This paper is concerned with the expected deviations in the production rate of natural 14C on the earth due to changes in solar activity. We review the published estimates of the global production rates of 14C due to galactic and solar cosmic ray particles, and present new estimates of the expected secular variations in 14C production, taking into account the latest information available on galactic cosmic ray modulation and long-term variations in solar activity.

We have investigated solar phenomena associated with unusual changes in the production rates of 14C in the atmosphere. 14C is produced in interactions of cosmic ray neutrons with nitrogen in the atmosphere. Intensity of the neutrons varies globally and fluctuates with time as a result of interactions of galactic cosmic rays which generate neutrons with plasma and magnetic fields of the solar wind. We estimate the total mean production rate of 14C for solar cycle 20, specifically 1965 to 1975, to be 2.25 ± 0.1 nuclei-cm−2sec−1 from galactic cosmic rays alone, with negligible integrated contribution from solar particle events. Annual averages of Rz, the Zurich sunspot number, and the production rate of 14C, n(14C), were related by n(14C) = 2.60–5.53 × 10–3 Rz ± 3 percent. The contribution of solar flare particles and the zero sunspot limit are discussed with relation to major fluctuations that appear in the radiocarbon versus dendrochronology over short (∼100 years) integration times.

Natural atmospheric 14C changes are caused by fluctuations in upper atmospheric 14C production rates (Q) that are related to earth geomagnetic field variations and changes in solar wind magnetic shielding properties. Climate variability may also be responsible for some of the changes because it influences exchange rates of 14C between the various terrestrial carbon reservoirs.

Upper atmospheric 14C production rates QM, in at/sec cm2 (earth), were calculated for the past 1200 years from the atmospheric 14C record and a carbon reservoir model. The changes in QM are compared in detail with the predicted Q variability derived from an Aa solar modulation mechanism and 20th century neutron flux observations. The influence of earth geomagnetic field changes on the magnitude of the solar wind modulation is discussed, and it is shown that the variations in this magnitude agree with the known differences in earth magnetic field intensity during the past 1200 years. The larger calculated QM oscillations during the sixth millennium bp also agree with this concept.

Solar wind magnetic as well as geomagnetic forces modulate the incoming cosmic ray flux and explain the main features of the atmospheric 14C record. It is argued that climatic fluctuation is not a dominant cause.

The oscillations between 3200 and 3700 BC, as measured by de Jong, Mook, and Becker, differ in rise time from those found for the current millennium.

A box-diffusion model for the carbon cycle is used to estimate the magnitude of 14C variations caused by changes of reservoir sizes and exchange fluxes in the global carbon system. The influence of changes in atmospheric CO2 concentration, biomass, CO2 exchange rate between atmosphere and ocean, and ocean mixing is considered. Steady-state 14C concentrations as well as the transients are calculated. For changing biomass, atmospheric CO2 levels and 13C/12C ratios are also calculated.

Carbon-cycle-induced 14C variations may have been significant in the transition period from Glacial to Postglacial when drastic changes in environmental conditions took place within short time periods, while they were probably less important during the climatically more stable Postglacial.

Changes of the oceanic circulation, as supposedly occurred, are considered the most important factor, besides variations of the production rate, affecting the global distribution of 14C. 14C variations due to changes of the atmospheric CO2 level or the air-sea-exchange probably did not exceed one to a few percent. Fluctuations of the forest biomass, which may have occurred between Glacial and Postglacial, hardly affected the 14C concentration over a long term.

Responses of the atmospheric 14C concentration are also calculated for variations of the 14C production rate by cosmic radiation. The following cases are considered: a step change, square-wave changes producing “wiggles”, and sinusoidal variations.

Palaeomagnetic field strength measurements for the last 50,000 years are summarized. The period before ~12,000 yr bp is characterized by low dipole moments, but high values are associated with the Lake Mungo polarity excursion between ~32,000 and ~28,000 yr bp. The variation since 12,000 yr bp, based on new results from Australia and published data from the Northern Hemisphere has a quasi-cyclic appearance with maxima at ~10,000 and ~3500 yr bp. The geomagnetic record is used to predict variations in atmospheric 14C concentration, and the results are compared with independent comparisons between 14C and other dating methods. Long-term variations in the 14C time-scale are readily explained by known geomagnetic changes.

It has become generally accepted during the last year that 14C fluctuations, the so-called “wiggles”, observed in wood, dated by its tree rings, do indeed exist. Furthermore the La Jolla measurements show that apart from experimental noise, they do not represent random red noise, but characteristic, recurring features. In 1971, Houtermans found indications for the existence of cyclic components and recent Fourier analyses of all the available data by Neftel and Hartwig show a 200-year component. Cyclic oscillations with other periods appear to be present during limited time intervals. The character of the oscillations is not harmonic. The time derivative of many fluctuations is remarkably constant and such that the 14C rises by 1 percent in about 20 years and decreases by 1 percent in slightly more than twice that length of time. The properties of the overall radiocarbon record have to be considered in attempts to explain the variations in terms of variations of the cosmic ray-production rate and changes of the geochemical distribution of radiocarbon.

Ionium dates for the Upper Lisan Formation in the Dead Sea valley average (10 ± 3) percent higher than a set of radiocarbon dates from the same profiles. No analytical explanation can be found so that the discrepancy may be real for the period 15,000 BP to 35,000 BP (conventional radiocarbon years). This would have implications for the chronology of the Upper Pleistocene.

Radiocarbon variations between 3900 and 2800 bc have been established in La Jolla and Groningen using oak tree rings from a 2350-year floating oak series. Comparison of these variations with the bristlecone pine 14C variations provides precise ages for tree-ring dates of Neolithic settlements of Switzerland and Germany over a period of 1400 years. 14C variations measured in Heidelberg in absolutely dated oak-ring series from ad 250 to 720 show trends similar to those of long-term growth variation of oaks during the same period of time. The influence of the climatic regime on oak growth of this period is discussed.

Information on the history of the atmospheric CO2 content and the 13C/12 and 14C/C ratios is recorded in natural ice. Measurements on samples from very cold accumulation regions show that CO2 is occluded not only in air bubbles, but also in the ice lattice. The two CO2 components are of similar size. It is very difficult to measure CO2 in the bubbles and CO2 in the ice lattice separately. By melting the samples and extracting the evolving gases in two fractions, it is possible to estimate CO2 concentration in the bubbles and the ice lattice. Enrichment or depletion of CO2 in the bubbles by exchange with the ice is difficult to estimate. Information about this effect is expected from 13C/12C analysis on the extracted CO2 fractions.

To investigate whether atmospheric CO2 content was different during the last glaciation than during the present one, sets of 16 and 20 samples distributed over the last 40,000 years from the two deep ice cores from Camp Century (North Greenland) and Byrd Station (West Antarctica) were measured. The time scales for the two cores are based on a rheological model. Results and conclusions are:

1. — The data series from both cores show similar trends correlated to a certain degree to the δ18O profiles.

2. — For both cores, the values for the CO2 concentration of the first fraction, considered to best represent the atmospheric composition, show lower values during glaciation than in the Holocene, with a minimum before the end of glaciation.

3. — Low CO2 concentrations in the first fractions (200ppm) of certain samples are a strong indication that the atmospheric CO2 concentration during last glaciation was lower than during the postglacial. These low concentrations indicate that, at that time, CO2 concentration in the atmosphere could have been lower than today by a factor of 1.5. Possible explanations for such a change in atmospheric CO2 content as well as its influence on atmospheric 14C/C ratio and on the radiation balance is discussed.

The concentration of radiocarbon in the earth's atmosphere is used to analyze the complex of problems associated with the study of various astrophysical, geophysical, and geochemical phenomena within the framework of the All-Union problem of Astrophysical Phenomena and Radiocarbon. Effects of the supernovae explosions, solar activity, geomagnetic field and climatic changes in the 14C level in tree-rings are considered.

A procedure is outlined for using cosmogenic 10Be variations in polar ice cores and lake or inland sea sediments to correct for 14C production variations in the past. Some of the requirements and problems associated with such a procedure are discussed.

The correlations between atmospheric radiocarbon cencentrations, on the one hand, and solar activity and the earth's magnetic field, on the other, as well as possible relations between surface temperatures and solar activity and the earth's magnetic field are considered. To draw a consistent picture of these relations, the following mechanisms for climatic changes are proposed: The earth's magnetic field and solar activity both influence the fluxes of charged solar particles into the upper stratosphere; higher surface temperatures are positively correlated with higher particles fluxes. Charged particles produce nitric oxide, which controls, to some extent, the ozone destruction and, thus, the ozone inventory in the upper layers of the stratosphere (above the 10 mbar level) in latitudes greater than about 60°. The varying ozone contents in the upper stratosphere affect the radiation balance and the temperatures on the earth's surface. It has been estimated that a reduction, or increase, respectively, of ozone in the layers above 10 mbar warms, or cools, respectively, the earth's surface. A change of ozone in the lower layers of the stratosphere has the opposite effect. The maximum net effect is in the order of ΔTs ≈ 0.3–1 K, in both directions. This hypothesis might be in accordance with an orbital origin of long-term climatic changes.

The absolute time of growth of a floating tree-ring series is determined by computer, by first fitting a curve to the 14C ages of tree-ring dated wood and then by matching the 14C ages of the floating tree-ring series to that curve. The results obtained by this matching procedure are given here for five floating European oak chronologies for which 14C dates have been published previously by Suess (1978). Three of these five floating series have been linked together dendrochronologically. There now remain three floating series covering a 4000-year time span from 4820 BC to 830 BC. Their 14C dates, matched by computer to the bristlecone pine chronology, provide a possibility of obtaining precise ages of oak wood series from Neolithic sites in Switzerland and South Germany by dendrochronologically cross-dating with the calibrated master chronologies.

High-precision 14C measurements are presented, carried out on single tree rings from a section of the floating South German Neolithic tree-ring chronology. They confirm the existence of pronounced medium-term variations in the order of 2 percent during the 33rd to 38th centuries BC.

These variations turn out to be very regular while the precision of 1.5‰ allows a comparison with a geochemical model calculation. Good agreement is acquired for an input function with a periodicity of about 150 years and an amplitude of approximately 30 percent in the 14C production rate.

High precision radiocarbon measurements (±2.5‰ STD) were carried out on absolutely dated European oak material. These showed that significant short-term atmospheric 14C variations of up to 2 percent occurred between AD 200 and 800. These changes, reflecting a 150-to 180-year periodicity, seem to correlate with the average sunspot activity around that time. In comparison with results obtained by other authors, a very regular pattern of natural atmospheric 14C variations is exhibited.

At the Ninth International Radiocarbon Conference a paper was presented concerning possible short-term 14C variations in 4000-year-old red deer antlers (Cervus elaphus) found in Neolithic flint mines in Norfolk, England (Burleigh and Hewson, 1980). It was argued, on archaeologic grounds, that the true age of the samples varied by a few years at most. Their radiocarbon ages, however, varied by a considerably greater amount than could be explained by the errors in the measurements. Duplication of the measurements confirmed this unexpected variation.

Farmer and Baxter (1972) claimed a significant correlation of atmospheric 14C levels in the northern hemisphere with sunspot number based on radiocarbon assay of single tree rings for the period 1829 to 1865. In contrast, Stuiver (1978) stated that a series of single-year Douglas Fir measurements did not show a statistically significant periodicity.

This paper reports measurements made on a series of single-year dendrochronologic samples of approximately the same age as the red deer antlers, which have been carried out at the British Museum Research Laboratory. The results show that the variation in 14C between rings is not statistically significant; some other explanation must be sought for the anomalous antler measurements.

The 14C age of charcoal and wood from Lebanon and central Europe is compared partly with the dendrochronologically determined age of the samples and partly with the archaeologically expected value. While the dendrologic values approximately confirm the correction curve of Ralph, Michael, and Han (1973), charcoal of burned layers seemes to be generally 2 to 3 centuries older than expected from contemporary archaeologically retrieved materials.

The secular patterns of Δ and δ13C measured in wood cellulose are compared with a palaeotemperature index published for the immediate growth region. 14C enrichment shows an overall decrease of ca 20% (Δ) during the 18th and 19th centuries with shorter term (decade) variations superimposed on the general trend. Measured Δ values correlate significantly with the average mean summer temperatures and indicate a linear coefficient of ca − 10.2% °C–1. Short-term variations in the natural concentration of 14C which are closely related to climatic change may reflect the latitudinal dependence for atmospheric mixing.

The secular trend in δ13C covers a range of 2% but variations of this magnitude reflect influences other than those directly associated with climatic change, eg, possible assimilation of recycled biospheric CO2 during early life and post AD 1850 anthropogenic effects. No evidence was found to suggest a significant correlation between δ13C and the mean atmospheric temperature during summer growth periods.