Download full-text PDF

Peatlands in the Earth’s 21st century climate system

Article · September 2011with797 Reads
DOI: 10.1139/a11-014
Peatlands occupy a relatively small fraction of the Earth’s land area, but they are a globally important carbon store because of their high carbon density. Undisturbed peatlands are currently a weak carbon sink (~0.1 Pg C y–1), a moderate source of methane (CH4; ~0.03 Pg CH4 y–1), and a very weak source of nitrous oxide (N2O; ~0.00002 Pg N2O–N y–1). Anthropogenic disturbance, primarily agriculture and forestry drainage (10%–20% of global peatlands), results in net CO2 emissions, reduced CH4 emissions, and increased N2O emissions. This likely changes the global peatland greenhouse gas balance to a C source (~0.1 Pg C y–1), a 10% smaller CH4 source, and a larger (but still small) N2O source (~0.0004 Pg N2O–N y–1). There is no strong evidence that peatlands significantly contributed to 20th century changes in the atmospheric burden of CO2, CH4, or N2O; will this picture change in the 21st century? A review of experimental and observational studies of peatland dynamics indicates that the main global change impacts on peatlands that may have significant climate impacts are (1) drainage, especially in the tropics; (2) widespread permafrost thaw; and (3) increased fire intensity and frequency as a result of drier climatic conditions and (or) drainage. Quantitative estimates of global change impacts are limited by the sparse field data (particularly in the tropics), the large variability present in existing data, uncertainties in the future trajectory of peatland use, interactive effects of individual impacts, and the unprecedented rates of climate change expected in the 21st century.
Peatlands in the Earths 21st century climate
Steve Frolking, Julie Talbot, Miriam C. Jones, Claire C. Treat, J. Boone Kauffman,
Eeva-Stiina Tuittila, and Nigel Roulet
Abstract: Peatlands occupy a relatively small fraction of the Earths land area, but they are a globally important carbon
store because of their high carbon density. Undisturbed peatlands are currently a weak carbon sink (~0.1 Pg C y1), a mod-
erate source of methane (CH4; ~0.03 Pg CH4y1), and a very weak source of nitrous oxide (N2O; ~0.00002 Pg N2ONy
Anthropogenic disturbance, primarily agriculture and forestry drainage (10%20% of global peatlands), results in net CO2
emissions, reduced CH4emissions, and increased N2O emissions. This likely changes the global peatland greenhouse gas
balance to a C source (~0.1 Pg C y1), a 10% smaller CH4source, and a larger (but still small) N2O source (~0.0004 Pg
1). There is no strong evidence that peatlands significantly contributed to 20th century changes in the atmospheric
burden of CO2,CH
2O; will this picture change in the 21st century? A review of experimental and observational stud-
ies of peatland dynamics indicates that the main global change impacts on peatlands that may have significant climate im-
pacts are (1) drainage, especially in the tropics; (2) widespread permafrost thaw; and (3) increased fire intensity and
frequency as a result of drier climatic conditions and (or) drainage. Quantitative estimates of global change impacts are lim-
ited by the sparse field data (particularly in the tropics), the large variability present in existing data, uncertainties in the fu-
ture trajectory of peatland use, interactive effects of individual impacts, and the unprecedented rates of climate change
expected in the 21st century.
Key words: peat, land use change, CO2,CH
Résumé : Les tourbières occupent une faible proportion de la surface terrestre, mais leur rôle dentreposage du carbone est
primordial. Les tourbières vierges sont un faible puits de carbone (~0,1 Pg C y1), une source modérée de méthane (CH4;
~0,03 Pg CH4y1), et une faible source doxyde nitreux (N2O; ~0,00002 Pg N2ONy
1). Les perturbations anthropogéniques,
surtout le drainage agricole ou forestier (10%20 % des tourbières), provoquent une augmentation des émissions de CO2et
N2O et une diminution des émissions de CH4. Ainsi, les tourbières deviennent une source de carbone (~0,1 Pg C y1), une
source de CH4diminuée de 10 %, et une source plus grande (mais toujours faible) de N2O (~0,0004 Pg N2ONy
1). Rien
n'indique clairement que les tourbières ont contribué significativement aux changements atmosphériques en CO2,CH
4et N2O
au cours du 20ième siècle, mais ce constat pourrait-il changer au cours du 21ième siècle? Une revue détudes expérimentales
et observationnelles indique que les impacts les plus significatifs des changements globaux sur le climat seront (1) le drai-
nage, spécialement dans les tropiques; (2) la fonte du pergélisol; et (3) une intensité et une fréquence accrue des feux à la
suite d'un drainage et (ou) de sécheresse. L'estimation quantitative des impacts des changements globaux sont limitées par la
rareté des données de terrain (surtout dans les tropiques), la grande variabilité des données, l'incertitude relative aux projec-
tions de changements d'utilisation des terres, l'interaction entre les impacts individuels, et le taux de changement climatique
sans précédent prévu pour le 21ième siècle.
Motsclés : tourbe, modification de lutilisation des terres, CO2,CH
[Traduit par la Rédaction]
1. Introduction
Peatlands are ecosystems with a surface layer of peat at
least 3040 cm thick, but often much thicker. Peat comprises
partially decomposed plant matter that has formed in place,
and is often saturated from its base to, or near to, the surface.
Hydrophytic vegetation that is adapted to these conditions is
predominant (e.g., Rydin and Jeglum 2006). Although classi-
fication criteria vary (e.g., Wheeler and Proctor 2000), fens
and bogs are widely recognized as the two main types of
Received 14 February 2011. Accepted 11 July 2011. Published at on 29 September 2011.
S. Frolking, J. Talbot, and C.C. Treat. Institute for the Study of Earth, Oceans, and Space, University of New Hampshire, Durham NH
03824, USA.
M.C. Jones.*Department of Earth and Environmental Sciences, Lehigh University, Bethlehem, PA USA.
J.B. Kauffman. Northern Research Station, USDA Forest Service, Durham NH USA.
E.-S. Tuittila. Peatland Ecology Group, Department of Forest Sciences, University of Helsinki, Helsinki, Finland.
N. Roulet. Department of Geography, McGill University, Montreal, QC, Canada.
Corresponding author: Steve Frolking (e-mail:
*Water and Environmental Research Center, University of Alaska Fairbanks, Fairbanks, AK.
Environ. Rev. 19: 371396 (2011) doi:10.1139/A11-014 Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
peatlands. Fens are typically wetter peatlands that receive
some water and nutrients from outside their own limits,
whereas bogs are generally drier, nutrient-poor systems pri-
marily fed via atmospheric inputs.
Globally, peatlands cover about 4 million km2(400 Mha)
and contain about 400600 Pg carbon (C) in peat (Lappalainen
1996; Rydin and Jeglum 2006; Tarnocai et al. 2009; Page et al.
2011; Yu et al. 2010).1The majority of this peat (75%80%)
is in boreal and sub-arctic peatlands, with about 10%15% in
tropical peatlands, primarily in SE Asia, and maybe about 10%
in temperate peatlands (Andriesse 1988; Lappalainen 1996).
Many boreal and sub-arctic peatlands are located in regions of
sporadic to continuous permafrost (Vitt et al. 2000; Tarnocai
2006; Smith et al. 2007). Peatlands have a very high C density
per unit area (~50 to >500 kg C m2). Most peat has accumu-
lated within the past 10 00020 000 years (Harden et al. 1992;
Yu et al. 2010). Like all ecosystems, peatlands have aesthetic,
educational, and recreational value. In addition, peatlands pro-
vide a unique habitat for specialized and (or) rare plants and
animals (Parish et al. 2008). Because of their slow decomposi-
tion rates, peatlands can store over thousands of years layered
traces of life such as pollen, spores, seeds, plant parts and
shells, and artifacts and human remains (Barber 1993), provid-
ing unique information about past environments, climates and
cultures. In this analysis, we will not consider likely changes
in any of these important peatland functions, but focus on the
climate system impacts. Peatlands contain a large, potentially
vulnerable C pool close to the atmosphere, that, if destabilized
could have a major impact on climate (Gorham 1991; Moore et
al. 1998; Limpens et al. 2008).
The relatively small area of peatlands (about 3% of the
global land surface) means that widespread changes in peat-
land surface properties (e.g., albedo, surface roughness) will
not have significant global consequences (though they could
be locally important). Instead, peatland functions impact the
global carbonclimate system primarily through the net ex-
change of two carbon-based greenhouse gases carbon diox-
ide (CO2)andmethane(CH
4), and, less significantly, nitrous
oxide (N2O). Most unmanaged or undisturbed peatlands are
weak sinks of atmospheric CO2(<1000 kg C ha1·y1)with
total annual photosynthesis slightly in excess of ecosystem re-
lease of C gases (CO2and CH4) and dissolved organic carbon
(DOC) export. Evidence for this comes from multi-year peat-
land C-balance studies (Roulet et al. 2007; Nilsson et al. 2008;
Koehler et al. 2011, Dinsmore et al. 2010), and from long-
term peat accumulation rates, determined by quantifying C
content in radiocarbon dated peat cores, which are typically
around 100400 kg C ha1·y1for northern peatlands (Gorham
1991; Turunen et al. 2002; Yu et al. 2009), and around 200
1·y1for tropical peatlands (Page et al. 2004;
Rieley et al. 2008). These cores show variability in C accumu-
lation over the Holocene, partly associated with climate varia-
bility (Mauquoy et al. 2008; Beilman et al. 2009; Yu et al.
2009). At the same time, most peatlands emit CH4,sincein
situ CH4production in the water-saturated peat generally ex-
ceeds CH4oxidation in the overlying thin unsaturated peat
layer (Reeburgh 2003). CH4emissions from different peatlands
can differ by several orders of magnitude, depending on hy-
drology, vegetation, and temperature (e.g., Bartlett and Harriss
1993; Christensen et al. 2003; Rydin and Jeglum 2006), but in
aggregate emissions average ~0100 kg CH4Cha
tlett and Harriss 1993; USEPA 2010). Unmanaged or undis-
turbed peatlands are also a generally very weak source of
N2O (e.g., Martikainen et al. 1993; USEPA 2010; however,
see a recent high-latitude study by Repo et al. 2009), and
this is typically ignored in assessments of climate impacts.
How do peatlands currently impact the climate system?
The net climate impact of current emissions has been as-
sessed using the Global Warming Potential (GWP) method-
ology to classify an individual peatland as a net greenhouse
gas source (CH4emission dominating) or sink (net CO2up-
take dominating), based on net annual emission rates, and a
choice of the time horizon (or GWP value) (Shine et al.
1990). Whiting and Chanton (2001) classified seven peat-
lands (sub-tropical to boreal) where they had measured
CO2and CH4fluxes as net greenhouse gas (or CO2-equiva-
lent) sources or sinks. For a 20-year time horizon, all seven
sites were classified as net greenhouse gas sources; for a
500-year time horizon, all were classified as net sinks; and
for a 100-year time horizon, the boreal sites were classified
as sources, and the temperate and sub-tropical sites as sinks.
Similar GWP results were found by Roulet (2000) for Cana-
dian peatlands, by Crill et al. (2000) for natural and man-
aged peatlands in Finland, if they excluded emissions from
storage and combustion of harvested peat, and by Friborg
et al. (2003) for a site in western Siberia. This GWP ap-
proach can be applied to the approximate mean behavior of
global peatlands (for now we ignore disturbed peatlands;
these are discussed in Section 3 of this review). Assuming
a mean uptake of about 300 kg C ha1·y1for 400 Mha of
peatlands (Table 1), the aggregate C uptake is about 0.1 Pg
1or 0.4 Pg CO2y1. Global CH4budgets estimate total
wetland CH4emissions to be 0.10.2 Pg CH4y1,withthe
majority of this from the tropics (USEPA 2010), where CH4
fluxes are dominated by nonpeatland wetlands (e.g., Bridg-
ham et al. 2006). If we approximate the total preindustrial
peatland CH4flux as 0.04 Pg CH4y1(or a mean flux of
~100 kg CH4ha1·y1; Table 1) and use a 100-year GWP
2-equiv kg1CH4(Forster et al. 2007) and
2flux to the atmosphere as pos-
itive, the net climate impact of contemporary annual peat-
land greenhouse gas fluxes is about +0.6 Pg CO2-equiv y1,
a net warming impact. Total N2O emissions are more uncer-
tain, but may be about 0.00002 Pg N2ONy
1(Table 1);
with a 100-year GWP value of 298 kg CO2-equiv kg1N2O
(Forster et al. 2007), this adds an additional 1% to the
CO2-equiv emissions.
Taking a different perspective, Frolking and Roulet (2007)
argued that the contemporary climate impact of the long-term
development of stable peatlands is a result of the effect of sus-
tained net emissions through the Holocene on the current
composition of the atmosphere. Since sustained emissions
cannot be evaluated with standard GWP calculations, net radi-
ative forcing from peatlands has been assessed using simple
atmospheric budget models and time series estimates of net
CO2and CH4fluxes (Laine et al. 1996; Minkkinen et al.
1Given the large uncertainties associated with most global-scale numbers related to peatlands and the nature of arguments put forward in this
paper, we will report values to one or, at most, two significant figures.
372 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
2002; Frolking et al. 2006), under the simplifying assumption
that these emissions from a single component of the earth sys-
tem (e.g., peatlands) can be considered a small perturbation to
an otherwise stable atmosphere. Given the short atmospheric
lifetime of CH4(~10 years), the atmospheric burden adjusts
to a persistent flux (e.g., peatlands through the Holocene) in
less than 100 years, so the contemporary atmosphere only
feelsthe effect of recent emissions. In a simple box model
of the atmosphere (e.g., Frolking and Roulet 2007) a peatland
flux of 0.04 Pg CH4y1and a 10 year atmospheric lifetime
result in a perturbation to the atmosphere of about 0.4 Pg
CH4(or ~200 ppbv (part per billion by volume)). The lifetime
of CO2in the atmosphere is more complicated than CH4, but
a fraction (~20%) of net emissions effectively remains in the
atmosphere for millennia (e.g., Archer and Brovkin 2008; So-
lomon et al. 2009). Using this simplifying assumption, the
current atmosphere deficitdue to Holocene peat accumula-
tion is ~20% of the accumulated peat (Frolking and Roulet
2007), or 100 Pg C (or about 50 ppmv (part per million by
volume) CO2). Weighting these two perturbations by their ra-
diative efficiencies (1.30 × 1013 Wm
0.0198 × 1013 Wm
2·kg1CO2; Ramaswamy et al. 2001),
gives a global net radiative forcing of about 0.7 W·m2,a
net cooling. Again, N2O emissions from undisturbed peat-
lands are likely to be low, and would reduce the net cooling
by <1% (annual emissions ~0.00002 Pg N2ONy
1, atmos-
pheric lifetime ~110 years, radiative efficiency ~3.96×1013
W·m2·kg1N2O; Ramaswamy et al. 2001).
Though these two estimates cannot be directly compared
one (GWP) is of the cumulative radiative impact, integrated
over 100 years, of a single years net emissions, while the
other (box model) is of instantaneous net radiative forcing in
a given year due to the accumulated net emissions over the
lifetime of the peatland they imply that stable peatlands
likely have a relatively small but persistent impact on the cli-
mate system, and have had an impact for millennia. Peatlands
were a part of the preindustrial C cycle and climate system,
both as a CO2sink and as a CH4source, and, a few hundred
(or thousand) years ago, net CO2and CH4emissions were
probably within a factor of two of contemporary emissions
from unmanaged or undisturbed peatlands. Though the prein-
dustrial terrestrial C cycle is typically portrayed as balanced
overall (e.g., fig. 7.3 in Denman et al. 2007), peatlands repre-
sent a persistent imbalance that is smaller than the uncer-
tainty in the global gross C cycle fluxes.
Is there evidence of significant changes in peatland
greenhouse gas balances? The large peatland C pool, accu-
mulated over the Holocene, must have played a significant
role in the Holocene global C cycle (e.g., Wang et al. 2009;
Yu 2011). Armentano and Menges (1986) estimated that peat
C losses from combustion and agricultural drainage were
about 6 Pg C over 17951980, mostly from Europe, temper-
ate North America, and the USSR. This is <10% of global
wood harvest over this period (Hurtt et al. 2006), and there
has been no strong argument for peatlands playing a major
role in the rise in atmospheric CO2over the past 150 years,
Table 1. Recent estimates of global peatland areas and areas disturbed by agricultural or forestry land uses, and contemporary net annual
peatland ecosystem greenhouse gas fluxes to the atmosphere (negative values imply net uptake from atmosphere).
Area (Mha) CO2(kg C ha1·y1)CH
4(kg CH4ha1·y1)N
2O (kg N ha1·y1)
Non-tropical 300 [J10] LU-0: 200 to 1000 [Fig. 2] LU-0: 10 to 700 [Fig. 2] LU-0 (omb.): <0.01 [Ma10]
400 [Y10] LU-0 (min.): 0.1±0.1 [Ma10]
Tropical 40 [Y10] LU-0: 1000±2000 [H11] LU-0: 30±10 [H11] LU-0: comparable to forested
temperate sites [C10]
40 (40 to 70)
[P10] LU-0: 4 to 60 [C10]
70 [J10] LU-0: 800±300 [Mu10] LU-0: 20±8 [R08]
LU-0: 1000 [R08] LU-0: 0 to 20 [Fig. 2]
Non-tropical 30 [J10] LU-A: 5000 (800 to 8000) [Ma10] LU-A: 2 (2 to 20) [Ma10] LU-A: 7 (0.3 to 60) [C10]
50 [J02] LU-F: 2000 [Ma10] LU-F: 7 (9 to 50) [Ma10] LU-A: 10 (1 to 40) [Ma10]
50 [L09] LU-F: 4 (0.3 to 20) [C10]
LU-A: 70% [L09] LU-F: 2 (0 to 30) [Ma10]
LU-F: 30% [L09]
Tropical 20 [J10] LU-A: 10000±3000 [H11] LU-A: 0.3 to 30 [C10] LU-Af: 100 (8 to 300) [C10]
LU-A*: 4000±1000 [Mu10] LU-A: 1±1 [R08] LU-Anf: 0.8 (1.2 to 5) [C10]
LU-A: 2000 [C10, H10] LU-Ap: 40 to 500 [C10] LU-F: 4 (0.6 to 15) [C10]
LU-F: 3000 to 8000 [C10] LU-Ap: 200±100 [H11]
LU-F: 0 to 20±10 [H11]
Note: Land use notations are LU-0: none; LU-A: agriculture (crop or pasture); LU-Af: fertilized agriculture; LU-Anf: nonfertilized agriculture; LU-Ap:
paddy rice; LU-F: forestry, including deforestation. Peatland use for forestry and agriculture generally involves draining. Additional peatland uses include
draining for peat harvest or other mining, inundation by water in reservoir construction, and linear disturbances by roads and seismic lines. Restored sites are
not considered in this table. omb., ombrotrophic; min., minerotrophic. All values are reported here to only one significant figure to remind the reader of the
large uncertainties. References ([Xnn]) are C10: Couwenberg et al. (2010); H10: Hooijer et al. (2010); H11: Hergoualch and Verchot (2011); J02: Joosten
and Clarke (2002); J10: Joosten (2010); L09: Laine et al. (2009); Ma10: Maljanen et al. (2010b); Mu10: Murdiyarso et al. (2010); P10: Page et al. (2011);
R08: Rieley et al. (2008); Y10: Yu et al. (2010).
*From peat; does not include initial loss from forest biomass removal (12 ± 0.4 × 105kg C ha1·y1) or land-clearing fire (1 ± 0.5 × 105kg C ha1·y1)
Per 10 cm drainage depth up to 50 cm, little impact thereafter [C10], or per 10 cm drainage depth up to 100 cm [H10].
Frolking et al. 373
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
though peatland fires have been documented as a strong CO2
source in certain years, e.g., Page et al. (2002) reporting a
strong CO2pulse following peat fires in Indonesia. Wetlands,
including peatlands, are and probably have been the domi-
nant natural source of CH4to the atmosphere (e.g., Mikaloff
Fletcher et al. 2004; Chen and Prinn 2006), and their initia-
tion and development has been used to explain Holocene dy-
namics in atmospheric CH4concentration (e.g., MacDonald
et al. 2006; Korhola et al. 2010; Singarayer et al. 2011), as
well as the inter-hemispheric concentration gradient (e.g.,
Chappellaz et al. 1997). Methane concentration in ice core
bubbles rose by ~100200 ppbv over about 50150 years in
conjunction with the rapid warming at the end of the last gla-
ciation about 14.6 ka BP (Severinghaus and Brook 1999) and
again at the end of the Younger Dryas at about 11.6 ka BP
(Severinghaus et al. 1998). In both cases the rise in CH4
seems to have lagged behind a more rapid warming, and in-
creased CH4emissions from tropical and high-latitude wet-
lands are considered to have been a possible source (Sowers
2006; Schaefer et al. 2006; Petrenko et al. 2009). However,
high temporal-resolution analysis of CH4concentration in
West Antarctic ice sheet divide ice from the late preindustrial
Holocene (LPIH; 2001000 BP or 10001800 AD) did not
find a very clear correlation between atmospheric CH4bur-
den and major climate variables, indicating that wetland or
peatland CH4emissions, in the global aggregate, have a
weak sensitivity to climate variation on the scale that has oc-
curred during the LPIH (Mitchell et al. 2011). In another ice
core analysis, Finkelstein and Cowling (2011) found a nega-
tive correlation between a pollen-based North American wet-
land index for the last 2000 years and CO2from the Law
Dome (Antarctica) ice core, and no correlation between the
wetland index and CH4.
The rise in atmospheric CH4concentration between 1750
and 1990 has been attributed to the dramatic increase in
anthropogenic emissions during that period (Denman et al.
2007). Since 1990, the growth rate of CH4concentration in
the atmosphere has slowed and become much more erratic (e.
g., Dlugokencky et al. 2009). The most likely overall explan-
ation for the slower growth rate is a general stabilization in
total emissions and the atmosphere approaching equilibrium
with that emission strength (Dlugokencky et al. 2003). A
number of explanations for the recent variability have been
put forward, including interannual variability in two natural
sources wetlands and biomass burning (e.g., Bousquet et
al. 2006). Based on several lines of evidence, Dlugokencky et
al. (2009) attribute the significant growth in concentration in
2007 (globally ~8 ppbv but higher at high northern latitudes
and in the tropics) to increased wetland CH4emissions at
high latitudes, which were warmer, and in the tropics, which
were wetter (again, most northern wetlands are peatlands,
while most tropical wetlands are not). The growth in concen-
tration was lower in 2008 globally ~4 ppbv but near zero
at high latitudes. Dlugokencky et al. (2009) concluded:
Since the growth rate returned to near zero in the polar
Northern Hemisphere during 2008, the Arctic has not yet
reached a point of sustained increased CH4emissions
from [thawing] permafrost and CH4hydrates.
Carbon dioxide and (or) CH4gas flux measurements by
eddy covariance or chamber methods have been conducted
on a number of undisturbed temperate and boreal/sub-arctic
peatlands, but few sites have year-round, multiple year re-
cords. Fewer studies have been done on undisturbed tropical
peatlands, and no year-round, multiple year records have
been collected (Couwenberg et al. 2010). The few peatland
sites with complete, multiyear C budgets (CO2+CH
DOC) show a weak but variable sink, averaging about 200
300 kg C ha1·y1(Roulet et al. 2007; Nilsson et al. 2008;
Koehler et al. 2011) or up to 700 kg C ha1for a British
peatland with light grazing (Dinsmore et al. 2010). Other
multiyear records of CO2exchange have similar net fluxes
(e.g., Lund et al. 2010). None of these records show any
signs of a clear trend in the C balance, and weather-driven
interannual variability is high enough that the records would
probably need to be much longer for a statistically significant
trend to emerge (Roulet et al. 2007). The current C accumu-
lation rates derived from the sparse contemporary measure-
ments in UK peatlands suggest that they may be lower than
those measured from peat cores for the last 150 years, but
this pattern might change as more complete C flux datasets
become available (Billett et al. 2010). Aquatic fluxes, includ-
ing DOC losses via leaching and erosion in peatlands with
heavy land-use impacts, e.g., grazing in the UK, may be a
significant C budget pathway for some peatlands (Billett et
al. 2010). Multi-year records of annual CH4emissions are
also too limited to reach any conclusions on trends (Rinne et
al. 2007).
Past peatland C accumulation rates are commonly deter-
mined by radiocarbon dating specific intervals in peat cores
(e.g., Charman 2002). Evidence for rapid peat accumulation
over intervals of few hundred years to millennia has been ob-
served in records from the northern hemisphere boreal re-
gion. The warm early Holocene contains some of the highest
peat accumulation rates in the boreal region (Yu et al. 2009),
and rapid accumulation rate changes have occurred on cen-
tennial timescales, which are attributed to climatic changes
(Yu et al. 2003a, 2003b). In Alaska, the highest peat accu-
mulation rates coincided with maximum insolation seasonal-
ity, suggesting a link exists between high peat accumulation
and warm growing season temperatures and cold, dry winters
(Jones and Yu 2010). Neoglacial cooling beginning about
3500 years ago represents a period where rates of peat accu-
mulation across the boreal region slowed significantly (Yu et
al. 2009) and in some cases ceased (Peteet et al. 1998). In the
tropics, coastal peatlands in Indonesia were most productive
at the end of the Pleistocene and into the early Holocene
(Page et al. 2004) with average peat accumulation rates al-
most twice the rate of boreal peatlands (Turunen et al.
2002). It remains unknown why tropical coastal peat accumu-
lation rates were high during that time but possibilities in-
clude that water tables stabilized and made conditions
favorable for peat accumulation, a change in monsoon fre-
quency or intensity, and a change in climate seasonality
(Page et al. 2004). Inference of paleo-methane fluxes from
peat core data is more problematic, although attempts have
been made to link paleo net primary production (Steinmann
et al. 2006) and lateral expansion rates (Korhola et al. 2010)
to past CH4flux.
Why might we expect a significant change in the peat-
land greenhouse gas balance in the 21st century? One rea-
son is that peatlands, like the worlds forests, which have
similar total C stocks but occupy about ten times as much
374 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
area, are under increasing land use pressure, particularly in
the tropics. A second is that peat C, unlike forest C, is
perched on a fundamental thermodynamic boundary i.e.,
the redox boundary between oxia and anoxia, which can
have a very strong influence on rates of CO2and CH4flux
(Beer and Blodau 2007). This boundary is dynamic at daily
to interannual time scales, but typically lies within tens of
centimetres of the average water table depth (generally shal-
low). Anoxic peat decomposes more slowly than oxic peat,
contributing to C accumulation. Anaerobic decomposition in
peatlands also generates CH4. Relatively small shifts in peat-
land hydrology will expose more or less peat to oxic condi-
tions, enhancing or restricting decomposition, CH4
production and oxidation, and nutrient mineralization. A sec-
ond thermodynamic boundary relevant to many northern
peatlands is the frozen/thawed boundary separating perma-
frost from the surface seasonally thawed (active) layer. Or-
ganic C in permafrost, whether in peat or mineral soil, is
relatively inert both physically and biogeochemically (Riv-
kina et al. 2004). When frozen peat thaws, it becomes more
readily decomposable and decomposition products become
much more susceptible to loss to the atmosphere, leaching or
thermokarst erosion. In addition, changes in permafrost state
can influence vegetation productivity and net C input into
peat (e.g., Camill et al. 2001; Turetsky et al. 2007).
Both of these thermodynamic boundaries are controlled by
the climate system, so carbonclimate feedbacks loops are in-
evitable climate change changes in energy and (or)
water balance changes in water storage and (or) change in
active layer thickness changes in metabolic pathways of
respiration and (or) oxidation changes in CO2and CH4
fluxes changes in atmospheric concentrations of CO2and
CH4further changes in climate. The feedbacks could be
positive or negative (more or less CH4emitted in combina-
tion with more or less net C uptake); they could be strong or
weak (depending on relative change in fluxes); and they
could be transient (years, decades, a century or two) or repre-
sent a persistent change in state/flux. In addition, peatlands
themselves exert controls on the rate at which both of these
boundaries can move through the strong influence of organic
soils on soil thermal and hydraulic regimes (e.g., Yi et al.
2007). Finally, significant movement of the freezethaw
boundary in peatlands is anticipated to affect peatland hy-
drology, i.e., the oxicanoxic boundary, so when both feed-
backs are present, they interact.
Because peatlands develop slowly and are generally resil-
ient ecosystems (e.g., Charman 2002), and because there is a
lack of long-term direct measurements of peatland C cycling,
anticipating the role of peatlands in the Earths 21st century C
balance is a considerable challenge. However, identifying
clearly what is known and what is not known about peatland
responses to changes in temperature and moisture and to other
disturbances is an important step towards quantifying the role
that peatlands might play in the global C cycle over the next
century. To achieve that goal, we first review observational
data on peatland greenhouse gas emission sensitivities to cli-
mate and other global change factors (e.g., land use, pollution
loading, elevated CO2). We then discuss and estimate likely
peatland responses to climate and global change impacts in
the 21st century. We conclude by discussing important issues
that need to be addressed to better understand the long-term
response of peatlands to disturbances.
2. Climate and global change impacts on
peatland CO2,CH
4, and N2O fluxes
Climate change impacts on peatlands include both immedi-
ate direct impacts (warmer or cooler, wetter or drier) and lon-
ger term indirect impacts (permafrost thaw, shifts in
vegetation community composition, changes in disturbance
rates). Additional global change impacts include elevated
CO2and ozone concentrations, enhanced nitrogen and sulfur
deposition, and land use.
2.1. Temperature and moisture sensitivities
Quantification of the response of peatland greenhouse gas
fluxes to temperature and moisture change can be based on
three different lines of evidence interannual variability at
a site, existing climatic gradients, and field manipulation
studies. Measurements in years with different weather can be
used to quantify short-term, direct impacts on gas fluxes of
warmer, wetter or drier conditions, but often both tempera-
ture and moisture vary (e.g., Bubier et al. 2005) so specific
impact attribution can be difficult. Measurements along cli-
matic gradients represent only long-term impacts of climate,
including presumably a plant community composition that is
in quasi-equilibrium with the local climate, and can be con-
founded by other overlapping gradients (e.g., N-deposition).
Manipulation studies, such as warming or draining can con-
trol for different climate change effects, but may more closely
represent changed climate than climate change i.e., the re-
sponse indicates what happens when a large disequilibrium
between ecosystem structure/function and climate is intro-
duced. For example, installing a drainage ditch can rapidly
and persistently lower a peatland water table by 1050 cm
(Paavilainen and Päivänen 1995), while a climate change
trend towards a drier climate would result in a trend towards
increased average water table depth, but not a step-function
change; this difference will likely impact vegetation re-
sponses and lead to different net greenhouse gas fluxes.
Results from global climate model simulations for the 21st
century consistently indicate warming for all regions with
abundant peatlands mean annual air temperature increases
by the end of the century of 28 °C in the boreal/subarctic,
26 °C in the temperate zones, and 25 °C in the tropics
(Christensen et al. 2007). Predictions of precipitation changes
are less consistent among models, but generally there is a
trend towards precipitation increases in the boreal/subarctic
and tropical Southeast Asia, and a drying in Amazonia
(Christensen et al. 2007). Hydrological simulations of climate
change impacts through 2050 by Fekete et al. (2010) suggest
that increased evapotranspiration due to warming will offset
much of the increase in precipitation, with little change in an-
nual runoff (which may be a better measure of peatland water
availability than precipitation). Runoff is predicted to in-
crease in Canada and Russia by ~10 mm·y1and in Southeast
Asia by ~40 mm·y1, and to decrease in South America and
Oceania by ~20 mm·y1; all of these changes are only a few
percent of contemporary runoff (Fekete et al. 2002). While
these changes in annual inputs and outputs are of interest, it
is their combination that controls changes in water stored in
Frolking et al. 375
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
peatlands, and it is the mean and seasonal changes in mois-
ture availability that are most critical for peatland structure
and function. Regional precipitation projections are for in-
creased precipitation in all major peatland regions (Christen-
sen et al. 2007), although potential changes in the seasonal
distribution of precipitation might lead to moisture deficit in
the peatlands that are located in driest end of the peatlands
climatic spectrum. McGill Wetland Model simulations of
21st century climate change impacts on a bog in temperate
eastern Canada and a fen in boreal Sweden (Wu 2009) indi-
cate a net change in the peatlands for warmer and drier con-
ditions, with fen moisture changes significantly greater than
those in bogs.
Field and laboratory manipulations generally show that in-
creased soil temperatures lead to increased CH4emission
from peat (Fig. 1a), although there is no clear relationship
between the magnitude of the temperature change and the
magnitude of the increase, especially for small temperature
increases (approx. 1 to 4 °C). Similarly, a compilation of an-
nual CH4emission from a variety of sites representing a
range of climatic settings shows no clear relationship be-
tween CH4emission and average annual temperature of the
site, though relatively high annual emissions (i.e., >100 kg
CH4ha1·y1) have not been measured at very cold or warm
sites (Tmean <2.5 °C or Tmean > 15 °C; Fig. 2a). Note that
there are very few annual CH4flux estimates for tropical
peatlands. Water table depth was shown to have an interac-
tive effect on the response of fluxes to increased temperatures
(Moore and Dalva 1993; Turetsky et al. 2008).
The relationship between net ecosystem exchange (NEE)
of CO2and a change in soil temperature is hard to address
because of the lack of data. To our knowledge, only one ma-
nipulative study clearly presents the impact of a soil warming
on NEE, in which the response was found to be insignificant
(Chivers et al. 2009), although many studies related an in-
crease in soil temperature with an increase in ecosystem res-
piration (ER) as CO2(Fig. 1b). Increased temperature could
lead to changes in plant production, but the response varies
depending on the plant functional type and the ecosystem
considered (Weltzin et al. 2001, 2003; Gunnarsson et al.
2004; Wiedermann et al. 2007; Dorrepaal et al. 2006; Aerts
et al. 2004). Sulman et al. (2010) found that temperature ex-
plained interannual variability in ER as measured by eddy
correlation only in fen sites, whereas Lund et al. (2010)
found that variation in average temperature between peatland
and wet tundra sites correlates with both gross primary pro-
duction (GPP) and ER, but not with their much smaller net
difference, NEE. Similarly, there is no clear relationship be-
tween NEE and mean annual temperature for the sites for
which an annual NEE value has been reported in the litera-
ture (Fig. 2b). No study was found that relates N2O emission
to temperature manipulations in peatlands.
Water table is known to exert a strong control on CH4
emissions from peatlands. Water table depth determines the
extent of the saturated peat, where CH4is generally pro-
duced, and the thickness of the overlying unsaturated portion,
where CH4is generally oxidized (Whalen 2005). Equally im-
portantly, especially over the longer term, water table has a
strong influence on vegetation composition, which also influ-
ences CH4flux (Bubier et al. 1995; Bubier 1995). Studies
conducted in the lab with peat cores, studies using micro-
and mesocosms, field manipulations, long-term post-drainage
studies, and responses of pristine ecosystems to extreme
weather events all show the same general patterns of re-
sponse of CH4fluxes to wetting and to drying (Fig. 1d).
Drier conditions reduce CH4emissions, which are completely
suppressed (reductions of ~100%) once the water table is
lowered by 2025 cm, whereas wetter conditions enhance
CH4emissions, often by several orders of magnitude,
although very wet conditions might reduce CH4emissions
once standing water replaces a flooded sedge vegetation
(Juutinen et al. 2003). The water table control on CO2fluxes
is less clear (Fig. 1e). Drier conditions generally lead to in-
creased ER, although in some cases respiration is suppressed
(Aerts and Ludwig 1997) or no change occurs (Chivers et al.
2009; Updegraff et al. 2001; Wickland et al. 2006). GPP and
NEE were less frequently measured than respiration in ma-
nipulation studies. However, the available data show a gen-
eral trend towards reduced GPP under drier conditions and
increased GPP under wetter conditions (Fig. 1e). In combina-
tion with the changes in respiration, this results in an appa-
rent decrease in NEE under drier conditions. Nitrous oxide
emissions are also partially controlled by water table depth
(Fig. 1e), with drier conditions increasing the release of N2O
from peatlands to the atmosphere by several orders of magni-
tude, although the overall emissions generally remain small.
The available full-year C flux data do not show a clear rela-
tionship between annual NEE or annual CH4flux and annual
precipitation (Fig. 2c,d), though the limited data seem to
suggest that high annual CH4flux does not occur where
there is high annual precipitation. This could have a biogeo-
chemical explanation, but also may be because of the limited
sample size.
It is important to note that water table manipulations gen-
erate an abrupt and persistent hydrological change, while cli-
mate change will be more gradual and variable. For example,
a recent study assessing the interannual variability in CO2
emissions based on eddy correlation measurements showed
that bogs and fens respond differently to interannual water ta-
ble depth variability with wetter conditions leading to lower
GPP and ER in fens and higher GPP and ER in bogs,
whereas NEE is not correlated with water table depth (Sul-
man et al. 2010). Lund et al. (2010) did not find any relation-
ship between water table depth and the CO2flux components
when comparing peatland and wet tundra sites from different
climatic settings. The time scale of a study is also a factor
short-term C balance response to hydrological change by ex-
tant plant and microbial communities is likely to be different
than the long-term response as the community composition
and structure adjusts to new environmental conditions (Laiho
2006). Plant and microbe community composition changes
have been observed, especially in the longest-running water
table manipulation experiments (Weltzin et al. 2001, 2003;
Strack et al. 2006; Strack and Waddington 2007; Peltoniemi
et al. 2009) or in studies involving long-term drainage (Mar-
tikainen et al. 1995; Silvola et al. 1996; Talbot et al. 2010;
Yrjälä et al. 2011), where drier conditions generally increase
the importance of the shrubs and decrease the importance of
Sphagnum mosses and sedges. This change in vegetation is
coupled with increased fungal and bacterial biomass (Jaatinen
et al. 2008) and decreased diversity in methanogenic and
methanotrophic communities (Yrjälä et al. 2011).
376 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
2.2. Nitrogen and sulfur deposition, and elevated CO2and
ozone sensitivities
Field studies of enhanced N-deposition on peatlands (either
fertilization studies or deposition gradient studies) have shown
changes in species composition, particularly loss of Sphagnum
mosses (Heijmans et al. 2001; Tomassen et al. 2004; Bubier et
al. 2007), which is considered a serious threat for peatland ca-
pacity to sequester C (Bubier et al. 2007; Limpens et al.
2008). Nitrogen deposition rates greater than 10 kg N ha1·y1
may lead to vegetation change (e.g., Vitt et al. 2003; Bragazza
et al. 2004; Limpens et al. 2008), and N fertilization has led
to increases in vascular plant growth and site productivity in
arctic sites (Shaver et al. 2001; Arens et al. 2008), but some
studies also indicate enhanced decomposition and accelerated
C loss from peatlands due to N input (Bragazza 2006; Gun-
narsson et al. 2008). Sulfur deposition has been shown to re-
duce peatland CH4emissions (Dise and Verry 2001; Granberg
et al. 2001; Gauci et al. 2002; Vile et al. 2003), as sulfate re-
duction is energetically more favorable than methanogenesis.
Reviewing the literature, Gauci et al. (2004) found that peat-
land CH4fluxdecreasedby15%45% for deposition rates of
The effect of elevated atmospheric CO2concentration has
been studied using mini-FACE (free air CO2enrichment) sys-
tems and lab peat incubations. Increasing the concentration to
twice the ambient level both in situ and in the lab resulted in
15%20% increases in CH4emissions from Finnish fens
(Saarnio et al. 1998, 2000; Saarnio and Silvola 1999). In
Sphagnum-dominated mires, the response to increased CO2
was temperature-dependent, with increased CH4emissions in
warm conditions and decreased emissions in cold conditions.
A doubling of the CO2concentration had no effect on the
CH4emission of an incubated core from a British fen,
although ER increased by 285% and N2O emissions in-
creased by 190% from a low base level (Kang et al. 2001).
The effect of elevated atmospheric ozone (O3) concentration
on peatland CO2and CH4emissions has seldom been as-
sessed. Rinnan et al. (2003) found that a 30% increase in O3
Fig. 1. Percent change in annual, seasonal or instantaneous fluxes of CH4,CO
2O in relationship to increased or decreased soil tem-
perature or water table depth, obtained via lab or field manipulations, or extreme weather events. See Table 2 for list of studies. Note that a
positive change in water table position refers to an elevated or shallower water table (if below the surface) or to deeper inundation (if
flooded). For CH4fluxes a percent change less than 100% means a switch from source to sink. CO2flux values represent a net uptake. A
negative percent change in CO2flux means less net emission or greater net uptake; a positive percent change in NEE means more net emis-
sion or less net uptake.
Frolking et al. 377
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
concentration increased ER and decreased NEE in peatland
microcosms, but had no clear effect on CH4emissions. Haa-
pala et al. (2011) found that chronic exposure to elevated O3
had only minor effect on peatland microcosm CO2exchange.
Haapala et al. (2009) concluded that elevated ultraviolet radi-
ation (UV-B; 280315 nm) is not likely to significantly affect
northern fen CO2balance.
2.3. Land use and land cover change
Global peatland area has probably declined by ~50 Mha
(10%15%) over the past 200 years (Bridgham et al. 2006,
2000; Tarnocai 2006; Joosten and Clarke 2002; Maltby and
Immirzi 1993). Peat formation has likely stopped on virtually
all converted peatlands. Excluding peatlands currently used
for agriculture, forestry, or peat harvest (all discussed further
in the following text) many converted peatlands may still be
losing C (as CO2, probably not as CH4) and those abandoned
after use for agriculture are generally sources of N2O (Malja-
nen et al. 2007, 2010a, 2010b).
The dominant land-use activity on remaining northern
peatlands is draining for agricultural use as cropland and pas-
ture and to enhance forest productivity (Table 1), with
smaller areas drained for peat harvest for fuel and other
uses. In southeast Asia, the most rapid rates of deforestation
and land cover change are occurring in the carbon-rich peat
forests and mangroves, where annual rates of deforestation
may be >2% for the former and nearly 8% for the latter
(Langner et al. 2007). Of the 27 Mha of peatland in SE
Asia, 13 Mha (47%) were deforested and mostly drained by
2006 (Hooijer et al. 2010); this was predominantly for con-
version to forest plantations for pulp wood, palm oil planta-
tions, agricultural expansion, and degradation from logging
(Verchot et al. 2010; Langner et al. 2007).
Drainage ditches are cut into the peat to enhance lateral
drainage and lower the water table, but unless the peat layer
is relatively thin, the peat is not fully drained. Typical drain-
age for forestry targets a water table 0.350.55 m below the
surface (Paavilainen and Päivänen 1995); for tropical planta-
tion agriculture target water table depths can be closer to 1 m
(Hooijer et al. 2010). As with the water table sensitivity re-
ported previously, general responses to drainage in managed
peatlands are enhanced decomposition of peat (heterotrophic
respiration) generating CO2(Maljanen et al. 2010b; Hooijer
et al. 2010; Salm et al. 2009), increased N2O emissions, and
reduced CH4emissions or even low rates of CH4uptake
(Maljanen et al. 2010b; Hooijer et al. 2010; Salm et al. 2009;
Inubushi et al. 2003), unless the cropping is paddy rice in the
tropics, in which case the peat soil has been reflooded and
CO2and N2O fluxes may decline and CH4emissions increase
(Furukawa et al. 2005). Long-term subsidence of drained
peats, resulting from an early, rapid consolidation of the peat
followed by persistent decomposition and emission of CO2,
can lower peats by 110 cm per year (Andriesse 1988).
The magnitude of the impact of drainage on greenhouse
gas emissions (see Table 1 for summary values) is influenced
by initial site fertility and water table depth, depth of drain-
age, and post-drainage management (Maljanen et al. 2010b).
Cropland sites that are regularly tilled and fertilized generally
are stronger sources of CO2and N2O than planted grasslands
or forests on drained peatlands (Maljanen et al. 2010b). In
addition, for cropland sites, net vegetation C inputs into the
soil are relatively low because of harvest removals. Peatlands
drained for forest production in northern Europe often have a
net C uptake due to enhanced productivity (until forests are
cut), and can even have net increase in peat C content be-
cause of an enhanced litter production (Minkkinen et al.
Fig. 2. Relationship between total annual CH4and CO2fluxes from different types of peatlands and the annual average temperature or the
total annual precipitation of the site studied (from climate normals reported by the authors). See Table 3 for list of studies. Note that all
annual CO2fluxes are negative (net uptake).
378 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
2002; Hargreaves et al. 2003). In Indonesian tropical forested
peatlands, land-use impacts (logging, conversion to oil palm
or Acacia plantation, conversion to crop or shrubland, or
burning) result in an initial loss of forest biomass of 100
200 t C ha1(Murdiyarso et al. 2010; Hergoualch and Ver-
chot 2011). Relative to virgin tropical peat forests, post-con-
version plant litter inputs are generally lower, losses owing to
fire and enhanced decomposition increase, and CH4emis-
sions generally decline, except for conversion to paddy rice
(Murdiyarso et al. 2010; Hergoualch and Verchot 2011).
In studies from Nordic managed peatlands, there is a gen-
eral inverse relationship between CH4and N2O emissions
if one is high, the other is low (Maljanen et al. 2010b), con-
sistent with basic oxidationreduction biogeochemistry (Li
2007). The mean CH4emission rate from Nordic peatlands
drained for agriculture is 2 kg CH4ha1·y1and, from forests
is 6 kg CH4ha1·y1(Maljanen et al. 2010b). In Indonesia,
net CH4fluxes from a drained forest and a drained deforested
peatland site were insensitive to water table changes below
about 0.2 m depth (Jauhiainen et al. 2008). N2O emissions
from drained peatlands can increase significantly (Fig. 1c),
up to 10100 kg N2ON ha·y1, particularly in tropical and
sub-tropical sites (e.g., Takakai et al. 2006; Terry et al.
1981), although not in all cases (Inubushi et al. 2003; Hadi
et al. 2005). In some cases, this increase can be at least partly
attributed to substantial N additions as mineral fertilizer
(these sites were not included in Fig. 1c).
Annual peat harvest for energy and horticulture is
~108m3·y1(Hood and Sopo 2000). If, at a harvest site, the
amount removed each year is 0.12.0 m (Silvan et al. 2010),
the annual harvest requires 0.0050.100 Mha (Joosten and
Clarke (2002) estimate about 0.001 Mha·y1). Nordic sites
are typically harvested for <10 years (Maljanen et al.
2010b), though this depends on harvest methods, which have
changed over time. Total C released directly from peat har-
vest, assuming all peat C is emitted as CO2in 1 year, and
peat bulk density is 100 kg·m3with 0.5 kg C kg1,is
0.005 Pg C y1. Without active restoration, vegetation coloni-
Table 2. Studies relating CH4,CO
2O fluxes to a change in water table depth or temperature as a result of a direct manipulation or an
extreme climate event (studies where the effects of interacting factors cannot be separated are not considered).
Location Peatland type(s) Study type Reference
Manipulation: Temperature
Ontario and Quebec (Canada) Bog, fen Lab incubations Moore and Dalva (1993)
Minnesota (USA) Bog, fen Mesocosms White et al. (2008); Updegraff et al. (2001);
Bridgham et al. (2008)
Alaska (USA) Rich fen Open top chambers Turetsky et al. (2008); Chivers et al. (2009)
Scotland (UK) Blanket bog Open top chambers Macdonald et al. (1998)
Västerbotten (Sweden) Poor fen Open top chambers Granberg et al. (2001)
Northern Sweden Blanket bog Open top chambers Dorrepaal et al. (2009)
Manipulation: Water table
Ontario and Quebec (Canada) Bog, fen Lab incubations Moore and Dalva (1993)
Wales (UK) Flushed gully mire Lab incubations Dowrick et al. (1999, 2006)
Vechtplassen (Netherlands) Quaking fens Lab incubations Aerts and Ludwig (1997)
Wales (UK) Flushed gully mire Microcosms Freeman et al. (1993)
Alaska (USA) Bog Microcosms Funk et al. (1994)
Colorado (USA) Subalpine fen Microsoms Chimner and Cooper (2003)
Minnesota (USA) Raised bog, fen Mesocosms Bridgham et al. (2008); White et al. (2008);
Updegraff et al. (2001)
Ontario (Canada) Bogs Mesocosms Blodau et al. (2004)
Scotland (UK) Blanket bog Open top chambers Macdonald et al. (1998)
Alaska (USA) Rich fen Open top chambers Turetsky et al. (2008); Chivers et al. (2009)
Wales (UK) Flushed gully mire Exp. drainage Dowrick et al. (1999, 2006)
Bavaria (Germany) Fen Exp. drainage and roofing Goldberg et al. (2010)
Quebec (Canada) Poor fen Exp. drainage Strack et al. (2004, 2006); Strack and Waddington
Quebec (Canada) Bog Exp. drainage Strack et al. (2009)
Minnesota (USA) Bog Exp. flooding Dise et al. (1993)
Ontario (Canada) Bogs Exp. flooding Scott et al. (1999)
Southern Sweden Fens Forestry drainage von Arnold et al. (2005)
Lakkasuo (Finland) Fens, bogs Forestry drainage Minkkinen et al. (1999); Nykänen et al. (1995);
Silvola et al. (1996); Martikainen et al. (1995)
Ontario (Canada) Fen, bogs Forestry drainage Roulet et al. (1993)
Finland Fen Forestry drainage Regina et al. (1996)
Finland Fen Lab incubations Regina et al. (1999)
Sumatra (Indonesia) Peat swamp Forestry drainage Furukawa et al. (2005)
Ontario (Canada) Bog Road drainage Talbot (2009)
Alaska (USA) Lowland Permafrost thaw (flood) Wickland et al. (2006)
Saskatchewan (Canada) Bog Permafrost thaw (flood) Turetsky et al. (2002b)
Central Finland Lacustrine fen Wet year Juutinen et al. (2003)
Frolking et al. 379
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
zation of cutover peatlands is very slow (decades), and the
systems tend to be dry because they were drained before har-
vest (Rochefort et al. 2003), so CH4emissions are low (Tuit-
tila et al. 2000), and CO2emissions are about 10002000 kg
CO2·C ha1·y1(Tuittila et al. 1999); this would be an addi-
tional 106105Pg C y1from the expanding cut-over area
(assuming no restoration). N2O emissions from peatlands
under active peat extraction have been measured at several
sites in Finland; emission rates are relatively low, 0.6 ±
0.6 kg N2ONha
1·y1(Maljanen et al. 2010b); this would
be an additional 106105Tg N2ONy
1from the expand-
ing cut-over area (assuming no restoration). So, the annual
peat harvest area is small enough that changes in gas fluxes
will not likely be globally significant.
Table 3. Studies reporting annual CO2and CH4balance for peatlands (EC, eddy covariance).
Location Latitude (°) Longitude (°) Peatland type Method Reference
Alaska 64.41N 148.19W Bog chambers Wickland et al. (2006)
Michigan 42.27N 84.01W Bogs chambers Shannon and White (1994)
Minnesota 47.32N 93.28W Bog EC Shurpali et al. (1993)
Minnesota 47.32N 93.28W Bog chambers Crill et al. (1988)
Minnesota 47.32N 93.28W Fen chambers Dise et al. (1993)
New Hampshire 43.12N 71.3W Fen chambers Frolking and Crill (1994)
Northwest Territories 61.8N 121.4W Fens chambers Liblik et al. (1997)
Northwest Territories 61.8N 121.4W Bogs chambers Liblik et al. (1997)
Ontario 51.30N 80.28W Fens chambers Roulet et al. (1994)
Ontario 51.30N 80.28W Bogs chambers Roulet et al. (1994)
Ontario 58.45N 94.09W Fens chambers Roulet et al. (1994)
Ontario 58.45N 94.09W Bogs chambers Roulet et al. (1994)
Ontario 45.41N 75.48W Bog chambers Roulet et al. (2007)
Quebec 45.33N 73.2W Bogs chambers Moore and Knowles (1990)
Quebec 54.48N 66.49W Fens chambers Moore and Knowles (1990)
Quebec 53.38N 77.43W Bog chambers Pelletier et al. (2007)
Ireland 51.55N 9.55W Blanket bog chambers Laine et al. (2007)
Finland 69.8N 27.16E Fen (aapa mire) EC Hargreaves et al. (2001)
Finland 61.50N 24.12E Fen EC Rinne et al. (2007)
Finland 67N 27E Fen chambers Huttunen et al. (2003)
Scotland 58.23N 3.40W Blanket bog EC Hargreaves and Fowler (1998)
Scotland 55.04N 4.25W Bog chambers Clymo et al. (1995)
Scotland 56.53N 2.32W Bog chambers Chapman and Thurlow (1996)
Sweden 63.44N 20.06E Bog chambers Waddington and Roulet (2000)
Sweden 68N 19E Bog chambers Svensson (1980)
Sweden 68N 19E Fen chambers Svensson (1980)
Sweden 68.28N 19.03E Fen EC Jackowicz-Korczyński et al. (2010)
Sweden 57.08N 14.45E Fens chambers von Arnold et al. (2005)
Slovenia 45.59N 14.30E Bog chambers Danevcic et al. (2010)
Siberia 57N 83E Bog chambers Panikov and Dedysh (2000)
Siberia 56.50N 82.49E Fen chambers Glagolev and Shnyrev (2008)
Siberia 56.51N 82.58E Bog chambers Friborg et al. (2003)
Malaysia 2.49N 111.51E Peat swamp forest chambers Melling et al. (2005)
Indonesia 2.20S 113.55E Peat swamp forest chambers Jauhiainen et al. (2005)
Indonesia 3.25S 114.41E Peat swamp forest chambers Inubushi et al. (2003)
Indonesia 2.12S 113.55E Peat swamp forest chambers Hirano et al. (2009)
Japan 36.55N 139.12E Bog chambers Inubushi et al. (2005)
China 35.39N 98.48E Bog chambers Jin et al. (1999)
Ontario 45.51N 75.52W Bog EC Roulet et al. (2007)
Saskatchewan 54.95N 112.57W Treed fen EC Syed et al. (2006)
Ireland 51.55N 9.55W Blanket bog EC Koehler et al. (2011)
Finland 69.08N 27.17E Fen EC Aurela et al. (2004)
Finland 61.49N 24.11E Fen EC + chambers Aurela et al. (2007)
Sweden 64.11N 19.33E Fen EC Sagerfors et al. (2008)
Sweden 56.27N 13.55E Bog EC Lund et al. (2007)
Sweden 68.22N 19.03E Palsa mire auto chambers Bäckstrand et al. (2010)
Siberia 56.51N 82.58E Bog chambers Friborg et al. (2003)
Thailand 6.4N 101.56E Peat swamp forest EC Suzuki et al. (1999)
380 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
Other land uses that impact peatlands include flooding by
artificial reservoirs (e.g., hydropower development), draining
for sub-peat mining activity, and linear disturbances that can
disrupt peatland hydrology (e.g., roads, pipelines, seismic
survey lines). The extent of these impacts is generally not
well-quantified, nor are the impacts on greenhouse gas emis-
sions; however, see Maljanen et al. (2010b) and Teodoru et
al. (2011) on the impacts of reservoir inundations. We as-
sume that, in the global aggregate, these areas and associated
greenhouse gas emissions are small compared to agriculture,
forestry, and deforestation on peatlands.
Based on the areas and fluxes summarized in Table 1, the
overall impact of land use activity on the global peatland
greenhouse gas budget, relative to the undisturbed calculation
presented in Section 1, is to switch the CO2balance from net
sink to net source of about 0.1 Pg C y1, reduce global peat-
land CH4emissions by ~10% (i.e., by ~0.005 Pg CH4y1),
and increase N2O emissions by a factor of 20, but only to
0.0004 Pg N y1, about 2% of total global N2O flux.
2.4. Fire and permafrost thaw
The past decades have seen a substantial increase in tropi-
cal peatland wildfires (Flannigan et al. 2009). Major fire
years in the tropics are associated with dry weather (van der
Werf et al. 2008), but it is the synergism of land use and
drought that promotes fires in tropical peat forests (Langner
et al. 2007; Page et al. 2002). Uncontrolled fires in tropical
peatlands emit 100400 Mg C ha1from burning peat and
up to 150 Mg C ha1from vegetation biomass (Couwenberg
et al. 2010; Murdiyarso et al. 2010). Peatland fires in Indone-
sia in 1997, a major fire year, are estimated to have released
12 Pg C, with about 80% of the release from burned peat,
and about 20% from burned aboveground forest biomass
(Page et al. 2002).
Major fire years in boreal peatlands are also associated
with warm, dry conditions (Turetsky et al. 2004). In boreal
and sub-arctic North America, peatlands experience a fire re-
turn interval of 75300 years (Zoltai et al. 1998), with drier
peatlands (bogs) experiencing more frequent fires and larger
C losses than wetter peatlands (fens). In continental western
Canada (Alberta, Saskatchewan, and Manitoba), peatlands
are on the dry limit of the boreal peatland domain (Vile et
al. 2011), and are burned by wildfires with a fire return inter-
val of 100150 years (Turetsky et al. 2004). In more humid
northern Europe, a fire frequency interval of <300 years was
found in peat cores in Estonia and Finland (Sillasoo et al.
2007; Väliranta et al. 2007). Wildfire emissions from boreal
peatlands are estimated at 3.1 Tg C y1(Benscoter and
Wieder 2003), with combustion losses in peatlands averaging
32± 4 Mg C ha1per fire event (Turetsky et al. 2004).
Biomass burning (e.g., forest and peatland fires) can be a
significant direct source of CH4to the atmosphere (Kasischke
and Bruhwiler 2003). Levine (1999) estimated C-gas emis-
sions from the large fires of Southeast Asia in 1997 as ~1%
CH4C (about 15% as COC, the rest as CO2C). This is
roughly consistent with estimates of combustion emissions
of ~10 Tg CH4(van der Werf et al. 2004; Bousquet et al.
2006). Takakai et al. (2006) measured similar N2O emissions
from burned and unburned tropical peat forests in 1 year, and
lower emissions from the burned forest in another year, and
concluded that the effect of forest fire on N2O fluxes
was neither consistent nor clear.[p. 672]
The net impacts of wildfire on peatland C emissions de-
pends on post-fire response of hydrology and vegetation. Bor-
eal fire scar soils tend to be wetter than the surrounding
landscape, probably due to decreased evapotranspiration rates
and thawing of permafrost (Bourgeau-Chavez et al. 2007);
plot-level studies of the vegetation communities following
wildfire also show an increase in species tolerant of wet condi-
tions (Benscoter et al. 2005; Sillasoo et al. 2011), perhaps
caused in part by a reduction in microtopography in bogs (Tol-
onen 1983). Wetter post-fire soil conditions have been corre-
lated with increased CH4emissions from burned sites
compared with net CH4uptake in unburned sites in Eastern
Siberia (Nakano et al. 2006). In Borneo, tropical fire scar soils
were drier in the dry season, likely caused by increased solar
loading and soil evaporation (Siegert and Ruecker 2000). This
should have an impact on net changes in CH4and N2Oemis-
sions, but there are limited data sets to quantify this impact.
Vegetation recovery partly determines the net effect of
wildfire on C cycling, but fire frequency is also key. Two
years after wildfire, burned peatland sites in Alaska had
lower rates of NEE and ER compared with unburned peat-
land sites because of a reduction in vegetation cover (ODon-
nell et al. 2009). Post-fire vegetation recovery has been found
to occur within 1520 years in boreal North American peat-
lands (Zoltai et al. 1998; Benscoter and Vitt 2008). In a
study in northern Europe, the time-scale of post-fire vegeta-
tion succession varied between some tens of years (weak
fires) to 250 years (severe fires), where after an initial wet
phase, the succession led towards a dry hummock commun-
ity (Sillasoo et al. 2011). A post-fire chronosequence study of
ombrotrophic treed bogs in Alberta by Wieder et al. (2009)
determined that sites were a C source for ~13 years following
a fire until the moss and shrub layers had recovered, and then
were an increasing C sink over the subsequent 60 years as
black spruce biomass accumulated below- and above-ground,
then declined to a weaker sink at about 100 years post-fire as
tree biomass stabilized. Integrated over an entire fire cycle,
the net C flux was small, so over a large region the impact
of fires on net peatland C balance will be most sensitive to
changes in fire frequency or intensity (Turetsky et al.
2002a). For fires associated with land use in the tropics, veg-
etation recovery depends on the subsequent land use, and
may not occur if the land is converted to another use.
Peatland organic soils serve to insulate permafrost from a
warming atmosphere and land surface, particularly drier om-
brotrophic bogs, and so permafrost thaw is generally slower
when thick organic soils are present (Lawrence et al. 2008;
Yi et al. 2007; Treat 2010; Wisser et al. 2011). In Canada,
more than one-third of peatlands are underlain by permafrost
(Tarnocai 2006). Smith et al. (2007) estimated that about
one-third of northern peatlands are in zones of continuous
permafrost, with another 40% of northern peatlands in dis-
continuous, sporadic, and isolated permafrost zones. Perma-
frost thaw has both direct and indirect impacts on
greenhouse gas emissions. In addition to a general impact of
warming (discussed previously), permafrost thaw has a direct
impact on net greenhouse gas emissions because thawed peat
is more susceptible to decomposition, and thus loss as CO2
and (or) CH4, than frozen peat (Dorrepaal et al. 2009). Indi-
Frolking et al. 381
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
rect effects include (i) higher water table in surface collapse
(thermokarst) than in pre-thaw (Zoltai 1993; Vitt et al. 1994;
Jorgenson and Osterkamp 2005; Johansson et al. 2006); (ii)
possible changes in watershed flowpaths that can lead to dry-
ing, though this has been observed more in lakes than peat-
lands (Yoshikawa and Hinzman 2003; Smith et al. 2005;
Riordan et al. 2006); (iii) changing vegetation composition
because of changes in site wetness, particularly changes in
abundance of sedge or other vegetation that can enhance
CH4emissions from negligible flux on permafrost palsas to
~0500 kg CH4ha1·y1in collapse features (e.g., Turetsky
et al. 2002b; Johansson et al. 2006; Wickland et al. 2006);
and (iv) enhanced C sequestration (or reduced loss) rates by
~3001000 kg C ha1·y1(Turetsky et al. 2007; Camill et al.
2001; Johansson et al. 2006; Wickland et al. 2006). Note that
Schuur et al. (2009) found, for a permafrost upland tundra
site, that enhanced summer productivity (C uptake) was off-
set by thawing and decomposition of deeper, older soil or-
ganic matter, to the extent that the site with deepest and
warmest thawing was a net C source. For peatlands, the im-
pact of deeper thawing on the C balance will be very depend-
ent on the depth of the thawed system water table.
3. Climate and global change impacts in the
21st century
We do not anticipate globally significant climate-driven
changes in peatland area in the 21st century. Rates of terres-
trialization and paludification are too slow (Heathwaite 1993;
Kuhry and Turunen 2006), although a small increase in peat-
land area because of thermokarst in former forests is likely
(Jorgenson et al. 2001). Rates of peatland disappearance (e.
g., due to climatic drying) are not well-quantified, but also
likely too slow to be globally significant in the next
100 years; rising sea levels may inundate low-lying coastal
peatlands, but the peatland area affected is not likely to be
globally significant within 100 years. Therefore, any large
impact of peatlands on the climate system in the 21st century
will be because of changes in their greenhouse gas balances,
through (1) direct climate change impacts; (2) indirect im-
pacts via permafrost thaw and wildfire; and (3) anthropogenic
impacts land use, land conversion, and pollution loading.
3.1. Direct climate change impacts on peatland net CO2
The data presented in Section 2 indicate that peatland
greenhouse gas emissions are more sensitive to changes in
water table depth than to changes in temperature. Conversely,
confidence in predictions of temperature changes are higher
than in predictions of precipitation changes (e.g., Meehl et
al. 2007), and changes in peatland temperatures will be pri-
marily a function of changes in air temperature (but also
peat water content, and, in the north, winter snowpack),
while changes in peatland water table depth will be a func-
tion of changes in precipitation, temperature, evapotranspira-
tion rates, vegetation composition, and landscape hydrology.
So, confidence in predictions of changes in temperature (or
at least its sign) are higher than predictions of changes in
water table depth (or even its sign).
Long-term C accumulation rates appear to be higher on
average in the low-elevation tropics than at high latitudes
(Page et al. 2011; but see Yu et al. 2010), but probably by
no more than a factor of two for a mean annual temperature
difference of 1030 °C. In the West Siberian Lowlands, Beil-
man et al. (2009) showed that non-permafrost sites in the
southern portion have accumulated C at about twice the rate
of permafrost sites in the northern portion (roughly 20 versus
10 g C m2·y1). On the other hand, current CO2uptake rates
do not show a relationship to mean annual temperature
(Fig. 2b), and eddy flux tower data do not show any strong
relationship between NEE and temperature. We conclude
that, if the current net uptake by undisturbed peatlands is
~0.1 Pg C y1(see Secion 2), then direct climate warming
impacts will increase this uptake by <0.1 Pg C y1, not large
enough to play a major role in the 21st century climate sys-
tem. A possible exception to this would be destabilization of
peatlands and rapid and sustained C loss. However, both the
weak and uncertain temperature dependence of CO2fluxes
(Fig. 1b) and the presence of peatlands in regions with mean
annual temperatures from <5 °C to >25 °C indicate that de-
stabilization is not likely to be a direct response to temperature;
destabilization could arise from temperature-induced permafrost
thaw, or from temperature- and precipitation-induced increases
in drought and (or) fire frequency and intensity, which are dis-
cussed in the following.
Short-term (less than a decade) water table manipulations
induce changes in net CO2flux (e.g., typically higher ecosys-
tem respiration and reduced GPP and NEE with drying; data
are more ambiguous for wetting; Fig. 1e), but over the longer
term several factors may override these short-term impacts.
First, and most importantly, the vegetation composition will
shift towards plant functional types that are better suited to
drier conditions (e.g., shrubs and trees) or wetter conditions
(e.g., sedge establishment in permafrost collapse features),
likely reversing any decline in NPP or even increasing NPP
above predisturbance rates (Minkkinen et al. 2002; Malmer
et al. 2005). Second, peat mineralization rates may decline
as the readily decomposable fraction is mineralized (as has
been seen in upland soil warming plots, see Davidson and
Janssens 2006), though it is difficult to predict how long this
will take (Dorrepaal et al. 2009). At the same time, changes
in vegetation composition will influence the tissue quality
and decomposability of new litter inputs into the peat. Third,
ongoing peat mineralization, particularly if it does not de-
cline, may lead to nutrient (e.g., N and P) mineralization, en-
hancing NPP to offset ongoing decomposition. Also post-
drainage subsidence and the compaction of surface peat can
make available nutrients from deeper stores (Laiho and Laine
1994). This has been observed, for example, in drained, treed
peatlands that are sufficiently minerotrophic to support in-
creased NPP (Minkkinen et al. 1999; Laiho et al. 2003).
However, much less is known about the ecological response
of tropical peatlands to drought. It has been established that
there is a synergism of drought and land use resulting in
greater areas of tropical peatlands burned during ENSO (El
Niño/La NiñaSouthern Oscillation)-mediated droughts
(Page et al. 2002), but the response of undisturbed tropical
peatlands to drought is not well studied.
We conclude that, although climate change thresholds to
peatland stability must exist, 21st century climate change im-
pacts on the net CO2balance of peatlands not experiencing
land use will likely be small relative to anthropogenic pertur-
382 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
bations to the global C cycle. Some peatlands will probably
show a change towards less C uptake or more C loss, and
others the opposite. These will be important to quantify if
ecosystem C becomes an economic commodity, but in the
global aggregate, they will not have a major impact on the C
budget. It is important to note, however, that 21st century
warming will be rapid and large relative to changes over the
past several millennia, and thus peatlands may behave in
ways that are not evident in either the contemporary or the
paleo record. This assessment also ignores potential impacts
of permafrost thaw, changes in natural disturbance rates, and
anthropogenic impacts (all discussed in the following).
3.2. Direct climate change impacts on peatland net CH4
We can estimate warming impacts on CH4flux in two ways.
Manipulation studies with temperature increases <10 °C all
show increases in CH4flux of ~10%80%, depending on site
water table, but the increase is not well correlated with the
amount of warming (Fig. 1a). Annual CH4flux does not corre-
late well with mean annual air temperature (Fig. 2a), but is
generally low at very cold sites and very warm sites (the warm
site conclusion may be a consequence of limited sampling, but
also may reflect the high lignin content of tropical peat provid-
ing a poor substrate for methanogens (Couwenberg et al.
2010)) These data indicate that warming impacts on CH4
fluxes are likely to be significant for high latitude wetlands
(those with contemporary mean annual temperatures <0 °C),
but do not indicate that warming in sub-boreal, temperate, or
tropical peatlands will necessarily cause significantly enhanced
CH4emissions, particularly in bogs.
In considering water table depth changes, we assume that
water tables will not rise substantially above the peat surface
(i.e., inundated area may increase, but the area with surface
ponding greater than a few cm will not increase substan-
tially). Thus, peatland water table increases are bounded by
contemporary water table depths; these are typically~20
40 cm for bogs, and 015 cm for fens (e.g., Zoltai et al.
1998). It is also important to bear in mind that a large and
persistent change in water table depth will likely lead to
changes in plant functional types, which may have a large
impact on CH4emissions (e.g., increasing or decreasing
sedge presence in a peatland will almost always lead to a sig-
nificant change in CH4emissions). Because we cannot pre-
dict peatland water table changes with confidence, we
cannot predict with confidence how peatland CH4emissions
might change due to 21st century climate change. We, there-
fore, pose an alternative question what area of peatlands
would need to experience what magnitude of water table
change to have a large impact on global peatland CH4emis-
sions? If all peatlands were to experience a 10 cm rise in
water table, CH4emissions would likely increase substan-
tially (~100%) from all but the wettest sites. Since the wettest
sites are typically strong CH4sources, we estimate an overall
increase in CH4flux of ~50%. If all peatlands were to expe-
rience a 10 cm drop in water table, CH4emissions would
likely decrease substantially (~50%100%) from all but the
driest sites. Because the driest sites are typically negligible
CH4sources, we estimate an overall decrease in CH4flux of
~75%. The total global CH4flux to the atmosphere is ~0.6
Pg CH4y1(e.g., Mikaloff-Fletcher et al. 2004), and peat-
lands may contribute 0.04 Pg CH4y1, so a significant
change in global CH4flux (>5% or > 0.03 Pg CH4y1)
would require a large change in peatland flux (>60%). Thus,
based on available data, we conclude that a globally signifi-
cant change in aggregate peatland CH4flux will require ei-
ther virtually all peatlands to get wetter or virtually all
peatlands to get drier.
3.3. Direct climate change impacts on peatland net N2O
The USEPA (2010) reported natural wetland N2O emissions
as a negligiblefraction of global natural-source emissions of
1. We estimate a net global undisturbed
peatland N2O flux of ~0.00002 Pg N2ONy
global N2O emissions from natural sources. However, we do
not have high confidence in this estimate, due to data limita-
tions both for tropical sites and for year-round measurements
at high latitudes, where spring thaw emission pulses may be
significant. Direct temperature impacts are not likely to have
as strong an impact on N2O emissions as water table change
impacts; water table manipulation data indicate that drying
could lead to large percentage increases to what are currently
low emission rates. The potential increase in N2O emissions
from direct climate change impacts is not likely to be signifi-
cant relative to the global N2Obudget.
3.4. Indirect climate change impacts fire and
permafrost thaw impacts
In northern regions, climate change is expected to lead to
increased fire frequency (Balshi et al. 2009), a longer fire
season (Flannigan et al. 2009), and to permafrost thaw (e.g.,
Lawrence and Slater 2005; Zhang et al. 2008). Boreal/subarc-
tic peatland ignition generally results from spreading of adja-
cent upland fires (e.g., Pitkänen et al. 1999, 2001; Zoltai et
al. 1998), so peatland fire frequency is also likely to increase.
In western Canada, a shortening of the fire return interval
to <60 years may lead to the change of fire-prone boreal
peatlands from net C sinks to net C sources (Wieder et al.
2009). In Alaska, there has been an increase in burn area
and the number of late season fires, but burn depth in flat
wet sites (peatlands) does not seem to have been affected by
this (Turetsky et al. 2010). In the tropics, fire frequency is
strongly and non-linearly correlated with drought intensity;
part of this strong relationship is a result of enhanced clear-
ing rates and accidental fires during dry periods (van der
Werf et al. 2008). Given the population and development
pressures, which are the main ignition sources in tropical
peat forests, any increase in drought severity and (or) fre-
quency is likely to result in increased fires in tropical peat-
lands, with potential for large CO2emissions and hence
greater C emissions from tropical peatlands associated with
global climate change. Biomass burning CH4emissions (cur-
rently estimated at 1% of global CH4fluxes; USEPA 2010)
will increase with increased fire frequencies. Wildfires in
boreal regions are likely to lead to increased CH4emissions
relative to low pre-fire bog emissions through increases in
soil moisture and a change in species composition (Nakano
et al. 2006; Tuittila et al. 2007).
Permafrost thaw impacts on peatland net C balance are
likely to be an increase in C sequestration rates during the
21st century, but CH4emissions are likely to also increase.
Frolking et al. 383
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
Several shallow core studies indicate that changes in C se-
questration rates can persist for more than 100 years (Robin-
son and Moore 2000; Camill et al. 2001); there is not
equivalent information about CH4flux impacts, but C se-
questration rates are influenced by vegetation composition
and moisture, so we assume CH4flux changes can also per-
sist for 100 years. To generate a rough estimate of the impact
of permafrost thaw in the 21st century, we assume a constant
decline in near-surface permafrost area over the next
100 years resulting in a reduction of 10 million km2(Law-
rence et al. 2008), and that 10%20% of that area has peat-
lands (Smith et al. 2007; Tarnocai 2006). If we further
assume that thaw results in an increase in C sequestration of
3001000 kg C ha1·y1and in CH4emission of 0500 kg
CH4ha1·y1, and that these effects persist for 100 years
post-thaw, we estimate that by the end of the 21st century a
net increase in C uptake of 0.03 0.1 Pg C y1, and a net
increase in CH4flux of up to 0.05 Pg CH4y1.
3.5. Anthropogenic impacts land use, land conversion,
and pollution loading
The trajectory of land use pressures on peatlands resem-
bles that of the worlds forests: the temperate zone was heav-
ily perturbed in the 19th and 20th centuries, while the tropics
have been under intense land use pressure for the last few
decades, and the boreal zone is relatively unperturbed (except
in northwestern Europe). We do not anticipate that land-use
and harvest pressures on northern peatlands are likely to in-
crease or decrease substantially in the coming decades. These
land use impacts on greenhouse gas fluxes (see Table 1) rep-
resent a climate system perturbation relative to unmanaged
peatlands, but not a significant change from emission rates
of the past few decades from northern peatlands. Possible ex-
ceptions to this static future include peatland inundation due
to reservoir construction, and mining of abundant oil/tar
sands in peatland-rich northern Alberta (Tarnocai et al.
2000; Timoney and Lee 2009); the current industrial foot-
print of this mining is only about 0.07 Mha, but has quad-
rupled in the past 20 years (Schindler 2010). The climate
impact of using peat as a fuel is comparable to fossil fuels
(Kirkinen et al. 2007, 2010), so there is not likely to be a
strong climate incentive for increased peat harvest for use as
a biofuel, though other economic factors may prevail. The
peatland area undergoing restoration/reclamation is not well
known, but Höper et al. (2008) estimate that ~0.1 Mha of
northern peatlands may be under restoration by 2025. We as-
sume that this area is not big enough to have a significant
impact on global peatland greenhouse gas fluxes.
Hooijer et al. (2010) estimate that current CO2emissions
from drained tropical peatland soils in Southeast Asia are
~0.10.2 Pg C y1, with the majority from Indonesia; this sit-
uation differs from land cover change over mineral tropical
soils, where little soil C is lost (Kauffman et al. 2009). Simi-
larly, Miettinen and Liew (2010) estimate annual emissions of
~0.08 Pg C y1from peatlands subjected to land use in Penin-
sular Malaysia and the Islands of Sumatra and Borneo. Hooijer
et al. (2010) predict that CO2emissions from these drained
peatlands will continue to increase for another decade (by
~10%) as more peatlands are drained for use, then slowly de-
cline over the 21st century as the shallower peat deposits are
fully oxidized. They identify a number of areas of uncertainty
in their calculations due to limited field studies and data. Ac-
cording to a 2005 Indonesian Government report, Indonesia
has plans for significant expansion of food production, palm
oil plantations, and timber and pulpwood plantations (Verchot
et al. 2010). To the extent that these land uses occur on tropi-
cal peatlands, the net C emissions will be ~600000 kg CO2C
ha1over a 50-year production rotation (Verchot et al. 2010) or
~10 000 kg C ha1·y1; establishment of plantations on previ-
ously degraded grasslands, conversely, would result in a net
uptake of ~20,000 kg CO2-C ha1over a 50-year production
rotation. Fire is associated with land conversion as a manage-
ment tool (e.g., Murdiyarso and Lebel 2007); fire emissions
are highly variable, and generally are much higher in dry/
drought years (Murdiyarso and Adiningsih 2007).
The IPCC Special Report on Emissions Scenarios (SRES)
emission scenarios projections of sulfur emissions trajectories
in the 21st century range from a mid-century peak at ~2
times present rates declining back to ~1 times present rates
to a slow decline to about 0.2 times present rates (Nakice-
novic and Swart 2000). For a smaller set of scenarios, Den-
tener et al. (2006) projected decreases in S deposition in
Europe and North America by 2030, and large increases in
southern and eastern Asia and South America. Gauci et al.
(2004) estimated that by 2030 sulfur deposition (pollution)
may be sufficient to reduce annual northern peatland CH4
emissions by ~5 Tg CH4y1. Dentener et al. (2006) also pre-
dicted changes in reactive N deposition in 2030: increase by
5%20% in northern North America, approximately constant
in Fenno-Scandia and Russia, decrease by 5%20% in Eu-
rope, increase by 50%200% in tropical SE Asia, and remain
approximately constant in Amazonia. Persistent elevated dep-
osition rates of reactive N deposition rates are very likely to
change the vegetation composition of peatlands, with declines
in Sphagnum and increases in vascular plants (Limpens et al.
2009; Vitt et al. 2003); the consequences of this for long-
term changes in peatland C sequestration rates, and CH4and
N2O emission rates are not well known.
4. Discussion and conclusions
4.1. Global change impacts on peatland greenhouse gas
Over the past century or two the dominant global change
impact on peatlands has been land use pressures. These gen-
erally lead to a net CO2loss, except for draining northern
peatlands for forestry, with overall net emissions of ~0.20.3
Pg C y1; this likely has more than offset the net uptake of all
undisturbed peatlands, making global peatlands overall a net
CO2source (Table 4). Land-use pressure generally leads to
reduced CH4emissions, probably by about 510 Tg CH4y1
to date, or about 10% of total undisturbed peatland flux and
1% of total global flux to the atmosphere. Note that paddy
rice on peat soils is ignored here, because it is likely included
in the rice agriculture category of the CH4budget. Land-use
pressure generally leads to increased N2O emissions, prob-
ably by about 0.5 Tg N2ONy
1, or up to about 10x the total
undisturbed peatland flux and 3% of total global flux to the
atmosphere. These values can all be expected to increase over
the first decades of the 21st century, due to ongoing or in-
creasing land use pressures in the tropics and persistent land-
use rates in northern peatlands.
384 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
In the context of the global climate system, climate change
impacts on peatland greenhouse gas emissions represent a
feedback. Direct feedbacks could be small and positive
(warming), small and positive or negative (heterogeneous
wetting/drying), or moderate and positive or negative (wide-
spread wetting or widespread drying). Indirect feedbacks op-
erating through permafrost thaw and changes in wildfire
frequency are likely to be small to large and positive in the
short term (permafrost thaw causing enhanced CH4emis-
sions), declining and perhaps becoming negative over the
longer term (permafrost thaw causing increased CO2uptake),
or small to moderate and positive (increased wildfire fre-
quency, permafrost thaw causing net CO2emissions), though
this would diminish over decades following a restabilization
of wildfire frequency.
Climate forcing is very likely to be strongest in the boreal/
sub-arctic, anthropogenic forcing is very likely to be stron-
gest in the tropics, and pollution loading strongest in the tem-
perate zone, at least for the next several decades. For CH4
emissions, the biggest potential increase comes from perma-
frost thaw, but the magnitude of this is quite uncertain. Most
other changes are likely to reduce the aggregate peatland CH4
flux (drier conditions, increased S-deposition) or increase it
only slightly (warming). For CO2emissions, the biggest po-
tential loss comes from land use pressure and fire in the
tropics, unless there is a major expansion in peatland cultiva-
tion or harvest in the boreal zone. Net emissions from boreal
fires will likely increase as well. The biggest potential in-
crease in CO2uptake is likely to be due to high latitude
warming. For N2O emissions, land use and management ac-
tivity will determine the outcome.
While peatlands will influence the Earths climate in the
21st century, they are not likely to be a climate-system tip-
ping element as defined by Lenton et al. (2008). The largest
likely impacts rapid peatland degradation following
draining in the tropics and perhaps permafrost thaw CH4
emissions will probably be less than current anthropo-
genic emissions. Contemporary and paleo data indicate that
the more general response of global peatlands to climate
change is likely to be small, though the change in net green-
house gas emissions may persist for centuries. However,
there are several reasons why these data may underestimate
the impact: (1) it is possible that strong responses will result
from interacting impacts of multiple forcings that can over-
come some of the internal (negative) feedbacks in peatlands
that contribute to peatland stabilization or self-regulation; (2)
the rate of climate change may be faster than the systems
have had to adapt to in the past, and perhaps faster than they
can adapt to, though manipulation studies such as warming
or draining represent abrupt changes; (3) new climate states,
particularly in the north, will be outside the range of Holo-
cene climates, and how peatlands will respond to novel cli-
mates is not well known; and (4) the observational data are
inadequate to show strong impacts perhaps the records
are too short in the case of the contemporary flux data, or
too incomplete in the case of the paleo data. At this time,
there has been only limited process-based modeling to ad-
dress this issue. One modeling result indicates that warming
could cause significant C loss over a century or two (Ise et
al. 2008), while another model indicates that under long-
term drying peatland C stocks may be relatively stable
(Frolking et al. 2010). Neither model fully represents the
complexity of peatland ecosystem responses to climate forc-
ing or other perturbations, nor the range in peatland types
that will be responding.
While the global impact of changes in peatland greenhouse
Table 4. Current greenhouse gas annual flux rate and emission estimate for global undisturbed and disturbed peatlands (excluding distur-
bances due to natural causes). The pre-industrial value assumes an additional 60 Mha of non-tropical peatlands, which have been drained and
lost over the past few centuries (see text). Flux rates are approximate central values (see Table 1 and Fig. 2). Tropical disturbed CH4flux
does not consider the high fluxes from paddy rice as the area involved is uncertain, and that flux is often included in agricultural greenhouse
gas accounting. Uncertainties are large due to limited sampling of gas fluxes across the diversity of peatland types and settings, and due to
inadequacies in global mapping of peatland areas (see Section 4.3), so all values are reported to only one significant figure.
Flux rates Emissions
Peatland category Area
(Mha) kg CO2-C ha1·y1kg CH4ha1·y1kg N2O-N ha1·y1Pg CO2Cy
1Pg CH4y1Pg N2ONy
Non-tropical 400 400 +100 +0.05 0.2 0.04 0.00002
Tropical 60 500 +20 +0.05 0.03 0.001 0.000003
460 -0.2 0.04 0.00002
Contemporary undisturbed
Non-tropical 300 400 +100 +0.05 0.1 0.03 0.00002
Tropical 40 500 +20 +0.05 0.02 0.0008 0.000002
Total undisturbed 340 -0.1 0.03 0.00002
Contemporary disturbed
Non-tropical crop 30 +5000 +2 +5 0.2 0.00006 0.0002
Non-tropical /
non-crop 10 2000 +7 +3 0.02 0.00007 0.00003
Tropical 20 +6000 +5 +10 0.1 0.0001 0.0002
Total disturbed 60 0.2 0.0002 0.0004
400 0.1 0.03 0.0004
Frolking et al. 385
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
gas fluxes may be small in the 21st century relative to other
anthropogenic emissions, in regions or countries with sub-
stantial peatland area and peat C stocks, these impacts will
not be small, and comprehensive national greenhouse gas re-
porting will require better quantification of peatland emissions.
4.2. Temporal considerations
Predicting 21st century climate change impacts on peat-
land greenhouse gas emissions requires discriminating be-
tween, and then comparing, short-term and long-term
impacts. Laiho (2006) suggests that highly variable peatland
responses to lowered water levels across a range of studies
may be due, in part, to differences in observation time scales.
Short-term changes in an environmental control such as water
level represent a disturbance to a system that was in relative
equilibrium with previous conditions, while long-term
changes result in adaptations to a new regime. Laiho (2006)
proposes that although disturbed systems will almost always
lose C in the short term, the long-term response can be
highly variable. Experimental manipulations typically last
only a few to perhaps 10 years, and so do not generally ob-
serve long-term responses. Other studies have used natural
spatial gradients, or manipulations-of-opportunity(e.g.,
long-term drainage for forestry) as analogs to changes over
time since these can potentially include long-term environ-
mental responses (though any short-term effects may not
have been observed and may no longer be detectable).
Do manipulations really represent climate change forcings?
This is particularly important for water table manipulations
i.e., is anthropogenic draining different than climatic drying?
If drying is gradual (e.g., either due to a warming-induced in-
crease in evapotranspiration with little change in water inputs
or a gradual downward trend in precipitation) peatland eco-
systems might slowly adjust to the change, without being
strongly destabilized, and impacts would be minor. For exam-
ple, in simulations with a long-term peat accumulation
model, decade- to century-long periods of reduced net pre-
cipitation resulted in a gradual loss of peat (more years with
negative NEE, but not years with large losses), so that net
peat accumulation slowed (or stopped or reversed) and the
water table depth did not drop substantially (Frolking et al.
2010). However, if the drying were the result of a multi-year
severe drought, with a rapid and large drop in water table, the
peatland might be destabilized, with larger impacts on the C
balance and a shift to a new state(e.g., Hilbert et al. 2000).
Even this new state could lead to stable peat accumulation,
however, as has happened in northern peatlands drained for
forestry. Therefore, it is important to identify the thresholds
of overall peatland stability. What climate or anthropogenic
forcing will sufficiently destabilize peatlands such that they
will (eventually) lose most or all of their stored peat C?
While peatlands are widespread, they are not present every-
where, and there is evidence in paleo-records from central
North America of peatland presence (inferred from Sphagna
spores) where no peatlands now exist (Halsey et al. 2000).
A challenge for large-scale modeling and analysis is ad-
equately representing both very small-scale spatial heteroge-
neity that can have an important impact on peatland C
cycling and greenhouse gas emissions (e.g., microtopogra-
phy, fire intensity, thaw collapse features), and temporal dy-
namics on decade to century scales (e.g., transitions in
vegetation composition, meso-scale spatial reorganization)
that will be a central factor in 21st changes in the role of
peatlands in the Earths climate system. These two scales
play an interacting role in some of the most uncertain antici-
pated changes (e.g., permafrost thaw collapse features, fire).
As some of the results reported in this review are based on
short-term, controlled experiments that do not take into ac-
count longer term and potentially cumulative impacts of
global change, there is a possibility that we underestimate
the extent of the changes that might affect the role of peat-
lands in the climate system.
Due to their different lifetimes in the atmosphere (~10 years
for CH4, ~110 years for N2O, and ~100200 years or longer
for CO2; Forster et al. 2007), short- and long-term impacts on
CO2or N2O emissions will overlap in the atmosphere, while
short- and long-term impacts on CH4emissions may result in
different short- and long-term climate impacts. Additional
blending of short- and long-term impacts will come from the
slow migration of permafrost thaw across the sub-arctic and
arctic so that at any given time the north will be a mix of
landscapes that are warming, thawing, recently thawed, and
decades post-thaw, and from the stochastic nature of low-
and high-intensity fire years.
4.3. Uncertainties and conclusion
Peatlands are generally understudied compared to other
major terrestrial ecosystems (forests, grasslands, croplands).
This is particularly true for tropical peatlands, where there
have been very few multi-year studies to quantify temporal
variability or trends. Only very preliminary work has been
done in African and South American peatlands, and at this
stage even their areal extent is not well quantified, much less
their C content, vegetation composition, or C cycle character-
istics. The short-term to long-term impact of manipulations
(changed climate) needs to be better understood so that it
can be related to climate change impacts. The cumulative
nature of multiple impacts can be complex (e.g., Dukes et
al. 2005). There have been some multiple impact studies
done on peatlands (e.g., Heijmans et al. 2001; Turetsky et
al. 2008; and Chivers et al. 2009) and peatland mesocosms
(e.g., Bridgham et al. 2008), but more work is needed on a
wider range of peatland types. The ranges of reported values
in Table 1 and Figs. 1 and 2 provide some indication of the
uncertainties associated with both peatland areas and green-
house gas flux rates. Typical uncertainties are ±20%
to ±100% and ranges can span a factor of two to more than
an order of magnitude. We consider the values reported in
Table 4 to be central estimates based on the literature, and
cannot provide a quantitative estimate of uncertainty; qualita-
tively, we estimate uncertainties in global aggregate peatland
areas to be on the order of ±25%, and in global aggregate
greenhouse gas fluxes to be on the order of ±50%.
Several issues were not addressed in this synthesis, includ-
ing other disturbances that might be important (e.g., vegeta-
tion migration, sea level rise). There are also other peat
accumulating wetlands, e.g., marshes (Knott et al. 1987) and
mangroves (Bouillon et al. 2008), that we have not discussed,
whose hydrology, ecology, and C cycling will be impacted
by 21st climate change, land use, and sea level rise, but
whose characteristic responses may be quite different.
Although we dont anticipate net CO2losses from peat-
386 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
lands in the 21st century to be large, the peatland carbon
pool is large enough that a net loss of 0.1 Pg C y1could
persist for several millennia, slowly but inexorably increasing
the atmospheric CO2burden and making it more difficult to
reach a target level by reducing direct anthropogenic emis-
sions. We, therefore, conclude that the role of peatlands in
the 21st century carbon budget will not be a carbon bomb,
but rather a relatively small and persistent contributor to the
atmospheric CO2burden, with occasional significant pulses
in years with extended droughts. Increased CH4emissions re-
sulting from permafrost thaw are likely to exceed decreased
emissions from land use. Non-permafrost peatland CH4emis-
sions will likely increase somewhat due to warming, but that
will be tempered or enhanced by drying or wetting, which
may be very heterogeneous across the planet. If agricultural
and forestry land use of peatlands increases, N2O emissions
will increase, but they will likely remain a small fraction of
the global N2O budget.
We thank the National Science Foundation Research Co-
ordination Networks (NSF RCN) program for support of the
Peatland Ecosystem Analysis and Training NETwork (Peat-
NET), which organized and funded several international meet-
ings focused on peatlands and the C cycle. We thank Dave
Lawrence for helpful comments on the manuscript and ideas
within it. In addition, SF and JT were supported by NSF
grants ATM-0628399 and ARC-1021300, National Aeronau-
tics and Space Administration (NASA) grant NNX07AH32G,
and the NASA IDS Program. Funding for NTR climate re-
search was provided by the Natural Sciences and Engineering
Research Council of Canada and the Canadian Foundation for
Climate and Atmospheric Sciences. MCJ was supported by
NSF grant ATM-0628455. CCT was supported by UNH
NASA Research and Discover Program and US DOE Office
of Science Graduate Fellowship (SCGF); EST was supported
by the Academy of Finland (project codes 218101, 140863).
Aerts, R., and Ludwig, F. 1997. Water-table changes and nutritional
status affect trace gas emissions from laboratory columns of
peatland soils. Soil Biol. Biochem. 29(11-12): 16911698. doi:10.
Aerts, R., Cornelissen, J.H.C., Dorrepaal, E., van Logtestijn, R.S.P.,
and Callaghan, T.V. 2004. Effects of experimentally imposed
climate scenarios on flowering phenology and flower production
of subarctic bog species. Glob. Change Biol. 10(9): 15991609.
Andriesse, J.P. 1988. Nature and management of tropical peat soils.
FAO Soils Bulletin 59, Rome.
Archer, D., and Brovkin, V. 2008. The millennial atmospheric
lifetime of anthropogenic CO2. Clim. Change, 90(3): 283297.
Arens, S.J.T., Sullivan, P.F., and Welker, J.M. 2008. Nonlinear
responses to nitrogen and strong interactions with nitrogen and
phosphorus additions drastically alter the structure and function of
a high arctic ecosystem. J. Geophys. Res. 113(G3): G03S09.
Armentano, T.V., and Menges, E.S. 1986. Patterns of change in the
carbon balance of organic soil-wetlands of the temperate zone. J.
Ecol. 74(3): 755774. doi:10.2307/2260396.
Aurela, M., Laurila, T., and Tuovinen, J.P. 2004. The timing of snow
melt controls the annual CO2balance in a subarctic fen. Geophys.
Res. Lett. 31(16): L16119. doi:10.1029/2004GL020315.
Aurela, M., Riutta, T., Laurila, T., Tuovinen, J., Vesala, T., Tuittila,
E., Rinne, J., Haapanala, S., and Laine, J. 2007. CO2exchange of a
sedge fen in southern Finlandthe impact of a drought period.
Tellus B Chem. Phys. Meterol. 59(5): 826837. doi:10.1111/j.
Bäckstrand, K., Crill, P.M., Jackowicz-Korczyñski, M., Mastepanov,
M., Christensen, T.R., and Bastviken, D. 2010. Annual carbon gas
budget for a subarctic peatland, Northern Sweden. Biogeosciences,
7(1): 95108. doi:10.5194/bg-7-95-2010.
Balshi, M.S., McGuire, A.D., Duffy, P., Flannigan, M., Walsh, J., and
Melillo, J. 2009. Assessing the response of area burned to
changing climate in western boreal North America using a
Multivariate Adaptive Regression Splines (MARS) approach.
Glob. Change Biol. 15(3): 578600. doi:10.1111/j.1365-2486.
Barber, K. 1993. Peatlands as scientific archives for past biodiversity.
Biodivers. Conserv. 2(5): 474489. doi:10.1007/BF00056743.
Bartlett, K.B., and Harriss, R.C. 1993. Review and assessment of
methane emissions from wetlands. Chemosphere, 26(14): 261
320. doi:10.1016/0045-6535(93)90427-7.
Beer, J., and Blodau, C. 2007. Transport and thermodynamics
constrain belowground carbon turnover in a northern peatland.
Geochim. Cosmochim. Acta, 71(12): 29893002. doi:10.1016/j.
Beilman, D.W., MacDonald, G.M., Smith, L.C., and Reimer, P.J.
2009. Carbon accumulation in peatlands of West Siberia over the
last 2000 years. Global Biogeochem. Cycles, 23(1): GB1012.
Benscoter, B., and Vitt, D. 2008. Spatial patterns and temporal
trajectories of the bog ground layer along a post-fire chronose-
quence. Ecosystems (N.Y.), 11(7): 10541064. doi:10.1007/
Benscoter, B.W., and Wieder, R.K. 2003. Variability in organic
matter lost by combustion in a boreal bog during the 2001
Chisholm fire. Can. J. For. Res. 33(12): 25092513. doi:10.1139/
Benscoter, B.W., Wieder, R.K., and Vitt, D.H. 2005. Linking
microtopography with post-fire succession in bogs. J. Veg. Sci.
16(4): 453460. doi:10.1111/j.1654-1103.2005.tb02385.x.
Billett, M.F., Charman, D.J., Clark, J.M., Evans, C.D., Evans, M.G.,
Ostle, N.J., Worrall, F., Burden, A., Dinsmore, K.J., Jones, T.,
McNamara, N.P., Parry, L., Rowson, J.G., and Rose, R. 2010.
Carbon balance of UK peatlands: current state of knowledge and
future research challenges. Clim. Res. 45:1329. doi:10.3354/
Blodau, C., Basiliko, N., and Moore, T.R. 2004. Carbon turnover in
peatland mesocosms exposed to different water table levels.
Biogeochemistry, 67(3): 331351. doi:10.1023/B:BIOG.
Blundell, A., and Barber, K. 2005. A 2800-year palaeoclimatic record
from Tore Hill Moss, Strathspey, Scotland: the need for a multi-
proxy approach to peat-based climate reconstructions. Quat. Sci.
Rev. 24(10-11): 12611277. doi:10.1016/j.quascirev.2004.08.017.
Bouillon, S., Borges, A.V., Castañeda-Moya, E., Diele, K., Dittmar,
T., Duke, N.C., Kristensen, E., Lee, S.Y., Marchand, C.,
Middelburg, J.J., Rivera-Monroy, V.H., Smith, T.J., III, and
Twilley, R.R. 2008. Mangrove production and carbon sinks: a
revision of global budget estimates. Global Biogeochem. Cycles,
22(2): GB2013. doi:10.1029/2007GB003052.
Bourgeau-Chavez, L.L., Kasischke, E.S., Riordan, K., Brunzell, S.,
Nolan, M., Hyer, E., Slawski, J., Medvecz, M., Walters, T., and
Frolking et al. 387
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
Ames, S. 2007. Remote monitoring of spatial and temporal surface
soil moisture in fire disturbed boreal forest ecosystems with ERS
SAR imagery. Int. J. Remote Sens. 28(10): 21332162. doi:10.
Bousquet, P., Ciais, P., Miller, J., Dlugokencky, E., Hauglustaine, D.,
Prigent, C., van der Werf, G., Peylin, P., Brunke, E., Carouge, C.,
Langenfelds, R., Lathière, J., Papa, F., Ramonet, M., Schmidt, M.,
Steele, L., Tyler, S., and White, J. 2006. Contribution of
anthropogenic and natural sources to atmospheric methane
variability. Nature, 443(7110): 439443. doi:10.1038/
nature05132. PMID:17006511.
Bragazza, L. 2006. Chapter 12 Consequences of increasing levels of
atmospheric nitrogen deposition on ombrotrophic peatlands: a
plant-based perspective. In Peatlands Evolution and Records of
Environmental and Climate Changes Elsevier. Edited by I.P.
Martini, A. Martinez Cortizas, and W. Chesworth. Vol. 9, pp. 271
Bragazza, L., Tahvanainen, T., Kutnar, L., Rydin, H., Limpens, J.,
Hájek, M., Grosvernier, P., Hájek, T., Hajkova, P., Hansen, I.,
Iacumin, P., and Gerdol, R. 2004. Nutritional constraints in
ombrotrophic Sphagnum plants under increasing atmospheric
nitrogen deposition in Europe. New Phytol. 163(3): 609616.
Bridgham, S.D., Ping, C.-L., Richardson, J.L., and Updegraff, K.
2000. Soils of northern peatlands: Histosols and Gelisols. In
Wetland Soils: Genesis, Hydrology, Landscapes, and Classifica-
tion. Edited by J. L. Richardson, and M. J. Vepraskas. CRC Press,
Boca Raton, FL, USA. pp. 343370.
Bridgham, S.D., Megonigal, J.P., Keller, J.K., Bliss, N.B., and
Trettin, C. 2006. The carbon balance of North American wetlands.
Wetlands, 26(4): 889916. doi:10.1672/0277-5212(2006)26[889:
Bridgham, S.D., Pastor, J., Dewey, B., Weltzin, J.F., and Updegraff,
K. 2008. Rapid carbon response of peatlands to climate change.
Ecology, 89(11): 30413048. doi:10.1890/08-0279.1.
Bubier, J.L. 1995. The relationship of vegetation to methane emission
and hydrochemical gradients in Northern Peatlands. J. Ecol. 83(3):
403420. doi:10.2307/2261594.
Bubier, J.L., Moore, T.R., Bellisario, L., Comer, N.T., and Crill, P.M.
1995. Ecological controls on methane emissions from a northern
peatland complex in the zone of discontinuous permafrost,
Manitoba, Canada. Global Biogeochem. Cycles, 9(4): 455470.
Bubier, J., Moore, T., Savage, K., and Crill, P. 2005. A comparison of
methane flux in a boreal landscape between a dry and a wet year.
Global Biogeochem. Cycles, 19(1): GB1023. doi:10.1029/
Bubier, J., Moore, T., and Bledzki, L. 2007. Effects of nutrient
addition on vegetation and carbon cycling in an ombrotrophic bog.
Glob. Change Biol. 13(6): 11681186. doi:10.1111/j.1365-2486.
Camill, P., Lynch, J.A., Clark, J.S., Adams, J.B., and Jordan, B. 2001.
Changes in biomass, aboveground net primary production, and
peat accumulation following permafrost thaw in the boreal
peatlands of Manitoba, Canada. Ecosystems (N.Y.), 4(5): 461
478. doi:10.1007/s10021-001-0022-3.
Chapman, S.J., and Thurlow, M. 1996. The influence of climate on
CO2and CH4emissions from organic soils. Agric. For. Meteorol.
79(4): 205217. doi:10.1016/0168-1923(95)02283-X.
Chappellaz, J., Blunier, T., Kints, S., Dällenbach, A., Barnola, J.,
Schwander, J., Raynaud, D., and Stauffer, B. 1997. Changes in the
atmospheric CH4gradient between Greenland and Antarctica
during the Holocene. J. Geophys. Res. 102(D13): 1598715997.
Charman, D. 2002. Peatlands and Environmental Change. John Wiley
& Sons, Chichester, UK.
Chen, Y.-H., and Prinn, R.G. 2006. Estimation of atmospheric
methane emissions between 1996 and 2001 using a three-
dimensional global chemical transport model. J. Geophys. Res.
111(D10): D10307. doi:10.1029/2005JD006058.
Chimner, R.A., and Cooper, D.J. 2003. Influence of water table levels
on CO2emissions in a Colorado subalpine fen: an in situ
microcosm study. Soil Biol. Biochem. 35(3): 345351. doi:10.
Chivers, M.R., Turetsky, M.R., Waddington, J.M., Harden, J.W., and
McGuire, A.D. 2009. Effects of experimental water table and
temperature manipulations on ecosystem CO2fluxes in an Alaskan
Rich Fen. Ecosystems (N.Y.), 12(8): 13291342. doi:10.1007/
Christensen, T.R., Ekberg, A., Ström, L., Mastepanov, M., Panikov,
N., Öquist, M., Svensson, B.H., Nykänen, H., Martikainen, P.J.,
and Oskarsson, H. 2003. Factors controlling large scale variations
in methane emissions from wetlands. Geophys. Res. Lett. 30(7):
1414. doi:10.1029/2002GL016848.
Christensen, J.H., Hewitson, B., Busuioc, A., Chen, A., Gao, X.,
Held, I., Jones, R., Kolli, R.K., Kwon, W.-T., Laprise, R., Magaña
Rueda, V., Mearns, L., Menéndez, C.G., Räisänen, J., Rinke, A.,
Sarr, A., and Whetton, P. 2007. Regional Climate Projections. In
Climate Change 2007: The Physical Science Basis. Contribution
of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change. Edited by S.
Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B.
Averyt, M. Tignor, and H.L. Miller. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA.
Clymo, R.S., Pearce, D.M.E., and Conrad, R. 1995. Methane and
Carbon Dioxide Production in, transport through, and efflux from
a eeatland. Philos. Trans. R. Soc. London, Ser. A. 351(1696): 249
259. [and discussion.] doi:10.1098/rsta.1995.0032.
Couwenberg, J., Dommain, R., and Joosten, H. 2010. Greenhouse gas
fluxes from tropical peatlands in Southeast Asia. Glob. Change
Biol. 16(6): 17151732. doi:10.1111/j.1365-2486.2009.02016.x.
Crill, P.M., Bartlett, K.B., Harriss, R.C., Gorham, E., Verry, E.S.,
Sebacher, D.I., Madzar, L., and Sanner, W. 1988. Methane flux
from Minnesota peatlands. Global Biogeochem. Cycles, 2(4):
371384. doi:10.1029/GB002i004p00371.
Crill, P., Hargreaves, K., and Korhola, A. 2000. The Role of Peat in
Finnish Greenhouse Gas Balances. Studies and Reports 10/2000,
Ministry of Trade and Industry, Helsinki, Finland.
Danevčič, T., Mandic-Mulec, I., Stres, B., Stopar, D., and Hacin, J.
2010. Emissions of CO2,CH
4and N2O from Southern European
peatlands. Soil Biol. Biochem. 42(9): 14371446. doi:10.1016/j.
Davidson, E.A., and Janssens, I.A. 2006. Temperature sensitivity of
soil carbon decomposition and feedbacks to climate change.
Nature, 440(7081): 165173. doi:10.1038/nature04514. PMID:
Denman, K.L., Brasseur, G., Chidthaisong, A., Ciais, P., Cox, P.M.,
Dickinson, R.E., Hauglustaine, D., Heinze, C., Holland, E., Jacob,
D., Lohmann, U., Ramachandran, S., da Silva Dias, P.L., Wofsy, S.
C., and Zhang, X. 2007. Couplings Between Changes in the
Climate System and Biogeochemistry. In Climate Change 2007:
The Physical Science Basis. Contribution of Working Group I to
the Fourth Assessment Report of the Intergovernmental Panel on
Climate Change. Edited by S. Solomon, D. Qin, M. Manning, Z.
Chen, M. Marquis, K.B. Averyt, M.Tignor, and H.L. Miller.
Cambridge University Press, Cambridge, United Kingdom and
New York, NY, USA.
Dentener, F., Drevet, J., Lamarque, J.F., Bey, I., Eickhout, B., Fiore,
388 Environ. Rev. Vol. 19, 2011
Published by NRC Research Press
Environ. Rev. Downloaded from by Beijing University of Aviation and Aerospace on 06/06/13
For personal use only.
A.M., Hauglustaine, D., Horowitz, L.W., Krol, M., Kulshrestha, U.
C., Lawrence, M., Galy-Lacaux, C., Rast, S., Shindell, D.,
Stevenson, D., Van Noije, T., Atherton, C., Bell, N., Bergman,
D., Butler, T., Cofala, J., Collins, B., Doherty, R., Ellingsen, K.,
Galloway, J., Gauss, M., Montanaro, V., Müller, J.F., Pitari, G.,
Rodriguez, J., Sanderson, M., Solmon, F., Strahan, S., Schultz, M.,
Sudo, K., Szopa, S., and Wild, O. 2006. Nitrogen and sulfur
deposition on regional and global scales: A multimodel evaluation.
Global Biogeochem. Cycles, 20(4): GB4003. doi:10.1029/
Dinsmore, K.J., Billett, M.F., Skiba, U.M., Rees, R.M., Drewer, J.,
and Helfter, C. 2010. Role of the aquatic pathway in the carbon
and greenhouse gas budgets of a peatland catchment. Glob.
Change Biol. 16(10): 27502762. doi:10.1111/j.1365-2486.2009.
Dise, N.B., and Verry, E.S. 2001. Suppression of peatland methane
emission by cumulative sulfate deposition in simulated acid rain.
Biogeochemistry, 53(2): 143160. doi:10.1023/A:1010774610050.