A model for waste-specific

life cycle inventories of

open burning of waste

Technical report
Zurich, September 2017

Gabor Doka
Doka Life Cycle Assessments, Zurich
LCI model of open burning 2

Author Gabor Doka, Doka LCA, Zurich, do@doka.ch

Review Ivan Muñoz of 2.-0 LCA consultants

Title picture Burning waste on a landfill in Peru in March 2012, by Alex Proimos. Creative Commons under
https://commons.wikimedia.org/wiki/File:Environmental_Disaster_(6983486731).jpg (9 Jul 2017)

Disclaimer The information contained in this report were carefully elaborated and compiled by the author according to best
knowledge and available data. Due to the broad scope of the task and the inherently variable nature of the
subject, errors or deviations cannot be excluded. For this reason the information contained in this report is
provided without warranty of any kind, express or implied, including but not limited to the fitness for a particular
purpose. In no event shall the author, Doka Life Cycle Assessments or the commissioner be liable for any claim,
damages or other liability, arising from the use, application or dissemination of this information.

Significant digits Figures in this report often feature several digits. This is not to imply that all the shown digits are really
significant or that the data displayed is very precise. Showing several digits helps to minimise the avoidable
accumulation of rounding mistakes along the chain of calculations performed here, and in possible future
studies referring to this data.

Copyright Copyleft © 2002-2017 Gabor Doka, do@doka.ch. Permission is granted to copy, distribute and/or modify this
document under the terms of the GNU Free Documentation License(http://www.gnu.org/copyleft/fdl.html),
Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, with no
Front- Cover Texts, and with no Back-Cover Texts. The reports are provided free of charge. They may be
distributed freely provided that they always distributed together, in whole and unchanged. The reports may not
be sold or offered for sale, or included with software products offered for sale. Companies that distribute
documents for profit are expressly prohibited from distributing the reports. The reports come with absolutely no
warranty. These are free documentations, and you are welcome to redistribute them under certain conditions.
See GNU General Public License (http://www.gnu.org/copyleft/fdl.html) for details.

Suggested citation: Doka G. (2017) A model for waste-specific life cycle inventories of open burning of waste. Doka Life Cycle
Assessments, Zurich, Switzerland. Available at http://www.doka.ch/publications.htm
LCI model of open burning 3

Contents
1 Introduction..........................................................................................................4

2 Basic concept ......................................................................................................4

3 Waste composition..............................................................................................5

3.1 Waste-specifity and burnability ............................................................................................................ 5

3.2 Assistant fuels....................................................................................................................................... 6

3.3 Recycled materials ............................................................................................................................... 6

4 Particulate emissions..........................................................................................6

5 Incomplete combustion products ......................................................................7

5.1 Nitrogen emissions ............................................................................................................................... 8

5.2 Carbon monoxide and methane .......................................................................................................... 8

5.3 Organic carbon compounds................................................................................................................. 9

5.4 Properties of exchanges .................................................................................................................... 10

6 Modelling parameters .......................................................................................12

6.1 Geography code ................................................................................................................................. 12

6.2 Subcategory of air emissions............................................................................................................. 12

6.3 Subcategory of soil emissions ........................................................................................................... 12

6.4 Share of thermal NOx......................................................................................................................... 12

6.5 Recycling rates ................................................................................................................................... 12

7 Annex A: Transfer Coefficients ........................................................................13

8 References .........................................................................................................15
LCI model of open burning 1. Introduction 4

1 Introduction
The LCA waste incineration model in Doka (2003-II, 2013) is based on modern waste incinerator
plants in Switzerland and Germany, with extensive flue gas purification. Here a process inventory
model for open burning of waste is presented. As with the previous instalments, the model focuses on
the emission of pollutants from specific waste materials. This model can be used to obtain emission
figures for specific waste materials when they are burned in the open without any kind of flue gas
cleaning. A first model of open burning was already developed in (Doka 2000).
The present tool was developed 2016–2017 by Gabor Doka in a project for Sustainable Recycling
Industries SRI, financed by the Swiss State Secretariat for Economic Affairs SECO, commissioned by
the Swiss Federal Laboratories for Materials Science and Technology EMPA, mandated by the
ecoinvent Association. The reports were reviewed by Ivan Muñoz of 2.-0 LCA consultants.

2 Basic concept
The present model for open, uncontrolled burning is conceptually based on a previous model
elaborated for controlled municipal incineration. The municipal waste incineration model in Doka
(2003-II, 2013) basically calculates waste-specific emissions by applying transfer coefficients for
chemical elements to the content of those elements in a specific waste material. The elemental waste
composition describes the input to the plant and the transfer coefficients describe the behaviour of
elements in the plant. In very basic terms this can be formulated in the following way:
Waste input Element × Transfer Coefficient Element,Output = Emission Element,Output
A distinction is made in the incineration model between burnable and unburnable waste. This concept
is also applied for the open burning model. Unburnable materials are assumed to go to the solid
residue, while burnable materials can also lead to air emissions. For the user of the model it is
therefore important to keep unburnable and burnable materials separate. The model allows for
inventories of mixtures of unburnable and burnable materials. The functional unit is always 1 kg waste
input.
A modern waste incinerator has several output streams, each of which are described with a set of
transfer coefficients. The outputs in the municipal incinerator model are bottom ash, boiler ash, ESP
ash, scrubber sludge, water emissions, and air emissions. Scrubber sludge and water emissions
contains the pollutants removed from flue gas by scrubber columns. For the open burning model the
transfer coefficients are reduced to two outputs: emissions to air and emissions to solid residues (Fig.
2.1). The transfer coefficients to air are obtained by summing up the transfer coefficients for scrubber
sludge, water emissions, and air emissions. These are the streams that would remain airborne without
scrubbers. The transfer coefficients to residues are the remaining transfer coefficients for bottom ash,
boiler ash, and ESP ash. These are the streams that will sink to the ground after incineration. Transfer
coefficients for burnable materials in open burning are shown in Tab. 7.1 on page 14.
Some recyclable metals might be separated from the residues. No further treatment is assumed after
recycling: the remaining residues stay on site and are inventoried as emissions to soil. The user can
choose if this shall be industrial soil (default) or agricultural soil (for agricultural biomass burning).
The air emissions are inventoried unfiltered and unchanged. From the non-gaseous compounds in the
flue gas a figure for total particulate mass is calculated. From this, waste-specific emissions are
derived for three particulate matter classes (PM>10; PM2.5-10; PM<2.5). The heavy metals and other
LCI model of open burning 3. Waste composition 5

elements are additionally inventoried as emissions in accordance to methodological choices in the

ecoinvent database.

Fig. 2.1 Flows and boundaries for the open burning process

3 Waste composition
3.1 Waste-specifity and burnability
The open burning model is waste-specific, which means that the elemental waste composition to a
large degree influences the inventoried emissions. Waste compositions should be divided into
burnable and unburnable parts by the user, since they have distinctly different behaviour. For instance
electronic waste should be divided into burnable parts (mostly plastics) and unburnable parts (bulk
metals, bulk minerals, ceramics, glasses etc.). Chemical elements contained in a burnable matrix, for
instance copper traces in plastics, belong to the burnable material and shall not be classified by the
user as being unburnable. A borderline case may be thin layers of unburnable material on burnable
materials. Here the user has to decide which fate seems most likely, or devise a burnable/unburnable
mixture.
LCI model of open burning 4. Particulate emissions 6

3.2 Assistant fuels

Fig. 3.1 Roasting of printed wiring boards over an open fire in Guiyu, China. Credit: Step Initiative /
EMPA Source: http://ewasteguide.info/images/4333

Sometimes open burning employs some assistant fuels, e.g. roasting of printed wiring boards on open
coal stoves. If the emissions of the fuel are to be included in the inventory, two methods can be
employed:
1) Two separate inventories are created for the waste and the assistant fuel.
2) Waste and fuel are defined as a mixture of materials and inventoried in one inventory.
In either case the elemental composition of the fuel must be specified. Also if the fuel is commercial
(e.g. coal, fuel wood) and not a waste itself, its upstream production chain should be included by
adding an appropriate fuel supply input in the inventory.

3.3 Recycled materials

Sometimes open burning is performed to isolate recyclables from a composite material, e.g. burning
off of cable insulation from copper cables (cf. title page). In such cases some target recyclables are
removed from the burn residues.
For iron, aluminium and copper the user can specify the recycling rate of these metals from waste
residues (cf.chapter 6.5 'Recycling rates' on page 12). The actually recycled fractions are then
subtracted before calculating the emissions to soil. The rates refer only to bulk metallic parts of these
elements, not to their oxides etc. nor the complete elemental content.
In case any other materials than iron, aluminium or copper are removed, e.g. metallic lead, it is
advised that the effectively recycled amounts are already excluded in the waste composition
definition, otherwise they would cause soil emissions, which actually do not occur.

4 Particulate emissions
A waste-specific figure for total particulate mass emitted to air is derived from the waste-specific
masses transferred to air. For each element an oxidation and an according weight increase is assumed.
 LCI model of open burning 5. Incomplete combustion products 7

The oxides of following elements emitted to air are summed up to obtain total particulate mass:
phosphor, boron, chlorine, bromine, fluorine, iodine, silver, arsenic, barium, cadmium, cobalt,
chromium, copper, mercury, manganese, molybdenum, nickel, lead, antimony, selenium, tin,
vanadium, zinc, beryllium, scandium, strontium, titanium, thallium, tungsten, silicon, iron, calcium,
aluminium, potassium, magnesium, and sodium. For example for average municipal waste this yields
a total particulate mass of 6500 mg/kg waste, which is close to a value found in literature (7300 mg/kg
in World Bank 2008).
Ecoinvent requires inventory of particulate mass in three size classes. Information in (Downard et al.
2015) on particulate emissions in open burning of tires is used there to obtain a particulate profile:
PM>10: 20.0w%
PM2.5-10: 28.0w%
PM<2.5: 52.0w%
This tentative profile is applied to the waste-specific amount of total particulate mass.

5 Incomplete combustion products

From a literature survey 1 various emission factors for open burning of municipal waste were compiled.
Tab. 5.1 shows the geometric mean of the found literature values. From the literature only values
which represent a municipal waste mixture were heeded. In the following, the way these factors were
used in the inventory are described.
For most pollutants simply a constant emission factor per kilogram of waste input is assumed. For
these pollutants it is assumed that it is less waste composition, but more the characteristics of
combustion (temperature, oxygen supply, burn reaction heterogeneities etc) that determine their
magnitude, which were called "process-specific" emissions in (Doka 2003-II) to contrast them to the
waste-specific emissions, which depend strongly on waste input composition.

Tab. 5.1 Geometric mean of literature survey of emission factors for open burning of municipal waste

Pollutant unit Emission to air

Nitrogen oxides mg/kg waste 2'700
Carbon monoxide mg/kg waste 38'600
Methane mg/kg waste 5'900
Dinitrogen monoxide mg/kg waste 116
Benzene mg/kg waste 980
Phenol mg/kg waste 15
Styrene mg/kg waste 528
Toluene mg/kg waste 372
Formaldehyde mg/kg waste 444
Acetaldehyde mg/kg waste 428
Benzene, ethyl- mg/kg waste 327
PAH, polycyclic aromatic hydrocarbons mg/kg waste 344
Polychlorinated biphenyls mg/kg waste 0.112

1
Gullett et al 2000, Costner 2006 , UNEP 2005, EPA 1992, Gullett and Touati 2003, Lemieux 2002, Leung et al 2006, Doka
2000, Lemieux et al. 2003, World Bank 2008, Downard et al. 2015, IPCC 2006-5
LCI model of open burning 5. Incomplete combustion products 8

5.1 Nitrogen emissions

A simplified model of NOx formation is applied here. The total NOx emissions to air have three
contributions:
- Fuel-NOx, which originates from the oxidation of nitrogen in the fuel.
- Thermal-NOx , which originates from N2 in combustion air, oxidised at high temperatures
>1000°C
- Prompt-NOx, which originates from N2 in combustion air reacting with free hydrocarbon
radicals (CH·).
Fuel-NOx
For this model the user-specified waste is the fuel. The fuel-NOx is therefore the part of nitrogen in
the waste that ends up as NOx in flue gas. According to the incineration model in (Doka 2003) 37.38%
of fuel nitrogen in untreated incineration (without DeNOx) is in NOx, the rest mostly in N2. The air-
bound nitrogen in a waste composition is converted with this factor to NOx .
Thermal-NOx
Formation of thermal NOx depends on a range of combustion parameters, most importantly
temperature and air excess, and not on the nitrogen content in fuel. Even in controlled incineration
there is ambiguous information of the share of thermal NOx on total NOx emissions, ranging from 20%
to 100% (Doka 2003-II:40). A value of 50% was used in the controlled incinerator model. No
literature information of thermal-to-total-NOx ratios could be found for the present open burning
model. It can be surmised that for open burning, combustion temperatures are lower and that the ratio
is lower than for controlled incineration. A tentative default value of 30% is chosen as a default value.
The user can set other values. The literature survey suggests a value of 2700 mg per kg waste for total
NOx emissions (Tab. 5.1). With the chosen default share of thermal NOx of 30%, an average emission
of 810 mg thermal NOx per kg waste results. The contribution from thermal NOx is added to the
contribution from fuel NOx in the inventory. Each waste composition receives an emission from
thermal NOx regardless of its nitrogen content (process-specific emission).
Prompt-NOx
No special modelling of prompt NOx is attempted, as information is even more scarce than for thermal
NOx. Also this is a process-specific emission depending more on the details of the combustion process
than on waste composition. Any forthcoming knowledge on prompt-NOx in open burning can be
covered by adjusting the thermal-NOx-share parameter accordingly.
Nitrous oxide N2O
The small amount of N2 O from Tab. 5.1 is inventoried as process-specific, constant emission.

5.2 Carbon monoxide and methane

Constant process-specific emissions of CO and CH4 are inventoried for each kilogram of waste,
regardless of composition. The formation of carbon dioxide from waste is corrected by the amounts of
carbon in CO and CH4.
LCI model of open burning 5. Incomplete combustion products 9

5.3 Organic carbon compounds

The range of emissions of benzene, phenol, styrene, toluene, formaldehyde, acetaldehyde, ethyl-
benzene, and polycyclic aromatic hydrocarbons are inventoried as constant process-specific emissions
for each kilogram of waste, regardless of composition.
The incineration model is focussed on fate of chemical elements and fate of individual compounds is
not modelled. Compounds are only inventoried in a summary fashion as process-specific emissions. If
there is reason to believe the burning could release significant amounts of specific compounds like
flame retardants or plasticizers these emissions should be estimated additionally.

5.3.1 Dioxin
Dioxin to air
Gullett et al. (2000) performed open burning tests on various municipal waste mixtures and found that
their total chlorine content is a fair predictor for emissions of dioxin, regardless of the forms of
chlorine.
Chlorine content in Dioxin emission
burnt mixture factor
mg/kg ng TEQ /kg waste
MSW baseline 2'000 71
MSW with 0% PVC 2'000 14
MSW with 1% PVC 8'000 201
MSW with 7.5% PVC 47'000 4916
MSW with 7% CaCl2 added 46'752 734

These measurements are used here to derive a power-law regression formula to coarsely estimate
waste specific dioxin generations based on chlorine content:

Eq. 5.1 Odioxin,air = 0.001683" ICl 1.296881

where,
Odioxin,air Output of dioxin in flue gas, in [ng/kg waste]
! ICl Input of chlorine in waste, in [mg/kg waste]

Please note that this formula is only valid for uncontrolled open burning. In proper incinerators
temperatures are higher, dioxin formation is lower and generally less dependent on chlorine content,
but more on incinerator design (Doka 2003-II:44).
In an LCA framework the dioxin emissions show not to be very relevant when compared to other
pollutants. For this reason the coarse estimations of dioxin emissions made here – including neglecting
to heed the exact form of chlorine – are of little overall importance. By comparison, emissions of
benzene from open burning are about five orders of magnitude more damaging than dioxin emissions,
or particulate emissions are about nine orders of magnitude more damaging than dioxin emissions
(weighted with ReCiPe'08 HA). Dioxin emissions are therefore inventoried only for completeness, but
they will likely not be relevant, but this depends on the applied Impact Assessment method.
LCI model of open burning 5. Incomplete combustion products 10

Dioxin in residues
Costner (2006) lists ranges of dioxin releases from open burning of municipal waste to air and to solid
residues. For a medium MSW composition average dioxin emissions to air are approximately 3.6
larger than average dioxin emissions to residues. This relation is used here to estimate dioxins in
residues (ultimately to soil) from the emissions of dioxins to air derived above.
Odioxin,air
Odioxin,residue = 3.6
Eq. 5.2

where,
Odioxin,residue Output of dioxin in solid residues, in [ng/kg waste]
! Odioxin,air Output of dioxin in flue gas, in [ng/kg waste]

5.4 Properties of exchanges

In Ecospold2 each exchange can – besides a name and an amount – have several properties describing
the exchange in more detail, like dry mass, water content, fossil carbon content etc. The tool adjusts
also some properties of exchanges. An overview is given in the following.
5.4.1 Waste input
The specific waste composition is a core aspect of these calculation tools and thus an extensive range
of properties are defined for the waste input, including trace amounts. The waste is however not
described in full, as for waste consisting of more fraction (complex wastes) detail is lost. This is
important to understand especially for complex wastes with heterogeneous fractions, i.e. mixtures of
burnable and inert waste or mixtures of degradable and less degradable fractions. It not possible to
back-engineer the composition of individual fractions from a given complex mixture's composition
alone. New waste fractions shall therefore be clearly documented in an accompanying report.
5.4.2 Emissions
Also elementary exchanges possess properties, describing their characteristics. Commonly this
information is constant (e.g. an emission of copper has a dry mass of 100%, a fossil carbon content of
zero etc.). For carbonaceous emissions, the distinction between fossil and non-fossil carbon can
become relevant for some purposes. In the ecoinvent database this issue is dealt with in two different
ways:
1. Different exchanges for fossil/non-fossil emissions: CO2, CH4, CO each have two different
exchanges with different names, e.g. "Carbon dioxide, fossil" and "Carbon dioxide, non-
fossil". This allows to assign different characterisation factors in LCIA, e.g. giving biogenic
CO2 a GWP of zero.
2. Different properties for fossil/non-fossil carbon within an exchange. The carbon content
property is split up in two variants, which must be set separately (the total carbon property
does exist in ecoinvent, but cannot be changed by a normal user). For instance an emission of
ethylene can be of fossil or non-fossil origin, or a mixture of both. All exchanges possess
these properties as defaults. The figures for the defaults were defined when the exchanges
themselves were defined. Currently (Jan 2017) the default properties of ecoinvent's emissions
are chiefly set to contain 100% fossil carbon, though for some emissions a 50/50 split was
employed (e.g. for NMVOC or methanol).
The previous waste tools of 2003 already discerned different emissions according to (1) like fossil and
non-fossil CO2, and this is maintained in the present models. For instance burning of plastic will lead
to fossil CO2 emissions, while burning of biomass waste will lead to non-fossil CO2. The prerequisite
 LCI model of open burning 5. Incomplete combustion products 11

for this is that the used waste material has a properly defined share of biogenic carbon in the waste
definition, cf. Calculation Manual, point 148.
Additionally in the tools of the current project also carbonaceous emissions like benzene or dioxins
receive adjusted properties, according to (2) above. E.g. the benzene emission associated with burning
paper will have its fossil carbon content property set to zero, while the benzene emission associated
with burning common plastic will receive a zero biogenic carbon content property. This is calculated
automatically from waste input properties and transfer coefficients. For the open burning model
following procedures are employed.
Share of non-fossil carbon in air emissions is derived from the mass of non-fossil carbon in the
burnable waste fraction(s) over total carbon in the burnable waste fraction(s), see Eq. 5.3. The
denominator is the sum of all carbon present in burnable waste fraction(s) and is calculated from
carbon content and the burnability flag of the employed waste fraction(s). The numerator is the sum of
all non-fossil carbon in the burnable waste fraction(s) and is calculated from carbon content, share of
non-fossil carbon and the burnability flag of the employed waste fraction(s). If a chosen waste
composition does not contain any carbon, a default share of 60.7% non-fossil carbon is used for
carbonaceous air emissions (see chapter 5.3 'Organic carbon compounds' on page 9).
$ ' $ '
%bair = && # f i " mC ,i " Bi " %bi )) && # f i " mC ,i " Bi ))
Eq. 5.3 % fraction i ( % fraction i (

where,
%bair Fraction of non-fossil, biogenic carbon in air emissions, [weight-%]
! fi Share of fraction i in selected waste (mixture), [weight-%]
mC,i Mass of all carbon in fraction i, [kg carbon/ kg wet waste]
Bi Indicator of burnability of fraction i, [0 or 1]
%bi share of biogenic carbon in fraction i, [weight-%]

Soil emissions derive from the ash residues of open burning. In the model these residues have two
sources: inert, unburnable waste materials which are assumed to go directly to ash, and the parts of
burnable waste which end up in ash according to the transfer coefficients. Also here the share of non-
fossil carbon is calculated by relating the sum mass of all non-fossil carbon ending up in residues to
the total mass of carbon in residues is calculated. Either is calculated from waste composition, the
burnability flag of the waste fraction(s) and the transfer coefficient for carbon into residues of 0.5%
(cf. Tab. 7.1 on page 14). If a chosen waste composition does not contain any carbon, a default share
of 30.32% non-fossil carbon is used for soil emissions, based on average municipal waste.
% ( % (
(
%bsoil = '' $ f i " mC,i " Bi " TKC + (1 # Bi ) " %bi ** ) ( )
'' $ f i " mC ,i " Bi " TKC + (1 # Bi ) **
Eq. 5.4 & fraction i ) & fraction i )

where,
%bsoil Fraction of non-fossil, biogenic carbon in soil emissions, [weight-%]
!
fi Share of fraction i in selected waste (mixture), [weight-%]
mC,i Mass of all carbon in fraction i, [kg carbon/ kg wet waste]
Bi Indicator of burnability of fraction i, [0 or 1]
TKC Transfer coefficient for carbon to soil in burnable materials, = 0.5%
%bi share of biogenic carbon in fraction i, [weight-%]
 LCI model of open burning 6. Modelling parameters 12

Note that the term (Bi · TKC + (1-Bi)) is equal to TKC for burnable mateials (when Bi is 1) and equals 1
for unburnable materials (when Bi is 0).

6 Modelling parameters
The model includes several parameters which characterise the open burning process. These are
informations that are more associated with a certain site or location, than a particular waste material.
6.1 Geography code
The open burning model is geographically independent, apart from the user-defined waste input
composition which can be dependent on location. The geographic code will be incorporated in the
Ecospold2 inventory.

6.2 Subcategory of air emissions

The user can choose if the air pollutants are released in high-population (urban) air, or in low-
population (rural) air. A choice of high-population air will not depict health burdens imposed on
workers in close proximity to the open burning site.
Entering the code "hi" chooses high-population air, anything else the default low-population air.

6.3 Subcategory of soil emissions

The user can choose if the soil pollutants are released to industrial soil, or in agricultural soil.
Agricultural soil can – depending on the Impact Assessment method – lead to larger human health
burdens.
Entering the code "agri" chooses agricultural soil, anything else the default industrial soil.

6.4 Share of thermal NOx

For the attribution of thermal NOx emissions, a ratio of (thermal NOx to total NOx) must be given. A
default of 30% is suggested. The function of this parameter in this model is explained in section
'Thermal-NOx' on page 8.

6.5 Recycling rates

For iron, aluminium and copper the user can specify how much of the metal in the waste will actually
be recycled from burn residues. A figure of 100% signifies that all of the element in bulk metallic form
is separated. The recycling rate defines how much of the maximally recyclable bulk metal is actually
recycled. The maximally recyclable bulk metal is part of the waste material definition.
An example: a waste has a total copper content of 1 gram per kg. Of that copper only 50% is in bulk
metallic form.2 The rest is in oxides and distributed in traces of materials. So maximally only 0.5 gram
of copper metal can be gained from this waste. Recycling rates are usually not perfect and small bits or

2
This figure is part of the waste material definition and refers to the make-up of the material, and not to its treatment.
LCI model of open burning 7. Annex A: Transfer Coefficients 13

devalued items cannot be recycled. If the recycling rate is set for instance at 70% it means that in this
case 0.35 gram of copper metal will be separated from residues (=0.5 g · 70%). The emissions to soil
will be corrected by this amount.
The user can define recycling rates for iron, aluminium and copper. For other recyclables see chapter
3.3 'Recycled materials' on page 6.

7 Annex A: Transfer Coefficients

The transfer coefficients for elements in burnable materials are shown in Tab. 7.1. The transfer
coefficients for elements in unburnable materials are 100% to residue.
The waste-specific transfer coefficients for a particular waste mixture of burnable and unburnable
components depend on the relative distribution of elements in burnable and unburnable materials and
is calculated from the previous two statements.
Uncertainty of transfer coefficients
The results from the open burning model must be taken as rough estimates for open burning. Open
burning is by its very nature largely uncontrolled can include a wide possible range of different
combustion conditions – from agricultural burning of straw on fields to smouldering waste fires. Large
differences in conditions can even occur within one burning event. The present model is intended for
open burning of municipal waste fractions and due to the large possible variability in burning
conditions the results bear considerable uncertainties. In (Doka 2003-II) the uncertainty of transfer
coefficients was derived from a generic formula, giving large uncertainties to small transfer
coefficients.

Eq. 7.1 GSDTK = N ! ln(mTK ) + 1

where,
GSDTK Geometric standard deviation of the transfer coefficient
mTK Mean value of transfer coefficient, in [kg/kg]
N constant

In (Doka 2003) the value for N was chosen so that the GSD2 value for the smallest mTK values (metals
to air) are 200%. I.e. at most the confidence interval for the transfer coefficient is assumed to stretch
from double to half of the mean value. This resulted in a value for N of -0.022.
For open burning it is assumed that due to burn condition heterogeneities the transfer coefficients tend
to be even more uncertain than in controlled municipal incineration. The value of N is decreased to
-0.0546, which results in maximal GSD2 values of 300%, i.e. for small transfer coefficients the
confidence interval stretches from one third to thrice the mean value. As always, for emissions the
uncertainty of transfer coefficients is combined with the uncertainties of waste composition.
LCI model of open burning 7. Annex A: Transfer Coefficients 14

Tab. 7.1 Mean transfer coefficients for elements in burnable materials (rounded to 5 digits)

Emissions to Emissions to air

residues
g/kg waste g/kg waste
O 38.343 961.66
H 0 1000
C 5.0151 994.98
S 856.43 143.57
N 10.012 989.99
P 993.13 6.8707
B 999.6 0.39996
Cl 332.89 667.11
Br 679.32 320.68
F 893.39 106.61
I 332.89 667.11
Ag 994.87 5.1321
As 991.48 8.5229
Ba 958.78 41.216
Cd 994.34 5.6578
Co 959.67 40.333
Cr 998.51 1.4912
Cu 998.9 1.0991
Hg 393.94 606.06
Mn 991.49 8.5121
Mo 995.01 4.9918
Ni 998.57 1.4267
Pb 991.6 8.4033
Sb 1000 0.0024032
Se 913.33 86.668
Sn 1000 0.0019596
V 922.79 77.209
Zn 992.56 7.4399
Be 998.11 1.89
Sc 999 1
Sr 999 1
Ti 989.4 10.603
Tl 998 2
W 1000 0
Si 1000 0.0046726
Fe 998.19 1.8119
Ca 976.39 23.61
Al 999.18 0.82428
K 985.18 14.822
Mg 997 2.9959
Na 975.37 24.631
LCI model of open burning 8. References 15

8 References
Web addresses indicate the source of electronic documents. The subsequent (date in brackets) refers to the date of retrieval.

Costner 2006 Costner P (2006) Update of Dioxin Emission Factors for Forest Fires, Grassland and Moor Fires, Open
Burning of Agricultural Residues, Open Burning of Domestic Waste, Landfills and Dump Fires.
International POPs Elimination Network, 15 November 2006
Doka 2000 Doka G (2000) Ökoinventar der Entsorgungsprozesse von Baumaterialien - Grundlagen zur Integration der
Entsorgung in Ökobilanzen von Gebäuden. Doka Ökobilanzen, Zürich. IEA CBS Annex 31: Energy
Related Environmental Impact of Buildings. Swiss Federal Office for Energy (Bundesamt für Energie).
Februar 2000
Doka 2003-II Doka G (2003) Life Cycle Inventories of Waste Treatment Services, part II: waste incineration. ecoinvent
report No. 13, part III. Swiss Centre for Life Cycle Inventories, Dübendorf, 2003.
http://www.doka.ch/13_II_WasteIncineration.pdf
Doka 2003-III Doka G (2003) Life Cycle Inventories of Waste Treatment Services, part III: Landfills - Underground
Deposits - Landfarming. ecoinvent report No. 13, part III. Swiss Centre for Life Cycle Inventories,
Dübendorf, 2003. http://www.doka.ch/13_III_Landfills.pdf
Doka 2013 Doka G. (2013) Updates to Life Cycle Inventories of Waste Treatment Services - part II: waste incineration.
Doka Life Cycle Assessments, Zurich, 2013. Available at
http://www.doka.ch/ecoinventMSWIupdateLCI2013.pdf
Downard et al. 2015 Downard J, Singh A, Bullard R, Jayarathne T, Rathnayake CM, Simmons DL, Wels BR, Spak SN, Peters T,
Beardsley D, Stanier CO, Stone EA (2015) Uncontrolled combustion of shredded tires in a landfill - Part 1:
Characterization of gaseous and particulate emissions. Atmospheric Environment 104 (2015) p. 195-204
EPA 1992 Emission Factor Documentation For AP-42 Section 2.4, Open Burning. U.S. Environmental Protection
Agency 1992
Gullett and Touati 2003 Gullett, B.K. and Touati, A. (2003). PCDD/F emissions from forest fire simulations, Atmospheric
Environment 37, p. 803-13.
Gullett et al 2000 Gullett, B. K. , P. Lemieux, C . Winterrowd, D . Winters. 2000. PC DD/F Emissions from Uncontrolled,
Domestic Waste Burning. Presented at Dioxin ’00, 20th International Symposium on Halogenated and
Environmental Organic Pollutants & POPs, held Aug 13-17 at Monterey, CA. Corrected revision of short
paper in Organohalogen Compounds 46: 193-196.
IPCC 2006-5 Guendehou GHS, Koch M, Hockstad L, Pipatti R, Yamada M (2006) 2006 IPCC Guidelines for National
Greenhouse Gas Inventories, Volume 5 Waste, Chapter 5: Incineration and Open Burning of Waste.
Intergovernmental Panel on Climate Change (IPCC). http://www.ipcc-
nggip.iges.or.jp/public/2006gl/pdf/5_Volume5/V5_5_Ch5_IOB.pdf (26 Aug 2010)
Lemieux 2002 Lemieux P (2002) Emissions of Organic Air Toxics from Open Burning. Prepared for United States
Environmental Protection Agency Office of Research and Development Washington, DC, EPA-600/R-02-
076, October 2002. http://nepis.epa.gov/Adobe/PDF/P1001G31.pdf
Lemieux et al. 2003 Paul M. Lemieux , Brian K. Gullett , Christopher C. Lutes , Chris K. Winterrowd & Dwain L. Winters
(2003) Variables Affecting Emissions of PCDD/Fs from Uncontrolled Combustion of Household Waste in
Barrels, Journal of the Air & Waste Management Association, 53:5, 523-531
Leung et al 2006 Leung A, Cai ZW, Wong MH (2006) Environmental contamination from electronic waste recycling at
Guiyu, southeast China. J Mater Cycles Waste Manag (2006) 8:21–33
UNEP 2005 cited in p74 Amoyaw-Osei Y, Agyekum OO, Pwamang JA, Mueller E, Fasko R, Schluep M (2011) Ghana
e-Waste Country Assessment. Ghana e-Waste Country Assessment, SBC e-Waste Africa Project.
"Standardized Toolkit for Identification and Quantification of Dioxin and Furan Release" (UNEP
Chemicals 2005, Edition 2.1)
World Bank 2008 Global Study for Purpose of Global World Bank Guidance Development - Solid Waste Management
Holistic Decision Modeling. Final Report. World Bank. Japan Country-Tied Fund, submitted by Nippon
Koei Co., Ltd., June 2008