[RADIOCARBON, VOL 22, No. 3, 1980, P 909-918] CARBON-14 AND CARBON-13. IN SOIL CO2 HELMUT DORR and KARL OTTO MUNNICH Institut fur Umweltphysik der Universitat Heidelberg, Federal Republic of Germany ABSTRACT. Carbon isotope measurements in soil CO2 are presented and discussed. Soil CO2 concentration and 13C profiles were measured using a new technique. A simple model describing the CO2 transport from the soil to the atmosphere is derived. The finding that CO, in the soil is richer in 13C than the CO2 leaving the soil is attributed to isotopic fractionation in molecular diffusion. INTRODUCTION 11401 In modeling local variations of the atmospheric -120 ratio due to anthropogenic sources (Levin, Munnich, and Weiss, 1980) there is some doubt as to the precise value of this figure for soil respiration CO2 which considerably influences the atmospheric level. Therefore, 14C soil respiration studies implemented years ago (Munnich, 1963; Munnich and Roether, 1963) have been resumed and are now being supplemented by CO2 concentration and 13C profile measurements in the uppermost 60cm of the soil. The findings are relevant to the understanding of soil respiration mechanisms of atmospheric CO2 balance as well as to the initial value in 14C groundwater dating (Munnich, 1968; Fontes and Gamier, 1979). Theoretical considerations Soil CO2 is produced by respiration of plant roots and bacteriaoxidizing dead organic matter, and takes place primarily in the uppermost half-meter of the soil. If we denote CO2 production per unit soil volume and time at depth z by q(z) we find, for the horizontally homogeneous case in the steady state, -(aj (z)/az) + q(z) = 0. Since, on the time average, practically all CO2 produced leaves the soil surface at z = 0, the CO2 flux density j (z) has its maximum (negative) value at z = 0 and gradually goes to zero at greater depth z. It is generally assumed (de Jong and Schappert, 1972) that gas transport in the soil is primarily by diffusion. It can, in fact, be shown that "pumping" by atmospheric pressure variations only occasionally plays a role where the groundwater table is at greater depth. Turbulent motion in the atmosphere above the soil surface usually is very effectively damped out in the soil because the average momentum relaxation time r in the 20µsec r (Munnich, 1968) giving eg, T 8v for an average pore radius of r = 50j, and a kinematic viscosity v = 0.15 cm`'/sec for air. This means a relaxation length x = v'- well below lmm for outside air motion velocity v. It has been argued, however, (Kraner, Schroeder, and Evans, 1964) that micro-oscillation of atmospheric pressure (or rather the corresponding spatial wave pattern) might produce a kind of slight eddy diffusion in the uppermost soil layers. soil pores is T 909 Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 ] I I I I I I 100 ml 4n NatOH f I I I I I I I I 1 --- I I I I I I I I I I I r--- / / / / / / / / / / / / / / / / / / / / / / / / / / / / / 1. Sampling Tubes Fig / / / / / B. A. Static collection of soil respiration CO2 by alkali absorption under Lundgardh's cup. B. Sampling technique for CO2 concentration and 12C depth distribution measurements. 40cm A. / / / Absorption System Pump & Gasmeter N O Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 C. D. I I I Pirani gauge ,J ' Pressure Transducer r --4Pt-Thermometer I C-13' I ample adsorption flow control pump ' I capillary rotating pump V sxz mm dry ice cold trap stainless steel spirals V li N cod rap i freeze out ' finger ice box sample HZSOs sample bottle(5oml C. CO2 recovery. 1111 heater(50 °c) absorption bottle used with technique b; the air inlet tube stays iii the bottle during acidification and CO2 D. CO2 recovery system for technique b; volumetric measurement of CO2 amount sampled. 912 Soils & Groundwater Diffusion transport of a substance through any medium is defined S.D where S is the solubility, ie, the equilibrium by its permeability' P partition factor of the substance between the medium in question and a reference medium; D is the diffusion constant. In our specific case the obvious reference medium is air which makes S = E, the air filled fraction of the total soil volume (porosity). The available (air-filled) porosity E and the volumetric fraction of soil moisture, F, add up to the total porosity Eo = E + F of the soil. The diffusion constant, D, on the other hand, would be identical to D0, the diffusion constant of CO2 in air, only if the soil consisted of a bundle of straight capillary tubes pointing in the direction of the concentration gradient. In reality, however, we have D = f3°Do with a factor < 1, considering the fact that the reduction of available diffusion cross-section, only on the average, is given by the available porosity, E. Microscopically, this cross-section fluctuates around the average value, larger voids alternate with bottlenecks where the soil grains touch each other. It can easily be shown, from the extreme example of very narrow bottlenecks, that the larger than average cross-section between the two bottlenecks do not compensate their impedance to diffusion. With reference to a similar mechanism, assuming that the diffusing substance cannot travel the direct route, the factor, k = l//3, is often called the tortuosity factor (see, eg, Penman, 1940; Zimmermann, Munnich, and Roether, 1967). EXPERIMENTAL METHOD Figure lA shows static collection by Lundgardh's inverted cup The CO2 diffusing out of the soil surface is collected under a tin container and absorbed quantitatively in three ceramic dishes containing a total of 300ml 4 normal sodium-hydroxide-solution. The average sampling time is two weeks. With this method, soil respiration rate, 13C, and '4C content of respiration CO., are measured. Figure l B shows CO., and 13C depth profile measurement. Twenty thin brass tubes, 70cm long, of 2mm outer diameter, carrying 12 inlet holes of 0.5mm diameter just behind the cone tip are driven to a preset depth into the soil. All probes are connected in parallel and are attached to a 5L Mariotte bottle serving as an air pump and as a gas meter, simultaneously. Soil CO, contained in the air stream is absorbed quantitatively in a 5Oml glass bottle. This absorption bottle contains a packing of stainless steel spirals 2 X 2mm in diameter (Vereinigte FullkorperFabriken, D-5412 Baumbach) which holds 4m1 of a 4 normal NaOH solution like a sponge exposing a very large absorption surface to the passing air (fig 1C). Total air flow rate is about IOL/hr (or 500cc/hr per individual probe). At this flow rate, the natural diffusion steady state in the soil cannot be disturbed. We show this by assuming that a sphere of radius, r, around the suction tip had been flushed free from CO2 by r The permeability P, is often incorrectly called a diffusion constant (see eg, Kraner, Schroeder, and Evans, 1964: "bulk diffusion constant"). This notation is tolerable only if a single and homogeneous medium is to be described (see Stiller and Carmi, 1975 for the treatment of the general case). Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Carbon-14 and carbon-13 in soil CO2 913 injecting C02-free air into the soil. Relaxation time, which restores the natural situation by diffusion, can be estimated from the lowest order relaxation time, T1, of the corresponding eigenfunction representation of the problem (Carslaw and Jaeger, 1959). We find 71 = r2,/(ir2D) r2/60 [min], which means that the disturbance produced by injection of ½L of air (with E = 0.25 this gives r = 7.8cm) decays with a relaxation time in the order of one min. Absorbed CO2 is recovered by adding 5ml of 5 normal sulfuric acid to the sampling bottle (fig 1D). The sample and acid bottles are heated from the bottom and cooled from the top. The stainless steel spiral packing, thus, is continuously flushed with water vapor enhancing the degassing of the system. The whole system is cleaned of air beforehand in the same way. The transfer of CO2 to the vacuum line is again by water vapor (flux controlled by the capillary) condensing in the first cold trap while the CO2 is being collected in the second. The crucial point of our absorption and recovering technique (a continuation and simplification of the technique used by Esser, ms, for atmospheric 13C work) is that there is absolutely no loss of absorbing solution, and the CO2 is recovered quantitatively afterwards. The amount of CO2 is measured volumetrically in a calibrated metal container with precision temperature and pressure meters. Using a Na2CO3 standard solution, the measurement of the CO2 amount sampled is reproducible to ± 1.20 STD. The system allows for checking of degassing of the absorbing solution to better than 0.1/0 of the CO2 amount processed. The 13C measurements are reproducible to ± 0.03/0. For 14C measurement setup, see Schoch and others (1980). RESULTS AND DISCUSSION respiration flux and temperature From our measurements the flux density of soil respiration CO2 supplied to the atmosphere is determined independently by two different techniques. Figure 2 shows both sets of data: the dots give the rates calculated from the concentration versus depth profiles of figure 4 as j = -P(dc/dz), while the histogram shows the amount of CO2 collected under the inverted cup divided by the collection time and the soil area covered by the cup. Figure 3 nicely shows how the flux depends on soil temperature, apparently being approximately doubled by each 5°C temperature increase. The flux density is calculated from the profiles with P = E/3D0 _ .042 cm2/sec, assuming the free porosity to be E _ .20, and the tortuosity reduction fractor, ,3, by which the gas diffusion constant, D0 _ .15cm2/sec, is to be multiplied to be /3 = 2/3 (Penman, 1940; Zimmermann, Munnich, and Roether, 1967a). In view of the uncertainty of the factors E and /3, the agreement between the two data sets is quite acceptable. The soil CO2 profiles have been fitted by a parabolic curve that occurs if constant CO2 production is assumed between 0 and 60cm depth. Inspection shows, however, that, eg, an exponential fit, assuming exponential decrease of source strength with depth at a scale height z 60cm, would be, likewise, possible. CO2 soil Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Soils & Groundwater 914 Note that the flux variation histogram derived from CO2 absorbed under the inverted cup crosses the other curve and gives comparatively lower fluxes at high respiration intensity. This is to be expected if absorption in the alkali solution is not fast enough: the total absorbing alkali surface in our case is just about equal to the soil surface area covered by the inverted cup. Thus, if CO., absorption were at its maximum rate (piston velocity wa = 7 00cm/hr, see Munnich, 1963) the steady state CO2 concentration under the cup would always be below atmospheric CO2 level, and the system would gain CO2 from the atmosphere. Ideally, the concentration under the cup should always be identical with the atmospheric concentration to avoid gain or loss due to diffusion w [Ch ] 125 10 75 50 25 23.4 23.5 22.6 22.7 1.8 Fig 2. Soil respiration CO2 flux density j vs date (1979) presented as the production velocity w = j/cA with standard atmospheric concentration cA = 330ppm. The dots are derived from measured CO2 concentration profiles (fig 4) with E = 0.2; = 2/3; Do = 0.15 cm2/sec. The histogram represents average flux density values for the time intervals indicated, measured directly by the amount of CO2 collected under an inverted cup. T [°C] 20 15 10 5 23.4 23.5 22.6 22.72C8 Fig 3. Soil temperature versus date (1979) at depth, 10cm, measured by a platinum thermometer. Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Carbon-14 and carbon-13 in soil CO2 915 around the rim of the cup. From the data in figure 2, we might estimate that the real absorption piston velocity is about 50cm/hr instead of 700. Statistically, there is a stagnant molecular diffusion layer of air over the absorber of about 10cm thickness. In the future, we shall try adjusting absorption velocity by an internal fan in order to force the cup concentration to approximately atmospheric level. One of us measured soil respiration CO2 under an inverted cup during the period, 1958 to 1962 (Munnich, 1963; Munnich and Roether, 1963) finding that the 14C level in soil respiration seemed to follow the increasing atmospheric level with a time lag of about 2 to 3 years. The present data (table 1) shows no significant difference in 14C level of soil 14G. c/co cic B5 40 40 30 30 20 m 10 10 5 5 15 30 45 601 15 C B7 40 30 m 20 20 10 10 5 5 0 45 6 CM o 45 0CM 45 6 CM 612 40 15 610 15 30 Fig 4. CO2 concentration versus depth profiles taken at the sample location of table 1. The parabolic fit to data points assumes constant CO2 production between 0 and 60cm depth (see text). With this assumption and the values of fig 2 for E, f3, Do one obtains the following CO2 source densities q: B5: 10 April 79; q = 24mmoles C02/m3 hr B7: 9 May 79; 29 B10:26 June 79; B12:11 July 79; 59 59 Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Soils & Groundwater 916 CO2 if compared with contemporary atmospheric levels. Although the lowermost 100m of the atmosphere are strongly influenced by soil respiration CO2 (Levin, Munnich, and Weiss, 1980) the 14C level in the atmospheric CO2 is obviously altered very little by this. This greatly simplifies the interpretation of man-made and Weiss,1980). 14C variations (Levin, Munnich, 13C. If the biospheric system under investigation is basically in a steady state, respiration CO2 derived from organic matter with 813C = 25% must, on the average, have 8130 = 25%, as well. This, of course, does not exclude time variations around the average value. Both our flux and profile data show such variations with time. However, particularly during the time of high respiration flux, ie, after mid-May, CO2 in the soil is richer in 13C than CO2 leaving the soil by about 3.5%. This is to be expected if transport is, as generally assumed, primarily by molecular diffusion: Craig (1954) notes that the diffusion constant of 13002 should be by 4% smaller than the one of 12C02. As in evaporation from a tree leaf (Zimmermann, Ehalt, and Munnich, 1967b), making water in the leaf heavier than water passing through it and being transpired to the atmosphere, we should find CO2 in the soil 4% heavier than that in the flux leaving the soil. No variation of the isotopic composition with depth should occur despite the strong increase of CO2 concentration with depth. This is easily verified by Fick's law, and it seems to be exactly what we observe (fig 5). This behavior has been noticed before (Fontes and Gamier, 1979), but, to our knowledge, has not been explained according to our methods. - - CONCLUSION The data presented shows that the technique used yields reliable information on soil respiration fluxes and isotopic composition. We TABLE 4C Sampling time 16.3 24.3 30.3.79 --10.4.79 2.4- 12.4.79 12.4 -19.4.79 19.4- 25.4.79 27.44.5 4.5.79 -14.5.79 10.4- 9.5.79 - 30.5.79 14.5-25.5.79 9.5 1.6- 8.6.79 30.5 -13.6.79 8.6-15.6.79 - (%) atm CO. HD 278 ±5 284 293 ± ± ± 285 ± 286 ± 281 ± z14C (%) soil respiration CO2 HD 5 5 5 (%) 6 - 289± 5 -24.6 293 ± 5 282 ±5 5 318±4 613C respiration CO2 ± 287 6 6 311±5 288 1 20.2 --24.8 25.8 25.9 13.6 26.6.79 296 ± 5 25.4 26.6-11.7.79 286± 5 74C and 13C data on soil CO2 collected by alkali absorption under an inverted cup (sample location on uncultivated sandy soil with loess loam admixtures 15km south of Heidelberg). 14C data on atmospheric CO2 samples collected in Heidelberg are also given for comparison. Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Carbon-14 and carbon-13 in soil CO2 15 60 CM 45 30 15 -22 -22 -24 - -26 -26 -28 - 28 B5 15 45 30 45 60 CM 45 60CA1 24 B10 d C-13 30 917 15 60CM 30 -22 -24 - 24 -26 -26 -26 -28 B7 dC-13 Fig 5. Soil S13C B12 C-13 profiles measured; see fig 4 for sampling dates. 23.4 2.5 2?.6 237 21.8 -221 -24 - 26 -28 d C-13 Fig 6. 3C in soil CO2 versus date of sampling (average values for individual profiles). Broken line shows average isotopic composition of CO2 leaving the soil (table 1). Note that after mid-May the standing crop in the soil is heavier isotopically than CO2 in the soil respiration flux. Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316 Soils & Groundwater 918 shall continue these measurements in combination with radon-222 measurements in the soil using the detector designed by Roether and Kromer, (1978). Some tentative results have already been obtained (Volpp, personal commun, 1979). One of the aims of this combined study is to provide better flux data for soil-born gases to calibrate a regional atmospheric model. This will be developed to connect atmospheric concentrations with fluxes to and from the earth's surface by using meteorologic data such as atmospheric stability (Levin, Munnich, and Weiss, 1980). The study also relates to the question of initial 14C content in groundwater. REFERENCES Carslaw, H S and Jaeger, J C, 1959, Conduction of heat in solids: Oxford, Clarendon Press, p 233. Craig, Harmon, 1954, Carbon 13 in plants and the relationships between carbon 13 and carbon 14 variations in nature: Jour Geology, v 62, p 122. Esser, Norbet, ms, 1975, Messung der absoluten Konzentration and des 13C-Gehalts von atmospharischem CO2: Thesis, Inst f Umweltphysik, Heidelberg. Fontes, J C and Gamier, J M, 1979, Determination of initial 14C activity of the total dissolved carbon: A review of the existing models and a new approach: Water Resources Research, v 15, no. 2, p 399-413. de Jong, E and Schappert, H J V, 1972, Calculation of soil respiration and activity from CO2 profiles in the soil: Soil Science, v 113, p 328-333. Kraner, H W, Schroeder, G L, and Evans, R D, 1964, Measurements of the effects of atmospheric variables on Rn-222 flux and soil-gas concentrations, in The natural radiation environment: Houston, Texas, Rice Univ Pub, p 191-215. Levin, Ingeborg, Miinnich, K 0, and Weiss, Wolfgang, 1980, The effect of anthropogenic CO2 and i4C sources on the distribution of 14C in the atmosphere, in Stuiver, Minze and Kra, Renee, eds, Internatl radiocarbon conf, 10th, Proc: Radiocarbon, v 22, no. 2, p 379-391. Munnich, K 0, 1963, Der Kreislauf des Radiokohlenstoffs in der Natur: Naturwiss, v6,p211-213. Isotopendatierung von Grundwasser: Naturwiss, v 55, no. 4, p 158-163. Wolfgang, 1963, A comparison of carbon 14 ages and tritium ages of groundwater, in Radioisotopes in hydrology: Vienna, IAEA, p 3971968, Munnich, K 0 and Roether, 406. Penman, H L, 1940, Gas and vapour movements in the soil: Jour Agric Science, v 30, no. 3, p 437-461. Roether, Wolfgang and Kromer, B, Field determination of air-sea gas exchange by continuous measurements of radon-222: Pure appl Geophys, v 116, p 476-485. Schoch, Hildegard, Bruns, Michael, Miinnich, K 0, and Miinnich, Marianne, 1980, A multi-counter system for high precision carbon 14 measurements, in Stuiver, Minze and Kra, Renee, eds, Internatl radiocarbon conf, 10th, Proc: Radiocarbon, v 22, no. 2, p 442-447. Stiller, M and Carmi, I, 1975, Water transport through Lake Kinneret sediments traced by tritium: Earth and Planetary Sci Letters, v 25, p 297-304. Zimmermann, Uwe, Ehalt, D and Miinnich, K 0, 1967, Soil water movement and evotranspiration: changes in the isotopic composition of the water, in Isotopes in hydrology: Vienna, IAEA, p 567-585. Zimmermann, Uwe, Miinnich, K 0, and Roether, Wolfgang, 1967, Downward measurement of soil moisture traced by means of hydrogen isotopes: Washington, D C, Am Geophys Union, mon 11, p 28-36. Downloaded from https://www.cambridge.org/core. IP address: 168.151.145.2, on 05 Dec 2017 at 22:15:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0033822200010316