18
d O

records Ice Volume
Every 10-m change in sea level produces an
~0.1‰ change in d18O of benthic foraminifer
 The age of most prominent d18O minima
 Correspond with ages of most prominent
reef recording sea level high stands
 Absolute sea levels estimates from
reefs
• Correspond to shifts in d18O
 Reef sea level record agreement with
assumption of orbital forcing
 125K, 104K and 82K events forced by
precession
Astronomical d18O as a Chronometer

Relationship between orbital forcing and d18O
so strong
18
 d O values can orbitally tune sediment age
 Constant relationship in time between
insolation and ice volume
 Constant lag between insolation change
and ice volume change
 Date climate records in ocean sediments
 In relation to the known timing of orbital
changes
Orbital Tuning


41,000 and 23,000
year cycles from
astronomically dated
insolation curves
 Provide tuning
targets
 Similar cycles
embedded in the
d18O ice volume
curves are matched
and dated
Now most accurate
way to date marine
sediments
Orbital-Scale Change in CH4 & CO2


Important climate records from last 400 kya
 Direct sampling of greenhouse gases in ice
Critical questions must be addressed
 Before scale of variability in records
determined
 Reliability of age dating of ice core?
 Mechanisms and timing of gas trapping?
 Accuracy of the record?
• How well gases can be measured?
• How well do they represent atmospheric
compositions and concentrations?
Vostok Climate Records




Illustrates strong correlation
between paleotemperature and
the concentration of
atmospheric greenhouse gases
Concentrations of CO2 and CH4
moved in tandem with
paleotemperatures derived
from stable isotope records
Mechanisms of relationships
poorly understood
To what extent did higher
greenhouse gases cause greater
radiative warming of the
Earth's atmosphere?
Dating Ice Core Records


Ice sheets thickest in center
 Ice flow slowly downward
 Then flows laterally outward
Annual layers may be preserved and counted
 Deposition of dust during winter
 Blurred at depth due to ice deformation
Reliability of Dating

Dust layer counting
 Best when ice deposition rapid
 Greenland ice accumulates at 0.5 m y-1
• Layer counting good to 10,000 years
 Antarctica ice accumulates at 0.05 m y-1
• Layering unreliable due to slow
deposition
 Where unreliable, ice flow models used
 Physical properties of ice
 Assumes smooth steady flow
• Produces “fairly good estimates” of age
Dust Layers

Greenland has two primary sources for dust
 Particulates from Arctic Canada and coastal
Greenland
 Large volcanic eruptions anywhere on the globe
Gas Trapping in Ice



Gases trapped during ice
sintering
 When gas flow to
surface shut down
 Crystallization of ice
 Depths of about 50 to
100 m below surface
 Gases younger than
host ice
Fast accumulation
minimizes age difference
(100 years)
Slow deposition maximizes
age difference (1000-2000
years)
Implication of Age Difference

If change in greenhouse gas concentrations
 Force changes in ice volume
 Gas concentration should lead ice volume
 Gas age is younger than ice age
 Therefore offset between changes in
atmospheric gas concentrations
• Which should be relatively rapid
 Closer to change in ice volume
• Which should be relatively slow
Reliability and Accuracy of Records

Can be evaluated by comparing instrumental record
 With records from rapidly accumulating ice sheets
 Instrumental records date to 1958 for CO2 and
1983 for CH4
• Mauna Loa Observatory (David Keeling)
NOAA/CMDL Air Sampling Network
35 Sampling stations or about half world-wide stations
CSIRO CH4 Sampling Network
Carbon Dioxide

Measurements of CO2 concentration
 Core from rapidly accumulating ice
 Merge well with instrumental data
Methane

Measurements of CH4 concentration
 Core from rapidly accumulating ice
 Merge well with instrumental data
CH4 and CO2 in Ice Cores

Given agreement between records from
rapidly accumulating ice
 Instrumental data
 Accuracy and variability about the
trends
 Assume that longer-term records
collected from ice cores
 Reliable for determining the scale of
variability
Orbital-Scale Changes in CH4


CH4 variability
 Interglacial maxima 550700 ppb
 Glacial minima 350-450
ppb
Five cycles apparent in
record
 23,000 precession period
 Dominates low-latitude
insolation
 Resemble monsoon
signal
• Magnitude of signals
match
Monsoon forcing of CH4
Match of high CH4 with strong monsoon
 Strongly suggests connection
 Monsoon fluctuations in SE Asia
 Produce heavy rainfall, saturate ground
 Builds up bogs
• Organic matter deposition and
anaerobic respiration likely
– Bogs expand during strong summer
monsoon
– Shrink during weak summer
monsoon

Alternative Explanation
High-latitude soils and continental margins
source of atmospheric methane
 CH4 stored in frozen soils (permafrost)
 Continental margin sediments (hydrates)
 Released during exceptionally warm summers
 Precessional changes in summer insolation
affects high latitudes
 Cycles of summer warming should also occur
on 41,000 year cycles
 Lack of 41,000 cycle in record argues
against high latitude source

Orbital-Scale Changes in CO2

CO2 record from Vostok
Interglacial maxima 280300 ppm
 Glacial minima 180-190 ppm
100,000 year cycle dominant
Match ice volume record
 Timing
 Asymmetry
 Abrupt increases in CO2
match rapid ice melting
 Slow decreases in CO2
match slow build-up of ice



Orbital-Scale Changes in CO2


Vostok 150,000 record
 23,000 and 41,000
cycles
 Match similar cycles
in ice volume
Agreement suggests
cause and effect
relationship
 Relationship unknown
 e.g., does CO2 lead
ice volume?
 Correlations not
sufficient to provide
definite evaluation
Problems with Records
Ice cores poorly dated
 CO2 older than ice by variable amount
 Greenland ice core well-dated (dust layers)
 Dust is CaCO3-rich
 Dissolution of CaCO3 releases CO2
 Precise timing between changes in CO2 and
ice volume uncertain
 New data provide better correlation
 Data do show that signals correlate
 Some causal link must exist
 Big question – how did CO2 vary by 30%?

Covariation Between CO2 and dD

Substantial mismatch in Vostok records (r2 = 0.64
over the last 150 kya)
If the dD change reflects a
Values shown normalized
proportional T drop, then more
to their mean values
than ½ of the interglacial-toduring the mid-Holocene
(5–7 kya BP) and the glacial change occurred before
removal of
last glacial (18–60 kyasignificant
BP)
atmospheric CO2
Clearly visible are the
disproportionately low
deuterium values during
the mid-glacial (60–80
Kya BP), the glacial
inception (95–125 Kya
BP), and the penultimate
glacial maximum (140–
150 Kya BP)
Temperature from Ice Cores

Snow falling on ice sheets under colder
temperatures is more negative
 A plot of the d18O of snow versus temperature
shows an excellent correlation
 Thus d18O serves as a paleothermometer
18
d O

in Ice Cores
Several factors in addition to temperature of
precipitation
 Affect the d18O of snow and ice on glaciers
Meteoric Water Line

dD and d18O in precipitation correlated
 Determined by evaporation/precipitation and rainout
 Mixture of equilibrium and non-equilibrium processes

Deuterium excess (d = dD – 8d18O) quantifies intercept
and disequilibria
Deuterium Excess in Marine Rain


Deuterium-excess value in marine environments
 Established at the site of the air-sea interaction
 The offset from equilibrium conditions
• Determined by the humidity deficit above
the sea surface
 This deuterium-excess value is conserved
during the rainout over the continents
If humidity deficit is known or can be modeled
 Can be used to correct dD/d18O of precipitation
 Determine more precisely ambient
temperature during precipitation
dD on Antarctica

Determine by the temperature, humidity and
dD of the vapor source region
 Cuffy and Vimeux (2001, Nature, 415:523527) showed using deuterium excess
 Mismatch is an artifact caused by
variations in climate of the vapor source
region
 Used a climate model and measured
deuterium excess
• Calculate Southern Hemisphere
temperature variations
Vostok Temperature and CO2

Deuterium excess corrected Southern Hemisphere
temperature correlate remarkably well with CO2
variations
Covariation of CO2 and
temperature have
r2 = 0.89 for last 150
kya and r2 = 0.84 for
last 350-400 kya
Implications of Results



CO2 is an important climate forcing on the Modern
Earth
Long-term synchrony of glacial-interglacial cycling
 Between Northern and Southern Hemispheres
 Due to greenhouse gas variations and
feedbacks associated with variations
Southern Hemisphere DT explained by
 CO2 variations
 Without considering changes in N. Hemisphere
insolation
 Delay between CO2 decrease and DT
 During last glacial inception only ~5,000 years
Unresolved Issues
Cuffy and Vimeux (2001) show that
 90% of DT can be explained by variations
in CO2 and CH4
 Reasonably firm grasp on causes of CH4
variations (Monsoon forcing)
 What produced CO2 variations?
 Variations are large – 30%
 Show rapid changes – drop of 90 ppm
from interglacial to glacial

Physical Oceanographic Changes in CO2

During glaciations physical properties change
 Temperature and salinity
 Affect solubility of CO2(aq) and thus pCO2
90% of the CO2
decrease unexplained
by physical processes
Exchange of Carbon


Carbon in rock reservoir exchanges slowly
 Cannot account for 90 ppm change in 103 y
Rapid exchange of carbon must involve nearsurface reservoirs
Changes in Soil Carbon


Expansion of ice sheets
 Covered or displaced forests
 Coniferous and deciduous trees
• Displaced forests replaced by steppes and
grasslands
– Have lower carbon biomass
Pollen records in lakes
 Indicate glacial times were dryer and less
vegetated than interglacial
 Estimates of total vegetation reduced by 25%
(15-30%) during glacial maxima
• CO2 removed from atmosphere did not go
into vegetation on land!
Where is the Missing Carbon?

Carbon from reduced CO2 during glacial times
 Not explained by physical properties of surface
ocean
 Did not go into biomass on land
 Must have gone into oceans
 Surface ocean not likely
• Exchanges carbon with atmosphere too
rapidly
• Most areas of ocean within 30 ppm of
atmosphere
– Glacial surface ocean must also have
been lower, like atmosphere
 Deep ocean only likely remaining reservoir