Everyone is entitled to their own opinion but not their own facts

Tom Quirk

A presentation to The Lavoisier Group’s 2007 Workshop

‘Rehabilitating Carbon Dioxide’ held in Melbourne on 29-30 June 2007

7 June Presentation Lavoisier Meeting

Opinions and Facts This presentation will be directed at the context in which various data sets are used to set the scene for the present and past climate.

1

Past Temperature Record

Figure 1 is the conventional presentation of the extensive ice core record. Figure 1 Temperature derived from Vostok ice cores Petit et al(1999)

4

2

0

-2 Temperature difference from -4 Present

-6

-8

-10 450,000

400,000

350,000

300,000

250,000

200,000

150,000

100,000

50,000

0

Years before Present

Figure 2 puts the ice core record in the context of present temperature variations from the Equator to the South Pole. Note the very large, 50 C variation at Vostok with a possible 4 C variation in the tropical oceans. Figure 2 Temperature derived from Vostok ice cores Petit et al(1999)

Temperature

Annual Summer Maximum

Annual Winter Minimum

Equatorial Temperature 50

25

Equator

"Average" Surface Temperature 15 C 0

-25

-50

-75

-100 450,000

400,000

350,000

300,000

250,000

200,000

Years before Present

150,000

100,000

50,000

0

Temperature C Pole

Figures 3 and 4 give a rough view of the global surface temperature distribution in summer and winter. Note that in the tropics there is little change in surface area for the high temperature bands. The tropical oceans are controlling and limiting temperature movements as a result of evaporation the Polar Regions have substantial change and particularly Antarctica shows extreme sensitivity moving from summer to winter. Figure 3 World Surface Temperature Distribution Source 1947-2007 data NCEP/NCAR 30.0%

25.0%

January July 20.0% Fraction of Surface 15.0% in Temperature Intervals 10.0%

5.0%

0.0% -58 -53 -48 -43 -38 -33 -28 -23 -18 -13

-8

-3

3

8

13 18

23 28

10

20

30

33

Average Temperature C

Figure 4 World Surface Temperature Distribution Source 1947-2007 data NCEP/NCAR 100%

January

90%

July

80% 70% % Surface below given temperature

60% 50% 40% 30% 20% 10% 0% -60

-50

-40

-30

-20

-10

0

40

Average Temperature C

The interesting question is how the ice core record might indicate temperature changes nearer to the equator and the “average” global temperature. It is clear that the tropical temperature range is substantially damped.

2

2

Sampling CO2 in Ice Cores

The statement is frequently made that, apart from the present, the level of CO2 in the atmosphere has not been above 300 ppm for the last 500,000 years. Figure 5 shows the Vostok measurement of the level of CO2. Figure 5 Carbon dioxide Petit et al(1999)

300.0

280.0

260.0

240.0 ppmv 220.0

200.0

180.0

160.0 450,000

400,000

350,000

300,000

250,000

200,000

150,000

100,000

50,000

0

Time before Present

A comparison of ice core measurements and contemporary “in atmosphere” measurements is shown in Figure 6. This is a careful comparison of C14 sampled from CO2 in the atmosphere and at the Law Dome in Antarctica. The sampling explores the age resolution of CO2 measurements in an elegant manner. Figure 6 C14 Measurements Levchenko (1996) Law Dome

NZ Atmosphere

1800 1700 1600 1500 1400

C14

1300 1200 1100 1000 900 800 1930

1940

1950

1960

1970

1980

1990

2000

Year

3

When snow falls it remains in a porous state for some time and air is able to diffuse through it. As the snow accumulates, the weight (pressure) compacts the snow finally sealing the spaces from the column above and then forming gas bubbles in the ice. The diffusing column is 70 to 100 metres deep. The partially compacted snow is referred to as Firn. The compacting process allows a mixing of the air over a period of years. By using the carbon pulse from weapons testing it has been possible to measure the effect of this compacting and sealing process for air sampling. This resolution or extended sampling time reduces the carbon 14 peak by an absolute 10 percent compared to the atmospheric measurements Figure 7 Law Dome Ice Age variation with Depth At a depth of 80 metres 0.7 and 1.3 years per metre Age Spread is 12 years or 10 metres of ice core Etheridge (1996)

1950 1940 1930

Law Dome Approx. Firn Depth

1920 1910 Year 1900 AD 1890 1880 1870 1860 1850 50.00

70.00

90.00

110.00

130.00

150.00

170.00

190.00

Depth metres

Ice accumulation at the Law Dome is 0.7 to 1.2 metres per year as recorded at a depth of 80 to 100 metres, Figure 7. As a result the diffusion process in the snow and firn averages out air sampling over a period of 12 years. Figure 8 Modelling C14 Measurements Measured Moving Window 12.5 years Normal Distribution 4 year Std Dev

1800

1600

1400

C14 1200

1000

800 1930

1940

1950

1960

1970

1980

1990

2000

Year 4

The result gives an indication of the age resolution in the Vostok and EPICA measurements The Vostok and EPICA cores show 0.01 to 0.02 metres accumulation per year, Figures 9 and 10. If the pressure column of Firn is the main control for the air sampling process then scaling the Law Dome results give an age resolution-averaging period of 100’s to 1,000 years at EPICA and Vostok. Figure 9 Ice Age variation with Depth 2000 1500 1000 500 0 Year AD/BC -500 -1000 -1500

Law Dome Taylor EPICA Dome C Vostok Approx. Firn Depth

-2000 -2500 0.00

20.00

40.00

60.00

80.00

100.00

120.00

140.00

Depth metres

There is an additional correction to be applied for samples collected at depths where pressure has caused the ice to flow and the apparent snowfalls are reduced to mm per year. The apparent compression is shown in Figure 10 with a correction. For a metre sample this may increase the sampling period by one to two hundreds of years. Figure 10

The results of the ice core sampling resolution are illustrated below in Figures 11 and 12. Figure 11 shows measurements from the Vostok ice cores while Figure 12 shows the possible underlying values when unpacked with the measuring resolution.

5

Figure 11 Measured CO2 340 320 300 280 260

CO2 ppmv

240 220 200

330,000

325,000

320,000

315,000

310,000

305,000

180 300,000

Age of Gas BP

Figure 12 Resolution Damping of CO2 Measurements Std Dev 200 Years, Window 1 metre of Ice Core

Underlying CO2 values

Ice Core Resolution added

Measured 340 320 300 280 260

CO2 ppmv

240 220 200

330,000

325,000

320,000

315,000

310,000

305,000

180 300,000

Age of Gas BP

It is not possible to compare peaks and valleys in CO2 measurements from Vostok or EPICA with contemporary atmospheric time series. There is a mismatch in gas age resolution. Peaks are flattened and valleys are filled for ice core measurements.

6

Thus on our contemporary timescale it is not possible to say that the CO2 level has not been above 300 ppm for the last 500,000 years. The same comment applies to comparing the “rapid” run up of contemporary CO2 levels with the ice core records where “sharp” pulses of less than 100 years may well be smoothed away.

3

Sampling CO2 Elsewhere.

There have been a number of techniques used to get better measurement resolution than ice core data. The following are examples. Figure 13 Taylor and Law Dome Ice Cores compared to Stomata Kouwenberg (2005) Stomata

CO2 ppm

420 400 380 360 340 320 300 280 260 240 220 200 180 800

900

Stomata + 1 SD

1000

1100

1200

Stomata - 1 SD

1300

1400

1500

Taylor Dome

1600

1700

Law Dome

1800

1900

2000

1800

1900

2000

Year AD

Figure 14 Taylor and Law Dome Ice Cores compared to Stomata Fit to Constant: Chi Sqrd = 20.6 for 30 D of F Taylor Dome

Law Dome

Stomata

Stomate Average 800 to1845

Stomata Average 1900 to 1990

420 400 380 360 340 320

CO2 300 ppm 280 260 240 220 200 180 800

900

1000

1100

1200

1300

1400

1500

1600

1700

Year AD

Figure13 shows CO2 measurement derived from leaf and needle stomata. The solid lines are as in the paper. Figure 14 shows the minimalist position where a constant, a straight line, is sufficient to fit the data. There is no serious difference of ice core and stomata measurements. Figures 15 and 16 below show a similar result where a constant is sufficient for a fit. However it is often useful to see if correlations exist and a plot of semivariance in Figures 17A and 17B implies that there may be some significant variations in the data of Figure 15. 7

Figure 15 Stomata derived CO2 Levels Wagner (2002) Measurements

Taylor Dome 340 320 300

CO2 ppmv

280 260

9000

8500

8000

7500

240 6500

7000

Age BP

Figure 16 Stomata derived CO2 Levels Fit to Constant: Chi Sqrd = 20.5 for 34 D of F Measurements

Average

Taylor Dome 340 320 300

CO2 ppmv

280 260

9000

8500

8000

7500

240 6500

7000

Age BP

Figure 17A

17B

Semi Variances for Stomata CO2 Levels 830 to 1990 AD Kouwenberg (2005)

Semi Variances for Stomata CO2 Levels Wagner (2002)

6500 to 9000 years BP

Semi Variances Variance X 2 Average

4000

Semi Variances

Variance X 2

Average

Best Fit

900 800

3500 700

3000 600

2500

Semi Variance

Semi 2000 Variance 1500

500 400 300

1000

200

500

100

0 -1

0

1

2

3

4

5

6

Separation of Measurements

7

8

0 -1

0

1

2

3

4

5

6

7

8

Separation of Measurements

8

The conclusion of this section is that data needs to be critically assessed for its significance. This is apart from the systematic errors, which can be overwhelming.

4

Carbon 14 and Atmospheric Mixing

Neutrons produced in the upper atmosphere by cosmic rays from outer space change Nitrogen14 to Carbon14 by replacing a proton with a neutron. The carbon then reacts with oxygen to form CO2. This is a continual process independent of other sources of carbon in the atmosphere. In the late 1950s, nuclear weapons testing in the atmosphere produced neutrons that changed N14 to C14. Atmospheric testing was stopped with the Partial Test Ban Treaty that came into force in October 1963. The weapons tests nearly doubled the amount of C14 in the atmosphere. Thus in 1964, the atmospheric carbon dioxide had been radio-labelled and this has been of great value in tracing the path of carbon dioxide in the biosphere. C14 has a half-life of 5,700 years and decays back to N14. Atmospheric measurements of carbon 14 have been made in the Northern and Southern Hemispheres. Figure 18 shows these along with the difference between North and South. The difference diminished with time showing global mixing with a half-life of about one year. Also note the very large summer to winter variations of CO2 in the Northern Hemisphere that is of the order of 35 ppmv. Figure 18 Northern and Southern Hemisphere Carbon Pulse 1000

Northern Hemisphere Southern Hemisphere

800

Difference

600

Carbon 14

400

200

0

-200

1955

1960

1965

1970

1975

1980

Year

The dilution shown in the falling C14 measurements is not caused by fossil fuel derived carbon, which has no C14. Assuming all the extra CO2 since 1965 was derived from fossil fuel and remains in the atmosphere does not give sufficient dilution to explain the measurements. 9

The explanation must be with the general exchanges between sources and sinks for atmospheric CO2. This is a well understood position. However the fall in the C14 peak is driven by exchanges with sources of reduced C14. This is thought to be from the deep ocean where transport times of hundreds of years yield CO2 that predates the weapons created C14. Figure 19 Dilution of C14 in Atmosphere C

14

i =(1-F).C

14 i-1

+F.C

14

returned

2000

1800 NZ Atmospheric CO2 Measurements

1600 C14 ratio to natural 1400 background X 1000 1200

Law Dome Ice Core Measurements F: 6.2% removed and replaced annually Returned Carbon

1000

800 1940

1950

1960

1970

1980

1990

2000

Year

The fall in C14 is exponential (Figure 19) and this implies that the sources and sinks that contain different levels of C14 keep pace with the steady increase in the total CO2 in the atmosphere.

5

Conclusions

The use of data derived from measurements is often location and context dependent. Presentations do not always make this clear. • Ice Core measurements of CO2 sample over time scales of 100’s to 1,000’s of years. Detail is lost compared to contemporary measurements. Valleys are filled in and peaks are reduced. There is a mismatch of techniques. • Data derived from alternative measure of atmospheric CO2 need to be carefully assessed • Northern and Southern Hemisphere atmospheric mixing takes up to 5 years. Detail may be seen in one hemisphere but lost in the other hemisphere. • Annual mixing shows continuous expansion of the capacity of sinks and sources for exchange of CO2. No evidence of constraints from C14 mixing. •

T W Quirk 7 June 2007

10