Replacing time with space: using laboratory fires to explore the effects of repeated burning on black carbon degradation

Replacing time with space: using laboratory fires to explore

the effects of repeated burning on black carbon degradation

Wade T. Tinkham

, Alistair M. S. Smith

, Philip E. Higuera

Jeffery A. Hatten

, Nolan W. Brewer

and Stefan H. Doerr

Warner College of Natural Resources, Colorado State University, Fort Collins, CO 80523, USA.

Idaho Fire Initiative for Research and Education (IFIRE), College of Natural Resources,

University of Idaho, Moscow, ID 83844, USA.

Department of Ecosystem and Conservation Sciences, University of Montana. Missoula,

MT 59812, USA.

College of Forestry, Oregon State University, Corvallis, OR 97331, USA.

Washington Department of Natural Resources, Olympia, WA 98504, USA.

Department of Geography, College of Science, Swansea University, Singleton Park,

Swansea SA2 8PP, UK.

Corresponding author. Email: [email protected]

Abstract. Soil organic matter plays a key role in the global carbon cycle, representing three to four times the total carbon

stored in plant or atmospheric pools. Although fires convert a portion of the faster cycling organic matter to slower cycling

black carbon (BC), abiotic and biotic degradation processes can significantly shorten BC residence times. Repeated fires

may also reduce residence times, but this mechanism has received less attention. Here we show that BC exposed to

repeated experimental burns is exponentially reduced through four subsequent fires, by 37.0, 82.5, 98.6 and 99.0% of BC

mass. Repeated burning can thus be a significant BC loss mechanism, particularly in ecosystems where fire return rates are

high, relative to BC soil incorporation rates. We further consider loss rates in the context of simulated BC budgets, where

0–100% of BC is protected from subsequent fires, implicitly representing ecosystems with varying fire regimes and BC

transport and incorporation rates. After five burns, net BC storage was reduced by as much as 68% by accounting for

degradation from repeated burning. These results illustrate the importance of accounting for BC loss from repeated

burning, further highlighting the potential conflict between managing forests for increasing soil carbon storage vs

maintaining historic fire regimes.

Additional keywords: carbon storage, CTO-375, ecosystems, fire regimes, soil incorporation.

Received 22

July 2015, accepted 2 November 2015,

published online 25 January 2016

Introduction

Biomass burning converts a portion of the faster cycling

organic matter to slower cycling black carbon (BC) (Lehmann

et al. 200 8), which has been shown to be resistant to biotic

degradation due to changes in its chemical structure (Schmidt

and Noack 2000; Masiello 2004). For the boreal biome, for

example, it has been estimated that .25% of carbon affected by

fire is converted into a slower cycling form (Santı´n et al.

2015a). Increased fire activity is predicted for many ecosys-

tems in response to climate change and will likely lead to

significant changes in the global carbon c ycle (Chapin et al.

2000; Westerling et al. 2006; Moritz et al. 2012; Flannigan

et al. 2013; IPCC 2013). Given the potential for long-term

carbon storage, this process has received significant attention

in recent decades (Goldberg 1985; Preston and Schmidt 2006),

especially given the growing interest in BC’s contribution to

the soil organic carbon (SOC) pool, which can exceed 50% in

some ecosystems (Le hmann et al. 2008; Schmidt et al. 2011;

Cheng et al. 2013).

Radiocarbon dating of BC in soils has shown residence times

in the order of hundreds to thousands of years (Ohlson et al.

2009; Egli et al. 2012). This variation is commonly attributed to

abiotic and biotic loss processes, which can significantly shorten

BC residence times (Czimczik and Masiello 2007; Zimmerman

2010). Repeated biomass burning has been postulated to also

reduce residence times of BC (Ohlson and Tryterud 2000;

Rovira et al. 2009), but this BC loss mechanism has received

less attention than most biotic mechanisms (Santı´n et al. 2013).

However, in forested ecosystems that have a high propensity to

produce BC through combustion of woody fuels (Hurteau and

Brooks 2011), the typical time between fires is in the order of

decades to centuries. This temporal scale limits the ability to

observe and quantify this process in a field setting, thus

emphasising the need for an experimental approach.

CSIRO PUBLISHING

International Journal of Wildland Fire

http://dx.doi.org/10.1071/WF15131

Journal compilation Ó IAWF 2016 www.publish.csiro.au/journals/ijwf

Research Note

Black carbon cycling

Fire-altered carbon from biomass burning exists on a recalci-

trance gradient, from partially charred vegetation to very recal-

citrant soot (Masiello 2004; Smith and Hudak 2005; Preston and

Schmidt 2006; Keiluweit et al. 2010). Past research into BC has

been in the fields of soil science, biogeochemistry and fire sci-

ence, which use differing terminology including char, charcoal,

soot and graphite that can cause confusion. As described in

Preston and Schmidt (2006), the term ‘pyrogenically altered

carbon’ is used to describe the entire spectrum of carbon that has

been thermally altered by fire, whereas BC often refers only to a

narrow range of pyrogenic carbon produced through thermal

exposure. Specifically, BC is defined here as the highly recalci-

trant material that exhibits significantly lowered H:C and O:C

ratios as compared to charcoal and charred biomass (Kuhlbusch

et al. 1996; Hatten and Zabowski 2009; Brewer et al. 2013).

Pyrogenic organic matter is another widely used term that

describes BC (Santı´n et al. 2015a). Confusion may arise from use

of the term ‘pyrogenic carbon’, which is used to describe BC

(Santı´n et al. 2013) due to its closeness to the Preston and Schmidt

(2006) terminology. In this study, we do not advocate the usage of

a specific set of terminology, but utilise the terminology associ-

ated with the analytical BC calculation method used in this paper

(Hatten and Zabowski 2009).

The coupling of BC production, soil incorporation, off-site

transport and BC loss via combustion creates a complex rela-

tionship between the frequency of burning in an ecosystem and

total BC storage. The molecular composition alone cannot

singularly predict carbon’s persistence in soils; rather, ecosys-

tem processes including fire intensity and frequency, soil

pedoturbation, and climatic fluctuations interact to play impor-

tant roles in its longevity (Czimczik and Masiello 2007; Jenkins

et al. 2014). Specifically, the physical protection of BC, for

example via soil incorporation or off-site transport, is required

for it to be preserved over hundreds to thousands of years

(DeLuca and Aplet 2008; Fang et al. 2014). However, the

literature describes a complex picture (Singh et al. 2014), in

part because of the aforementioned challenges associated with

different definitions of what constitutes BC. Some studies have

indicated that due to soil stabilisation processes, BC levels can

persist over periods of ,50 years in fallow soils (Vasilyeva et al.

2011). Other studies have indicated rapid losses over decadal

time scales (Nguyen et al. 2008).

With each new fire, BC is generated, but the existing BC

from prior fires may also be consumed if it remains near the soil

surface and exposed to the fire (Ohlson and Tyrterud 2000;

Czimczik et al. 2003;

Rovira et al. 2009). Two recent field

studies measured consumption of pre-existing BC (charcoal) by

fire in contrasting environments. Santı´n et al. (2013) found

median mass losses of ,15% of BC samples placed within the

surface of the organic layer consumed in a boreal forest fire;

although the organic layer was consumed – exposing the

charcoal – the samples survived. In contrast, when considering

burying charcoal within the mineral layer Schmidt and Noack

(2000) suggested that thermal degradation is not usually

achieved at mineral soil depths .30 mm. Although the impor-

tance of soil insulation to prevent thermal degradation is clear,

there is a lack of research quantifying soil incorporation rates of

pyrogenic carbon (Nocentini et al. 2010). Preliminary projec-

tions suggest that it takes decades to centuries for fire residues to

be adequately incorporated into a soil matrix to ensure protec-

tion from fires (Eckmeier et al. 2007; Lehmann et al. 2008),

considerably longer than fire return intervals in most dry

temperate forests (Littell et al. 2009 ). Given fire frequencies

in many ecosystems are expected to increase due to climatic

warming (Flannigan et al. 2013), it is important to address how

repeated burning may influence BC pools (Czimczik and

Masiello 2007; Schmidt et al. 2011; Santı´n et al. 2015b). Saiz

et al. (2014) evaluated fire as a BC consumption mechanism in

Australian savannah woodlands, finding mass losses of ,8% of

BC when particles were placed on the soil surface in a prescribed

fire. The low mass loss rates reported in the literature are

attributed to a combination of soil insulation and fuels properties

that could limit consumption (Santı´n et al. 2013; Saiz et al.

2014). We present results from a controlled laboratory experi-

ment as a case study that quantifies the loss of BC biomass from

an initial burn when exposed to four subsequent fire events. We

hypothesise that neglecting the BC losses associated with

repeated burning may lead to substantial overestimation of soil

BC storage if soil incorporation does not occur to protect the

particles from further pyrolytic degradation. In turn, accounting

for BC losses as a function of repeated burning will lead to less

BC being available for incorporation into the passive SOC pool,

and these losses will likely have substantial effects on net soil

BC estimates.

Methods

Sample collection and construction of fuelbeds

To test our hypothesis, we selected woody surface fuels that had

been controlled for their particle size and moisture content,

allowing for a controlled laboratory combustion experiment. The

constructed fuelbeds represented a masticated fuel matrix (i.e.

mechanically shredded and distributed tree biomass) within a

western North American temperate conifer forest dominated by

western white pine (Pinus monticola), Douglas-fir (Pseudotsuga

menziesii), and lodgepole pine (Pinus contorta

). Full details on

sample collection, characterisation and fuelbed construction are

provided in a prior study that evaluated the influence of fuel

moisture on BC production rates (Brewer et al. 2013), and here

we provide a brief summary. Fuels were sampled and collected

from an 8-ha stand within the Clearwater National Forest

(46847

N, 119828

W) that included western white pine,

Douglas-fir and lodgepole pine. During mastication the wood

was chipped into predominately small-diameter particles

(,7.6 cm) although some larger diameter particles (.7.6 cm)

were also observed (Brewer et al. 2013). Fifteen fuelbeds were

constructed from this material to match the total surface fuel

loading (5835 g m

2

) and particle size class distribution (7.6–2.5,

2.5–1.3, 1.3–0.6 and 0.6–0.3 cm diameter, and litter) observed in

the field. This total surface fuel loading value is in line with other

observed fuel loadings in mesic mixed conifer systems of the

north-western United States (Kreye et al. 2014).

Initial burn methodology

Fire experiments were conducted at the Idaho Fire Initiative for

Research and Education laboratory located in a climatically

B Int. J. Wildland Fire W. T. Tinkham et al.

controlled environment (Brewer et al. 2013; Smith et al. 2013).

The experimental burn and residue collection and analysis

methodologies followed the procedures detailed by Brewer et al.

(2013). Burns were considered to be complete when mass loss

had ceased, as measured using a Sartorius EB Series scale

(precision: 1 g, range: 0.0005–65.0000 kg, Goettingen, Ger-

many). Following combustion, post-fire residues were sieved

into .6 mm, 1–6 mm and ,1-mm size classes and weighed

using a Sartorius scale (precision: 0.1 g, range: 0.1–2000.0 g),

with two ,1-g sub-samples collected for BC proportion anal-

ysis. BC was quantified using thermo-chemical methods adap-

ted from CTO375 protocols, which isolate a graphitic and

biologically resistant portion of the pyrogenic carbon (Gus-

tafsson et al. 1997; Masiello 2004; Hatten and Zabowski 2009;

Sa´nchez-Garcı´a et al. 2012). Following elemental analysis of

BC proportions (CTO375

BC(%)

), the BC mass (BC

mass

, g) was

calculated using Eqn 1 (Hatten and Zabowski 2009):

mass

¼ CTO375

BC ð%Þ

½pre

mass

 post

mass

ð1Þ

where pre

mass

(g) and post

mass

(g) are the original fuel mass and

the mass of post-fire residues.

Repeated burns methodology

To quantify how the post

mass

residues from the initial burns

persisted under subsequent burns, each burnt fuelbed (n ¼ 15)

was subsequently exposed to four consecutive burn trials. We

acknowledge that ideally we should observe BC incorporation

and losses in a field setting (e.g. Santı´n et al. 2013); however we

contend that this experimental approach allowed us to simulate

long fire return intervals that otherwise could only be inferred

from modelling studies. In each of the subsequent burn trials a

consistent litter fall was included as a layer of solely pine nee-

dles consisting proportionally (by mass) of lodgepole pine and

ponderosa pine (Pinus ponderosa), which are common early

seral species in the study region. The pine needles were added to

the top of the fuelbed and selected in lieu of other litter com-

ponents due to ease of replication, with the mass of needles

increased to account for other missing components (e.g. leaves,

twigs). These pine needle fuelbeds were constructed to resemble

observed rates of litter fuel loading in the fuel collection site,

which fall at the upper bound of litter fuel loading (1700 g m

2

)

in temperate conifer forests in western North America (Law

et al. 2003; Hyde et al. 2011). Fuel moisture was controlled by

placing prepared fuelbeds in a drying oven before combustion

(Table 1; Brewer et al. 2013).

In each repeated burn after the initial characterisation, only

the residues .6 mm in size underwent additional elemental

analysis for assessing BC proportions. This component was then

carried forward into the subsequent burn trials, mixed through-

out the pine needle fuelbeds. Residues ,6 mm were not carried

forward, as these were indistinguishable from the newly burnt

pine needles. To calculate percentage BC remaining [Eqn 2]in

subsequent burns, BC masses were standardised against BC

produced in the initial burn, by:

 BC

ð2Þ

where BC

is the normalised remaining BC after burn number i

(from 2–5); BC

is the BC produced from burn number 1 and BC

is the remaining BC after burn number i.

Statistical analysis and modelling

A repeated-measures ANOVA was used to test for differences in

post-fire residue and BC masses for each burn trial. When

Mauchly’s sphericity assumption was not met, the Greenhouse–

Geisser statistic was used. A Bonferroni post hoc test was used to

compare the main effect of burn number. To generalise our

results and construct a range of partially protected BC budgets,

we fit a negative exponential model, y ¼ a*e

robust to outliers,

to the individual mass loss percentages of the exposed .6-mm

residues, where y is the predicted remaining BC mass (% of the

original mass), and x is the burn number (where x ¼ 1 for the

initial burn). When a ¼ 259.0 (239.5 – 278.6, 95% prediction

bounds) and b ¼0.9476 (1.008 to 0.8877), this model

explained 98% of the variability in our observed data. The

observed BC mass loss rate was used to develop a BC budget

through the five burns. Further, a series of hypothetical BC

budgets, representing varying degrees (0–100%) of protection

of the total BC produced in the initial burn are produced and

discussed. Statistical analyses were conducted using IBM SPSS

predictive analytics software (version 19), and Matlab technical

computing software (version 7.11.1). For this modelling, the BC

produced in the combustion of the added litter layer is not

included, as it is indistinguishable from the fine woody BC.

Results and discussion

Following the initial burns, BC biomass in the burn residues

were divided into three size classes (,1, 1–6 and .6 mm), each

containing significantly different proportions of the residues:

,61, 31 and 8%. The .6-mm BC from the initial burns declined

Table 1. Mean (s.d.) fuelbed characteristics and burn conditions for the five experimental burns (n 5 15 replicates for each burn number)

Burn 1 represents the initial fire, which produced all of the .6-mm residues used in the subsequent burns represented as burns 2–5. All P-values ,0.001

reported from repeated-measures ANOVA

Burn number Bulk density (kg m

3

) Fuel loading (g m

2

) Consumption (%) Fuel moisture (%) Temperature (8C) Relative humidity (%)

1 102.1 (9.3) 5829.7 (211.7) 90.6 (2.6) 10.0 (3.5) 17.8 (7.3) 38.4 (12.3)

2 58.7 (6.3) 2107.2 (242.6) 45.5 (13.1) 9.7 (3.9) 16.7 (2.7) 36.1 (11.0)

3 48.8 (8.4) 1771.9 (201.7) 57.4 (18.1) 11.0 (3.3) 21.5 (2.9) 34.3 (5.2)

4 45.9 (4.4) 1752.5 (97.7) 57.3 (15.4) 9.5 (4.6) 25.4 (5.2) 30.2 (6.9)

5 52.5 (16.9) 1763.3 (128.4) 61.8 (18.1) 10.2 (4.7) 21.5 (2.3) 33.6 (6.4)

Repeated burning on black carbon residence times Int. J. Wildland Fire C

significantly across the four subsequent burns (Table 2, Fig. 1).

Following Burn 5, only 1% of the exposed .6-mm BC produced

in the initial burn (Burn 1) remained. Integrating the charred

residues throughout the litter layer probably made the charred

particles more susceptible to thermal degradation, as the thermal

insulation (e.g. provided by being at the bottom of the litter layer

or within the O-horizon – soil layers with high amounts of

organic matter) would have been reduced (Santı´n et al. 2013).

However, in a field setting, charred residues could remain at the

interface of the O-horizon and mineral soil, or the material could

be mixed throughout the O-horizon; depending on the pedo-

turbation intensity and soil texture, eventually the char will

become incorporated into mineral soil (Gavin 2003). Conse-

quently, our estimates are likely to represent an upper bound of

losses associated with repeated burning, and our results high-

light the need for future research on pedoturbation within the

O-horizon and influence of soil textures on BC soil incorpo-

ration. Further, use of the CTO375 method of characterising the

quantity of BC sets a fairly rigid definition that excludes part of

the BC continuum, especially were BC is produced under low

charring temperatures (Baldock and Smernik 2002; Chen et al.

2014). Nonetheless, our results support the prevailing hypoth-

esis that in forest types with high-frequency burning and little to

no soil incorporation, repeated burning can be a significant

mechanism for BC loss (Ohlson and Tryterud 2000; Preston and

Schmidt 2006; Rovira et al. 2009). Our results also indicate that

in as few as two repeated burns, the majority (,80%, Table 2,

Fig. 1) of the exposed BC produced in an initial fire can be lost.

Given that the best estimates of BC loss rates through biotic and

abiotic (non-pyric) mechanisms range from ,1% to 37% over

100 years (Zimmerman 2010), our results highlight repeated

burning as a potentially significant mechanism of carbon loss.

To estimate the compounding effects of BC loss through

repeated burning, we used the experimental loss rates of the

exposed carbon (Fig. 1) to construct a BC budget spanning five

burns (Fig. 2; Ohlson and Tryterud 2000; Czimczik and

Masiello 2007; Zimmerman 2010). This budget assumes all

material ,1 mm is protected from future burning, implicitly

representing immediate off-site wind transport, as is common in

grassland and savannah ecosystems, or incorporation of fine

particles into the soil and litter matrix (Rumpel et al. 2009).

Protecting the ,1-mm BC left 39% of the BC produced in an

Table 2. Median (s.d.) production and loss rates associated with the post-fire residues and black carbon (BC) (n 5 15)

Burn 1 represents the initial fire, which produced all of the .6-mm residues used in the subsequent burns represented as burns 2–5. All P-values ,0.001

reported from repeated-measures ANOVA. Homogenous subsets as identified by Bonferonni post hoc analysis are identified as a, b, c, d, and e

Burn number Residues (g m

2

) Residue remaining (%) BC (g m

2

) BC remaining (%)

1 198.1 (77.3) a 100.0 (0.0) 0.0650 (0.0290) a 100.0 (0.0)

2 113.7 (50.9) b 55.6 (17.1) 0.0400 (0.0180) b 63.0 (23.0)

3 78.8 (46.6) c 38.5 (15.0) 0.0140 (0.0100) c 17.5 (13.4)

4 54.8 (34.9) d 24.2 (13.3) 0.0010 (0.0006) d 1.4 (0.8)

5 39.5 (30.7) e 15.7 (13.3) 0.0007 (0.0004) d 1.0 (0.6)

100

Black carbon mass (%)

Observations

y ⫽ a e

95% pred. int.

Burn number

Fig. 1. Black carbon loss with repeated burning. Burn number 1 represents

total carbon produced after the initial burn (100%). For each repeated burn

(Burn number 2–5), 15 replicates are shown as observations. The solid black

line represents a robust fit of the model y ¼ a e

, and the dashed lines represent

95% prediction intervals. The fitted model explains 98% of the variability in

the observations (r

adj

¼ 0.98), when a ¼ 259.0 and b ¼0.9476.

100%

186%

253%

315%

376%

2345

Burn number

100

200

Net black carbon (%)

300

400

BC, Burn 1

BC, Burn 2

BC, Burn 3

BC, Burn 4

BC, Burn 5

Fig. 2. Black carbon (BC) budget based on the percentage of BC remaining,

reported in Table 1. The net BC contributing to the budget of each burn

number is tracked via varying shades of grey. Each burn contributes 100%

of the BC generated in the initial burn; 61% of this BC is protected

from degradation and 39% is consider exposed through subsequent burns

following the observed loss rates.

D Int. J. Wildland Fire W. T. Tinkham et al.

initial burn available for further thermal degradation. This

scenario also assumes that an equivalent masticated fuel loading

was reached between each burn, representing a best-case esti-

mate for dry forest types, were net BC increases at a near-linear

rate of ,70% per burn (Fig. 2). By the fifth burn, BC storage is

381% of that created in the initial burn, whereas an estimate

ignoring BC loss from repeated burning would predict a value of

500%. The ‘missing’ 119% represents the tradeoff between

generating new BC while consuming existing BC with each

successive burn. We acknowledge that a full accounting of the

residual size categories would provide more accurate estimates

of BC loss rates; however, given analytical limitations of

separating these fine char fractions from pine needle residues

in the repeated burns, this was not feasible. As grass fires likely

produce charcoal residues smaller than 6 mm, the conclusions of

this experiment are limited to ecosystems with woody vegeta-

tion (i.e. trees and shrubs), which represents a continuum from

savannah to mesic forest.

The degree to which repeated burning influences a BC

budget inherently reflects assumptions on BC protection from

future burning (e.g. via soil incorporation or off-site transport;

Rumpel et al. 2009; Dittmar et al. 2012; Santı´n et al. 2013). To

generalise our results and explore their sensitivity, we used

the fitted model in Fig. 1 to calculate net BC storage under

five different scenarios, where 0%, 25%, 50%, 75% and 100% of

the BC produced in the initial burns is protected from future

burning (Fig. 3). As expected, BC budgets are highly sensitive to

protection rates: at 0% protection, net BC asymptotes around

160% after four burns, implying an upper limit to BC storage.

Under the scenario with 75% BC protected, values fail to

asymptote and reach 416% following the fifth burn (Fig. 3).

Implicitly, these scenarios are both sensitive to, and represent

a series of, generalised environmental conditions that dictate

both BC protection rate (% per year) and the rate of burning (fire

per year) in a given ecosystem. For example, although savannah

ecosystems are frequented by fires every 3–5 years, mesic

forests can exhibit fire return intervals of 200þ years. Although

less well known, soil incorporation and off-site wind transfer

rates of BC in these systems likewise vary (Rumpel et al. 2009;

Nocentini et al. 2010; Kasin and Ohlson 2013), principally due

to differences in soil exposure, topography, precipitation, tem-

perature and wind regimes.

We posit that if burning rates are faster tha n the protec tion

rate re quired to safeguard BC, then most BC produced in a burn

will still be exposed during subsequent burns. This may

represent a scenario of 0% or 25% BC prote ction between

burns, which may be comparable to the short fire return

intervals found in pine–savannah systems if off-site transport

through w ind and water erosion is minimal; this sce nario may

not be commonplace, given these particles are often considered

very susceptible to off-site transport by wind and water

processes in grasslands and savannahs. In contrast, if similar

tradeoffs were observed in other systems such as the longer fire

return inte rvals common to borea l forests, such systems may

exhibit enough time between burns for BC to be incorpora ted

sufficiently far into the litter and duff or mineral soil to be

protected during subsequent burns. Such ecosystems are more

likely represented by a scenario of 75% or 100% BC protection

(Santı´n et al. 2015b).

We contend that ecosystem variations affecting the ratio of

these two processes (fires per percentage BC protected), that is,

the rate of burning (fires per year) and rate of BC protection

(% per year), could determine the sensitivity of soil BC budgets

0% BC protected

162%

247%

331%

416%

500%

12345

Burn number

25% BC protected

50% BC protected

75% BC protected

100% BC protected

BC, Burn 1

BC, Burn 2

BC, Burn 3

BC, Burn 4

BC, Burn 5

100

200

100

300

200

100

400

Net black carbon (%)

300

200

100

500

400

300

200

100

Fig. 3. Black carbon (BC) budgets based on varying modelled scenarios;

each budget assumes that a set level of BC, varying from 0–100% in 25%

steps, is protected from subsequent burning. The remaining exposed BC

follows the fitted model in Fig. 1 to calculate loss rates. Each scenario

implicitly represents varying soil incorporation rates and fire frequencies

across a range of fire-prone ecosystems.

Repeated burning on black carbon residence times Int. J. Wildland Fire E

to repeated burning. Assessment of this ratio could enable

studies to determine if and when net BC storage reaches an

asymptote, but will require understanding the protection and

production rates of residues of different sizes in different

ecosystems. If frequent fire consumes BC faster than it can be

protected by soils or off-site transport, then those forested

systems with short fire return intervals would have lower

amounts of BC compared with forested systems with long fire

return intervals. However, the per-fire production rate of black

carbon is a function of other factors such as fire severity and

biomass production between fires, which complicate studies

trying to determine the variability in this relationship (Jauss

et al. 2015).

Ultimately, these processes likewise dictate the feasibility of

using fuel treatments and fire hazard management (prescribed,

wildland fire use fires, etc.) as tools to increase soil BC storage

(DeLuca and Aplet 2008; Santı´n et al. 2015b). Given the

tradeoff between BC production and consumption in subsequent

burns, maintaining ‘natural’ fire regimes in ecosystems histori-

cally characterised by high-frequency fires (e.g. burning once

every several years to decades), may be at odds with maximising

soil BC storage (Cheng et al. 2013).

Conclusion

This study highlights the potential importance of physical loss of

BC through repeated burning, adding combustion as a key

mechanism to previous work demonstrating BC loss through

biological and physical degradation. For BC that remains in situ

to be most effective as a net carbon sink, it must be incorporated

into the mineral soil matrix before subsequent burning. Our work

is a first step towards quantifying BC loss rates from repeated

burning to more accurately model long-term BC storage in soil

organic pools, but projecting the long-term impacts on carbon

budgets requires more precise estimates of BC protection rates.

Biogeochemical models that track BC should be sensitive to the

combined effects of burning, soil incorporation and off-site

transport rates, as exemplified by the ratio of these processes.

Acknowledgements

The combustion facility was built using funds provided by NSF Idaho

EPSCoR Program (EPS-0814387). The Joint Fire Sciences Program funded

the BC analysis (11–3-1–30). This work was partially funded by the National

Aeronautics and Space Administration under award NNX11AO24G.

References

Baldock JA, Smernik RJ (2002) Chemical composition and bioavailability

of thermally altered Pinus resinosa (Red pine) wood. Organic Geochem-

istry 33(9), 1093–1109. doi:10.1016/S0146-6380(02)00062-1

Brewer NW, Smith AMS, Hatten JA, Higuera PE, Hudak AT, Ottmar RD,

Tinkham WT (2013) Fuel moisture influences on fire-altered carbon in

masticated fuels: an experimental study. Journal of Geophysical

Research 118, 31–40. doi:10.1029/2012JG002079

Chapin FS, III, McGuire AD, Randerson J, Pielke R, Sr, Baldocchi D,

Hobbie SE, Roulet N, Eugster W, Kasichke E, Rastetter EB, Zimov

SA, Running SW (2000) Arctic and boreal ecosystems of western North

America as components of the climate system. Global Change Biology 6,

211–223. doi:10.1046/J.1365-2486.2000.06022.X

Chen CP, Cheng CH, Huang YH, Chen CT, Lai CM, Menyailo OV,

Fan LJ, Yang YW (2014) Converting leguminous green manure into

biochar: changes in chemical composition and C and B mineralization.

Geoderma 232–234, 581–588. doi:10.1016/J.GEODERMA.2014.06.021

Cheng CH, Chen YS, Huang YH, Chiou CR, Lin CC, Menyailo OV (2013)

Effects of repeated fires on ecosystem C and N stocks along a fire

induced forest/grassland gradient. Journal of Geophysical Research 118,

215–225. doi:10.1002/JGRG.20016

Czimczik CI, Masiello CA (2007) Controls on black carbon storage in soils.

Global Biogeochemical Cycles 21, GB3005. doi:10.1029/2006GB002798

Czimczik CI, Preston CM, Schmidt MWI, Schulze ED (2003) How surface

fire in Siberian Scots pine forests affects soil organic carbon in the forest

floor: stocks, molecular structure, and conversion to black carbon.

Global Biogeochemical Cycles 17, 1020. doi:10.1029/2002GB001956

DeLuca TH, Aplet GH (2008) Charcoal and carbon storage in forest soils of

the Rocky Mountain West. Frontiers in Ecology and the Environment 6,

18–24. doi:10.1890/070070

Dittmar T, de Rezende CE, Manecki M, Niggemann J, Ovalle ARC,

Stubbins A, Bernardes MC (2012) Continuous flux of dissolved black

carbon from a vanished tropical forest biome. Nature Geoscience 5,

618–622. doi:10.1038/NGEO1541

Eckmeier E, Gerlach R, Skjemstad JO, Ehrmann O, Schmidt MW (2007)

Minor changes in soil organic carbon and charcoal concentrations

detected in a temperate deciduous forest a year after an experimental

slash-and-burn. Biogeosciences 4, 377–383. doi:10.5194/BG-4-

377-2007

Egli M, Mastrolonardo G, Seiler R, Raimondi S, Favilli F, Crimi V,

Krebs R, Cherubini P, Certini G (2012) Charcoal and stable soil organic

matter as indicators of fire frequency, climate and past vegetation in

volcanic soils of Mt. Etna, Sicily. Catena 88(1), 14–26. doi:10.1016/

J.CATENA.2011.08.006

Fang Y, Singh B, Singh BP, Krull E (2014) Biochar carbon stabiligy in

four contrasting soils. European Journal of Soil Science 65(1), 60–71.

doi:10.1111/EJSS.12094

Flannigan MD, Cantin AS, de Groot WJ, Wotton M, Newbery A, Gowman

LM (2013) Global wildland fire season severity in the 21st century.

Forest Ecology and Management 294, 54–61. doi:10.1016/J.FORECO.

2012.10.022

Gavin DG (2003) Forest soil disturbance intervals inferred from soil

charcoal radiocarbon dates. Canadian Journal of Forest Research 33,

2514–2518. doi:10.1139/X03-185

Goldberg ED (1985) ‘Black carbon in the environment: properties and

distribution.’ (John Wiley & Sons: New York, NY)

Gustafsson O, Haghseta F, Chan C, MacFarlane J, Gschwend PM (1997)

Quantification of the dilute sedimentary soot phase: implications for

PAH speciation and bioavailability. Environmental Science & Technol-

ogy 31(1), 203–209. doi:10.1021/ES960317S

Hatten JA, Zabowski D (2009) Changes in soil organic matter pools and

carbon mineralization as influenced by fire severity. Soil Science Society

of America Journal 73, 262–273. doi:10.2136/SSSAJ2007.0304

Hurteau MD, Brooks ML (2011) Short- and long-term effects of fire on

carbon in US dry temperate forest systems. Bioscience 61(2), 139–146.

doi:10.1525/BIO.2011.61.2.9

Hyde JC, Smith AMS, Ottmar RD, Alvarado EC, Morgan P (2011) The

combustion of sound and rotten coarse woody debris: a review. Interna-

tional Journal of Wildland Fire 20, 163–174. doi:10.1071/WF09113

IPCC (2013) Fifth Assessment Report (AR5) of the United Nations Inter-

governmental Panel on Climate Change (IPCC). Available at http://ipcc.

ch/report/ar5/index.shtml [Verified 15 December 2015].

Jauss V, Johnson M, Krull E, Daub M, Lehmann J (2015) Pyrogenic carbon

controls across a soil catena in the Pacific Northwest. Catena 124, 53–59.

doi:10.1016/J.CATENA.2014.09.001

Jenkins ME, Bell TL, Norris J, Adams MA (2014) Pyrogenic carbon:

the influence of particle size and chemical composition on soil carbon

release. International Journal of Wildland Fire 23, 1027–1033.

doi:10.1071/WF13189

F Int. J. Wildland Fire W. T. Tinkham et al.

Kasin I, Ohlson M (2013) An experimental study of charcoal degradation

in a boreal forest. Soil Biology & Biochemistry 65, 39–49. doi:10.1016/

J.SOILBIO.2013.05.005

Keiluweit M, Nico PS, Johnson MG, Kleber M (2010) Dynamic molecular

structure of plant biomass-derived black carbon (biochar). Environmen-

tal Science & Technology 44, 1247–1253. doi:10.1021/ES9031419

Kreye JK, Brewer NW, Morgan P, Varner JM, Smith AMS, Hoffman CM,

Ottmar RD (2014) Fire behavior in masticated fuels: a review. Forest

Ecology and Management 314, 193–207. doi:10.1016/J.FORECO.

2013.11.035

Kuhlbusch TAJ, Andreae MO, Cachier H, Goldammer JG, Lacaux J-P,

Shea R, Crutzen PJ (1996) Black carbon formation by savanna fires:

measurements and implications for the global carbon cycle. Journal of

Geophysical Research 101, 23651–23665. doi:10.1029/95JD02199

Law BE, Sun OJ, Campbell J, van Tuyl S, Thornton PE (2003) Changes in

carbon storage and fluxes in a chronosequence of ponderosa pine. Global

Change Biology 9, 510–524. doi:10.1046/J.1365-2486.2003.00624.X

Lehmann J, Skjemstad J, Sohi S, Carter J, Barson M, Falloon P, Coleman K,

Woodbury P, Krull E (2008) Australian climate–carbon cycle feed-

back reduced by soil black carbon. Nature Geoscience 1, 832–835.

doi:10.1038/NGEO358

Littell JS, McKenzie D, Peterson DL, Westerling AL (2009) Climate and

wildfire area burned in western U.S. ecoprovinces, 1916–2003. Ecologi-

cal Applications 19, 1003–1021. doi:10.1890/07-1183.1

Masiello CA (2004) New directions in black carbon organic geochemistry.

Marine Chemistry 92, 201–213. doi:10.1016/J.MARCHEM.2004.06.043

Moritz MA, Parisien M-A, Batllori E, Krawchuk MA, Van Dorn J, Ganz DJ,

Hayhoe K (2012) Climate change and disruptions to global fire activity.

Ecosphere 3, art49. doi:10.1890/ES11-00345.1

Nguyen BT, Lehmann J, Kinyangi J, Semernik R, Riha SJ, Engelhard MH

(2008) Long-term black carbon dynamics in cultivated soil. Biogeo-

chemistry 89, 295–308. doi:10.1007/S10533-008-9220-9

Nocentini C, Certini G, Knicker H, Francioso O, Rumpel C (2010) Nature

and reactivity of charcoal produced and added to soil during wildfire

are particle-size dependent. Organic Geochemistry 41, 682–689.

doi:10.1016/J.ORGGEOCHEM.2010.03.010

Ohlson M, Tryterud E (2000) Interpretation of the charcoal record in

forest soils: forest fires and their production and deposition of

macroscopic charcoal. The Holocene 10, 519–525. doi:10.1191/

095968300667442551

Ohlson M, Dahlberg B, Okland T, Brown KJ, Halvorsen R (2009)

The charcoal carbon pool in boreal forest soils. Nature Geoscience 2,

692–695. doi:10.1038/NGEO617

Preston CM, Schmidt MWI (2006) Black (pyrogenic) carbon: a synthesis of

current knowledge and uncertainties with special consideration of boreal

regions. Biogeosciences 3, 397–420. doi:10.5194/BG-3-397-2006

Rovira P, Duguy B, Vallejo VR (2009) Black carbon in wildfire-affected

shrubland Mediterranean soils. Journal of Plant Nutrition and Soil

Science 172, 43–52. doi:

10.1002/JPLN.200700216

Rumpel C, Ba A, Darboux F, Chaplot V, Planchon O (2009) Erosion

budget and process selectivity of black carbon at meter scale. Geoderma

154, 131–137. doi:10.1016/J.GEODERMA.2009.10.006

Saiz G, Goodrick I, Wurster C, Zimmermann MPN, Bird MI (2014)

Charcoal re-combustion efficiency in tropical savannas. Geoderma

219–220, 40–45. doi:10.1016/J.GEODERMA.2013.12.019

Sa´nchez-Garcı´a L, Cato I, Gustafsson O

(2012) The sequestration sink of

soot black carbon in the Northern European Shelf sediments. Global

Biogeochemical Cycles 26, GB1001. doi:10.1029/2010GB003956

Santı´n C, Doerr SH, Preston CM, Bryant R (2013) Consumption of residual

pyrogenic carbon by wildfire. International Journal of Wildland Fire 22,

1072–1077. doi:10.1071/WF12190

Santı´n C, Doerr SH, Preston CM, Gonza´lez-Rodrı´guez G (2015a)

Pyrogenic organic matter production from wildfires: a missing sink

in the global carbon cycle. Global Change Biology 21, 1621–1633.

doi:10.1111/GCB.12800

Santı´n C, Doerr SH, Kane E, Masiello C, Ohlson M, Preston C, de la Rosa

A, Dittmar T (2015b) Towards a global assessment of pyrogenic carbon

from vegetation fires. Global Change Biology . doi:10.1111/GCB.12985

Schmidt MWI, Noack AG (2000) Black carbon in soils and sediments:

analysis, distribution, and current challenges. Global Biogeochemical

Cycles 14, 777–793. doi:10.1029/1999GB001208

Schmidt MWI, Torn MS, Abiven S, Dittmar ST, Guggenberger G, Janssens

IA, Kleber M, Ko¨gel-Knabner K, Lehmann J, Manning DAC,

Nannipieri P, Rasse DP, Weiner S, Trumbore SE (2011) Persistence

of soil organic matter as an ecosystem property. Nature 478, 49–56.

doi:10.1038/NATURE10386

Singh N, Abiven S, Maestrini B, Bird JA, Torn MS, Schmidt MWI (2014)

Transformation and stabilization of pyrogenic organic matter in a

temperate forest field experiment. Global Change Biology 20(5),

1629–1642. doi:10.1111/GCB.12459

Smith AMS, Hudak AT (2005) Estimating combustion of large downed

woody debris from residual white ash. International Journal of Wildland

Fire 14, 245–248. doi:10.1071/WF05011

Smith AMS, Tinkham WT, Roy DP, Boschetti L, Kremens RL, Kumar SS,

Sparks AM, Falkowski MJ (2013) Quantification of fuel moisture

effects on biomass consumed derived from fire radiative energy retrie-

vals. Geophysical Research Letters 40, 6298–6302. doi:10.1002/

2013GL058232

Vasilyeva NA, Abiven S, Milanovskiy EY, Hilf M, Rizhkov OV, Schmidt

MWI (2011) Pyrogenic carbon quantity and quality unchanged after

55 years of organic matter depletion in a Chernozem. Soil Biology &

Biochemistry 43(9), 1985–1988. doi:10.1016/J.SOILBIO.2011.05.015

Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW (2006) Warming

and earlier spring increase western US forest wildfire activity. Science

313, 940–943. doi:10.1126/SCIENCE.1128834

Zimmerman A (2010) Abiotic and microbial oxidation of laboratory-

produced black carbon (biochar). Environmental Science & Technology

44, 1295–1301. doi:10.1021/ES903140C

www.publish.csiro.au/journals/ijwf

Repeated burning on black carbon residence times Int. J. Wildland Fire G