ORIGIN AND AGE OF CARBON IN THE CELLULOSE OF MID-LATITUDE TREE RINGS

Abstract

Cellulose of tree rings is often assumed to be predominantly formed by direct assimilation of CO₂ by photosynthesis and consequently can be used to reconstruct past atmospheric ¹⁴C concentrations at annual resolution. Yet little is known about the extent and the age of stored carbon from previous years used in addition to the direct assimilation in tree rings. Here, we studied ¹⁴C in earlywood and latewood cellulose of four different species (oak, pine, larch and spruce), which are commonly used for radiocarbon calibration and dating. These trees were still growing during the radiocarbon bomb peak period (1958–1972). We compared cellulose ¹⁴C measured in tree-ring subdivisions with the atmospheric ¹⁴C corresponding to the time of ring formation. We observed that cellulose ¹⁴C carried up to about 50% of the atmospheric ¹⁴C signal from the previous 1–2 years only in the earlywood of oak, whereas in conifers it was up to 20% in the earlywood and in the case of spruce also in the latewood. The bias in using the full ring of trees growing in a temperate oceanic climate to estimate atmospheric ¹⁴C concentration might be minimal considering that earlywood has a low mass contribution and that the variability in atmospheric ¹⁴C over a few years is usually less than 3‰.

INTRODUCTION

Tree rings play a fundamental role in radiocarbon dating. Due to their absolute dating and annual resolution, tree rings provide a continuous radiocarbon reference for dating any carbonaceous sample of organic origin over the last ca. 14,000 years (Reimer et al. 2020). Recent technological improvements in radiocarbon measurements are now allowing sub-ring resolution (Xu et al. 2015; Cain et al. 2018; Kudsk et al. 2018; Turney et al. 2018; Svarva et al. 2019; McDonald et al. 2019) opening new opportunities for intra-annual dating. For such an application, questions regarding sub-ring radiocarbon variation and its variability across sites and species need to be clarified and understood.

Annual rings of most trees in the middle and high latitudes are composed of two parts, the earlywood (EW), formed at the beginning of the growing season, and the latewood (LW), whose formation begins in the summer. Moreover, since these two parts have multiple functions (water transport vs structural support) and display different cell anatomical characteristics, they also undergo cell formation processes with different timing, rates and requirements (Rathgeber et al. 2016, Rathgeber et al. 2022) that vary among species (Chen et al. 2022) and environments (Rossi et al. 2016). Phenological studies showed that in ring-porous deciduous species (e.g., oak) and deciduous conifers (e.g., larch) the process of earlywood formation starts around bud-burst time (e.g., Moser et al. 2009; Dox et al. 2022), i.e., when photosynthesis is still low and new assimilates for structural use are not yet available. This initial growth is mainly sustained by large pools of nonstructural carbohydrates (NSC), which are stored in the perennial tissues of trees and can be mobilized through the parenchyma/medullary rays. Medullary rays extend over many annual rings into the center of the tree and, in contrast to these, are alive over many years. A considerable part of the mobilizable NSC (especially starch) is deposited in these cells during the winter dormant period. However, these pools are a mixture of NSC with different ages (see Gessler and Treydte 2016). The contribution of reserves accumulated from the previous year or years to initial growth has been also supported by studies using tree-ring stable isotopic information (Helle and Schleser 2004; Gessler and Treydte 2016; Boettger and Friedrich 2009). The latewood, on the other hand, begins formation in the summer, when the leaves/needles are fully functional. Thus, growth and metabolic activity are mainly maintained by recent assimilates, although there are still some uncertainties about how this pattern is modified under stressful conditions such as extreme droughts.

This mix of ¹⁴C signatures from old and new assimilates within a single tree-ring naturally raises the question of the extent to which the carbon of an annual ring matches the atmospheric ¹⁴C concentration of the year or more precisely the season of growth. This consideration has large relevance nowadays, as ¹⁴C is now more easily measured even at intra-ring resolution. The strongest observed intra- and interannual atmospheric ¹⁴C fluctuations occurred during the atmospheric bomb testing period between 1955 and 1962, and the successive decline in atmospheric ¹⁴C. This period provides an ideal framework to tackle questions regarding the amount of incorporation of previous years carbon in a tree-ring in addition to the direct uptake from the atmosphere.

Several studies have used the bomb-spike period to investigate the age of the atmospheric ¹⁴C uptake in tree rings. Grootes et al. (1989) studied the 1962–1964 tree rings in spruce from the US Pacific coast and concluded “that there is little or no contribution of stored photosynthate to radial growth.” Olsson and Possnert (1992) observed that earlywood ¹⁴C concentration of Swedish oak was clearly lower than the atmospheric values for 1962 and 1963 indicating significant use of stored carbon from previous years. In a more recent study by (Kudsk et al. 2018), no influence of stored carbon was found in either earlywood or latewood in Danish oak by measuring the ¹⁴C content of the cellulose for the years 1954–1970.

MATERIAL AND METHODS

Sample Selection

For our study, we selected a deciduous ring-porous species (oak), a deciduous coniferous larch, and two evergreen coniferous species with different needle retention times (5–7 years for spruce, 4–5 years for pine). For each species, we analyzed one individual. In addition, we compared lowland to highland trees, the latter of which have a shorter growing season (Figure 1). The selected species are among the most commonly used for the development of the radiocarbon calibration data sets (Stuiver et al. 1986; Reimer et al. 2004; Bayliss et al. 2020):

1. (1) Oak (Quercus robur L.) from Rust in the Upper Rhine valley, Germany (48°16’04,6’’N; 7°42’09,5’’E; 160 m a.s.l.) – ring-porous deciduous tree species.

2. (2) Pine (Pinus nigra Arnold) from Botanical Garden at Stuttgart-Hohenheim, Germany (48°42’37,8’’N; 9°12’25,4’’E; 380 m a.s.l.) – evergreen conifer

3. (3) Larch (Larix decidua Mill.) from Lötschental, Switzerland, (46°23’30.42’’N; 7°45’40.8’’E; 1300 m a.s.l.) deciduous conifer, 1-year needles

4. (4) Spruce (Picea abies (L.) Karst.) from Lötschental, Switzerland (46°23’30.42’’N; 7°45’40.8’’E; 1300 m a.s.l.) evergreen conifer

Figure 1 Geographic locations of the sites of the four species (one tree each) selected in this study, and of the Central European ¹⁴C background site in Vermunt, Austria.

The oak from Rust was a healthy, ca. 130-year-old, living tree from a hardwood floodplain forest in the “Taubergiessen” nature reserve in the Upper Rhine valley. The oak was a predominant tree from the overstory with an exposed crown. The black pine comes from the Botanical Garden of the University of Hohenheim. Before it fell over in a storm, it was ca. 140-year-old and ca. 25-m-high solitary tree without any understory.

We divided the individual tree rings into three subsections, which are supposed to represent three easily defined sections of the wood increment: Ring-porous deciduous tree species (oak): one section of earlywood (EW1), two sections of latewood (LW1, LW2). Conifer trees (pine, larch and spruce): Two sections of earlywood (EW1, EW2) and one section of latewood (LW) (Figure 2).

Figure 2 Scheme of sampling: A) Conifer (larch), B) oak. Due to the differences in the proportions of each whole ring represented by the earlywood and the latewood in conifers and oak, the earlywood (EW) in conifers has been divided into two sections (EW1, EW2) and latewood (LW) in one section. The oak tree rings were divided into one earlywood (EW) and two latewood (LW1, LW2) sections. In oak, the broad medullary rays (MR) are excluded for analysis.

In addition to the longitudinally oriented wood cells, the annual rings of all wooden species are interspersed with radially oriented medullary rays, which extend over many annual rings into the center of the tree, and which are alive over many years (Gessler and Treydte 2016). Because of the extended life of these medullary ray cells and to avoid any possible contamination by translocation of older or younger carbon, we have removed the medullary rays in the oak annual rings as best we could by preparation. This is possible with oak wood because the medullary rays are present in a number of large bundles and can therefore be easily identified and removed under the light microscope. This is not feasible with conifer wood because the numerous medullary rays that are only a few cell rows wide are spread over the entire cross-section.

Growth periods

For pine and oak trees, we estimated the mean earlywood and latewood growth periods based on continuous dendrometric measurements at Hohenheim Botanical Garden in 1990’s and 2000’s aligned with leaf phenological observation data of the investigated tree species at sample sites Hohenheim and Rust, which cover the entire period of dendrometric measurement and leaf phenological observations (Holz et al. 2010). We took into account that in northern hemisphere conifers cell enlargement and wall thickening are completed several weeks earlier than lignification (subsequent incorporation of lignin into the cell walls made of cellulose), which continues into late autumn (Cuny et al. 2015). This is of particular interest here because we only measure ¹⁴C in the cellulose, as is standard procedure for high-precision ¹⁴C dating. For the larch and spruce trees from the Lötschental, we estimated the period of tracheid secondary wall deposition to EW1, EW2 and LW as the average of cellular observations (from weekly microcores) performed for 4 trees per species at the same sites during the period 2008 to 2013.

The calculations have been performed according to the method described in Cuny et al. (2019) and Pérez-de-Lis et al. (2022). Primary data, namely repeated cellular observations performed on stem microcores, were obtained from Cuny et al. (2019). The cellular observations allowed determination of the time of tracheid enlargement and secondary wall deposition. Since cellulose deposition of the secondary wall includes most of the cellulose used in the ring structure, and since lignin deposition co-occurs with a temporal offset of few days (see Rathgeber et al. 2022), the timing of cellulose deposition and the observations of secondary wall deposition are given in Table 1. The phenology of the 2008-2013 observations were assumed to be an acceptable estimate for the phenology 1962-1965 since difference in the timing of peak climatic signal in wood cell anatomical features between early and late twentieth century have been shown to only have a minimal shift of a few days (Carrer et al. 2017).

Table 1 Years of ¹⁴C measurement and estimated growth periods for the subsections of the annual rings.

Sample Preparation and AMS Measurement

All samples were pretreated for holocellulose using the BABAB technique (Němec et al. 2010). The pretreatment starts with a base step overnight at 60°C (4% NaOH), followed by an acid step using 2M HCl, an additional base step before the final acid step. In between each step the samples were always washed two to three times with ultrapure water.

Holocellulose was then extracted by using sodium chlorite, where hydrochloric acid (HCl) was added until pH 3. The NaClO2 solution was repeatedly replaced after 2 hours until the remaining holocellulose was white. Then the samples were washed twice with ultrapure water and dried before being combusted in an elemental analyzer (Microcube, elementar) and graphitized in an AGE III system (Ionplus) (Wacker et al. 2010). The oak and the pine samples were measured at the CEZA laboratory, Mannheim, Germany (sample identifier starting with “MAMS”) (Kromer et al. 2013) while the spruce and larch samples were measured in ETH Zurich, Switzerland (Wacker et al. 2010). Blank measurements were done on phthalic anhydride as a measurement blank, but also on wood samples (process blanks) undergoing the same pretreatment together with the samples and containing no ¹⁴C (lignite from Reichwalde in Germany and Kauri wood from New Zealand) (Sookdeo et al. 2020).

Time Series of the Central European Atmospheric ¹⁴C Concentrations

Essential for this study is the ¹⁴C concentration of continental background air in Central Europe during the interval of ring growth over the respective years. For the northern hemisphere monthly ¹⁴C data were combined into a sector NH1 from atmospheric measurements from Norway, Austria (Vermunt) and China Lake (USA) (Hua et al. 2013) We decided to compare the tree-ring ¹⁴C to only the ¹⁴C measurements in the Austrian mountains (Vermunt, Montafon, Figure 1) for these reasons:

1. (1) Only in Vermunt did the ¹⁴C measurements start by 1959, covering fully the interval of our samples (Levin et al. 1985), with several days of sampling at the center of each month of the year during the interval considered in this study.

2. (2) The Vermunt data could be compared to the 1966 to 1983 data of a site close to the oak site (50 km distance), Schauinsland at 1100 m asl above the Rhine valley, and close agreement has been found (Levin et al. 1985, Figure 2).

RESULTS

The data are shown in Figure 3 for the full set of the four species, and in Figure 4 for the extended interval of 1965 to 1972 for the oak and pine species. The full ¹⁴C dataset of all species and ring sectors is given in the supplementary table.

Figure 3 F¹⁴C of earlywood (open symbol) and latewood (filled symbol) of all four species in the interval of 1962–1965, compared to data of an atmospheric sampling site (Vermunt, Germany). Growth periods indicated by horizontal bars.

Figure 4 F¹⁴C in earlywood (open symbol) and latewood (filled symbols) of oak and pine (not corrected for fossil CO₂) in the interval 1958–1972.

We observed considerable differences between the species:

1. (1) F¹⁴C in latewood of oak agreed well with the Vermunt ¹⁴C activity, but the earlywood was considerably lower in the years 1962 and 1963, and higher for 1965 and the following years. The EW fraction of the oak tree shows up to 11% lower values the biggest offsets compared to the atmosphere compared with all the other tree-ring fractions.

2. (2) F¹⁴C in earlywood and latewood of pine follow closely the transitions of Vermunt, but they are shifted significantly to lower values than Vermunt (mean difference 1959-1972 ΔF¹⁴C=0.025). The difference could come from lower ¹⁴CO₂ of soil carbon (heterotrophic respiration) or from fossil CO₂. Measured CO2 levels in Stuttgart-Hohenheim (near the airport Stuttgart and a major highway) in the year 2002 were 29 ppm higher compared to the natural reserve site in the Rhine valley, the location of the oak. Hence, depletion of the atmospheric ¹⁴C content by fossil fuel CO2 is the plausible explanation, and it is corrected for in the model calculations of the section below. After correction, an insignificant contribution from NSC from previous years can be seen in the F¹⁴C signature of the earlywood as well of the latewood. Soil carbon has a longer average residence time, so we should observe low bomb test ¹⁴C in this contribution before 1965, and higher ¹⁴C than atmospheric ¹⁴C later, which is not the case.

3. (3) The F¹⁴C in the deciduous larch follows more closely the atmospheric values of Vermunt in 1962 and 1963 than the deciduous oak tree, though here also the EW is a bit lower. In the following years 1964 and 1965 the F¹⁴C of EW follows closely the atmosphere.

4. (4) The F¹⁴C in the spruce in general, but more specifically in the latewood of 1962 and 1963, is significantly lower than Vermunt F¹⁴C. At this site, fossil CO₂ influence can be excluded because of the Alpine clean air site and the correct larch latewood values from the same location, so the earlywood ¹⁴C results can therefore be considered unbiased by fossil-fuel CO₂. While the F¹⁴C for 1964 is still a little bit lower, for 1965 it agrees well with the atmosphere.

Modeling

To better understand and interpret our obtained results, a simple model was applied to compare the measured ¹⁴C content of the tree-ring sections with the theoretical ¹⁴C concentration assuming direct assimilation of CO₂ by photosynthesis and continuous integration into the cellulose. For the model, we assumed the amount of woody biomass and thus also the cellulose production (Cuny 2015) to be gaussian shaped around mid-summer (July 7th). The width of the distribution was adjusted to the assumed growth period of the trees at the different locations, with a shorter period for the high-elevation trees from the Lötschental (full width at half maximum: 0.25 yr^–1) and a longer period for the low-elevation trees from Southern Germany (fwhm: 0.31 yr^–1).

We then calculated the theoretical radiocarbon concentration (atmospheric F¹⁴C, Table 2) assuming the formed cellulose was based on the direct assimilation of atmospheric CO₂ and on a rapid transport from source to sink (Epron, 2012). This was done by integrating accordingly the interpolated atmospheric ¹⁴C concentrations from Vermunt (Levin 1985) over the growth periods for the equivalent EW and LW fractions (see table 2, ring portion).

Table 2 Measured and modeled radiocarbon concentrations expected for direct atmospheric (Levin et al. 1985) uptake in the tree rings. Also shown are values for a storage pool with NSC reserves that can be remobilized and incorporated into the cellulose cell wall of current year, with a best-fit annual exchange rate to estimate the integration of pre-aged carbon in addition to direct uptake from the atmosphere.

The measured F¹⁴C can now be compared to the atmospheric F¹⁴C (direct assimilation). We assume here, that the atmosphere is well mixed within the region of study, an assumption supported by atmospheric measurements from Jungfraujoch (Levin et al. 2013), which compare well with the ones from Schauinsland (as long as there are no local sources of fossil CO₂). The difference we explain with a variable contribution (Table 2, fraction NSC) of a constant NSC storage pool that is refreshed every year at a specific exchange rate R (annual exchange rate of NSC storage pool, see also Table 2). ${F^{14}}{C_{NSC}}$ of the NSC storage pool for a specific year (y) is calculated by a recursive procedure from the previous year (y - 1):

$${F^{14}}{C_{NSC}}\left( y \right) = \left( {1 - R} \right)*{F^{14}}{C_{NSC}}\left( {y - 1} \right) + R*{F^{14}}{C_{Atm}}\left( {y - 1} \right)$$

where the replaced fraction R of the NSC storage pool has the same signature as the atmosphere ( ${F^{14}}{C_{Atm}}).\;$ The ${F^{14}}{C_{NSC}}{\rm{\;}}$ is calculated for every year starting with the year 1951 (run-in of at least 10 years for comparison with measured data). For the best possible fit, we find that the exchange rate per year of the storage pool tends to be a bit higher for the oak and pine (R = 70%) than for the spruce and larch (R = 50%) from the elevated sites in the Lötschental. A low exchange rate will result in lower values than the atmosphere after 1963, while a high exchange rate will result in higher values. The exchange rate was thus fitted to give the best results for the years 1964 and 1965. In a last step, the fraction NSC for a specific year is estimated assuming the measured F¹⁴C of the cellulose ( ${F^{14}}{C_{Meas}})\;$ is the sum of carbon directly assimilated from the atmosphere and of NSC from the storage pool, with their respective F¹⁴C signatures:

$$f\left( y \right) = {{{F^{14}}{C_{Meas}}\left( y \right) - {F^{14}}{C_{Atm}}\left( y \right)\;} \over {{F^{14}}{C_{NSC}}\left( y \right) - {F^{14}}{C_{Atm}}\left( y \right)\;}}$$

Note that the results should only be considered significant, when the atmospheric F¹⁴C is significantly different from the F¹⁴C of the NSC storage pool, which is primarily the case for 1963 and 1964.

IMPLICATIONS AND DISCUSSION

The F¹⁴C measured in tree rings follows relatively closely the atmosphere during the time of growth. However, specific fractions as for example the EW fraction of the oak, deviate significantly from the atmosphere by up to 11%. The average deviation over the full annual rings is, however, much lower and only up to 3%.

A simple model was used to quantify the amount of stored carbon used by the tree to explain the observed discrepancies. The applied model, though neglecting naturally occurring variabilities (e.g., Carbone et al. 2013; Richardson et al. 2013), seems to predict relatively well the average use of NSC for specific trees.

Deciduous Trees

Unsurprisingly, because a newly formed earlywood is a prerequisite for the water supply during refoliation, the deciduous trees incorporate significant amounts of stored carbon from previous years in the EW, while the latewood is essentially built up from carbon of the current year (Michelot et al. 2012; Richardson et al. 2015). Especially the oak tree uses ca. 50% of the amount of stored carbon. This agrees well with the observation that a large portion of the EW is formed before budburst (and therefore before photosynthesis starts). Our results, however, are not in agreement with observations done on similar F¹⁴C measurements on EW and LW of Danish and Swedish oak trees (Kudsk et al. 2018).

The stored carbon in oak trees seems to be of relatively young age because the F¹⁴C of the EW of 1965 is already increased compared with the atmosphere. Therefore, an average age for the carbon incorporated greater than 1–2 years is not indicated for the years analyzed (we estimated an average annual exchange rate of 70%). The deciduous conifer larch, being physiologically very different from the oak, also shows a detectable influence of stored carbon on the F¹⁴C of the first EW fraction. At 25% on average, the amount is only half as large as for the oak tree (where the EW fraction is also smaller). The storage pool in the larch seems to integrate over somewhat longer timescales, as the EW for 1965 the F¹⁴C was still not above the atmosphere, which would be expected for very short storage times.

Evergreen Conifer Trees

Even if the pine tree shows a significant influence on the F¹⁴C from the enhanced local anthropogenic CO₂ emission of fossil fuels, which we can correct for by an estimated fossil fuel contribution of 2%, the corrected F¹⁴C-values in all tree-ring sections follow the atmosphere very closely. Therefore, no strong influence of stored carbon is visible, but a small contribution of stored carbon on the EW or the LW of less than 10% cannot be excluded, confirming data of Grootes et al. (1989) and Svarva et al. (2019) and physiological observations that evergreen conifers obtain their carbon for the wood structure mainly from current photosynthesis of the previous year’s needles and less from reserve carbon (Michelot et al. 2012).

The spruce from the Lötschental shows again a clear contribution of stored carbon (up to 20%), but primarily in the LW, not in the EW. While the evergreen conifer tree spruce already assimilates CO₂ from the atmosphere in early spring, the amount is still expected to be small and thus a larger contribution of stored carbon was also expected on EW.

Implications for Tree Physiology

Several studies reported that earlywood growth, particularly in deciduous species, is supported by NSC pools from previous years as photosynthate production is still low. Gaudinski et al. (2009) reported that the age of these NSC was maximum of 1 year in white and chestnut oak whereas Muhr et al. (2016) found a mean age of the C of 3–5 years in sugar maple. Our results at the intra-annual level also indicate a substantial contribution of previous year reserves to earlywood. Such results also agree with previous studies that found a significant reduction of NSC pools (starch) at the beginning of the growing season in oak (Michelot et al. 2012). However, the values of the year 1965 can be a good example of how carbon from more than one previous year could be incorporated. Oak latewood sections showed a better fit with Vermunt measurements reinforcing the fact that growth will thus be mainly governed by recent assimilates. However, in oak the wood density of the earlywood (0.47 g/cm³, SD ± 0.27 g/cm³), which consists mainly of very large-lumen vessels, is clearly less than that of the latewood (0.74 g/cm³, SD ± 0.15 g/cm³) (Mayr et al. 2003), which consists mainly of very dense fiber cells and latewood vessels. The mass proportion of the earlywood in the total ring is therefore much lower than the mass proportion of the latewood. However, this depends to a considerable extent on the width of the annual rings. The width of the total annual rings shows clear fluctuations, caused by the site and weather-related growth conditions. However, the weather-related fluctuations are essentially only reflected in the latewood width, while the earlywood width shows very little year-to-year fluctuation.

Here, information from various approaches in plant synthesis is helpful, particularly in studying the age and seasonal use of actual versus stored NSCs in different species. There is a broad range of literature; here we refer to just two major publications and the references therein. Michelot et al. (2012) studied oak, pine and beech at lowlands east of Paris. They observed the beginning of oak growth 14 days before budburst, and a significant decrease of starch in this interval. For pine, the beginning of pine growth appeared coincident with budburst, and no change in NSC is seen. Richardson et al. (2015) studied ¹⁴C of NSC on the outermost 20 rings of white pine and red oak in a temperate forest of the northeastern United States. The age of the NSC in the most recent five years was found to be less than 1 year, while older stem wood had NSC ages of more than 5 years. Grootes et al., 1989 found use of less than 15% of stored photosynthate, but a contribution of 13% to 28% CO₂ from decay of biological material, in a Sitka spruce at the Pacific coast of Washington.

Specific Implications for Annual ¹⁴C Series on  oak tree-rings

Since the absolute amount of earlywood of ring-porous oak wood is largely independent of the growing conditions, its relative proportion compared to the latewood has implications for high-resolution ¹⁴C data measured on annual tree rings as demonstrated here. With narrow annual rings, the proportion of earlywood is significantly higher than with wide annual rings. Concerning the total annual ring, it is, therefore, to be expected that with narrow annual rings, i.e., with higher proportions of earlywood, the proportion of carbon from reserve materials in the cellulose will be higher than with wide annual rings, where the proportion of latewood with its higher density is significantly higher (Mayr et al. 2003).

However, before the bomb ¹⁴C interval, the difference of F¹⁴C within 4 years or less in the atmosphere amounts to only a few permille, except for solar particle events (SPE), where it can be up to 12 permille in one year. Therefore, the uptake of ¹⁴C in earlywood, especially of oak, must be considered in the period of rapid and strong fluctuations during the bomb tests and SPEs, but it is not statistically significant elsewhere. For most applications, total oak tree rings can therefore be used for producing high- resolution (even annual) ¹⁴C time series.

CONCLUSIONS

Using the bomb ¹⁴C signal (especially the years 1962–1965) generated during the nuclear bomb tests starting in the 1950s, our analyses were able to identify the contribution of NSCs from previous years before the growth year of the tree rings. In general, depending on the species mostly earlywood is affected. With regard to the suitability for high resolution (annual, resp. sub-annual) ¹⁴C-series, we can summarize that there are clear differences between the different tree species we have investigated.

For the ring-porous deciduous oak, the cellulose of the earlywood shows a substantial contribution of stored carbohydrates from the previous year(s), whereas the latewood is formed exclusively from currently photosynthesized C. The earlywood of the deciduous conifer larch also shows a significant proportion of C from previous years, but to a much lesser extent than the oak, and the latewood follows the atmosphere closely. The evergreen pine, on the other hand, essentially does not seem to incorporate significant amounts of previous year(s) carbohydrates in either earlywood or latewood, even if there is some uncertainty here due to the increased fossil C input at that site. Finally, for spruce, a small proportion of reserve C is discernible in the earlywood, but the spruce also incorporates “older” stored carbohydrates into the latewood.

However, for pre-bomb years, earlier signals in earlywood are usually beyond the ¹⁴C measurement error (1.5 to 2‰ at best), because of (1) the low contribution of earlywood to the total carbon mass of a ring, and (2) max. 15 permille/year change during SPEs, and 20 permille/60 years at strong solar minima. Hence, total tree-ring ¹⁴C data are still useful for ¹⁴C calibration using annual ¹⁴C data of trees growing in a temperate oceanic climate. Only in special cases, such as comparisons of annual ¹⁴C data sets between different regions/latitudes, the temporal differences of uptake of seasonal ¹⁴C and incorporation of NSC in earlywood suggest focusing on latewood only, as documented here for oak and larch.