Charles D. Keeling

The concentration and isotopic abundances of atmospheric carbon dioxide in rural areas

Fifty samples of rural air collected near the Pacific coast of North America have been analysed for carbon dioxide reporting, in addition to concentration in air, the isotopic abundances of C13 and O18. A correlation observed for all samples between C13 isotope abundance and concentration in air can be explained by assuming an initial composition for atmospheric carbon dioxide of 0.031 volume per cent in air, C13C12 ratio−7.0 per mil., to which is added carbon dioxide of plant origin with a ratio of approximately −23 per mil. Minimum concentrations and associated carbon isotope ratios at different stations show very little variation (0.0307–0.0316 per cent, −6.7 to −7.4 per mil) and are believed to be representative of Pacific maritime air. Oxygen isotope abundances are approximately the same as for carbon dioxide in chemical equilibrium with average ocean water, but individual samples show variations which generally do not correlate with changes in concentration in air and are as yet unexplained.

THE CONCENTRATION AND ISOTOPIC ABUNDANCES OF CARBON DIOXIDE IN RURAL AND MARINE AIR

Abstract An earlier study of the concentration and isotopic abundances of C13 and O18 of carbon dioxide in rural air ( Keeling , 1958) has been extended by reporting the analysis of 106 additional samples of rural air and thirteen samples of air collected over tropical waters of the eastern Pacific Ocean. At locations far removed from terrestrial plants and the influence of cities, the concentration and C13 abundance of carbon dioxide in the air are found to be nearly constant; but the O18 abundance of carbon dioxide, under various circumstances, appears to show a systematic variation with air temperature, ocean water temperature, or season. Extreme values are, concentration: from 0.0303 to 0.0320 volume per cent of original air; C 13 C 12 ratio: from −6.7 to −7.4 per mil; O 18 O 16 ratio: −0.8 to −0.6 per mil. A correlation between C13 abundance and concentration of carbon dioxide previously observed for forest air is again observed. On the basis of this correlation, the C 13 C 12 ratio of the carbon dioxide released by the forest plants has been computed and is found to vary for different stations between −21 and −26 per mil. The O18 abundance of carbon dioxide in forest air is observed to be variable but shows no simple relationship with other measured quantities.

Ocean pCO2 calculated from dissolved inorganic carbon, alkalinity, and equations for K1 and K2: validation based on laboratory measurements of CO2 in gas and seawater at equilibrium

Abstract The partial pressure of carbon dioxide in the oceans surface waters, precisely expressed as the fugacity ( f CO 2 ) is determined from dissolved inorganic carbon (DIC) and total alkalinity (TA), and the first and second dissociation constants of carbonic acid, K 1 and K 2 . The original measurements of K 1 and K 2 reported by Mehrbach et al. [Mehrbach, C., Culberson, C.H., Hawley, J.E., Pytkowicz, R.M., 1973. Measurement of the apparent dissociation constants of carbonic acid in seawater at atmospheric pressure. Limnol. Oceanogr. 18, 897–907] are reformulated to give equations for p K 1 and p K 2 (p K =−log 10 K ) as a function of seawater temperature and salinity, consistent with the “total hydrogen ion” concentration scale: p K 1 =3633.86/T−61.2172+9.67770 ln T−0.011555 S+0.0001152 S 2 p K 2 =471.78/T+25.9290−3.16967 ln T−0.01781 S+0.0001122 S 2 By equilibrating solutions of seawater with gas mixtures of known composition, we demonstrate that the above formulations of K 1 and K 2 give calculated f CO 2 values that agree with equilibrated values to 0.07±0.50% (95% confidence interval, f CO 2 up to 500 μatm). Formulations of K 1 and K 2 based on other studies resulted in calculated f CO 2 values approximately 10% lower than the measurements. Equilibrations at f CO 2 above 500 μatm yielded measured f CO 2 values higher than calculated values by on average 3.35±1.22% (95% confidence interval). The cause for the f CO 2 dependence of the results is not known. The uncertainties in p K 1 and p K 2 were combined with the analytical uncertainties typical of contemporary measurements of DIC and TA to reveal the expected reliability of seawater f CO 2 calculated from these parameters. For example, an uncertainty of 1.0 μmol kg −1 in DIC and 2 μmol kg −1 in TA (1 standard deviation (s.d.)) will result in uncertainty of the calculated f CO 2 of 1% or ±3.5 μatm at 350 μatm (1 s.d.).

RECENT TRENDS IN THE C-13-C-12 RATIO OF ATMOSPHERIC CARBON-DIOXIDE

THE 13C/12C ratio of atmospheric carbon dioxide has decreased by approximately 0.6‰ over 22 yr according to new direct measurements reported here. Our results offer a way of establishing whether 13C/12C ratios of tree rings1–6 are representative of atmospheric 13CO2 variations. We have carried out both isotopic (at Groningen) and concentration (at La Jolla) measurements of atmospheric CO2 on air samples obtained during 1977 and 1978 at three widely spaced locations: La Jolla, California (33°N, 117°W), Fanning Island (4°N, 159°W) and the South Pole. Sampling, instrumental, and analytical procedures closely matched a similar study carried out 22 yr earlier by Keeling7,8.

Interannual variations in satellite‐sensed vegetation index data from 1981 to 1991

Normalized difference vegetation index (NDVI) data processed from measurements of advanced very high resolution radiometers (AVHRR) onboard the afternoon-viewing NOAA series satellites (NOAA 7, 9, and 11) were analyzed for spatial and temporal patterns comparable to those observed in atmospheric CO2, near-surface air temperature, and sea surface temperature (SST) data during the 1981–1991 time period. Two global data sets of NDVI were analyzed for consistency: (1) the land segment of the joint NOAA/NASA Earth Observing System AVHRR Pathfinder data set and (2) the Global Inventory Monitoring and Modeling Studies AVHRR NDVI data set. The impact of SST events was found to be confined mostly to the tropical latitudes but was generally dominant enough to be manifest in the global NDVI anomaly. The vegetation index anomalies at latitudes north of 45°N were found to exhibit an increasing trend. This linear trend corresponds to a 10% increase in seasonal NDVI amplitude over a 9 year period (1981–1990). During the same time period, annual amplitude in the record of atmosphere CO2 measured at Point Barrow, Alaska, was reported to have increased by about 14%. The increase in vegetation index data between years was especially consistent through the spring and early summer time periods. When this increase was translated into an advance in the timing of spring green-up, the measure (8±3 days) was similar to the recently published estimate of about 7 days in the advance of the midpoint of CO2 drawdown between spring and summer at Point Barrow, Alaska. The geographical distribution of the increase in vegetation activity was consistent with the reported patterns in springtime warming and decline of snow cover extent over the northern hemisphere land area.

Spatiotemporal patterns of carbon‐13 in the global surface oceans and the oceanic suess effect

A global synthesis of the 13C/12C ratio of dissolved inorganic carbon (DIC) in the surface ocean is attempted by summarizing high-precision data obtained from 1978 to 1997 in all major ocean basins. The data, mainly along transects but including three subtropical time series, are accompanied by simultaneous, precise measurements of DIC concentration and titration alkalinity. The reduced isotopic ratio, δ13C, in the surface ocean water is governed by a balance between biological and thermodynamic processes. These processes have strongly opposing tendencies, which result in a complex spatial pattern in δ13C with relatively little variability. The most distinctive feature in the spatial distribution of δ13C seen in our data is a maximum of δ13C near the subantarctic front with sharply falling values to the south. We attribute this feature to a combination of biological uptake of CO2 depleted in 13C (low δ13C) and air-sea exchange near the front and upwelling further south of waters with low δ13C resulting from the remineralization of organic matter. Additional features are maxima in δ13C downstream of upwelling regions, reflecting biological uptake, and minima in the subtropical gyres caused by strongly temperature dependent thermodynamic isotopic fractionation. At the time series stations, two in the North Atlantic Ocean and one in the North Pacific, distinct seasonal cycles in δ13C are observed, the Pacific data exhibiting only about half the amplitude of the Atlantic. Secular decreases in δ13C caused by the invasion of isotopically light anthropogenic CO2 into the ocean (the 13C Suess effect) have been identified at these time series stations and also in data from repeated transects in the Indian Ocean and the tropical Pacific. A tentative global extrapolation of these secular decreases yields a surface oceanic 13C Suess effect of approximately −0.018‰ yr−1 from 1980 to 1995. This effect is nearly the same as the 13C Suess effect observed globally in the atmosphere over the same period. We attribute this response to a deceleration in the growth rate of anthropogenic CO2 emissions after 1979, which subsequently has reduced the atmospheric 13C Suess effect more than the surface ocean effect.

The Suess effect: 13Carbon-14Carbon interrelations

Abstract The Suess Effect is a term which has come to signify the decrease in 14C in atmospheric CO2 owing to admixture of CO2 produced by the combustion of fossil fuels. This term is here extended, as a concept, to the shifts in isotopic ratio of both 13C and 14C in any reservoir of the carbon cycle owing to anthropogenic activities. To explain this generalized Suess Effect a four reservoir global model of the natural carbon cycle is developed in which isotopic fractionation and radioactive decay are fully taken into account. The model includes the cases in which the deep ocean is treated either as a single undifferentiated box model reservoir or is vertically differentiated with eddy diffusion governing the transport of carbon. Also, the governing equations are expressed with sufficient generality to apply simultaneously to both rare isotopes. In so far as possible, the model is expressed without approximation of the isotopic processes even though this leads to non-linear differential equations to describe the rates of change of rare isotopic carbon within carbon reservoirs. Linear approximations also developed and solved using the method of Laplace transforms. The sensitivity of the predicted Suess Effects to uncertainties in the assigned values of the model parameters is investigated in detail, including estimates of some of the effects of linearizing the governing equations. The approximation of Stuiver, in which the atmospheric Suess Effect is assumed to be 0.018 times the corresponding effect for 14C, is examined in detail and shown to arise when both isotopic fractionation and radioactive decay are left out of the model. This approximation, although correct as to order of magnitude, is found to be too imprecise to be recommended in modeling studies. As found in previous work, the predicted atmospheric Suess Effect for 13C for a given airborne fraction of industrial CO2 is of similar magnitude whether the land biosphere has been a net source or sink of carbon during recent times. On the other hand, the corresponding effect for a surface ocean water is considerably smaller than otherwise if the land biosphere has been a source of CO2 instead of a sink. The model is thus useful in indicating the need to consider isotopes in several reservoirs simultaneously. Although the emphasis is on formulating models rather than surveying and interpreting data, observational data are summarized and compared with model predictions. The oceanic data are seen to be too meager as yet to help settle the question of biospheric response to mans activities.

Global net carbon exchange and intra-annual atmospheric CO2 concentrations predicted by an ecosystem process model and three-dimensional atmospheric transport model

A generalized terrestrial ecosystem process model, BIOME-BGC (for BIOME BioGeoChemical Cycles), was used to simulate the global fluxes of CO 2 resulting from photosynthesis, autotrophic respiration, and heterotrophic respiration. Daily meteorological data for the year 1987, gridded to 1° by 1°, were used to drive the model simulations. From the maximum value of the normalized difference vegetation index (NDVI) for 1987, the leaf area index for each grid cell was computed. Global NPP was estimated to be 52 Pg C, and global R h was estimated to be 66 Pg C. Model predictions of the stable carbon isotopic ratio 13 C/ 12 C for C3 and C 4 vegetation were in accord with values published in the literature, suggesting that our computations of total net photosynthesis, and thus NPP, are more reliable than R h . For each grid cell, daily R h was adjusted so that the annual total was equal to annual NPP, and the resulting net carbon fluxes were used as inputs to a three-dimensional atmospheric transport model (TM2) using wind data from 1987. We compared the spatial and seasonal patterns of NPP with a diagnostic NDVI model, where NPP was derived from biweekly NDVI data and Rh was tuned to fit atmospheric CO 2 observations from three northern stations. To an encouraging degree, predictions from the BIOME-BGC model agreed in phase and amplitude with observed atmospheric CO 2 concentrations for 20° to 55°N, the zone in which the most complete data on ecosystem processes and meteorological input data are available. However, in the tropics and high northern latitudes, disagreements between simulated and measured CO 2 concentrations indicated areas where the model could be improved. We present here a methodology by which terrestrial ecosystem models can be tested globally, not by comparisons to homogeneous-plot data, but by seasonal and spatial consistency with a diagnostic NDVI model and atmospheric CO 2 observations.

The Carbon Dioxide Cycle: Reservoir Models to Depict the Exchange of Atmospheric Carbon Dioxide with the Oceans and Land Plants

The following survey of carbon dioxide in nature places major emphasis on describing how the injection of CO2 into the atmosphere by man’s industrial activity has perturbed the natural carbon cycle on a global scale. In a sense, this injection is a mammoth geochemical experiment. It permits us to observe the transient response of the air, the oceans, and the biosphere to a major disturbance taking place over the interval of only a few years. Our quantitative understanding of the carbon cycle is thus repeatedly challenged and refined.

Atmospheric Carbon Dioxide, the Southern Oscillation, and the Weak 1975 El Niño

The observed rate of change of the atmospheric carbon dioxide concentration at the South Pole, Fanning Island, Hawaii, and ocean weather station P correlates with an index of the southern oscillation and with El Ni�o occurrences. There are changes at all four stations that seem to be in response to the weak 1975 El Ni�o. Thus, even poorly developed El Ni�o events may affect the atmospheric carbon dioxide concentration.