tree thermometers and commodities: historic climate indicators

Environment International Vol. 2, pp. 317-333. Pergamon Press Ltd. ! 979. Printed in Great Britain

Tree Thermometers and Commodities: Historic Climate Indicators

L.M. Libby and L.J. Pandolfi Environmental Science and Engineering, University of California at Los Angeles, Los Angeles, CA 90024, U.S.A.; Global Geochemistry, Santa Monica, CA, U.S.A.

In four modern trees, hydrogen and oxygen isotope ratios track the modern temperature records. In a 2000-yr sequence of a Japanese cedar, there are the same periodicities of variation o f D / H and

O18/O 16 as have been found in O18/O 16 in a Greenland ice well. The same periodicities are found in u ran ium and organic carbon concentrat ions versus depth in a sea

core f rom the Santa Barbara Channel , and in carbon-14 variations in a sequence o f Bristlecone pine from southern California.

In a 2000-yr sequence o f Japanese cedar and in a 1000-yr sequence o f European oak D / H and O18/O 16 are related to each other by a slope of 8, just as they are in world-wide precipitation.

In a 72-yr sequence o f Sequoia gigantea, measured year by year for its oxygen isotope ratios, the 10.5-yr cycle o f sunspot numbers found, but not the 21-yr cycle o f sunspot magnet ism; this we believe indicates that the sun is affecting the ear th ' s climate with non-magnet ic particles, probably photons.

All these phenomena are related to periodic changes in sea surface temperature caused by periodic changes in the sun.

History

Long term changes in precipitation, caused by changes in climatic temperature, are well documented in polar ice caps: the heavier of the stable isotopes is depleted in ice laid down in the ice age by comparison with present day ice. In 1970 we extended this concept to trees, suggesting that they, also, are thermometers. Trees grow from water and atmospheric CO 2. In trees which grow on rain water, isotope variations in their rings should be climate indicators because the isotope composition in rain and CO2 varies with temperature.

Libby (72) had previously calculated the theoretical temperature coefficients of the stable isotope fractio- nations in the manufacture of wood from CO2 and H20, finding that the coefficients are small compared with those measured in rain and snow (IAEA, 68, 70, 71, 73, 75).

Whole wood, lignin, or cellulose, wood being about 25% lignin on the average (Gould, 66), were considered for measurement; however, isotope fractionation at climatic temperatures is a function of the frequencies of the chemical bonds (Libby, 72, and references therein). Herzberg (45) suggests, "One would expect the - - C - - H bond to have essentially the same electronic structure and therefore the same force constant in different molecules, and similarly for other bonds. This is indeed observed." For the -C-H bonds the vibrational frequencies in lignin and in cellulose are almost equal, but in fact differ by 6°/0 (Herzberg, 45, see Table, p. 195)

31'7

because cellulose is a multiple alcohol ( H - - C - - O - - H ) n and lignin is a polymer containing both aromatic and aliphatic carbons connected to hydrogen. Therefore, assuming thermodynamic equilibrium, variations of the lignin concentration in tree rings might affect the hydrogen isotope ratio by as much as 25% of 6%, namely by as much as 1'5%.

Likewise for the - - C - - O - - H bonds, the vibrational frequencies are equal in lignin and in cellulose within a few percent (Herzberg , 45), but lignin, different from cellulose, also contains ether linkages, ( - - C - - O - - C - - ) . The - - C - - O - - H linkage has a C - - O bond distance of 1,427 A, and the ther linkage has a - - C - - O bond distance of 1.43 A (Weast, 62); therefore, the presence of lignin, containing 14% oxygen of which 16% is ether linked (Gould, 66) might affect the isotope ratio of oxygen in wholewood by 0.3o70 x 25% x 14o70 x 60%, equal to 6 × 104%, by variation from its average concentration of 25°70.

The carbon-carbon bond linkage in cellulose is 1.541 • A and in the aromatic groups of lignin is 1.395 ,~, a difference of 10% (Weast, 62), therefore the presence of lignin, containing 75°/o carbon (Gould, 66) of which 60% is aromatic, may affect the carbon isotope ratio by 10%0 × 75o70 × 60% × 25%, or about 1°70, by variation from its average concentration of 25%. On these numerical arguments, the decision to measure stable isotope ratios in whole wood was made.

Many tree ring sequences can be counted with an accuracy of about one year. Those which are not yet tied

318 L.M. Libby and L.J. Pandolfi

to modern sequences by overlapping ring patterns (said to be " f loa t ing" ) , can be dated in favorable cases with an accuracy of about 30 yr by radiocarbon dating, depending on the age and the number of radiocarbon measurements which are made. But to prove the hypothesis that trees are thermometers , comparison of measurements of stable isotope ratios in the tree rings with mercury thermometer records near where the trees grew was made. Thus, we could not use bristlecone pines, because there is no lengthy temperature record for hundreds of miles near their home in the White Mountains of California. Because the longest temperature records are in Europe, we obtained a German oak from the only tree ring laboratory then existing in Europe, the laboratory of Bruno Huber in Munich (Huber, 69), where the oak rings were counted and labeled with the numbers of the years in which they grew.

From the laboratory of K.Y. Kigoshi in Tokyo, we obtained a counted 2000-yr ring sequence of a cedar f rom the southern tip of Japan, in which Kigoshi made 50 radiocarbon measurements for verification. Although radiocarbon dating's accuracy is only about 40 yr, by making 50 measurements the verification achieves an accuracy of 40/(50)½ or about 6 yr,

The temperature records needed for comparison with the German oaks exist at nearby Basel and Geneva, extending back more than two centuries, and in middle England, extending back three centuries. Temperature records for the cedar exist at nearby Miyazaki, Japan, since 1890. In addition, there are surrogate climate records for the far east in the form of records of dates of cherry trees blooming, of first freezing dates of lakes, and of number of snowy days per year.

The magnitude of the isotope variations could be predicted in trees because, for 15 yr, the International Atomic Energy Agency in Vienna (IAEA) has measured and published the stable isotope variations in rain and snow, month by month, for some 100 world- wide weather stations, including those in Germany, Austria, and Japan.

There is compelling evidence that when sapwood passes into heartwood it becomes sealed against sap and therefore against isotope exchange with sap. It has been shown, at least in species having tight rings, that the sap does not interact with the heartwood. Huber (35) has shown, using biological dyes in many tree species, that water conduction remains limited to the outermost annual ring. In addition, radiocarbon injected into or ingested by a living tree does riot exchange with neighboring rings. Furthermore, we have observed that when Sequoia gigantea wood was soaked for 37 days in an air a tmosphere saturated at 100% humidity, and previously labeled to delta Dsmow = 1170 parts per thousand, no deuterium exchange was observed even though water saturation was optically evident. This is reasonable to expect because wood is a polymer, containing large molecules, largely cross linked inter- nally and to each other with hydrogen bonds. Further- more, the bonds of cellulose and lignin are strong by comparison with most other molecular bonds, so that isotope exchange with water or dissolved CO2 becomes very improbable (Libby, 76).

To separate lignin and cellulose from each other requires wet chemistry. The review article of Taube (56) describes exchange of hydroxyl oxygen ( - - O - - H ) , carbonyl oxygen ( - - C - - O ) , and ether oxygen ( - - C - - O - - C ) , under many conditions of acidity and alkalinity in aqueous solution, and with ketones, alde- hydes, and alcohols. Besides isotope exchange in homo- geneous systems, it is also known that isotope exchange occurs when wood is finely ground and soaked in hydrogen-containing liquids (Sepall, 61) which is a heterogeneous system. Taking these facts into con- sideration, we decided to use only dry chemistries for evolution of H 2 and CO2 f rom ground whole wood.

For the first tree sequence (Libby, 73a, b, 74a, b, 76) D / H was measured by reacting sawdust with uranium to produce H2, 99°70 quantitatively. For measurement of OI8/O16, a modified method of Rittenberg and Pontecorvo (Rittenberg, 56) was used by carrying it out at high temperatures, quantitatively at a temperature of 525°C or higher. To measure the stable isotope ratio in carbon, sawdust was burned to completion in oxygen.

Whether the oxygen in tree rings comes from water or from CO2 is irrelevant, because Cohn and Urey (38) showed that isotopic equilibrium between the two substances is obtained in a damp atmosphere within a few hours.

For measurement of the oaks we used, perforce, a mass spectrometer of somewhat low accuracy, and achieved the accuracy to demonstrate that trees are thermometers by making many measurements on each sample. On the tree sequencies which were measured later, a high precision spectometer with accuracies of _+ 0.1 parts per thousand (ppt) for O18/O 16 and C13/C12, and _+ 2 ppt for D / H was used. The measurements are expressed in terms of 6 D and 618, defined in the last part of this paper.

Sample preparation

A groove is milled perpendicular to the tree rings, that is, along the radius of the tree; sawdust from each few rings is collected into an individual vial with a camel 's hair brush. The vials are dried at 50°C and capped off to protect the dried sawdust from damp air.

To evolve CO2 for measurement of O18/O16: pump for 4 h on 3 mg of sawdust mixed with 120 mg Hg CI 2 in a vacuum sealed container. Heat at 525°C for 4 h; if the temperature is lower than 525°C, production of CO2 does not quantitatively remove oxygen. Boil with triple-distilled quinoline until quinoline turns yellow. Freeze in ethanol dry-ice-slurry at -120°C. Pass gas through two dry ice-acetone traps.

To evolve H 2 for measurement of D / H : burn 5 mg dry sawdust in 1 atm Q2 in a cupric oxide furnace at 750°C. Use oxygen purified over silica gel and cupric oxide to ensure that the 02 is hydrogen free. Freeze out H 2 0 and CO2 in a liquid oxygen trap, then release CO2 at dry ice temperature. React H20 vapor on clean uranium shavings at 950°C, thus producing H 2 quantitatively.

To evolve CO2 for measurement of C13/CI2: burn 3 mg sawdust in dry, clean oxygen gas.

Tree thermometers and commodities 319

Recent trees and thermometer records

To prove that trees are thermometers, measured values were needed on modern trees whose rings have been correctly counted, and to obtain lengthy mercury thermometer records from places near where the trees grew, preferably from the same altitude. The oldest temperature records in the World Weather Records (W.W.R. , 66) are for Basel and Geneva, each at about 300 m altitude. From the laboratory of Bruno Huber we obtained an oak grown in middle Germany at an altitude of about 300 m, which had been correlated with the fiducial tree ring sequential pattern developed by Huber and his colleagues (Huber, 69). Later samples of similarly calibrated German oaks from his students, B. Becker and D. Eckstein (Libby, 74b, 76) were also obtained.

4.8

4 . 5

ENGLAND'S WINTER _ • ~

TEMPERATURE • • • DEGREES CENTIGRADE

200 • , • . • ,

The isotope ratios in four modern trees of four different species at four different altitudes, latitudes, and longitudes, were measured (i.e.: German oaks Quercus petraea, Bavarian fir Abies alba, Japanese cedar Cryptomeria japonica, and California redwood Sequoia gigantea) and compared with local temperature records from mercury thermometers. For the parts of the oaks and the cedar extending beyond the beginning of mercury thermometer records, the measured isotope ratios were compared with surrogate evidence of climate change such as lateness of cherry tree bloomings, number of snowy days per year, and number of days per year when lakes were frozen. Signi- ficant correlations were found.

In Fig. 1, carbon, hydrogen and oxygen isotope measurements for the German oaks are compared with English temperature records (Manley, 53, 59) back to

e ~ I ~ •

q #• " ~

• ID

8 D ppt 10C

18 ppt

'~13 ppt

3.6

• ",,j

Ak 4 • & •

• A~AA •

®

%

8

7 L I

6

i 5 x

X X XxXx

4 ; )0( X x l X

3 XX

2 X

I X 1750

1750 1800 1850

• • &

I

co~., 4,

c~ .3o,

xx X X

X x X

1800 x x 1850

YEAR - AD

x x x

xXx x

/

I

x

YX xx X x x~

1900 1950

1900 1950 2000

Fig. i. Deuterium, carbon, and oxygen isotope ratios vs temperature in a German oak, 1700-1950 A.D.


the time of the invention of the mercury thermometer in the late seventeenth century. Temperature records of Basel and Geneva f rom 1850 resemble those of middle England and thus fit the isotope ratios well. The English thermometer records are shown here because they extend further into the past. The carbon isotope ratios for the years 1890-1950 have been corrected for the effect of fossil CO2 production (Suess effect); the maximum correction, that for the year 1950 A.D. , was taken as 8.4°7o or a maximum increase of 2.1 ppt in C13/ Cl2, because in wood from rings of 1920 A.D. , two radiocarbon dates were measured by Berger, U.C.L.A. , as 375 + 35 yr old with respect to 1950 A.D. , whereas the actual age is 1950-1920= 30 yr, corresponding to 4.2% dilution of atmospheric C1402 by inert CO2 produced by man ' s burning of coal and oil up to 1920. In 1950 the correction for CJ4 in this particular tree should be 8.4°70, and the correction for C13 dilution should be 8.4°/o of 25 ppt or 2.1 ppt.

Direct measurements of O]8/O16 in rain and snow have been made and are available in publications of the International Atomic Energy Agency (IAEA, 69-75); Fig. 2 shows the 180 and D correlations with temperature for Austrian stations for the past 15 yr. The isotope concentration of precipitation varies similarly with temperature in many other places as shown by plots of the IAEA isotope measurements against air temperature.

Figure 3 shows the O18/O 16 ratio for a Bavarian fir, Abies alba, the rings of which were counted by Becker and Siebenlist (Becker, 70) compared with temperature records made near where the tree grew, both coming f rom 1000 m altitude on the north slopes of the Alps. Due to the mountainous terrain, local differences in temperature may be expected.

Figure 4 shows O18/O 16 in a Cryptomeria japonica compared with a local temperature record, in southern Japan, from the weather station at Miyazaki. The tree grew at 1350 m altitude whereas Miyazaki is at sea level.

Figure 5 shows O18/O 16 in a modern part of a Sequoia gigantea which grew in the Giant Forest of Sequoia and King's Creek Canyon National Park,

Three Rivers, California, at 1940 m altitude (Daughterty, 76; Michael, 76) compared with air temperatures for the same years measured in Yosemite National Park, about 100 miles south, at an altitude of 1200 m. Summer air temperatures are compared because the Sequoias in King's Canyon grow for the most part in summer when many inches of rain fall from clouds coming from the Gulf of Mexico.

Old trees and surrogate evidence

In Fig. 6 oxygen measurements are shown for oaks which grew before thermometers were invented, indicating warm intervals at A.D. 1530, 1580, and 1650, and cold periods at about A.D. 1700 and 1800 in agreement with Bergthorsson's deductions of climatic variations in Iceland (Bergthorsson, 62), and also in agreement with the l~istorical evidence of severe climate deteriorations in the First and Second Little Ice Ages in Europe.

In Fig. 7 oxygen and hydrogen isotope ratios are shown in a Japanese Cryptomeria japonica also from the island of Yaku at the southern tip of Japan, as is the modern cedar (Fig. 4), for the years 1370-1900 A.D. , compared with temperatures deduced from old diaries for Japan and China.

In Fig. 8 the oxygen isotope ratios are shown for the same Sequoia gigantea (Fig. 5) for the years 1750-1975 A.D. It shows the same negative slope indicating deter- iorating climate as does the Japanese cedar in Fig. 7, amounting to a long term decline of about 1.6°C in Miyazaki in the last 2000 yr. This is significant especially considering that there was a decline of about 10°C in the last ice age.

Climate periods

Fourier t ransforms have been conducted (Blackman, 58) for the data in Figs 7 and 8 to deduce the power spectra of periodicities; the results are shown in Figs 9 and 10. The same periods are found in deuterium and

DELTA 018/016 TEMPERATURE DEGREES C DELTA DEUTERUIM

o

~ ~- - O 0 • OQ • O 0 Q • 0 0 ~ •

"r " " " : , " . . . . " ' *

"" i . . 2 ; "." . .

Fig. 2. Oxygen and deuterium isotope ratios vs temperature in Austrian precipitation.

Free t h e r m o m e t e r s and c o m m o d i t i e s 3 2

-20.5

~- -21.0 tD

%

-21.5 t~

-22,0

m - - AVERAGE ANNUAL TEMPERATURES AT HOHENPEISSENBERG °C

018/016 RATIOS IN BAVARIAN FIR

1.

I I I I I 1700 1750 1800 1850 1900

A.D. YEARS

Fig. 3. Oxygen i so tope ra t ios vs t e m p e r a t u r e in a Bavar i an fir.

Y A K U SUGI

- 2 2

7.0

g 6.5 m

6.0 ~m

5.5 C -i1

5.0 1950

<--7

- 6 0

-7O

a - 8 0

-90 --

-100 - -

1860

~ 016 pdb

Z • 8DsMow - - - - 5 YEARS A V E R A G E OF

I I I TEMPERATURE AT M I Y A Z A K !

1880 i 9 0 0 1920 1941) 1960 1 ~ 0

YEARS (ANNO DOMINI )

-23

Z - 2 4 ~,~

- 2 5

- 2 6

Fig. 4. Oxygen and deu te r ium i so tope ra t ios vs t e m p e r a t u r e in m o d e r n Cryptomeria japonica.

O

O

-22

225

-23

717 JULY - AUGUST- SEPTEMBER

YOSEMITE TEMPERATURE

8 o'% '~ ~ . . / x , ~?'- \ \\

2 , ' i x , ,

1920 19130 19i40 19150 1960 19170 YEARS A D

71 o

?0 z-

69 co bA

I

B8 $ -

g

6B ~_ . t980

.>00

700

195

25 5 18 9 1910

Fig. 5. Oxygen i~otope ra t ios vs t e m p e r a t u r e in Sequoia gigantea.

- - 17 .10

u 9_

- - 16 .80 ~ I-

- - 1 6 . 5 0 m

- - 16 .20


- 1 6 0

-17 .0

aa

~ -18.0 - -

~.- -19 .0 - -

-20 .0 - -

-21.0 - -

-22 .0 - -

-23 .0 - -

1350

MARBURG, GERMANY, OAK

5 SAMPLE ( ~ 3 0 yr. ) RUNNING AVERAGE

• (4-7 yr.) SAMPLE

. . . . . . AVERAGE A N N U A L TEMPERATURE IN ENGLAND

SPESSART OAK NO, 2

X 5 SAMPLE (30 yr . ) RUNNING AVERAGE

SPESSART OAK NO, 1

7 SAMPLE (30 yr.) RUNNING AVERAGE

- - - - SMOOTHED BY EYE

I : "g

• X I: "

• X X

• " ' ' " -", :'.

x x ~" x~ x ,~ • ,?~" / . " 1'

\ /

/.: • ~ I , d •

-, ::. .,J •

I I I I I I I' I "~1 I I ,40o ,45o isoo ,s~o ,600 ,65o ,7o0 ,7s0 ,800 ,sso i~00

A. D. YEARS

1 it

:! "1 :!

. :-

~ r r i"

I 1950

- - 9 . 5

- - 9 . 4

m m

- - 9 . 3 o m

z

-- 9.2

-- 9.1

2OOO

Fig. 6. Oxygen isotope ratios in a sequence of German oaks, 1350- 1950 A.D.

oxygen in the Japanese cedar, within experimental error, as indicated from the widths of the peaks of the power spectra (Fig. 9; also Table 1).

Samples consist of wood from about 5 yr each, so one cannot expect to find meaningful evidence for periods of less than about 40 yr in these data. The Japanese cedar spans only about 1800 yr, so no meaningful evidence for periods of more than about 250 yr can be expected. In making the Fourier transforms, the deviations of the isotope ratios from the long term slopes were used - - that is, we have corrected for the slopes. We have tested the meaningfulness of the

-50

Y A K U SUGI

-70

~E

-90

-21

-22

-23

~ -24

~ -25

-26

-27 L ~ I k I 200 400 600

[ (~ D = 8 t~ol + CONST 8

800 1000 1200 1400 1600 1800

YEARS (ANNO DOMINI)

Fig. 7a. Deuterium and oxygen isotope ratios in a Japanese Crypto- meriajaponica (Yaku Sugi) from Yaku island.

periods in Table 1 by manufacturing data sets consisting of a number taken from the least squares fit to the data, namely the straight line fit, plus a random number

-22.0

,201 A/r 1 '25 "

P OI V , , 1.23 5 ~ 1400 1500 1600 A.D. " o

0 ~ .25 -(¢)

I 2 . 2 5 °C

l .40

.45 MOO 1500 1600 A.D

Fig. 7b. Oxygen isotope ratios in C. japonica calibrated for temperature change using temperatures obtained from time of cherry tree blooms (a), and from first freezing of Lake Biwa (b), and from

number of snowy days per year (c).


-20 I

- 2 1 ~ _ ~ 018/016:- 0091 l + CONSTANT

~ -22

S~ -23 - • e ,e C~ ee, ,

-24

1?50 1800 1850 1900 1950 2000

0

000

2050 YEARS A D

Fig. 8. Oxygen isotope ratios in 250 yr of Sequoia gigantea ring sequence from California. The horizontal lines represent measured values. The black circles represent values predicted from the recon- struction using periods, phases, and amplitudes obtained from the Fourier transform. The prediction has been projected into the future.

2~ r 1154 + 13y 2.32 b n JAPANESE CEDAR

] ~1 FOURIER TRANSFORM OF 618 2106 F 271 + 11y tl AVERAGE SLOPE SUBTRACTED

O~ 1.55 / 69 + 3y

~ 1.29

~ 1.03 ...1 59 + 14y

0.77 1

o

0"26 ~ 1 ' I I ] t ~

• 1.55

0 0.0039 0.0283 0.0528 0.0772 0.1017 0.1261

0.0161 0.0406 0.0650 0.0894 0.1139 2~/YEARS

Fig. 9a. Fourier transform of the oxygen isotope ratio vs time in C. japonica, transformed into power vs reciprocal period.

3.21

2.74

O~ 2.18

1.64

1.08

0.54

0 0.0127

JAPANESE CEDAR FOURIER TRANSFORM OF 6D AVERAGE SLOPE SUBTRACTED

- /,,S o7,7 /// '°°-""

| (1 ~6 +_ tt

0.038~ 0 . ~ 3 7 0.08~2 0.11~7 0.0255 0.0510 0 .076, o . t01~ 0 . t 2 , ,

2 7i"/YEA RS

Fig. 9b. Fourier transform of the deuterium isotope ratio vs time in C. japonica, transformed into power vs reciprocal period.

POWER SPECTRUM OF SEQUOIA GIGANTEA 1749 A.D - 1975 AD

15

x

l0 o_

o @ o

05

3 3 4

532y

/ J

k I0 2'0 4'0 60

142.8y

,&o 200 3~ 460 PERIOD IN YEARS

Fig. 10. Fourier transform of oxygen isotope ratio vs time is S. gigantea, transformed into power vs per iod

varying over the numerical range o f the deviat ions from the line. Each set o f manufactured data was subjected to the Fourier transform. In each case, in the transforms for 30 such manufactured data sets, no peaks o f significance were generated, indicating a conf idence level for the periods in Table 1 o f better than 90°7o.

The Fourier transforms were performed in the standard way, i .e. no s m o o t h i n g or filtering was employed . Subtract ion o f the data from the least squares fit r emoves the constant or linear term charac-

terising a Markovian process . Fourier transform o f the differences from the linear fit suppresses the enhancement o f both the power and ampl i tude spectra at low frequencies .

The periods derived from the transform o f the oxygen i so tope ratio in the S. gigantea appear to agree with those in the C. japonica as shown in Table 2. In the Sequoia, the samples were taken for as few as 2 yr at a t ime, so that it is possible to have conf idence that the 33- yr period found in it is meaningful . But because the tree


Table 1. Periods and respective powers evaluated in Fourier transforms of oxygen and hydrogen isotope ratios in a Japanese cedar A.D. 140-1950 (ref. 2) and in oxygen isotope ratios in the Greenland ice cap,

A.D. 1200-1950

(180/160)* (D/H)* (180/160)

(ice and tides) Per iod(yr) Relative Per iod(yr) Relative Per iod(yr) Relative

power power power

5 4 ± 3 0.58 5 2 ± 5 1.00 55 - - 5 7 ± 3 0.65 5 7 ± 3 0.60 62 - - 594-3 0.70 614-2 0.40 68 - - 6 9 ± 3 1.50 6 6 ± 4 1.64 78 strong 924-2 0.52 8 7 ± 3 1.00 86 - - 964-2 1.29 954-2 2.18 100 - -

154±6 2.58 153±7 2.18 142 - - 1744- 7 1.29 176 ± 7 2.00 179 strong 2 0 2 ± 7 1.45 205± 9 1.70 - - - - 271 ±11 1.80 273± 20 1.64 285 - -

Table 2. Periods, powers and phases observed in the Fourier transforms of D / H and O18/O16 measured in a Japanese Cedar and in a

California Sequoia

Cryptomeria Sequoia Japonica Gigantea 168 A.D. - 1885 A.D. 1749 A.D.- 1975 A.D. Yaku Island, Japan Three Rivers, Cal ifornia

r 157yr 143yr P2.72/100 1.59/106 ¢ 1.76 rad 1.73 rad

r 70.0 yr 78.5 t P 1.47/100 1.09/106 0 2.93 rad 5.86 rad

r 59.1 yr 53.2yr P0.79/103 1.46/106 0 3.09 rad 3.23 rad

- - 3 3 . 4

P - - 1.59/100 - - 3 . 0 6

r - - Years; P - - Power; O-- Radians.

has been analysed only for about a 225-yr span, we can not ask for meaningful evidence of periods of more than 143 yr, and although it appears to agree with the period of 157 yr found in the C. japonica, and although in our experiments with artificially generated isotope ratios in the Fourier t ransform no such large amplitudes were produced, still one must be cautious.

In the Fourier t ransform, amplitudes and phases are generated for each period. We have put the periods, amplitudes and phases back into the summation of transcendental functions which represents the function of the data, using this expression to generate the original data. By running this function into the future we have made a prediction of the climate to be expected in King's Canyon; the prediction is that the climate will continue to deteriorate on the average, but that after our present cooling-off of more than the average decay in climate, there will be a temporary warming-up followed by a greater rate of cooling-off.

Naturally, a complete analysis of the 3000-yr span of the S. gigantea will yield a more reliable prediction of the future climate for King's Canyon.

Eight of the periods found in the oxygen and hydrogen isotope ratios of the Japanese cedar are also evidenced in variations of the oxygen isotope ratio versus depth in the Greenland ice (Dansgaard, 71, as listed in Table 1 (Libby, 77)). The remarkable agreement between our tree records of oxygen and hydrogen isotopes and the ice record of oxygen isotopes shows itself in yet another way: the D / H and O18/O 16 ratios for the Japanese cedar have been found to be significantly correlated in opposite phase to the C14 variations in bristlecone pines of southern California (Libby, 76b) as measured by Suess (70). Similarly Dansgaard (71) found the oxygen isotope record in Greenland ice to be significantly correlated in opposite phase with bristlecone carbon-14. Furthermore, the Fourier t ransform of the C14 variations shows four of the same periods as in our tree and in the Greenland ice; see Table 1.

From the correlations of these four sets of data it is concluded that (1) The calculation of Dansgaard (71) of the age of ice versus depth in the Greenland ice cap seems to be correct with an error of not more than a couple of years, at least over the last 800 yr. (2) The climate variations in Greenland, southern Japan, and southern California have had the same periodicities for the last 800 yr or more. A logical explanation for the global nature of these correlations is that all four are related to variations of the sun, which cause variations in the temperature of the sea surface, thus causing variations in the isotopic composit ion of water vapor. This water vapor distills o f f the sea surface, nourishes trees, is stored as wood in tree rings, and forms the annual layers of the icecap. The variations of the sun are fur thermore related to mass of the total biosphere, causing the small variations in C14 content of the bristle- cones (as well as variations in solar wind).

Two additional data sequences (Kalil, 76, 78) versus age are reported in which the same periodicities are revealed, namely the organic carbon and uranium concentrations in an ocean-bot tom sediment core taken from the Santa Barbara Basin of f the coast of California. Preservation of annual varves in this anoxic sediment provides a record of the age of the sediments, there being one varve or distinct layer for each year. The concentrations, versus depth, of organic carbon and uranium were measured in a continuous sequence of samples, each containing 7 yr of sediment, in sea core PT-8G, spanning the years 1264-1970 A.D. The age of sediment versus depth in the core was determined by comparison of its varve patterns with those in core 214 in which the varves had been counted (Doose, 78; Soutar, 69).

The age determination allowed Fourier t ransforms to be made, t ransforming concentration versus depth into signal power and amplitude versus the period expressed in years. The sea core spans 700 yr, with each sample containing seven years of sediment, so that evidence for periods between 40 and 200 yr should be meaningful (Pandolfi , 78).

Eight periods found in uranium and organic carbon variations in the sea core are also found in stable isotopes of hydrogen and oxygen in the tree, and in oxygen isotope variations in the ice, enriching the inter- pretation of climate variations on a global scale, and

Tree t h e r m o m e t e r s and c o m m o d i t i e s 325

enriching the attr ibution of these periodicities to variations of the sun causing changes in sea surface temperature. The temperature of the sea surface determines the abundance of life, and therefore, the abundance of organic matter which falls to the ocean bot tom and binds uranium ions in the sea water to it as it falls. (It is well known that in uranium ores, the amount of organic carbon is proport ioal to the amount of uranium.)

The recently completed C l A P study of the stratosphere (Grobecker, 75) indicated mechanisms by which the sun 's UV may influence temperature on the ear th 's surface. In particular, solar energy absorbed in the stratosphere is rapidly converted by chemical reactions, producing a variety of chemical species which reradiate both to space and to the ground. Thus, climate on the ground is sensitive to variations in stratospheric chemical components . The solar light is known to vary by factors of 2 and 3 in intensity at short wave lengths, affecting the concentrations of those species. Reradiation to the ground is especially sensitive to ozone and water concentrations in the stratosphere, and their variations are only just beginning to be assessed. Study of stable isotopes in trees and of uranium and organic carbon in sea sediments as a function of time in the past may be a very powerful way of studying past misbehaviours of the sun.

The most well known misbehaviour, i.e. variation, o f the sun is varying the number of its sun spots, with a change in number between the north and south hemispheres of the sun every 10.5 yr, and a completed magnetic sun spot cycle every 21 yr on the average. Sup- posing that the sun's variations have caused the stable isotopes in our tree, the stable oxygen isotopes in the ice cap, the organic carbon and uranium in the icecap, and the carbon-14 in the bristlecone pines to vary with the same periodicities, then it seemed reasonable to look for evidences of the 10.5 and 21-yr sunspot cycles in our trees. In order to find meaningful evidence for such short periods, it is necessary to measure the stable isotope ratios in every single year of a tree ring sequence.

Hurt (77) measured the oxygen isotope ratio in a ring sequence of 72 yr in a B.C. sample of Sequoia gigantea. He chose this sequence because the rings are unusually wide and consequently easy to separate. The results are shown in Fig. 11. The Fourier t ransform of the ratios, shown in Fig. 12, has significant power signals at 6.69, 10.4, and 37.4 yr. To determine the significance of these signals we generated 30 random number sequences of 72 numbers each having values between -18 and -21 ppt, the range of the values of the measurements in Fig. 11. The Fourier t ransforms made of these computer experiments showed peaks with more than half the amplitude of those in Fig. 12, indicating that the periods of 6.69 and 10.4 yr are real, with a confidence of 90°70 or better (no peaks in 30 experiments with random number sets).

The reality of the large peak at 37.4 yr might be ques- tioned because 37.4 yr is a large fraction of 72 yr; however, it appeared in the Fourier t ransform of the oxygen isotope ratios of the modern part of the same Sequoia (1749-1975 A.D.) as 33.4 yr, so it is believed to

be real. Furthermore, the data sequence was cut from 72

-18 0C I ~ I I I r

-190C * ° •

~ - , , o ,Jo ' ,;o ,;o ,~'o 200 ,',o ,o;o Y E A R S B C

Fig. 11. Oxygen isotope ra t io vs t ime in S . g i g a n t e a , m e a s u r e d for each year .

700

600

500

40O r~

300

2OO

I00

SPECTRAL DISTRIBUTION OF 8018 (SOLID LINE) IN TREE & RANDOM NUMBERS (CROSSES)

x x x x x x ) ~ ~ x x x

x x x Xx x x x' x ~ x × ~ x ,,-,, x x ' I ~ " " ~,,/~' ] x~x " x

/ XxX x x x

x.,~-kj, , , .~=...

3 5 IO 20 30 50 mOO

PERIOD (yrs)

Fig. 12. Four ie r t r a n s f o r m , power vs per iod , o f oxygen isotope rat ios vs t ime, m e a s u r e d year by year.

to 63 yr, and its Fourier t ransform was made, revealing that the 37-yr signal remained unaffected, and not an artifact of a half harmonic of 72 yr.

For comparison with the periodicities in the sunspot cycle, Fourier t ransforms of the 350 yr of the sunspot counts were made (Waldemeier, 61). By assuming all sunspot counts to be positive numbers, (Cohen, 74), the power spectrum shown in Fig. 13 was obtained, with signals at 9.94, 10.47, and 11.06 yr, for a strong signal averaging to 10.6 yr. By assuming sunspot counts to have opposite magnetic polarity in the northern f rom the southern hemisphere of the sun (Hill, 77), the power spectrum shown in Fig. 14, with a strong signal at the well known period o f 21 yr, was obtained.

The fact that the oxygen isotope power spectrum in Fig. 12 does not show a significant signal at 21 yr, but shows a strong signal at 10.4 yr (which could easily be 10.5 yr within our errors) indicates that the influence of

326 L .M. Libby and L.J. Pandolfi

50-

40-

x ec 50- k ~

2 20-

5 10

IZ06, I0.47 ~ 994y

SPECTRAL DISTRIBUTION OF SUNSPOT CYCLE (ASSUMING ALL SUN SPOT COUNTS TO BE POSITIVE NUMBERS)

50 100 200 400

PERIOD (yrs)

supportive o f the conclusions of the C l A P study (Grobecker, 75) that solar UV rays influence the abundances o f many kinds o f chemical species in the stratosphere, and therefore influence the amount o f solar energy they absorb and reradiate to earth. Thus, they also influence the surface temperature of the earth and especially the surface temperature o f the oceans. It is the surface temperature o f the oceans which produces the variations in the phenomena discussed: the isotope ratios in rain and hence in tree rings, the isotope ratios in the Greenland ice cap, the organic carbon content and the uranium content o f sea bottom sediments; and its variations o f the sun which produce variations in solar wind flux incident on the earth's atmosphere producing variations in the carbon-14 content o f atmospheric CO2 and hence in the carbon-14 content o f tree rings (Hurt, 78).

Fig. 13. Fourier transform of sunspot numbers, counted year by year , for 250 yr, transformed into power vs period, assuming all sunspot

numbers to be identical in magnetic polarity.

X

Ct7 u j

o

2,50

200

SPECTRAL DISTRIBUTION OF "~ SUNSPOT CYCLE (ASSUMING

SUNSPOT NUMBERS TO HAVE OPPOSITE SIGNS IN N & S HEMISPHERES)

15(

IOC ~,

10 50 100 PERIOD (yrs.)

280

Fig. 14. Fourier transform of sunspot numbers, counted year by year, for 250 yr, transformed into power vs period, assuming sunspots to have opposite magnetic polarity in northern and southern hemi-

spheres.

the sunspot cycle on the earth's climate is effected by neutral particles (which are affected by the number o f spots but not by their magnetic polarity because they are neutral), probably photons. Neutral particles are not deflected by the earth's magnetic field, so they are able to come to the earth symetrically from either hemisphere o f the sun.

The signals in the oxygen isotope record from the S. gigantea's tree rings at 6.69 and at 37.4 yr may or may not be related to periodicities o f the sun. The sun has many ways to vary, apart from the sun spot cycle, such as fluctuations in frequency o f solar flares and plages, and misbehaviors o f the overall magnetic field of the sun and o f the solar corona.

The conclusions of Hurt's study o f year-by-year oxygen isotope ratios in 72 yr o f S. gigantea are thus

Other climate indicators: commodities, prices, wages

The biological reservoir o f organic carbon and the influence o f climate on its variations are o f major importance. The variations in yields o f bio-organic commodit ies , and the variations in their prices, and the variations in the ability o f the human population to buy the bio-organic commodit ies (wages), caused by changes in climate are discussed. Therefore, we identify the changes in these historic quantities as indicators o f historic climate change.

The blue crab (Callinectes sapidus) catches in the Chesapeake Bay have been increasing (Franklin, 77) for the years 1922-1976 A .D . , see Fig. 15. A Fourier transform was made o f the yield in millions o f pounds per year (1 million pounds = 454 tonnes) versus years into amplitude versus period in years, see Fig. 16, and principle cyclic periods o f 8.6, 10.7, and 18.0 yr were found.

IOO

o

o 60

50

40

i'i

- - Annual Calch Of Blue Crab In Chesapeake Bay ..... Predicted

,/,,,, .:""

/ i :=

195

150

~05 ,..o

360 . u

315 t,_ 0

2 7 0 ~ .._1

225 ~

180

155

t9'20 19'50 19'40 t9'50 1960 19'70 1980 19~90 -- 9 0 YEARS A D

Fig. 15. Annual catch of blue crab measured in millions of pounds per year in Chesapeake Bay, 1920-1976 A . D .


mr"

>--

h l O - -

Z

J

Z

b---

d

SPECTRUM OF ANNUAL CATCH OF BLUE CRAB IN CHESAPEAKE BAY

~80

107

J I i i i I

6 8 10 I I i

io 3'o 4o so G'o 7o PERIOD IN YEARS

Fig. 16. Fourier transform of annual catch of blue crab vs time into amplitude vs period.

findings of the CIAP study (Grobecker, 75), in that climate on the ground is affected by variations in the chemical species in the stratosphere whose concentration varies as the sun's spectrum and intensity varies.

Why temperature and rainfall near Chesapeake Bay should be affected by variations of the tidal forces is not so clear. However, the atmosphere and stratosphere are pulled away from the earth by tidal forces just as are the waters of the earth. These forces vary by as much as 10070 during the tidal periods (Rinehart, 72), resulting in density variations in the stratosphere with the same periods; the consequent density variations may affect the relative rates of stratospheric chemical reactions, causing disturbances of temperature and rainfall on the ground with the tidal periodicities.

The periods of blue crab and their absolute phases and amplitudes listed in Table 3 have been used to reconstruct the annual yield of blue crab, Y, and project it into the future (shown by the dashed line in Fig. 15). The prediction formula is:

Y = ~ A sin (2 r t t / r+epn)

The Fourier t ransform of the record of annual air temperature versus time shows principle periods at 7.4, 9.8, and 17.5 yr; the Fourier t ransform of the record of annual rainfall at Philadelphia shows periods of low rainfall of 8.3, 9.5, and 18.5 yr (Landsberg, 75, 76).

Tidal forces have periods of 8.8 and 18.6 yr (Rinehart, 72). The variations in the tidal forces arise because the orbital planes of the moon and the earth are slightly inclined with respect to each other and because the sun, moon, and earth form a mutual rotating gravitational system so that the magnitude of the tidal force depends on their relative positions.

For comparison, the periods of enhanced blue crab yield, of maxima of air temperatures at Philadelphia (which is close to the Chesapeake Bay), of minima of rainfall at Philadelphia, and of enhanced tidal forces (leading to high tides) are listed in Table 3. The explanation offered for the agreement among the periods so listed is that high tides wash nutrient into the surface waters of the Bay, and higher temperatures warm the surface waters, and minimum rainfall allows the surface waters to become more saline, all of which factors are conducive to crab growth.

Variation of temperature and rainfall with the sunspot periodicity of 10.5 yr is explained by the

Table 3. Periods, relative amplitudes and relative phases, observed in the Fourier Transforms of annual temperature and rainfall records in Philadelphia, in annual blue crab catch in Chesapeake Bay, and cal-

culated for earth-sun-moon tidal forces

Annual average of Annual average of Philadelphia Philadelphia temperature rainfall (minima)

Annual average of Earth- Blue Crab Catch in Sun-

Chesapeake Bay Moon- Tidal Forces

T ~ A r ~ A r ~ A 17.5 0 1.0 18.5 0 1.0 18.0 0 1.0 18.6 9.8 1.9 1.2 9.5 4.1 1.1 10.7 0.9 0.5 - - 7.4 1.4 0.9 8.3 2.5 0.9 8.6 2.1 0.3 8.8

where ~n is the phase of the nth period r n and A n is its amplitude.

The prediction in Fig. 15 shows the recovery of the crab crop already observed in 1977 (Franklin, 77), to increase until 1982 and to remain at high yield until 1988, followed by a decrease. The absolute phases and amplitudes are to be regarded as inaccurate, however, because their values depend sensitively on the accuracy of the input data. This may be the reason for the disagreements of the relative phases shown in Table 3 (Hurt , 79).

As with all of our Fourier t ransforms of real data, the statistical significance of the periods so revealed were tested by generating appropriate sets of Markovian data, each datum consisting of a constant, a, plus a random number e j, where the random number ej is small compared with a, and varies in such a way that the artificial data set so generated ranges between the maximum and minimum of the values of the real data. Each set so generated is subjected to the Fourier t ransform, peaks of amplitude corresponding to those found for the real data are noted. If, in 30 such experiments, no comparable peaks are found, it is concluded that the periods found in the true data set have a significance of better than 1/30, namely better than 90°70 confidence level. This has been the case of all our Fourier t ransforms of sets of real data presented.

Records of delicacies such as crab extend for less than 100 yr. The commodi ty for which quantitative records of price (contrasted to yield) extend over the longest span, to our knowledge, is European wheat. Historic prices were found for England, Netherlands, France, and northern Italy since 1250 A.D. in units of guilders per 100 kg, see Fig. 17 (Veenman, 38). In all four countries there were many-fold increases in prices at around 1600 and 1800 A.D. , caused by something more general than local political and economic manoeu- vering.

gld. / lO0 kg E/~r~ I. ~5 ,V_t._

"~, ~?.. i

12 ~ : • '

/~.. ,.," ~ , . , . . . .. ~.

6 . - / :.: i:" ": ........ "2" ;' ..-" " / . . F r ; :

I / I

1:200 1500 1400 1500 o 16~0 , i IT00 1800 1900 YEARS A.D.

..=-,

=~ L L I

Fig. 17. Price of wheat in Italy, France, Netherlands, and England, expressed in guilders per 100 kg., vs time, 1200-1900 A.D.

Economists agree that there exists a definite relation between the market price of a commodity, such as wheat, and the quantity available (Samuelson, 70). Thus the price of wheat, in historic times of plenty and in times of short supply, should be an indicator of abundance in good climate and of shortage in bad climate, and therefore an indicator of climatic changes.

The prices of wheat and wages (ability to pay for wheat) from 1250 A.D. to the present, are compared with variations of average air temperatures from 1650 A.D. to the present as a calibration of the modern part of the wage and price record. The population increased during this time span; thus corrections for this were made. It was also necessary to correct for inflation of money. In Fig. 18 are shown the prices per troy ounce for sterling silver (90% pure) and for fine (pure) gold, in English money, since 1250 A.D. Sterling prices are more basic than gold prices because sterling was the only official coin of the realm until 1774.

I000 I

125

Ioo

50

io [ ,

1200 1400 1600 1800 2000 ,

1200


1400 1800 1800

YEARS A.D.

100

20

2000

Fig. 18. Cost o f sterling silver and o f pure gold vs time, expressed in English pennies per troy ounce, 1000-1950 A.D.

One way of normalizing grain prices to take into account inflation is to compute the ratio Rs of wheat prices to price of sterling. A second way of normalizing is to compute the amount of grain which a laborer's daily wage can buy, namely Rw, the ratio of labor's daily wage to wheat price. Rs and Rw are shown plotted in Fig. 19 versus centuries. R s has been computed from data in Fig. 17 and 18. Rw has been taken from Meredith (39) and Steffen (01). The third part of Fig. 19 shows a 25-yr running average of winter air temperatures in England computed from Manley (74).

The warming trend between 1800 and 1930 is generally displayed in temperature records of Holland, Edinburgh, Stockholm, Vienna, Berlin, Copenhagen, Greenland, Ukraine, Siberia, Basel and Geneva; in the United States the trend is evident in New Haven, Philadelphia, St. Paul, St. Louis and Washington, D.C. (Ladurie, 71). The amplitude of the change varies locally from 1 to 2°C. In particular, the warming of the entire northern temperate zone is esti- mated at 0.64-0.70°C (Ladurie, 71) for the entire period.

m c , , ~ i B _ j

x

H ~STORIC F RST SECOND ~'/ OLD LITTLE LITTLE !

ICLIMATE ~, ICE AGE ICE AGE i

t ~, I ' . . ; '~ ,^,/ , . . ,

D " ' ' ' ' ' ' ' ' ' ' ' ' D

~ . 01

o~ 02

05

1300 1400 1500 1600

44 ~ ¢ .~ z t ~ a

42 ~ =

4o~ 3 8 ~

1700 -1800 1900

=I[

4 5 ~ , P

. ,o .4" ~ . . % .~'.,4,. 40 ~ •

.... • I n'oo ' m'oo ' 35

Y E A R S A D

Fig. 19. Variations in amount o f wheat purchasable with daily wage of an agricultural laborer, a carpenter, compared with amount necessary to nourish a family, and compared with cost of wheat, and with

temperatures 1650-1900 A.D.

The major declines of Rw and R s with a minimum at 1630 A.D. are known as the First Little Ice Age, a time famous for famines killing millions of people, cold


summers, bitter winters, failed harvests, wars and civil unrest. The declines of Rw and Rs reaching a minimum at around 1800 A.D. correspond to the Second Little Ice Age, which caused extensively documented civilian hardships.

The purpose was to obtain the temperature scale shown in Fig. 19, applicable to times as far back as 1250 A.D. when there were no thermometer records, in order to compare the temperature record so obtained with the record of Tree Thermometers . To this end, we measured ARw, ARs, and A T (from Fig. 19), f rom peaks to valleys, since 1700 A.D. and f rom these, computed the temperature coefficients A 7~/ARw and A T/ARs, and using these coefficients, we computed the temperatures for 1250-1600 A.D. The resultant temperature scale, shown on the right hand ordinate of Fig. 19, (Libby, 77), is in agreement with that obtained f rom variations of the stable isotope ratios in the European oaks.

The bio-organic reservoir

The biosphere is an important reservoir of carbon, derived f rom atmospheric CO 2, which it both obtains f rom the atmosphere and returns to the atmosphere as it decays. Its mass depends on the CO2 content of the atmosphere, on air temperature, on sea surface temperature, and concentration of carboxyanhydrase in the surface sea water (namely, on climate). The dependence on rate of absorption of CO2 into surface sea water has been shown by Berger and Libby (Berger, 70). In turn, the concentration of CO2 in the a tmosphere depends on the mass of the biosphere and its rate of decay, and on the carboxyanhydrase concentration in the atmosphere. In future predictions of the rate of increase of CO2 partial pressure in the a tmosphere it will be important to include the interaction with the bio-organic reservoir and the catalysation of its absorpt ion into the sea by means of the action of carboxyanhydrase, conside- rations which have not been taken account of in past computat ions.

In Table 4 are listed data which describe the interaction of the system: atmospheric CO2 + CO2 in ocean water + CO2 stored in the bio-organic reservoir. From these data, it is shown (Libby, 73) that changes in climate by +_ 10°C can cause changes in the carbon-14 content o f bio-organic matter , and therefore in

Table 4.

Total atmospheric CO 2 = 0.13 g C/cm 2 Mixed layer of ocean ( 100 m thick surface waters) contains 0o 15 g C /

cm 2 0.013 g C /cm 2 exchanged from atmosphere to oceans each year 0.001 g C/cm 2 returned to atmosphere per year by plant decay Storage time in biosphere is about 140 yr Storage time in oceans is about 1000 yr

apparent carbon-14 age of as much as _+ 100 yr. Such changes are well able to account for Suess's "wiggles" in the carbon-14 concentrations versus time measured in the 5000-yr sequence of bristlecone pines constructed by W. Ferguson. This interaction between climate and the biosphere is the likely explanation for the agreement of the periods listed in Table I.

The slope of eight

The linear relationship of the isotope ratios of hydrogen and oxygen with a slope of 8 in rain and snow was discovered by Craig (61) to be

8D = 8818 + const.

where

518 = [ ( 0 1 8 / 0 1 6 ) / ( 0 1 8 / 0 1 6 ) s m o w - - I ] X 10 a

5D = [(D/I ' t ) / (D/H)smo w - 11 X l 0 3

and where the ratios labeled " s m o w " are those for standard mean ocean water.

This relation has been abundant ly verified for rain and snow collected world-wide, and measured by the IAEA since 1969 (IAEA, 69-75). The IAEA data are shown in Fig. 20 as a plot o f 5 D versus 518 with the characteristic slope of 8.

This slope can readily be computed f rom laboratory measurements (e.g. Stewart, 1975 and references quoted therein); for the temperature range 0°-35°C these measurements can be represented as,

( D / H ~ i q / (D/H)v -- - 9 X 10 -4 T + 1.1035

( 018 / 01 6)liq / ( 0 1 ~ 16 /0 ) v = - 9 X 10"5T+1.01135

where the subscripts " l i q " and " v " mean liquid and vapor, and where Tis the temperature in degrees C.

So, for rain (made from vapor distilled from the ocean),

5D = [(D/H)v / (D//-/)liq - I ] X 103 = I0 a ]

(- 9 X 10 -4 t + 1.1035)

= I11 16 8/0 l 51S [(0 /0 )v / (01 6)liq - 1] X 103 = 10 a ]

( - 9 × 1 0 5 + 1 . 0 1 1 3 5 ) .

Thus the derivatives of D and 5 ia with respect to temperature are,

A(5D) / (AT) = 0.9 / (-- 9 X 10 -4 + 1.1035) 2

A(51 s) ] (AT) = 0.09 / ( - 9 X 10"ST + 1.01135) 2 .

The ratio of these two derivatives gives the slope of 8 of the rainwater curve in Fig. 20. In fact the IAEA data yield a far more accurate determination of the slope than can be made in the laboratory.

330 L .M. Libby and L.J. Pandolfi

Using the above laboratory relations one computes the values o f 8 D and ~18 expected for rain made by a single distillation from the ocean, and for rain made by a subsequent distillation for sea rain which fell on the land. These values, listed in Table 4, show that when rain and snow exhibit very large depletions o f the heavy isotopes, e.g. 8 D o f -200 and -300 ppt, two or three distillations have occurred. It is known from measurements o f tritium in rain (Levinthal, 68) that two or three distillations occur in the US between evaporation from the Pacific Ocean and precipitation in the midwestern US. The agreement between the number o f distillations deduced from stable isotopes and deduced from radio- tritium is gratifying.

For every point on the curve in Fig. 20 characterising rain and snow, there is a corresponding average air temperature, as may be seen in Fig. 21. Consequently, a similar dependence on the slope o f 8 occurs in trees which ingest rain. The slope o f 8 was found in a Japanese cypress (Libby, 76) measured for the years 1500-1970 A.D. But in the sequence of German oaks which we measured, (Libby, 73a, b, 74a, b) a slope o f 8 was not found at first.

It is now stated that a slope o f 8 occurs in all o f the trees which we have measured. The explanation is as fol lows. For the oaks, all isotope ratios were measured using w o o d for the years 1712-1714 A.D. as the standard (as described in Libby (74b)). In those years

~ 9 5 3 - 19631 I N T E R N A T I O N A L A T O M I C E N E R G Y A G E N C Y , V I E N N A , 1969

--30(] - - E N V I R O N M E N T A L ISOTOPE D A T A NO 1; WORLD S U R V E Y OF ISOTOPE C O N C E N T R A T I O N IN P R E C I P I T A T I O N j 9 5 3 -- 1963 T E C H N I C A L REPORT SERIES NO. 96

O z

- 2 0 0 0 I

- 1 s o

O/H

~ -1~ I B

-50

+50

P >,,."

• ~ . r ~

S '~o,

C. T A Y L O R

~ 1÷3°C, C. T A Y L O R )

D = 7.795018 + 4.14 0.99132 C O R R E L A T I O N C O E F F I C I E N T t . 084 ERROR ON INTERCEPT

N = 1 1 0 1

/

. 1 0 0 ÷10 +6 0 - 5 - - 1 0 - 1 5 - 2 0 - 2 5 - 3 0 - 3 5 - 4 0 - 4 6

6{018/016) iN PARTS PER T H O U S A N D

--S0

• B E T H E L . A L A S K A 61 ° I1 L A T = T E H E R A N . I R A N 36 o N L A T o GENOA, I T A L Y 44 ° N L AT = NEW D E L H I , I N D I A 28.( o N L A T

• K H A R T O U M . S U D A N 15 6 ° N L A T S E Y C H E L L E S 4.6 ° N L A "

• NORD G R E E N L A N D 81.6 ° N L A T

o GOOSE B A Y . N E W F U N O L A N O 53 ° N L A T GREENE ~)AL. G R E E N L A N D 810 N L A T

= A Z O R E S P O R T U G A L 37.8 ° N L A T V I E N N A . A U S T R I A 48 ° N L A T

• K A R A C H I , P A K I S T A N 25 ° N L A T WIND H O E K . S. A F R I C A 22 .6 o S L A T

• ASCENSION. I S L A N D 8 ° N L A T D A R ES S A L A A M . T A N Z A N I A 7 ° S L A T

• ISFJORO. N O R W A Y 78 ° N L A T • L I S T A . N O R W A Y 58 ° N L A T • V A L E N T I A , I R E L A N D 52 ° N L A T * R E K J A V I K . I C E L A N D 64 ° N L A T • G I B R A L T A R . U K 36 ° N L A T n S T U T T G A R T , W. G E R M A N Y 49 ° N L A T - B A H R A I N . P E R S I A N G U L F 26 ° N L A T

• B O M B A Y . I N D I A 19 o N L A T S A L I S B U R Y . R H O D E S I A 18 ° S L A T

• DEEP ICE WELL . B Y R D S T A T I O N . A N T A R C T I C A , S. EPSTEIN, R.P. SHARP. A.S. COW 1000 B.P. TO 75.000 B. a. ( E S T I M A T E D ) ~.CIENCE. 168. 1 5 7 0 - 1 5 7 2 (1970)

• TROPOSPHERIC VAP( )R . 5 K M A L T I T U D E . C.B. T A Y L O R . I N S - R - - 1 0 7 . FEB 1972. INST N U C L E A R SCI, LOWER H U T T . NEW Z E A L A N D (PREPRINT)

Fig. 20. Deuterium isotope ratio vs oxygen isotope ratio for world- wide precipitation (IAEA data) showing the slope of 8.


- 4 0

- 3 5

- 3 0

- 2 5

-20 ̧ ¢o

- 1 5

<

< <.

< • (

"1< " ?

~ O t 8 - - 9 . 3 6 + 0 . ~ 7 T - 0 . 0 7 3 T 2

n = 213 1 9 6 6 - 1 9 6 7 I A E A

, ! I • : v V

i 0 O I

-~o ~00

-5

0 . . . . . .

+ 5

< < I

< •

QV • J O - ~ x~ o ~ v ~ -

L E G E N D :

• 71.3°N; 156.8°W

• 35.1 ° N; 111.7°W

x 41.8 ° N ; 8 7 . 8 ° W

1 . 4 ° S: 48.6 ° W

v 6 5 . 3 ° S: 6 4 . 3 ° W

• 75.5 ° S; 26 .7°W

+ 76 .5°N; 68.8 ° W

> 14.8 ° N; 121 ° E

28.5 ° N; 77.25 ° E

< 81 .4 ° N; 17° W

• 36 ° N; 129 ° E

27.4 ° S; 153.1 ° E

78.1 ° N 13.6 ° E

35.7 ° N; 51.3 ° E

j 40 .0 ° N; 32.9 ° E c

o

F ® 0

- 3 0

R--0401 (U)

--20 10 0 10 20

M O N T H L Y A V E , TEMPI °C

3 0

Fig. 21. Oxygen isotope ratio in world-wide precipitation vs monthly average air temperature, showing that for every point on the line in

Fig. 20 there is a corresponding average air temperature.

Europe was very cold, at the bot tom of the First Little Ice Age according to our tree measurements (see Figs 1 and 2) and according to European thermometer records. The wood used as the standard was consequently greatly depleted in heavy isotopes, in deuterium by -180 ppt, because the rain ingested was correspondingly depleted.

When the measured, listed, and published isotope measurements, referred to wood of 1712-1714, are referred instead to ocean water (smow), then 6 Dsmow and 618(smow) are related to each other by the slope of 8 for the oak sequence just as they are in the cedar sequence, see Fig. 22.

The relation between the tree 6 D and smow 6 D is, in ppt,

6 D s m o w = 0 . 8 2 6 D t r e e -- 180

which follows from the relations, for deuterium.

8 D , m o . = 10 3 × ( R t r o e / R * m o * -- 1)

8Dtree = 10 3 X (Rtree ~ R ' t r e e - 1)

+ 1)R* (10 "a 6Dtree + 1)R"#tree ( l O ' S 6 D s m o w sraow =

-200

•o tSO

~ -I00;

50 0

× J4P4NESE CEDAR (t50-1970 A 0) • GERMAN OAK I (1712-t950 A D) o GERM4N OAK 11 (1550 t725 ~ D) o o~/ %

• " .Z ~-~'~ • i ; : o o

-5 10 J5 -20 25 DELL& (018/016) pot (SMOW)

Fig. 22. Deuterium isotope ratio vs oxygen isotope ratio measured in our sequence of German oaks and in our Japanese C. japonica. All ratios are related to each other by the slope of 8, namely according to D / t t = 8018/O t6 + constant. We have subtracted off the constant in order to make all points lie on the line for world-wide precipitation

(IAEA)


so that for 1712-1714, when /iOtree = 0 and ~5 Osmow = -180,

R ' t r e e / R * m o w = 1 - 0.180 = 0.82.

Finally, it is predicted that for all trees which subsist on rainwater, the deuterium and oxygen isotope ratios will be related linearly in their rings with a slope of 8 (Libby, 78), namely,

D / H = 8 01~/016 +constant.

The constant is different in the oaks and in the cedar f rom its nearly zero value in sea water; in plotting Fig. 22, the appropr ia te constant was substracted f rom the tree measurements in order to make all the measured points lie on the sea water curve; this subtraction has no effect on the slope of 8, o f course.

Recommendations

The history of climate should be studied in all radiocarbon dating laboratories, in the following way. For each sample on which a radiocarbon date has been made, the deuterium and oxygen isotope ratios should be determined as well. This should be done for wood samples and for organic material in samples of soil, as well as any other plant material. Then f rom the stable isotope ratios, climate can be deduced, and from radiocarbon measurement , the age of the sample is determined. These data should be computerized and used to deduce world-wide climate horizons versus past time. Finally, climate periodicities should be determined and future climate variations predicted from them.

Acknowledgement - - We thank Willard F. Libby for his advice in the sample preparation, in particular how to combust the wood samples quantitatively, for checking all our numbers, and for collaboration in proving the generality of the slope of 8.

References

B. Becket and V.G. Siebenlist. (1970) Tree ring sequence in Bavarian Fir. Flora 159, 310-346. R. Berger and W.F. Libby. (1970) Exchange of CO 2 between atmosphere and sea water. In: Global Effects of Environmental Pol- lution, S.F. Singer, Ed., D. Reidel Publishing Co. P. Bergthorson. (1962) Ice record in Rejkavik harbor. Proc. Conf. Climate 1 Ith to 16th Centuries, Aspen, Colorado, National Center for Atmospheric Research, Air Force Cambridge Research Laboratories. R.B. Blackman and S.W. Tukey. (1958) The Measurement o f Power Spectra, Dover, N.Y.; and see also Weast (1973) Handbook of Chemistry and Physics, 54th edition, Chemical Rubber Company, Cleveland, Ohio. T.J. Cohen and P.R. Lintz. (1974) Periodicities in sunspot numbers. Nature 250, 398-399. M. Cohn and H.C. Urey. (1938) Exchange of oxygen isotopes between CO~ and H20. J. Am. Chem. Soc. 60, 679-687. W. Dansgaard, S.J. Johnson, H.B. Clausen, and H.J. Langway. (1972) Ice record of oxygen isotopes at camp century. Nature 227, 482-483. G.V. Daugherty. (1976) Acting Superintendent, United States Depart-

ment of Interior, National Park Service, Three Rivers, California, collected the Sequoia gigantea and sent it to us. DARPA, Defense Advanced Research Projects Agency. (1971) Contract Numbers F44620-73-0025, F44620-72-0029, monitored by Air Force Office of Scientific Research. B.A. Franklin. (1977)New York Times, June 27, p. 14. Gerhard Herzberg. (1945 ) Infrared and Raman Spectra o f Polyatomic Molecules, Van Nostrand and Reinhold, New York. R.F. Gould, Ed. (1966) Lignin Structure and Reactions, Advances in Chemistry Series, No. 59, American Chem. Soc., Washington D.C. A.J. Grobecker. (1975) Ed. in chief, Monograph No. 1, Climatic Impact Assessment Program, Dept. of Transportation, Washington D.C. J.R. Hill. (1977) Periodic analysis of sunspot records. Nature 266, 151-153. B. Huber. (1935) Ring width sequence in German oaks. Ber. at. Bot. Gesell. 53, 711-719. K.M.B. von Huber and V.G. Siebenlist. (1969) Unsere tausendj~ihrige Eichen Jahrring-Chronologie durch schnittliche 57 (10-150)-fach belegt. Sitzberichte Abteiling I, Ostereichische Akademie der Wissenschaft 178, Band Ibis 4, Heft 37-42, Wien. P.H. Hurt. (1978a) Stable oxygen isotopes in tree rings as an indicator of solar variability. Masters Thesis, Chemistry Department, Universi- ty of California at Los Angeles. P.H. Hurt, L.J. Pandolfi and L.M. Libby. (1978b) Sunspot cycle in tree ring isotope record. Submitted for publication. P.H. Hurt, L.M. Libby, L.J. Pandolfi, L.H. Levine, and W.A. Van Engel. (1979) Periodicities in blue crab population of Chesapeake Bay. Climatic change 2, 75-78. I.A.E.A. (1969-1975) Environmental Isotope Data, No. 1-5. International Atomic Energy Agency, Vienna, Austria. E.K. Kalil and I.R. Kaplan. (1976) Uranium and organic carbon concentrations in a deep sea core. Ph.D. Dissertation, Univ. of Calif. at Los Angeles. E.K. Kalil and I.R. Kaplan. (1978) Uranium and organic carbon concentrations in a deep sea core. Marine Chemistry, in press. E.L. Ladurie. (1971) Times of Feast," Times o f Famine. Doubleday, Garden City, New York. H. Landsberg. (1975) Time Series o f Temperatures and Rainfall From Records in the Eastern U.S., Reduced to Philadelphia. University of Maryland, College Park, Maryland. W.H. Landsberg and R.E. Kaylor. (1976) Cyclic analysis of temperature and rainfall records for the eastern U.S.J. interdiscip. Cycle Res. 7, 237-243. L.M. Libby. (1972) Multiple thermometry in paleoclimate and historic climate. J. Geophys. Res. 77, 4310-4317. L.M. Libby and L.J. Pandolfi. (1973a) Calibration of two isotope thermometers in a European oak using official weather records. Pro* ceedings of International CLIMAP Conference, Norwich, England, pp. 21-39.

L.M. Libby and L.J. Pandolfi. (1973b) Calibration oflsotope Ther- mometers in an Oak Tree Using Official Weather Records. Colloques Internationaux du Centre National de la Recherche Scientifique, Gif- sur-Yvette, France. L.M. Libby. (1974a) Final technical report on historical climatology. DARPA order no. 1964-1, U.S. Air Force Office of Scientific Research, R & D Associates, Santa Monica, California. L.M. Libby and L.J. Pandolfi. (1974b) Temperature dependence of isotope ratios in tree rings. Proc. Natn. Acad. Sci. 71, 2482-2486. L.M. Libby and L.J. Pandolfi. (1976a) Isotopic tree thermometers; correlation with radiocarbon. J. Geophys. Res. 81, 6377-6381. L.M. Libby, L.J. Pandolfi, P.N. Payton, J. Marshall 3rd, B. Becker and V.G. Siebenlist. (1976b) Isotopic tree thermometers. Nature 261, 284-288. L.M. Libby and L.J. Pandolfi. (1977a) Climate periods in tree, ice, and tides. Nature 266,415-417. L.M. Libby. (1977b) Correlation of historic climate with historic prices and wages. Indian J. Met. 28. W.F. Libby, L.M. Libby, and L.J. Pandolfi. (1978) The Slope o f Eight. To be published. G. Manley. (1974) Central air temperatures: monthly means 1659- 1937. J. R. met. Soc. 100,389-405. G. Manley. (1953). The mean temperature of central England 1698- 1952. J. R. Soc. 79, 242-261. G. Manley. (1959) Temperature trends in England 1698-1957. Archs. Meteorol. Geophys. Bioklimatol. Set. B.A.K., Band 9, Heft 3/4, Vienna.


H. Michael. 0976) University Museum, University of Pennsylvania, counted the rings and dated them by comparison with the fiducial ring sequence pattern developed at the University of Arizona. H.O. Meredith. (1939) Outline o f the Economic History o f England. Sir Pitman, London, Chart B opposite page 353. L.J. Pandolfi, E.K. Kalil, P.R. Doose, L.H. Levine and L.M. Libby. (1978) Climate periods in trees and in a sea sediment core. Phys. Earth Planet. Interiors, in press. A. Rebello. (1976) Composition of wood. In: Environmental Biogeo- chemistry, J. Nriagu, Ed., 2 vol., Ann Arbor, MI. H.S. Rinehart. (1972) Tidal periods in yellowstone geysers. Science 177,346-347. P.A. Samuelson. (1970) Economics, 8th Edition, pp. 55-74, McGraw Hill, New York. O. Sepall and S.G. Mason. (1961) Deuterium exchange between

cellulose and water. Can. J. Chem. 39, 1934-1943. G.F. Steffen. (1901) Studien zur Geschichte der englischen Lohnarbei- ter mit besonderer Berticksichtigung der Veranderungen lhrer Lebens- haltung. Hobbing & Bychle, Stuttgart, 3 vols. H. Taube. (1956) Oxygen isotopes in chemical studies. Ann. Rev. nucl. Sci. 6, 277-302. H. Veenman and Zonen. (1938) De Landbouw in Brabants Westhoek in het Midden van de Achttiende Eeuw. Agronomisch Historische Bij- dragen, Wageningen, Netherlands. M. Waldemeier. (1961) The Sunspot Activity in the Years 1610-1960. Zurich Schulthen, Ziirich, Switzerland. R.C. Weast, Ed. 0962) Handbook o f Chem. and physics. Chem. Rubber Co., Cleveland, Ohio, 45th edition, pp. F90 and F91. World Weather Records. (1966) Vol. 2, Europe, U.S. Dept. of Commerce, Environmental Services Administration, Washington DC.

tree thermometers and commodities: historic climate indicators

Documents