Science of the Total Environment 370 (2006) 190 – 206

www.elsevier.com/locate/scitotenv

Formation, transformation and transport of black carbon

(charcoal) in terrestrial and aquatic ecosystems
M.S. Forbes a,⁎, R.J. Raison b,c , J.O. Skjemstad a,c
a
CSIRO Land and Water, PMB2, Glen Osmond, South Australia 5064, Australia
b
ENSIS, P.O. Box E4008, Kingston, ACT 2604, Australia
c
CRC for Greenhouse Accounting, GPO Box 475, Canberra 2601, Australia
Received 9 February 2006; received in revised form 7 June 2006; accepted 16 June 2006
Available online 24 July 2006

Abstract

Black carbon (BC) is ubiquitous in terrestrial environments and its unique physical and chemical properties suggest that it may
play an important role in the global carbon budget (GCB). A critical issue is whether the global production of BC results in
significant amounts of carbon (C) being removed from the short-term bio-atmospheric carbon cycle and transferred to the long-
term geological carbon cycle. Several dozen field and laboratory based studies of BC formation during the burning of biomass have
been documented. Findings are difficult to interpret because they have been expressed in an inconsistent manner, and because
different physical and chemical methods have been used to derive them. High error terms documented in many of these studies also
highlight the problems associated with the quantification of the amount of biomass C consumed in fire, the amount of residue
produced and the constituents of that residue.
To be able to estimate the potential for BC as a carbon sink, issues regarding its definition, the methods used in its identification
and measurement, and the way it is expressed in relation to other components of the carbon cycle need to be addressed. This paper
presents BC data in a standard way; BC production as a percentage of the amount of C consumed by fire (BC/CC), which can be
readily integrated into a larger carbon budget. Results from previous studies and new data from Australian ecosystems were
recalculated in this way. As part of this process, several BC estimates derived solely from physical methods were discarded, based
on their inability to accurately identify and quantify the BC component of the post-fire residue. Instead, more focus was placed on
BC estimates obtained by chemical methods. This recalculated data lowered the estimate for BC formation in forest fires from 4%
to 5% to < 3% BC/CC. For savannah and grassland fires a value of < 3% is consistent with reported data, but considerable variation
among estimates remains. An updated flow-chart linking the sources, fluxes and pools of BC formed in the terrestrial environment
with the aquatic and marine environments, and estimates of mean residence times for BC are also presented.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Black carbon; Fire; Charcoal; Global carbon cycle; Carbon sequestration

1. Introduction

Charred materials are the product of the incomplete

combustion of vegetation and fossil fuels, resulting from
⁎ Corresponding author. Tel.: +61 8 83038444; fax: +61 8 83038550. gas phase processes or pyrolysis in the residues of liquid
E-mail address: [email protected] (M.S. Forbes). or solid fuels (Goldberg, 1985). The continuum of
0048-9697/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.scitotenv.2006.06.007
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 191

combustion products such as char, ash and charcoal are forest and 4% (BC/initial biomass) for grasslands in
commonly referred to as black carbon (BC). BC is a Australia (Harwood and Jackson, 1975; Flinn et al.,
heterogeneous mixture of residues that have contrasting 1979; Raison et al., 1985a; Graetz and Skjemstad,
chemistry and thus resistance to further biological or 2003), and 1–3% (BC/C exposed) for tropical rain-
chemical degradation. Together they occur ubiquitously forests (Barbosa and Fearnside, 1996; Kuhlbusch et al.,
in soils and other terrestrial sediments, and via fluvial 1996; Fearnside et al., 1993, 1999, 2001; Gráça et al.,
and atmospheric transport, in marine sediments as well. 1999). These estimates are not easy to compare because
Over the past few decades BC concentrations in the they have been derived by various methods and they are
earth's atmosphere and biosphere have become of expressed in different terms, resulting in significant
interest because; in aerosol form they are strong problems of interpretation, and in difficulties in using
absorbers of solar radiation, they can provide a record them to calculate a global cycle for BC. Kuhlbusch (1999)
of palaeo-environments in sediment and ice cores, and highlighted such problems by providing two estimations
they may also be a sizable contributor of oxygen to the of global BC production. The first, 40–179 Tg year− 1 was
atmosphere over geological time frames (e.g. Seiler and based on BC/C exposed conversion factors, compared to
Crutzen, 1980; Crutzen and Andreae, 1990; Kuhlbusch an estimate of 50–270 Tg year− 1 derived from BC/CO2
and Crutzen, 1996; Kuhlbusch, 1995, 1998, 1999; emitted conversion factors. Thus a standard expression for
Schmidt and Noack, 2000; Graetz and Skjemstad, 2003; the production of BC is needed in order to be able to
Schmidt, 2004; Druffel, 2004; Preston and Schmidt, synthesise information, and to calculate the effect of BC
2006). BC, in particular the charcoal component, is also on the GCB. Expressing BC formation as a percentage of
important because it represents one of the few ways that the amount of total carbon actually consumed (BC/CC) in
carbon can be rendered relatively inert, such that it can a fire event (including carbon converted to CO2, CO, ash,
not easily recombine with oxygen to form CO2. Hence BC, smoke and soot) provides a meaningful measure of
there is a strong potential for BC to act as a significant the effects of combustion in a range of fire types, and
removal (sink) of carbon from the more rapid bio-atmo- would allow for BC estimates to be easily incorporated
sphere carbon cycle to the slower (long-term) geological into any GCB.
carbon cycle. A major reason why the quantification of BC is signif-
Seiler and Crutzen (1980) were the first to assess the icantly affected by the method of assay (Novakov et al.,
role of BC as a sink in the global carbon budget (GCB). 1997; Schmidt and Noack, 2000) is that BC by definition
Based on past demographic and land use data they is a range of materials from partly charred plant matter to
estimated global BC production to be 500–1700 Tg fully graphitised soot particles (Jones et al., 1997). This
year− 1 (1 Teragram = 1 Mt) or 20–30% of the total has resulted in many ambiguous terms being used to
biomass burned. These estimates were found to be too define charred particles, especially those that occur in
high and were revised down to 200–600 Tg year− 1 by sediments. Procedural inaccuracies can also occur during
Crutzen and Andreae (1990). Experimental determina- BC determinations, resulting in large uncertainties in the
tions of the BC content of various post-fire biomass weight and/or volume estimations of the various pre-fire
residues allowed Kuhlbusch and Crutzen (1995) to biomass components and of the post-fire residues.
provide a further reduced range of 76–241 Tg year− 1. In Furthermore, the manner in which the amount of BC
a follow-up publication, Kuhlbusch and Crutzen (1996) produced has been expressed is not uniform. It has been
again reduced the estimate by 20–40% to 44–194 Tg expressed as a percentage of the total biomass consumed
year− 1, noting work undertaken by Fearnside et al. (1993), (BC/TBC), the amount of carbon that is exposed or
which demonstrated the high variability (0.5–4.5%) of available to the fire (BC/CE) or the amount of carbon
charcoal formation in fires associated with the clearing of combusted (BC/CO2). Therefore, when estimating BC
Amazonian rainforests. production and comparing various reported data, it is
Various field and lab-based estimates of BC important that the definition of what is being measured
production (in particular charcoal and ash) from and how it is measured and reported are critically assessed.
biomass burning have been made, including the follow- Jones et al. (1997) proposed a set of standard
ing: 0.5–2.0% (BC/CO2 emitted and BC/initial bio- terminologies applicable to all disciplines that deal
mass) for savannas (Kuhlbusch et al., 1996; Kuhlbusch with fire-altered plant matter. This covered partially
and Crutzen, 1996; Carter et al., 1998), 0.7–6% (BC/C burned post-fire residues and ash (which contains some
exposed) for Northern Hemisphere boreal forests organic materials), smoke, soot and the significantly
(Kuhlbusch and Crutzen, 1996; Clark et al., 1998; altered charcoal and other BC components. Using this
Czimczik et al., 2003; Lynch et al., 2004), 1–15% for terminology, BC was defined as any form of fire-derived
192 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

polymeric, aromatic to graphitic carbon fraction in fire- Table 1

altered material or smoke particulates. Charcoal is any Methods for estimating post-fire biomass consumption, for collection
of postfire residues, and their limitations
blackened plant-derived material that has had its
chemical composition and ultrastructure significantly Method and Summary Commentry
altered as a result of heating in a fire, but retains a Gravimetric Determinations
recognizable atomic structure of the parent plant, even if Estimation of biomass prior to Can determine the amount of
and after a fire event post-fire residue created and thus
only in a fragmentary form. Hence, based on these
calculation of the percentage of
definitions, charcoal is a component of BC. Another the biomass consumed
major component of the BC continuum is ash, defined Utilisation of two adjacent plots, Accuracy of biomass estimations
by Jones et al. (1997) as the mineral-rich powdery or replicate sampling of micro- is questionable, especially if
residue that remains on site after a fire. It is debateable plots before and after fire paired plot approach used
Best suited to biomes with
whether or not the BC forms of either graphite or soot
non-woody biomass
are part of the BC continuum, because they are not (savannas and grasslands)
formed by different mechanisms to charcoal and ash. Problematic when woody
Yet even though ash and charcoal are components of the materials are present
BC continuum, they are not entirely comprised of BC. Post-fire residue can be Post-fire residue is the most
collected by various difficult to accurately measure
As will be discussed later, other mineral and organic
means including
components can contribute to many post-fire residues. 1. Trays or Need to extrapolate the weights
Therefore, the clarification of the BC component in traps7,10,11,12,13,14 from trays to total area
post-fire residues such as charcoal and ash is essential Fine ash may be blown off trays
for a better appreciation of the effect that BC has on the Can collect other components
2. Manually2,3,4,5,6,10,15 Fine charred materials
GCB. In some studies there is clear definition between
will be lost, and there can be
the components, whilst in many there is not. In the soil contamination
ensuing literature review, the production of BC residues 3. Vacuum cleaner1,9 Collects more than post-fire
(predominantly in the form of charcoal and ash) from a residue (ie silts)
range of vegetation fires are discussed after the BC
Volumetric Determinations
production has been converted to a common basis of
Based on estimations of Can be erroneous as wood
carbon consumed by fire (BC/CC). The limitations on biomass prior to and density can vary in pre- and
some of the methods used to derive these estimations are after a fire event post-fire residue
also discussed. Woody mass estimated Can use sampling based on
from volume8 line intersects, or all logs in
area can be measured
2. Methodology Can be determined for all woody Line intersect method must
components or by sampling account for biomass
2.1. Determining BC production based on line transects heterogeneity within area
Loss of log volume in fire is Problems arise in the
Both laboratory and field-based approaches have used to estimate determination of what is
BC production incorporated into post-fire
been used to estimate the production of BC, in the form
residue
of ash and charcoal, from biomass burning (e.g. Problems occur if some logs
Harwood and Jackson, 1975; Flinn et al., 1979; Raison are completely combusted, final
and McGarity, 1979; Raison et al., 1985a; Barbosa and volume can not be determined
Fearnside, 1996; Kuhlbusch et al., 1996; Fearnside et al., Note:
1993, 1999, 2001; Carter et al., 1998; Clark et al., 1998; The volumetric method can be used in conjunction with the gravi-
Gráça et al., 1999; Ohlson and Tryterud, 2000; González metric method for areas where both coarser woody and non-woody
materials exist.2,4
et al., 2002; Czimczik et al., 2003; Graetz and Skjemstad,
These methods can allow reasonable quantification of the amount of
2003; Lynch et al., 2004). Despite procedural differences biomass consumed, and collection of post-fire residue created from a
between these studies, they are all based on a common fire; but they are severely flawed when it comes to the identification
theme of comparisons of C in biomass prior to a burning and quantification of BC in the post-fire residue (see Table 3).
event and C in components of post-fire residues. The References: 1. Graetz and Skjemstad (2003), 2. Fearnside et al. (1999),
3. Fearnside et al. (1993), 4. Fearnside et al. (2001), 5. Gráça et al.
amount of C in pre-fire fuels (biomass) is estimated
(1999), 6. Barbosa and Fearnside (1996), 7. Lynch et al. (2004), 8.
either gravimetrically or volumetrically (Table 1). These Tinker and Knight (2000), 9. Kuhlbusch et al. (1996), 10. Carter et al.
two methods can also be used to approximate the various (1998), 11. Clark et al. (1998), 12. Ohlson and Tryterud (2000), 13.
post-fire residues. However, methods utilising various Hurst et al. (1994), 14. Raison et al. (1985a), 15. Raison et al. (1985b).
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 193

Table 2 the quantification of BC is useful as they can to some

Methods for identifying and quantifying post-fire residues extent overcome the inaccuracies associated with purely
Method and Summary Commentry visual identification of post-fire residues. Nonetheless,
Thermal/chemical Oxidation these above-mentioned methods are always used in
Residue concentrated by conjunction with the gravimetric and volumetric
two step method methods, which are essential in the determination of
(1) Treatment with aqueous Thermal/chemical oxidation method
the amount of biomass fuel actually consumed, and the
solvents to remove as much can cause pyrolysis of carbon and in
inorganic and organic C turn the formation of BC-like amount of post-fire residues formed. We briefly discuss
as possible compounds, and thus estimation of some of the benefits and problems associated with the
(2) Remaining OC removed BC in the post-fire residue physical and chemical techniques used in estimating BC
by thermal treatment at production.
340 °C in pure oxygen
for 2 h
BC is determined by Produces quite accurate estimation 2.2. Gravimetric quantification of BC
measuring the H/C ratio of BC amount in the post fire
via elemental analysis residue but slow and expensive Gravimetric (or direct) determinations of BC produc-
of the residue3 procedure tion (charcoal and/or ash) are based on the quantification
A lower H/C ratio indicates a Enables estimation of non-BC
of the weight of fuel consumed and the weight of post-fire
higher degree of component in post-fire residues
aromaticity leading up to residue produced by fire. Then using estimates of the
elemental or carbon content of biomass (45–50%), charcoal (60–70%)
graphitic carbon and ash (15–20%, but can be highly variable), these
values can be adjusted to provide an estimation of BC
Loss on Ignition
produced as a percentage of carbon consumed. However,
Post-fire residue weighed Problem in defining the
and combusted5 heterogeneous charred residue significant errors can emerge regarding mass estimations
components for pre-burn fuels and post-burn residues due to
The difference between the Mineral component in ash can create heterogeneity in biomass distribution, and problems in
pre and post combustion an error in estimations accurately recovering post-fire residues.
weight is the organic
The retrieval and quantification of the BC in resi-
fraction including the
charred component dues can be achieved by various means. It can be
collected by hand (Kuhlbusch et al., 1996; Barbosa and
13
C-NMR Spectroscopy Fearnside, 1996), collected in traps that are placed into
Quantitates Aromatic-C Accurate but costly method for the land surface (Raison et al., 1985a; Hurst et al.,
component1,2,3 identifiying different organic
components in post-fire residue
1994; Carter et al., 1998; Lynch et al., 2004) or
including BC retrieved by vacuuming (Harwood and Jackson, 1975;
Kuhlbusch et al., 1996; Graetz and Skjemstad, 2003).
Visual Quantification All three methods are problematic as they can result in
Determined via Heterogeneity of BC on microscope the collection of not only components of the BC
identification of point slides can lead to errors in percentage
continuum, but other materials. For instance the vacuum
counting of charcoal % estimates
on slides viewed under collection of BC generally results in samples being
microscope6,7 contaminated with very fine sand regardless of care
References: 1. Skjemstad et al. (1999), 2. Czimczik et al. (2003), 3. taken (Graetz and Skjemstad, 2003), thus overestimat-
Kuhlbusch et al. (1996), 4. González et al. (2002), 5. Carter et al. ing BC production. The manual collection of post-fire
(1998), 6. Pitkänen et al. (1999), 7. Schmidt and Noack (2000). residues can result in much of the smaller charcoal and
ash particles remaining on site, causing BC production
forms of chemical analysis are required to reliably quan- to be underestimated (Fearnside et al., 1999). The trap
tify the composition of the post-fire residue components. method is also problematic because it is more effective
Methods for estimating the BC component of post- in collecting very fine particles, but does not recover the
fire residues (Table 2) include chemical oxidation and charcoal that has yet to be deposited in the soil, such as
H/C ratio determination (Kuhlbusch, 1995), the identi- standing burnt shrubs and trees (Lynch et al., 2004).
fication of molecular markers via 13C-NMR spectros- Traps, such as small metal trays may also have fine ash
copy (Skjemstad et al., 1999; González et al., 2002; blown off them. Hence all these collection methods
Czimczik et al., 2003) and loss on ignition (LOI) have the potential to produce erroneous BC percentage
determinations (Carter et al., 1998). Their application in estimations.
194 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

2.3. Volumetric quantification of BC of fuel consumption. The benefit in the application of a

chemical method is that it identifies the quantity of BC
Volumetric-based determinations of the production based on more stringently defined chemical properties,
of BC components by fire are similar in principle to the and is consistent in application. This reduces error in the
gravimetric method, yet are more applicable to the identification and estimation of the BC in post-fire residue.
charcoal produced from woody fuels with diameters Examples of chemical determinations of BC include
greater than 10 cm (Gráça et al., 1999; Fearnside et al., Kuhlbusch (1995), and Kuhlbusch and Crutzen (1996)
1999). This method is based on estimating the volume of who measured BC (defined as a highly poly-aromatic to
biomass for a given area prior to the burn event, elemental or graphitic carbon fraction in the particulate
compared to the volumes of unburnt material and material) in different post-fire residue fractions using a
charcoal (BC) that exist after. The difference between two-step pre-treatment method that removed all inor-
the volume of total wood prior to the fire and the volume ganic and organic carbon. BC was then quantified based
of unburnt wood present after can be used to estimate on its H/C ratio determined by elemental analysis.
the amount of wood combusted. The volume of charcoal Czimczik et al. (2003) took advantage of a naturally
produced can also be directly evaluated and expressed occurring surface fire to estimate BC production. Ana-
as a percentage of the fuel burnt or the fuel available. lysing both unburnt and burnt samples, the authors
Tinker and Knight (2000) use this volumetric method to determined mass, inorganic and organic elemental
calculate the amount of coarse woody debris (CWD) concentrations and calculated the corresponding stocks.
converted to charcoal for a fire in a North American Secondly they characterized the bulk molecular com-
boreal forest. Utilising the formula for a tapered cylinder position of organic carbon (OC) using 13 C-NMR
they made volume estimations for each fallen log with a spectroscopy, and then via integration of spectra quan-
diameter > 7.5 cm, in a designated burn area and across tified the aromatic carbon fraction. From this they were
several transects (using the Line Intersect Method). A able to calculate the conversion rate of OC to BC.
post-fire estimation of the volume of charcoal was made The use of 13C-NMR spectroscopy to accurately
in the same manner, with the charcoal produced (8%) determine BC content (e.g. Skjemstad et al., 1999)
being expressed as a percentage of the initial volume of allowed Graetz and Skjemstad (2003) to formulate a
biomass available for combustion (BC/CE). method for estimating the percentage of BC produced
Problems regarding this type of BC estimation can per unit C combusted for grassy vegetation. This ap-
occur particularly in cases where logs are completely proach compares carbon to silicon ratios (C/Si) in
combusted during the burn event, or where charcoal is unburnt plant material with those for charcoal and ash
lost from the charred log. These are dealt with by the residues. The procedure is based on the assumption that
identification and measurement of log shadows and silicon, a component of plant tissues, particularly in
trenches from which volumes can be estimated. The tropical grasses, occurs as silica (SiO2) in discrete
variability in charcoal depth around some logs and the phytoliths. Phytoliths are distinctive, well preserved and
diameter reduction of charred logs during the burn can non-volatile; hence they are conserved during the burn-
also lead to erroneous calculations. As a result of these ing process. Once C/Si values are obtained for plant
ambiguities and errors, authors using this method material and charred material, and assuming that the
generally consider results to be an underestimate of inert element Si is conserved during the burning process,
charcoal formation. then the ratio of carbon in the charred material to that in
the plant can be used to calculate the amount of carbon
2.4. Chemical quantification of BC consumed by the fire (Graetz and Skjemstad, 2003).
This calculation could be considered a lower range
The difficulties and uncertainties in estimating BC estimate due to the possibility that some of the fine
production using both gravimetric and volumetric field charcoal is re-distributed in smoke, or lost into the soil.
studies, and the imprecise definition of BC and its To use this approach, soil contamination by Si must be
components has resulted in the use of more specific avoided. The use of Ca in a similar manner to Si is also
chemical methodologies (Kuhlbusch and Crutzen, 1995; effective in the estimation of fuel consumption and BC
Kuhlbusch et al., 1996; Carter et al., 1998; Graetz and production in ash residues (Raison et al., 1985b). This is
Skjemstad, 2003; Czimczik et al., 2003). These chemi- due to the fact that the volatilisation of Ca is likely to be
cally based methods (Table 2) have been applied negligible in most vegetation fires.
predominantly as a second step in the quantification Carter et al. (1998) applied a different method to
process following gravimetric or volumetric estimations estimate the amount of OC converted to charcoal (BC),
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 195

Table 3 during a study of the rate of formation of carbonised leaf

Carbon and mineral contents of vegetation, charcoal and ash residues on five separate burnt plots of tropical pasture in
Sample Total Carbon Total Mineral southeast Queensland. Pasture (biomass) yields were
Fraction (%) Content (%) measured using conventional gravimetric methods, and
Vegetation then the fire residue collected on trays was sorted into
Spinifex Trodia sp B1 45.05 6.61 unburnt and burnt materials, dried at 105 °C and
Sorghum intrans B1 44.4 4.44
weighed. Following this, the burnt residue was tested for
Terminalia sp B1 46.31 5.76
Eucalyptus tetradonta 49.19 7.34 LOI (525 °C to constant weight) to determine the
Ironwood Chlorostrihyum sp 53.49 3.77 percentage of ‘inorganic component’ or ‘mineral
Acacia aneura Lila Vale B2 47.82 3.46 content’ in the material, which remains after combus-
Eucalyptus populnea B2 49.08 6.03 tion. The authors considered that charcoal was equiv-
Hetropogon sp B2 42.79 11.57
alent to the amount of mass lost during the LOI.
Themeda sp (dry) B2 42.44 11.62
Eucalyptus sp B2 44.72 6.39 The existence of a significant mineral component in
Cenchrus sp (Buffel Grass) B2 39.88 12.21 post-fire residues is highlighted in Table 3, which
Unknown sp Pentland B2 42.54 10.66 presents carbon and mineral contents of vegetation,
Terminalia sp ‘Cabbage’ B2 45.28 6.09 charcoal and ash residues for several fire events in the
Mitchell Astrebla sp Aramac B2 41.93 12.78
Northern Territory, Australia. These data identify a
Evagrostis eviapoda Lila Vale 41.43 16.52
Charcoal from woods and grasses higher mineral content and a lower carbon (BC) content
S1-35 (woody) 84.88 2.38 for ash residues in comparison to charcoal residues, with
S1-61 (woody) 75.1 2.84 the mineral content in ash comprising over 60% in some
S1-76 (woody) 82.3 4.24 instances. This has implications for BC determinations
S1-116 (woody) 79.78 1.81
of post-fire residues, in particular for ash. If this mineral
S1-119 (woody) 78.87 1.38
S1-149 (woody) 90.5 2.37 component is not accounted for when estimating BC
S1-125 (grasses) 63.78 9.02 production, then it is highly probable that results will be
S1-176 (grasses) 57.32 20.98 overstated.
S1-130 (grasses) 57.59 24.63 Besides mineral contents, there is also evidence that
S1-140 (grasses) 56.19 24.36
different forms of ash residues vary in their organic C
S1-161 (grasses) 54.64 25.29
Ash content. For example, ‘black’ ash residues are known to
S2-81 12.66 59.45 contain large amounts of organic C. Khanna et al. (1994)
S2-83 12.74 60.18 found that black ash produced during the open burning
S2-85 16.31 46.26 of mixed eucalypt litter had a total C content of 34%.
S2-103 14.9 54.87
S2-104 15.75 53.63
Black ash can also show very high LOI values (Table 4).
S2-105 14.5 56.22 Black ash for several eucalypt forest fuel components
S2-106 17.44 49.27
S2-108 17.92 58.13
S2-109 17.51 50.7 Table 4
S2-110 18.79 50.53 Properties of unburnt fuel components and black and grey ash from
S2-111 21.12 47.97 fire in a sub-alpine Eucalyptus panciflora forest (after Raison et al.,
S2-112 18.23 53.27 1985a,b)
S2-113 12.51 63.8
S2-114 11.13 62.07 Component % Field % LOI (loss on ignition) % Ca % Mg
S2-115 10.54 60.51 Combustion in Laboratory
S1-82 18.34 58.91 Thin Bark
S1-114 34.39 44.64 unburnt 0 99 0.14 0.11
S1-118 18.37 62.06 black ash 78 96 0.62 0.25
S1-171 14.91 64.66 grey ash 99.5 15 25.3 9.3
S1-62 12.49 61.09
S2-116 13.21 62.71 Small Twigs
S2-117 13.75 60.76 unburnt 0 99 0.26 0.07
S1-123 22.25 57.19 black ash 52 97 0.54 0.14
S1-170 26.14 52.18 grey ash 98 55 12.95 2.32
S1-79 43.9 37.05
S1-122 20.84 57.67 Wood
unburnt 0 99 0.06 0.02
Percentage mineral contents were determined by X-ray fluorescence
spectrometry and represent the total major elements expressed as oxides. grey ash 99.7 17 24.3 2.52
196 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

showed very high LOI, and low mineral (Ca, Mg) estimated charcoal production using a similar combina-
concentrations. Grey ash resulting from nearly complete tion of the gravimetric and volumetric methods (Table 5).
combustion in the field, still showed significant LOI Two of these studies (Barbosa and Fearnside, 1996;
(15–55%), but had very high mineral contents. Clearly, Gráça et al., 1999) also estimated the amount of ash
the yield of ash from vegetation fires cannot be regarded produced. Two further studies by Kauffman et al. (1995,
as representing the yield of BC. LOI can also over- 1998) identified and quantified the ash in post-fire
estimate charcoal yields, unless combustion has been residue produced during the burning of cleared grazing
nearly complete, and the post-fire residues contain only pastures. Relatively consistent results were evident for
carbonised organic C. all of these studies with estimates of between 0.7% and
Despite the increased reliability that chemical 2.9% of the carbon in the initial biomass exposed to fire
methods provide over visual and gravimetric methods, (BC/CE) being converted to charcoal and an additional
there is still debate over their accuracy. For example 0.3–2.7% to ash.
Simpson and Hatcher (2004) demonstrated that many However, for the majority of these studies, the post-
documented measurements of BC based on the thermal fire residue was manually collected and the BC
oxidative method have inadvertently overestimated the components were only identified visually. This would
amount of BC. Oxidative measurements identified the have resulted in not only quantities of very fine charcoal
formation of pyrogenic carbon that falls under the on the ground not being collected, but also non-charcoal
operational definition of BC in samples that were components of the post-fire residue possibly being
relatively BC free. identified as charcoal. Hence it is highly likely that
these estimations of charcoal production have a
3. Direct estimations of BC production considerable uncertainty. This is reflected in the error
range quoted in some of these studies, for example
3.1. Studies of BC, charcoal and ash production charcoal production of 4.3 ± 5.9 t ha− 1 by Fearnside et al.
(2001). To be able to compare these estimates with those
About 23 studies focussing on the direct estimation of determined for other biomes, and to be better able to
BC, charcoal and ash quantities produced from a pres- integrate them into broader carbon budgets, the results
cribed burning event have been documented (Table 5). have been re-expressed as the amount of carbon in the
These direct estimations have been restricted to a narrow residue as a percentage of the amount of carbon
group of biomes, including the tropical rainforests of the consumed by fire (BC/CC). The result is a new range
Brazilian Amazon (Fearnside et al., 1993, 1999, 2001; of 2.9–7.8% for charcoal and 0.8–5.3% for ash
Barbosa and Fearnside, 1996, Kauffman et al., 1995, production, both are significantly higher, and cover a
1998; Gráça et al., 1999), the arid and sub-tropical sa- broader range than originally quoted by the authors.
vanna and grassland regions of the southern hemisphere
(Hopkins, 1965; Hurst et al., 1994; Kuhlbusch et al., 3.3. BC estimates for boreal forest burning
1996; Graetz and Skjemstad, 2003), the temperate forests
of Australia (Harwood and Jackson, 1975; Flinn et al., Estimations of BC production for boreal forest burns
1979; Raison et al., 1985a; Graetz and Skjemstad, 2003) have been derived using a wider range of methods than
and the boreal forests of the northern hemisphere (Clark reported for the tropical rainforest studies. In all, six
et al., 1998; Pitkänen et al., 1999; Tinker and Knight, studies have been reported, two each from Siberia (Clark
2000; Ohlson and Tryterud, 2000; Czimczik et al., 2003; et al., 1998; Czimczik et al., 2003), North America
Lynch et al., 2004). Several other studies have been (Tinker and Knight, 2000; Lynch et al., 2004) and
reported, however difficulties were encountered in Scandinavia (Pitkänen et al., 1999; Ohlson and Tryterud,
estimating the actual BC production. In the following 2000). Three of the studies (Clark et al., 1998; Czimczik
section, literature which provides sufficient quantitative et al., 2003; Lynch et al., 2004) applied similar
detail is reviewed and findings compared. methodology to those used in the tropical rainforest
studies, to determine biomass fuel loads. Two of these
3.2. BC estimates for the burning of cleared tropical studies determined the BC residue in the same manner,
rainforest the third (Czimczik et al., 2003) used 13C-NMR spectral
analysis to quantify the aromatic C component of the
Five studies undertaken on prescribed burns in the post-fire residue and thus estimate the amount of BC
Amazon Basin (Fearnside et al., 1993, 1999, 2001; produced. Results for all three suggested that approxi-
Barbosa and Fearnside, 1996; Gráça et al., 1999) mately 1–2% of the total biomass available (BC/TBC)
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 197

Table 5
Estimations of Charcoal (BC) production from prescribed fire events
Vegetation Type Where Fuel Type Method Residue Type % Estimate
in study
Australian Forests
Managed forest fires1 Australia Understory and litter layer Chemical; Si/C based estimates charcoal 6.0B
Unmanaged forest fires1 Australia Understory and litter + canopy Chemical; Si/C based estimates charcoal 10.0B
Clearing fires1 Australia Felled woody vegetation Chemical; Si/C based estimates charcoal 8.0B
Forest Fire19 Australia Logged mixed forest coupe Gravimetric and Volumetric Ash 1.5B
Forest Fire20 Australia Logged Pinus radiata Forest Gravimetric traps Ash 5.9B,F

Tropical Deforestation in Amazon Basin

Original Forest; fall and burn2 Brazil 6 classes of felled rainforest Gravimteric and Volumetric (LIS) charcoal 1.3C
Original Forest; fall and burn3 Brazil trunks, branches, leaves and detritus Gravimteric and Volumetric (LIS) charcoal 2.7C
Original Forest; fall and burn4 Brazil 10 classes of felled rainforest Gravimteric and Volumetric (LIS) charcoal 1.7C
Original Forest; fall and burn4 Brazil 10 classes of felled rainforest Gravimteric and Volumetric (LIS) charcoal 0.9C
Original Forest; fall and burn4 Brazil 10 classes of felled rainforest Gravimteric and Volumetric (LIS) charcoal 1.8C
Original Forest; fall and burn5 Brazil 11 classes of felled rainforest Gravimteric and Volumetric (LIS) charcoal 2.9C
Second burn of orig remains6 Brazil twigs, branches, trunks and palms Gravimteric and Volumetric (LIS) charcoal 0.7C
Secondary Forest (no original)6 Brazil twigs, branches, trunks and palms Gravimetric charcoal 1.2C
Original Forest; fall and burn18 Brazil Woods, rootmat and fine fuels Gravimteric and Volumetric (LIS) Ash 1.1C
Original Forest; fall and burn18 Brazil Woods, rootmat and fine fuels Gravimteric and Volumetric (LIS) Ash 1.3C
Original Forest; fall and burn18 Brazil Woods, rootmat and fine fuels Gravimteric and Volumetric (LIS) Ash 2.1C
Original Forest; fall and burn18 Brazil Woods, rootmat and fine fuels Gravimteric and Volumetric (LIS) Ash 0.8C
Original Forest; fall and burn5 Brazil 12 classes of felled rainforest Gravimteric and Volumetric (LIS) Ash 0.3C
Cleared Grazing pasture6 Brazil All pasture 2–3 cm above surface Gravimetric Ash 0.07C
Cleared Grazing pasture18 Brazil All pasture 2–3 cm above surface Gravimteric and Volumetric (LIS) Ash 0.5C
Cleared Grazing pasture18 Brazil All pasture 2–3 cm above surface Gravimteric and Volumetric (LIS) Ash 1.27C

Boreal Forests
Crown fire∼Boreal forest7 Canada Floor, org duff + canopy Gravimetric (traps) charcoal 2A
Boreal forest8 Siberia Floor material inc CWD <2 cm 13
C-NMR of aromatics BC 0.7C
Boreal forest9 NW Usa Downed CWD >7.5 cm Volumetric (LIS and destructive) charcoal 8A
Temperate and Boreal Forest10 Global Various woods, straws, grasses Lab based grav on H/C ratio BC 2∼6A
Boreal Forest15 Sweden Floor, org duff + canopy Gravimetric (traps) charcoal 0.24 t ha− 1,A

Savanna and Sub-tropical Grassland

Savanna11 Sth Afr Grass, litter and herbs (var dens) Grav (trap and vac) + chem (H/C) BC 0.6∼1.5E
Savanna10 Global Various woods, straws, grasses Lab based grav on H/C ratio BC 1∼2D
Savanna17 Nigeria Herbs and 6 parts of tree cover gravimetric Ash 2.4A
Savanna13 Australia Tropical Grasses gravimetric (traps) charcoal <1.0A
Savanna16 Australia Grass tree and litter gravimetric (traps) Ash 9A

Mediterranean Forests
Mediterranean Forest12 Spain estimation 13
C-NMR & Py-CG/MS techs BC 7.5–3.4 kg
ha− 1 year− 1
Mediterranean Forest12 Spain estimation 13
C-NMR & Py-CG/MS techs BC 1.9–0.7 kg
ha− 1 year− 1
References: 1. Graetz and Skjemstad (2003), 2. Fearnside et al. (1999), 3. Fearnside et al. (1993), 4. Fearnside et al. (2001), 5. Gráça et al. (1999), 6.
Barbosa and Fearnside (1996), 7. Lynch et al. (2004), 8. Czimczik et al. (2003), 9. Tinker and Knight (2000), 10. Kuhlbusch and Crutzen (1996), 11.
Kuhlbusch et al. (1996), 12. González et al. (2002), 13. Carter et al. (1998), 14. Clark et al. (1998), 15. Ohlson and Tryterud (2000), 16. Hurst et al.
(1994), 17. Hopkins (1965), 18. Kauffman et al. (1995), 19. Harwood and Jackson (1975), 20. Flinn et al. (1979), 21. Raison et al. (1985a).
A
Estimate is carbon produced as a percent of the carbon exposed to fire.
B
Estimations of the amount of charcoal carbon produced as a percentage of the carbon consumed.
C
The amount of carbon in charcoal/ash/BC produced as a percentage of the amount of carbon in biomass available.
D
Experimental BC (defined by H/C ratio) formation divided by C volatilized as CO2 (Kuhlbusch and Crutzen, 1996).
E
Total BC from ash and particulate residue (Kuhlbusch et al., 1996).
F
Post-fire residue collected consisted of ash and unburnt materials.
G
Biomass determined prior to burn event was only the litter component.
198 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

for burning was converted to BC. Expressing these data The other three studies were not as easy to compare
as the carbon in BC as a percentage of carbon con- because they measured the production of charcoal
sumed by fire (BC/CC) gave a more variable result of under different circumstances. For example, the Tinker
1.5–3.1%. and Knight (2000) study described earlier, was

Table 6
Revised estimations of BC production as a percent of carbon consumed
Post Fire Residue Characteristics Location % of total C avaliable Estimated carbon Total BCF Yield % of C in BC component of
consumed by fire consumed t ha− 1 t ha− 1 (±se) Residue charcoal or ash as
% of C consumedH
Charcoal Formation in Boreal Forests
Lynch et al. (2004) Canada 53.5 21.4 0.831(0.12) 70% 2.70%
Clark et al. (1998) Siberia 55 16.7 0.74 80% 3.54%
Czimczik et al. (2003) Siberia 48.5 4.2 0.06 n/aC 1.43%
Ohlson and Tryterud (2000) Scandinavia 0.24 n/a
Pitkänen et al. (1999) Scandinavia 1.42 n/a

Charcoal Formation during Tropical Rainforest Clearing

Fearnside et al. (1993) Brazil 24.9 35.9 3.50 75% 7.31%
Barbosa and Fearnside (1996) Brazil 44.8 56.5 2.50 65% 2.88%E
Fearnside et al. (1999) Brazil 42.0 56.0 2.20 75% 2.95%
Gráça et al. (1999) Brazil 38.0 53.3 6.4 (2.7) 65% 7.81%
Fearnside et al. (2001) Brazil 28.3 54.7 4.3 (5.9) 75% 5.86%

Ash Formation during Tropical Rainforest Clearing

Kauffman et al. (1995) Brazil 50.3 71.9 8.8 (6.6) 19% 2.33%
Kauffman et al. (1995) Brazil 51.2 108.3 10.9 (1) 21% 2.11%
Kauffman et al. (1995) Brazil 42.0 57.9 9.4 (1.4) 33% 5.35%
Kauffman et al. (1995) Brazil 55.8 99.3 7.2 (1.3) 20% 1.45%
Gráça et al. (1999) Brazil 38.0 53.3 5.7 (1.0) 7% 0.75%D,G

Ash Formation during burning of Pastures in Amazon Basin

Barbosa and Fearnside (1996) Brazil 42.1 11.6 0.25 (0.19) 16% 0.35%D
Kauffman et al. (1998) Brazil 81.2 11.2 2.05 (0.17) 20% 3.66%
Kauffman et al. (1998) Brazil 56.8 15.7 3.00 (0.28) 20% 3.82%
Kauffman et al. (1998) Brazil 25.4 19.9 4.78 (0.56) 20% 4.80%

Ash Formation during burning of Australian forests

Harwood and Jackson (1975) Australia 55.0 162.1 12.20 20% 1.51%
Flinn et al. (1979) Australia 55.0 155.4 30.0 (4.5) 20% 5.90%
Raison et al. (1985a) Australia 55.0 9.5 4.52 20% 9.50%

BC Formation in Savanna burns

Kuhlbusch et al. (1996)A St Afr 91.2 2.7 0.03 n/aC 1.11%A
Kuhlbusch et al. (1996)B St Afr 78.1 2.1 0.02 n/aC 0.95%B

Ash Formation in Savanna burns

Hopkins (1965) Nigeria 65.0 2.4 0.48 16% 3.20%
Hurst et al. (1994) Australia 100.0 2.6 0.48 (0.004) 20% 3.69%
(This study) Australia n/a n/a n/a 55% 2.50%

Charcoal Formation for Tropical Grass Fires

Carter et al. (1998) Australia 99.5 1.6 0.01 19% 0.12%
Carter et al. (1998) Australia 99.1 1.6 0.03 26% 0.48%
Carter et al. (1998) Australia 99.0 2.0 0.05 18% 0.45%
Carter et al. (1998) Australia 99.3 2.3 0.06 19% 0.49%
Carter et al. (1998) Australia 99.4 3.2 0.05 22% 0.34%

Charcoal Formation for Coarse Woody Debris (>7.5 cm in diameter) in Boreal Forest
Tinker and Knight (2000) nw USA 16.0 52.0 6.40 85% 10.46%

Notes to Table 6
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 199

concerned solely with the production of charcoal from overestimates since the post-fire residue that was
the combustion of coarse woody debris, which was quantitated by Flinn et al. (1979) and Raison et al.
found to be 8% of total biomass available (BC/TBC), (1985a,b) consisted of both ash and unburnt materials.
or 10.5% of carbon consumed (BC/CC). Ohlson and
Tryterud (2000) estimated the production of charcoal 3.5. BC estimations for savanna and grassland burning
from a 35 ha area of boreal forest to be 0.235 t ha− 1,
however the pre-fire biomass was not estimated, The five documented studies on BC production from
therefore a percentage could not be determined. The savanna and grassland fires have been restricted to
same problem occurred in the work reported by southern Africa (Hopkins, 1965; Kuhlbusch et al., 1996)
Pitkänen et al. (1999), who estimated production of and Australia (Hurst et al., 1994; Carter et al., 1998;
1.42 t ha− 1 for a prescribed burn covering 0.2 ha area Graetz and Skjemstad, 2003). For all studies, gravimetric
of a Scandinavian forest. A recent review of the C methods were used to estimate fuel consumption;
cycle in boreal forests by Preston and Schmidt (2006) however, three separate chemical methods were used to
estimates that 1–7% of mass burned is converted to BC determine the amount BC produced during the burn
during burning, with 7–17 Tg BC year− 1 being event. Kuhlbusch et al. (1996) quantified the amount of
produced as solid residues and 2–2.5 Tg BC year− 1 BC produced to be 1% of carbon exposed to fire by
as aerosols globally. utilising the H/C ratio method. Graetz and Skjemstad
(2003) determined the amount of charcoal (BC) residue
3.4. BC estimations for Australian forest burning by evaluating C/Si ratios for burnt grasslands in
Australia and estimated a production rate of 4% of
Three studies that involved the determination of post- carbon consumed, which has subsequently been revised
fire residues resulting from forest burns in Australia are to 2.5% (J. Skjemstad, personal communication). Carter
documented. Harwood and Jackson (1975), while et al. (1998) first determined the amount of fire residue
investigating the loss of elemental nutrients to the produced via gravimetric measurements, then estimated
atmosphere during burning of logging residues in the BC component of the residue by igniting the residue
Tasmania found that the remaining ash residue repre- at 525 °C to determine the loss on ignition (LOI),
sented 1.93% of the total biomass exposed to the fire. suggesting between 0.4% and 1.2% of the total fuel load
Flinn et al. (1979) studied nutrient losses during the was converted to BC. However as previously mentioned
burning of logging slash in a Pinus radiata forest in the LOI method is likely to produce overestimated
southwest Victoria, and found that the combined residue values, as the difference in weight prior to and after
of ash and unburnt material, which remained on site after combustion reflects not only BC but other organic
the fire collectively, represented 16% of the pre-fire fuel carbon components and hydrated minerals as well.
load. Raison et al. (1985a,b) studied nutrient transfers When these estimates are revised to reflect BC as a
during low-intensity prescribed burns in sub-alpine percentage of the total carbon consumed (BC/CC), then
eucalypt forests near Canberra, Australia. In one study BC produced as charcoal for southern African savannas
where litter beds from Eucalyptus panciflora were burnt is approximately 1%, while BC produced as charcoal for
on large metal trays, 74% of the initial fuel was tropical Australian grasslands is between 0.1% and
consumed and the following proportions (% of fuel 0.5%. The amount of BC in ash residue, as a percent of
consumed) of ash were produced: <2 mm (finer mineral total carbon consumed (BC/CC) was 3.7% and 3.2% for
ash, 10.7%); 2–6 mm (containing more unburnt material Australian and African savannas respectively. A new
in the ash, 15.6%); 6–16 mm (larger partly combusted value of 2.5% for BC produced from burning of sub-
fragments, 9.1%). These values must be regarded as tropical grasslands in Northern Australia is presented

Notes to Table 6
A
Kuhlbusch et al. (1996) data for 7 heading fires.
B
Kuhlbusch et al. (1996) data for 11 backing fires.
C
These studies directly estimated the amount of BC present post-fire rather than the total amount of residue.
D
Ash determinations for these studies were in conjunction with charcoal determinations.
E
Estimation of charcoal formation for 4 subsequent burns of same plot spanning 10 years.
F
BC estimates here are stipulated as charcoal, BC or ash.
G
Ash stock has likely been diminished by 8 mm of rain prior to collection, hence its production is underestimated.
H
% expression derived from the formula ((total post-fire residue × % of C in residue)/total C consumed) × 100.
200 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

here for comparison (Table 6); it was derived by 4. Global dynamics of BC: production, turnover
applying the Si/C ratios for data presented in Table 3. and pools
This new data suggests that the production of BC in ash
residues for savanna and grassland burns is likely to be 4.1. Global estimates of BC
in the range of up to 3% (BC/CC).
Table 6 summarises the recalculated BC production Literature dealing with the global BC cycle has
estimations, expressing BC as a percentage of carbon increased in recent years (e.g. Kuhlbusch, 1999;
consumed during the fire. Results for BC as charcoal Schmidt and Noack, 2000; Druffel, 2004; Schmidt,
include 1.4–3.5% for boreal forests, 1.0–1.1% for 2004; Masiello, 2004). The majority of the estimations
African savannas, 0.1–0.5% for Australian tropical of global BC production presented have been modelled
grasslands, 2.9–7.8% for Amazon Basin rainforest on only a few studies, in particular the work un-
clearing and a very high figure of 10.5% for coarse dertaken by Kuhlbusch and Crutzen (1995). There has
woody debris burnt in the boreal forests of North also been an assumption in many of these studies that
America. For some studies of ash formation, the C the degradation of charcoal through microbial or
content of the ash was not directly measured, resulting inorganic reactions over time is negligible. However,
in the inability to calculate the BC content of the this assumption leads to difficulty in balancing the
residues. However, for the studies where C was global BC budget. Using the BC production rates
measured, the range in C content is reasonably tight, proposed by Seiler and Crutzen (1980), Crutzen and
averaging 20%. This average value was therefore Andreae (1990), Kuhlbusch and Crutzen (1995, 1996)
applied to the studies where C data was not available, along with the assumed substantial resistance to
to enable an estimate of BC formation to be made. degradation, Schmidt and Noack (2000) concluded
Results in Table 6 identifying BC produced in ash are that BC should represent a significant proportion of all
2.5–3.7% for semi-arid savannas, <9.0% for Australian the organic carbon buried in soils and sediments. For
forests, 0.8–5.4% for Amazon Basin rainforests and instance, if BC has been produced since the Last Glacial
0.4–4.8% for cattle pastures in the Amazon basin. It Maximum at the same rate as it is currently, then it should
appears that the amount of BC produced as charcoal is account for between 25% and 125% of the total soil
higher for the woody biomes, such as tropical and organic C pool (Masiello and Druffel, 1998). These
boreal forests, compared to grassland and savannas, and estimates are very high when compared with measure-
that ash BC is higher in grasslands and savannas. ments of BC in soils: on average 3–7% of soil organic C
However, many estimations of BC production in (SOC) in Southeast Asia (Rumpel et al., 2006a), <30% in
charcoal, especially those derived using gravimetric Australia (Skjemstad et al., 1996), <45% in Germany
and volumetric methods and in excess of 4–5%, have (Schmidt et al., 1999), and <35% in the U.S. (Skjemstad
significant margins of error. For example, results et al., 2002). Using a different method of BC analysis,
derived for tropical rainforest burns include 4.3 ± 5.9 t Glaser and Amelung (2003) determined that <18% of
ha− 1 (Fearnside et al., 2001) and 6.4 ± 2.7 t ha− 1 (Gráça SOC was BC in native prairie soils of North America.
et al., 1999). In comparison, chemically derived Simpson and Hatcher (2004) identified a substantially
estimations of BC production, which are probably lower BC content of 0.63% for a Floridian peat. Thus,
more accurate, are significantly lower for both savanna assuming that the current global BC production estimates
fires (1%; Kuhlbusch et al., 1996) and boreal forests are reasonable, and that transport of BC from continents is
(1.5%; Czimczik et al., 2003). lower than the rate of production, then a significant
The higher recalculated estimates (> 4%) for the quantity of BC appears to be undergoing degradation in
production of BC in ash also have significant error these environments.
margins resulting from the physical methodology used.
Further, two results for ash production which are much 4.2. The degradation of BC
lower (Gráça et al., 1999; Barbosa and Fearnside, 1996)
than the average (Table 6) were calculated after the The conclusion that BC is chemically and biologi-
charcoal component had been removed. The charcoal cally inert and has long turnover times is based on the
data from these studies are also presented in Table 6. For following evidence: it exhibits a high degree of
other studies, the charcoal component was included in resistance to a range of chemical oxidants (eg Skjemstad
the ash. As such, these lower values are more likely to et al., 1996; Bird and Gröcke, 1997), it has been
reflect the true values of BC in the ash components in preserved for long periods in the geological record
post-fire residues. (Cope and Chaloner, 1980; Herring, 1985) and it exists
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 201

at depth within soil profiles where in many cases very long lived. The most resistant components seem to
residence times exceed 1000 years (e.g. Berger and have an average turnover time, of the order of some
Libby, 1966; Saldarriaga and West, 1986; Glaser et al., thousands of years (Preston and Schmidt, 2006), thus
2001; Schmidt et al., 2002; Krull and Skjemstad, 2002). representing the most stable biomass-derived material
Yet, rather than being completely inert, it is more likely involved in the biospheric C cycle.
that the degree to which BC is resistant to breakdown is In order to fully gauge the influence of BC on the
dependent on the varying extent to which organic global carbon cycle, a sound understanding of the
components in the post-fire residue are altered by production and degradation rates of BC are required.
thermal processes (Baldock and Smernik, 2002). Further, the assessment of fluvial and aeolian transport of
Photochemical and microbial breakdown are the two BC needs to be understood at fine scales. This will allow
primary mechanisms proposed for BC degradation for uncertainties and discrepancies regarding estimates
(Goldberg, 1985), although erosion and translocation of of BC fluxes between the atmosphere, biosphere and
BC within the soil profile have also been suggested as oceans to be minimised. The result will be a more
other possible explanations of BC loss (Czimczik et al., accurate global black carbon budget, and better under-
2003). Potter (1908) first identified the oxidation of standing of the role of BC as a sink in the GCB.
charcoal (BC) by microorganisms. Since then, several
studies have investigated the microbial degradation of BC 4.3. Fluxes and pools of BC
(Shneour, 1966; Winkler, 1985; Scott et al., 1986;
Haumaier and Zech, 1995; Hofrichter et al., 1999; A synthesis of the fluxes and pools of BC in a global
Baldock and Smernik, 2002; Hamer et al., 2004). Shneour context are given in Table 7. Global estimates of the initial
(1966) found that over a 96-day period, 2% of artificial fluxes of BC produced from biomass burning include 40–
graphite was oxidised in non-sterile soils. Winkler (1985) 241 Tg year− 1 to soils (Schmidt and Noack, 2000) and 5–
suggested that organic components of charred remains 6 Tg year− 1 to the atmosphere (Kuhlbusch, 1999).
would break down in an acidic, anaerobic lake environ- Erosion of soil systems results in sediments being
ment. Scott et al. (1986) and Hofrichter et al. (1999) continuously enriched in BC (Rumpel et al., 2006a;
identified several wood-decaying fungi with the ability to Chaplot et al., 2005). This can be attributed to BCs light
decompose low-grade coals, and Baldock and Smernik nature and minor interactions between BC particles and
(2002) suggested that the rate of BC degradation is the soil matrix (Rumpel et al., 2006b). A large proportion
determined by the temperature that existed during its of this eroded BC is deposited near the site of production
formation. These studies provide clear evidence that BC; (Whitlock and Millspaugh, 1996; Schmidt and Noack,
including charcoal and other products of incomplete 2000), and specifically, parts of the landscape that act as
combustion undergo some level of degradation. sinks including hill slopes and valley bottoms (Chaplot et
Given that BC does degrade, whether or not BC al., 2005). Despite this, the long-range transportation of a
contributes to an ongoing carbon sink depends on the portion of the BC occurs through rivers and the
rate of its degradation. Bird and Cali (1998) concluded atmosphere to the oceans (Goldberg, 1985; Suman,
that charred particles are slowly degraded, while Hamer 1988), with the majority doing so via fluvial mechanisms.
et al. (2004) identified at least a small fraction of BC that Several studies have been undertaken to better quantify
decomposed in the short-term. Bird et al. (1999) the secondary BC fluxes from the continents to the ocean
suggested an extreme view that at least in well-aerated sediments via fluvial (Griffen and Goldeberg, 1983;
tropical soil environments, charcoal and oxidation- Suman, 1986; Gustafsson and Gschwend, 1998; Mitra et
resistant elemental carbon (OREC) can be significantly al., 2002; Mannino and Harvey, 2004) and aeolian (Ogren
degraded on decadal to centennial timescales. The and Charlson, 1983; Suman, 1988; Penner et al., 1993)
estimated rate of charcoal decomposition in soils transport.
examined in a boreal forest was significant at 5–9 kg Studies in Central Panama have shown that only 5%
C ha− 1 year− 1 (Harden et al., 2000). BC can be found at of charcoal (BC) produced by agricultural burning
depth in numerous soil profiles, providing 14C ages in entered the Gulf of Panama, and 97% of that was via
excess of a thousand years (e.g. Berger and Libby, 1966; fluvial transport (Suman, 1988). Mannino and Harvey
Saldarriaga and West, 1986; Glaser et al., 2001; Schmidt (2004) also suggested that river–estuary systems are
et al., 2002; Krull and Skjemstad, 2002), and it also important exporters of colloidal BC to the ocean, and
occurs in deep marine sediments (Verado, 1997; Mitra et al. (2002) suggested that the Mississippi River
Masiello and Druffel, 1998; Middleburg et al., 1999). alone discharged 5% of all the BC that is buried annually
This suggests that at least some components of BC are in the ocean. These findings add considerable weight to
202 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

Table 7
Estimations of the fluxes and pools within the global black carbon cycle
FORMATION OF BC (This study)
Regional Vegetation Fires (BC/CO2 released) % Total BC (Tg year− 1)C (BC/CC)%A
Amazon rainforest biomass burning1 1.0–4.1 5 ∼ 3.0
Savanna biomass burning2 0.6–2.0 4–40 <3.0
Northern Hemisphere boreal forest burning3 5.0–7.0 9–20 ∼ 3.0
Australian forest fires4 4.0–6.0 0.08–0.8 ∼ 3.0
Other Anthropogenic Sources (BC/CE)% Total BC (Tg year−1)C
Shifting agriculture2 1.5–3.0 15–60B
Permanent deforestation3 1.5–3.0 7–54B
Agricultural waste2 1.0–2.0 5–16
Fuel wood2 2.5–4.5 8–21
Fossil fuel combustion5 ? 12–24
Global Production (BC/CO2 released year−1)2 50–270
Global Production (BC/CEbiomass year−1)2 40–179
FLUXES OF BC
Initial Fluxes % BC produced Total BC (Tg year−1)C
Flux to atmosphere6 <20 6–28
from fossil fuels7 0.2–0.4 6–9
from biomass8 2.0–20.0 5–6
From biomass >80 40–241
burning to soils2
Secondary Fluxes (BC as % of OC transfer) Total BC (Tg year−1)C
In river OC to ocean (global)9 10 <20
In river OC to ocean ( NE USA)10 4–7D ?
In river OC to ocean ( Panama)11 5 ?
Atmospheric transport to ocean (Panama)11 3 1.2
BC to coastal oceans8 3–12 1.4–5.4
BC to open oceans8 1–10 2.4–4.8
BC from oceans to Marine sediments11 5–20 10
BC in oceanic POC to sediment surface11 10 10
STOCKS AND TURNOVER OF BC BC/TOC (%) Pg of BCC Residence Time (ky)
Ocean DOC pool12 4–22 ? <14
Marine sediments8 2–21 400–1000 10–20
Coastal sediments8 5–38 2000–5000 10–20
Soils13,14,15 10–40 ? 0.5–5+
Polar and Greenland sea ice9 (0.1–2.1 μg kg− 1) 10–20
CE is carbon exposed to fire, CC is carbon consumed by fire.
References: 1. Fearnside (2000), 2. Kuhlbusch (1999), 3. Preston and Schmidt (2006), 4. Graetz and Skjemstad (2003), 5. Penner et al. (1993), 6.
Crutzen and Andreae (1990), 7. Warneck (1988), 8. Schmidt and Noack (2000), 9. Druffel (2004), 10. Mannino and Harvey (2004), 11. Suman
(1986), 12. Masiello and Druffel (1998), 13. Skjemstad et al. (1996), 14. Glaser and Amelung (2003), 15. Schmidt et al. (1999).
A
Reworked estimates expressing BC as percentage of carbon consumed (this study).
B
These values incorporate BC estimations for various regional vegetation fire particularly Amazon forest burning.
C
Tg = teragrams = 1012 g; Pg = petagrams = 1015 g.
D
Includes fossil fuel component.

recent model calculations, which suggest that more than Despite the fact that the majority of BC transport from
90% of BC deposition in the oceans occurs on the terrestrial environments to the ocean is via fluvial
continental shelves, and that the magnitude of this mechanisms and that the bulk of the BC is deposited
transfer to shelf sediments is similar to fluvial transport on the continental shelf margins, the transfer of BC to
estimates of continental BC (Suman, 1997). Schmidt and deep ocean sediments via aeolian processes is also
Noack (2000) estimated the flux of BC to coastal oceans important. Ogren and Charlson (1983) calculated that
is only 1.4–5.4 Tg year− 1 (< 10% of total annual much of the BC that does enter the atmosphere could stay
formation, Table 7) and that this coastal sedimentary there for a period of between 40 h in wet climates to
BC pool is 2000–5000 Pg (Petragrams = 1 × 1015 g). > 7 days in clean, dry regions. As a result, some
Clearly, most of the BC remains in terrestrial systems submicrometer-sized particles remain airborne and
where it either accumulates in soils and sediments, or become part of the background planetary aerosol,
undergoes degradation processes. eventually reaching the ocean surface and finally the
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 203

ocean sea floor (Verardo and Ruddiman, 1996; Masiello global C cycle based on the data reviewed and presented
and Druffel, 1998). BC fluxes to the open ocean are in this paper.
estimated to be 2.4–4.8 Tg yr − 1 , a value that
incorporates both aeolian input as well as contributions 5. Conclusions
from fluvial sources that do not end up on the continental
shelves. The rate at which BC enters marine sediments is Components of Black Carbon (BC), in particular ash
estimated to be 10 Tg yr− 1 (Suman, 1997) contributing and charcoal, represent a significant amount of post-fire
to a BC pool of between 400 and 1000 Pg. The BC pool residues and are produced by the incomplete combustion
in marine sediments is significant and is considered to of vegetation or fossil fuels (Goldberg, 1985). These forms
comprise 12–31% of sedimentary organic carbon (SOC) of BC occur ubiquitously in soils and other terrestrial
in the deep ocean, and to have a residence time of sediments, and via fluvial and atmospheric transport, in
thousands of years (Masiello and Druffel, 1998). The coastal and deep marine sediments as well. Various
fact that such considerable amounts of BC exist in deep methods have been applied in the quantification of BC
ocean sediments even though it only receives a small production from biomass burning. Gravimetric and
percentage of what is produced annually, adds further volumetric determinations are the two main methods that
weight to the argument that BC in the oceans can be are used to estimate BC production. The gravimetric
considered inert and thus a sink for atmospheric CO2. methods are better suited to finer biomass materials such
Fig. 1 presents an updated schematic representation of as grasses, while the volumetric methods is more appro-
the estimated fluxes and pools of BC that constitute the priate for coarser woody material with diameters usually

Fig. 1. Schematic representation of the Global Black Carbon Cycle (GBCC). Modified and updated from Kuhlbusch and Crutzen (1995) and Schmidt
and Noack (2000). (A) BC/CO2 ; (B) BC/CE ; (C) BC/CC ; (D) percentage estimates in sediments and oceans are BC/TC. ? indicates no data currently
available.
204 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

greater than 10 cm. However, both methods are sus- savanna and grassland fires a value of <3% is consistent
ceptible to error through heterogeneities in biomass and with reported data, however considerable variation
the lack of accurate methods for the collection and identi- among these estimates remains.
fication of the post-fire residues. Chemical techniques A large portion (>80%) of the BC that is produced by a
such as 13C-NMR spectroscopy, chemical/thermal oxida- fire event remains proximal to the site where it was
tion treatment and the evaluation of H/C, Si/C and Ca/C formed. It is then incorporated into the soil where it can
ratios are helpful in identifying and quantifying the remain for long periods of time. However, BC can also be
amount of BC present in post-fire residues. These types transported via fluvial and atmospheric pathways to
of methods are particularly beneficial in attempting to marine sediments, with the majority moving through the
overcome the inaccuracies associated with physical fluvial system. This results in most of the particulate BC
collection and identification techniques. transported to the oceans being deposited on the coastal
Documented research regarding the production of shelves, while a smaller portion continues on to the deeper
BC from biomass burning has been restricted to ocean sediments. Another fraction of the particulate BC
Amazonian rainforests, Australian temperate forests, produced is dispersed into the atmosphere. With residence
semi-arid grasslands and savannas of the southern times that can exceed 7 days, much of this component of
hemisphere and the boreal forests of the northern hemi- BC is transported to the oceans and ultimately contributes
sphere. In most instances, the post-fire material in the to the BC fraction of deep ocean sediments, where it is
form of charcoal and/or ash was estimated, while BC very stable.
was directly quantified on only a few occasions. The Despite the evidence that photochemical and micro-
results from these studies have been expressed in bial processes can degrade BC, research such as in-
various forms making comparisons between them im- cubation studies has shown that BC degradation through
possible. Hence in this review, we expressed rates of biological factors is very slow. Other characteristics of
BC production as a percentage of the carbon consumed BC also support slow degradation; BC can comprise up
during the burn event (BC/CC). This revealed that to 40% of the OC in terrestrial soils and between 12%
production rates of BC as charcoal were 1.4–3.5% in and 31% of the OC in deep ocean sediments, and has
boreal forest fires, 0.9–1.1% for African savannas, 0.1– radiocarbon ages in soils in excess of thousands of
0.5% for Australian tropical grasslands and 2.9–7.8% years. Hence, BC appears to have a significant half-life,
for Amazonian rainforest burns. Production rates of BC in the order of thousands of years. This relative inertness
as ash residues are 2.5–3.7% of carbon consumed by means that the projected < 3% of the carbon converted to
fire (BC/CC) for savannas and grasslands, 0.3–4.8% for BC during forest, savanna and grassland fires, must be
cattle pastures in the Amazon Basin, 0.75–5.3% for considered a significant component of the global carbon
Amazon rainforest burns and 1.5–9.5% for Australian cycle with a very slow turnover.
forests. It is apparent from these data that more BC
occurs as ash residues in grass and savanna fires, and Acknowledgements
more in charcoal after boreal and tropical forest fires.
Further examination of these production rates, We thank the Co-operative Research Centre (CRC)
including the identification of high error terms has for Greenhouse Accounting for providing funding for
allowed for the removal of several over-estimates and this review and to Dr. D. Graetz (CSIRO Marine and
for a new set of lower average BC production rates to be Atmospheric Research) for supplying many of the
proposed. BC production as charcoal for boreal, Aus- Australian charcoal and ash samples. Thanks also due to
tralian and tropical forest fires is generally less than 3% the staff of the CSIRO Library, Waite, South Australia
of the amount of carbon consumed in the burn, while for their help in obtaining the relevant literature.
BC as ash is most likely lower than 2%. The opposite is
evident for grass and savanna fires with BC as charcoal References
contributing less than 1%, while the BC in ash is in a
range of 2–3%. Overall, BC production is estimated to Baldock JA, Smernik RJ. Chemical composition and bioavailability of
be less than 4–5% of the carbon that is consumed during thermally altered Pinus resinosa (Red pine) wood. Org Geochem
fire events, with fires in woody fuels producing more 2002;33:1093–109.
BC residue (<5%) than savannas and grassland fires Barbosa RI, Fearnside PM. Pasture burning in Amazonia: Dynamics of
residual biomass and the storage and release of aboveground
(<3%). Analysis of post-fire residues using chemical carbon. J Geophys Res 1996;101:25,847.
methods, considered more accurate, results in estimates Berger R, Libby WF. UCLA Radiocarbon Dates V. Radiocarbon
of BC production of <3% BC/CC for forest fires. For 1966;8:467–97.
 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206 205

Bird MI, Cali JA. A million-year record of fire in sub-Saharan Africa. Gráça PMLA, Fearnside PM, Cerri CC. Burning of Amazonian forest
Nature 1998;394:767–9. in Ariquemes, Rondonia, Brazil: biomass, charcoal formation and
Bird MI, Gröcke DR. Determination of the abundance and carbon burning efficiency. For Ecol Manag 1999;120:179–91.
isotope composition of elemental carbon in sediments. Geochim Graetz RD, Skjemstad JO. The charcoal sink of biomass burning on
Cosmochim Acta 1997;61:3413–23. the Australian continent. CSIRO Atmospheric Research Technical
Bird MI, Moyo C, Veenendaal EM, Lloyd J, Frost P. Stability of elemental Paper, vol. 64; 2003.
carbon in a savanna soil. Glob Biogeochem Cycles 1999;13:923–32. Griffen JJ, Goldeberg ED. The fluxes of elemental carbon in coastal
Carter JO, Howden SM, Day KA, McKeon GM. Soil carbon, nitrogen marine sediments. Limnol Oceanogr 1983;20:456–63.
and phosphorous and biodiversity in relation to climate change. Gustafsson O, Gschwend PM. The flux of black carbon to surface
Evaluation of the impact of Climate Change on Northern sediments on the New England continental shelf. Geochim
Australian Grazing Industries, vol. 139A. Final report for the Cosmochim Acta 1998;62(3):465–72.
Rural Industries Research and Development Corporation on Hamer U, Marschner B, Brodowski S, Amelung W. Interactive
Project DAQ; 1998. p. 204–5. priming of black carbon and glucose mineralisation. Org Geochem
Chaplot VAM, Rumpel C, Valentin C. Water erosion impact on soil 2004;35:823–30.
and carbon redistributions within uplands of Mekong River. Glob Harden JW, Trumbore SE, Stocks BJ, Hirsch A, Gower ST, O'Neill
Biogeochem Cycles 2005;19(4):20–32. KP, et al. The role of fire in the boreal carbon budget. Glob Change
Clark JS, Lynch J, Stocks BJ, Goldammer JG. Relationships between Biol 2000;6:174–84.
charcoal particles in air and sediments in west-central Siberia. Harwood CE, Jackson WD. Atmospheric losses of four plant nutrients
Holocene 1998;8:19–29. during a forest fire. Aust For 1975;38(2):92–9.
Cope MJ, Chaloner WG. Fossil charcoal as evidence of past Haumaier L, Zech W. Black carbon-possible source of highly aromatic
atmospheric composition. Nature 1980;283:647–9. components of soil humic acids. Org Geochem 1995;23:191–6.
Crutzen PJ, Andreae MO. Biomass burning in the tropics: impact on Herring JR. Charcoal fluxes into sediments of the North Pacific Ocean:
atmospheric chemistry and biogeochemical cycles. Science the Cenozoic record of burning. In: Sundquist ET, Broecker WS,
1990;250:1669. editors. The Carbon Cycle and Atmosphere CO2: Natural Variations
Czimczik CI, Preston CM, Schmidt MWI, Schulze E. How surface fire Archean to Present. Washington, DC: GU; 1985. p. 419–42.
in Siberian Scots pine forests affects soil organic carbon in the forest Hofrichter M, Ziegenhagen D, Sorge S, Ullrich R, Bublitz F, Fritsche
floor: Stocks, molecular structure, and conversion to black carbon W. Degradation of lignite (low ranked-coal) by lignolytic
(charcoal). Glob Biogeochem Cycles 2003;17(1): 20(1)–20(14). basidiomycetes and their peroxidase system. Appl Microbiol
Druffel ERM. Comments on Black Carbon in the global carbon cycle. Biotechnol 1999;52:78–84.
Mar Chem 2004;92:197–200. Hopkins B. Observations on savanna burning in the Olokemeji Forest
Fearnside PM. Global warming and tropical land-use change: Reserve, Nigeria. J Appl Ecol 1965;2:367–81.
greenhouse gas emissions from biomass burning, decomposition Hurst DF, Griffith DWT, Cook GD. Trace gas emissions from biomass
and soils in forest conversion, shifting cultivation and secondary burning in tropical Australian savannas. J Geophys Res (Atmos)
vegetation. Clim Change 2000;46:115–58. 1994;99:16,441­56.
Fearnside PM, Leal Filho N, Fernandes FM. Rainforest burning and the Jones TP, Chaloner WG, Kuhlbusch TAJ. Proposed Bio-geological
global carbon budget: biomass combustion efficiency and charcoal and Chemical Based Terminology for Fire-altered Plant Matter. In:
formation in the Brazilian Amazon. J Geophys Res (Atmos) 1993;98 Clark JS, Cachier H, Goldammer JG, Stocks BJ, editors. Sediment
(D9):16733–43. Records of Biomass Burning and Global Change, vol. 1. Berlin:
Fearnside PM, Gráça PMLA, Filho NL, Rodrigues FJA, Robinson JM. Springer-Verlag; 1997. p. 9–22.
Tropical forest burning in Brazilian Amazonia: measurement of Kauffman J, Cummings D, Ward D, Rabbit R. Fire in the Brazilian
biomass loading, burning efficiency and charcoal formation at Amazon: 1. Biomass, nutrient pools, and losses in slashed primary
Altamira, Para. For Ecol Manag 1999;123:65–79. forests. Oecologia 1995;104:397–408.
Fearnside PM, Gráça PMLA, Rodrigues FJA. Burning of Amazonian Kauffman J, Cumming D, Ward D. Fire in the Brazilian Amazon: 2.
rainforests: burning efficiency and charcoal formation in forest Biomass, nutrient pools, and losses in cattle pastures. Oecologia
cleared for cattle pasture near Manaus, Brazil. For Ecol Manag 1998;113:415–27.
2001;146:115–28. Khanna PK, Raison RJ, Falkiner RA. Chemical properties of ash
Flinn DW, Hopmans P, Farrell PW, James JM. Nutrient loss from burning derived from Eucalyptus litter and its effects on forest soils. For
of Pinus radiata logging residue. Aust For Res 1979;9:17–23. Ecol Manag 1994;66:107–25.
Glaser B, Amelung W. Pyrogenic carbon in native grassland soils Krull ES, Skjemstad JO. δ13C and δ15N profiles in 14C-dated Oxisol
along a climosequence in North America. Glob Biogeochem and Vertisols as a function of soil chemistry and mineralogy.
Cycles 2003;17(2):1064. Geoderma 2002;1890:1–29.
Glaser B, Haumaier L, Guggenberger G, Zech W. The ‘Terra Preta’ Kuhlbusch TAJ. Method for Determining Black Carbon in Residues of
phenomenon: a model for sustainable agriculture in the humid Vegetation Fires. Environ Sci Technol 1995;29(10):2695–702.
tropics. Naturwissenschaften 2001;88:37–41. Kuhlbusch TAJ. Black carbon and the Carbon Cycle. Science
Goldberg ED. Black carbon in the Environment: Properties and 1998;280:1903–4.
Distribution. New York: John Wiley and Sons; 1985. 198 pp. Kuhlbusch TAJ. Black carbon and the Global Carbon and Oxygen
González JA, González-Vila FJ, Polvillo O, Almendros G, Knicker H, Cycle. Abstracts of Ninth Annual Goldschmidt conference.
Salas F, et al. Wildfire and black carbon in Andalusian Boston, MA: Harvard University; 1999.
Mediterranean forest. In: Viegas DX, editor. Forest fire research Kuhlbusch TAJ, Crutzen PJ. Toward a global estimate of black
and wildland fire safety: Proceedings of IV International carbon in residues of vegetation fires representing a sink of
Conference on Forest Fire Research. Wildland Fire Safety Summit, atmospheric CO2 and a source of O2. Glob Geochem Cycles
Luso, Coimbra, Portugal, 18–23 November; 2002. 198 pp. 1995;9(4):491–501.
206 M.S. Forbes et al. / Science of the Total Environment 370 (2006) 190–206

Kuhlbusch TAJ, Crutzen PJ. Black Carbon, the Global carbon Cycle, tropical sloping land under slash and burn agriculture. Geoderma
and Atmospheric Carbon Dioxide. Chapter 16. In: Levine JS, 2006a;130:35–46.
editor. Biomass Burning and global change, vol. 1. Cambridge Rumpel C, Chaplot V, Planchon O, Bernadoux J, Valentin C, Mariotti
Massachusetts: The MIT Press; 1996. p. 160–9. A. Preferential erosion of black carbon on steep slopes with slash
Kuhlbusch TAJ, Andreae MO, Cachier H, Goldammer JG, Lacaux JP, and burn agriculture. Catena 2006b;65(1):30–40.
Shea R, et al. Black carbon formation by savanna fires: Saldarriaga JG, West DC. Holocene fires in the northern Amazon
Measurements and implications on the global carbon cycle: basin. Quat Res 1986;26:358–66.
SAFARI 92. J Geophys Res 1996;101:23,651. Schmidt MWI. Carbon budget in the black. Nature 2004;427:305–6.
Lynch JA, Clark JS, Stocks BJ. Charcoal production, dispersal and Schmidt MWI, Noack AG. Black carbon in soils and sediments:
deposition from the Fort Providence experimental fire: interpreting Analysis, distribution, implications, and current challenges. Glob
fire regimes from charcoal records in boreal forests. Can J For Res Biogeochem Cycles 2000;14(3):777–93.
2004;34:1642–56. Schmidt MWI, Skjemstad JO, Gehrt E, Kögel-Knaber I. Charred
Mannino A, Harvey HR. Black carbon in estuarine and coastal organic carbon in German chernozemic soils. Eur J Soil Sci
dissolved organic matter. Limnol Oceanogr 2004;49(3):735–40. 1999;50:351–65.
Masiello CA. New directions in black carbon organic chemistry. Mar Schmidt MWI, Skjemstad JO, Jäger C. Carbon isotope geochemistry
Chem 2004;92:201–13. and nanomorphology of soil black carbon: Black chernozemic
Masiello CA, Druffel ERM. Black carbon in Deep-Sea Sediments. soils in central Europe originate from ancient biomass burning.
Science 1998;280:1911–3. Glob Biogeochem Cycles 2002;16(4):71–7.
Middleburg JJ, Nieuwenhuize J, Van Breugel P. Black Carbon in Scott CD, Standberg GW, Lewis SN. Microbial solubilization of coal.
marine sediments. Mar Chem 1999;65:245–52. Biotechnol Prog 1986;2:131–9.
Mitra S, Bianchi TS, McKee BA, Sutula M. Black Carbon from the Seiler W, Crutzen PJ. Estimates of gross and net fluxes of carbon
Mississippi River: Quantities, Sources, and Potential Implications between the biosphere and the atmosphere from biomass burning.
for the Global Carbon Cycle. Environ Sci Technol 2002;36: Clim Change 1980;2:207–47.
2296–302. Shneour EA. Oxidation of graphic carbon in certain soils. Science
Novakov T, Cachier H, Clark JS, Gaudichet A, Masclet P. Characterization 1966;151:991–2.
of particulate products of biomass combustion. In: Clark JS, Cachier Simpson MJ, Hatcher PG. Overestimates of black carbon in soils and
H, Goldammer JG, Stocks B, editors. Sediment records of biomass sediments. Naturwissenschaften 2004;91:436–40.
burning and global change. NATO ASI Series. Series I. Global Skjemstad JO, Clarke P, Taylor JA, Oades JM, McClure SG. The
Environmental Change, Berlin, Germany: Springer; 1997. p. 117–43. chemistry and nature of protected carbon in soil. Aust J Soil Res
Ogren JA, Charlson RJ. Elemental carbon in the atmosphere: cycle and 1996;34:251–71.
lifetime. Tellus 1983;35B:241–54. Skjemstad JO, Taylor JA, Smernik RJ. Estimation of charcoal (char) in
Ohlson M, Tryterud E. Interpretation of the charcoal record in forest soils. Commun Soil Sci Plant Anal 1999;30:2283–98.
soils: forest fires and their production and deposition of Skjemstad JO, Reicosky DC, Wilts AR, McGowan JA. Charcoal carbon
macroscopic charcoal. Holocene 2000;10(4):519–25. in U.S. agricultural soils. Soil Sci Soc Am J 2002;66:1249–55.
Penner JE, Eddleman H, Novakav T. Toward the development of a global Suman DO. Charcoal production from agricultural burning in central
inventory for black carbon. Atmos Environ 1993;27A(8):1277–95. Panama and its deposition in the sediments of the Gulf of Panama.
Pitkänen A, Lehtonen H, Huttunen P. Comparison of sedimentary Environ Conserv 1986;13:51–60.
microscopic charcoal particle records in a small lake with Suman DO. The flux of charcoal to the troposphere during the period
dendrochronological data: evidence for the local origin of micro- of agricultural burning in Panama. J Atmos Chem 1988;6:21–34.
scopic charcoal produced by forest fires of low intensity in eastern Suman DO. Marine sediments: a reservoir for black carbon and their use
Finland. Holocene 1999;9:559–67. as spatial and temporal records of combustion. In: Clark JS, Cachier
Potter MC. Bacteria as agents in the oxidation of amorphous carbon. H, Goldammer JG, Stocks B, editors. Sediment records of biomass
Proc R Soc Lond 1908;80B:239–59. burning and global change. NATO ASI Series. Series I. Global
Preston CM, Schmidt MWI. Black (pyrogenic) carbon in boreal Environmental Change, Germany: Springer Berlin; 1997. p. 169–88.
forests: a synthesis of current knowledge and uncertainties. Tinker DB, Knight DH. Coarse woody debris following fire and
Biogeosci Disc 2006;3:211–71. logging in Wyoming lodgepole pine forest. Ecosystem
Raison RJ, McGarity JW. Changes in carbon and nitrogen fractions of 2000;3:472–83.
graminaceous straws after burning. Plant Soil 1979;51:311–8. Verado DJ. Charcoal analysis in marine sediments. Limnol Oceanogr
Raison RJ, Khanna PK, Woods PV. Mechanisms of element transfer to the 1997;42:192–7.
atmosphere during vegetation fires. Can J For Res 1985a;15:132–40. Verardo DJ, Ruddiman WF. Late Pleistocene charcoal in tropical
Raison RJ, Khanna PK, Woods PV. Transfer of elements to the Atlantic deep-sea sediments: climatic and geochemical signifi-
atmosphere during low-intensity prescribed burns in three Australian cance. Geology 1996;24:855–7.
sub-alpine eucalypt forests. Can J For Res 1985b;15:657–64. Warneck P. Chemistry of the natural atmosphere. San Diego:
Rumpel C, Alexis M, Chabbi A, Chaplot V, Rasse DP, Valentin C, et al. Academic; 1998. 614 pp.
Black carbon contribution to soil organic matter composition in