Click for next page ( 104


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 103
CHAPTER 11 ISOTOPIC TRACER TECHNIQUES FOR MEASUREMENT OF PHYSICAL PROCESSES IN THE SEA AND THE ATMOSPHERE1 HARMON CRAIG, Scripps Institution of Oceanography, University of California, La Julia, California I. Introduction THROUGHOUT this report reference has been made to the need for a fundamental understand- ing, on a long-term basis, of mixing phenomena in the ocean and the atmosphere. In a general sense the ocean and the atmosphere may be re- garded as a two-phase system, in which the phases are separated by the fundamental dis- continuity of the ocean-atmosphere interface. Each phase is further divided into two parts by a second order discontinuity; the atmosphere, divided by the tropopause at about 12 km, into the troposphere and the stratosphere, etc., and the oceans, divided by the thermocline at some 100 meters, into an upper and lower layer. The basic problems in determining the effects of both radioactive waste disposal and the dis- persal of debris from nuclear explosions may be formulated in terms of a single objective. Given the ocean-atmosphere system under normal steady state conditions, and given some sub- stance introduced at any point in one of the four designated zones, we wish to be able to predict quantitatively the concentration of the substance as a function of latitude, longitude, altitude or depth, and time. These problems thus involve studies of (1) the intra-phase mix- ing, above, below, and across the second-order discontinuities within each phase, and (2) the inter-phase mixing across the ocean-atmosphere interface, with the aim of predicting the effects of perturbations on the system. The dominant mixing processes in the vari- ous spheres are processes of mass movement or turbulent mixing. In such processes, for ele- ments undergoing no change of phase, there is little or no separation of components, and thus, in general, isotopic tracer techniques may in- 1 Contribution from the Scripps Institution of Oceanography, New Series, No. 902. volve a wide range of materials of quite differ- ent chemistry. It is this phenomenon of mass movement dominance, and the relative unim- portance of diffusive transfer except in special cases, that makes the tracer technique so power- ful; a tagged isotope for each element is not required, and one can choose for each particular study the elements most useful for tracing the movement of a mass of heterogeneous material. As mentioned at several points in this report, both artificial and natural isotopic tracers may be used for the study of transfer phenomena in the sea and the atmosphere. Artificial tracers are of value in such studies because they allow the investigator to introduce perturbations in the system at convenient times and places; they are especially valuable for the study of short term fluctuations in local systems. However, for the general understanding of mass transfer phe- nomena, artificial tracers are of value mainly as experimental checks on deductions based on other data, with the exceptions of a few special cases to be described below. The reason for this is that mass transfer phenomena are by nature subject to long term periodic fluctuations such as convection, with periods often longer than the time range available for observation. A sec- ond reason for this is the high cost of radioiso- topes and the large amounts of activity required to tag adequately the large masses of water nec- essary for ocean studies. Revelle, Folsom, Goldberg and Isaacs (1955) have discussed in their Geneva report the prob- lems involved in adapting radioisotope tracer techniques to transfer studies in the ocean, and the requirements for usable isotopes. If one introduces some 10 curies of a gamma emitter in solution at some point below the thermocline, it is found that within a reasonable duration of observation time the activity will be concen- trated in a layer of the order of 1 meter thick 103

OCR for page 103
104 Atomic Radiation and Oceanography and Fisheries spread over a horizontal area of radius r. They find that with the best of present instruments, the horizontal spread in which the concentration of the introduced radioactivity can be deter- mined corresponds to r= 1 km. With the possi- bility of improved instrumentation, and the use of specially selected nuclides, it may be possible to raise the area of detection and determination of activity concentration to about 100 km2, an area which is still negligible with respect to oceanic expanses. The radioisotopes suitable for such measure- ments must of course have a half-life compati- ble with the mixing rates to be studied and yet short enough so as not to constitute a perma- nent hazard, namely of the order of a week to a month. Moreover, they must be available in multi-curie amounts at reasonable cost, should form soluble ionic species in sea-water, have a high specific activity, and, for instrumental rea- sons, should be gamma emitters with energy be- tween .2 and 1.3 Mev. Revelle, et al., were able to list three such isotopes, which, together with half-life, cost, and other data, are listed in the following table: Cost Half- per Isotope life curie Kb" ... 19.5 day $1000 Im ... 8.0 day 750 Ba1" ... 12.8 day 500 Gamma energy Mev Specific activity available 9 me/gram 1.1 Carrier-free 0.36,0.72 Carrier-free 0.16,0.54 Comparison of the cost of these isotopes with the maximum area of detection cited above shows that the study of large-scale transfer phe- nomena in the oceans, using deliberately intro- duced artificial radioactivity in the form of spe- cific isotopes, is so costly as to be infeasible with the estimated best instrumentation which will be available in the near future. It is evident that such isotopes are at best adapted only to short-term, small scale studies of local phe- nomena. The use of mixed fission products on a large scale, discussed elsewhere in this report, is somewhat more feasible but is beset with many difficult problems of transportation and handling. From these considerations it seems evident that the critical data in studies of atmospheric and oceanic mixing and interaction will come from the use of the naturally occurring isotopic tracers, which reflect in their material balance adjustments the differential rates of transfer from source, through reservoir, to sink, and loss by decay. It is from these transfer rates, adjusted to the steady state geochemical and geophysical cycles of the various elements, that we can hope to gain an understanding of the long period variations in natural transfer phe- nomena. The importance of gaining a clear un- derstanding of the long period transfer rates, when problems such as storage of potentially hazardous radioactive wastes and cumulative ef- fects of nuclear detonations are considered, can- not be overemphasized. In the following sections we discuss the pres- ent status of our knowledge of the distribution and properties of the various naturally occurring isotopes which are useful for studies of atmos- pheric and oceanic transfer phenomena. In ad- dition, mention is made of the nuclides pro- duced in nuclear detonations and supplied by reactors which have properties such that they are also useful in such studies and which have been studied to some extent. II. Distribution of naturally occurring isotopes of elements adapted for transfer studies In this section we discuss the production and occurrence of radioactive and stable isotopes showing measurable isotopic variations, and the distribution factors which determine their rela- tive concentrations in natural materials. Carbon 14 Carbon 14 is formed in the atmosphere by the reaction of neutrons with nitrogen, i. e. the neutrons being the result of the interaction of primary cosmic rays with the atmosphere (Libby, 1955). The carbon 14 is naturally ra- dioactive, decaying by ^-emission back to nitro- gen 14 with a half-life of 5570 years. Thus the half-life is so short that radiocarbon depends, for its existence, on the continual production in the stratosphere, with which it is presumably in steady state. The assumption of a steady state condition for at least the last 15,000 years is justified by the observation that radiocarbon dates on historic samples agree with the calen- dar dates. The steady state production rate, which is equal to the steady state disintegration rate, can be calculated from measurements on the neutron flux in the lower stratosphere and compared with the observed specific activity of

OCR for page 103
Chapter 11 105 Tracer Studies of the Sea and Atmosphere carbon. Anderson (1953) who has made the most recent and detailed considerations of the production rate, finds a rate of 2.6 carbon 14 atoms per cm2 and per sec. The carbon 14 atoms are oxidized to CO2 and thus enter the normal geochemical and biologi- cal cycles of carbon via the atmosphere. The distribution through the atmosphere and the terrestrial plants is rapid, and the steady state radiocarbon concentration in these reservoirs is taken as the basis for the so-called "modern" specific activity of carbon, namely about 15 dis- integrations per minute per gram of carbon. On the other hand, the transfer of carbon from the atmosphere to the sea is slow enough, compared to the half-life, to produce a signifi- cant difference between the predicted and ob- served activity of carbon in the surface layers of the oceans. Carbon, as one of the lighter elements, is subjected to natural fractionation of its isotopes in the various reactions it under- goes in its biogeochemical cycle (cf. section on stable isotope variations, below). The steady state isotopic separation of the stable isotopes C12 and C15 produces a C12 concentration in surface ocean water bicarbonate and shell car- bonate which is about 2.5 per cent higher than the C12 concentration in terrestrial plants. It is thus known that the C14 concentration in ocean bicarbonate and carbonate should be about 5 per cent higher than the concentration in land plants, namely about 15.75 disintegrations per minute per gram of carbon (Craig, 1954). In actual fact, however, measurements show that the specific activity of bicarbonate and carbon- ate from the ocean is about the same as the spe- cific activity of land plants. Thus the atmos- pheric C14 activity has been increased 5 per cent by slow exchange of CO2 between atmos- phere and sea, resulting in an "apparent age," relative to wood standards, of 400 years for the bicarbonate and carbonate shells in the surface layers of the ocean (Craig, 1954, 1957 (a) ). Some 10 measurements have now been made on marine plants, animals, and sea-water from the Atlantic (Suess, 1954) and from the New Zea- land area (Rafter, 1955) which indicate that the radiocarbon "age" of surface marine car- bonate is about 400 years; it is thus clear that radiocarbon age determinations made on deep ocean waters must all be referred to this base- line, rather than to the modern specific activity displayed by the terrestrial plants and the atmos- phere. The evaluation of the exchange time of CO2 between atmosphere and sea from data on the natural distribution of C14, is discussed in Sec- tion IV of this paper. The most recent and accurate measurement of the absolute radiocarbon concentration is that of Suess (1955), based on comparisons with an absolute standard obtained from the National Bureau of Standards. Suess finds a concentra- tion of 1.238 x 10~12 atoms of C14 per atom of carbon for average 19th century wood, corrected for decay to the present date and corrected for isotopic fractionation. Based on this value and a half-life of 5568 years, we give below the amounts of C14 in metric tons, and the activi- ties in megacuries present in the major reser- voirs on the earth (Craig, 1957 (a), calcu- lated from his Table 1). The figures for the atmosphere and terrestrial living matter are nor- malized for isotopic fractionation, while the or- ganic and inorganic carbon in the ocean was assumed to have an average age of 600 years relative to corrected 19th century wood, or 200 years relative to surface ocean bicarbonate (see Section IV, this paper). Total C" Total activity Reservoir metric tons megacuries Atmospheric CO, 0.96 4.4 Terrestrial living matter + humus 2.2 11.0 Ocean: Total organic matter. 3.8 17.6 Ocean: Total inorganic car- bon 49.8 228.6 Totals 56.8 261.6 The total activity of radiocarbon present on the earth thus corresponds to some 260 mega- curies, practically all of which is in the ocean. Using Anderson's figure for the production rate, cited above, and the decay constant of 3.945 x 10-" sec-1, the calculated total inventory of radiocarbon on the earth is 78.4 tons, which differs from the figure of 56.8 metric tons, ob- tained in Table II, by about 28 per cent. How- ever, the production rate, as estimated from cos- mic ray data and the counting of atmospheric neutrons, is uncertain to at least 20 per cent be- cause of the uncertainty in the reactor flux from which the neutron counters are calibrated. More recent estimates of the production rate are lower than the figure cited above and all that can be said about the agreement between the calculated

OCR for page 103
106 Atomic Radiation and Oceanography and Fisheries and predicted radiocarbon inventories is that they agree within present limits of error. Tritium Tritium (H5) is made in the upper atmos- phere, primarily in the "stars" or nuclear ex- plosions produced by the collisions of primary cosmic ray particles with the atmospheric molecules; it is naturally radioactive, decaying by /T emission to helium 3 with a half-life of about 12.5 years (Kaufman and Libby, 1954). The T atoms "burn" very quickly to HTO and enter the precipitation — evaporation cycle of water. A very small amount of tritium is pro- duced in rocks by the nuclear reaction of lithium with neutrons produced by spontaneous fission of uranium and from (o,n) reactions (Morrison and Pine, 1955) ; the production of tritium by this process is insignificant relative to the atmospheric production. Detailed studies of the distribution of tritium in natural waters have been made by Libby and his co-workers at Chicago. The natural concen- tration of tritium (before thermonuclear tests) in continental waters averages about 5x10-15 atoms of tritium per atom of hydrogen. (Fol- lowing Libby's usage, such a concentration will hereafter be referred to as 5 tritium units, ab- breviated as T.U.) The concentration in oceanic rains is about 1 T.U., while in the surface waters of the ocean itself the concentration appears to be as low as 0.2 T.U. The sea is, of course, the ultimate resting place of the tritium formed in the atmosphere, and the low concentration in the oceanic rains relative to continental rains is principally due to tritium removal by direct molecular exchange with the sea surface (see below). Kaufman and Libby (1954) calculated the tritium production rate in the atmosphere by equating it with the rate at which tritium dis- appears from the atmosphere into the ocean, taken as the sum of the tritium entering the ocean by run-off from continental rains and the tritium entering directly via oceanic rains. For this calculation only the average run-off and ocean precipitation figures, and measured av- erage tritium content of such waters, are needed. They obtained a net production rate, averaged over the earth's surface, of .12 T atoms per cm2 per second. Von Buttlar and Libby (1955) measured many more rain samples, and also analyzed 5 samples of ocean water, from which they could estimate the tritium content of the water vapor which evaporates from the sea surface. Using this latter figure they calculated the production rate over the oceans, assuming that tritium is lost from the atmosphere only by oceanic rain, and gained by production and oceanic evaporation, and obtained a figure of 0.11 to 0.12 T atoms per cm2 per sec. A similar calculation was made for the production rate over land, assuming tritium is lost from the continental atmosphere only by continental rains running off into the ocean, and gained by pro- duction, and by transport of ocean vapor onto the continents. Using the tritium data for av- erage Mississippi Valley rains, they obtained a figure of 0.16. Their estimated world average production rate is 0.14 with a probable un- certainty of less than 20 per cent. This value agrees precisely with the expected world pro- duction rate calculated by Currie, Libby, and Wolfgang (1956) from their experimental measurements on tritium production in nitrogen and oxygen by bombarding protons of 450-Mev and 2-Bev energies. Previous experiments and calculations by Fireman and Rowland (1955) gave an expected production rate of 0.2 T atoms/cm2 sec, also in good agreement with the rate apparently observed. However, the tritium production rate must be a good deal higher than the figures given above. Von Buttlar and Libby calculated that, with such a production rate, and with the ob- served surface sea concentration of about 0.24 T.U., then the mixed layer of the sea is about 100 meters deep if one assumes that all the tritium of the sea is in the mixed layer. Though this depth is consistent with observational data on the sea, such a calculation assumes that the mixed layer is sealed off from the deep sea so that no tritium mixes below the thermocline, and the question then arises as to just how much mixing across the thermocline does, in fact, occur. As discussed by Wooster and Ketchum in a separate paper in this report, various observa- tions on ocean currents and on the heat flux through the ocean floor, indicate that the deep ocean water turns over, or mixes with surface water, in times of the order of a few hundred years. Assuming a generalized two-layer model of the sea, consisting of a shallow mixed layer about 75 meters deep on the average, and a

OCR for page 103
Chapter 11 107 Tracer Studies of the Sea and Atmosphere homogeneous deep sea below the thermocline marking the interface between the layers, Craig (In press (a) ) derived equations relating the production rate of a radioactive isotope to the concentrations of the isotope in the two layers of the sea and the mixing time through the thermocline. (These functions are discussed briefly in a separate paper by the writer in this report, in which calculations on the disposal of fission products in the sea and their ultimate steady state concentrations are discussed.) The applications of such calculations to the distribu- tion of radiocarbon in the atmosphere and sea were demonstrated; these results are discussed in Section IV of this paper. Application of such calculations to the dis- tribution of natural tritium (Craig, 1957 (b) and manuscript in preparation) shows that for reasonable internal mixing rates of the sea, most of the world inventory of tritium must actually be in the deep sea below the thermocline. Thus for a deep water replacement time, or residence time of a water molecule in the deep sea before mixing into the surface layer, of 0 to 1000 years, and with a surface concentration of 0.24 T.U., the tritium flux into the sea must be between 7.6 and 0.3 atoms cm2/sec. For the most reasonable deep sea residence time of the order of a few hundred years, the flux must be somewhere between 0.4 and 0.8. It is found that about £ of the total tritium in the sea is below the thermocline, with a deep-sea tritium concentration of about 0.014 tritium units. The tritium production rate over the North American continent was recalculated (Craig, op. cit.) by taking into account the removal of tritium from the continent by the outgoing water vapor which does not condense over the land. This calculation gives a world average production rate of from 0.6-0.8 after correction for the latitudinal geomagnetic effect on the incoming cosmic rays. A tritium production rate of this order of magnitude indicates an average deep-sea residence time of water of about 250 years, for a simple two-layer ocean. Calculations based on a second-order ocean model in which the deep sea reservoir is exposed to the atmosphere at high latitudes would give a longer residence time relative to the mixed layer of the sea because of direct entry of tritium from the atmosphere to the deep sea. (See the discussion of radiocarbon residence times in Section IV of this paper.) However, if the bulk of the tritium is not produced by cosmic radiation, but by solar accretion (see below), the world average pro- duction may be as high as 1.7 atoms cm2/sec because the geomagnetic correction applies only to tritium produced by cosmic rays in the trop- osphere. The calculated production rate over the oceans of about 0.14 is obtained by considering only the transfer of tritium into the sea by rainfall. Since rainfall appears to account for only about one-tenth of the tritium which ac- tually enters the sea, it appears that the trans- fer of tritium from atmosphere to sea by direct molecular exchange across the sea surface is about 9 times as effective as the scrubbing action of precipitation. A production rate of 1.4 atoms of tritium/ cm2sec means that the world inventory of trit- ium, before thermonuclear tests, was about 20 kg of tritium, or 200 megacuries, essentially all of which is in the ocean. However, from the experimental data obtained by the workers cited above on the production of tritium by the action of protons on nitrogen and oxygen, it appears very doubtful that the cosmic ray production rate can be much higher than about 0.2. In fact it is probably necessary to assume that tritium is produced on the surface of the sun and is directly accreted into the earth's atmosphere, rather than being a secondary re- sult of the action of the cosmic ray protons on the atmosphere, as postulated by Feld and Craig (Craig, 1957 (b)). From a study of the fall-out rate of strontium 90 pushed into the stratosphere by large atomic detonations, Libby (1956a, b) calculates the stratospheric residence time of strontium to be about 10 years (cf. Section IV of this paper). Since at least half of the tritium production should take place in the stratosphere even if all the production is due to the action of pro- tons on the earth's atmosphere, slow mixing through the tropopause will pile up tritium in the stratosphere in the same way that slow exchange across the sea surface builds up the radiocarbon concentration in the atmosphere. One-sixth of the atmosphere is above the tropo- pause on the average, but the water vapor con- centration is so low that only about 0.3 per cent of the total water vapor in the atmosphere is in the stratosphere; thus the tritium concentra- tion of the stratospheric water vapor will be

OCR for page 103
108 Atomic Radiation and Oceanography and Fisheries much higher than that of the tropospheric va- por, which averages about 1 T.U. From the strontium data we assume that the mixing time of water vapor through the tropopause is at least 10 years. Assuming a tritium production rate of 1.4, half of which is in the stratosphere, the trit- ium concentration of stratospheric water vapor is then calculated to be at least 300,000 tritium units. This is an astounding concentration fac- tor relative to tropospheric water vapor. Re- cently the present writer and F. Begemann analyzed a series of samples of atmospheric molecular hydrogen for deuterium and tritium content respectively. Mass spectrometric meas- urements showed that all samples contained about 2-10 per cent less D than ocean water, falling just in the range of meteoric waters, and containing far too much deuterium to represent thermodynamic equilibrium with water vapor. These data confirmed a few previous measure- ments (cf. Harteck, 1954) which showed that the molecular hydrogen in the atmosphere must form by direct photodissociation of water vapor in the region around 70 km altitude, rather than by bacterial decomposition of organic matter which has been shown to produce hydrogen in isotopic equilibrium with water. We may thus assume that the tritium content of stratospheric molecular hydrogen is about the same as that of the stratospheric water vapor. Assuming that the hydrogen is statistically distributed in the atmosphere, so that £ is above, and % below, the tropopause, and taking again the mixing time through the tropopause as 10 years, we then calculate the tritium con- tent of the molecular hydrogen in the trop- osphere. This figure is found to be 100,000 tritium units, probably as a minimum figure because of slow vertical mixing from the base of the stratosphere to the 70 km level where the hydrogen is made, and because of the indi- cation that more than half the tritium is found initially in the stratosphere. The tritium con- tents measured by Begemann on a dozen sam- ples of tropospheric hydrogen range from 50,- 000 to 100,000 tritium units, averaging about 80,000 T.U., in excellent agreement with the calculated value when the various uncertainties are considered. It thus appears that the high tritium content of tropospheric hydrogen can be satisfactorily explained by purely geophysical reasoning based on the stratosphere-troposphere exchange time as estimated from Libby's Sr*° data, and the known concentration of water vapor in the stratosphere. This explanation seems more likely than the intricate series of photochemical mech- anisms proposed by Harteck (1954) which at best may account for a tritium concentration of about 1000 T.U. in the molecular hydrogen. Beryllium 7 Beryllium 7 is formed in cosmic ray stars, the peak production occurring at about 15 km. It decays by electron capture to lithium 7 with a half-life of about 53 days. The discovery, and the elucidation of the geochemical history, of this cosmic ray produced nuclide is due to Arnold and Al-Salih (1955). Once formed in the atmosphere, the beryl- lium burns to the nonvolatile BeO or possibly Be (OH)2, either of which diffuses until en- countering a dust particle and adhering thereon. It is thus a tracer for the atmospheric dust, on which it is washed out of the atmosphere by rain, ultimately going into the ocean. Arnold and Al-Salih detected radioberyllium in 22 rain and snow samples from Chicago and Indiana, the average absolute assay being 6x 10" atoms/ liter. The estimated world-wide average pro- duction rate is 0.04 atoms per cm2 per second, based on estimated rates of transfer and mix- ing in the stratosphere and troposphere. Most of the mixing rates involved are of the order of magnitude of the half-life, which makes calculation of the production rate difficult but greatly enhances the utility of the isotope for studying atmospheric processes, especially when used in conjunction with tritium. A detailed discussion of the beryllium 7 pro- duction rate and atmospheric residence time has recently been given by Benioff (1956). He calculates the production rate to be 5.0 atoms/ cm2min in the stratosphere and 1.3 atoms/cm2- min in the troposphere, and he finds that a stratospheric residence time of the order of years is required to match these production rates. Thus his stratospheric residence time agrees with that found by Libby for fission products. Beryllium 10, a /T emitter with a half-life of 2.5 x 100 years, is also formed in the cosmic ray stars. J. R. Arnold has recently identified this isotope in deep sea sediment samples (manuscript in press); it should be of great

OCR for page 103
Chapter 11 109 Tracer Studies of the Sea and Atmosphere importance, because of the long half-life, for the dating of such sediments. Deuterium and Oxygen 18 Deuterium and oxygen 18 are stable isotopes of hydrogen and oxygen respectively, and it is now well known that the isotopes of these elements, as well as of other light elements such as carbon, nitrogen, and sulphur, are fractionated, or separated, by chemical and physical processes in natural systems. Since the fractionation factors for stable isotopes are measurable and/or calculable for many separa- tion processes, and since the magnitude of these factors is mainly a function of temperature and process, the stable isotopes are extremely well adapted for the study of natural transfer rates in the geochemical cycles of their elements. The concentrations of these isotopes show rather wide variations in different natural ma- terials, these variations generally ranging from a few tenths of a per cent to a few per cent. In this report we shall mainly be concerned with the distribution of these isotopes in marine and fresh waters and in the atmosphere. Craig and Boato (1955) have recently reviewed the present status of natural isotopic studies, and reference is made to that paper for a more extended discussion. VAPOR PRESSURES AND RELATIVE ABUNDANCES OF THE ISOTOPIC WATER MOLECULES Relative p (mm Hg) abundance , ^ , Species (ocean water) Mass 30° C 100° C H.O 1 18 31.5 760 HDO 1/3230 19 29.4 741 H,OU 1/500 20 31.3 756 The above table shows the three most prom- inent members of the family of isotopic water molecules, their masses, relative abundances in average ocean water, and their vapor pressures at two temperatures. Other members of the family are much less abundant and can be neglected. One sees from the table that the vapor pressures are not a direct function of the molecular weight; the vapor pressure differ- ence between HDO and H2O is 10 times larger than the vapor pressure difference between H.,019 and H2O, at 30°C. The isotopic separa- tion in an evaporation or condensation process is directly proportional to these vapor pressure differences, so that in water vapor in equi- librium with water at 30°, the percentage de- pletion in deuterium, relative to the water, is ten times larger than the percentage depletion in oxygen 18. The natural isotopic variations are customarily given in terms of per mil enrichment or deple- tion relative to a standard, similar to the way the density parameter is given in an oceano- graphic temperature-salinity diagram. The data are presented in terms of a function 8, defined as follows: «(%)=[ (Rsample/Rstd) -1] x 1000 where R is the isotopic ratio 012/010 or D/H. In the case of deuterium, however, the quantity in the brackets is multiplied by 100 and the 8 values are given in per cent, because of the ten times higher isotopic separations encountered. Rstd here refers to the isotopic ratio in average ocean water. Since H2O1S is the most volatile isotopic species, the water vapor over the oceans is depleted in the heavy isotopes relative to the surface ocean water. As this vapor moves over the continents, the first rain to fall out is en- riched in the heavy isotopes relative to the vapor, again because of the higher volatility of the lightest species. Removal of the heavy iso- topes, in the form of rain, then causes the vapor to become continually depleted in deuterium and oxygen 18. In general enough rain falls out of an air mass over the oceans so that by the time the mass reaches the continents the rain is already "lighter" in isotopic content than ocean water, and as the air mass moves inland and poleward the rain which falls out becomes more and more depleted in deuterium and oxygen 18. In a recent study by Craig (ms. in prepara- tion) several hundred fresh water samples from all over the world were analyzed for deuterium and oxygen 18 concentration. The deuterium concentration varies by about 30% relative to mean ocean water, 8D ranging from +3 to — 27%, while the oxygen 18 concentration varies by only 4%, SO15 ranging from +6%0 to — 34 %c. The delta values for the two iso- topes show a linear correlation such that 8D = 9SO12, corresponding to the vapor pressure difference ratio at about 25 °C. The reason for the high value of the average temperature at which liquid and vapor equilibrate in the at- mosphere is as yet unknown; the uncertainty

OCR for page 103
110 Atomic Radiation and Oceanography and Fisheries in the vapor pressure data is such that the value could hardly be less than about 20°. The delta values for fresh waters show a general correlation with latitude or distance from the ocean; there is a general decrease in the heavy isotope concentration as the latitude varies from equatorial to polar, reflecting the continuous loss of vapor from the poleward moving air masses. Isotopic variations such as mentioned above can be measured quite simply and precisely with the mass spectrometer, and it is evident, from the ranges of variation cited, that such studies on meteoric waters can provide a wealth of information concerning meteorological trans- fer and mixing phenomena in the atmosphere. The average water vapor of the earth has roughly the composition 8D=—10%, 8O19= —11 %0, but large variations, related to the amount of liquid water which has condensed out of the vapor, occur, and thus such studies are directly adapted to problems of water vapor transport over both the oceans and continents. The situation in the oceans themselves is somewhat more complicated. The oxygen iso- topic composition of ocean waters has been studied by Epstein and Mayeda (1953), and the deuterium variations in the same samples by Friedman (1953) ; these writers also an- alyzed nine fresh water samples and first eluci- dated the D-O12 relationship in natural waters. The surface layers of the oceans are in general enriched in the heavy isotopes relative to mean ocean water because of the net storage of H2O10 in the stagnant and circulating fresh water and vapor; the extent of this enrichment reflects the hold up at the boundary of the mixed surface layer, namely the thermocline. On the other hand, the deeper layers of the ocean are depleted in deuterium and oxygen 18, relative to mean ocean water, because of the influx of glacial melt water in polar latitudes, the glacial waters having 8 values at the lightest ends of the ranges cited in the preceding paragraphs. Thus the oceans are isotopically upside down with the heavy isotopes concentrated at the surface, and the isotopic composition parameters in general correlate with salinity. Epstein and Mayeda (op. cit.) showed that the salinity-oxygen 18 variations in marine waters were consistent with a model in which the oceanic precipitation is progressively de- pleted in the heavy isotopes as a function of the extent of precipitation from the local atmos- pheric reservoir. Salinity, of course, is uniquely related to the direct amount of fresh water removed by evaporation or added by meltwater dilution, but the relationship in the case of isotopic composition is more complex. This is because the isotopic composition of fresh water precipitating over the oceans, or added by run- off or melting of ice, is variable, depending on the history of the air mass from which it was precipitated. The correlation between isotopic composition and salinity is therefore more or less local, reflecting the particular relations ob- taining on the average in the area. As a result, the isotopic composition parameters, rather than being simply transforms of salinity, and thus not inherently very useful for the study of transfer problems, become important parame- ters for such studies because of the reflection of areal conditions in a manner different from, but related to, the salinity parameter. Examples of this effect are discussed in Part IV, where applications to transfer studies are treated. The isotopic composition of atmospheric oxygen is an interesting case of adjustment of a reservoir composition to steady state non- equilibrium biogeochemical transfer processes. Oxygen would exist in the atmosphere in the absence of living plants because of photodisso- ciation of water vapor in the atmosphere, with subsequent escape of hydrogen from the earth. However, oxygen is cycled through the bio- sphere so rapidly that its isotopic composition, rather than reflecting its mode of formation, may be adjusted to a steady state balance be- tween photosynthetic formation and respiratory uptake. The oxygen produced in photosyn- thesis is in isotopic equilibrium with the water taken up by the plants and is very close in isotopic composition to this water; however the atmospheric oxygen is some 23%e enriched in oxygen 18 relative to average ocean water. Lane and Dole (1956) have measured the preferential uptake of oxygen 16 by various animals and land plants and concluded that the net fractionation is such as to account quantita- tively for the atmospheric oxygen composition. Respiration in the oceans shows a much smaller selective oxygen 16 uptake (Rakestraw et al., 1951; Dole et al., 1954) and the isotopic com- position of oxygen dissolved in ocean water is variable and dependent on the amount of oxy- gen which has been taken up from the local

OCR for page 103
Chapter 11 111 Tracer Studies of the Sea and Atmosphere reservoir. There is some doubt as to whether the data of Lane and Dole can actually yield a material balance without invoking some spe- cial mechanisms relating the productivities of the oceans and the land, and a good deal of further study on this question is needed. The intent here is to point out that the isotopic transfer rates involved in this problem of the isotopic composition of atmospheric oxygen, and the variations in the isotopic composition and amounts of oxygen dissolved in ocean waters, may well be important parameters for the study of transfer phenomena in the oceans and the atmosphere and the interaction between them. Carbon 13 About one per cent of natural carbon con- sists of the stable isotope C12; the ratio C1S/C12, and thus effectively the Cls concentration, in natural material shows a range of variation of about 6 per cent. The details of the natural variation have been described (Craig, 1953, 1954), and reference is made to these papers for extended discussion. The delta values for carbon are referred to a standard which has the composition of average limestone; on this scale the characteristic compositions of natural materials are shown below: Material Limestones and shell 0 Ocean bicarbonate — 1.5 Atmospheric COi — 7 Marine biosphere —13 Terrestrial biosphere — 25 Coal _25 Petroleum _ 28 Shales —28 The difference between the compositions of atmospheric carbon dioxide and ocean bicar- bonate probably reflects the isotopic equilibrium constant for the exchange of carbon isotopes between these compounds; the other variations shown in the table are due to kinetic factors which cause a selection of the isotopes in the various processes involved in the biogeochemi- cal cycle of carbon. The carbon 14 variations caused by such processes should be almost exactly twice the C15 values shown above, and, as noted previously, the knowledge of the C12 variations has been of great value in under- standing the transfer rates and mixing phenom- ena involved in the distribution of radiocarbon. A particularly fertile field for study is the marine biosphere and the phenomena involved in the isotopic partition of carbon between carbonate and organic matter. One critical parameter in the kinetic processes involved is the rate of uptake of CO2 by photosynthesis versus the relative rates of CO2 replenishment by mixing and by reassociation of bicarbonate ions, and such studies may well lead to an improved knowledge of the carbon flux through local ecological systems and the interaction of the local system with the general marine reser- voir. Keeling (manuscript in preparation) has studied the isotopic variations in carbon dioxide over the land, and has found that the isotopic parameters are critical indicators of the atmos- pheric transfer phenomena through local bio- topes, as a result of the large difference in isotopic composition between normal atmos- pheric carbon dioxide and carbon dioxide pro- duced in respiration during the night. III. Contribution of radioisotopes to the geo- sphere by nuclear fission and detonations The steady state isotopic distributions dis- cussed in the preceding section have, in the case of radioactive elements, been altered to some extent by contribution to the geosphere of radioisotopes produced in nuclear fission in both reactors and nuclear detonations. Such contributions, rather than being detrimental to the study of natural transfer phenomena, have, on the whole, provided extra parameters of great value for such studies. It is of course obvious that addition of such elements under carefully controlled conditions in selected loca- tions and at planned times would have con- tributed a great deal more to our knowledge of geophysical phenomena than the actual dis- persal of the material has resulted in; never- theless it is possible, even though working in almost total ignorance of the amounts of ma- terial added, to deduce a great deal of valuable information about mixing rates and even to make detailed studies of certain specific prob- lems. The fission of uranium in reactors and nu- clear weapons results in a great variety of elements distributed mass-wise into a spectrum known as the fission yield curve; the propor- tions of the various masses produced are a unique function of the atomic mass and vary

OCR for page 103
112 Atomic Radiation and Oceanography and Fisheries little with neutron energy or substitution of plutonium for uranium 235. For our purposes, the elements of most interest produced by fis- sion are krypton 85, strontium 90, and cesium 137, ranging in half-life from 10 to 33 years; tritium, which is not a fission product, is also of great importance. Measurable additions to the geosphere of tritium, strontium 90, and krypton 85 have been noted and are discussed in Part IV in connection with general applica- tions to transfer study phenomena. In this sec- tion we estimate the total amounts of these nuclides which have been produced; these esti- mates are also of some interest in connection with the magnitude of the disposal problem, both present and future. One result of the advent of nuclear fission is that all the krypton in the atmosphere has become contaminated with radiokrypton. De Vries (1956) has measured the specific activity of atmospheric krypton, taken in March of 1955, as 25,000 counts per minute per mole. The activity is due to contamination with kryp- ton 85, which decays by y9~ emission with a half-life of ten years. From DeVries' measure- ment, we readily calculate that 56.4 moles, or 4700 grams, of Kr25 are now present in the atmosphere, and in ignorance of the rate of production, we make only a small error by assuming this figure as the total amount of radiokrypton produced and not correcting for decay. From the fission yield of 0.24 per cent for this isotope, it appears that some 23,500 moles, or 5500 kg, of U225 and plutonium have undergone fission since the advent of anthropo- genetic fission, resulting in an atmospheric krypton activity of 2 megacuries. It is assumed, as seems reasonable, that all fission produced krypton finds its way into the atmosphere. From the fission yield data, we calculate the total amounts of radiostrontium and cesium which have been produced; the data for the three elements are shown below. Only the strontium and cesium produced by detonation of nuclear weapons will escape into the atmos- phere and be deposited over the surface of the earth and sea. Because the krypton figure is uncorrected for decay, and because some kryp- ton must have gone directly into the strato- sphere and is not included in the measured activity, the values all represent lower limits and should be slightly larger. Total activity produced by all Fission fission yield (mega- Radioisotope Half-life "(',.'.) curies) Krypton 85 10 years 0.24 2 Strontium 90 28 years 5.0 15 Cesium 137 33 years 6.3 16 Libby (1956a, b) has given detailed discus- sions of the fall-out patterns of strontium 90 and cesium 137, based on the Project Sunshine measurements on world-wide samples. Geo- physically, the most significant finding is that, as mentioned previously, the residence time in the stratosphere of material pushed through the tropopause is about 10 years. The most recent measurements on the distribution of fission products from nuclear explosions (Libby: Address before American Association for the Advancement of Science, Washington, D.C., October 12, 1956) indicate that the amount of strontium 90 scattered over the surface of the earth is now equivalent to an average activity concentration of about 16 millicuries per square mile. In addition, the amount now held in the stratosphere is equivalent to another 12 milli- curies per square mile. The total amount so far distributed is thus about 5.6 megacuries of Sr'°, of which about 2.4 are still in the stratosphere, and, assuming purely statistical distribution, some 2.3 megacuries have fallen directly into the sea, while about 0.9 megacuries have fallen on the land surface. Because of the similar half-life and fission yield, the figures for cesium 137 will be almost identical to those for stron- tium 90. Comparing these figures with the ones given in the above table, we see that roughly 5.6/17.4 or 'J' of all the solid fission products so far produced, by all fission, have been dis- tributed over the atmosphere, the land, and the sea, by atomic weapons testing. The most important of these elements for studying mixing rates in the sea should be cesium 137, which being soluble, should be an excellent tracer for the mixing rate of surface ocean water down through the thermocline. Krypton 85 should ultimately prove important for atmospheric mixing studies, especially for comparison of mixing rates of gaseous and solid elements across the tropopause.

OCR for page 103
Chapter 11 113 Tracer Studies of the Sea and Atmosphere Thermonuclear weapons may also be expected to produce some carbon 14 because of the neutrons released into the atmosphere in the explosion. A contemporary sample of grass, collected in the summer of 1955 in S.W. Kansas by the writer, was analyzed for C14 content by M. Rubin at the U.S. Geological Survey labora- tory. This grass was found to be about 2.5 per cent higher in C14 content than the 19th century wood, corrected for age, used as the U.S.G.S. radiocarbon standard (Suess, 1955). The samples and standard were analyzed for C12 content by the writer and the results cor- rected for isotopic fractionation, and the sample was counted twice. Thus the measurement is quite precise, and probably indicates that the atmospheric radiocarbon content has risen about 2 per cent above normal at the present time, due to thermonuclear neutron production. For future geochemical studies with natural radio- carbon it will be important to monitor continu- ously the activity of contemporaneous plants and atmospheric carbon dioxide, though the effect will be insignificant for radiocarbon dat- ing studies for some time yet. The situation with tritium is different. The earliest rain ever analyzed for tritium content fell in Chicago in May of 1951; since October 1952 Libby and his co-workers at Chicago have produced an essentially continuous record of the tritium content of Chicago rain, and have analyzed a great many other samples from many parts of the world. Their data show that there was no significant production of tritium in the November 1952 Ivy test (Kaufman and Libby, 1954). However, the March 1954 Castle ther- monuclear tests produced an increase in tritium concentration of Chicago rain from an average value of 9 to a maximum value of 450 atoms T/1019 atoms H; i.e., a factor of 50 (von Buttlar and Libby, 1955). Even more striking was the discovery that the tritium content of southern hemisphere waters showed no signifi- cant increase in tritium concentration, and snow samples collected from the Antarctic as late as February of 1955 showed that during this in- terval no significant amounts of artifically pro- duced tritium had crossed the equator (Bege- mann, 1956). Begemann's recent data show that the tritium rained out of the northern hemispheric atmos- phere with a mean-life of about 40 days for the decrease in tritium concentration; as late as the end of 1955 the tritium concentration of Chicago rain was still about three times normal. In Section II above it was concluded that the world inventory of tritium was about 20 kg, with about 5 kg in the mixed layer of the sea, and about 15 kg in the deep sea. The Chicago data show that the tritium content of the sur- face ocean waters has increased by at least a factor of four, indicating that the order of magnitude of 20 kilograms of man-made trit- ium has so far rained out into the ocean. Thus the amount of tritium produced by man is now about equal to the natural steady-state inventory. IV. Applications of tracer techniques to the study of physical processes in the sea and atmosphere In this section we describe a few of the more obvious applications of the tracer techniques and isotopes described in the previous sections to specific problems of transfer phenomena in the oceans and atmosphere. The topics are sub- divided in terms of the isotopes discussed, in order to facilitate reference to points in pre- ceding sections and parts of this section. Carbon 14 Carbon 14 is perhaps the most useful of the isotopic tools available for geophysical and geochemical studies, especially when used in conjunction with oxygen 18 data; the 5700 year half-life and the universal distribution of carbon in organic and inorganic reservoirs make it ideal for such purposes. The most obvious application of immediate interest is the dating of the bicarbonate of the deep-sea waters, in order to determine the mixing rate of the oceans. Unfortunately, only one definitive set of measurements of this type has been made, namely the U. S. Geological Survey laboratory measurements of waters east of the Lesser An- tilles at approximately 57° W. and 16° N., made by M. Rubin (personal communication). These data are shown below: Depth (meters) Surface 640 1640 1750 Carbon 14 age (years) .. 652 .. 634 .. 628 841

OCR for page 103
114 Atomic Radiation and Oceanography and Fisheries The absolute values are probably not better than ±150 years, but the relative values are more precise. The age of the surface bicar- bonate is somewhat older than the 400-year average age mentioned in Part II as the result of slow transfer of atmospheric carbon into the sea, perhaps as a result of local conditions; however, the important figure is the age dif- ference between surface and deeper waters and it is unfortunate that still deeper waters were not sampled. The importance of a great many vertical profiles of this sort from both oceans, and their fundamental import for knowledge of the mixing rates in the ocean, is obvious. Because of the requirements of steady state balancing, the amounts of water transferred, per unit time, downward and upward through the thermocline in the sea must be equal, but because the mixed layer contains only about 2 per cent of the sea, this balance requires that a water molecule remain, on the average, some 50 times longer below the thermocline than above. As a consequence of this relationship, an uncertainty of 10 years in the residence time of material in the mixed layer results in an uncertainty of 500 years in the residence time of the material below the thermocline, consid- ering the world average rate of general cross- thermocline mixing of the substance. As we shall see below, the half-life of radiocarbon happens to be so long, that considerations of the extensive data on C" distribution in the atmosphere, biosphere, and mixed layer of the sea, do not yield important information on the internal mixing rate of the ocean itself. In fact, the distribution of tritium above the thermocline of the sea furnishes a much more precise esti- mate of the general turnover time of water in the deep sea. Thus the application of radiocarbon analysis to mixing problems within the sea itself can be made only by actually getting below the mixed layer and studying the deep-sea distribu- tion of C14 directly; such studies, coupled with chemical analyses and physical data serving as parameters for the identification of continuous water masses, will probably prove to be the most fruitful method for the delineation of large scale mixing phenomena in the sea. On the other hand, the distribution of radio- carbon in the atmosphere and mixed layer of the sea is strongly dependent on the rate of exchange of carbon dioxide between the atmos- phere and sea, and from a study of the relation- ship between the exchange rates and the radio- active decay rate, it is possible to derive rather precise values for the flux of carbon into the sea and downward through the thermocline. For such calculations it is necessary to assume a model of the atmosphere-sea system based on simplifying assumptions as to the nature of the sea below the thermocline. Calculations of this type, outlining the factors affecting the natural distribution of radiocarbon, have recently been made by Suess (1953), Arnold and Anderson (1957), Craig (1957 (a)), and Revelle and Suess (1957). The conclusions of these papers, though reached by various means, are quite similar, and we shall briefly summarize the gen- eral results. There are two empirically observed effects, of different origin, by which factors affecting the natural distribution of radiocarbon may be evaluated. The first of these is the observation that the carbon in the surface layers of the sea (bicarbonate, shell, and organic matter) has an apparent age of about 400 years relative to the terrestrial wood used as standards for radio- carbon dating. The second is the observation that contemporaneous wood has a radiocarbon activity some 2 per cent lower than the activity of 19th century wood, corrected for age to the present date. This decrease in activity, reflect- ing the contribution of C" free CO2 to the atmosphere by the combustion of fossil fuel, was first found by Suess (1953) and we shall refer to it as the Suess effect. The "apparent age" of carbon in the mixed layer of the sea has been measured on Atlantic ocean samples (and one Pacific sample) by Suess (1954), and on Pacific ocean samples around New Zealand by Rafter (1955). The average age determined by Suess is 430 years, while that of the Pacific samples was reported by Rafter as only 290 years. However, the Suess measurements are relative to the 19th century wood standard, corrected for decay to the present, while the Rafter measurements were made relative to a contemporaneous stand- ard which has suffered a decrease in activity due to the Suess effect. Measurement of the effect in the New Zealand standard shows that 110 years must be added to the ages reported by Rafter (1955) in order to correct for this effect and make the ages comparable to those reported by Suess (Rafter, manuscript in press).

OCR for page 103
Chapter 11 115 Tracer Studies of the Sea and Atmosphere Thus the average ages reported for the two oceans are in almost exact agreement, and we may consider the 400 year apparent age well established as a world-wide phenomenon. The 400 year apparent age of mixed-layer carbon is simply a less meaningful way of stating that the radiocarbon activity of mixed- layer carbon is 5 per cent lower than the ac- tivity in modern wood, uncontaminated by the Suess effect. Actually it is observed that the activities in wood and in surface ocean carbon are measured to be the same, but the measure- ments must be corrected for natural isotopic fractionation in the physical and chemical proc- esses involved in the carbon cycle (see section on carbon 13 variations). Marine shells con- centrate carbon 13 by 2.5 per cent relative to terrestrial wood, and must therefore concentrate carbon 14 by 5 per cent; since this concentra- tion factor is not observed, we see that the activity of carbon in the mixed layer is, in fact, 5 per cent lower than expected. The relation- ships between carbon 13 and carbon 14 varia- tions expected on theoretical grounds, and on the basis of laboratory measurements, were dis- cussed in detail by Craig (1954) who showed that the 5 per cent discrepancy must be the result of slow transfer of carbon from the atmosphere to the sea, and cannot be explained by any other cause. Rafter (1955) verified the conclusion that the carbon 14 difference be- tween atmospheric CO2 and wood must be twice the carbon 13 difference, by direct measurement. The exchange rate of carbon dioxide between atmosphere and sea may be deduced from con- siderations of the steady state relationships be- tween the exchange rate and the radioactive decay rate; this type of evaluation is independ- ent of considerations based on the magnitude of the Suess effect and the kinetics of the transient state. The general equations govern- ing the transfer of a radioactive isotope between its various exchange reservoirs have been given by Craig (1957(a)) in terms of the rela- tionship between the uniform activity which would be observed if all of the sea and the atmosphere were mixed together at a rate in- finitely faster than the radioactive decay rate, and the percentage deviations from this uni- form activity which are actually observed in the different reservoirs. Mixing rates are ex- pressed in terms of the residence time of a molecule in a particular reservoir, which cor- responds to the operational definition of flush- ing time or replacement time, used by oceanog- raphers, and, for the first order processes with which we are concerned, to the reciprocal of the exchange rate constant. The constant radioactive decay rate of carbon 14 furnishes a built-in clock which monitors the transfer rate of carbon between its various reservoirs. For example, if a barrier is inter- posed between the atmosphere and sea, so that the transfer rate of carbon between these reser- voirs is slowed down, the radiocarbon atoms formed in the atmosphere have less probability of getting into the sea and thus of leaving the atmosphere by physical removal. However, the steady state requires that the total number of C" atoms leaving the atmosphere by all mech- anisms be equal to the production of radio- carbon atoms by the cosmic rays, and thus the number undergoing radioactive decay in the atmosphere must increase. The number of radioactive atoms decaying per unit time is a constant fraction of the total number present (the exponential decay law), and therefore the piling up of radiocarbon in the atmosphere because of such an exchange barrier results in an increase in the number decaying in just the way required to maintain the steady state secu- lar equilibrium with the production rate. The percentage increase in the C14 activity of the atmosphere is a function of the ratio between the exchange rate and the decay rate, or, what is the same thing, between the atmospheric residence time and the radioactive mean life. Considering then, the percentage change in the radiocarbon activity of atmospheric CO2 and terrestrial wood, relative to the activity which would characterize these materials in the hypothetical state of infinitely rapid mixing between atmosphere and sea, it is found that for each year of residence time of a CO2 molecule in the atmosphere as a result of slow exchange, the atmospheric activity will increase by 0.74 per cent. The activity in the sea would, of course, decrease as a result of the slower transfer of radiocarbon into the ocean, but since there is some 60 times as much carbon in the sea as in the atmosphere, the percentage de- crease of activity in the sea will be only 1/60 of the atmospheric increase, namely about 0.01 per cent, which is not observable. We can make a more detailed model of the carbon exchange system by breaking the sea

OCR for page 103
116 Atomic Radiation and Oceanography and Fisheries up into a two-layer ocean, taking the upper layer to be about 75 meters deep corresponding to the average mixed layer as actually observed in the sea. (The 75 meter estimate was made by Dr. Warren Wooster, who kindly studied the question of the average depth of the mixed layer over the year in the various areas of the oceans.) The lower layer, extending to a depth of 4000 meters on the average, is termed for convenience the "deep sea," though it is of course obvious that such a uniform layer has little resemblance to the actual structure of the sea below the thermocline. Nevertheless, it is found that the consequences of such an assump- tion about the nature of the deep sea are not serious insofar as affecting the validity of the calculations on the atmospheric residence time, and the treatment of the relationships existing between the atmosphere, mixed layer, and main body of the sea, is of course improved im- mensely by assuming such a model. If we then add a barrier between the mixed layer and the deep sea, representing slow mixing across the thermocline, the radiocarbon is further piled up in both the atmosphere and the mixed layer, in the same manner as previously described. Calculation shows that the activities in the atmosphere and mixed layer are both increased by about 1.2 per cent, relative to the case of a rapidly mixed, uniform sea, for each 100 years of residence time in the deep sea, or, what is almost the same thing, for each 100 years of "age" of the deep water. The activity in the deep sea is reduced by 0.05 per cent for each 100 years of deep-sea residence time. For deep-sea residence times up to several thousand years, the interpolation of a mixing barrier at the thermocline in the sea causes very close to the same activity increase in both the atmosphere and the mixed layer, and thus the activity difference observed between the atmos- phere and mixed layer is sensitive only to the atmosphere-sea exchange rate for internal mix- ing times of the sea of the order of a few thousand years or less. The physical evidence discussed by Wooster and Ketchum in a sep- arate paper in this report, and the tritium cal- culations cited previously in this paper, clearly show that the average mixing time of the sea is at least within this range. Thus the figure cited above of a 0.74 per cent increase in atmospheric activity for each year of atmospheric residence time, indicates that the residence time of a CO2 molecule in the atmosphere, before entering the sea, is about 7 years, corresponding to the 5 per cent activity difference between carbon in the atmosphere and in the mixed layer of the sea. An independent calculation of the atmos- pheric residence time can be made by consider- ing only the steady-state material balance in the atmosphere as a function of the production rate of radiocarbon, taken as (2 ± .5) C" atoms/ cm2 sec, and the rate at which carbon enters the sea. This calculation gives an atmospheric resi- dence time of about 6 years. Considering the errors to be assigned the numerical values in both these calculations, it appears that the best value of the atmospheric residence time of car- bon dioxide may be taken as 7±3 years, cor- responding to a rate constant £0_m = 0.14, where k is the fraction of the carbon in the atmosphere transferred to the mixed layer per year (Craig, 1957 (a)). The average annual exchange flux of carbon dioxide, into and out of the sea each year, is thus found to be about 2x10-' moles per square centimeter of sea surface. This rate is lower by a factor of 10,000 than the rate re- cently obtained by Dingle (1954) from consid- eration of the various rate constants involved, and the discrepancy thus serves to emphasize the power of natural isotopic studies to yield quantitative data, as compared with more tra- ditional methods. An entirely independent calculation of the atmospheric residence time, not based on steady- state considerations, may be made from the magnitude of the so-called Suess effect described previously. It is known that since the begin- ning of the industrial revolution, man has added an amount of carbon dioxide to the atmosphere by fuel combustion equivalent to about 12 per cent of the amount originally present. The degree of dilution of radiocarbon activity in contemporaneous wood by incorporation of C14- free COZ, measured relative to the activity of 19th century wood, is then a measure of the rate at which the dead CO, has been removed from the atmosphere into the sea. The first measurements of this effect, made by Suess (1953), indicated a dilution of about 3 per cent, and from these data Suess deduced an atmospheric CO2 residence time of 20-50 years. More recent and extensive measurements by Suess (1955) have shown that the figure of

OCR for page 103
Chapter 11 117 Tracer Studies of the Sea and Atmosphere 3 per cent is higher than the average world- wide figure, and represents an increased local contamination in trees growing near sites of industrial activity. The latest measurements in- dicate a world-wide effect of about 1.7 per cent. Revelle and Suess (1957) have discussed the relationships between the exchange rate, the Suess effect, the effect of an increase in the atmospheric CO2 content on the atmospheric and oceanic reservoirs, and the buffering effect of the sea water alkalinity on carbon transients. They conclude that, all things considered, the residence time of CO2 in the atmosphere, rela- tive to exchange with the sea, is of the order of 10 years. Though the uncertainty in their esti- mate is a good deal larger than in the case of the steady-state considerations discussed above, the close agreement of the figures obtained by these different considerations is gratifying, and indicates that the factors governing the natural distribution of radiocarbon are now fairly well understood. The size of the terrestrial biosphere and the annual rate of photosynthesis on land have been estimated by Schroeder and Noddack, and from their figures it appears that the terrestrial plants consume about 3 per cent of the atmospheric CO2 per year, corresponding to an atmospheric residence time before entrance into the bio- sphere of 33 years. With a residence time of 7 years, prior to exchange into the sea, the total residence time of a CO2 molecule in the atmos- phere is 6 years, after which it goes either into the sea (9 chances out of 11) or into the ter- restrial biosphere (2 chances out of 11). Thus the carbon dioxide flux into the sea is about 4.5 times larger than the flux into the biosphere, and about 82 per cent of the CO2 leaving the atmosphere goes into the sea, while only about 18 per cent goes into the terrestrial plants. This ratio represents a considerable departure from previous estimates, and indicates that the spatial distribution of plants and soils is probably not the dominant factor in determining the steady- state CO, concentration in the atmosphere. In fact it appears more likely that the spatial pat- tern of absorption and release of CO2 by the sea, and the seasonal variations in this pattern, are the dominant factors. The various considerations outlined above are all consistent with any deep-sea residence time of carbon up to a few thousand years, and do not yield any closer estimate for this figure. Recent unpublished data by Broecker and co- workers at the Lamont Geological Observatory indicate that the bicarbonate of deep ocean waters probably averages about 8 per cent lower in C14 content than the surface mixed layer, corresponding to a radiocarbon "age" of the order of 670 years. However, considerations by Craig (in press), based on a second order oceanic model in which the deep sea reservoir is exposed to the atmosphere in high latitudes, show that about half of the radiocarbon in the deep sea is derived directly from the atmosphere. The other half enters the deep sea from the surface mixed layer of the ocean by the mixing and interchange of water. Because of this dual source of radiocarbon, the residence time calculated for carbon in the deep sea is only about half of the actual resi- dence time of a water molecule in the deep sea relative to the mixed layer; thus the deep-sea residence time of water relative to the mixed layer is probably of the order of 1000 years as a world-wide average. However the actual inter- pretation of such residence times in the sea is quite complicated, and reference is made to the paper cited above for a detailed discussion of carbon and water residence times. t Deuterium and Oxygen 18 As discussed previously, the stable isotopes are of great value in the study of ocean water mixing as additional parameters related to salin- ity. One particular case in which information can be gained from such studies is the problem of meltwater dilution of the oceans in the polar regions. A salinity decrease can be caused by addition of fresh water from river runoff, or from the melting of sea ice, and from salinity data alone these sources cannot be differentiated. However, the isotopic composition of the two sources is quite different; the sea ice should have a composition quite similar to that of the ocean water, while, as shown above, the runoff of rivers in polar areas is greatly depleted in deuterium and oxygen 18 relative to ocean water. Thus from consideration of salinity and isotopic data taken together, a quantitative eval- uation of the mixing conditions can be made. Friedman of the U. S. Geological Survey is currently studying such problems with deute- rium analyses of Atlantic waters. The isotopic data should also be useful in material balance

OCR for page 103
118 Atomic Radiation and Oceanography and Fisheries studies over various sections of the oceans, because of the latitudinal decrease in deuterium and oxygen 18 concentration of oceanic water vapor, and the known temperature dependence of the isotopic selection in evaporation. Craig, Boato, and White (1956) have shown how deuterium and oxygen 18 measurements can be usesd to determine the proportions of juvenile or magnetic water to reheated ground water in thermal springs, and volcanic steam. These isotopes, together with tritium, have im- portant applications to practically all hydrologic problems, and the exploitation of such tech- niques has barely begun. Tritium and Strontium 90 As described in Part II, the tritium measure- ments made by Libby and his co-workers fur- nish an independent value for the mixing rate in the sea; more detailed studies will surely provide important information on the oceanic mixing phenomena. The production of tritium in thermonuclear explosions provides an iso- topic tracer for determination of atmospheric mixing times across the face of the earth and storage times in the atmosphere. The measurement of the world-wide distribu- tion of strontium 90 produced by nuclear deto- nations has been done by W. F. Libby and E. A. Martell at the University of Chicago. The results of their work have recently been described by Libby (1956 a, b). The radio- nuclides produced by low-yield kiloton weapons, and part of the activity produced by the higher- yield megaton weapons, are distributed within the troposphere in a belt corresponding to the latitude of the test site. This material has a tropospheric life which is a function of particle size; some of the activity may circle the earth two or three times within the hemisphere in which it was produced before being washed out of the atmosphere. However, the mean life of this tropospheric material is only a few weeks. More interesting is the fact that Libby and Martell find that half or more of the radio- strontium produced by the megaton weapons is distributed over both hemispheres and falls out much more slowly, the mean storage time in the atmosphere being of the order of ten years. They conclude that this material is car- ried up into the stratosphere, above the tropo- pause, where it is mixed horizontally in a time comparable to the storage time at this level. The contrast between the distribution of megaton weapon produced radiostrontium and tritium is extremely significant. As noted in Part III, Begemann and Libby find that the artificially produced tritium is confined to a single hemisphere and is rapidly washed out of the atmosphere; this material thus follows the pattern of the activities which remain in the troposphere. The tritium and fission prod- uct data thus show that over a period of months there is virtually no cross-hemispheric mixing in the troposphere, but that over a period of years the stratosphere is well-mixed horizontally. The failure to detect tritium carried up into the stratosphere with the megaton weapon produced radiostrontium may be due to the instantaneous combustion of tritium to HTO by the catalytic action of the oxides of nitrogen produced in the blast (Harteck, personal communication). As water, the tritium may be frozen out at the lower cold trap, in the tropopause, where the temperature is about — 70°C, and thus pre- vented from entering the stratosphere. On the other hand, Martell points out (per- sonal communication), that the thermal energy of the fireball is still quite large by the time a fireball produced by a megaton weapon has risen to the height of the tropopause. In order for HTO to condense and thus be trapped be- low the tropopause, it is necessary to assume that the lighter constituents of the fireball have diffused into the cooler outer layers. Martell suggests that if such is the case, then the actual explanation may be that the portion of the cloud containing the HTO may not have sufficient thermal energy to penetrate the tropopause, and as a result, this portion of the cloud merely expands horizontally below the tropopause. V. Conclusions From the discussion in the preceding parts of this report, it is apparent that the advent of manmade nuclear reactions introduced a series of geophysical and geochemical experiments on a vast scale. It is fortunate that the introduction of such experiments came at a time when geo- chemists were well underway towards the under- standing of natural transfer phenomena by means of studies based on naturally occurring isotopes in their steady state biogeochemical cycles. It should be clear that the need for this knowledge is such that every effort should be

OCR for page 103
Chapter 11 119 Tracer Studies of the Sea and Atmosphere made to prevent irreversible procedures which might eliminate the opportunity to study such mixing at the natural level where evaluation of the long term variables is possible. On the other hand, it is also evident that the introduction of artificially produced radioisotopes into the geo- sphere has been productive of a great deal of new knowledge that might otherwise not have been obtained. The importance of continuous monitoring of the levels of such substances as tritium cannot be overemphasized. As an example of this, it may be pointed out that one reason that carbon 14 is such a powerful tool for the evaluation of ocean-atmosphere interaction that we have relatively precise records on just how much dead carbon has been produced by the combustion of fossil fuels; were this information not available the use of radiocarbon in such studies would be exceedingly difficult, if not impossible. From the carbon 14 inventory discussed in Part II, and assuming an average depth of about 150 meters for the oceanic thermocline, it appears that about 4 per cent of the carbon 14 in the sea lies above the thermocline; this corresponds to an activity of about 10 mega- curies. It is thus evident that introduction of artificially produced radiocarbon in 10,000 curie amounts above the thermocline would begin to produce a critical level which would interfere with the natural radiocarbon studies of such fundamental importance. Introduction of 100- 1000 curie amounts above the thermocline would produce activity sites which could be traced for years, but such experiments could not be done more than once every decade or so if the natural level is to be preserved. It would thus seem highly desirable that some international body be constituted to record and monitor the material put into the sea and the atmosphere as wastes and for tracer experi- ments. It is a truism to point out that a con- taminated laboratory is rather easily replaced, but that the laboratory of the earth scientists is not easily renovated. CONCLUSIONS ANDERSON, E. C. 1953. The production and distribution of natural radiocarbon. Ann. Rev. Nuclear Science 2:63-78. ARNOLD, J. R., and H. A. AL-SALIH. 1955. Beryllium-7 produced by cosmic rays. Science 121:451-453. ARNOLD, J. R., and E. C. ANDERSON. 1957. The distribution of carbon-14 in nature. Tellus 9:28-32. BEGEMANN, F. 1956. Distribution of artifi- cially produced tritium in nature. Nuclear Processes in Geologic Settings: Proceed- ings of the Second Conference. National Academy of Sciences — National Research Council Publication 400:pp. 166-171. BENIOFF, P. A. 1956. Cosmic-ray production rate and mean removal time of Beryllium-7 from the atmosphere. Phys. Rev. 104: 1122-1130. CRAIG, H. 1953. The geochemistry of the stable carbon isotopes. Geochim. et Cosmochim. Acta 3:53-92. 1954. Carbon 13 in plants and the relation- ships between carbon 13 and carbon 14 variations in nature. /. of Geology 62: 115-149. 1957 (a). The natural distribution of radio- carbon and the exchange time of carbon dioxide between atmosphere and sea. Tel- lus 9:1-17. 1957 (b). Distribution, production rate, and possible solar origin of natural tritium. Phys. Rev. 105:1125-1127. In press. The natural distribution of radio- carbon: II. Mixing rates in the sea and residence times of carbon and water. Tellus. CRAIG, H., and G. BOATO. 1955. Isotopes. Ann. Rev. Phys. Chem. 6:403-432. CRAIG, H, G. BOATO, and D. E. WHITE. 1956. The isotopic geochemistry of thermal wa- ters. Nuclear Processes in Geologic Set- tings: Proceedings of the Second Confer- ence. National Academy of Sciences — National Research Council Publication 400:pp. 29-38. CURRIE, L. A., W. F. LIBBY, and R. L. WOLF- GANG. 1956. Tritium production by high- energy protons. Phys. Rev. 101:1557- 1563. DE VRIES, H. 1956. Purification of CO2 for use in a proportional counter for "C age measurements. Appl. Sci. Res. (B) 5: 387^00. DINGLE, A. N. 1954. The carbon dioxide exchange between the North Atlantic ocean and the atmosphere. Tellus 6:342-350. DOLE, M., G. A. LANE, D. P. RUDD, and D. A. ZAUKELIES. 1954. Isotopic composition

OCR for page 103
120 Atomic Radiation and Oceanography and Fisheries of atmospheric oxygen and nitrogen. Geo- chim. et Cosmochim. Acta 6:65-78. EPSTEIN, S., and T. K. MAYEDA. 1953. Varia- tion of O19 content of waters from natural sources. Geochim. et Cosmochim. Acta 4:213-224. FIREMAN, E. L., and F. S. ROWLAND. 1955. Tritium and neutron production by 2.2- Bev protons on nitrogen and oxygen. Phys. Rev. 97:780-782. FRIEDMAN, I. 1953. Deuterium content of natural water and other substances. Geo- chim. et Cosmochim. Acta 4:89-103. HARTECK, P. 1954. The relative abundance of HT and HTO in the atmosphere. /. Chem. Phys. 22:1746-1751. KAUFMAN, S., and W. F. LIBBY. 1954. The natural distribution of tritium. Phys. Rev. 93:1337-1344. LANE, G. A., and M. DOLE. 1956. Fractiona- tion of oxygen isotopes during respiration. Science 123:574-576. LIBBY, W. F. 1955. Radiocarbon dating. Univ. of Chicago Press, Chicago: 2nd edition, 175 pp. 1956(a) Radioactive fallout and radioactive strontium. Science 123:656-660. 1956 (b) Radioactive strontium fallout. Proc. Nat. Acad. Sci. 42:365-390. MORRISON, P., and J. PINE. 1955. Radiogenic origin of the helium isotopes in rock. Ann. New York Acad. Science 62:69-92. RAFTER, T. A. 1955. "C variations in nature and the effect on radiocarbon dating. New Zealand ]. Sci. Tech. (B) 37:20-38. RAKESTRAW, N., D. RUDD and M. DOLE. 1951. Isotopic composition of oxygen in air dissolved in Pacific ocean water as a func- tion of depth. /. Amer. Chem. Soc. 73: 2976. REVELLE, R., T. R. FOLSOM, E. D. GOLDBERG, and J. D. ISAACS. 1955. Nuclear science and oceanography. Intern. Conf. on Peace- ful Uses of Atomic Energy, Geneva, also Contr. Scripps Inst. Ocean. N. S. 794: 22 pp. REVELLE, R., and H. E. SUESS. 1957. Car- bon dioxide exchange between atmosphere and ocean and the question of an increase in atmospheric CO2 during the past dec- ades. Tellus 9:18-27. SUESS, H. E. 1953. Natural radiocarbon and the rate of exchange of carbon dioxide between the atmosphere and the sea. Nu- clear Processes in Geologic Settings. Na- tional Academy of Sciences — National Research Council Publication, pp. 52-56. SUESS, H. E. 1954. Natural radiocarbon meas- urements by acetylene counting. Science 120:5-7. 1955. Radiocarbon concentration in modern wood. Science 122:415-417. VON BUTTLAR, H., and W. F. LIBBY. 1955. Natural distribution of cosmic-ray pro- duced tritium. 2. /. Inorganic and Nuclear Chem. 1:75-91.