INTRODUCTION
Radionuclides represent unique tracers of environmental processes, which help to study not only processes of their origin, but also their spread in the atmosphere and in Earth’s other reservoirs up to their final storage in isotope archives that record their temporal evolution (in tree rings, glaciers, stalagmites/stalactites, corals, freshwater and marine sediments) (e.g., Lal Reference Lal1999; Beer et al. Reference Beer, McCracken and von Steiger2012; Reimer et al. Reference Reimer2020; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021). Except for radionuclides of natural origin (e.g., primordial radionuclides such as 232Th, 235U, 238U, and products in their decay chains), and of anthropogenic origin (e.g., 3H, 14C, 90Sr, 129I, 137Cs, Pu isotopes, etc.), released to the environment mainly after atmospheric nuclear weapons tests (global fallout), nuclear accidents (Chernobyl, Fukushima) or from nuclear reprocessing facilities (e.g., Sellafield in UK and Le Hague in France) (Povinec et al. Reference Povinec, Hirose and Aoyama2013, Reference Povinec, Hirose, Aoyama and Tateda2021), an important group of radionuclides, applied in environmental studies, are cosmogenic radionuclides, produced by interactions of cosmic-ray particles mainly in the Earth atmosphere (e.g., 3H, 7Be, 10Be, 14C, 22Na, 26Al, 36Cl, etc.) (Lal and Peters Reference Lal, Peters and Sittle1967; Beer et al. Reference Beer, McCracken and von Steiger2012; Masarik and Beer Reference Masarik and Beer1999, Reference Masarik and Beer2009).
The radioactive isotope most frequently used in environmental studies has been radiocarbon (14C) due to its suitable half-life (5700 yr) for studying processes over the last 50,000 years, availability in environmental archives, and existing high-sensitivity detection methods, namely accelerator mass spectrometry (AMS). After its production in the atmosphere as a result of interactions of secondary cosmic-ray particles (mainly neutrons) with nitrogen and oxygen, radiocarbon oxidizes to CO and later to CO2, which, as a result of exchange processes, enters the biosphere and the ocean. The best 14C archives are therefore materials with annually growing mass records, such as tree rings. Radiocarbon has also been important due to its specific stratosphere-troposphere-biosphere mixing, exchange of carbon dioxide with surface ocean, and finally its sequestration into the deep ocean, which is the main reservoir of radiocarbon (Siegenthaler et al. Reference Siegenthaler, Heimann and Oeschger1980; Bard et al. Reference Bard, Raisbeck, Yiou and Jouzel1997; Levin and Hesshaimer Reference Levin and Hesshaimer2000). Radiocarbon has therefore become a very useful cosmogenic radionuclide, important for better understanding of exchange processes in the environment.
Environmental studies with radionuclide tracers are built on a foundation of many scientific fields, starting from astrophysics (cosmic rays, their production, and variations in the heliosphere), via nuclear physics (nuclear reactions, cross-sections, simulations of production rates), analysis of samples in suitable archives, and finally carbon cycle modeling (e.g., Lal Reference Lal1999; Beer et al. Reference Beer, McCracken and von Steiger2012; Masarik and Beer Reference Masarik and Beer1999, Reference Masarik and Beer2009; Levin and Hesshaimer Reference Levin and Hesshaimer2000). Radionuclides, especially the cosmogenic ones, have been widely used in astrophysics, in research on climate change, in environmental pollution studies, in dating of geological samples, and in dating of historical events and cultural heritage samples (e.g., Reimer et al. Reference Reimer2020; Usoskin et al. Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020; Mekhaldi et al. Reference Mekhaldi, Adolphi, Herbs and Muscheler2021; Povinec et al. Reference Povinec, Hirose, Aoyama and Tateda2021).
The closest source of cosmic radiation is the Sun (solar cosmic rays, SCR), emitting mainly protons during solar flare eruptions, which are accelerated in the solar corona and interplanetary space by shocks driven by coronal mass ejections to maximum energies of about 10 GeV. Such solar proton events (SPEs), also called solar energetic particle (SEPs) events (since they contain a few % of He and other nuclei), produce short but very intensive fluxes of energetic particles that could enter the Earth’s atmosphere (Shea and Smart Reference Shea and Smart2000; Usoskin Reference Usoskin2013; Usoskin et al. Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020).
Galactic cosmic rays (GCR) are mostly composed of protons (87%) and α-particles (12%), with a smaller contribution (1%) from heavier nuclei, having proton energies between 1 – 1015 MeV (Dunai, Reference Dunai2010). After entering the heliosphere, they are significantly influenced by the Sun due to turbulent solar wind, accompanied by heliospheric magnetic fields. The GCR in the heliosphere, due to changes in the solar activity (mainly by the 11-year solar cycle), undergo large temporal variations in their intensity and energy. This modulation effect is important for proton energies between 0.020–20 GeV, having a maximum at 500 MeV, representing a variation in GCR flux by an order of magnitude over a solar cycle. The 11-year cycle in GCRs is delayed from a month up to two years with respect to the sunspot numbers which indicate the solar activity (Usoskin Reference Usoskin2017).
As GCR and SCR are charged particles, they are affected by the Earth’s magnetic field, which causes a shielding effect expressed by the cutoff rigidity Footnote 1 required for a particle to reach the Earth surface at a given location. Cosmic rays with higher rigidity are able to penetrate deeper into the atmosphere before being deflected or stopped by the Earth’s magnetic field (especially at the poles, where the shielding effect of the geomagnetic field is weaker) (Masarik and Beer Reference Masarik and Beer1999, Reference Masarik and Beer2009).
Cosmic rays produce in the atmosphere cascades of secondary particles, including nucleons, which then participate in nuclear reactions with atmospheric atoms, producing cosmogenic radionuclides. After their formation, the radionuclides enter into atmosphere-hydrosphere-biosphere exchange processes, and finally they are deposited in isotope archives, carrying signals that were affecting their production and transport in the environment (e.g., Masarik and Beer Reference Masarik and Beer2009; Beer et al. Reference Beer, McCracken and von Steiger2012; Reimer et al. Reference Reimer2020; Shea and Smart Reference Shea and Smart2000; Usoskin Reference Usoskin2013; Usoskin et al. Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020; Mekhaldi et al. Reference Mekhaldi, Adolphi, Herbs and Muscheler2021).
In this paper, we review past radiocarbon variation studies in the biosphere (mainly tree rings), which were associated with changes in solar activity during 1785–1950, mostly represented by 11-year solar cycles (SCs), also called Schwabe cycles. As the Sun modulates the intensity of GCR in the heliosphere, it was reasonably expected that it could have an impact on the 14C variations in the atmosphere and biosphere.
Stuiver (Reference Stuiver1961, Reference Stuiver1965) was the first (using data of Willis et al. Reference Willis, Tauber and Münnich1960) to notice an inverse correlation between 14C levels in tree rings and Sun activity (sunspot numbers), suggesting a 14C cycle with a period of ca. 100 yr (the Centennial solar cycle). The well-known Suess wiggles (Suess, Reference Suess1980) had a characteristic period of ca. 200 yr (also called the Suess cycle). In both cases, the amplitude of the 14C variations was estimated at 10–20‰.
The 11-Year Solar Cycles
The 11-year SCs have been recorded by regular observation of the number of sunspots on the Sun, well documenting the solar activity after AD 1700, although the first observations started in 1610. During the Maunder Grand solar minimum (1621–1718), almost no sunspots were observed, and later two weaker minima, the Dalton (1797–1823) and the Gleissberg (1878–1933), were identified (Figure 1) (Petrovay Reference Petrovay2010; Gao Reference Gao2016). The Centennial solar cycle (1913-2019) and the Modern solar activity maximum (1933-2008) can also be identified in the sunspot number record presented in Figure 1. It is interesting to note that during the Modern solar maximum the solar activity was higher than the sunspot number record (reconstructed from 14C measurements in tree rings) over the last 8000 years (Solanki et al. Reference Solanki, Usoskin, Kromer, Schüssler and Beer2004).
Figure 1 International sunspot numbers for the 19th and 20th century, indicating 11-year solar activity cycles, including Dalton (1797–1823) and Gleissberg (1878–1933) solar activity minima. The Centennial solar cycle (1913–2019) and the Modern solar activity maximum (1933–2008) can also be identified. Source: Royal Observatory of Belgium (https://www.sidc.be/SILSO/home).
The solar activity has an impact on the flux of GCR in the heliosphere, which anti-correlates with 11-year SC and produces regular variations in the intensity of GCR in the Earth atmosphere. Although the 11-year SC is the most prominent cycle observed in solar activity, short-term 14C variations are difficult to observe because of the complex mechanism of 14C production by cosmic rays, its transport from the atmosphere to the ocean and biosphere, and climatic and reservoir influences. Lingenfelter and Ramaty (Reference Lingenfelter, Ramaty and Olsson1970) did the first calculations of the 14C production rates during the 11-year SC, which could vary from solar minimum to solar maximum by 20–30%. Recently, more precise calculations by Usoskin et al. (Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020) for the strongest 11-year SC (1954–1964, Figure 1) showed 14C production rates ranging from 1.3 to 1.9 atoms cm-2 s–1, (the varying factor of 1.5). However, because of the complicated carbon cycle, which acts as a buffer filtering the 14C signal (Siegenthaler et al. Reference Siegenthaler, Heimann and Oeschger1980; Bard et al. Reference Bard, Raisbeck, Yiou and Jouzel1997), the 11-year 14C cycle is attenuated by a factor of ca. 100 to the magnitude of 1–3‰.
Damon et al. (Reference Damon, Long and Wallick1973a,b) performed the first high-precision 14C measurements in annual tree-ring samples (1940–1954), which showed Δ14C amplitude variation during the 11-year SC of around 2‰. These results were later confirmed in several other investigations (e.g., Povinec Reference Povinec1977, Reference Povinec1987; Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a, Reference Burchuladze, Pagava, Povinec and Usačev1980b; Stuiver Reference Stuiver1980, Reference Stuiver1982; Stuiver and Quay Reference Stuiver and Quay1980; Povinec et al. Reference Povinec, Burchuladze and Pagava1983; Attolini et al. Reference Attolini, Galli, Nanni and Povinec1989; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021).
Recently, another interesting phenomenon was investigated, when in addition to regular 14C variations during the 11-year SCs, rapid 14C increases up to 12‰ were found in tree rings, which could be associated with the emission of solar protons (e.g., Miyake et al. Reference Miyake, Nagaya, Masuda and Nakamura2012, Reference Miyake, Masuda and Nakamura2013; Usoskin, Reference Usoskin2013; Usoskin et al. Reference Usoskin, Kromer, Ludlow, Beer, Friedrich, Kovaltsov, Solanki and Wacker2013, Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020; Mekhaldi et al. Reference Mekhaldi, Muscheler, Adolphi, Aldahan, Beer, McConnell, Possnert, Sigl, Svensson, Synal, Welten and Woodruff2015, Reference Mekhaldi, Adolphi, Herbs and Muscheler2021; Park et al. Reference Park, Southon, Fahrni, Creasman and Mewaldt2017; Wang et al. Reference Wang, Yu, Zou, Dai and Cheng2017; Jull et al. Reference Jull, Panyushkina, Miyake, Masuda, Nakamura, Mitsutani, Lange, Cruz, Baisan, Janovics, Varga and Molnár2018; O’Hare et al. Reference O’Hare, Mekhaldi, Adolphi, Raisbeck, Aldahan and Anderberg2019; Reimer et al. Reference Reimer2020; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021).
SAMPLES AND METHODS
Thanks to advanced radiometric methods for radiocarbon measurements, it has been possible to study even fine radiocarbon variations in the biosphere, which could be associated with the 11-year SC (Povinec Reference Povinec1977, Reference Povinec1987). A lime tree (Tilia cordata) from the Prešov region, Slovakia (48º52’N, 21º10’E) was used to develop the radiocarbon time series. The preparation of samples and 14C counting using a low-background proportional counter was described by Povinec (Reference Povinec1972, Reference Povinec1977, Reference Povinec1978). 14C in Georgian wines (ca. 38ºN, 45ºE), obtained from the Tbilisi Museum of Wines, was analyzed using liquid scintillation decay counting (Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a, Reference Burchuladze, Chudý, Eristavy, Pagava, Povinec, Šivo and Togonidze1989). The typical standard deviations of the single measurements were 3‰ (at 1σ). Several cross-check measurements between the Bratislava and Tbilisi laboratories showed results that were indistinguishable within the statistical uncertainties.
The development of the AMS radiocarbon dating method has greatly benefited 14C measurement in tree-ring studies thanks to decreasing the sample size below 1 mg of carbon and reducing the uncertainty of measurements down to 1.5‰ (e.g., Povinec et al. Reference Povinec, Litherland and von Reden2009; Miyake et al. Reference Miyake, Nagaya, Masuda and Nakamura2012; Mekhaldi et al. Reference Mekhaldi, Muscheler, Adolphi, Aldahan, Beer, McConnell, Possnert, Sigl, Svensson, Synal, Welten and Woodruff2015; Jull et al. Reference Jull, Panyushkina, Miyake, Masuda, Nakamura, Mitsutani, Lange, Cruz, Baisan, Janovics, Varga and Molnár2018; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021).
Radiocarbon results are presented as Δ14C values that are calculated relative to the 14C oxalic acid standard of NIST (National Institute of Standards and Technology, Gaithersburg, USA), following the Stuiver and Polach (Reference Stuiver and Polach1977) approach.
RESULTS AND DISCUSSION
Radiocarbon Variations in Tree Rings and Wines (1902–1954)
We shall compare 14C results obtained in tree-ring and wine samples collected in the first half of the 20th century that were measured by radiometric (Damon et al. Reference Damon, Long and Wallick1973a, Reference Damon, Long and Wallick1973b; Povinec Reference Povinec1977, Reference Povinec1987; Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a, Reference Burchuladze, Pagava, Povinec and Usačev1980b; Stuiver Reference Stuiver1980, Reference Stuiver1982; Stuiver and Quay Reference Stuiver and Quay1980; Attolini et al. Reference Attolini, Galli, Nanni and Povinec1989; Stuiver and Braziunas Reference Stuiver and Braziunas1993) and AMS techniques (Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021). As the GCRs are modulated by the Sun activity during the 11-year SC, it has been challenging to search for 11-year 14C cycles in the biosphere. It was believed that the 11-year 14C cycle would be difficult to observe, as short-term changes in radiocarbon production rates in the atmosphere will be too small to be observed in the biosphere due to a large reservoir buffer effect (Damon et al. Reference Damon, Long and Wallick1973a, Reference Damon, Long and Wallick1973b). However, radiocarbon measurements in tree rings and wines made in the 1970s showed that although the 11-year 14C cycle is quite small, it is observable (Damon et al. Reference Damon, Long and Wallick1973a, Reference Damon, Long and Wallick1973b; Povinec Reference Povinec1977; Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a, Reference Burchuladze, Pagava, Povinec and Usačev1980b; Stuiver Reference Stuiver1980, Reference Stuiver1982; Stuiver and Quay Reference Stuiver and Quay1980). The Δ14C levels in tree rings and wines corresponded well to the levels in the atmosphere at the time the tree rings (and grapes) were growing (Povinec et al. Reference Povinec, Chudý and Šivo1986; Burchuladze et al. Reference Burchuladze, Chudý, Eristavy, Pagava, Povinec, Šivo and Togonidze1989; Svetlik et al. Reference Svetlik, Povinec, Molnár, Vána, Sivo and Bujtás2010; Kontuľ et al. Reference Kontuľ, Ješkovský, Kaizer, Šivo, Richtáriková, Povinec, Čech, Steier and Golser2017, Reference Kontuľ, Svetlik, Povinec, Brabcová and Molnár2020). The radiometric 14C results were later confirmed with AMS measurements carried out with much better precision (Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021).
Unfortunately, after 1950, the natural 11-year 14C cycle has been disrupted by anthropogenic 14C produced during atmospheric nuclear weapons tests, which peaked in the biosphere in 1964 (ca. 100% above the natural level) (see, e.g., Burchuladze et al. Reference Burchuladze, Chudý, Eristavy, Pagava, Povinec, Šivo and Togonidze1989).
Another anthropogenic effect observed during the 19th and 20th century in the 14C levels in the atmosphere and biosphere was due to growing amounts of CO2 from fossil fuels that do not contain radiocarbon (Suess effect) (Suess Reference Suess1955). Fortunately, as we shall discuss later, we can detrend the Suess effect from the annual Δ14C data, and then compare the results with the sunspot number record.
Figure 2 shows radiocarbon variations in Bratislava tree rings (Povinec Reference Povinec1977, Reference Povinec1987) and Georgian wines (Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a) during five successive 11-year SCs for 1902–1954. We can clearly see a decreasing trend in Δ14C levels due to the Suess effect, as well as the anticorrelation in radiocarbon variations with 11-year SCs (represented by sunspot numbers). The non-statistical scatter of the results and the cyclic trend through the data indicate a possible dependence of Δ14C data on the 11-year SC.
Figure 2 The Δ14C record in tree rings (Povinec Reference Povinec1977, Reference Povinec1987) and wines (Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a) for the 20th century. The International sunspot number record is also shown (Royal Observatory of Belgium, http:/side-be/silso).
We applied several spectral analysis techniques to evaluate 14C variations in tree-ring and wine samples. The spectral analysis of Bratislava tree-ring Δ14C data for the 17th and 18th solar cycles showed an average 3‰ amplitude, an average 10-year periodicity and an average 3-year time lag, i.e., a delay in the Δ14C record following the sunspot data (Table 1) (Povinec Reference Povinec1977; Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a). The delay between the Δ14C maxima and the solar activity minima by ca. 3 years is caused by a time lag from the production of 14C in the atmosphere to its uptake by the biosphere.
Table 1 Radiocarbon variation parameters for 11-year SCs in tree-ring and wine samples (1902–1954).
References:
1 Damon et al. Reference Damon, Long and Wallick1973a
2 Povinec Reference Povinec1977
3 Attolini et al. Reference Attolini, Galli, Nanni and Povinec1989
4 Stuiver and Braziunas Reference Stuiver and Braziunas1993 (14C data)
5 Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021 (14C data)
6 Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a
The Bratislava Δ14C tree-ring data presented in Figure 2 were also analyzed using a cyclogram method, which is different from conventional power spectrum analysis with integrated harmonic components over time (Attolini et al. Reference Attolini, Galli, Nanni and Povinec1989). The cyclogram method enables one to follow the amplitude and phase variations with time for any component of a given periodicity. The Δ14C phase and amplitude cyclograms were the most stretched for a period of 9 years, while the cross-cyclogram vectors showed the most regular behavior for a period of 10.5 years. The rmsFootnote 2 Δ14C amplitude computed through the cyclogram was 3‰ and the mean phase shift was 3 years (Table 1), confirming the anticorrelation dependence of Δ14C on sunspot numbers (Attolini et al. Reference Attolini, Galli, Nanni and Povinec1989).
Harmonic analysis of Tbilisi wine samples for four 11-year SCs (1913–1954) of Δ14C (t) data rows for wines and sunspot numbers (SN(t)) showed Δ14C amplitudes of ca. 4‰ for different solar cycles (the average amplitude is 4.3 ± 0.2‰) with average periodicity of 11 years (Table 1). The time shift between the sunspot maxima and the delaying Δ14C minima was 3–7 years (depending on the solar cycle), with an average value of 4 years. The correlation coefficients were between –0.6 and –0.9 (Burchuladze et al. Reference Burchuladze, Pagava, Povinec and Usačev1980b). This anticorrelation between the decreasing solar activity and increasing 14C production in the atmosphere is related to the modulation of GCR by the solar activity, as during the solar minimum there is an increased flux of GCR in the heliosphere and therefore also in the atmosphere of the Earth.
Furthermore, we applied spectral analysis, correlation analysis, and harmonic analysis to compare the Δ14C results for wines (Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a) and tree-ring samples from the NW Pacific coast of USA (Stuiver and Braziunas Reference Stuiver and Braziunas1993) using the same analysis technique. The results of the spectral analysis showed that for both types of samples, the autocorrelation functions had a quasi-periodic behavior with an 11-year period (at the 95% confidence interval). The autocorrelation functions had argument values UW = 10 ±1 yr, Utrees = 11 ± 1 yr, Uwines = 11 ± 1 yr. The harmonic analysis showed the amplitude of Δ14C variations in wines and tree rings for different solar cycles of 4.0–4.5‰ (±0.8‰) and 0.7–2.0‰ (±0.5‰), respectively. The average amplitude for wines and tree rings was 4.3 ± 0.2 and 1.5 ± 0.5‰, respectively. The average lag between SN(t) and Δ14C(t) rows for wines and tree rings was 4 and 2 years, respectively. The basic power of SN(t) and Δ14C(t) spectra was concentrated for a period of 11 years.
Recently Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) published a comprehensive data set of annual 14C measurements in tree rings that covered the last millennium. We performed a harmonic analysis of the Δ14C data for the period of 1902–1933 (tree rings from England) which showed the Δ14C amplitudes from 0.8 to 1.8‰ (the average value of 1.2 ± 0.5‰), the average periodicity of 12 ± 1 years, and the average time lag of 3 years (Table 1).
The results of the discussed radiocarbon studies during the five 11-year SCs for the period of 1902–1954 are presented in Table 1. We can see that a reasonable agreement within uncertainties has been obtained between samples of different origin (tree rings from Slovakia, England and NW Pacific coast, and wines from Georgia). The average Δ14C amplitude derived from tree-ring samples is 2.2 ± 0.7‰, the average periodicity is 11 ± 1 years, and the average time lag between the sunspot numbers and Δ14C records is 3 ± 1 years. Within uncertainties, these results are the same as those obtained for the wine samples.
Radiocarbon Variations in Tree-Ring Samples (1800–1950)
We performed a harmonic analysis of the Δ14C tree-ring results for the period of 1783–1954 using the data of Stuiver and Braziunas (Reference Stuiver and Braziunas1993) and Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) obtained by gas counting and AMS, respectively. The purpose of the study was to compare the 11-year 14C SCs measured in wood samples from the NW Pacific coast of the USA (Stuiver and Braziunas, Reference Stuiver and Braziunas1993) and from England (Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021). A comparison of Δ14C tree-ring data presented in Figure 3 shows good agreement between both sets of data, although there are some differences. Generally, the Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) data after 1840 are ca. 5‰ lower than the Stuiver and Braziunas (Reference Stuiver and Braziunas1993) data, which could be due to the regional Suess effect. The NW Pacific is expected (also due to ocean influence) to be less affected by fossil fuel CO2 contribution than the England location. It is interesting to note that the decline in 14C levels for the years 1833–1880 and 1905–1952 had the same trend (–0.11‰ yr–1 and –0.44‰ yr–1 for the NW Pacific coast, and –0.16‰ yr–1 and –0.37‰ yr–1 for the western Europe (England), respectively), which is in reasonable agreement with modelling predictions of the Suess effect (Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021).
Figure 3 Δ14C tree-ring results (with 2σ error bars) for the period of 1783–1954 obtained by Stuiver and Braziunas (Reference Stuiver and Braziunas1993) for the NW Pacific USA (gas counting) and for 1783–1933 obtained by Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) for England (AMS). The International sunspot number record is also shown (Royal Observatory of Belgium, http:/side-be/silso).
The results of harmonic analysis of tree-ring samples covering twelve 11-year SCs (1798–1933) of Δ14C data and sunspot numbers are presented in Table 2. The average Δ14C amplitudes derived from the data from Stuiver and Braziunas (Reference Stuiver and Braziunas1993) and Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) were very similar (1.2 ± 0.5‰ and 1.1 ± 0.5‰, respectively). The periodocities were in the range of 8–14 and 9–15 yr with mean values of 10.7 ± 2.1 and 11.5 ± 1.7 yr, respectively. The average time lags between the sunspot maxima and delaying Δ14C minima were in the range of 1–7 and 2–9 yr with the average values of 4 ± 2 and 5 ± 2 yr), respectively. Therefore, the Δ14C results obtained for the NW Pacific coast (measured by gas counting) and for England (measured by AMS) trees were the same within uncertainties.
Table 2 Radiocarbon variation parameters for 11-year solar cycles #5–16 in tree-ring samples (1798–1933).
1 Stuiver and Braziunas Reference Stuiver and Braziunas1993
3 Average Δ14C amplitude
4 Average periodicity
5 Average time lag
Figure 3 also shows elevated Δ14C levels during the Dalton minimum (1797–1823, covering two 11-year SCs with peak sunspot numbers around 75, Figure 1), which in the literature, similarly to the Maunder minimum, is also called Grand solar minimum (e.g., Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021). Recently, several papers discussed parameters of the newly established Gleissberg minimum (1878–1933), with peak sunspot numbers in the range of 100–175 during cycle peaks, Figure 1 (e.g., Petrovay Reference Petrovay2010; Komitov and Kaftan Reference Komitov and Kaftan2013; Gao Reference Gao2016). Using the Gleissberg sunspot number filter (Gleissberg (Reference Gleissberg1967), the length of this minimum was estimated to cover five 11-year SCs (Petrovay Reference Petrovay2010; Gao Reference Gao2016). Although this minimum was heavily affected by the Suess effect, especially after 1900, the Δ14C data presented in Figure 3 clearly demonstrate its presence. For a better evaluation of its parameters, Figure 4 shows the detrended course of Δ14C for the period of 1835–1951, after the corresponding corrections of the Suess effect. The detrended Δ14C excess during the Gleissberg minimum in the Stuiver and Braziunas (Reference Stuiver and Braziunas1993) and Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) data is ca. 7‰ (similarly as in the case of the Dalton minimum). The duration of the minimum in Δ14C data is, however, shorter: two 11-year SCs instead of five SCs (1878–1933) predicted by Petrovay (Reference Petrovay2010). This could be due to the fact that Δ14C data covering the Gleissberg minimum were heavily affected by the Suess effect, therefore, more precise 14C measurements would be required to improve its parameters.
Figure 4 Detrended Δ14C course for the period of 1835–1951 (extracted from Figure 3 after Suess effect corrections), confirming the existence of the Gleissberg minimum. To make trends and variations easier to see, error bars are not shown (see Figure 3). The International sunspot number record is also shown (Royal Observatory of Belgium, http:/side-be/silso).
In the case of the Dalton minimum, the Δ14C data cover three SCs as predicted (Petrovay Reference Petrovay2010). Actually, the normalized 14C production rates calculated for the Dalton minimum by Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) may also suggest three SCs, although the generally accepted duration of the Dalton minimum covers only two cycles (1797–1823). Compared to the Maunder Grand solar minimum, the Dalton minimum could only be called a “semi-grand minimum”, and the Gleissberg minimum should be an even weaker minimum (Petrovay Reference Petrovay2010). Although it is longer in duration than the Dalton minimum, the number of sunspots is greater (> 100 in each SC). As the Δ14C data presented in Figures 3 and 4 clearly demonstrate the existence of the Gleissberg minimum, we should take it into account when discussing impacts of solar activity on the biosphere.
Solar Proton Events
Solar proton events (SPEs), represented by emissions of energetic protons from the Sun, could influence the course of 11-year 14C SCs in the atmosphere and biosphere (Lingenfelter and Ramaty Reference Lingenfelter, Ramaty and Olsson1970; Lingenfelter and Hudson Reference Lingenfelter, Hudson, Pepin, Eddy and Merrill1980), since the anticorrelation between the 14C variations and the solar activity could be affected, especially if total proton fluences are large (>1010 p cm–2), and their rigidity is also high (>100 MV) (Povinec Reference Povinec1975; Konstantinov et al. Reference Konstantinov, Levchenko, Kocharov, Mikheva, Chechini, Galli, Nani, Povinec, Rugiero and Salomoni1992).
In the 1970s we therefore searched for possible elevated Δ14C levels in annual tree-ring data that could be associated with SPEs observed during the 1940s (Figure 5) (Povinec Reference Povinec1975, Reference Povinec1977). As the registered SPEs were too weak (both in proton fluencies and proton rigidities), the effects in 14C production could be unobservable. However, these were the only registered SPEs that we could compare with our Δ14C dataset.
Figure 5 Total proton fluences during solar proton events with different proton rigidities, observed up to 1962. (Sources: Dorman and Mirosnicenko Reference Dorman and Mirosnicenko1968; Shea and Smart 2020; Usoskin et al. Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020.)
The first interesting observation was that the large SPEs were usually observed before or after the maximum of the 11-year SC (Figure 5). This was also the case of the 23 February 1956 SPE that was observed close to the maximum of the 11-year SC. Unfortunately, as we already mentioned, natural 14C levels during 1956 were already affected by global fallout from atmospheric nuclear bomb tests (Burchuladze et al. Reference Burchuladze, Chudý, Eristavy, Pagava, Povinec, Šivo and Togonidze1989), and therefore we could compare Δ14C data only with SPEs that were observed in the first half of the 20th century.
The first documented solar bursts were observed by Forbush (Reference Forbush1946) on February 28 and March 7, 1942. These SPEs were observed, however, two years before the solar minimum, i.e., at a time when the effects from these flares would be superimposed on the high GCR background observed during the 11-year SC. The second important point is that due to Forbush decreases (Forbush Reference Forbush1946), which represent rapid decreases in the GCR intensity following a coronal mass ejection when the magnetic field of the plasma solar wind sweeps the GCR away from the Earth, the 14C production by the GCR should also decrease. Although it is expected that the SPEs of 1942 were much weaker than the 1956 SPE, it was interesting to search in the experimental Δ14C data presented in Figure 2 for possible effects. We see in the tree-ring record two peaks with elevated Δ14C values of 5‰ above the GCR background, which should be associated with the increased GCR background after the preceding solar activity minimum. Interestingly, the Δ14C data of wine samples presented in Figure 2 show elevated Δ14C values in 1943 (two measurements of –22.5 and –24.3‰, Burchuladze et al. Reference Burchuladze, Pagava, Povinec, Togonidze and Usačev1980a). However, the expected decline of Δ14C levels after this maximum is missing, therefore this peak cannot be associated with SPE. The Stuiver and Braziunas (Reference Stuiver and Braziunas1993) Δ14C data (Figure 3 and 4) show a minimum value for 1943, and then rising values until 1947, i.e., this Δ14C record cannot be associated with the SPE of 1942.
The second SPE was observed on 25 July 1946, which was a single event comparable to each of the two 1942 events, also accompanied by the Forbush decrease (Forbush, Reference Forbush1946). This flare originated before the solar maximum, and therefore its possible impact could be better observable because of a lower GCR background. The Δ14C wine data show during 1947 and 1948 an elevation of 5‰ above the GCR background (Figure 2), however, the Δ14C tree-ring data (Figure 2) and Stuiver and Braziunas (Reference Stuiver and Braziunas1993) data (Figure 3 and 4) do not show any increased values above the GCR background.
The SPE observed on 19 November 1949 (Adams and Braddick Reference Adams and Braddick1951) was the strongest (both in proton fluences and rigidities) when compared with the 1942 and 1946 events. The wine Δ14C record (Figure 2) shows a minimum value for 1951 (0‰) and then an increase in 1952 (2.5‰), i.e., the data do not follow the expectations. It is likely that the increased Δ14C value in 1952 was already influenced by the bomb effect, similarly as the Stuiver and Braziunas (Reference Stuiver and Braziunas1993) data (Figure 3). After 1955 (during the Modern solar activity maximum, Figure 1), many large solar proton flares occurred, however, because of the bomb effect, possible SPE impacts on 14C production cannot be investigated (Burchuladze et al. Reference Burchuladze, Chudý, Eristavy, Pagava, Povinec, Šivo and Togonidze1989).
The origin of SPEs that originated before 1940 were not well documented. We should also mention that large solar superstorms producing large geomagnetic disturbances were not always accompanied by large SPEs. Although during the SPEs of 1942, 1946 and 1949 solar particles were registered together with radio disturbances, during some other solar super storms, such as the one observed in February 1958, the SPE was not registered. The special case is the Carrington event observed in September 1859, which was recognized as probably the biggest solar superstorm observed till now (Green and Boardsen Reference Green and Boardsen2006). However, the Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) and Stuiver and Braziunas (Reference Stuiver and Braziunas1993) Δ14C data (Figures 3 and 4) do not show any peak that could be associated with this event (actually, they show decreasing Δ14C values after 1859).
Recent AMS analysis of annual tree rings revealed sharp increases in Δ14C of 5–12‰ at AD 775 and 993 (Miyake et al. Reference Miyake, Nagaya, Masuda and Nakamura2012, Reference Miyake, Masuda and Nakamura2013). Usoskin et al. (Reference Usoskin, Kromer, Ludlow, Beer, Friedrich, Kovaltsov, Solanki and Wacker2013), using more precise calculations of 14C production rates, suggested the AD 775 event as possible consequence of a giant SPE. Furthermore, Usoskin et al. (Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020) compared these SPEs with the one observed on 23 February 1956, which was the strongest directly observed SPE to date (109 p m–2) with a very hard proton spectrum (rigidities >300 MV, Figure 5). The calculated Δ14C signal from this SPE (due to attenuation of high-frequency signals by the carbon cycle) would be, however, only 0.2‰. As this value is comparable with the present uncertainty of 14C analysis by AMS (1.5‰; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021), the effect of the SPE of 23 February 1956 (and similar SPEs) would be undetectable. The SPEs observed by Miyake et al. (Reference Miyake, Nagaya, Masuda and Nakamura2012, Reference Miyake, Masuda and Nakamura2013), and also other events observed by other groups, and later confirmed by 10Be and 36Cl records in ice cores as well (Mekhaldi et al. Reference Mekhaldi, Muscheler, Adolphi, Aldahan, Beer, McConnell, Possnert, Sigl, Svensson, Synal, Welten and Woodruff2015, Reference Mekhaldi, Adolphi, Herbs and Muscheler2021; Park et al. Reference Park, Southon, Fahrni, Creasman and Mewaldt2017; Wang et al. Reference Wang, Yu, Zou, Dai and Cheng2017; Jull et al. Reference Jull, Panyushkina, Miyake, Masuda, Nakamura, Mitsutani, Lange, Cruz, Baisan, Janovics, Varga and Molnár2018; O’Hare et al. Reference O’Hare, Mekhaldi, Adolphi, Raisbeck, Aldahan and Anderberg2019; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) should therefore be stronger than the 1956 SPE by about a factor of at least 25 (Usoskin et al. Reference Usoskin, Kromer, Ludlow, Beer, Friedrich, Kovaltsov, Solanki and Wacker2013, Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020).
Stuiver and Braziunas (Reference Stuiver and Braziunas1993) and Brehm et al. (Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) data (Figures 3 and 4) do not show clear peaks that could be associated with SPEs during 1800–1950. The disadvantage of 14C for searching for SPEs is that small Δ14C peaks would be difficult to observe in tree-ring samples because the carbon cycle damps high-frequency atmospheric signals by about a factor of 300 (Usoskin et al. Reference Usoskin, Koldobskiy, Kovaltsov, Rozanov, Sukhodolov, Mishev and Mironova2020). The 10Be signal, because of simpler atmospheric transport and its storage in ice, would have a better sensitivity than 14C (although it is a product of high-energy spallation reactions). It is expected, however, that 36Cl in ice cores could be an even better isotope tracer to identify SPEs, having advantages of 10Be and 14C (it is a low-energy product), and because high-sensitivity analysis of 36Cl can be achieved with AMS (Mekhaldi et al. Reference Mekhaldi, Adolphi, Herbs and Muscheler2021).
The previous skepticism (which was common during the 1970s–1980s) about the possibility to observe short-term 14C variations in tree rings has been dispelled, and many new well-defined Δ14C increases have been investigated in tree-ring data covering the last millennia thanks to the availability of tree ring and ice core samples, and the high-precision AMS technique. As the Δ14C levels in the atmosphere are currently close to zero due to decline in the bomb effect (Kontuľ et al. Reference Kontuľ, Povinec, Richtáriková, Svetlik and Šivo2022; Povinec et al. Reference Povinec, Kontuľ, Ješkovský, Kaizer, Richtáriková, Šivo and Zeman2024a, Reference Povinec, Kontuľ, Ješkovský, Kaizer, Kvasniak, Pánik and Zeman2024b), this opens a new window to study recent impacts of SPEs on 14C production in the atmosphere and biosphere using much better datasets on the characteristics of SPEs, mostly obtained from satellites.
The SPEs have been of major concern to society with regard to possible negative impacts on electronic infrastructure as well as for long-term space missions. Therefore, better understanding of the frequency of their occurrence in past millennia, as well as of the emitted proton fluxes, will help in prediction and preparation for their future impacts.
CONCLUSIONS
Cosmogenic radiocarbon has proved to be an important tracer for studying solar activity impacts on the biosphere, including 11-year solar activity cycles, as well as impacts of solar protons on 14C production in the atmosphere and its transport to the biosphere. Observations of thirteen 11-year 14C solar activity cycles (1798–1944) in annual tree-ring samples from the NW Pacific coast of USA (Stuiver and Braziunas Reference Stuiver and Braziunas1993) and from England (Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021) discussed in this review showed the average Δ14C amplitude of 1.3 ± 0.3‰, the average periodicity of 11 ± 1 yr, and the average time lag between the sunspot numbers and Δ14C records of 3 ± 1 yr. The Δ14C data after 1840 for England were ca. 5‰ lower than the data for the NW Pacific coast, which could be due to the regional Suess effect. The decline in 14C levels for the years 1833–1880 and 1905–1953 had the same trend at the two sites (–0.11‰/year and –0.44‰/year for the NW Pacific, and –0.16‰/year and –0.37‰/year for Europe, respectively).
The Δ14C tree-ring data discussed for the NW Pacific (USA) and England clearly demonstrated the presence of the Dalton minimum (1797–1823), as well as the existence of the new Gleissberg minimum (1878–1933). The observed Δ14C excess during the Dalton and Gleissberg minima was ca 7‰. Although the Gleissberg minimum was heavily affected by the Suess effect (especially after 1900), the Δ14C data (for the first time) clearly demonstrate its existence, and therefore the Gleissberg minimum should be taken into account when discussing impacts of solar activity on the biosphere.
The discussed Δ14C tree-ring data for the period of 1800–1950 did not show any elevated levels above the GCR background, which could be associated with solar proton events, because they were ca. at least an order of magnitude smaller than the giant events producing sharp increases in annual tree-ring Δ14C data during the first millennium AD (Miyake et al. Reference Miyake, Nagaya, Masuda and Nakamura2012, Reference Miyake, Masuda and Nakamura2013; Brehm et al. Reference Brehm, Bayliss, Christl, Synal, Adolphi, Beer, Kromer, Muscheler, Solanki, Usoskin, Bleicher, Bollhalder, Tyers and Wacker2021). There have been many discussions about the 14C records associated in the past with giant SPEs, also accompanied with 10Be and 36Cl records in ice cores, which have been of great concern to society with regard to possible negative impacts on electronic infrastructure, as well as for long-term space missions.
Investigations carried out during the last 50 years confirmed that solar phenomena (solar activity cycles and solar proton events) are an important and practical part of radiocarbon research aimed at better understanding the Sun’s behavior and its impact on the Earth’s environment.
ACKNOWLEDGEMENTS
A support under the Operational Program Integrated Infrastructure for the project “Advancing University Capacity and Competence in Research, Development and Innovation (ACCORD)”, (ITMS2014+:313021X329, co-financed by the European Regional Development Fund), and the support from the Slovak Research and Development Agency (projects APV-15-0576 and APVV-21-0377), and the Slovak Science and Grant Agency (project VEGA-1/0625/21) is highly acknowledged.
