Exploring the LGMPI change in methane:

1
Exploring the LGM-PI change in methane Is it
possible to differentiate between changes in
sources and changes in sinks? James Levine,
Eric Wolff, Anna Jones and Manuel Hutterli
(BAS) Oliver Wild (Univ. Lancaster), Glenn
Carver and John Pyle (Univ. Cambridge)
2
Outline Methane and the ice record - the
motivation Methane in the AntBL (and
ArcBL) - model results Chemical sensitivities
in the AntBL - chemical signals
3
Methane and the ice record
B/A
D-O8
YD
4
Methane and the ice record
5
Methane and the ice record
6
Changes in sources vs. changes in sinks Model
studies suggest changes in methane sources (i.e.
emissions) cannot fully account for the observed
change in methane concentration (Kaplan, 2002
Valdes et al., 2005) Changes in methane sinks,
specifically changes in OH, could have also
contributed OH could have changed as a result
of changes in NMVOC emissions from
vegetation The bottom line is, methane
emissions and/or oxidising capacity (OH) must
have changed between the LGM and the PI but
their relative contributions remain
uncertain We aim to identify any atmospheric
chemical signal which could differentiate
between changes in methane emissions/OH and
assess the chance it is preserved in the ice
7

• Sensitivity experiments with p-TOMCAT
• 3-D global model of tropospheric chemistry and
  transport
• Driven by ECMWF winds, temperatures and
  humidities
• Resolution of 2.8x 2.8 on 31 levels
  (surface - 10 hPa)
• 52 chemical species (43 advected), 174
  gas-phase reactions
• HOx/NOx chemistries of CH4, C2H6, C3H8 and
  isoprene
• Emissions (PI Valdes et al., 2005) wet and
  dry deposition

8
Sensitivity experiments with a CTM
Antarctic boundary layer (AntBL) lowest model
level south of 70S
9
Methane in the AntBL and ArcBL
PI (control) Sink 2 (DENMVOCs) Source 2
(0.83 x ECH4) Sink 1 (2.5 x POH) Source 1
(0.55 x ECH4)
10
Methane in the AntBL and ArcBL
PI (control) Sink 2 (DENMVOCs) Source 2
(0.83 x ECH4) Sink 1 (2.5 x POH) Source 1
(0.55 x ECH4)
Model run to equilibrium using meteorology from a
single year (1997)
11
Methane in the AntBL and ArcBL
PI (control) Sink 2 (DENMVOCs) Source 2
(0.83 x ECH4) Sink 1 (2.5 x POH) Source 1
(0.55 x ECH4)
Model run to equilibrium using meteorology from a
single year (1997) Then run for further three
years (1998-2000) with respective meteorology
12
Methane in the AntBL and ArcBL
PI (control) Sink 2 (DENMVOCs) Source 2
(0.83 x ECH4) Sink 1 (2.5 x POH) Source 1
(0.55 x ECH4)
Model run to equilibrium using meteorology from a
single year (1997) Then run for further three
years (1998-2000) with respective
meteorology Paired source/sink expts
designed to achieve same change in AntBL methane
13
Methane in the AntBL and ArcBL
PI (control) Sink 2 (DENMVOCs) Source 2
(0.83 x ECH4) Sink 1 (2.5 x POH) Source 1
(0.55 x ECH4)
Reduction in methane upon switching off emissions
from vegetation half the observed LGM-PI
change ? At least half the LGM-PI change must be
due to a change in methane emissions ..and/or a
change in OH due to some other factor besides a
change in vegetation
14
Chemical sensitivities in the AntBL
But the real question is, what else changes as
methane changes? Does any chemical species
show substantially different sensitivities to the
changes in methane emissions/OH (leading to
comparable changes in AntBL methane)? (And, if
so, what is the likelihood it is preserved in the
ice record?)
15
Chemical sensitivities in the AntBL
16
Chemical sensitivities in the AntBL
17
Chemical sensitivities in the AntBL

• But many species can be discounted on the basis
  that
• they are too reactive to be preserved in the ice
  record (e.g. radicals)
• they exhibit negligible concentrations in the
  AntBL (e.g. isoprene)
• their interpretation requires unknown
  information (e.g. NMVOCs)
• they show very similar sensitivities to changes
  in methane emissions/OH

18
Chemical sensitivities in the AntBL
2.5 x POH 0.55 x ECH4
DENMVOCs 0.83 x ECH4
19
Chemical sensitivities in the AntBL
2.5 x POH 0.55 x ECH4
DENMVOCs 0.83 x ECH4
PAN decreases in Sink 1 and Sink 2 but hardly
changes in Source 1 and Source 2 ? Could provide
an indicator of changes in OH (and/or emissions
from vegetation) we plan to examine PI PAN
budget, and changes therein in the sensitivity
experiments Could be issues re the preservation
of PAN in the ice thermal decomposition,
photolysis
20
Chemical sensitivities in the AntBL
2.5 x POH 0.55 x ECH4
DENMVOCs 0.83 x ECH4
H2O2 increases in Sink 1 and decreases in
Source 1/2.. but also decreases in Sink 2 ? H2O2
may be of limited use in differentiating between
changes in methane emissions and changes in
OH, if the latter are driven by changes in
emissions from vegetation Frey et al. (2006)
predict a well-preserved, long-term H2O2 record
at Divide site (WAIS)
21
Chemical sensitivities in the AntBL
2.5 x POH 0.55 x ECH4
DENMVOCs 0.83 x ECH4
Assuming steady state is established between HCHO
production (OH CH4) and loss (OH, J), DOH
can be estimated from DHCHO and DCH4
(Staffelbach et al., 1991) ? DHCHO and DCH4
give reasonable estimates of DOH But (post-)
depositional processes complicate HCHO
preservation (Hutterli et al., 2002)
22
Chemical sensitivities in the AntBL
2.5 x POH 0.55 x ECH4
DENMVOCs 0.83 x ECH4
O3/OH ratio shows substantially different
sensitivities in the sink/source experiments ?
If the mass-independent fractionation of oxygen
isotopes in sulphates is related to the O3/OH
ratio (Savarino et al. 2000), it could provide a
useful chemical signal
23

• Conclusions
• PAN could provide an indicator of changes in
  OH (and/or emissions from vegetation)
• ? Further work needed to examine PI PAN budget
  and changes therein
• DHCHO and DCH4 give reasonable estimates of
  DOH (Staffelbach et al., 1991)
• ? Possible use during periods in which HCHO
  preservation remains constant
• DO3/OH (DMIF of oxygen isotopes in sulphates
  Savarino et al., 2000) shows potential

24

• Conclusions
• Changes in NMVOC emissions from vegetation can,
  at most, account for roughly half
• the observed change in methane between the LGM
  and the PI
• ? The results of new experiments (not shown)
  suggest combined effect of DT and DH2O
• is roughly equal and opposite to the maximum
  effect of changes in vegetation
• Much of the observed change in methane must be
  due to a change in methane
• emissions and/or a change in OH for some
  other reason (..NOX?)
• ? Still much uncertainty re influence of
  vegetation competing effects of DT and DCO2

25
Acknowledgements Paul Valdes (Univ. Bristol)
Providing PI emissions (Valdes et al.,
2005) Nicola Warwick (Univ. Cambridge) Impleme
nting PI emissions in p-TOMCAT