ORIGINAL RESEARCH
published: 10 Oc obe 2018
doi: 10.3389/ ea .2018.00127
F on ie s in Ea h Science | www. on ie sin.o g 1Oc obe 2018 | Volume 6 | A icle 127
Edi ed by:
Sa ish Myneni,
P ince on Uni e si y, Uni ed S a es
Re iewed by:
Ronny Laue wald,
F ee Uni e si y o B ussels, Belgium
Dimi is Pou sanidis,
Founda ion o Resea ch and
Technology Hellas, G eece
*Co espondence:
S e an H. Doe
[email p o ec ed]
Special y sec ion:
This a icle was submi ed o
Biogeoscience,
a sec ion o he jou nal
F on ie s in Ea h Science
Recei ed: 15 Decembe 2017
Accep ed: 10 Augus 2018
Published: 10 Oc obe 2018
Ci a ion:
Doe SH, San ín C, Me ino A,
Belche CM and Bax e G (2018) Fi e
as a Remo al Mechanism o
Py ogenic Ca bon F om he
En i onmen : E ec s o Fi e and
Py ogenic Ca bon Cha ac e is ics.
F on . Ea h Sci. 6:127.
doi: 10.3389/ ea .2018.00127
Fi e as a Remo al Mechanism o
Py ogenic Ca bon F om he
En i onmen : E ec s o Fi e and
Py ogenic Ca bon Cha ac e is ics
S e an H. Doe 1*, C is ina San ín1,2, Agus ín Me ino3, Clai e M. Belche 4and G eg Bax e 5
1Depa men o Geog aphy, College o Science, Swansea Uni e si y, Swansea, Uni ed Kingdom, 2Depa men o
Biosciences, College o Science, Swansea Uni e si y, Swansea, Uni ed Kingdom, 3Depa men o Soil Science and
Ag icul u al Chemis y, Uni e si y o San iago de Compos ela, Lugo, Spain, 4Wild i e Lab, Ha he ly Labo a o ies, Uni e si y o
Exe e , Exe e , Uni ed Kingdom, 5Wild i e Ope a ions Resea ch, FPInno a ions, Hin on, AB, Canada
Py ogenic ca bon (PyC, cha coal) is p oduced du ing ege a ion i es a a a e o
∼116–385 Tg C y −1globally. I ep esen s one o he mos deg ada ion- esis an
o ganic ca bon pools, bu i s long- e m a e and he p ocesses leading o i s deg ada ion
emain subjec o deba e. A equen ly highligh ed po en ial loss mechanism o PyC is
i s consump ion in subsequen i es. Howe e , only h ee s udies o da e ha e es ed
his hypo hesis wi h epo ed losses o <8–37%, wi h he e ec s o PyC chemical
cha ac e is ics and i e condi ions on PyC loss in wild i es emaining unexplo ed. To
add ess his, we placed ma e ials wi h di e en deg ees o he mal and chemical
ecalci ance (A: wild i e cha coal, B: slash-pile cha coal, C: pine wood and D: ceda
wood) on he g ound su ace jus p io o a high-in ensi y and a low-in ensi y bo eal
o es wild i e. Mass losses we e highly a iable and dependen on i e- and sample
cha ac e is ics. Mass losses ac oss bo h i es (as % o d y weigh ) we e o A: 66.5
±25.2, B: 41.7 ±27.2, C: 78.2 ±14.9, and D: 83.8 ±18.9. Mass loss co ela ed
signi ican ly wi h maximum empe a u e (Tmax) eco ded on sample su aces ( =0.65,
p=0.01), bu only weakly ( =0.33) wi h ime >300◦C. Mass losses also showed
a signi ican nega i e co ela ion ( = −0.38, p=0.05) wi h he mal ecalci ance
(T50) de e mined using Di e en ial Scanning Calo ime y (DSC) and Tmax wi h cha coal
e lec ance (Ro) de e mined a e he i es ( =0.46, p=0.05). Losses in he high-in ensi y
i e we e signi ican ly highe (p=0.05) han in he low-in ensi y i e, bu he la e had a
highe a e o con e sion o uel o PyC. Ou esul s demons a e ha exposu e o i e can
indeed be a signi ican emo al mechanism o PyC ha emains exposed on he g ound
a e a p e ious i e. The losses ound, howe e , a e likely o ep esen an ex eme uppe
ange as mos PyC p oduced in a i e would no emain exposed on he g ound su ace
by he ime he nex i e occu s. Ou da a also demons a e, o eal wild i e condi ions,
he (i) con as ing esis ance o di e en PyC ypes o combus ion and (ii) con as ing
ne PyC losses be ween di e en i e in ensi ies. The DSC and e lec ance (Ro) esul s
suppo he use ulness o hese analyses in e lec ing he mal deg ada ion esis ance
and empe a u e exposu e unde ac ual wild i e condi ions.
Keywo ds: black ca bon, bo eal o es , ca bon balance, cha coal, cha coal e lec ance, wild i e, managemen
bu n, he mal analysis
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
INTRODUCTION
Py ogenic ca bon (PyC, also known as cha coal, py ogenic
o ganic ma e o black ca bon) is a ubiqui ous o ganic esidue
esul ing om he incomple e combus ion o o ganic uels
du ing ege a ion i es (bo h na u al and an h opogenic i es).
I comp ises a con inuum o py ogenic o ganic compounds
anging om pa ly cha ed ma e ial o soo (San ín e al., 2016a).
I s chemical p ope ies esul in some o he PyC ma e ials
being highly esis an o biological b eakdown, making PyC
one o Ea h’s mos deg ada ion- esis an o ganic ca bon pools.
Es ima es o i s mean esidence ime ange om decades o
cen u ies (e.g., S einbeiss e al., 2009; Bi d e al., 2015) o
millennia (e.g., The enon e al., 2010; de La on aine e al., 2011),
and, impo an ly, consis en ly one o wo o de s o magni ude
longe han hose o hei unbu n p ecu so s (B uun and EL-
Zehe y, 2012; San os e al., 2012; Naisse e al., 2014). The mos
ecen es ima es o PyC global p oduc ion ange om be ween
63 and 140 Tg C y −1(Bi d e al., 2015) and 116–385 Tg C
y −1(San ín e al., 2016a), which is equi alen o ∼0.1–0.6%
o he annual e es ial ne p ima y p oduc i i y (Hus on and
Wol e on, 2009).
P oduc ion o PyC and i s subsequen a e, he e o e, ha e
impo an implica ions o he global ca bon cycle (Land y and
Ma hews, 2017). Howe e , i s long- e m a e in he en i onmen
and he speci ic bio ic and abio ic p ocesses leading o i s
deg ada ion emain he subjec o much deba e (Bi d e al., 2015;
San ín e al., 2016a). A po en ially majo mechanism highligh ed
in he li e a u e o he abio ic e es ial b eakdown o PyC has
been i s in si u combus ion in subsequen i es (e.g., Ohlson and
T y e ud, 2000; Czimczik e al., 2005; P es on and Schmid , 2006;
Czimczik and Masiello, 2007; Kane e al., 2010; Foe eid e al.,
2011). PyC p oduced in a i e and no anspo ed o si e by wind
o wa e , o p o ec ed h ough i s ans e o su icien dep hs
below li e o soil su ace laye s, can ac as uel and be subjec o
combus ion in subsequen i es. This abio ic loss mechanism o
PyC could be pa icula ly impo an in en i onmen s wi h sho
i e- e u n in e als, low bio u ba ion and opog aphy, which
limi , espec i ely, i s e ical mo emen and o -si e anspo .
The e icacy o his p oposed loss mechanism, howe e , has
emained la gely un es ed and was based, un il ecen ly, on
li le mo e han specula ion. Indeed, o ou knowledge, his “ i e
loss” hypo hesis has, o da e, been es ed empi ically in only he
ollowing h ee s udies. In a high-in ensi y expe imen al bo eal
o es i e, San ín e al. (2013) ound median losses o cha coal
pieces (2–4 cm size) placed in he o es loo laye (∼2 cm dep h)
o be <15%. Saiz e al. (2014) epo ed mean losses o <8% o
cha coal pieces (0.5–1.5 cm size) placed on he soil su ace in a
p esc ibed i e in an open sa annah woodland. Using coni e ous
uels ypical o he wes e n US, Tinkham e al. (2016) ound ha
epea ed bu ning o li e beds in he labo a o y educed he
PyC mass (>6 mm size) o med in he i s i e, by 37% in he
subsequen i e.
Gi en ha ege a ion i es as well as PyC ma e ials a y
widely in hei cha ac e is ics, he cu en s udy aimed o
explo e he ole o i e and PyC cha ac e is ics as po en ially
impo an a iables in de e mining PyC emo al a es du ing
i es. We hypo hesized ha PyC wi h di e en deg ees o
ca boniza ion exposed o i es o con as ing in ensi ies would
esul in di e en PyC losses du ing bu ning. Thus, wo di e en
ypes o PyC (wild i e and slash-pile cha coal om jack pine
wood) and, o compa ison, uncha ed wood (jack pine and
wes e n ed ceda ), we e exposed o wo con as ing ypes
o wild i es: a high-in ensi y c own i e and a low-in ensi y
su ace i e. Tempe a u e exposu e o he samples du ing he
i es was de e mined using he mocouples a ached o sample
su aces, and he chemical and physical cha ac e is ics o he
exposed and, whe e ele an , e e ence (unbu n ) samples
de e mined using he mog a ime ic-di e en ial scanning
calo ime y, elemen al- and cha coal su ace e lec ance
analyses.
MATERIALS AND METHODS
S udy A ea and Expe imen al Se up
Field expe imen s we e ca ied ou in he G ea Sla e Lake egion
o he Canadian No hwes Te i o ies in summe 2015. This
bo eal egion has a d y, sub-humid con inen al clima e wi h
an annual p ecipi a ion o ∼300 mm, a wild i e season usually
be ween May and Sep embe , and a i e ecu ence in e al in
he egion o 150 o 200 yea s. The o es is domina ed by black
sp uce (Picea ma iana) and jack pine (Pinus banksiana) wi h
o en hick o ganic soil laye s. The opog aphy is la gely la wi h
an ele a ion be ween 100 and 200 m.a.s.l. (Alexande e al., 2004;
S ocks e al., 2004).
Two ypes o PyC and wo ypes o uncha ed wood ma e ials
we e used in he i e exposu e expe imen s (Figu es 1A,B):
(A) Wild i e cha coal gene a ed du ing an expe imen al wild i e
(June, 2012) in a ma u e jack pine s and in his egion
and collec ed in June 2013 (61◦34’55” N; 117◦11’55” W, see
San ín e al. (2015), o si e and i e cha ac e is ics).
(B) Cha coal om a jack pine slash-pile (a pile o slash gene a ed
in clea ing/ha es ing) bu n in he win e o 2011/12 and
collec ed in June 2012 a a si e adjacen o whe e wild i e
cha coal was sampled. Due o i s p oduc ion in a long-
bu ning slash pile i e, his ype o cha coal was specula ed in
p e ious wo k (San ín e al., 2013) o exhibi a highe deg ee
o cha ing and he e o e g ea e esis ance o combus ion
when exposed o a subsequen i e han wild i e cha coal.
(C) Unbu ned jack pine down wood (small dead b anches)
collec ed in 2012 om an unbu ned a ea in he same s udy
egion.
(D) Unbu ned blocks cu om wes e n ed ceda (Thuja plica a)
wood we e included o compa ison as a low-densi y wood
used in a complemen a y s udy ocusing speci ically on
cha coal e lec ance (Belche e al., 2018).
A–C pieces we e 2–3 cm in diame e and 2–4 cm in leng h. D
blocks we e 3 ×3×3 cm. All we e ee o ba k. Se s we e made
wi h one piece o each ype (A+B+C+D), and he ou pieces
we e linked oge he by hin wi e and wi h a me al label added o
acili a e eloca ion a e i e (Figu e 1B). A o al o 30 se s we e
placed on o he o es loo <24 h p io o exposu e o wo ypes
F on ie s in Ea h Science | www. on ie sin.o g 2Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
FIGURE 1 | (A) A sample se placed on he o es loo p io o he high-in ensi y i e. The samples a e connec ed by a hin wi e o acili a e eloca ion. The hicke
wi es a e he mocouples. (N.B. The pine wood sample appea s da k, bu was ee o ba k. The wild i e cha coal is no isible and was loca ed jus below he image
ame.) (B) Examples o cha ed samples a e he high-in ensi y i e. (C) The high in ensi y i e. (D) Sampling a e he low-in ensi y i e whe e much o he s anding
ege a ion emained una ec ed. The igu e is published wi h he consen o he depic ed indi idual ( he lead au ho ) in (D).
o bo eal wild i e condi ions: a high-in ensi y c own i e and a
low-in ensi y su ace i e (i.e., 15 se s o each i e). Indi idual
pieces we e si ua ed a leas 5 cm apa (Figu e 1A).
K- ype he mocouples (≤1 mm diame e ) we e a ached
di ec ly o he su ace o he samples and connec ed o da a
logge s (Lasca , Easylog) bu ied in he adjacen soil o eco d
he empe a u e a he sample su ace- lame in e ace du ing
he i e (a 10 s in e als). Gi en he limi ed numbe o logge s
a ailable, we chose o a ach he mocouples o only one ype o
cha coal (A: wild i e cha coal) on he basis ha any di e ences
in empe a u e/du a ion da a be ween he wo ypes o cha coal
migh be expec ed o be less han be ween he cha coal and
he con as ing ypes o wood. All B, C, and D samples had
he mocouples a ached.
Immedia ely a e he i es samples we e iden i ied
(Figu e 1B), he mocouples and logge s emo ed and samples
placed in sealed con aine s o subsequen labo a o y analysis.
Wild i e Bu n Expe imen s
(1) The high-in ensi y c own i e was an expe imen al i e
conduc ed on 15/6/2015 as pa o he Canadian Bo eal
Communi y Fi eSma P ojec a F . P o idence, (61◦34’55” N;
117◦10’13” W) and aimed a simula ing wild i e condi ions.
The expe imen al plo (∼2,000 m2) was a ma u e s and o
black sp uce and jack pine, o igina ing om a s and- eplacing
wild i e in 1931, wi h a ee densi y (li e and dead, >1.5 m
high) o app oxima ely 5,500 s ems ha−1and an a e age canopy
ee heigh o 11 m (a ypical ma u e s and heigh in his
bo eal egion). The unde s o ey was e y spa se, he o es loo
composed o li e , mosses and lichens, wi h some down wood
p esen . Fi een se s o A-D samples (Figu es 1A,B) we e placed
andomly a leas 1 m apa om each o he wi hin an a ea o 15
×15 m. The i e was as mo ing wi h a o al bu n ime o <5
mins and lame leng hs >5 m abo e canopy heigh (Figu e 1C).
The o e all i eline in ensi y was es ima ed as ∼8,000 kW m−1,
which is wi hin he ypical ange o high-in ensi y bo eal c own
i es in Canada (de G oo e al., 2009). The i e esul ed in he
homogenous and comple e consump ion o he canopy, hin
b anches and he unde s o y, wi h no unbu n o low se e i y
pa ches emaining wi hin he plo .
(2) The low-in ensi y i e was si ua ed ∼170 km eas o he
high-in ensi y i e si e (60◦49′38′′ N; 114◦24′28′′ W). I was a
slow-mo ing, su ace wild i e (Fi e 28, Hay Ri e ) a ec ing a
mixed black sp uce and jack pine s and o compa able s and
cha ac e is ics o he high-in ensi y i e, bu wi h a pea y soil laye
o up 50 cm dep h. This ligh ning-caused wild i e bu n ac oss he
si e on 29/6/2017 du ing calm a mosphe ic condi ions, ad ancing
a <5 m h−1. This allowed su icien ime o place samples and
he mocouples wi h logge s ahead o one o he lanks o he
i e. Fi een se s o A-D samples we e placed 2 m apa along
a ansec pa allel o he i e on . This su ace i e led o he
pa chy smolde ing combus ion o he o es loo o a dep h o up
o 30 cm (Figu e 1D), cha ing o ees o ∼1 m heigh and he
F on ie s in Ea h Science | www. on ie sin.o g 3Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
“ o ching” o a ew indi idual ees, wi h some a eas emaining
unbu ned. The i eline in ensi y was es ima ed o no exceed 500
kW m−1based on lame-leng h/in ensi y ela ionships o jack
pine o es (C uz and Alexande , 2010). Due o he pa chy na u e
o his bu n, no all samples we e exposed o i e and, bu in
some cases, deep bu ning o he o ganic soil laye also led o he
des uc ion o logge s and loss o da a.
Fu he de ails abou he i es a e gi en in Table 1.
Labo a o y Analyses
The ai -d y weigh o all samples (N=120) was de e mined
be o e and a e exposu e o i e, and mass loss calcula ed as:
Mass loss (%)=100∗(1 −pos i e weigh
p e i e weigh )
O he o al 120 samples, 90 had he mocouples a ached, and
o hose, 19 logge s we e los o did no eco d p ope ly du ing
he i es. A subse o 40 samples (5 o each ype A-D o
each i e) spanning a ep esen a i e ange o mass losses and
empe a u e/du a ion eco ds, and p e- i e e e ence samples (n
=1–5 o each ype A–D) we e selec ed o u he analyses
o allow he po en ial ela ionships be ween mass loss, sample
cha ac e is ics, and empe a u e exposu e du ing he i es o be
examined.
Di e en ial Scanning Calo ime y (DSC)
These analyses we e ca ied ou using a Me le Toledo
ins umen o de e mine he he molabili y (Me ino e al.,
2014) o he di e en samples be o e and a e he exposu e
o he i es. App oxima ely 4 mg o g ound sample was placed
in aluminum pans unde a low o d y ai (unde O2 lux;
low a e, 50 mL s−1) a a scanning a e o 10◦C min−1.
The empe a u e anged be ween 50 and 600◦C. Samples o
indium (mp: 156.6◦C) we e used o calib a e he calo ime e .
Samples we e analyzed in duplica e. The hea o combus ion
(Q) was de e mined by in eg a ing he DSC he mog aphs wi h
espec o ime o e he exo he mic egion (150<T<600◦C).
The egion T<150◦C was no conside ed as i is domina ed
by endo he mic eac ions associa ed wi h wa e loss. The a eas
unde he he mog aphs we e di ided in o h ee empe a u e
egions ep esen ing di e en le els o esis ance o he o ganic
ma e ial o he mal oxida ion (Me ino e al., 2014, 2015): be ween
150 and 375◦C (Q1), be ween 375 and 475◦C (Q2) and be ween
475 and 600◦C (Q3). The empe a u e a which 50% o he o al
ene gy is eleased unde he gi en condi ions (T50) was also
de e mined (Ro i a e al., 2008).
Elemen al Composi ion
To al ca bon (C), ni ogen (N) and hyd ogen (H) con en s
(%) we e de e mined using a LECO elemen al analyze . Values
epo ed we e co ec ed by mine al ash con en (as de e mined
by DSC): e.g., C ash-co ec ed =(C∗100)/To al OM loss.
The pe cen age o he C emaining in samples a e i e
(%C em) compa ed o hei ini ial C con en was also calcula ed
as ollows using he ash-co ec ed C% alues:
%C em =100 ∗pos i e C concen a ion g/g∗pos i e mass (g)
p e i e C concen a ion g/g∗p e i e mass(g)
Fo he ceda and pine wood pieces (C & D sample ypes) ha
we e ully cha ed ollowing i e, his igu e is equi alen o
he con e sion a e o i e-a ec ed C o PyC (%PyC). This a e
quan i ies he ac ion o i e-a ec ed C ha is e ained as PyC in
cha ed ma e ial a he han emi ed o he a mosphe e (San ín
e al., 2015).
TABLE 1 | Expe imen al condi ions and esul ing PyC losses in p e ious- and he cu en s udies (n.d., no da a; T, ambien ai empe a u e, and RH, ela i e humidi y
immedia ely p io o igni ion).
S udy En i onmen PyC sou ce and size o
pieces
Su ace uel load (kg
m−2)
PyC size ange and
loca ion
A mosphe ic condi ions
(T, RH, wind speed) & i e
in ensi y
Mean PYC
losses (% mass)
Saiz e al., 2014 Sa anna Acacia aneu a, ∼700◦C
slash bu n cha coal
0.8–1.0 ∼0.5–1.5 cm soil
su ace
26–36◦C; 17–45%; ∼5 m
s−1;n.d.
<8
San ín e al., 2013 Bo eal o es Pinus banksiana, slash-pile
cha coal
9.4 2–4 cm inside li e 28◦C; 22%; 3 m s−1,
∼8,000 kW m−1
23
(median <15)
Tinkham e al.,
2016
Labo a o y Pseudo suga menziesii,
Pinus mon icola, P. con o a
wood chips, expe imen al
lab bu n*
2.1 <6 mm−7 cm inside
li e
17◦C; 36%; <1 m s−1; n.d. 36*
This s udy
(high-in ensi y i e)
Bo eal o es Pinus banksiana, slash-pile
cha coal)
>7 1–3 cm li e su ace 28◦C; ∼32%; ∼2.5 m s−1,
6,000–8,000 kW m−1
64
Pinus banksiana, wild i e
cha coal
84
This s udy
(low-in ensi y i e)
Bo eal o es Pinus banksiana, slash-pile >10 1–3 cm li e su ace ∼28◦C; ∼25%; <1 m s−1,
<500 kW m−1
17
Pinus banksiana, wild i e
cha coal
50
*PyC was quan i ied only in he ac ion >6 mm using a me hod which only conside s pa o he PyC spec um.
F on ie s in Ea h Science | www. on ie sin.o g 4Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
Cha coal Re lec ance
Cha coals ha e been obse ed o ha e a ying abili ies o e lec
ligh when s udied unde oil using a e lec ance mic oscope,
which has been shown o e ol e h oughou py olysis (e.g.,
Belche and Hudspi h, 2016). The e o e, we an icipa ed ha
cha coal e lec ance migh a y be ween he ypes o i es ha
c ea ed o modi ied he PyC examined in his s udy. In o de o
assess his, he PyC samples we e sliced in o sec ions using a band
saw and embedded in polyes e esin, hen g ound and polished
using a Buehle Me aSe 250 g inde –polishe (Buehle , Necka ,
Ge many). The samples we e placed in he esin such ha he
op su ace exposed o he i e was he polished ace ollowing
he me hod desc ibed in Belche and Hudspi h (2016). The
polished samples we e analyzed unde oil (RI 1.514) a 238◦C,
using a Zeiss Axio-Scope A1 op ical mic oscope, wi h a TIDAS-
MSP 200 mic ospec ome e (SMCS L d, Baldock, UK). Thi y
manual pho ome ic e lec ance measu emen s we e aken a
cell-wall junc ions ac oss he polished su ace o all 35 su icien ly
sized pos - i e samples and he mean cha coal e lec ance (Ro%)
calcula ed o each sample.
S a is ical Analysis
S a is ical analyses we e pe o med wi h he so wa e IBM
SPSS S a is ics 20. The le el o signi icance used o all es s
was 5% (i.e., α=0.05). Pea son p oduc -momen co ela ion
coe icien s ( ) we e calcula ed o es linea co ela ions be ween
pai s o a iables. Va iables examined we e: mass loss (%),
TG-DSC pa ame e s (T50, Q1, Q3), C, N, H con en s (%),
ini ial C emaining (%Ci), and Ro (%), as well as maximum
empe a u e eco ded on samples (Tmax), ime >300◦C (s),
in eg al o du a ion >300◦C (◦Cs). The h eshold o 300◦C
was chosen because is widely conside ed he empe a u e abo e
which cha ing begins in woody uels (D ysdale, 2011), he main
chemical ans o ma ions o biomass-de i ed ma e ials du ing
hea ing s a a ound 200–300◦C (Keiluwei e al., 2010), and i
ha ing been used in p e ious s udies ocusing on PyC (e.g., San ín
e al., 2016b, 2017).
To analyze he di e ences in mass loss (%) be ween sample
ypes, one- ac o ANOVA was pe o med using sample ype
as he independen ac o . Be o e he ANOVAs, he equali y
o a iances o he s udied g oups was es ed by he Le ene’s
homoscedas ici y es and he pos -hoc Duncan’s mul iple ange
es (g oups had equal a iances) we e pe o med o iden i y
classes wi h signi ican ly di e en means.
RESULTS
Sample P ope ies Be o e he Fi es
The wo unbu ned wood samples used in his s udy ( ypes C and
D) showed he lowes % C, he highes con en s o he molabile
biomass and he lowes he mal ecalci ance (Q1 >33%; T50
<420◦C), wi h he ceda wood (D) being mo e he molabile han
he pine wood (Table 2,Figu e 2). In compa ison, bo h cha coal
ypes (A and B) showed much highe C con en and he mal
ecalci ance, and lowe he molabili y (Q1 <21%; T50 >447◦C).
As expec ed, he slash-pine cha coal (B) p esen ed, by a , he
highes C% and he mal ecalci ance o all he samples be o e
i e (Table 2).
TABLE 2 | A i hme ic mean alues and s anda d de ia ion (i alicized) o elemen al composi ion and ou comes o he mal analysis (Di e en ial Scanning Calo ime y; DSC)
be o e and a e i e, and o mass loss and e lec ance (Ro) a e i e, o he ou di e en sample ypes o samples o he high-in ensi y (H) and low-in ensi y (L) i es.
P e- i e Pos - i e
C
(%)
N
(%)
H
(%)
T50
(◦C)
Q1
(%)
Q3
(%)
C
(%)
N
(%)
H
(%)
T50
(◦C)
Q1
(%)
Q3
(%)
Ro
(%)
Mass
loss (%)
Wild i e
cha coal
72.5 0.23 2.9 447 20.5 43.8 H 80.2 0.26 2.31 486 9.5 50.6 1.73 77.9
0.9 0.03 0.2 4.1 0.1 3.6 4.3 0.03 1.04 14.5 1.0 12.3 0.74 14.1
L 78.8 0.43 2.43 468 14.1 43.9 0.92 53.7
6.4 0.30 0.51 19.0 6.6 14.0 0.36 31.0
Slash-pile
cha coal
80.2 0.16 2.6 510 6.4 71.4 H 86.8 0.11 1.29 523 6.9 76.4 2.80 52.9
n.a n.a n.a n.a n.a n.a 2.0 0.17 0.24 22.5 1.8 9.1 0.54 19.4
L 85.8 0.17 1.36 519 5.9 76.3 2.37 24.8
2.2 0.14 0.24 14.7 1.4 5.8 0.98 22.7
Pine
wood
46.1 0.07 5.4 417 33.8 25.8 H 76.3 0.29 2.74 470 13.8 46.9 1.47 89.0
n.a n.a n.a n.a n.a n.a 2.2 0.03 0.36 7.1 2.5 4.8 0.26 6.6
L 73.2 0.38 2.83 459 18.4 39.6 0.97 65.9
2.3 0.06 0.41 10.4 4.2 6.9 0.27 9.0
Ceda
wood
50.9 0.23 n.d. 395 40.5 4.6 H 75.6 0.32 2.65 458 19.1 35.2 1.46 96.2
0.2 0.04 1.3 0.5 1.6 3.6 0.03 0.20 3.1 3.6 3.7 0.34 3.8
L 72.9 0.34 2.75 453 21.1 32.5 1.33 70.4
1.2 0.09 0.28 10.5 5.2 7.1 0.16 19.3
T50 index: loss o 50% o ene gy du ing DSC; Q1 and Q3: pe cen age con ibu ions o combus ion hea s ob ained by DSC analysis. No e ha mean mass loss alues gi en he e
ep esen a subse o samples analyzed o speci ic cha ac e is ics gi en in his able. Mean mass losses o all samples a e gi en in Table 4.
F on ie s in Ea h Science | www. on ie sin.o g 5Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
FIGURE 2 | Compa ison o he di e en ial scanning calo ime y cu es (a i hme ic mean alues o 5 samples each) o he ou samples ypes be o e and a e he i e.
TABLE 3 | Pea son p oduc -momen co ela ion coe icien s ( ) be ween selec ed chemical and e lec ance cha ac e is ics o he s udied samples, and key i e
pa ame e s.
MaxT In Secs >300◦C Mass loss% Ro% C% H% T50
MaxT 1 0.228 0.644** 0.462* 0.351 −0.486** 0.374
In Secs>300◦C 1 0.483** 0.548** 0.168 −0.322 0.216
Mass loss% 1 −0.148 −0.418** 0.190 −0.379*
Ro% 1 0.768** −0.785** 0.722**
C% 1 −0.900** 0.908**
H% 1 −0.856**
T50 1
Only he mos ele an co ela ions a e shown he e. Fo co ela ions among o he a iables analyzed see Table S1. Numbe o samples: 28–47 (see Table S1)
*Signi ican a he 0.05 le el.
**Signi ican a he 0.01 le el.
Mass Losses and Thei Co ela ions Wi h
Tempe a u e/Du a ion Reco ds
Conside ing da a om bo h i es oge he (N=101), wild i e
cha coal ( ype A), exhibi ed a signi ican ly highe mean mass loss
(66.5 ±25.2%) compa ed o slash-pile cha coal (Type B, 41.7
±27.2%). Pine wood ( ype C) showed highe mean mass losses
(78.2 ±14.9%) compa ed o bo h cha coal ypes A&B, and ceda
wood ( ype D) he g ea es losses (83.8 ±18.9%). Howe e , gi en
he high a iabili y in he da a, he di e ences be ween mean
mass losses be ween he g oups o wild i e cha coal, pine wood
and ceda wood we e no signi ican .
Fo he combina ion o all he samples wi h empe a u e da a
associa ed (N=71) and o he h ee i e pa ame e s analyzed,
Tmax showed he s onges posi i e co ela ion ( =0.64) wi h
mass loss (Table 3,Figu e 3). Co ela ion be ween mass loss and
he in eg al >300◦C was also signi ican , bu weake ( =0.48).
No signi ican co ela ion was ound be ween mass loss and ime
>300◦C ( =0.33) (Supplemen a y Table 1).
When compa ing he wo i es (Figu e 4), he mean
maximum empe a u e (Tmax) eco ded on all samples in he
high-in ensi y i e was 855◦C ( ange 661–1005◦C; N=39),
whe eas in he low-in ensi y i e i was only 434◦C ( ange 40–
792◦C; N=32). Indeed nea ly all Tmax alues eco ded on
samples in he high-in ensi y i e exceed hose o low-in ensi y
i e (Figu e 4). The high-in ensi y i e also exhibi ed longe
mean esidence imes (>300◦C) han he low-in ensi y i e,
F on ie s in Ea h Science | www. on ie sin.o g 6Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
FIGURE 3 | Mass loss e sus maximum empe a u e eco ded o all ceda
wood, pine wood and wild i e cha coal samples (N=71; slash-pile cha coal
samples did no ha e empe a u e da a associa ed). G een ci cles,
high-in ensi y i e; blue ci cles, low-in ensi y i e.
al hough hese di e ences be ween i es we e less p onounced
(Figu e 4). Mean sample exposu e du a ions >300◦C we e
245 s ( ange 70–570 s) and 142 s (0–620 s) o he high- and
low-in ensi y i e espec i ely, and he mean in eg als >300◦C
we e 53,656◦C s ( ange 20,689–122,798◦C s) and 18,589◦C s
( ange 0–99,327◦C s) o hese espec i e i es. All samples
exposed o he high-in ensi y i e, and mos samples exposed
o he low-in ensi y i e expe ienced mass losses, wi h hese
mean losses being signi ican ly highe in he wood pieces
(80.7 ±16.8%; n=49) han in cha coal pieces (54.7 ±
28.5%; n=52) and in he high-in ensi y han in he low-
in ensi y i e (80.4 ±18.6%; n=47 s. 53.7 ±27.4%; n
=54). These signi ican di e ences in losses be ween i e
ypes ag ee wi h hei con as ing empe a u e eco ds (Table 4,
Figu e 4). Rega ding he di e ences in mass losses be ween
he wo i es o he di e en sample ypes, mass losses we e
consis en ly smalle in he low-in ensi y han in he high-
in ensi y i e and in bo h cases ollowing he o de D>C>A>B
(Figu e 5). The a iabili y o hese losses was highe in he low-
in ensi y i e in all cases excep ype D, he slash-pile cha coal
(Figu e 5).
The mal-, Elemen al-, and Re lec ance
Cha ac e is ics
Resul s o he elemen al-, DSC- and e lec ance analyses o
each i e a e gi en in Table 2, and DSC cu es o each
sample ype in Figu e 2. As highligh ed be o e, he DSC
analyses show ha , p io o he i es, he slash-pile cha coal
was he mos he mally ecalci an (i.e., leas he molabile)
ma e ial used in his expe imen (mean T50 =510◦C)
ollowed by wild i e cha coal (447◦C), pine wood (417◦C),
and ceda wood (395◦C). This o de is e lec ed also in
he inc easing alues o Q1 and dec easing alues o Q3
(Table 2).
A e sample exposu e o i e, he same o de is e ained,
bu wi h subs an ial inc eases in he mal ecalci ance o he
wood samples ollowing cha ing, and only modes inc eases
o he cha coal samples (Table 2,Figu e 2). These inc eases in
ecalci ance we e g ea e in he high- han he low-in ensi y i e,
which ma ches he g ea e he mal exposu e o he samples in he
o me (Table 4).
The C concen a ion also inc eased in all sample ypes
ollowing i e exposu e, bu his change was much mo e
p onounced o he wood samples han cha coal samples
(Table 2). Gi en ha exposu e o i e does no only lead o
mass- and he e o e C losses h ough combus ion, bu also o
ela i e en ichmen in C in he emaining PyC h ough py olysis,
we also calcula ed he % o PyC emaining in ela ion o he
o al C con en o he samples p io o exposu e (Table 4).
Mean alues o PyC% emaining we e, espec i ely o high-
and low-in ensi y i es: 26.2 s. 48.3 (wild i e cha coal), 44.8
s. 80.8 (slash cha coal), 15.7 s. 54.2 (pine wood) and 8.2 s.
30.8 (ceda wood). As highligh ed in he me hodology, o he
ully cha ed wood samples C&D, his pa ame e ep esen s he
uel C o PyC con e sion a e used in San ín e al. (2015). I is
in e es ing o no e ha he low-in ensi y i e esul ed in a highe
mean PyC p oduc ion a io (54.2 and 30.8%) han he high-
in ensi y i e (15.7 and 8.2%) o he pine and ceda wood samples
espec i ely.
Rega ding N and H concen a ions, N inc eased and H
dec eased a e i e o mos sample ypes (Table 2). Cha coal
e lec ance (Ro) ollowing he i es was highes o slash-pile
cha coal wi h alues o he o he sample ypes being a ound
hal his alue (Table 2). Values ollowing he high-in ensi y
i e o e all exceed hose ollowing he low-in ensi y i e wi h
espec i ely mean Ro alues o 2.80 s. 2.37 (slash cha coal), 1.73
s. 0.92 (wild i e cha coal), 1.47 s. 0.97 (pine wood), and 1.46 s.
1.33 (ceda wood).
Co ela ions Be ween Sample
Cha ac e is ics, Cha coal Re lec ance and
Fi e P ope ies
When conside ing all samples exposed o he i es oge he ,
signi ican co ela ions we e ound be ween some sample
chemical and e lec ance cha ac e is ics, and wi h i e p ope ies
(Table 3). Fo example, T50 showed posi i e co ela ions wi h
C% ( =0.91), Ro ( =0.72) and a nega i e co ela ion wi h mass
loss (−0.38) and wi h H% ( = −0.86). C% co ela ed wi h Ro
( =0.72) and mass loss (−0.42). Maximum empe a u e a he
uel- lame in e ace (Tmax) showed no only a posi i e co ela ion
wi h mass loss ( =0.64) as p esen ed in mo e de ail abo e,
bu also wi h Ro ( =0.46) and H% ( = −0.49). The du a ion
in eg al o exposu e >300◦C was simila ly well co ela ed o
Ro ( =0.55), and weake (bu s ill signi ican ly) co ela ed o
mass loss han Tmax ( =0.48). All o he ela ionships be ween
i e cha ac e is ics, including exposu e du a ion >300◦C, showed
weake and non-signi ican ela ionships wi h sample chemical
and e lec ance cha ac e is ics (Table S1).
F on ie s in Ea h Science | www. on ie sin.o g 7Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
FIGURE 4 | Maximum empe a u es (◦C) (le ) and esidence imes o T >300◦C ( igh ) eco ded a he sample- lame in e ace du ing he low- and high-in ensi y i es
o all sample ypes combined (N=71). Ci cles indica e ou lie s.
TABLE 4 | Mean alues (and ange) o empe a u e cha ac e is ics, mass losses and % PyC emaining, sepa a ed by sample and i e ypes.
Tmax (◦C) Time >300◦C (s) In eg al >300◦C (◦Cs) Mass loss (%) PyC emaining (%)
HIGH-INTENSITY FIRE
Wild i e cha coal (n=12; 6) 882 283 61197 82 26.2
(735–1005) (70–410) (20689–101113) (54–100) (6.0–53.6)
Slash-pile cha coal (n=15; 5) n.d. n.d. n.d. 63 44.8
(33–100) (19.4–73.7)
Pine wood (n=13; 5) 833 249 53,132 86 15.7
(661–1003) (80–570) (21406–122798) (57–100) (5.1–31.7)
Ceda wood (n=14; 3) 851 208 47681 98 8.2
(747–1002) (90–460) (22932–97238) (90–100) (4.7–14.2)
LOW-INTENSITY FIRE
Wild i e cha coal (n=10; 5) 390 16 13685 56 48.3
(50–603) (0–590) (0–43290) (2–82) (17.2–97.1)
Slash-pile cha coal (n=15; 5) n.d. n.d. n.d. 17 80.8
(12–64) (39.5–100)
Pine wood (n=11; 5) 432 90 14286 41 54.2
(85–715) (0–300) (0–51011) (5–80) (31.5–65.8)
Ceda wood (n=11; 4) 475 172 27349 55 30.8
(40–792) (0–620) (0–99327) (5–99) (16.3–44.1)
% PyC emaining is he ac ion o C emaining compa ed o he o al C con en o he samples p io o i e exposu e. (The pos - i e wood samples selec ed o chemical analysis we e
ully cha ed, he e o e all he C in hem is conside ed PyC). Sample size (n) applies o all pa ame e s excep o PyC emaining whe e he second i alicized igu e applies.
n.d., no da a.
DISCUSSION
Mass Losses Compa ed o P e ious
S udies
The only wo p e ious s udies examining PyC ecombus ion
unde ield condi ions epo ed mean mass losses o <8% o
cha coal pieces o ∼0.5–1.5 cm (Saiz e al., 2014) and o 23%
(median alue <15%) o cha coal pieces o 2–4 cm (San ín e al.,
2013). In he cu en s udy, mass losses we e subs an ially highe
excep o he slash-pile cha coal exposed o he low in ensi y
i e, which had a mean loss o 17.0% (median 11.8%) (Table 4).
In he high-in ensi y i e, slash-pile cha coal showed mean losses
o 63.6% (median 58.6). Wild i e cha coal exhibi ed losses o
84.0% (median 81.9%) and 49.7% (median 58.6%) in he high-
and low-in ensi y i es, espec i ely (Table 4). Thus, all losses
eco ded he e ac oss he wo ypes o PyC and i es (i) exceed
hose p e iously epo ed om in-si u ield s udies and, excep
o he slash-pile cha coal in he low in ensi y i e, (ii) exceed also
he mean losses o 36% o PyC agmen s >6 mm epo ed in he
labo a o y s udy by Tinkham e al. (2016). The da a p esen ed in
F on ie s in Ea h Science | www. on ie sin.o g 8Oc obe 2018 | Volume 6 | A icle 127
Doe e al. Fi e as a Remo al Mechanism o Py ogenic Ca bon
FIGURE 5 | Mass losses o he di e en ype o samples o he (A) high - and
(B) low-in ensi y i e. Ci cles and s a s indica e ou lie s. To al N=101.
Table 1 allows compa ison o he cu en s udy wi h i e and PyC
cha ac e is ics in hese p e ious s udies.
The compa a i ely low losses in he s udy by Saiz e al. (2014)
(<8%; Table 1) could be a e lec ion o a ela i ely high he mal
esis ance o hei PyC combined wi h a low capaci y o he i e
o lead o PyC combus ion. The sou ce ma e ial, Acacia aneu a,
has oughly wice he densi y (1 g cm−3) o pine wood and was
p oduced in a slash bu n, which ypically ha e much longe
lame esidence imes han wild i e and he e o e lead o a g ea e
he mal deg ada ion esis ance o he cha coal p oduced han in
a wild i e (Massman e al., 2008). The i e, al hough conside ed
in ense in he con ex o sa anna i es, had he lowes uel load
and highes wind speed o all expe imen al i es lis ed in Table 1
and hence is likely o ha e led o a bu n o compa a i ely low
in ensi y and lame esidence ime.
The expe imen al c own i e examined by San ín e al. (2013)
was ca ied ou in he same uel complex and s udy egion as he
cu en s udy and used he same sou ce o slash-pile cha coal.
Fuel load, a mosphe ic condi ions and i eline in ensi y (∼8,000
kW m−2) we e also b oadly compa able o he high in ensi y
i e o cu en s udy (Table 1). The main di e ence be ween
hese wo s udies is ha , in he cu en s udy, PyC pieces we e
placed on he o es loo su ace, whe eas in San ín e al. (2013)
hey we e placed a ∼2 cm dep h in he o es loo laye . In
he 2013 s udy, his could ha e esul ed in he ini ial p o ec ion
om he mal exposu e o he cha coal pieces un il he li e laye
i sel was consumed o his dep h. This may ha e con ibu ed
o he subs an ially lowe combus ion o he slash-cha coal o
23% epo ed by San ín e al. (2013) compa ed o he 64% in he
high-in ensi y i e o he cu en s udy (Table 1).
The mass losses o bo h slash-pile cha coal in he high-
in ensi y i e and o wild i e cha coal in bo h i es o he
cu en s udy a e subs an ially g ea e han he 36% ound in
he labo a o y bu ns o Tinkham e al. (2016) (Table 1). Thei
expe imen , howe e , is no di ec ly compa able o he cu en
s udy no only due o i being ca ied ou unde labo a o y
condi ions using mas ica ed uels. They also conside ed only PyC
pieces >6 mm and wi hin hose, only he ac ion ha is esis an
o chemical deg ada ion based on he CTO375 me hod (Ha en
and Zabowski, 2009). In addi ion, he sample placemen on he
li e su ace, he subs an ially highe uel loads and wa me and
d ie a mosphe ic condi ions in he cu en s udy a e all mo e
conduci e o a g ea e combus ion comple eness han hose in
he labo a o y bu n by Tinkham e al. (2016).
Mass Losses and Thei Co ela ions Wi h
Tempe a u e/Du a ion Reco ds
Ou esea ch design allowed us o link PyC losses di ec ly o
empe a u e eco ded on he uel- lame in e ace o he espec i e
PyC pieces. Tha mass losses show a easonably s ong co ela ion
wi h Tmax ( =0.65; Table 3) is pe haps no su p ising. The
he mocouple ip (≤1 mm diame e ) e lec s he empe a u e
a he uel- lame in e ace, which is in luenced by he ene gy
ecei ed om he su ounding combus ion combined wi h
ha eleased om he bu ning PyC piece i sel . Tmax could
he e o e be expec ed o be a leas somewha ela ed o he
mass loss o a PyC piece, wi h highe empe a u es e lec ing
g ea e losses. Wha may be su p ising is ha mass losses
a e only weakly co ela ed wi h exposu e du a ion >300◦C (
=0.33, no signi ican ; Table S1). Th ee Hund ed deg ee is
widely conside ed he h eshold abo e which cha ing begins
in woody uels (D ysdale, 2011) and hence he du a ion abo e
his h eshold could be expec ed o co ela e mo e closely wi h
mass loss han Tmax. Gi en ha he co ela ion be ween mass
loss and he in eg al >300◦C was s onge ( =0.48) han o
ime >300◦C (and signi ican ), sugges s ha he as e a e o
combus ion associa ed wi h he p ocess ha gene a es highe
empe a u e eadings in he mocouples is mo e impo an in
he o e all mass loss p ocess obse ed he e han he o al ime
du ing which mass loss occu s. This no ion is suppo ed by
a s udy examining o ganic ma e in a soil block a ec ed by
expe imen al labo a o y bu ns, whe e he ime >300◦C also
showed no co ela ion wi h C loss (Me ino e al., 2018).
F on ie s in Ea h Science | www. on ie sin.o g 9Oc obe 2018 | Volume 6 | A icle 127
