Annual 10Be Record for 1510–1701 CE Obtained From Endogenic Travertine at Baishuitai, China: A New Proxy Record of Annual Solar Activity

The 10Be record in laminated travertines is a potential proxy for reconstructing past solar activity down to the annual scale; however, correcting for the potential influence of climatic or environmental variations remains challenging. Here, we present an annually resolved 10Be record using travertines from Baishuitai, China, covering the period from 1510 to 1701 CE, along with environmental proxies, to evaluate climatic influences and implement corrections to accurately reconstruct solar activity. We demonstrate that the 10Be deposition in travertines exhibits two environmental impacts: the transport efficiency of atmospheric 10Be into travertine and the additional 10Be inflow from overland flow associated with rainfall. We show that these impacts can be corrected based on iron and potassium contents. The resulting corrected record agrees with ice‐core and tree‐ring records, demonstrating the feasibility of using such carbonate sediment 10Be records to reconstruct past solar activity.


Introduction
Cosmogenic 10 Be is a proxy to yield pivotal information on solar activity variations over time (e.g., Steinhilber et al., 2012; reviews by Usoskin (2023)).It is mainly produced in the lower stratosphere and upper troposphere via galactic or solar cosmic-ray-induced spallation reactions involving atmospheric constituents (Lal & Peters, 1967).The 10 Be production rate in the atmosphere is proportional to the cosmic ray intensity, which are modulated by changes in solar magnetic activity and geomagnetic field strength (Lal & Peters, 1967).After production, 10 Be adsorbs onto the surface of aerosols and is scavenged from the atmosphere mainly through the precipitation (Heikkila et al., 2011).The mean residence time of 10 Be is 1-2 years in the stratosphere, whereas it stays only a few weeks in the troposphere (Heikkila et al., 2011).Thereafter, 10 Be undergoes the process of the geochemical cycle and is stored in natural archives.
To date, a number of different polar ice core 10 Be series with resolutions typically ranging from 5 to 50 years have been obtained to reconstruct the centennial to millennial solar variability during the late Pleistocene and the Holocene (e.g., Adolphi et al., 2014;Bard et al., 1997;Delaygue & Bard, 2010;Horiuchi et al., 2008Horiuchi et al., , 2016;;Raisbeck et al., 1981).However, ice core 10 Be records with annual resolution covering several centuries that facilitates the reconstruction of annual or decadal-scale solar variations are still scarce.There are only two 10 Be records with annual resolutions covering past six centuries (Beer et al., 1990(Beer et al., , 1998;;Berggren et al., 2009) and a few short-term annual 10 Be records for detecting extreme solar proton events (e.g., Mekhaldi et al., 2015;Miyake et al., 2015Miyake et al., , 2019;;O'Hare et al., 2019;Paleari et al., 2022).Moreover, the reconstruction of annual solar activity using ice core 10 Be records is constrained by the challenge of chronological control precision due to ice flow and thinning effects over time (Beer, 2000).In addition, climatic influences on the atmospheric mixing, transport and deposition of 10 Be could result in non-solar signals in ice core 10 Be records (Beer et al., 2013;Heikkila et al., 2011;Pedro et al., 2012;Zheng et al., 2023).Using 10 Be from other archives would help complement and extend the existing 10 Be records.
10 Be in travertine, as discovered in our previous studies (Miyahara et al., 2020;Xu et al., 2019), has considerable potential as a proxy for solar variability.Travertine deposits are widely distributed worldwide, often having clear annual layers with high deposition rates (e.g., 1-20 mm/yr in Baishuitai, China; Liu, 2014), likely spanning back to 1 Ma (Capezzuoli et al., 2013).Unlike ice samples, the annual layers in travertines do not become thinner with increasing depth; thus, they are potential archives for high-resolution solar activity reconstructions that include a considerable number of years.In addition, 10 Be records from travertines distributed at mid-to low-latitude regions may contribute to the understanding of the production, transport, and deposition of 10 Be.However, correcting possible climatic influences on travertine 10 Be record remains challenging.A 16-year-long annual 10 Be record of modern travertine from Baishuitai, China, yielded a significant correlation with the global 10 Be production rate (Xu et al., 2019), but the amplitude of travertine 10 Be variation was nearly 1.7 times larger than that of the global production rate, even after correcting for possible climatic and environmental impacts using 9 Be and potassium (K) contents, suggesting that potential remains for improvement in correcting these impacts.Longer travertine records are therefore required to validate or improve the methodology for correcting climatic and environmental impacts.
In this study, we present an annually resolved 10 Be record from Baishuitai travertine covering 1510-1701 CE, with environmental proxies, discuss the possible influence of climatic or environmental variations on the travertine 10 Be record, and propose an improved correction methodology for extracting the atmospheric 10 Be production signal from the travertine.We then compare the obtained 10 Be time series with existing solar activity records to assess its potential for reconstructing past solar activity.

Sampling Site
Baishuitai is located in Shangri-La County, Yunnan Province, southwest China (Figure S1a in Supporting Information S1), at an elevation of approximately 2,380-3,800 m (Zhao et al., 1998).The mean annual precipitation is ∼750 mm, with >75% of the precipitation occurring from May to October.The average annual temperature is ∼8°C (Sun et al., 2014).The Baishuitai area is a typical constructive karst landscape.The limestone is heavily fractured, resulting from strong geotectonic activity and rock weathering, providing favorable conditions for the infiltration of rainfall and inflow of groundwater.Owing to the presence of a fault-controlled drainage cut-off, there are numerous springs enriched with HCO 3 and Ca 2+ ions, upsurging at approximately 2,600 masl, such as S1 and S2 in Figure S1b of the Supporting Information S1.The discharge of the spring water is perennial and exhibits only slight seasonal variation (∼50-60 L/s) depending on the timing, duration, and intensity of the monsoon (Sun & Liu, 2010).After the groundwater emerges, a large amount of CO 2 is degassed, and the aqueous solution becomes supersaturated relative to calcium carbonate.Therefore, a travertine platform of approximately 3 km 2 with a thickness of several tens of meters has been formed ∼200 m from the springs (Liu et al., 2003) (Figure S1b in Supporting Information S1).
During the rainy season, heavy rainfall dilutes spring water and decreases the deposition rate while promoting the incorporation of eroded soils in overland flow.These processes cause the travertine layer to become thinner and dark brown.During the dry season, the travertine deposition rate increases because of the higher evaporation rate, which, coupled with a lower amount of mineral sediment, makes the layer thicker and lighter.Since the main water source is the spring water supplied by precipitation, and because the majority of atmospheric 10 Be is retained in precipitation, this continuously adds atmospheric 10 Be to the travertine.Environmental conditions at Baishuitai allow annual layers well structured and suited for deriving data with a solid timescale.

Measurements of 10 Be in Annual Layers
A travertine record (27°30′15″N, 100°2′1″E) was obtained from the fossil travertine area at Baishuitai, approximately 200 m from spring S1 (Figure S1b in Supporting Information S1).It was approximately 250 cm long and had a well-defined, laminated structure (Figure S1c in Supporting Information S1).We identified 192 pairs of white and dark layers across the blocks.After drying at 60°C for 1 week and cleaning the surface, we cut the block with a slab saw and ground each sample to powder in a stainless-steel mortar following Xu et al. (2019), and collected 192 individual white (W-series, W1-W192) and 192 dark (D-series, D1-D192) samples (see Text S1 and Figure S3 in Supporting Information S1 for details).D-series are the deposits formed during the rainy season (May-October), while W-series are those deposited during the dry season (November-April of the subsequent year) (Liu et al., 2010).As highlighted by Xu et al. (2019), 10 Be in the white layers inherits the production signal better than 10 Be in the dark layers; thus, the W-series samples were analyzed in this study.Please note, however, that due to the possible presence of the conduit or reservoir underground, 10 Be in the white layers may partially reflect the 10 Be production rate during the preceding rainy season.
The extraction of 10 Be was conducted following procedures given in Xu et al. (2019).To improve the recovery rate of Be, ∼0.5 ml Fe(NO 3 ) 3 solution (1,000 ppm, Wako) was added with ammonia water (NH 3 •H 2 O) to promote the co-precipitation at the first precipitation process.We then conducted the measurement using the Accelerator Mass Spectrometory (AMS) at the University of Tokyo (Matsuzaki et al., 2015).We repeated the measurements for each sample at least twice to achieve a precision of less than 10%.The 10 Be/ 9 Be value and error for each sample were calculated based on the weighted mean and standard deviation of repeated runs.

Measurement of Environmental Proxies
To explore the influence of environmental factors on 10 Be deposition, trace elements, oxygen isotope, and travertine residue (the detrital material transported by the overland flow) were measured.The 9 Be content in travertines was measured using atomic absorption spectrometry at Hirosaki University, and standard addition techniques, following Horiuchi et al. (2016), were applied for samples W1-W11, W55, W124, and W153-W192, while the calibration curve method was used for the remaining samples.Owing to a limited data collection period, the 9 Be content measured using the latter method was based on a single measurement, with a uniform 25% margin of error, leaving room for improvement.The iron (Fe) and potassium (K) contents were measured using an inductively coupled plasma-optical emission spectrometer at East China Normal University.Analytical errors were generally less than 5%.δ 18 O in travertine was measured using a Finnigan MAT252 instrument at the Institute of Geochemistry, Chinese Academy of Sciences, following the procedures in Sun and Liu. (2010).The standard deviation (1σ) of the δ 18 O measurement was below 0.20‰.The residues from dissolving travertines in acetic acid were dried at 105°C for 24 hr and weighed.

Dating
The sequence of the travertine record was first determined by counting the annual layers.The radiocarbon content of the plant roots found in the 165th layer (D165) was then measured to determine the rough absolute age of the travertine record.Given that plant roots in this layer were delivered by overland flow rather than later growth, their radiocarbon age can be used to approximate the time of travertine deposition.Finally, the travertine 10 Be record was compared with the record of 14 C production rate obtained by Brehm et al. (2021) to estimate the precise absolute age of the travertine record.More details can be found in Supporting Information S1 (Text S2).

Results
All data obtained are shown in Figure 1.The annual 10 Be concentration ( 10 Be conc ) in the travertine varies between 1.19 × 10 5 and 6.28 × 10 5 atoms/g, with a mean of 1.81 × 10 5 atoms/g (Figure 1a).The anomalous 10 Be conc was found in W165 and was 3.5 times higher than the mean 10 Be conc , which may be caused by 10 Be contamination from the surrounding environment (see Text S3 in Supporting Information S1 for further details).

10.1029/2023GL107434
The approximate10 Be flux in the travertine ranges from 54 to 469 atoms/m 2 /s, with a mean of 175 atoms/m 2 /s (see Text S4 in Supporting Information S1 for details).There was a strong positive correlation between 10 Be flux and thickness (R = 0.86, p < 0.001), whereas there was no significant correlation between thickness and 10 Be conc in the travertine (R = 0.09, p = 0.75) (Figures S6 and S7 in Supporting Information S1), suggesting that the longterm trend of 10 Be flux in travertine largely reflects the travertine accumulation rate.Note that the correlation coefficient was estimated using Pearson's correlation analysis, while the p-value was estimated using a non-parametric random-phase test (see Text S5 in Supporting Information S1 for details, Ebisuzaki, 1997).
The δ 18 O in the travertine exhibits a gradual negative trend toward the deeper layers (Figure 1f), indicating that the amount of local precipitation was larger (Sun & Liu, 2010;Sun et al., 2022).The increase in the rainfall amount leads to the enhancement of groundwater discharge, resulting in a faster spring water flow rate.Therefore, it can be considered that the thickness of the travertine layers, showing an increasing trend toward older ages (Figure 1g), is caused by this increment of spring water flow.The content of K and residue in travertines, incorporated associated with the overland flow (Xu et al., 2019), also supports that the amount of rainfall was higher during the time (Figures 1d and 1e).

Climatic/Environmental Influence on 10 Be Deposition and Its Correction
Regional climatic or environmental changes affect the deposition of 10 Be and then on the 10 Be accumulating in travertine.For example, contamination from eroded soil transported in overland flow may change the 10 Be concentrations in the travertine (see Figure S8 in Supporting Information S1, Xu et al., 2019).Also, it is possible that the transport of 10 Be via the spring water is changed in time.Here, we assessed these effects and corrected them by comparing the recorded 10 Be concentration to the environmental proxy time series from the same travertine.

Effect of Iron on the Transport Efficiency of Atmospheric 10 Be Into Travertine
Figure 2a shows a scatter diagram of Be isotopes ( 10 Be and 9 Be, both normalized) and the Fe content in the travertine.It exhibits a significant positive correlation between 10 Be and Fe (R = 0.40, p < 0.001), as well as between 9 Be and Fe (R = 0.70, p < 0.001).With the exception of W165, the rates of change coincided with each other.These results confirm that iron may play a role in incorporating 10 Be into travertines.Fe(OH) 3 colloids are good absorbents for 10 Be (McHargue & Damon, 1991;Willenbring & von Blanckenburg, 2010), and those leached from the soil above the limestone terrain act as a medium for long-distance migration of 10 Be (McCarthy & Zachara, 1989) from the subsurface to the river via spring water (Figure S8 in Supporting Information S1).Therefore, the efficiency of the transport of 10 Be, which inherits the atmospheric production signal, toward the travertine platform depends on the amount of Fe(OH) 3 in the spring water.An additional influence may arise during the crystallization of calcium carbonate from spring water to travertine, as local precipitation may leach some Fe(OH) 3 colloids from the surrounding environment, which then may enter the travertine through overland flow.However, we observed a high degree of dispersion in the relationship between Fe and K or between Fe and the residue in the travertine (Figures 2b and 2c).The K content was used to monitor the strength of the overland flow in our previous study (Xu et al., 2019).The residue collected by dissolving travertine in acetic acid is the soil carried by the overland flow; thus, the amount of residue can be utilized to estimate the strength of the overland flow.This result suggests that Fe from overland flow is small and negligible at the studied site and we, therefore, assume that the effect of Fe on 10 Be occurs during the transport of atmospheric 10 Be into spring water and rivers.
The δ 18 O fluctuations in the travertine revealed that local precipitation varied during period examined.As local precipitation increases, more Fe(OH) 3 colloids are leached from the soil, and the transport efficiency of 10 Be in the travertine changed.Thus, a correction is needed.Normalizing meteoric 10 Be with 9 Be has become a standard practice to compensate for such environmental impact on 10 Be records obtained from terrestrial sediments (e.g., Horiuchi et al., 2016;Menabreaz et al., 2011;Wittmann et al., 2012), and it may be applied to travertine 10 Be because 10 Be and 9 Be agree with Fe in travertines.However, not all of the 9 Be in the travertine may be from the spring water and could have been contaminated by overland flow.We estimated that the addition of soil materials carried by overland flow may account for up to 8% of the increase in 9 Be content in travertine (see Table S1 in Supporting Information S1 for detailed estimation).In such cases, using 9 Be to normalize 10 Be may overcorrect the influence of the transport efficiency of 10 Be.Therefore, we used the Fe content for correction.Helz and Valette-Silver. (1992) used a similar strategy.
To correct transport efficiency of 10 Be, normalizing parameter (α Fe ) was applied to 10 Be concentration (Table S2 in Supporting Information S1). Figure S9a in Supporting Information S1 shows the variation of 10 Be concentration before and after normalizing by α Fe .

Effect of Overland Flow on 10 Be Concentrations
In addition to the effect of iron on 10 Be transport efficiency, overland flow caused by local precipitation may also affect the 10 Be concentration in travertines (Xu et al., 2019).For example, δ 18 O was lower during the formation of the W140-192 (Figure 1), during which the residue amount was also increased.Rainfall-induced surface runoff erodes and transports soil particles, with subsequent incorporation into the travertine.Because 10 Be is adsorbed on the surface of soil components owing to its particle affinity (Willenbring & von Blanckenburg, 2010), an increase in the incorporation of soil particles into the travertine enhances the 10 Be content.We found that there exist weak positive correlations between 10 Be conc (norm)/α Fe and the amount of residue (R = 0.29, p = 0.003) (Figure 2d),  10 Be conc (norm)/α Fe versus residue, and (e) 10 Be conc (norm)/α Fe versus K from the same travertine.Note that the regression line is fitted from the scatter points excluding the red diamond/circle (W165).Significance levels of the correlations in (a), (d), and (e) were calculated using a random phase test (Ebisuzaki, 1997).and between 10 Be conc (norm)/α Fe and K (R = 0.26, p = 0.025) (Figure 2e), suggesting a possible effect of overland flow on 10 Be deposition.
Although the overall effect of overland flow on the travertine platform seems much smaller than the previous sampling site (Xu et al., 2019), correlation between 10 Be conc (norm)/α Fe and K in Figure 2e suggests that the 10 Be inflow associated with occasional heavy rainfall events is non-negligible.For example, the concentration of 10 Be in W119 was ∼50% higher while K and residue contents were also markedly higher than in neighboring layers, although Fe content was relatively stable (green shading in Figure 1).This suggests that some augmentation of 10 Be concentration may occur associated with the 10 Be inflow from the overland flow.Therefore, correction is required.We use K to estimate the 10 Be component ( 10 Be EOF ) owing to the effect of overland flow (Figure 2e), following the same strategy as Xu et al. (2019) (see Text S6 in Supporting Information S1 for further details).The regression equation is: Then, an atmospheric 10 Be signal ( 10 Be ATM ) can be extracted by subtracting 10 Be EOF from 10 Be conc (norm)/α Fe : 10 Be ATM = 10 Be conc (norm)/α Fe 10 Be EOF + 1 (2) The variation of 10 Be conc (norm)/α Fe before and after the K correction is shown in Figure S9b of the Supporting Information S1, revealing that 3%-24% of the variability in the travertine 10 Be record was attributed to the overland flow effect (Table S2 in Supporting Information S1).
In summary, we corrected the effect of iron on the transport efficiency of atmospheric 10 Be into travertine and the effect of overland flow on 10 Be concentration in travertine using the associated Fe and K contents, respectively.Note that we decided not to include W165 in the above correction analyses due to our current limited understanding on the extreme event (see Text S3 in Supporting Information S1 for further details).The missing data point for W165 in the 10 Be time-series was interpolated using a cubic spline for the subsequent analyses.

Solar Activity Variation Derived From the Travertine 10 Be Record
To evaluate the centennial-scale solar signal in the travertine 10 Be record, we compared the corrected travertine 10 Be concentration ( 10 Be ATM ) with the NGRIP 10 Be flux record (Berggren et al., 2009), the stacked ice 10 Be concentration record based on the NGRIP and Dye-3 records (see McCracken & Beer, 2015), and 14 C production rate record obtained from tree rings (Brehm et al., 2021) as shown in Figure 3.Note that the travertine 10 Be record in this study was dated as 1510-1701 CE (see Text S2 in Supporting Information S1 for details).We used the timescale transfer function suggested by Adolphi and Muscheler. (2016) to correct the possible offset of the ice core 10 Be records for comparison with the travertine 10 Be record.The travertine 10 Be record exhibited a declining trend between 1520 and 1600 ( 1.1%/decade) and reached its lowest level at approximately 1600, and exhibited an overall upward trend afterward (1.6%/decade), with a smaller rate of change than that of the other referred records (Table S5 in Supporting Information S1).This long-term variation in the travertine 10 Be record agrees with that of the 14 C production rate, likely reflecting the Suess cycle of solar activity (about 210 years).
Furthermore, the travertine 10 Be record exhibits a decadal variation overliad on the long-term variation.The peak of such decadal-scale variations coincided well between the travertine 10 Be and 14 C records, such as at approximately 1527, 1586, 1623, 1642, and 1689 (see Figure 3).They also coincide with those in the ice core 10 Be record, although there are some discrepancies between the travertine 10 Be and ice-core 10 Be records at certain ages (e.g., around 1530-1545 and 1565-1575) (shaded in gray in Figure 3).These discrepancies might result from the disturbances of 10 Be precipitation by local climatic impacts on either or both of the records.For example, in the case of Greenland ice core 10 Be records, Pedro et al. (2012) and Zheng et al. (2020) have suggested that they may be affected by the North Atlantic Oscillation (NAO) and the Pacific-North American Pattern (PNA).Climatic impact may also disturb the 10 Be precipitation in the low to mid-latitude regions, and further investigations are needed to specify the cause of the discrepancies.The power spectrum and wavelet analysis of the travertine 10 Be record showed significant peaks at 11.3 years (Figure S10 in Supporting Information S1), suggesting that the evident decadal variation reflects the Schwabe cycle of solar activity.To further assess the Schwabe solar cycle of the travertine 10 Be record, we compared the travertine 10 Be record with the ice core 10 Be flux (Berggren et al., 2009), the stacked ice core 10 Be concentration (McCracken & Beer, 2015), and 14 C production records (Brehm et al., 2021) (Figure S11 in Supporting Information S1).The 14 C production time series was downsampled to an annual resolution.All time series were high-pass filtered for <35-year and then filtered using a 5year running average.The length of solar cycle recorded by travertine 10 Be ATM was not constant, as suggested by sunspot (Solanki et al., 2004) and 14 C analyses (Miyahara et al., 2004(Miyahara et al., , 2008(Miyahara et al., , 2021)).The travertine 10 Be record, the ice-core 10 Be records, and the 14 C production rate record in tree rings exhibited decadal variations that were mostly consistent with each other.These results demonstrate that the 10 Be record in laminated travertine can be a new tool to trace the past 11-year solar cycles.Some discrepancies were seen around 1530-1545 and 1565-1575 between the travertine 10 Be and ice core 10 Be records, again possibly due to the climatic impact as mentioned above.
In addition, the amplitudes of the Schwabe cycles in the travertine 10 Be ATM time series (∼9%) were generally smaller than that of the ice-core 10 Be flux (∼17%) and 10 Be stack (∼13%) time series during the studied period (Figure S11 in Supporting Information S1).Note that these amplitudes are all based on 5-year running averages of the data.One possible explanation for such amplitude differences is the latitudinal dependence of the 10 Be production variability (Poluianov et al., 2016) and incomplete atmospheric mixing (Zheng et al., 2024), as the travertine samples were obtained from a relatively low latitude (27.5°N), while the ice 10 Be records were obtained from samples at a much higher latitude (>65°N).However, earlier researches have found that ice-core 10 Be records show similar amplitude with the theoretically expected global 10 Be production rate record during the modern period, as well as with the 14 C production record during the Holocene, suggesting a small latitudinal effect (Adolphi et al., 2023;Zheng et al., 2020).Further examinations are needed, based on 10 Be series from multiple locations, to verify the cause of the discrepancy in the amplitudes.
In summary, the agreement between the travertine 10 Be record and the ice-core 10 Be or tree-ring 14 C records supports the idea that the Baishuitai travertine 10 Be record reflects variations in atmospheric production, and thus, that of solar activity.

Conclusions
In this study, we obtained a new high-precision annual 10 Be record for 1510-1701 CE using the travertine obtained from Baishuitai, China.Iron was suggested to control the transport efficiency of atmospheric 10 Be into travertine.Meanwhile, the impact of overland flow was relatively weak for the study site.The influence of the transport efficiency and the effect of overland flow on 10 Be concentration in travertine can be assessed using the associated Fe and K contents.The corrected travertine 10 Be record exhibited a good agreement on centennial and decadal scales with the ice-core 10 Be and tree-ring 14 C records, reflecting the Suess and Schwabe cycles of solar activity.We conclude that the 10 Be in travertine can be a good proxy for past solar activity.

Figure 1 .
Figure 1.Time series of 10 Be concentration and environmental proxies from the Baishuitai travertine.(a) 10 Be concentration, (b) 9 Be, (c-g) Fe, residue, K, δ 18 O, and thickness, respectively.The extremely high 10 Be concentration at the 165th layer is highlighted by gray shading.The cyan shading indicates the 119th layer.The red pentagram indicates the location of the sample for radiocarbon dating.

Figure 2 .
Figure 2. Scatter diagrams of (a) Be isotopes ( 10 Be conc and 9 Be) versus iron (Fe), (b) Fe versus potassium (K); (c) Fe versus residue; (d) 10 Be conc (norm)/α Fe versus residue, and (e)10 Be conc (norm)/α Fe versus K from the same travertine.Note that the regression line is fitted from the scatter points excluding the red diamond/circle (W165).Significance levels of the correlations in (a), (d), and (e) were calculated using a random phase test(Ebisuzaki, 1997).

Figure 3 .
Figure 3.Comparison of the travertine 10 Be and existing solar activity records during the study period.(a) Stacked ice annual 10 Be record (Ice_ 10 Be stack ) based on the NGRIP core 10 Be concentration from Berggren et al. (2009) and the Dye-3 core 10 Be concentration from Beer et al. (1990) (see McCracken & Beer, 2015); (b) NGRIP core 10 Be flux (Ice_ 10 Be flux ) (Berggren et al., 2009); (c) Corrected travertine 10 Be concentration (Tra_ 10 Be ATM ) (this study); (d) Global 14 C production rate derived from tree ring (Tree_ 14 C pro ) (Brehm et al., 2021); (e) Group sunspot numbers reconstructed by Svalgaard and Schatten.(2016).The fine lines represent annual data, and bold lines represent the 5-year running average.The yellow shadows represent the periods when the travertine 10 Be record matches other records, while the gray shadows indicate the periods the travertine 10 Be record disagrees with the ice core 10 Be records.Note that the timescale transfer function suggested byAdolphi and Muscheler. (2016) was used to correct the possible offset in ice core 10 Be records for comparison with the travertine 10 Be record.