy converters have been carried out by
various authors (see for example Refs. [5–9], and [10] for a review).
A drawback of the setting of system boundaries and, as a consequence, the omission
of processes outside these boundaries, is the introduction of a systematic truncation
error. The magnitude of this error varies with the type of product or process considered,
but can be of the order of 50% [11]. More importantly, it is not significantly reducible
by extending the system boundary [12, 21]. One way to avoid such significant errors is
to complement a conventional process analysis with an input–output analysis, resulting
in a hybrid life-cycle assessment method. In this work we employ a tiered hybrid energy
analysis, where the direct and downstream energy requirements (for construction,
use, and end-of-life), and some important lower order upstream requirements of the
functional unit are assessed in a detailed process analysis. System completeness is
achieved by covering the remaining higher-order requirements (for materials
extraction and manufacturing) using input–output analysis. Input–output-based
energy analyses of wind-energy converters can be found in Refs. [13,14].
Input–output analysis is a top-down economic technique, which uses sectoral
monetary transactions data to account for the complex interdependencies of indus-
tries in modern economies. The result of generalised input–output analyses is a 1Ân
factor multiplier, that is embodiments of production factors (such as water, labour,
energy, resources and pollutants) per unit of final consumption of commodities
produced by n industry sectors. A multiplier m can be calculated from a 1Ân vector f
containing sectoral production factor usage, and from a nÂn direct requirements
matrix A according to
m ¼ fðI À AÞÀ1;
ð1Þ
where I is the nÂn unity matrix. In this work, m is expressed in producers’ prices,
containing margins, but no sales tax (see Ref. [15]). The factor inventory F (scalar) of
a given product or process represented by a nÂ1 commodity inputs vector y and a
scalar Fd of direct factor usages is then simply
F ¼ my þ Fd:
ð2Þ

---

M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
123
my represents the indirect usage of factors embodied in all inputs into the product
and process.
An introduction into the input–output method and its application to environ-
mental problems can be found in Refs. [16–18]. A more elaborate treatment of the
theory underlying hybrid LCA is provided by Suh and Huppes [19]. For further
details on input–output-based LCA techniques, see Refs. [20–22].
4. Data sources
This work is based on a life-cycle study of the E-40 wind turbine by Pick and
Wagner [23,24]. We adopt all technical details, site conditions and cost from the
German example, and translate these into the Brazilian context. This means that we
examine a wind turbine identical to the one operating in Germany, but (partly or
fully) produced and operating in the Brazilian economy.
Considering Eqs. (1) and (2), this study requires three types of data: (1) factor use
statistics f on energy and CO2, (2) direct requirements matrices A, and (3) a common
commodity input vector y, preferably expressed in a common currency (here: US$).
For Brazil, data on industrial and residential energy usage for 24 primary and sec-
ondary fuels are regularly published by the National Department for Energy Develop-
ment [25]. Hydroelectricity was valued as hydro-potential and not converted into a
thermal equivalent. The direct requirements matrix is part of the input–output tables
published by the Instituto Brasileiro de Geografia e Estatı´stica [26], detailing 80 com-
modities and 43 industries. Finally, CO2 contents of fuels were taken from Ref. [27].
For Germany, data on energy and CO2 are contained in the input–output tables
published by the Statistisches Bundesamt [28,29]. The latter distinguishes 59 com-
modities and industries.
All freight transport was examined on the basis of distances between and within
the countries, using energy and greenhouse-gas intensities calculated for Australia
[30]. The impact of the disposal of the turbine was assumed to be negligible (cf. [31]
and [32]). Finally, input cost y for the E-40 turbine in 1995 DM could be derived
from data documented by Pick and Wagner [24]. These costs were converted into
1995 US$ using an exchange rate of 1.5 1995DM/1995US$ [33]. Costs for Brazilian
production were obtained by adjusting the German cost data with the lower per-
centage of primary inputs (such as wages and capital) in Brazil.2 In 1995, the
Brazilian Real (R$) was pegged to the US$.
5. Uncertainties
Input–output analysis suffers from uncertainties arising from various sources, the
most important of which are source data errors due to unreliable sampling, reporting
and imputation, the aggregation of input–output data over different producers, and
2 Primary inputs in Brazil and Germany represent 40 and 60% of total input, respectively [26,28].

---

124
M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
the aggregation of input–output data over different products supplied by one
industry sector (allocation error) [12].
The Brazilian statistical bureau (IBGE) does not keep comprehensive standard error
estimates on the direct requirement coefficients. However, since these coefficients arise
from industry surveys that are similar to those conducted in Australia, we apply
Australian error terms for source data and aggregation errors. Allocation errors for
the Brazilian input–output model were estimated by comparing the energy multipliers
of similar commodity groups (construction materials and metal products; chemicals
and plastic products; food products; services) with their aggregate value. The estima-
tion of standard errors for multipliers as in Eq. (1) is not so straightforward, because
standard errors of the Leontief inverse L=(IÀA)À1 cannot be calculated analytically,
but by using a Monte-Carlo technique to simulate the propagation of uncertainties
[34]. In this work, errors for Brazilian energy and CO2 multipliers were calculated
from 24,000 energy multipliers generated from 3000 simulation runs.
Given the resulting standard errors Ámj, and the standard errors Áyi and ÁÈd of
y and Èd, the total standard error ÁÈ of the factor inventory È is then
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
X
n
X
n
DF ¼
ðykDmkÞ2þ
ðmkDykÞ2þDF2:
ð3Þ
d
k¼1
k¼1
If the standard errors Ámk, Áyk, and ÁFd are assumed to be stochastic, the total
standard error ÁÈ decreases with increasing number of non-zero entries in y, that is,
with increasing detail of the breakdown of the inputs for the product or process. In
order to minimise the relative standard error of the factor inventory, it is therefore
important to (1) obtain a breakdown of the inputs that is as detailed as possible, and
(2) obtain important direct factor inputs with low relative standard errors ÁFd/Fd.
In a process-type LCA, strategy (1) is not applicable, because of systematic errors
due to the truncation of the system. For these non-stochastic errors, a decrease of
the overall error with increasing detail does not occur. For further details on the
uncertainty calculus, see Ref. [12].
6. Results
We examine five scenarios:
1. production and operation in Germany;
2. production (except foundation) in Germany, operation in Brazil;3
3. 1999: production of generator and nacelle in Germany, remaining parts and
operation in Brazil;3,4
3 Operation in Brazil occurs either in Parana´—a southern interior state of Brazil, or Ceara´—a northern
coastal state of Brazil.
4 Production in Brazil occurs in Sorocaba in the state of Sa˜o Paulo, about 500 km from Parana´ and
about 3000 km from Ceara´.

---

M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
125
4. production and operation in Brazil;3,4
5. production and operation in Brazil,3,4 but assuming a high proportion of
recycled steel.
For each scenario, we consider five installation options featuring different loca-
tions (with different tower heights and foundation masses, see Table 1). The annual
output in Brazil was calculated based on German data in Pick and Wagner’s report,
but using an average wind speed of 7 m/s (instead of 5 m/s for German sites [35]).
Further technical details can be found in Pick and Wagner’s report [24].
For the hypothetical scenario 5, we have assumed that 75% of Brazil’s steel produc-
tion is from scrap steel via the Electric Arc Furnace (EAF) route, and 25% from pri-
mary ore via the Basic Oxygen Furnace (BOF) route (as opposed to 24%/74% in 1995
[36]). The EAF route is more energy efficient than the BOF route, and in the case of
Brazil’s hydroelectric capacity cause substantially less CO2 emissions. This scenario was
calculated by adjusting the Brazilian energy matrix using input data provided by Wor-
rell et al. [36], and subsequently calculating a revised Leontief inverse.
The differences in the primary-energy embodiment of wind turbines produced in
Germany and Brazil are considerable (Table 2).5 The main reason for these dif-
ferences is the higher conversion efficiency of the Brazilian electricity generation
system (above 90%). An evaluation of all scenarios using a fossil-fuel equivalent of
Brazil’s hydraulic energy yielded energy embodiments that were similar to those of
German-produced turbines (around 12,000 GJ). It is remarkable that the energy
Table 1
Technical characteristics of the examined wind turbine and site options (after Ref. [24]).
Coastal 44
Coastal 55
Near-coastal Inland 55
Inland 65
Tower height (m)
44
55
55
55
65
Foundation mass (t)
132.7
163.8
163.8
150.2
185.8
Transport distance, landa (km)
3000/100
3000/100
3000/200
500/800
500/800
Transport distance, seab (km)
8000/10130
8000/10130
8000/10130
8000/10130
8000/10130
Annual output (kWh) Brazil
3,558,926
3,748,666
2,910,409
2,196,404
2,420,131
Annual output (kWh) Germany
1,296,985
1,366,132
1,060,645
800,439
881,972
Coastal=state of Ceara´, inland=state of Parana´
a The two numbers represent the distances from the place of production/delivery to the place of
installation (first number—land transport within Brazil from the place of the production in Sorocaba
either to Parana´ (500 km; I-55, I-65) or to Ceara´ (3000 km; C-44, C-55, NC-55); second number—land
transport from the nearest seaport to Parana´ (800 km; I-55, I-65) or Ceara´ (100 km, C-44, C-55; 200 km,
NC-55).
b These two numbers represent the distances between Germany and the seaport in Ceara´ (8000 km;
C-44, C-55, NC-55) or the seaport Santos (10,130 km; I-55, I-65).
5 Note that the cumulative energy requirement for German production obtained in this work is almost
twice as high as Pick and Wagner’s estimate (6000–7000 GJ). This discrepancy is due to the remarkable
low energy intensities used in the latter study (around 5 MJ/DM). A back-of-the-envelope calculation of
an average German energy intensity by dividing total primary energy consumption by GDP yields about
8.9 MJ/DM. This shows that Pick and Wagner’s values are probably too low.

---

126
M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
Table 2
Total and specific energy requirements (GJ and MJ/kWhel) for the production and operation of the E-40
under different scenarios in Brazil and Germany
Scenario
C-44
C-55
NC-55 I-55
I-65
C-44 C-55 NC-55 I-55 I-65
(GJ)
(MJ/kWhel)
P&O in Germany
11,263 12,568 12,326 12,330 12,938 0.43
0.46
0.58
0.77 0.73
P Germany O Brazil
11,627 13,029 12,835 13,055 13,797 0.16
0.17
0.22
0.30 0.29
P Germany and Brazil, O Brazil
9525 10,607 10,326 10,147 10,733 0.13
0.14
0.18
0.23 0.22
P&O in Brazil
8094
8827
8547
8486
9214 0.11
0.12
0.15
0.19 0.19
P&O in Brazil, recycled steel
6289
6667
6368
6322
6834 0.09
0.09
0.11
0.14 0.14
C=coastal (Ceara´); I=inland (Parana´); NC=near coastal; O=operation; P=production.
embodiments vary stronger with the production scenario than with site conditions
(tower high, foundation, mass) and transport distances. Looking at the specific energy
requirements (see the right-hand block of Table 2), the differences become even
more pronounced: the best and the worst option differ by a factor of more than 8.
This is because the total output is higher on average for Brazilian sites (see Table 1).
In general, the largest amount of energy is consumed for the tower (approximately
30–40% of the total), followed by the generator (25–30%) and the nacelle (10–15%).
Transport energy is consistently below 5% of the total energy requirement (Fig. 3, left
diagram). The shares of the components vary only slightly with the installation option
and country of production, except for scenario 5, where steel components (tower and
nacelle) consume relatively small shares due to the assumed efficient steel-industry.
Fig. 3. Shares of components in the specific energy and CO2 requirements; installation option C-44 (1—P
and O in Germany; 2—P Germany, O Brazil; 3—Real case; 4—P and O in Brazil; 5—P and O in Brazil,
recycled steel).

---

M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
127
While the transport energy requirement seems surprisingly low, our finding is
supported by Wenzel [37] and Gu¨rzenich et al. [2]. The latter authors arrive at
transportation between Germany and India representing between 4 and 5.4% of the
cumulative energy requirement of wind turbines. Examining the effect of explicitly
considering imports and foreign emissions on the life-cycle CO2 emissions on the
German production of passenger cars, computers and food items, Wenzel [37] finds
that in spite of long distances, CO2 emissions from transport form a relatively minor
part of total emissions (+1!2% for cars and computers, and around +6% for
food items). If, however, foreign energy production (especially electricity) was
explicitly taken into account, CO2 requirements changed significantly (À9% for cars
and computers; food was not examined). Wenzel concludes that within effects of
trade on CO2 emissions, and within reduction potentials, differences in production
structure are more important than increased transport requirements.
CO2 embodiments vary substantially with the location of production (Table 3):
the more components are produced in Brazil, the more favourable the score on
emissions. CO2 emissions for production and operation of wind turbines in
Brazil under present conditions are a factor of 5 lower than those for Germany.
This remarkable difference is entirely due to the differences between the energy-
supply systems of the respective economies, as depicted in Fig. 2. Once again,
these differences become even more pronounced in terms of the specific CO2
requirement. Even if the wind speed conditions were reversed, the better sites in
Germany would not be able to compensate for the advantages of the Brazilian
production-system.
The shares of components in the specific CO2 requirements vary significantly with
the production scenario. Shifting production of components from Germany (1, 2) to
Brazil (3) reduces every share except that of the nacelle and the generator, which in
1999 were still produced in Germany. Once production is fully transferred to Brazil
(4), transport, installation and operation assume a larger share, because these tasks
still rely on fossil fuels only. Their share even increases if the steel sector becomes
more efficient (5).
Table 3
Total and specific CO2 requirements (t and kg/kWhel) for the productions and operation of the E-40
under different scenarios in Brazil and Germany
Scenario
C-44 C-55 NC-55 I-55
I-65
C-44
C-55
NC-55 I-55
I-65
(tons of CO2)
(kg of CO2/kWhel)
P and O in Germany
1176 1315 1290
1291 1358 0.045 0.048 0.061
0.081 0.077
P Germany O Brazil
1053 1183 1178
1186 1244 0.015 0.016 0.020
0.027 0.026
P Germany and Brazil, O Brazil
599
628
588
580
564 0.008 0.008 0.010
0.013 0.012
P and O in Brazil
204
202
196
195
212 0.003 0.003 0.003
0.004 0.004
P and O in Brazil, recycled steel
134
119
114
113
123 0.002 0.002 0.002
0.003 0.003
C=coastal (Ceara´); I=inland (Parana´); NC=near-coastal; O=operation; P=production.

---

128
M. Lenzen, U. Wachsmann / Applied Energy 77 (2004) 119–130
Table 4
Standard errors of energy and CO2 embodiments in Tables 2 and 3 [after Eq. (3), in %]
Scenario
C-44
C-55
NC-55
I-55
I-65
P and O in Brazil
26.1
28.1
29.1
29.1
30.1
P and O in Brazil, recycled steel
23.9
25.4
25.8
26.9
28.8
C=coastal (Ceara´); I=inland (Parana´); NC=near-coastal; O=operation; P=production.
Given standard errors for multipliers mk and input costs yk
Dm
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
k
Dyk
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
% 20%;
%
Var C
ð Þ þ Var L
ð Þ ¼
5
ð %Þ2þð16%Þ2 % 17% ;
ð4Þ
mk
yk
with C and L representing component and labour costs, respectively, and further
assuming an additional standard error of 50% for mis-allocation of wind-turbine
components to input–output categories, and considering n%3 significant compo-
nents (tower, generator, nacelle) standard errors of total energy and CO2 require-
ments can be estimated to be about
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
!2
!2
DF
u Dmk
Dyk
p À1
% t
ffiffiffi
þ
þ 50%2
n
% 28% :
ð5Þ
F
mk
yk
Table 4 contains results from a more detailed calculation of standard errors for all
installation options in Brazil.
7. Conclusions
Over the past 20 years, electricity demand in Brazil grew much faster than GDP
and overall energy demand. Especially the residential sector features a substantial
potential for further growth, due to unsatisfied demand. At the same time, hydro-
power plants—currently supplying more than 90% of Brazil’s electricity—are
increasingly perceived by investors as expensive, controversial and risky [38].
With this in mind, Schaeffer and Szklo [38] conclude that a future electricity mix
that meets both least-cost and environmental protection criteria will feature a sig-
nificant proportion of wind turbines. The indirect energy requirements for a poten-
tial transition to this mix can potentially form a substantial part of the national
energy consumption [10]. The results of this work can be used to determine to what
extent projected levels of available power during such a transition are being over-
estimated if energy embodied in power plants is not taken into account.
On a more hypothetical note, differences in specific CO2 embodiments may cause
the production of wind turbines (or other commodities) to move to CO2-efficient
economies such as Brazil, if CO2 emissions were to be penalised sufficiently, and
accounted for in a life-cycle context (‘‘positive leakage’’). Geographical variability of
the embodied CO2 emissions is at least comparable to the variability across wind

---

Document Outline

• Wind turbines in Brazil and Germany: an example of geographical variability in life-cycle assessment
  • Introduction
  • The E-40 in Brazil and Germany
  • Methodology: hybrid IO-LCA
  • Data sources
  • Uncertainties
  • Results
  • Conclusions
  • References

---