JOURNAL OF QUATERNARY SCIENCE (2012)
ISSN 0267-8179. DOI: 10.1002/jqs.2568
Flooding of the continental shelves as a contributor
to deglacial CH4 rise
ANDY RIDGWELL,1 MARK MASLIN2* and JED O. KAPLAN3
1
BRIDGE, School of Geographical Sciences, University of Bristol, Bristol, UK
2
Department of Geography, University College London, London, UK
3
ARVE Group, Environmental Engineering Institute, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland
Received 17 August 2011; Revised 11 June 2012; Accepted 14 June 2012
ABSTRACT: The transition from the last glacial and beginning of Bølling–Allerød and Pre-Boreal periods in particular
is marked by rapid increases in atmospheric methane (CH4) concentrations. The CH4 concentrations reached during
these intervals, 650–750 ppb, is twice that at the last glacial maximum and is not exceeded until the onset of
industrialization at the end of the Holocene. Periods of rapid sea-level rise as the Last Glacial Maximum ice sheets
retreated and associated with ‘melt-water pulses’ appear to coincide with the onset of elevated concentrations of CH4,
suggestive of a potential causative link. Here we identify and outline a mechanism involving the flooding of the
continental shelves that were exposed and vegetated during the glacial sea-level low stand and that can help account
for some of these observations. Specifically, we hypothesize that waterlogging (and later, flooding) of large tracts of
forest and savanna in the Tropics and Subtropics during the deglacial transition and early Holocene would have
resulted in rapid anaerobic decomposition of standing biomass and emission of methane to the atmosphere. This novel
mechanism, akin to the consequences of filling new hydroelectric reservoirs, provides a mechanistic explanation for
the apparent synchronicity between rate of sea-level rise and occurrence of elevated concentrations of ice core CH4.
However, shelf flooding and the creation of transient wetlands are unlikely to explain more than 60 ppb of the
increase in atmospheric CH4 during the deglacial transition, requiring additional mechanisms to explain the bulk of the
glacial to interglacial increase. Similarly, this mechanism has the potential also to play some role in the rapid changes
in atmospheric methane associated with the Dansgaard–Oeschger cycles. Copyright # 2012 John Wiley & Sons, Ltd.
KEYWORDS: continental shelf; dansgaard-oeschger cycles; deglaciation; methane; sea-level rise.
Introduction
Polar ice cores reveal a pronounced variability in atmospheric
methane (CH4) concentrations over the glacial–interglacial
cycles of the past few hundred thousand years, with fluctuations
between glacial lows of 350 ppb and highs of 700–800 ppb
during interglacials (Brook et al., 2000; Petit et al., 1999;
Stauffer et al., 2002). A striking feature of deglacial transitions
is that after an initial period of slow increase the atmospheric
CH4 concentration rises rapidly to a maximum value within
the space of a few hundred years (Brook et al., 2000; Petit
et al., 1999) (Fig. 1). To date, explanations for the driver of
the observed century-scale CH4 increases include wetland
expansion (Brook et al., 2000; Severinghaus et al., 1998, 1999;
Maslin and Burns, 2000), destabilization of marine methane
hydrates (clathrates) (Kennett et al., 2000), and permafrost
melting and terrestrial hydrate destruction (Kalin and Jirikowic,
1996; Kennett et al., 2003).
The ‘smoking gun’ evidence for the marine hydrate
hypothesis (Kennett et al., 2000) – a strong correlation between
the timing of slope slumping and CH4 episodes (Maslin et al.,
2004; Owen et al., 2007) and a significant shift in global carbon
isotope records (Maslin and Thomas, 2003) – has proved
elusive. Changes in the isotope ratios of CH4 can also help in
pinpointing sources of methane during times of abrupt change
in concentrations (Sowers, 2006; Melton, 2010). For instance,
methane hydrates appear to have a more distinctive dD
signature compared to emissions from global wetlands than
they do in d13C hydrates have a mean d13C value of 62.5%,
extremely similar to values of 62% and 56.8% to 58.9%
for boreal and tropical wetlands respectively; in contrast,
dD-CH4 values are 190% and 360 to 380% for hydrates
*Correspondence: M. Maslin, as above.
E-mail: [email protected]
Copyright ß 2012 John Wiley & Sons, Ltd.
and wetlands respectively (Whiticar and Schaefer, 2007).
During the glacial period the GISP dD-CH4 record is
isotopically heavier (Sowers, 2006), as would be expected
with a greater role expected for geologic, hydrate or biomass
burning methane sources. Progressing towards the Holocene
the ice core record of atmospheric CH4 becomes isotopically
lighter and is generally consistent with wetland sources
assuming a greater role.
However, the wetlands hypothesis has also been criticized
because the expansion and maturation of wetlands might not be
fast enough to explain the rapidity of the observed CH4 rise and
synchronicity with proxies for a change to warmer and wetter
conditions (Kennett et al., 2003). Indeed, while global climate
plus wetland models have been successful in accounting for
the overall magnitude of glacial–interglacial CH4 variability
(Singarayer et al., 2011), they do so primarily as a function of
orbital configuration and atmospheric CO2 concentration,
which both change on a multi-millennial rather than centuryscale timescale that characterized the rapid methane transitions. High-resolution ice core records from Greenland also
show that atmospheric CH4 varied on a millennial scale
throughout the last glacial, for which the accepted explanation
is episodic wetland expansion (e.g. Brook et al., 1996, 2000;
Dällenbach et al., 2000; Tzedakis et al., 2009). Although
climate change and particularly temperature increase in driving
higher emissions per unit area rather than necessarily
expanding total area can help explain the rapidity of the
observed CH4 rise (van Huissteden, 2004), we note that some of
the most prominent Dansgaard–Oeschger warming events,
such as Marine Isotope Stage (MIS) 2, 19 and 20, have little
corresponding CH4 response. That some Dansgaard–Oeschger
warmings presumably drive substantial wetland expansion or
emissions intensity, but apparently not others, hints that an
additional contributor to millennial-scale variability in CH4
may be important.
JOURNAL OF QUATERNARY SCIENCE
Figure 1. Observed sea-level and atmospheric CH4 changes at the last
deglaciation. The GISP2 CH4 record is shown at the top (red) (Brook
et al., 1996) and contrasted with a compilation of indicators of relative
sea-level change from the last glacial to late Holocene (bottom – blue)
(compiled by Mark Siddall, pers. comm., from: Bard et al., 1996;
Edwards et al., 1993; Fleming et al., 1998; Hanebuth et al., 2000;
Peltier and Fairbanks, 2006; Yokoyama et al., 2000). Highlighted are
intervals of more rapid sea-level rise associated with MWP-1A, MWP1B and the 19 ka MWP (dark grey) as well as intervals of relatively
slower sea-level rise (light grey). Shown for comparison is the GISP2
d18O local temperature change proxy (Brook et al., 1996) (middle –
black). ‘BA’ and ‘YD’ delineate the Bølling–Allerød (ca. 15.0–12.8 ka)
and Younger Dryas (12.8–11.5 ka), respectively. This figure is available
in colour online at wileyonlinelibrary.com/journal/jqs.
Motivated by the above together with the apparent close
correlation between intervals of sea-level rise and occurrence of
elevated atmospheric CH4 (Fig. 1), we draw upon observations made of the fate of biomass inundated in the creation
of hydroelectricity generation schemes and propose a new
mechanism linking trace gas emissions with a changing water
table and sea-level rise (conceptually illustrated in Fig. 2).
A Shelf-flooding mechanism for transient
wetland creation and CH4 release
During the flooding of hydroelectric reservoirs there is a rapid
decomposition of labile vegetation and soil carbon fractions,
which produces an intense pulse of CH4 to the atmosphere
(Barros et al., 2011; Duchemin et al., 1995; Galy-Lacaux et al.,
1997, 1999). This phenomenon is particularly pronounced
in the Tropics, where higher temperatures reduce oxygen
solubility and enhance metabolic rates. Indeed, the CH4 (and
CO2) emissions from hydroelectricity schemes are sufficiently
large that the long-term climatic impact of some schemes could
approach (or even exceed) that of fossil fuels power plants with
the same installed capacity (Fearnside, 1997; Delmas et al.,
2001; Rosa et al., 2004; Kemenes et al., 2007; Guerin and
Abril, 2007; Zheng et al., 2011; Chen et al., 2011). We suggest
that an analogous methane release might take place during the
Copyright ß 2012 John Wiley & Sons, Ltd.
transition from glacial to Holocene as a consequence of
episodic rapid sea-level rise driven by fluctuations in the size
of ice sheets and releases of meltwater.
Dated coral reef terraces (e.g. Fig. 1) suggest that during the
last deglaciation sea level rose in two relatively abrupt steps of
24 and 28 m (Fairbanks, 1989) and at rates of 3.5–4.5 cm a1
(Hanebuth et al., 2000). Analysis of siliclastic depositional
systems on the Sunda Shelf, which afford a finer temporal
resolution than reef terraces, indicates a significantly faster
rate of rise during these ‘meltwater pulse’ (MWP) events – a
sustained mean rate of 5.3 cm a1 between 14.6 and 14.3 ka
(Hanebuth et al., 2000). These rises in sea level would have
flooded large expanses of the continental shelf that had
been vegetated since the previous (glacial maximum) low-stand
(see Fig. 3).
Terrestrial ecosystems are unlikely to be ‘suddenly’ inundated and permanently covered by the sea in this process.
Instead, an initial rise in water table would cause soils to
become waterlogged and anoxic, with labile soil carbon
fractions together with fresh litter derived from the dying
vegetation being quickly degraded and converted to CH4 and
CO2 (step B in Fig. 2). The soils would subsequently become
increasingly saline, with the high sulphate content of sea water
only later inhibiting methane release (step C in Fig. 2). Modern
coastal mangrove environments are observed to be net CH4
sources to the atmosphere both in the Tropics (e.g. Purvaja
and Ramesh, 2001; Mukhopadhyay et al., 2002; Purvaja et al.,
2004; Kristensen et al., 2008) and in temperate regions
(Livesley and Andrusiak, 2012), suggesting that a steady (but
much smaller) release of CH4 is possible even after soils are
fully inundated by sea water. However, whether mangroves
could colonize and keep pace with a rapidly landward-moving
coastline is questionable. Hence, for the purpose of this paper,
we assume CH4 emissions cease once sea level is higher than
the (soil) surface (step D in Fig. 2).
We quantify the potential for this mechanism in helping
explain rapid increases in atmospheric CH4 concentrations as
follows. The area of shelf flooded is calculated from global
elevation data (ETOPO2, 2001) integrated between 408 S and
408 N (Fig. 4A) on the basis that the flooding of reservoirs
in higher boreal latitudes results in significantly less CH4
production per square metre than is the case at warmer, lower
latitudes (Barros et al., 2011; Duchemin et al., 1995). For the
density of tropical biomass that is flooded we use the timedependent predictions made by a dynamic global vegetation
model (DGVM) run over the deglacial transition (Kaplan et al.,
2002) and integrated over the same latitude band (408).
Combined plant and soil biomass evolves from 10 kgC m2 at
the end of the last glacial (20 ka) to just under 24 kgC m2
during the termination (ca. 13 ka), before falling to under
6 kgC m2 during the Holocene (Fig. 4B). This behaviour
reflects a combination of a shift in the locus of shelf flooding
with increasing sea level and a CO2- and climate-driven change
in biome type, with tropical forest being predominantly
drowned during the deglacial transition, followed by more
savanna and semi-desert biomes during the early Holocene.
The flooding of tropical reservoirs has been estimated to
result in 5% of the total initial carbon biomass being released
to the atmosphere as CH4 (Galy-Lacaux et al., 1999). From
hydroelectric reservoirs in general, CH4 emissions account for
6% of all carbon emissions (the reminder as CO2) (Barros
et al., 2011). The conversion fraction might be somewhat
higher in the case of shelf flooding because the main conduit
of CH4 to the atmosphere – bubbling (‘ebullition’) (Bastviken
et al., 2004) – is strongly anti-correlated with water depth
(Bastviken et al., 2004; Keller and Stallard, 1994), with CH4
emissions being most efficient from shallow lake sediments
J. Quaternary Sci. (2012)
FLOODING OF THE CONTINENTAL SHELVES AND DEGLACIAL CH4 RISE
Figure 2. Hypothesized sequence of events
linking rapid sea-level rise and elevated
atmospheric partial pressure of methane
(pCH4). (A) Initial late-glacial sea-level stand.
The tropical and subtropical shelves are
exposed and vegetated. Atmospheric CH4 is
low. (B) Deglacial transition. Sea-level rises,
with the associated rise in the water table
leading to soil waterlogging over vast tracks
of coastal land. Rapid anaerobic decomposition of soil and above-ground biomass carbon results in elevated methane emissions to
the atmosphere. Atmospheric CH4 rises until
the increased loss rate through hydroxyl
radical removal re-balances the increased
emission rate from the shelves. This situation
persists for the duration of the period of sealevel rise. (C) Sea level begins to stabilize.
With little new land area becoming waterlogged, CH4 emissions fall and cease entirely
once high-sulphate seawater inundates what
was previously land. (D) Post-deglacial conditions with no significant shelf CH4 sources.
This figure is available in colour online at
wileyonlinelibrary.com/journal/jqs.
(Bastviken et al., 2008). The extremely rapid rate of water-level
rise in hydroelectric dams (1–10 m a1) means that ebullition
ceases to be effective much earlier than CH4 production stops.
In contrast, the rise in the water table on the continental
shelves during the deglacial transition would occur at a much
slower rate (0.01–0.1 m a1), suggesting that ebullition would
represent a much more effective CH4 transport mechanism
throughout the period of CH4 production, with the net flux to
the atmosphere potentially approaching the rate of methanogenesis (i.e. 100% efficient transport) (Devol et al., 1988).
We here chose a 10% biomass carbon ! CH4 conversion
efficiently but recognize that this is arguably the greatest
uncertainty in our analysis. Finally, we make a simple estimate
of the enhancement in atmospheric CH4 concentration (DCH4)
that would be attained after a few decades of steady emission
from shelf sources by assuming an atmospheric residence
time equal to modern (9.4 years) (Houghton et al., 2001); i.e.
neither allowing for changes in CH4 lifetime as a function of
atmospheric mixing ratio (e.g. Osborn and Wigley, 1994) nor
impacts of changes in volatile organics from plants (e.g. Valdes
Copyright ß 2012 John Wiley & Sons, Ltd.
et al., 2005). It should be noted that we make a simplified
steady-state calculation in which biomass carbon is assumed
converted to CH4 instantaneously and the resulting sea-level
rise dictated emission rate balanced against oxidative loss in
order to estimate DCH4. Instead, taking into account decay with
time of CH4 fluxes (Barros et al., 2011) from any one square
metre of recently waterlogged shelf would alter the results
only very marginally. Hence, as long as the total assumed
conversion efficiency can be assumed invariant, it does not
matter overly whether emissions occur predominantly within a
single year or over several decades, as long as the timescale of
emissions decline is short compared to the duration of intervals
of rapid sea-level rise (500 years).
Results and discussion
The results of our analysis are shown in Fig. 4(C). Methane
release through the shelf-flooding mechanism is strongly
dependant on initial sea level, which helps explain features
of the ice core record. Because the distribution of the surface
J. Quaternary Sci. (2012)
JOURNAL OF QUATERNARY SCIENCE
Figure 3. Timing of inundation of the continental shelves during the last deglaciation in steps of 2000 years.
area of the continental shelf is not uniform with altitude (or sea
level), a unit rise in sea level under ‘intermediate’ glacial
conditions (sea level between 20 and 50 m below modern)
inundates a much greater area compared to an identical sealevel rise occurring during peak glacial conditions. In addition,
carbon density on the shelves increases over the course of the
deglacial transition in response to increases in precipitation
(climate amelioration) and CO2-driven fertilization of productivity in the DGVM (Kaplan et al., 2002) (Fig. 4B). The
resulting increase in biomass carbon flooded per unit rise in sea
level translates into a greater atmospheric CH4 increase (DCH4)
(Fig. 4C). For the MWP-1A and MWP-1B sea-level rise events
we calculate maximum DCH4 values of 24 and 54 ppb,
respectively, equivalent to 15–30% of the observed change at
the start of the Bølling–Allerød and end of the Younger Dryas.
The distribution of flooded biome is spatially highly heterogeneous and, as simulated by Kaplan et al. (2004), is dominated
in terms of area (excluding latitudes outside of 408) by
tropical broadleaf forest and tropical shrubland situated in
Southeast Asia (Fig. 3). Instead, taking the global mean
drowned biomass carbon density but still restricting shelf area
to 408, the MWP-1A and MWP-1B DCH4 estimates become
32 and 62 ppb, respectively (Fig. 5).
Although more than 50% of drowned biomass lies outside of
408 at the time of MWP-1B, the mean CH4 fluxes from high
latitude (>308) are on average an order of magnitude lower
than from reservoirs at tropical latitudes (Barros et al., 2011),
meaning that restricting our analysis to lower latitudes is
unlikely to result in any substantive underestimate. For
instance, one might introduce an explicit latitudinal temperature modifier which would not only allow the smaller
contributions form cold, high latitudes to be taken into account
but would also account for the effect of glacial ocean surface
tropical temperatures being some 28C cooler (MARGO
Project Members, 2009) and hence reducing somewhat the
Copyright ß 2012 John Wiley & Sons, Ltd.
rate of methogenesis in the Tropics. A rather greater source of
uncertainty occurs because our predicted DCH4 scales directly
with the rate of sea-level rise – underestimates made here in the
maximum rate of sea-level rise will produce CH4 emission
estimates that are too low (and vice versa). Indeed, we note
that the uncertainty in estimated rates of sea-level rise for
MWP-1A allows for an upper limit of 16 cm a1; although a
value this high seems unlikely (Hanebuth et al., 2000), it would
provide a larger (but necessarily shorter in duration) increase in
methane. However, loss of pre-existing (at the Last Glacial
Maximum) low-lying freshwater wetlands and their associated
CH4 emissions during flooding should also to be taken into
account (Tranvik et al., 2009; Teodoru et al., 2012), which
we have not done in this simple analysis, creating an overprediction in our estimates. Finally, alternative assumptions
regarding the conversion efficiency of organic matter to
methane (here 10%) may be equally (if not more so) valid –
our DCH4 estimates (Figs. 4 and 5) can be simply linearly scaled
to reflect this.
Interestingly, and in contrast to the response associated
with MWP-1A and MWP-1B, there is very little change in
atmospheric CH4 at the end of the last glacial associated with
the ‘19 ka’ meltwater event (Fig. 1). This is despite there being
evidence for a sea-level rise of 10 m occurring over an interval
‘considerably shorter than 500 years’ (Clark et al., 2004). In our
model, we predict no more than 7 ppb associated with this
event (Fig. 4), even when allowing for a 0.04 m a1 rate of rise
(i.e. 10 m in 250 years). The reason for this is that the
hypsometry of the continental slope and shelf (Fig. 4A) at this
time is not conducive to the creation of large expanses of
temporary wetlands. Indeed, a unit increase in sea level,
starting from 120 m relative to modern, affects an area
about half that when starting at 90 m and almost four times
less than at 40 m. Thus our hypothesis is consistent with
maximum increases in methane recorded in ice cores
J. Quaternary Sci. (2012)
FLOODING OF THE CONTINENTAL SHELVES AND DEGLACIAL CH4 RISE
Figure 4. Sensitivity of land area and atmospheric CH4 to past rises in sea level. (A) The first derivative of surface area of the Earth vs. altitude
(equivalent to area inundated per unit rise in sea level). Global elevation data (ETOPO2, 2001) (diamonds) is integrated between 408 S and 408 N. A
Gaussian fit to relative sea level below 20 m (red line) is used for further analysis. The approximate rises in sea level associated with the 19 ka
meltwater event and MWP-1A and MWP-1B are shaded. (B) Predicted change with time in the mean carbon density of flooded tropical shelf biomass
(filled squares) together with a smoothed fit (solid line) (Kaplan et al., 2002) as a function of sea level. (Time is converted to sea level with the data of
Fairbanks, 1989.) (C) Contours are of atmospheric CH4 increase (DCH4) (ppb) assuming 10% biomass carbon ! CH4 conversion efficiency and plotted
as a function of rate of sea-level rise and initial sea level (relative to modern). Shaded regions delineate the range of calculated DCH4 associated with
19 ka MWP (2 cm a1 lower-end sea-level rise estimate represented by grading to white for 2), MWP-1A (vertical band represents the mean estimate
(Hanebuth et al., 2000), while the graded-fill box represents (part) of the 1s uncertainty limit), MWP-1B assuming the same rate of sea-level rise (and
uncertainty) as for MWP-1A. This figure is available in colour online at wileyonlinelibrary.com/journal/jqs.
correlating with the periods when the shelf-flooding mechanism is at its greatest sensitivity to sea level. The fact that a larger
proportion of the shelf area is in the Northern Hemisphere
compared to the Southern Hemisphere also leads us to predict a
stronger north–south atmospheric CH4 gradient during warm
periods and in accord with observations (Dällenbach et al.,
2000).
It should be noted that in the shelf-flooding hypothesis we do
not envisage inundation by sea water as the direct driver of
waterlogging, vegetation die-off and decomposition of organic
matter, as the higher sulphate (SO2
4 ) content of sea water
(29 mmol kg1) would tend to inhibit CH4 emissions by
providing an alternative electron acceptor for CH4 oxidation by
sulphate-reducing bacteria. However, occasional flooding with
Figure 5. Sensitivity to biomass assumptions. (A) Predicted change with time in the
mean carbon density of global shelf biomass
(filled circles). The smoothed fit used in the
analysis is delineated by a dashed line. (B)
Contours of atmospheric CH4 increase (DCH4)
(ppb) as per Fig. 4(C). This figure is available
in colour online at wileyonlinelibrary.com/
journal/jqs.
Copyright ß 2012 John Wiley & Sons, Ltd.
J. Quaternary Sci. (2012)
JOURNAL OF QUATERNARY SCIENCE
sea water would not make CH4 emissions impossible. For
instance, a one-off flooding to a depth of 10 cm of sea water,
for example, would deliver 100 kg m2 sea water with
2
an average sulphate content of 2.9 mol SO2
4 m . Because
biomass density is typically greater than 10 kgC m2 (Fig. 4B),
i.e. 833 mol C m2, the degree to which microbial sulphate
reduction could inhibit CH4 release is equivalent to less
than 1% of total carbon, or 10% of our assumed conversion
to CH4.
Even using slightly more generous assumptions, such as
including shelf carbon outside of the Tropics and Subtropics
(i.e. latitudes outside of 408) or increasing the assumed
conversion efficiency (10%) of organic carbon to methane, it is
clear that shelf flooding cannot explain the entire deglacial rise.
Other sources, particularly tropical wetland expansion and
perhaps changes in the emissions of other reactive trace gases
(Maslin and Burns, 2000; Valdes et al., 2005; Kaplan et al.,
2006) must presumably progressively kick in during the
deglacial transition. In addition, shelf flooding is unable to
explain the full duration of elevated CH4 events such as the
Bølling–Allerød, because the duration of a rate of particularly
rapid sea-level rise (5 or 10 cm a1) would be limited to 500 or
250 years, respectively, whereas the entire Bølling–Allerød
CH4 ‘high’ persists for about 2000 years. However, the
relatively transient CH4 ‘overshoot’ observed at the beginning
of the Pre-Boreal period in the ice core record (Dällenbach
et al., 2000) is not easy to understand in terms of a rapid
expansion followed by almost immediate contraction of
wetlands, while the main phase of northern peatland expansion
did not occur until almost 11 ka (MacDonald et al., 2006).
Instead, the overshoot is consistent with an interval of rapid
tropical shelf flooding occurring on top of a somewhat more
gradual but sustained increase in wetland emissions (Singarayer
et al., 2011).
Finally, while the potential contribution to elevated
atmospheric CH4 due to shelf flooding is relatively modest,
there may be additional biogeochemical consequences of
rapidly inundating vast tracks of shelf ecosystems. For instance,
globally, a 10 m rise in sea level could drown up to 25 Gt C of
biomass which, if fully degraded and converted to CO2, is
equivalent to 12 ppm CO2 in the atmosphere (based on
Fig. 4A, B). This is comparable with the observed magnitude of
CO2 variability during the glacial (Stauffer et al., 2002) and
suggesting a potentially important role of shelf flooding in
accounting for such features of the ice core record. For the
deglacial transition, Montenegro et al. (2006) found that
the magnitude and timing of potential shelf carbon CO2 release
are comparable to the atmospheric CO2 increase. However,
equilibrium of the released CO2 with the ocean and subsequent
geochemical interaction with deep-sea sediments would
reduce the magnitude of the ice core record explained via
this mechanism by 80% or more (Archer et al., 1998). On the
timescale of deglaciation, carbonate compensation (Ridgwell
and Zeebe, 2005) would further reduce the atmospheric CO2
increase that would be driven by this process. Finally, land area
made available by the retreating ice sheets in conjunction with
fertilization of terrestrial productivity by higher atmospheric
concentrations (Kaplan et al., 2002) would provide a strong
negative feedback.
Conclusions
While the overall increase in atmospheric methane concentrations between the last glacial and the Holocene may be
mainly due to a reorganization of the global methane cycle,
including the redistribution of low- and high-latitude wetlands,
the initiation of thermokarst lakes and changes in the
Copyright ß 2012 John Wiley & Sons, Ltd.
magnitude of the tropospheric sink, none of these mechanisms
can easily explain the extremely rapid rises during the last
deglaciation, which occurred in less than 200 years. The shelfflooding hypothesis presented here provides a mechanism to
provide rapid generation of methane with an isotopic signature
that is similar to ‘normal’ wetlands. It also provides a
mechanism to explain the rapid rises in atmospheric methane
during the Dansgaard–Oeschger warming events, as they are
associated with very rapid increases in sea level but may be too
short for continental wetlands to respond.
Acknowledgements. AR acknowledges a University Research Fellowship from the Royal Society. MM acknowledges a Royal Society
Wolfson Merit Award. JOK received support from the Swiss National
Science Foundation (PP0022_119049) and the Italian Ministry of
Research and Education (RBID08LNFJ).
Abbreviations. DGVM, dynamic global vegetation model; MWP,
meltwater pulse.
References
Archer D, Kheshgi H, Maier-Reimer E. 1998. Dynamics of fossil fuel
neutralization by Marine CaCO3. Global Biogeochemical Cycles 12:
259–276.
Bard E, Hamelin B, Arnold M, et al. 1996. Sea level record from Tahiti
corals and the timing of deglacial meltwater discharge. Nature 382:
241–244.
Barros N, Huszar V, Cole JJ., et al. 2011. Carbon emission from
hydroelectric reservoirs linked to reservoir age and latitude. Nature
Geoscience 4: 593–596.
Bastviken D, Cole J, Pace M, et al. 2004. Methane emissions from lakes:
dependence of lake characteristics, two regional assessments, and a
global estimate. Global Biogeochemical Cycles 18: GB4009.
Bastviken D, Cole JJ, Pace ML, et al. 2008. Fates of methane from
different lake habitats: connecting whole-lake budgets and CH4
emissions. Journal of Geophysical Research 113: G02024.
Brook EJ, Sowers T, Orchardo J. 1996. Rapid variations in atmospheric
methane concentration during the past 110,000 years. Science 273:
1087–1091.
Brook EJ, Harder S, Severinghaus J, et al. 2000. On the origin and timing
of rapid changes in atmospheric methane during the last glacial
period. Global Biogeochemical Cycles 14: 559–572.
Chen H, Yuan X, Chen Z, et al. 2011. Methane emissions from the
surface of the Three Gorges Reservoir. Journal of Geophysical
Research – Atmospheres 116: D21306.
Clark PU, McCabe AM, Mix AC, et al. 2004. Rapid rise of sea level
19,000 years ago and its global implications. Science 304: 1141–
1144.
Dällenbach A, Blunier T, Flückiger J, et al. 2000. Changes in the
atmospheric CH4 gradient between Greenland and Antarctica during
the Last Glacial and the transition to the Holocene. Geophysical
Research Letters 27: 1005–1008.
Delmas R, Galy-Lacaux C, Richard S. 2001. Emissions of greenhouse
gases from the tropical hydroelectric reservoir of Petit Saut (French
Guiana) compared with emissions from thermal alternatives. Global
Biogeochemical Cycles 15: 993–1003.
Devol AH, Richey JE, Clark WA, et al. 1988. Methane emissions to the
troposphere from the Amazon floodplain. Journal of Geophysical
Research 93: 1583–1592.
Duchemin É, Lucotte M, Canuel R, et al. 1995. Production of the
greenhouse gases CH4 and CO2 by hydroelectric reservoirs of the
boreal region. Global Biogeochemical Cycles 9: 529–540.
Edwards RL, Beck JW, Burr GS, et al. 1993. A large drop in atmospheric
14 12
C/ C and reduced melting in the Younger Dryas, documented with
230
Th ages of corals. Science 260: 962–968.
ETOPO2. 2001. US Department of Commerce, National Oceanic and
Atmospheric Administration, National Geophysical Data Center, 2minute Gridded Global Relief Data. www.ngdc.noaa.gov/mgg/fliers/
01mgg04.html [22 June 2012].
J. Quaternary Sci. (2012)
FLOODING OF THE CONTINENTAL SHELVES AND DEGLACIAL CH4 RISE
Fairbanks RG. 1989. A 17,000 year glacio-eustatic sea level record:
influence of glacial melting rates on the Younger Dryas event and
deep ocean circulation. Nature 342: 637–642.
Fearnside, P.M., 1997. Greenhouse-gas emissions from Amazonian
hydroelectric reservoirs: the example of Brazil’s Tucuruı́ Dam as
compared to fossil fuel alternatives. Environmental Conservation 24:
64–75.
Fleming K, Johnston P, Zwartz D, et al. 1998. Refining the eustatic
sea-level curve since the Last Glacial Maximum using far- and
intermediate-field sites. Earth and Planetary Science Letters 163:
327–342.
Galy-Lacaux C, Delmas R, Jambert C, et al. 1997. Gaseous emissions
and oxygen consumption in hydroelectric dams: a case study in
French Guyana. Global Biogeochemical Cycles 11: 471–483.
Galy-Lacaux C, Delmas R, Kouadio G, et al. 1999. Long-term greenhouse gas emissions from hydroelectric reservoirs in tropical forest
regions. Global Biogeochemical Cycles 13: 503–517.
Guerin F, Abril G. 2007. Significance of pelagic aerobic methane
oxidation in the methane and carbon budget of a tropical reservoir.
Journal of Geophysical Research – Biogeosciences 112: G03006.
Hanebuth T, Stattegger K, Grootes PM. 2000. Rapid flooding of the
Sunda Shelf: a late-glacial sea-level record. Science 288: 1033–1035.
Houghton JT, Jenkins GJ, Ephraums JJ (eds). 2001. Climate Change
2001: The Scientific Basis. Cambridge University Press: Cambridge
UK.
Kalin,R. M., Jirikowic JL. 1996. A plausible hydrological scenario for
the Bølling–Allerød atmospheric methane increase. Holocene 6:
111–118.
Kaplan JO, Prentice IC, Knorr W, et al. 2002. Modeling the dynamics
of terrestrial carbon storage since the Last Glacial Maximum. Geophysical Research Letters 29: 2074.
Kaplan JO, Folberth G, Hauglustaine D. 2006. Role of methane and
biogenic volatile organic compound sources in late glacial and
Holocene fluctuations of atmospheric methane concentrations. Global Biogeochemical Cycles 20: 2016.
Keller M, Stallard RF. 1994. Methane emission by bubbling from Gatun
Lake, Panama. Journal of Geophysical Research – Atmospheres 99:
8307–8319.
Kemenes A, Forsberg BR, Melack JM. 2007. Methane release below a
tropical hydroelectric dam. Geophysical Research Letters 34:
L12809.
Kennett JP, Cannariato KG, Hendy IL. 2003. Methane Hydrates
in Quaternary Climate Change: The Clathrate Gun Hypothesis.
American Geophysical Union: Washington, DC.
Kennett J, Cannariato KG, Hendy IL, et al. 2000. Carbon isotopic
evidence for methane hydrate instability during quaternary interstadials. Science 288: 128–133.
Kristensen E, Flindt MR, Ulomi S. et al. 2008. Emission of CO2 and CH4
to the atmosphere by sediments and open waters in two Tanzanian
mangrove forests. Marine Ecology – Progress Series 370: 53–67.
Livesley SJ, Andrusiak SM. 2012. Temperate mangrove and salt marsh
sediments are a small methane and nitrous oxide source but important carbon store. Estuarine Coastal and Shelf Science 97: 19–27.
MacDonald GM, Beilman DW, Kremenetski KV, et al. 2006. Rapid
early development of circumarctic peatlands and atmospheric CH4
and CO2 variations. Science 314: 285.
MARGO Project Members. 2009. Constraints on the magnitude and
patterns of ocean cooling at the Last Glacial Maximum. Nature
Geoscience 2: 127–132.
Maslin MA, Burns S. J. 2000. Reconstruction of the Amazon Basin
effective moisture availability over the past 14,000 years. Science
290: 2285–2287.
Maslin MA, Thomas E. 2003. Balancing the deglacial global carbon
budget: the hydrate factor. Quaternary Science Reviews 22: 1729–
1736.
Maslin MA, Owen M, Day S, et al. 2004. Linking continental-slope
failures and climate change: testing the clathrate gun hypothesis.
Geology 32: 53–56.
Maslin MA, Owen M, Betts R, et al. 2010. Gas hydrates: past and future
geohazard? Philosophical Transactions of the Royal Society A 368:
2369–2393.
Copyright ß 2012 John Wiley & Sons, Ltd.
Melton JR. 2010. Methane stable carbon isotope dynamics spanning the
Last Deglaciation. PhD thesis, University of Victoria, BC. http://
hdl.handle.net/1828/2327 (accessed 30 July 2012).
Montenegro A, Eby M, Kaplan JO, et al. 2006. Carbon storage on
exposed continental shelves during the glacial-interglacial transition.
Geophysical Research Letters 33: [DOI: 10.1029/2005GL025480].
Mukhopadhyay SK, Biswas H, De TK, et al. 2002. Impact of Sundarban
mangrove biosphere on the carbon dioxide and methane mixing
ratios at the NE Coast of Bay of Bengal, India. Atmospheric Environment 36: 629–638.
Osborn TJ, Wigley TML. 1994. A simple model for estimating methane
concentration and lifetime variations. Climate Dynamics 9: 181–193.
Owen M, Day S, Maslin MA. 2007. Late Pleistocene submarine mass
movement: occurrence and causes. Quarterly Science Reviews 26:
958–978.
Peltier WR, Fairbanks RG. 2006. Global glacial ice volume and Last
Glacial Maximum duration from an extended Barbados sea level
record. Quaternary Science Reviews 25: 3322–3337.
Petit JR, Jouzel J, Raynaud D, et al. 1999. Climate and atmospheric
history of the past 420,000 years from the Vostok ice core, Antarctica.
Nature 399: 429–436.
Purvaja R, Ramesh R. 2001. Natural and anthropogenic methane
emission from coastal wetlands of South India. Environmental
Management 27: 547–557.
Purvaja R, Ramesh R, Frenzel P. 2004. Plant-mediated methane
emission from an Indian mangrove Purvaja. Global Change Biology
10: 1825–1834.
Ridgwell A, Zeebe RE. 2005. The role of the global carbonate cycle in
the regulation and evolution of the Earth system. Earth and Planetary
Science Letters 234: 299–315.
Rosa LP, dos Santos MA, Matvienko B, et al. 2004. Greenhouse gas
emissions from hydroelectric reservoirs in tropical regions. Climatic
Change 66: 9–21.
Severinghaus JP, Sowers T, Brook EJ, et al. 1998. Timing of abrupt
climate change at the end of the Younger Dryas interval from
thermally fractionated gases in polar ice. Nature 391: 141–146.
Severinghaus JP, Brook EJ. 1999. Abrupt climate change at the end of
the last glacial period inferred from trapped air in polar ice. Science
286: 930–934.
Singarayer JS, Valdes PJ, Friedlingstein P, et al. 2011. Late Holocene
methane rise caused by orbitally controlled increase in tropical
sources. Nature 470: 82–85.
Sowers T. 2006. Late Quaternary atmospheric CH4 isotope record
suggests marine clathrates are stable. Science 311: 838–840.
Stauffer B, Flückiger J, Monnin E, et al. 2002. Atmospheric CO2, CH4
and N2O records over the past 60 000 years based on the comparison
of different polar ice cores. Annals of Glaciology 35: 202–208.
Teodoru CR, Bastien J, Bonneville M-C, et al. 2012. The net carbon
footprint of a newly created boreal hydroelectric reservoir. Global
Biogeochemical Cycles 26: GB2016.
Tranvik LJ, Downing JA, Cotner JB, et al. 2009. Lakes and reservoirs as
regulators of carbon cycling and climate. Limnology and Oceanography 54: 2298–2314.
Tzedakis PC, Pälike H, Roucoux KH, et al. 2009. Atmospheric methane,
southern European vegetation and low-mid latitude links on orbital
and millennial timescales. Earth and Planetary Science Letters 277:
307–317.
Valdes PJ, Beerling DJ, Johnson CE. 2005. The ice age methane budget.
Geophysical Research Letters 32: L02704.
van Huissteden J. 2004. Methane emission from northern wetlands in
Europe during Oxygen Isotope Stage 3. Quaternary Science Reviews
23: 1989–2005.
Whiticar M, Schaefer H. 2007. Constraining past global tropospheric
methane budgets with carbon and hydrogen isotope ratios in ice.
Philosophical Transactions of the Royal Society A 365: 1793–1828.
Yokoyama Y, Lambeck K, DeDeckker P, et al. 2000. Timing of the Last
Glacial Maximum from observed sea-level minima. Nature 406:
713–716.
Zheng H, Zhao X, Zhao T, et al. 2011. Spatial–temporal variations of
methane emissions from the Ertan hydroelectric reservoir in southwest China. Hydrological Processes 25: 1391–1396.
J. Quaternary Sci. (2012)