1.  18 O records Ice Volume Every 10-m change in sea level produces an ~0.1‰ change in  18 O of benthic foraminifer Every 10-m change in sea level produces an ~0.1‰ change in  18 O of benthic foraminifer  The age of most prominent  18 O minima  Correspond with ages of most prominent reef recording sea level high stands  Absolute sea levels estimates from reefs Correspond to shifts in  18 OCorrespond to shifts in  18 O  Reef sea level record agreement with assumption of orbital forcing  125K, 104K and 82K events forced by precession

2. Astronomical  18 O as a Chronometer Relationship between orbital forcing and  18 O so strong Relationship between orbital forcing and  18 O so strong   18 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

3. Orbital Tuning 41,000 and 23,000 year cycles from astronomically dated insolation curves 41,000 and 23,000 year cycles from astronomically dated insolation curves  Provide tuning targets  Similar cycles embedded in the  18 O ice volume curves are matched and dated Now most accurate way to date marine sediments Now most accurate way to date marine sediments

4. Orbital-Scale Change in CH 4 & CO 2 Important climate records from last 400 kya Important climate records from last 400 kya  Direct sampling of greenhouse gases in ice Critical questions must be addressed 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 gases can be measured? How well do they represent atmospheric compositions and concentrations?How well do they represent atmospheric compositions and concentrations?

5. Vostok Climate Records Illustrates strong correlation between paleotemperature and the concentration of atmospheric greenhouse gases Illustrates strong correlation between paleotemperature and the concentration of atmospheric greenhouse gases Concentrations of CO 2 and CH 4 moved in tandem with paleotemperatures derived from stable isotope records Concentrations of CO 2 and CH 4 moved in tandem with paleotemperatures derived from stable isotope records Mechanisms of relationships poorly understood Mechanisms of relationships poorly understood To what extent did higher greenhouse gases cause greater radiative warming of the Earth's atmosphere? To what extent did higher greenhouse gases cause greater radiative warming of the Earth's atmosphere?

6. Dating Ice Core Records Ice sheets thickest in center Ice sheets thickest in center  Ice flow slowly downward  Then flows laterally outward Annual layers may be preserved and counted Annual layers may be preserved and counted  Deposition of dust during winter  Blurred at depth due to ice deformation

7. Reliability of Dating Dust layer counting Dust layer counting  Best when ice deposition rapid  Greenland ice accumulates at 0.5 m y -1 Layer counting good to 10,000 yearsLayer counting good to 10,000 years  Antarctica ice accumulates at 0.05 m y -1 Layering unreliable due to slow depositionLayering unreliable due to slow deposition  Where unreliable, ice flow models used  Physical properties of ice  Assumes smooth steady flow Produces “fairly good estimates” of ageProduces “fairly good estimates” of age

8. Dust Layers Greenland has two primary sources for dust Greenland has two primary sources for dust  Particulates from Arctic Canada and coastal Greenland  Large volcanic eruptions anywhere on the globe

9. Gas Trapping in Ice Gases trapped during ice sintering 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) Fast accumulation minimizes age difference (100 years) Slow deposition maximizes age difference (1000-2000 years) Slow deposition maximizes age difference (1000-2000 years)

10. Implication of Age Difference If change in greenhouse gas concentrations 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 rapidWhich should be relatively rapid  Closer to change in ice volume Which should be relatively slowWhich should be relatively slow

11. Reliability and Accuracy of Records Can be evaluated by comparing instrumental record Can be evaluated by comparing instrumental record  With records from rapidly accumulating ice sheets  Instrumental records date to 1958 for CO 2 and 1983 for CH 4 Mauna Loa Observatory (David Keeling)Mauna Loa Observatory (David Keeling)

12. NOAA/CMDL Air Sampling Network 35 Sampling stations or about half world-wide stations

13. CSIRO CH 4 Sampling Network

14. Carbon Dioxide Measurements of CO 2 concentration Measurements of CO 2 concentration  Core from rapidly accumulating ice  Merge well with instrumental data

16. Methane Measurements of CH 4 concentration Measurements of CH 4 concentration  Core from rapidly accumulating ice  Merge well with instrumental data

17. CH 4 and CO 2 in Ice Cores Given agreement between records from rapidly accumulating ice 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

18. Orbital-Scale Changes in CH 4 CH 4 variability CH 4 variability  Interglacial maxima 550- 700 ppb  Glacial minima 350-450 ppb Five cycles apparent in record Five cycles apparent in record  23,000 precession period  Dominates low-latitude insolation  Resemble monsoon signal Magnitude of signals matchMagnitude of signals match

19. Monsoon forcing of CH 4 Match of high CH 4 with strong monsoon Match of high CH 4 with strong monsoon  Strongly suggests connection Monsoon fluctuations in SE Asia Monsoon fluctuations in SE Asia  Produce heavy rainfall, saturate ground  Builds up bogs Organic matter deposition and anaerobic respiration likelyOrganic matter deposition and anaerobic respiration likely –Bogs expand during strong summer monsoon –Shrink during weak summer monsoon

20. Alternative Explanation High-latitude soils and continental margins source of atmospheric methane High-latitude soils and continental margins source of atmospheric methane  CH 4 stored in frozen soils (permafrost)  Continental margin sediments (hydrates) Released during exceptionally warm summers 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 Cycles of summer warming should also occur on 41,000 year cycles  Lack of 41,000 cycle in record argues against high latitude source

21. Orbital-Scale Changes in CO 2 CO 2 record from Vostok CO 2 record from Vostok  Interglacial maxima 280- 300 ppm  Glacial minima 180-190 ppm 100,000 year cycle dominant 100,000 year cycle dominant Match ice volume record Match ice volume record  Timing  Asymmetry  Abrupt increases in CO 2 match rapid ice melting  Slow decreases in CO 2 match slow build-up of ice

22. Orbital-Scale Changes in CO 2 Vostok 150,000 record Vostok 150,000 record  23,000 and 41,000 cycles  Match similar cycles in ice volume Agreement suggests cause and effect relationship Agreement suggests cause and effect relationship  Relationship unknown  e.g., does CO 2 lead ice volume?  Correlations not sufficient to provide definite evaluation

23. Problems with Records Ice cores poorly dated Ice cores poorly dated  CO 2 older than ice by variable amount Greenland ice core well-dated (dust layers) Greenland ice core well-dated (dust layers)  Dust is CaCO 3 -rich  Dissolution of CaCO 3 releases CO 2 Precise timing between changes in CO 2 and ice volume uncertain Precise timing between changes in CO 2 and ice volume uncertain  New data provide better correlation Data do show that signals correlate Data do show that signals correlate  Some causal link must exist Big question – how did CO 2 vary by 30%? Big question – how did CO 2 vary by 30%?

24. Covariation Between CO 2 and  D Substantial mismatch in Vostok records (r 2 = 0.64 over the last 150 kya) Substantial mismatch in Vostok records (r 2 = 0.64 over the last 150 kya) Values shown normalized to their mean values during the mid-Holocene (5–7 kya BP) and the last glacial (18–60 kya BP) 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) If the  D change reflects a proportional T drop, then more than ½ of the interglacial-to- glacial change occurred before significant removal of atmospheric CO 2

25. Temperature from Ice Cores Snow falling on ice sheets under colder temperatures is more negative Snow falling on ice sheets under colder temperatures is more negative  A plot of the  18 O of snow versus temperature shows an excellent correlation  Thus  18 O serves as a paleothermometer

26.  18 O in Ice Cores Several factors in addition to temperature of precipitation Several factors in addition to temperature of precipitation  Affect the  18 O of snow and ice on glaciers

27. Meteoric Water Line  D and  18 O in precipitation correlated  D and  18 O in precipitation correlated  Determined by evaporation/precipitation and rainout  Mixture of equilibrium and non-equilibrium processes  Deuterium excess (d =  D – 8  18 O) quantifies intercept and disequilibria

28. Deuterium Excess in Marine Rain Deuterium-excess value in marine environments 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 surfaceDetermined 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 If humidity deficit is known or can be modeled  Can be used to correct  D/  18 O of precipitation  Determine more precisely ambient temperature during precipitation

29.  D on Antarctica Determine by the temperature, humidity and  D of the vapor source region Determine by the temperature, humidity and  D of the vapor source region  Cuffy and Vimeux (2001, Nature, 415:523- 527) 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 variationsCalculate Southern Hemisphere temperature variations

30. Vostok Temperature and CO 2 Deuterium excess corrected Southern Hemisphere temperature correlate remarkably well with CO 2 variations Deuterium excess corrected Southern Hemisphere temperature correlate remarkably well with CO 2 variations Covariation of CO 2 and temperature have r 2 = 0.89 for last 150 kya and r 2 = 0.84 for last 350-400 kya

31. Implications of Results CO 2 is an important climate forcing on the Modern Earth CO 2 is an important climate forcing on the Modern Earth Long-term synchrony of glacial-interglacial cycling Long-term synchrony of glacial-interglacial cycling  Between Northern and Southern Hemispheres  Due to greenhouse gas variations and feedbacks associated with variations Southern Hemisphere  T explained by Southern Hemisphere  T explained by  CO 2 variations  Without considering changes in N. Hemisphere insolation  Delay between CO 2 decrease and  T  During last glacial inception only ~5,000 years

32. Unresolved Issues Cuffy and Vimeux (2001) show that Cuffy and Vimeux (2001) show that  90% of  T can be explained by variations in CO 2 and CH 4 Reasonably firm grasp on causes of CH 4 variations (Monsoon forcing) Reasonably firm grasp on causes of CH 4 variations (Monsoon forcing)  What produced CO 2 variations?  Variations are large – 30%  Show rapid changes – drop of 90 ppm from interglacial to glacial

33. Physical Oceanographic Changes in CO 2 During glaciations physical properties change During glaciations physical properties change  Temperature and salinity  Affect solubility of CO 2 (aq) and thus pCO 2 90% of the CO 2 decrease unexplained by physical processes

34. Exchange of Carbon Carbon in rock reservoir exchanges slowly Carbon in rock reservoir exchanges slowly  Cannot account for 90 ppm change in 10 3 y Rapid exchange of carbon must involve near- surface reservoirs Rapid exchange of carbon must involve near- surface reservoirs

35. Changes in Soil Carbon Expansion of ice sheets Expansion of ice sheets  Covered or displaced forests  Coniferous and deciduous trees Displaced forests replaced by steppes and grasslandsDisplaced forests replaced by steppes and grasslands –Have lower carbon biomass Pollen records in lakes 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 CO 2 removed from atmosphere did not go into vegetation on land!CO 2 removed from atmosphere did not go into vegetation on land!

36. Where is the Missing Carbon? Carbon from reduced CO 2 during glacial times Carbon from reduced CO 2 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 rapidlyExchanges carbon with atmosphere too rapidly Most areas of ocean within 30 ppm of atmosphereMost areas of ocean within 30 ppm of atmosphere –Glacial surface ocean must also have been lower, like atmosphere  Deep ocean only likely remaining reservoir