DENDROCHRONOLOGY AND RADIOCARBON DATING

Abstract

Both dendrochronology and radiocarbon (¹⁴C) dating have their roots back in the early to mid-1900s. Although they were independently developed, they began to intertwine in the 1950s when the founder of dendrochronology, A. E. Douglass, provided dated wood samples for Willard Libby to test his emerging ¹⁴C methods. Since this early connection, absolutely dated tree-rings have been key to calibration of the Holocene portion of the ¹⁴C timescale. In turn, ¹⁴C dating of non-calendar-dated tree-rings has served to place those samples more precisely in time, advance development of long tree-ring chronologies, and bring higher resolution to earlier portions of the ¹⁴C calibration curve. Together these methods continue to shape and improve chronological frameworks across the globe, answering questions in archaeology, history, paleoclimatology, geochronology, and ocean, atmosphere, and solar sciences.

HISTORICAL DEVELOPMENT OF DENDROCHRONOLOGY

Growth rings are a prominent internal feature in the stems of woody plants. We may casually notice them in a wide variety of places in our human-built environment such as wooden doors and furniture, or perhaps on an outdoor hike through some natural area where stumps provide a window into the interior of remnant trees. Historical accounts indicate awareness of tree-rings by ancient scholars, such as Theophrastus in 322 BCE, who contemplated their annual formation, and Leonardo Da Vinci in the late 1400s CE, who considered weather as influencing their formation (Speer 2010). At least a dozen other scientists in Europe and North America in the 1700s and 1800s made observations or measurements on tree-rings related to climate, weather events, ecology, and human activities (Speer 2010). However, it was not until investigations in the early 1900s by Andrew Ellicott Douglass, considered to be the “founder of modern dendrochronology” (Figure 1A), that a scientific field of study began to emerge.

Figure 1 Early tree-ring and radiocarbon interactions: (A) Andrew Ellicott Douglass with his Cycloscope (1935), designed with the over-riding aim of discovering predictable cycles of solar activity in patterns of tree-ring growth, with emphasis on the 11-year solar cycle and its impact on climate (see Webb 1993 for further details). (B) Centennial Stump from California’s Sierra Nevada used by Willard Libby in the first radiocarbon calibration, the “Curve of Knowns” featured in his Nobel prize speech in 1960. Note the large notches along the top edge of the sample created by the radiocarbon sampling. (C) Sample from “Broken Flute Cave,” an Ancestral Puebloan cliff dwelling in the Prayer Rock district of the Navajo Nation in Arizona, also used in Libby’s Curve of Knowns. (All images reproduced with permission from The Laboratory of Tree-Ring Research, University of Arizona.)

Douglass was an astronomer, originally based at the Lowell Observatory in Flagstaff, Arizona, near the turn of the 20th century. Among his research interests were the Sun and solar activity manifested in sunspots. Reasoning that solar activity could influence Earth’s weather, he investigated tree-rings as indicators of year-to-year climate variability that could then be related to sunspots. During his time in Flagstaff, he noticed the growth rings of ponderosa pine, often on the ends of logs in lumber yards, showing variability in their ring width. He found that inter-annual variability of ring-size contributed to distinctive ring patterns (sequences) that were present in trees over a large region and appeared to be related to year-to-year precipitation variability (Dean 1997). He took a faculty position at the University of Arizona in 1906 to advance his scholarly interests in astronomy while continuing to work with tree-rings, particularly related to archaeological dating in the U.S. Southwest (Douglass 1929). He launched the journal Tree-Ring Bulletin (now Tree-Ring Research) in 1934, and in 1937 he established a new department known as the Laboratory of Tree-Ring Research (LTRR) at the University of Arizona.

Douglass systematically developed field/laboratory methods, principles, and terminology of dendrochronology, which remain in use today. In particular, his principle of “crossdating” (matching of patterns of ring widths among trees to establish absolute dates, see Figure 2) is the linchpin to successful tree-ring studies (Dean 1997; Leavitt et al. 2012). Rather than just “counting rings,” crossdating is necessary to identify the temporal correspondence between rings with certainty. Furthermore, crossdating identifies “missing rings” and extra rings that are not annual (so-called “false rings”). The method not only applies to matching patterns in contemporaneous living trees, but also to building extended tree-ring-width chronologies. Such long chronologies are produced by matching tree-rings of living trees to patterns in older wood samples from stumps and wood lying on the ground surface, wood in various historic and prehistoric constructions, and logs preserved in waterlogged sediments (see Figure 2). On one hand this crossdating is made possible by the imprint of weather influence on tree growth in a region, and on the other, tree-rings can offer a suitable proxy for climate reconstructions because of this influence. These aspects of dendrochronology are profound and have tremendously and increasingly contributed to historical and archaeological dating, and to a wide variety of investigations involving climate, hydrology, ecology, and geomorphology. Douglass’s student and eventual colleague, Edmund Schulman (Figure 3A), was the first to begin developing long tree-ring chronologies from bristlecone pine (Pinus longaeva, Pinus aristata) in California’s White Mountains, realizing the great longevity of these trees in that high-elevation mountain environment. Following Schulman’s death, C. Wesley (Wes) Ferguson (Figure 3B) was the central bristlecone investigator for almost three decades, arranging expeditions to the White Mountains to extend the chronology. Stimulated by Douglass’s successes, Bruno Huber began tree-ring work in Germany in the 1940s, eventually applying the technique to historic and prehistoric buildings (Becker 1992), much as Douglass had successfully done with the archaeological ruins in the American Southwest. Tree-rings have now been studied on all continents (Zhao et al. 2019), even from ancient and petrified wood in Antarctica (e.g., Taylor and Ryberg 2007). Calendar dated ring-width records spanning hundreds and thousands of years have been constructed for multiple regions and the long Northern Hemisphere tree-ring chronologies are paralleled by Southern Hemisphere records.

Figure 2 Illustration of the tree-ring crossdating method. Ring-width patterns from areas with common climate forcing show matching patterns of growth, which can be overlapped from successively older samples to develop an extended “master” chronology of ring-width variability. Here the oldest rings in the living tree are shown to match the pattern of growth in the outer rings of a standing dead tree, and in turn the inner rings of the standing dead tree match the outer rings of a beam used in construction of a building. (Image redrawn by C. Pearson based on a composite of images reproduced with permission of LTRR and P. I. Kuniholm, University of Arizona.)

Figure 3 The early long tree-ring chronologies. (A) Edmund Schulman with bristlecone sample #4779 in 1957; (B) Charles “Wes” Ferguson measuring a bristlecone pine dating 2963 BCE to 278 CE; (C) 20-g sample (10 years) of bristlecone pine prepared at LTRR for requests from radiocarbon dating labs in 1964 and published by Suess (1967); (D) Mike Baillie working on the Irish oak chronology; (E) Irish bog oaks, Garry Bog, (inset) trees from Hillsborough Co. Down, Trinity College Dublin and Coagh Co. Tyrone showing a matching pattern of wide rings, the last being 1580 CE; (F,G) Bernd Becker extracting trees for the European oak and pine chronology. (Images A–C reproduced with permission from The Laboratory of Tree-Ring Research, University of Arizona. D and E provided by M.G.L. Baillie. F & G provided by B. Kromer.)

The longest chronologies for the Northern Hemisphere are from Germany (>12,300 years, oak [Quercus petraea, Quercus robur] and Scots pine [Pinus sylvestris L.]), the European Alps (>9000 years, pine [Pinus cembra], larch [Larix decidua], spruce [Picea abies]), California (>8800 years, bristlecone pine [Pinus longaeva, Pinus aristata]), Finland (>7500 years, Scots pine [Pinus sylvestris L.]), Ireland (>7200 years, oak [Quercus petraea, Quercus robur]), and Siberia (>4000 years, Larch [Larix sibirica]) (see Becker 1992; Kromer 2009; Nicolussi 2009; Eronen 2002; Hantemirov and Shiyatov 2002; Friedrich et al. 2004; Pilcher et al. 1984; Brown and Baillie 2012; Salzer 2019). In the Southern Hemisphere, the longest securely dated sequences are from New Zealand (>4400 years, kauri [Agathis australis]) and Tasmania (>4000 years, Huon pine [Lagarostrobos franklinii C. J. Quinn]) (Boswijk et al. 2006, 2014; Cook et al. 2006). The diverse applications of these multi-functional, calendar-dated tree-ring data continue to enrich the histories of these regions in multiple ways but as remarked by Dean (1997), the primary applications of dendrochronology are for “the dating of past events and the reconstruction of past environmental conditions.” Similarly, tree-ring chronologies strike both of these chords with respect to radiocarbon (¹⁴C) applications in that they are used to infer the ¹⁴C concentrations in past atmospheres (thereby providing data for carbon cycle investigations and for solar/magnetic reconstruction) and to improve accuracy of dates obtained by radiocarbon methods.

IMPORTANCE OF TREE RINGS TO ¹⁴C DATING

Douglass established the LTRR just a decade or so before Willard Libby at the University of Chicago first developed the radiocarbon dating method and what, in addition to transforming understanding of past timelines, would also become a means to study the solar activity through time, the very motivation for much of Douglass’s research. So, since the very inception of these two remarkable disciplines, there has been a kind of symbiosis, with one technique feeding into the other and used in powerful combinations to advance numerous scientific frontiers. This relationship continues to advance and evolve into the present and will no doubt be a focus for future applications.

Calibration and Chronologies

The critical value of dendrochronology for radiocarbon dating can be distilled as follows: Anywhere around the world, where trees form annual growth rings that can be calendar-dated using the techniques of dendrochronology, each tree-ring offers a dated sample of atmospheric radiocarbon. Although there are some caveats to this, such as for oak (Quercus sp.), which begins yearly growth using stored carbon from the previous growth season (Pilcher 1995), this basic premise underpins the primary value of dated tree-rings for radiocarbon calibration: Tree-rings offer an independently dated, exact measure of changing radiocarbon levels through time. The first clear demonstrated use of this came in the early development of the radiocarbon dating method (Arnold and Libby 1949) and Libby’s “Curve of Knowns” (Libby 1961), featured in his Nobel Prize lecture in 1960, which included measurements of dendrochronologically dated tree-rings from Centennial stump and Broken Flute cave (Figure 1B,C) beginning the long history of LTRR providing tree-ring samples to internal and external ¹⁴C researchers (Leavitt and Bannister 2009). Work to test and refine the radiocarbon dating method continued in diverse ways using tree-rings from many growth locations between 1950 and 1970. Spruce (Picea sp.) from Alaska, white pine (Pinus strobus) and incense cedar (Calocedrus decurrens) from North America and Cedrela sp. from the Peruvian Amazon were used by Suess (1955) to compare radiocarbon concentrations in modern wood. Then de Vries (1958, 1959) using North American Douglas-fir (Pseudotsuga menziesii) and German oak, detected 1–2% fluctuations in ¹⁴C activity along with the decline in ¹⁴C activity over the most recent 100 years related to ¹⁴C dilution by carbon dioxide from combustion of ¹⁴C-free fossil fuels, which Suess (1955) had previously identified (known as the “Suess effect”). Suess (1965) then went on to verify de Vries’ original observations of ¹⁴C wiggles (the “de Vries effect”) using dendrochronologically dated North American sequoia (Sequoiadendron giganteum), Douglas-fir and ponderosa pine (Pinus ponderosa). Stuiver (1965) also independently verified the de Vries wiggles at high-resolution using North American Douglas-fir, thus tree-rings were central to establishing the first links between these features and solar activity that modulates ¹⁴C production (Stuiver 1961; Damon and Peristykh 2000). This work also underpins the premise for radiocarbon wigglematch dating, where several radiocarbon dates sampled from a non-calendar dated (or floating) tree-ring sequence, spaced by exact ring counts, can be used to find a more precise fit of the wiggles (slopes and plateaus) against the radiocarbon calibration curve, so narrowing the possible date range for end point (outermost ring) of that sequence. Wiggles in the calibration data-sets also limit the dating precision and accuracy achievable for particular time periods.

Meanwhile dendrochronological work on North American bristlecone pine (Schulman 1954; Pritchett 2021) had led to these remarkable long-lived trees becoming the basis for the world’s first multi-millennial tree-ring time-series (Ferguson 1969). The value of this material for radiocarbon calibration and testing the method was also realized by Suess (1967), who demonstrated that for the period between 4100 BCE and 1500 BCE (extended back to 5200 BCE, Suess 1970), the radiocarbon content of 80 multi-year blocks of calendar-dated bristlecone pine tree-rings (see Figure 3C for a typical sample) was between 6 and 9% higher than what had been calculated for the same time period using the radiocarbon half-life. This correction and time-series was used in some of the first attempts to synchronize archaeological sequences that were critical to old world chronology (Clark 1978) and to secure one of the first-ever radiocarbon wiggle-matches (Ferguson et al. 1966) of non-calendar secured tree-ring sequences from the archaeological pile-dwelling site of Burgäschisee in Switzerland. In this study, ¹⁴C in the calendar-dated bristlecone pine sequence was used to secure ¹⁴C measurements from the undated Swiss samples, assigning them to the 38th century BCE with an error less than 40 years.

As dendrochronology became a globally utilized science, multiple tree-ring sequences were developed in different regions, opening up more and more opportunities for cross-pollination of radiocarbon and tree-ring research. At Queen’s University Belfast in Northern Ireland, the idea to produce a ¹⁴C timescale for Holocene peat and lake deposits in Northern Ireland and to answer questions raised by the bristlecone pine-based calibration curve (Suess 1970) was in fact a driving force behind the development of the long Irish tree-ring chronology that utilized oak trees preserved in Irish bog and terrestrial environments (Baillie 2009; Figure 3D,E). As work progressed on the construction of the Irish tree-ring chronology, in Germany, Burghart Schmidt and Bernd Becker, and separately Hubert Leuschner and Axel Delorme, were working on parallel chronologies, published back to 2000 BCE in 1982 (Becker and Schmidt 1982) and back to 4004 BCE (Leuschner and Delorme 1984) using oaks and pines retrieved from riverine flood deposits (Figure 3F,G). At the time of Becker and Schmidt’s (1982) publication, the Irish chronology was back to 5289 BCE (Baillie 2009) and an initial comparison of these two records across a 1000-year test period raised an issue. The two groups of tree-ring patterns matched with certainty, but the dates applied to each chronology did not synchronize, revealing a 71-year discrepancy in the dating applied to the two records. Which was correct? Here, radiocarbon “wiggle-matching” of sections of the German chronology to the bristlecone pine calibration curve, played a part in resolving the discrepancy, by revealing that the German record was off-set by ca. 70 years relative to the bristlecone pine, almost the exact discrepancy revealed by the dendrochronological comparison with the Irish oak record. As a result the dendrochronological issue (which it turned out was because dating for the pre-550 BCE portion of Becker and Schmidt’s chronology was based on a single site chronology produced and misdated by Ernst Hollstein [1980]) was swiftly identified and resolved (Pilcher et al. 1984; Baillie and Pilcher 1987), resulting in the joint publication of an agreed European oak chronology back to 5289 BCE (Pilcher et al. 1984). This happened to coincide exactly with the publication of the alternate German oak record, developed by Leuschner and Delorme (1984), which did not contain the 71-year error included in Becker and Schmidt’s record, and so provided a fully independent dendrochronological confirmation of the Pilcher et al. (1984) chronology. Following this, the German chronology was expanded back to 9420 BCE (see table 4.1 in Becker 1993; Becker and Schmidt 1990). Meanwhile, combined dendrochronological and radiocarbon effort continued towards the creation of a series of high-precision radiocarbon calibration curves for terrestrial Northern Hemisphere samples using ¹⁴C measurements from Irish and German oak, German pine and North American conifers including bristlecone pine. These were measured primarily at the Seattle, Belfast, Heidelberg, and Arizona radiocarbon laboratories and extended back to 13,300 cal BP by 1986 (Stuiver et al. 1986; Pearson et al. 1986). The Hohenheim oak and pine chronology, currently the world’s longest calendar-dated tree-ring sequence at 12,460 years (Friedrich et al. 2004), facilitated the further extension of this calendar-dated ¹⁴C record for calibration in more recent years (Stuiver et al. 1998; Reimer et al. 2009, 2013), in combination with the European Preboreal Pine and Swiss chronologies. This is now reinforced by the inclusion of single-year subfossil pine data from the French Alps (Reinig et al. 2020) to extend the latest iteration back to 14,226 ± 4 cal BP (Reimer et al. 2020).

Out of these previously described early cross-pollination events between dendrochronology and radiocarbon dating also came some of the first research into interhemispheric calibration (McCormac et al. 1998, 2002; Hogg et al. 2002), the study of regional ¹⁴C offsets (McCormac et al. 1995—see Reimer et al. 2020 for a detailed discussion of these) and explorations of the complexities of ¹⁴C mixing in the intertropical convergence zone (McCormac et al. 2004).

The Southern Hemisphere has a larger ocean surface area than the Northern Hemisphere (ca. 60% compared to ca. 40%, respectively) and greater wind velocities. The effects of this on ocean/atmosphere transfer mean that natural levels of ¹⁴C in the southern troposphere are usually lower than in the northern troposphere. This means that radiocarbon ages for terrestrial samples from the Southern Hemisphere can be expected to measure older than contemporary terrestrial samples in the Northern Hemisphere by ca. 40 years. Radiocarbon measurements on contemporary pairs of tree-ring samples from Northern Hemisphere (Quercus petraea) and Southern Hemisphere (Libocedrus bidwillii/Manoao colensoi) trees at the Belfast and Waikato radiocarbon laboratories (McCormac et al. 1998, 2002; Hogg et al. 2002) for the period AD 1850–950 however showed that a fixed “offset” should not be applied to Northern Hemisphere radiocarbon calibration data in order to use it to calibrate Southern Hemisphere radiocarbon measurements. Instead, the Southern Hemisphere required its own separate calibration curve. Tree-ring measurements from New Zealand, Chile and South Africa combined to form the basis of the first (and subsequent) curve iterations (SHCal02—McCormac et al. (2002); SHCal04—McCormac et al. (2004); SHCal13—Hogg et al. (2013)). In the most recent iteration, SHCal20 (Hogg et al. 2020) 14 new tree-ring data sets are added in the 2140–0, 3520–3453, 3608–3590, and 13,140–11,375 cal BP time ranges. The first three of these periods provide calendar-dated calibration reference material, and the latter brings annual resolution data via a floating tree-ring sequence. In between these periods, the current SH curve still lacks direct SH observations and relies on the corresponding sections of the Northern Hemisphere curve as a modeling basis. Fortunately, dendrochronology of New Zealand kauri in particular offers many future possibilities to fill in some of these gaps and extend measurements further back in time (Boswijk et al. 2014; Lorrey et al. 2018).

Other Synergies and Applications

The transition between the Northern and Southern Hemispheric atmospheres lies along the Intertropical Convergence Zone (ITCZ). Seasonal shifts in the ITCZ may entrain atmospheric CO₂ from the Northern and Southern Hemisphere to sites in this region within a given year and be impacted by a number of climatic forcings. Evidence for the migration of the ITCZ on multi-decadal to millennial time scales has been seen in a wide range of proxy records (e.g., Jacobel 2017), but tree-ring ¹⁴C records offer excellent potential for fine-scale geographic coverage and high-resolution reconstructions. Studies utilizing ¹⁴C distribution in Mexican tree-rings have shown the influence of the North American Monsoon on the position of the ITCZ (Beramendi-Orosco et al. 2018), and these data also offer much potential for correcting archaeological chronologies within this region where calibration uncertainties are high due to the atmospheric mixing of ¹⁴C (see Marsh et al. 2018). Another beneficial feedback between disciplines here is that typically the tropics have a scarcity of trees suitable for traditional dendrochronological methods, so a combination of radiocarbon and dendrochronology can be used to first confirm the presence of annual increments in trees (Santos et al. 2020) and then to go on to explore how these increments might be used in tracing the ITCZ and the forces governing its movement. In particular, use of the ¹⁴C bomb spike, caused by nuclear bomb testing in the 1960s CE, which almost doubled atmospheric ¹⁴C levels at this time before a steady decline due to ocean surface transfer processes and carbon cycling, produced an artificial tracer with which to date more recent organic material. Vieira et al. (2005), used the bomb spike to determine ages of tropical trees from the Brazilian Amazon without clear tree-ring structures, inferring tree growth rates and their consequences to carbon cycle modeling of forest biomass turnover. Radiocarbon dating and dendrochronology techniques have also been applied to see if African baobab trees, which have hollow inner cavities, could be dated using size–age relationships (Patrut et al. 2011) and if monumental olive trees in Spain (Camarero et al. 2021) and Israel (Bernabei 2015) were as old as purported by local populations. Olive trees are frequently encountered in key contexts in the ancient Mediterranean and Near East and have been the focus of much recent study (Ehrlich et al. 2018, 2021) to aid with use in the dating of archaeological contexts (e.g., Friedrich et al. 2006, 2014; Cherubini et al. 2013).

THE SOLAR AND GEOMAGNETIC CONNECTION

Past Solar Activity

The solar activity that so fascinated Douglass can now be explored directly through radiocarbon measurements from tree-rings and compared with documented visual observations of sunspots over the past four centuries, and instrumental data from recent decades (Solanki et al. 2013). The significance of ¹⁴C in tree-ring chronologies as a tool to reveal the temporal variability of solar activity over past millennia was realized early in the history of ¹⁴C studies and pioneered largely using coarser resolution ¹⁴C from multi-year blocks of tree-rings. However, as discussed by Stuiver (1961) and Suess (1965), and summarized by Lingenfelter (1963), the precision and temporal resolution of these ¹⁴C data were insufficient to discriminate production and carbon cycle (oceanic) causes of the observed ¹⁴C variations.

In the late 1970s a precision of 1.5 to 2‰ was reached in the ¹⁴C laboratory of Seattle, and Stuiver and Quay (1980) published a detailed and statistically convincing comparison of ¹⁴C in decadal tree-ring samples, observing sunspot numbers (averaged over 11-year cycles) from 1620 to 1880 CE. They also found strong ¹⁴C maxima corresponding to three grand solar minima, the Wolf, Spörer, and Maunder, with very low solar activity for several decades, and discussed the heliomagnetic modulation of galactic cosmic rays, leading to production changes of the cosmogenic isotopes ¹⁴C, ¹⁰Be, and ³⁶Cl. Using IntCal98, Solanki et al. (2004) reconstructed decadal sunspot numbers back to 11,000 years BP and determined the distribution of grand solar maxima and minima. Also based on IntCal98, Usoskin et al. (2007) analyzed the statistics of grand minima and maxima, concluding that the occurrence of these events is characterized by a stochastic/chaotic process, and that they represent special states of the solar dynamo.

Solar Cycles and ¹⁴C Production Spikes

Initial measurements of ¹⁴C in annual rings over intervals of 10–20 years were limited by the precision of the measurements. After early inconclusive efforts to identify the solar cycle in 20th century tree-rings, Damon et al. (1973) concluded that radiocarbon “measurement errors [did] not allow precise determination of the relatively small amplitude of the atmospheric radiocarbon variation due to the 11-yr solar cycle.” Statistically robust observations of the solar 11-year cycle in ¹⁴C had to wait until the early 1990s, when Stuiver and Braziunas (1993) measured annual ¹⁴C content in tree-rings for 1510–1954 CE. At those times, low-level gas counting detectors required 20–30 g of wood per ring and extended counting times of a week or more per sample. Hence, only a rather limited interval of 1510–1954 CE could be considered; as noted by Minze Stuiver (1993): “the counting time for producing the 440-yr single-year series reported here is identical to that needed for an 8800-yr bidecadal chronology.” Stuiver and Braziunas (1993) found a statistically significant correlation between sunspot numbers and annual ¹⁴C data during 1715–1948 CE.

Fortunately, the AMS technique is ideal for annual ¹⁴C measurements on wood samples, even single tree-rings, because it can work with a low sample mass (<50 mg) with counting times per sample of only a few hours, i.e., two orders of magnitude lower compared to radiometric low-level counting (LLC). Routine precision of <2‰ in AMS has only become a reality in the past decade (Wacker et al. 2020), leading to several series of centennial-long annual ¹⁴C data sets. This work has been further propelled by the discovery of an incredible increase of Δ¹⁴C of 15‰ in single tree-rings covering the years 774 to 775 CE by Miyake et al. (2012). McCormac et al. (2008) had previously noted this increase within a 10-yr resolution time series based on the Irish oaks, reporting a rapid enrichment of ¹⁴C between 765 and 775 CE, but the discovery that this change was in fact an abrupt event occurring between two single years opened a wide range of new research. Initially, a supernova, gamma-ray burst, or an extreme solar event were discussed as a possible cause for this phenomenon, but later comparison with ¹⁰Be (Usoskin et al. 2013) and additionally ³⁶Cl (Mekhaldi et al. 2015) led to an unambiguous identification of an extreme solar proton event (SPE) as the source. The realization that such events can be identified, and their magnitude and re-occurrence tracked via annual ¹⁴C has led to a global scale search of annual (and sub-annual) tree-ring ¹⁴C for more such events and other solar cycles, solar maxima and minima, not visible in the previous coarser-resolution records (Miyake et al. 2013, 2017; Güttler et al. 2015; Neuhäuser et al. 2015; Sukhodolov et al. 2017; Jull et al. 2018; Uusitalo et al. 2018; Scifo et al. 2019; Friedrich et al. 2019).

So far two more SPEs have been confirmed: 993 CE (Miyake et al. 2013) and 660 BCE (Park et al. 2017; O’Hare et al. 2019; Fahrni et al. 2020), and three more candidates have been tentatively identified. These are at 3372 BCE (Wang et al. 2017; but this was not confirmed by Jull et al. 2021), 1052 CE and 1279 CE (Brehm et al. 2021) but they need further confirmation with independent tree-ring chronologies. The global nature of the 775 and 993 CE events has meanwhile been confirmed in a study involving annual ¹⁴C series of 44 tree-ring chronologies of five continents (Büntgen et al. 2018), demonstrating the unique “signature year” character of SPEs to link chronologies worldwide. Replicate annual measurements in the respective century of the 1052 and 1279 CE events were published by Kudsk et al. (2019).

The longest and most precise annual ¹⁴C dataset so far has been created by Brehm et al. (2021) using 13 oak timbers from buildings in the UK and Switzerland (full details are given in the supplement of Brehm et al. 2021) and covering the period 969–1933 CE. It shows the persistence of the 11-yr Schwabe cycle throughout the last millennium, with a Δ¹⁴C amplitude of 0.9‰ during solar maxima and 0.6‰ during minima, and an average length of 10.4 yr. Details of the Wolf, Spörer, and Maunder minima have been studied in annual ¹⁴C data by Eastoe et al. (2019), Fogtmann-Schulz et al. (2019, 2020), and Moriya et al. (2019). These same data were used by Land et al. (2020) to connect with the broader scope of dendrochronology by comparing the record of solar forcing with climatic impact on tree-ring growth for the same years. This is an exciting study in that it combines annually resolved paleoclimatic and solar proxies from the same calendar-dated tree-ring sequence and shows a direct influence of the Schwabe cycle on climate. It reaffirms that ¹⁴C in tree-ring chronologies can provide information on solar variability on time scales of years to centuries and millennia.

Earth’s Magnetic Field

The Earth’s magnetic field protects Earth against the solar wind and helps it shield against cosmic rays that produce radiocarbon in the upper atmosphere, with implications for Earth radio communications, satellite networks and life on Earth (Channell and Vigliotti 2019). The magnetic field strength, position, and polarity have long been known to vary over millions of years based on measurement of remanent magnetism in rocks (Kono 2007, remanent magnetism is the permanent magnetism in rocks, resulting from the orientation of the Earth’s magnetic field at the time of rock formation in a past geological age) but multi-millennial-scale production changes of cosmogenic isotope abundance are also considered to be caused mainly by these geomagnetic variations (Beer et al. 2012). High-resolution production records showing the “secular trend” of modulation by the Earth’s magnetic field were originally derived from radiocarbon measurements on long tree-ring sequences as part of the early calibration-curve development efforts, which revealed long-term quasi-sinusoidal variation in radiocarbon production (Sonett and Finney 1990). Between 1970 and 1990 atmospheric ¹⁴C reconstructions from tree-rings were used to infer changes of the global magnetic dipole moment, assuming a constant carbon cycle (Damon and Linick (1986); Stuiver et al. (1991) and references therein). Geomagnetic measurements on sediments and rocks are now the primary source of information about Earth’s magnetic field. Calculating this field requires a record of global dipole moment that is modeled from localized data covering part of the globe, which has only recently converged through three different models (e.g., Nilsson et al. 2014).

Solar Reconstructions and ¹⁴C Spike Dating

The discussion of the full sequence of solar variability, solar-terrestrial interactions and the response of the climate system is outside the scope of this paper (see Haigh (2007) and Gray et al. (2010) for reviews). The first steps involving reconstruction of sunspots and of total solar irradiance (TSI) or solar spectral intensity (SSI) have been gradually developed in response to the progress in determining atmospheric ¹⁴C over the Holocene and the availability of ¹⁰Be data from polar ice cores. Steinhilber et al. (2012) reconstructed TSI from a combination of 2000-yr high-pass filtered ¹⁴C in IntCal09 and ¹⁰Be from seven ice cores in Greenland and Antarctica, and compared TSI to the δ¹⁸O record of Dongge cave, China, thought to represent a signal of the Asian monsoon, revealing a significant correlation. Wu et al. (2018a, 2018b) presented the latest reconstruction of sunspots and TSI (Figure 4) and SSI back to 6755 BCE from ¹⁴C and ¹⁰Be. The range of the TSI variability on a millennial scale is determined to be ca. 0.11% (1.5 W m⁻²).

Figure 4 Total solar irradiance reconstructed from ¹⁴C (IntCal09) and six ¹⁰Be data sets from polar ice core archives. (Reproduced from Wu et al. 2018a.)

The high-resolution solar data, and the 774–775 CE event in particular, have been used as a chronological anchor point to secure the dating of ice-core records (Sigl et al. 2015) and to improve the dating of a number of “floating” tree-ring sequences (i.e., sequences that are not dendrochronologically anchored to the present with exact calendar dates using the tree-ring dating methods previously described). The replication of the 774/775 CE event in multiple global locations has also proved useful for exploring ideas about latitudinal, regional and laboratory offsets (Büntgen et al. 2018). Wacker et al. (2014) demonstrated how the 774–775 CE marker event could be used effectively (along with traditional dendrochronological techniques) to provide a precise and accurate construction date for wood samples from the St. John convent in Müstair, Switzerland, a UNESCO Word Heritage Site. Similarly, anchoring a tree-ring sequence on the 774–775 CE event allowed Oppenheimer et al. (2017) and Hakozaki et al. (2018) to date the volcanic eruption of Changbaishan, on the China–North Korea border, to late 946 CE. It also allowed Pearl et al. (2020) to establish and secure a paleoenvironmentally significant 2500-year-long Atlantic white cedar (Chamaecyparis thyoides) chronology from the Northeastern United States. The revolutionary potential of this approach in archaeological contexts has been well emphasized by Dee and Pope (2016) and has been demonstrated to dramatic effect when Kuitems et al. (2020) used the event to resolve the origins of the Uyghur monument of Por-Bajin in Russia, dating construction to the summer of 777 CE and resolving decades of debate.

The ¹⁰Be Connection

Changes in ¹⁴C over time are directly related to atmospheric ¹⁴CO₂, which registers production changes caused by the solar/magnetic variability previously described. This variability simultaneously produces a record of ¹⁰Be that is preserved as a parallel record in the polar ice cores. This can provide multiple advantages. Prior to the beginning of the continuous tree-ring-based ¹⁴C calibration data (i.e., before 14,000 cal BP), the temporal resolution of the ¹⁴C data from speleothems, lake and marine sediments, and corals is coarser, leading to some attenuation of the atmospheric ¹⁴C variability, which is thus less accurate for calibration purposes. However, floating tree-ring series, found for various intervals in glacial times in both hemispheres, can be independently fixed in time by matching the ¹⁴C from such samples with the ¹⁰Be ice record that extends beyond the secure tree-ring record. Once in the correct temporal position, the tree-ring ¹⁴C can then be used to improve the resolution of the calibration data for such time periods. For example, bidecadal ¹⁴C measurements across 2000 rings of New Zealand kauri around Heinrich event 3 (Turney et al. 2016), placed in time relative to the ¹⁰Be record from the GRIP ice cores, now refines the IntCal20 calibration data (compared to IntCal13) for the period ca. 30,000 cal BP. In that case, some ambiguity over the temporal ¹⁴C position relative to the ¹⁰Be record was resolved by the use of D-O (Dansgaard–Oeschger) phase 3, as recorded in the Cariaco climate proxy, to confirm the dating. Similarly, kauri was also measured in decadal blocks across 1300 years around the time of the Laschamps geomagnetic minimum event, showing the strong rise of Δ¹⁴C caused by the event. The link to a ¹⁰Be reconstruction provided an ice core-based age of 42,500 cal BP, which was 1000 years younger compared to the Hulu cave ¹⁴C ages (Turney et al. 2010), possibly resulting from carbon cycle changes causing different signals in ¹⁴C. A detailed discussion is presented in Staff et al. (2019). Meanwhile, Scots pine and other sub-fossil trees from Northern Italy sampled at 5- and 10-year resolution show large age variations ca. 12,400 ¹⁴C years BP during the Bølling warm phase, which were not evident in the coarser resolution IntCal13 data, but could clearly be seen in the GRIP ¹⁰Be record. These tree-ring ¹⁴C series were included in IntCal20, resulting in a wide calendar age bias for the onset of Greenland Interstadial 1 (Bølling chronozone), shown in Fig. 9 of Adolphi et al. (2017). Finally, for the Holocene period, where tree-ring ¹⁴C is calendar secured, the comparison with ¹⁰Be from the ice cores has instead been used to fine tune the chronological precision of the ice cores (e.g., Adolphi and Muscheler 2016) and to explore these combined records to improve our understanding of solar dynamics and to quantify the solar influence on climate (e.g., Steinhilber et al. 2012).

WHAT’S NEXT?

The future directions for tree-ring based radiocarbon research will likely include large-scale creation and replication of multi-millennial, multi-regional time series of annual (or sub-annual) measurements of ¹⁴C from individual dated tree-rings (e.g., Kudsk et al. 2019; Fogtmann-Schulz 2019; Friedrich et al. 2020; Pearson et al. 2018, 2020b; Fahrni et al. 2020; Brehm et al. 2021). Such data are now well established in terms of feasibility/quality of measurement (Sookdeo et al. 2020; Wacker et al. 2010), dynamic multi-purpose functionality and scientific value. There now seems a strong basis (Friedrich et al. 2020; Pearson et al. 2018, 2020b) for such data to be used to refine the structure of future iterations of the radiocarbon calibration curves as these community-based resources continue to evolve and improve (Reimer et al. 2020; Hogg et al. 2020). The current Northern Hemisphere IntCal data now include annual data from trees from across Europe, North America, and Japan (for a comprehensive review and future recommendations see Bayliss et al. 2020), and it seems likely that both the geographic and temporal spread of such data (with inter-laboratory replication) will increase for both hemispheres. In the Southern Hemisphere there is particular potential for earlier time periods, in combination with ¹⁰Be, from New Zealand’s swamp kauri (preserved in anoxic bog environments similar to the Irish oaks) that span a remarkable range of time periods extending back over 70,000 years (Turney et al. 2010; Lorrey et al. 2018). It also seems likely that an increasing number of annually resolved floating tree-ring series will be created to more finely delineate climatic events in Northern Hemisphere calibration data e.g., Capano et al. (2020).

While multi-regional, annually resolved ¹⁴C time series undoubtedly have a future role in calibration, such data are multi-functional and the driving force for their creation may also (instead) come from applications to explore and predict the occurrence of solar/magnetic phenomena (Miyake et al. 2017; Park et al. 2017; O’Hare et al. 2019; Fahrni et al. 2020), gain new insights into the role of solar forcing in different climatic regions (Land et al. 2020), and be used to synchronize with other high-resolution paleoclimatic proxy evidence (e.g., Reinig et al. 2020). The continued two-way feedbacks between dendrochronology and radiocarbon will likely see both an increased use of single-year tree-ring based ¹⁴C determinations to secure floating tree-ring sequences, either using SPEs (e.g., Kuitems et al. 2020) or using other subdecadal structure revealed by annual ¹⁴C (e.g., Pearson et al. 2020a) or to provide independent verification of the dendrochronological dating of others (discussed by Büntgen et al. 2018). The same annual tree-ring based ¹⁴C data will also be used to explore further questions of small-scale regional or latitudinal ¹⁴C variability (see Reimer et al. (2020), Bayliss et al. (2020) for in-depth discussion, plus data in Büntgen et al. (2018), Pearson et al. (2020a, 2020b). There are also considerable new possibilities to trace the movement of the ITCZ through time using ¹⁴C in tropical tree-rings as a means of climatic modeling, as well as in accessing parallel, highly resolved contemporaneous records of marine ¹⁴C, by crossdating tree-ring and marine sequences (Black et al. 2019).

This brief overview of the interwoven history of dendrochronology and radiocarbon advancements, exploration of cosmogenic isotopes, and study of solar magnetic activity, has demonstrated the complexity and productivity of the interactions between these fields. The high scientific relevance of the data generated has stimulated substantial progress in building millennia-long chronologies, greatly improving precision in isotope techniques, and advancing models of heliomagnetic impacts on the Earth’s climate. The data have also revealed the previously unknown presence and intensity of extreme solar proton events, potentially harmful to technology on Earth. Future pathways will undoubtedly result in an impressive further array of interdisciplinary research.