Temperature-Dependent Oxygen Isotope Fractionation in Plant Cellulose Biosynthesis Revealed by a Global Dataset of Peat Mosses

Introduction

The oxygen isotope ratio (¹⁸O/¹⁶O, expressed as δ¹⁸O) of plant cellulose (hereafter δ¹⁸O_c) is a valuable tracer to characterize the environmental conditions during tissue growth (Barbour, 2007; Sternberg, 2009). It has been used as a proxy to reconstruct past climate conditions in tree rings (Libby et al., 1976), lake sediments (Edwards and McAndrews, 1989), and peat deposits (Hong et al., 2000). Earlier studies established empirical relationships between δ¹⁸O_c and climatic parameters, such as temperature, relative humidity, and precipitation amount for a given region (Libby et al., 1976; Treydte et al., 2006; Managave et al., 2010). An improved understanding of physiological and biochemical processes enabled developing mechanistic models to quantitatively decipher the δ¹⁸O_c signal for paleoclimate reconstruction (Roden et al., 2000; Kahmen et al., 2011). It has been well recognized that the source water for vascular plants, in most cases precipitation, is taken up by roots for photosynthesis in leaves, while the leaf water is enriched in ¹⁸O relative to precipitation due to plant transpiration through stomata. The carbohydrate product (e.g., sucrose) that has been further enriched in ¹⁸O by biochemical fractionation is then translocated to the site of cellulose synthesis, such as stem, where additional biochemical exchange occurs with stem water. As such, plant δ¹⁸O_c signal essentially reflects precipitation δ¹⁸O (hereafter δ¹⁸O_p) with an offset governed by both physiological and biochemical processes (Richter et al., 2008; Sternberg, 2009).

The current mechanistic models have made progress in depicting the physiological effects including how different climatic parameters affect the degree of ¹⁸O enrichment in leaf water and imprint the δ¹⁸O_c signal (Kahmen et al., 2011). However, less is known about the biochemical effects that are equally important in shaping the final δ¹⁸O_c signal. The combined physiological and biochemical effects could be described by the following equation (Barbour and Farquhar, 2000):

$${\mathrm{\delta }}^{18}{O}_c={\mathrm{\delta }}^{18}{O}_p+{\mathrm{\Delta }}_{leaf}(1-p_x{p}_{ex})+{\mathrm{\epsilon }}_{bio}(1)$$

where δ¹⁸O_c and δ¹⁸O_p are the oxygen isotope composition of cellulose and precipitation, respectively; Δ_leaf is the degree of ¹⁸O enrichment in leaf water known as the physiological effect; p_x is the proportion of unenriched water at the site of cellulose synthesis; p_ex is the proportion of exchangeable oxygen in carbohydrate substrates destined for cellulose synthesis; and ε_bio is the biochemical enrichment (or fractionation) factor associated with the exchange of oxygen atoms between carbonyl group and medium water. The p_x is expected to be unity for stem cellulose (Cernusak et al., 2005; Kahmen et al., 2011), but for leaf cellulose it is expected to be less than unity (Helliker and Ehleringer, 2002; Barbour, 2007; Kahmen et al., 2011; Song et al., 2014b; but see Liu et al., 2017), or even has been assumed to be zero (Tuthorn et al., 2014). The p_ex and ε_bio have been successfully constrained using different approaches to be 0.42 ± 0.3 and 27 ± 3‰, respectively (DeNiro and Epstein, 1981; Sternberg and DeNiro, 1983; Sternberg et al., 1986; Yakir and DeNiro, 1990; Roden et al., 2000). These biochemical factors are a net result of complex biochemical processes in cellulose synthesis (Waterhouse et al., 2013) and seem to contain less climatic information, thus these factors are almost always treated as constants in mechanistic modeling. For instance, Helliker and Richter (2008) incorporated global tree-ring δ¹⁸O_c data into mechanistic model and then inferred that boreal tree-canopy temperature must be tremendously higher than ambient temperature to explain large observed Δ_leaf from their data. They also found that the inferred tree-canopy temperature throughout a large range of latitudes converges to ∼21°C. This conclusion, however, was reached based on constant ε_bio and p_ex in their inverse modeling. Later, Sternberg and Ellsworth (2011) challenged this conclusion by showing that the ε_bio is also temperature-dependent and increases with decreasing temperature. They heterotrophically generated cellulose from starch substrate at different temperatures and found that the ε_bio could be as large as 31‰ when temperature is 5°C. However, Zech et al. (2014a, b) indicated that the temperature-dependence of ε_bio shown in that laboratory experiment might be an artifact from a variable p_ex which, albeit not statistically significant (Sternberg, 2014), is temperature-dependent instead.

Here we use a global compilation of Sphagnum (peat mosses) δ¹⁸O_c data to test the hypothesis that the ε_bio is temperature-dependent. Sphagnum mosses are the keystone bryophyte species in peat-forming ecosystems, and these mosses are adapted to cool, waterlogged, and nutrient-poor conditions. Almost half of the boreal peatland areas are covered and built by Sphagnum mosses (Rydin and Jeglum, 2013). There have been a number of studies exploring the potential of stable isotopes in Sphagnum as proxies for peat-based paleoclimate reconstruction (e.g., Ménot-Combes et al., 2002; Daley et al., 2010; Loader et al., 2016; Granath et al., 2018; Xia et al., 2020). In general, Sphagnum mosses have many unique physiological characteristics with which their tissue δ¹⁸O_c could closely track δ¹⁸O_p with an offset dictated by ε_bio. First, Sphagnum mosses are important peat-forming plants in bogs where moisture is predominantly derived from precipitation. Second, Sphagnum mosses have no stomata and no root systems. They maintain a simple water use strategy without the ability to control evapotranspiration. The water loss on Sphagnum surface is balanced either by new precipitation input or by capillary movement of water from below (Rydin and Jeglum, 2013; McCarter and Price, 2014). Therefore, Sphagnum mosses passively utilize water stored in surface unsaturated and oxic peat known as the acrotelm without the ability to access water at depth as vascular plants do. Third, Sphagnum mosses have the extraordinary capacity to store water and, more importantly, could protect the metabolic water from evaporation loss (Xia et al., 2020). Sphagnum have specialized large hyaline (water-filling) cells on their leaves and outer cortex of stems, and in some species these hyaline cells completely enclose photosynthetic cells (Loisel et al., 2009). Most of the water is, however, external capillary water held in pore spaces between leaves (Murray et al., 1989; Rydin and Jeglum, 2013). As a result of these multiple layers of water storage in Sphagnum mosses, the metabolic water in photosynthetic cells of Sphagnum living in local wet habitats could be isolated and weakly impacted by evaporative enrichment of ¹⁸O (Xia et al., 2020). Fourth, Sphagnum mosses are opportunist photosynthesizers, growing and synthesizing new cellulose as long as conditions allow even in near-freezing temperature. This physiological trait allows us to explore the sensitivity of ε_bio to air temperatures that are too low to allow the growth of vascular plants (Clymo and Hayward, 1982). In summary, the newly compiled dataset and analysis presented in this study offer a simple but effective test of the relationship between ε_bio and temperature, while circumventing the complex physiology of vascular plants.

Data and Methods

Sphagnum Oxygen Isotope Data

We compiled all available Sphagnum tissue δ¹⁸O measurements reported in literature, supplemented with some new measurements to fill regional gaps. The final dataset includes 785 individual δ¹⁸O measurements from 173 peatland sites (Figure 1). Key site information and references are summarized in Table 1, while raw data and metadata are in the Supplementary Dataset 1. Most of these sites are bogs from boreal and sub-arctic regions, but a few bog sites are located in the tropical and subtropical regions or the Southern Hemisphere. A few fen sites were also included in the dataset. The main Sphagnum species are the common bog species, S. magellanicum, S. fuscum, S. cuspidatum, and S. capillifolium. In some studies, the local microtopography information (such as hummock or hollow), or water table depth (WTD) for collected Sphagnum samples were available.

FIGURE 1

Figure 1. Global map of peatland regions (green areas; Yu et al., 2010) and sites with Sphagnum δ¹⁸O_c data presented in this study. Sites are color-coded to represent different regions: Alaska (blue), Canada and Contiguous US (orange), Europe (green), New Zealand (yellow), Patagonia (pink), Temperate East Asia (purple), Tropics and Subtropics (brown), and West Siberia (red).

TABLE 1

Table 1. Summary information about the compiled Sphagnum δ¹⁸O data.

The exact materials of Sphagnum mosses used for isotope analysis differ among studies and include “capitula” (the cluster of new branch leaves at the tip of moss), “capitula cellulose,” “newest 1-cm stem cellulose,” “leaf,” “leaf cellulose,” “stem cellulose,” “whole plant cellulose,” and “5-cm whole plant cellulose.” On the basis of our pilot experiments and literature data, we applied an offset correction to raw δ¹⁸O data that were measured from non-cellulose (“capitula” and “leaf”), or stems (“newest 1-cm stem cellulose” and “stem cellulose”), and re-scaled these raw δ¹⁸O data to their equivalent capitula or leaf δ¹⁸O values with the involved uncertainty propagated in the following analysis. Correction is not required for δ¹⁸O data that were measured from “whole plant cellulose” or “5-cm whole plant cellulose” because our pilot experiment (n = 6) found that leaf tissues represent 90.3 ± 5.0% (error is the standard deviation, 1σ) dry mass of Sphagnum whole plant and have a similar cellulose yield (45.6 ± 4.7%) as stem tissues (48.8 ± 8.0%). The details about our cellulose extraction method can be found in Supplementary Text 1.

Precipitation Oxygen Isotope Data and Climate Data

We used the Online Isotopes in Precipitation Calculator (OIPC; Bowen, 2020) to obtain monthly δ¹⁸O_p for every Sphagnum sampling location in the dataset. In the OIPC, δ¹⁸O_p at a given location is modeled based on latitude and altitude plus a residual term that is interpolated using existing δ¹⁸O_p observations (Bowen et al., 2005). To our knowledge, this model provides the best estimate for the site-specific δ¹⁸O_p for our synthesis as there is no on-site precipitation δ¹⁸O measurement prior to Sphagnum sample collection except at three sites. The uncertainty of OIPC is depicted as the confidence interval for mean annual δ¹⁸O_p prediction and is large in regions where δ¹⁸O_p observations are scarce (Bowen and Revenaugh, 2003), but the confidence interval for monthly δ¹⁸O_p prediction is not available in the OIPC. In addition, the OIPC is for predicting long-term mean value for monthly and annual δ¹⁸O_p at a given location but does not account for their inter-annual variability. This inter-annual variability is an important uncertainty source in determining ε_bio for Sphagnum as these samples were collected in different years and in some studies only their recent growth was harvested. To estimate this uncertainty, we used the archived outputs from three nudged isotope-enabled global circulation models (IsoGSM, LMDZ4, and GISS ModelE; accessible from https://data.giss.nasa.gov/swing2/swing2_mirror/; Risi et al., 2012) to calculate the multi-model mean standard deviation of monthly δ¹⁸O_p for each sample location. Therefore, we used the OIPC model that had been validated against observational δ¹⁸O_p data to derive the site-specific δ¹⁸O_p but used the isotope-enabled global circulation models nudged on reanalysis data to estimate the uncertainty associated with its inter-annual variability (at least for 31 years). We acknowledge that the accuracy of monthly δ¹⁸O_p generated in the OIPC is unknown, but we used the prediction error for mean annual δ¹⁸O_p as a reference.

We used the NASA MERRA-2 monthly 2-m air temperature (T2M in M2IMNXASM; GMAO, 2015b) data to infer the relevant growth temperature for each Sphagnum sample that was collected in different year and month. The MERRA-2 is an atmospheric reanalysis product for satellite era (beginning in 1980) and has a resolution of 0.5° for latitude and 0.625° for longitude. The raw temperature data were corrected for the altitude difference between each sampling site and the corresponding MERRA-2 grid assuming a lapse rate of −0.0065°C/m. The MERRA-2 grid altitude is inferred from the constant model parameter surface geopotential height (PHIS in M2C0NXASM; GMAO, 2015a) as PHIS/g, where g is the standard gravity, 9.8 m/s². Similarly, we also used the MERRA-2 monthly precipitation amount (PRECTOTCORR in M2TMNXFLX; GMAO, 2015c) data that were combined with OIPC δ¹⁸O_p data to calculate the metrics of amount-weighted δ¹⁸O_p that represent the isotopic composition of precipitation input for Sphagnum. However, the MERRA-2 products do not provide the uncertainty of reanalysis data other than the temporal variance of monthly mean.

Data Analysis

We first calculated the range and standard deviation of site-specific δ¹⁸O_c data in one-time sampling campaign to characterize the within-site variability and evaluated if measured WTD or microtopographical locations would explain the within-site δ¹⁸O_c variability. Then we calculated the apparent enrichment factor (ε_app) for each δ¹⁸O_c datum as δ¹⁸O_c minus its corresponding δ¹⁸O_p, with the latter value depending on the type of materials and sampling time. We also similarly determined the relevant growth temperature for each sample from the MERRA-2 dataset. For δ¹⁸O_c data derived from the recent growth of Sphagnum (capitula or the newest 1-cm stem), we used amount-weighted δ¹⁸O_p of the sampling month and the previous month (δ¹⁸O_p2) to pair with temperature of the sampling month (T₁) for calculating the ε_app. We assumed that the precipitation input in these recent 2 months was relevant to imprint the δ¹⁸O_c signal in the tissue of recent growth, but growth was only restricted to the course of recent 1 month. To evaluate the sensitivity of calculated ε_app to the choice of relevant time span for δ¹⁸O_p and temperature, we also used δ¹⁸O_p of the recent 1 month or 3 months (amount-weighted), together with the arithmetic average temperature of the recent two or 3 months as a pair. For δ¹⁸O_c data derived from the whole plant of Sphagnum, we used the amount-weighted δ¹⁸O_p of the entire recent 12 months (δ¹⁸O_p12) to pair with the arithmetic average temperature of the most recent growing season defined as the months with MERRA-2 temperature above 0°C during the recent 12-month period. We assumed that precipitation input during all these months either as rainfall or snowfall were relevant to imprint the δ¹⁸O_c signal in whole plant, but growth itself was only restricted to the course of growing season. To minimize, but not necessarily eliminate the influence of evaporative enrichment that has caused the within-site variability in δ¹⁸O_c and ε_app, we used the site-specific minimum δ¹⁸O_c to calculate the site-specific minimum ε_app, i.e., ε_appmin, for each site as an approximation for ε_bio. The precision of ε_appmin was assessed using the Gaussian error propagation combining the analytical uncertainty in δ¹⁸O_c, the uncertainty related to offset correction from raw δ¹⁸O data, and the climatological uncertainty in amount-weighted δ¹⁸O_p values. The accuracy of ε_appmin was assessed using the OIPC prediction error for mean annual δ¹⁸O_p. Pairing the calculated ε_appmin and temperature data, we explored their relationship using a second order polynomial regression. Finally, we compared the result with previous study and discussed the implications for paleoclimate and plant physiology studies.

Results

Our compiled Sphagnum tissue δ¹⁸O data ranged from 11.5 to 26.6‰ (relative to VSMOW). The uncertainty of δ¹⁸O analysis was represented by either the long-term precision of measurements (1σ) in different laboratories or the standard deviation (1σ) of replicate measurements, which was less than 0.6 and 2.0‰, respectively. The difference in δ¹⁸O between extracted cellulose and untreated bulk tissue of the same Sphagnum material was small, and we determined the offset to be 0.69 ± 0.83‰ for capitula (error is the standard deviation, 1σ; Figure 2A). Kaislahti Tillman et al. (2010) reported that this offset is a negative value (Figure 2A), likely due to the fact that they measured Sphagnum materials from peat rather than from growing plants while untreated bulk tissues in peat probably had been contaminated by other chemical components in peat. The difference in δ¹⁸O between leaf cellulose (or capitula, which is formed by branch leaves) and stem cellulose of the same Sphagnum material was also small and well constrained in multiple studies as 0.62 ± 0.51‰ (Figure 2B). This tissue-specific offset might be due to different water pools at the site of leaf and stem cellulose synthesis (Moschen et al., 2009). The capitula medium water might be subtly more enriched in ¹⁸O than the matrix water below capitula where stem cellulose is synthesized. After offset correction, the median within-site ranges of δ¹⁸O_c data were 1.0, 1.5,1.8, and 2.3‰ for sites with two, three, four, and at least five measurements, respectively (Figure 3A). The within-site standard deviations (1σ) of δ¹⁸O_c data were consistently between 0.7 and 0.8‰, and are not influenced by the number of measurements per site (Figure 3B). There is a pattern that more sites have a positive correlation between δ¹⁸O_c and WTD (Figure 3C) and that Sphagnum collected from hummock locations have higher δ¹⁸O_c than from hollow locations (Figure 3D), but few are statistically significant.

FIGURE 2

Figure 2. Oxygen isotope offset between different Sphagnum materials. (A) Biplot showing the relationship between δ¹⁸O_c and untreated bulk tissue δ¹⁸O. Data reported from Kaislahti Tillman et al. (2010) were derived from subfossil Sphagnum branches or stems in peat and are plotted below the 1:1 line, indicating that δ¹⁸O_c are lower than untreated bulk tissue δ¹⁸O in peat. Data from this study were derived from fresh capitula material and are plotted mostly above the 1:1 line, indicating that δ¹⁸O_c are higher than untreated fresh capitula δ¹⁸O. (B) Biplot showing the relationship between branch/capitula δ¹⁸O_c and corresponding stem δ¹⁸O_c. Data reported from Moschen et al. (2009), Kaislahti Tillman et al. (2010), and Finsinger et al. (2013) were derived from subfossil Sphagnum branches or stems in peat, while data from Jones et al. (2019) and this study were derived from fresh plants. The results tend to show that branch or capitula δ¹⁸O_c are higher than stem δ¹⁸O_c and the offset (branch/capitula δ¹⁸O_c minus stem δ¹⁸O_c) resembles a Gaussian distribution as shown in the inset histogram.

FIGURE 3

Figure 3. Within-site variability of Sphagnum δ¹⁸O_c data. (A) Range of within-site δ¹⁸O_c data for each site. (B) Standard deviation of within-site δ¹⁸O_c data for each site. In both (A,B), results are separated for the number of measurements per site. Individual site data are separated for bogs and fens and for recent growth and whole plant parts. Also shown are median and lower/upper quartile values of the range and the standard deviation as cross symbols. The numbers of sites are indicated on top of plot in (A). (C) Histogram of site-specific correlation coefficient between within-site δ¹⁸O_c and WTD data from studies by Loader et al. (2016), Granath et al. (2018), Jones et al. (2019), and Xia et al. (2020) in which WTD measurements are available and there are at least three data points. The significant correlations with p-value lower than 0.05 are shown as dark gray color. Positive correlation indicates that high δ¹⁸O_c correspond to deep WTD. (D) Within-site δ¹⁸O_c data from studies by Taylor (2008) and Brader (2013), plus a site from this study, in which samples were collected at microtopographical locations of hummock and hollow. Data from different sites are separated by horizontal lines. The “ns” denotes “not significant” difference at p < 0.05 level in Student’s t-test between hummock and hollow δ¹⁸O_c data for sites with at least two data points for both hummock and hollow.

For all data from the recent growth of Sphagnum, there was a clear trend that the ε_app increases with decreasing temperature, especially when T₁ is below 10°C (Figure 4A). The whole-plant data complemented the recent-growth data in the temperature range from 6 to 15°C (Figure 4B), but their ε_app did not show a clear trend with temperature. In both recent-growth and whole-plant data synthesis, outliers were mainly those data collected from fens and their ε_app could be either higher, or lower than the rest of dataset. Unlike bogs that are rain-fed, fens are mainly fed by groundwater and stream water that that could complicate the source water δ¹⁸O for Sphagnum. In addition, Sphagnum inhabiting fens are not under desiccation stress and could develop different physiological strategy compared to those in bogs. These fen data were thus not included in the following analysis.

FIGURE 4

Figure 4. The relationship between ε_app and temperature. Symbols are color-coded to represent different regions as in Figure 1. (A) All data for the recent growth of Sphagnum are plotted using the pair of δ¹⁸O_c–δ¹⁸O_p2 and T₁. (B) All data for the whole plant of Sphagnum are plotted using the pair of δ¹⁸O_c–δ¹⁸O_p12 and T_gs. (C) The biplot between the site-specific ε_appmin and growth temperature for bog sites with at least three measurements. The 95% prediction interval of OIPC-based mean annual δ¹⁸O_p at each sampling location is shown as error bar with arrow caps. The precision (2σ) of ε_appmin is denoted as error bar without caps. These data were fitted by a second order polynomial regression shown by the dash-dotted line. Also shown in these figures are the second order polynomial regressions fitted by data from heterotrophic generation of cellulose and data from submerged aquatic plants by Sternberg and Ellsworth (2011) as well as their extrapolations. Note that the full range of temperature (to 30°C) in laboratory experiments by Sternberg and Ellsworth (2011) is shown to illustrate that the sensitivity of ε_app to temperature is very weak at high temperature.

Merging both recent-growth and whole-plant data from bogs and plotting the site-specific ε_appmin for sites with at least three data points versus corresponding growth temperature, the overall relationship that the ε_appmin increases with decreasing temperature was slightly different from the regression models by Sternberg and Ellsworth (2011; Figure 4C). The second order polynomial regression from our data indeed corroborated the trend that the increase in ε_appmin was steeper when temperature was lower, but the sensitivity was very weak for temperature range from 10 to 15°C. Our additional analysis indicated that this pattern was not affected by pairing different time spans of precipitation input and temperature for recent-growth dataset (Supplementary Figure 1). The ε_appmin data points were considerably scattering relative to the trend line, but the deviations were acceptable considering the propagated uncertainty range. Notably, the accuracy of ε_appmin was robust compared to their precision, except in a few cases, as shown by comparing the two types of error bars (Figure 4C).

Discussion

Within-Site Variability in Sphagnum Cellulose Oxygen Isotopes

The δ¹⁸O_c variability for a specific site reflects differential evaporative enrichment of ¹⁸O in metabolic leaf water for Sphagnum inhabiting different microtopographical locations in peatlands, as samples from the same site should receive the same precipitation input (Xia et al., 2020). This evaporation component—which could theoretically be expressed as Δ_leaf (1–p_xp_ex) in Eq. (1)—has modified the δ¹⁸O_p signal before the precipitation source water was incorporated into cellulose synthesis. For most sites, the δ¹⁸O_c data were not measured from the same Sphagnum species, but recent studies suggested that the species effect of leaf anatomy or physiology (biotic factor) does not directly cause the within-site δ¹⁸O_c variability (El Bilali and Patterson, 2012; Xia et al., 2020). Instead, different Sphagnum species favor different microhabitats (abiotic factor) such as drier hummocks, or wetter hollows that indirectly control the degree of evaporative enrichment of ¹⁸O in metabolic leaf water and thus the δ¹⁸O_c signal (Xia et al., 2020).

The result that most sites, if WTD data were available, displayed positive correlations between δ¹⁸O_c and WTD suggests that Sphagnum samples associated with a deeper water table would register a higher degree of evaporative enrichment, although few of these correlations were statistically significant (Figure 3C). Similar result was also found for data collected at hummock and hollow locations within a site (Figure 3D), despite that the actual WTD measurements and the size of these microtopographical features were unknown. The mechanism behind the connection between evaporative enrichment of ¹⁸O in leaf water and WTD is more complex than a simple corollary that drier habitats with deeper water table cause a higher evaporation rate of leaf water and a higher enrichment of ¹⁸O. In fact, the metabolic leaf water is technically difficult to collect as Sphagnum and other bryophytes store a large proportion of water in external capillary spaces (Dilks and Proctor, 1979). Through measuring Sphagnum water content and “bulk” leaf water δ¹⁸O, our previous study proposed that water stored in external capillary spaces of Sphagnum would evaporate first, thereby indirectly protecting internal metabolic leaf water from evaporative loss. As such, drier Sphagnum mosses would have a smaller proportion of external capillary water and as a consequence the metabolic leaf water is also subject to some degree of evaporation (Xia et al., 2020). That being said, the negative correlations between δ¹⁸O_c and WTD also existed in the dataset (Figure 3C), although in at least a few cases these negative correlations were probably an artifact from a narrow range of WTD that did not represent a well-defined moisture gradient (Supplementary Figure 2). In addition, at a few sites Sphagnum inhabiting hollow could have higher δ¹⁸O_c than hummock (Figure 3D). These opposite observations might be due to the decoupling between Sphagnum water content and WTD. Hollow-inhabiting Sphagnum have a higher collective surface area and these Sphagnum are more often susceptible to desiccation (Turetsky et al., 2008; Rydin and Jeglum, 2013) and could have a lower water content despite a shallower WTD, in particular after a sustained period of lacking precipitation. Studies that reported these data did not measure the Sphagnum water content nor had any description of their microhabitats, thus this specific hypothesis could not be directly evaluated.

Unlike vascular plants for which Δ_leaf is governed by leaf-air interaction through stomata and is highly dependent on relative humidity and transpiration rate (Zech et al., 2014a), there is no appropriate physiological model to mechanistically derive the Δ_leaf in Eq. (1) for Sphagnum mosses. Instead, the expression of Δ_leaf in Sphagnum should be considered in terms of the partitioning between internal and external leaf water rather than the atmospheric humidity. This water storage partitioning could explain why Sphagnum δ¹⁸O_c data are less variable at site-specific scale (Figure 3B) compared to their measured “bulk” leaf water δ¹⁸O that could be highly variable or highly enriched in ¹⁸O (Ménot-Combes et al., 2002; Price et al., 2009; Loader et al., 2016; Xia et al., 2020). This “water buffer” mechanism might be unique for Sphagnum mosses adapted to waterlogged peatlands, but not for other bryophytes (Xia et al., 2020). For example, the moss species (Polytrichum strictum and Chorisodontium aciphyllum) from aerobic moss peatbanks in the Antarctic Peninsula showed very high and highly variable δ¹⁸O_c values (Royles et al., 2013, 2016), which are only possible with an unbuffered metabolic leaf water. Due to the presence of external water buffer for Sphagnum mosses, we propose that the evaporative enrichment of ¹⁸O in their metabolic leaf water, despite being difficult to be directly measured or modeled, is a fairly small quantity compared to ε_bio. From a process-based understanding, the evaporative enrichment should approximate to zero for Sphagnum inhabiting very wet habitats or having persistent external water layer, in which case the metabolic leaf water is close to being completely protected by external capillary water from evaporation. For individual sites where Sphagnum δ¹⁸O_c data displayed a considerable variability, the site-specific minimum δ¹⁸O_c value represents the case that the metabolic leaf water is least modified by evaporative enrichment. Therefore, we hereby justify that the corresponding ε_app, i.e., ε_appmin, serves as an alternative measure of ε_bio for each of those sites. Future studies should aim to further refine the processes in how the isotopic signals of diverse Sphagnum water pools (e.g., Ooki et al., 2018) are modified evaporatively or transferred into newly formed cellulose in controlled laboratory experiments (Price et al., 2009; Brader et al., 2010).

Uncertainty in Constraining the Time Span of Precipitation Input and Sphagnum Growth

Defining the time span of growth and precipitation input for collected Sphagnum samples presents a main challenge in refining the relationship between ε_app and temperature. The whole plant of Sphagnum likely integrated the growth for the time span of entire growing season, which is best defined as those unfrozen months (above 0°C). The threshold value of 0°C has been used to define bioclimatic variables such as the photosynthetically active radiation during the growing season (Gallego-Sala et al., 2010; Loisel et al., 2012; Charman et al., 2013). We thus used the T_gs, the arithmetic average temperature of the months above 0°C, as the growth temperature for the whole plant of Sphagnum. In fact, Sphagnum growth during the months of freezing temperature does occur if snow cover was thick and persistent enough to insulate Sphagnum from winter frost (Dorrepaal et al., 2004; Genet et al., 2013), as snow cover shading is not is not a major limiting factor for Sphagnum growth (Clymo and Hayward, 1982; Küttim et al., 2020). That being said, biomass production of Sphagnum during freezing months is much reduced. Considering this, using the T_gs rather than the average temperature of every month of a year, could avoid underestimating the growth temperature for Sphagnum if the freezing season was very long or cold. However, the cumulative temperature—such that is defined as growing degree days above zero in bioclimatic studies (Charman et al., 2013)—might be a more accurate measure of the total warmth Sphagnum have experienced. Studies found that Sphagnum productivity at a given site is enhanced during the warmest season (Krebs et al., 2016; Küttim et al., 2020), although a few studies suggested that Sphagnum growth could be limited by summer desiccation stress and photoinhibition (Lindholm, 1990; Dorrepaal et al., 2004; Genet et al., 2013). Certainly, there is a tendency that T_gs would underestimate the mass-weighted growth temperature for the whole plant of Sphagnum.

Although the freezing season is not considered relevant for Sphagnum growth, the accumulation of winter precipitation as snowpack during the freezing season could be an important component of source water for Sphagnum for the following snowmelt season. Indeed, direct water sampling of peatland water table showed δ¹⁸O values close to amount-weighted mean annual δ¹⁸O_p (Ménot-Combes et al., 2002; Xia et al., 2020). For δ¹⁸O_c data measured from the whole plant of Sphagnum, the amount-weighted δ¹⁸O_p12, which should be the most relevant representation for precipitation input, was used to derive the ε_app to pair with the T_gs in our analysis.

Unlike the whole plant, the capitula and the newest 1-cm stem only represent the recent growth of Sphagnum right before sample collection, and these tissues could register the recent input of δ¹⁸O_p in cellulose. For example, Daley et al. (2010) collected Sphagnum capitula every two to three months and found that the δ¹⁸O_c signals showed variations of more than 4‰ that were in general tracking the variations of δ¹⁸O_p over seasonal scale at two of three sites in northwestern Europe, but at another site the variations were less than 2.5‰, despite similar δ¹⁸O_p seasonality at these locations according to the OIPC model. These observations imply that Sphagnum capitula regenerate and refresh their δ¹⁸O_c signal at bimonthly and quarterly timescales, but the rate of δ¹⁸O_c renewal for capitula appears to be variable. This isotopic renewal rate may depend on biotic factors such as Sphagnum productivity, which has been routinely determined by measuring Sphagnum stem height growth (Clymo, 1970). Field measurements indicate that stem height growth is enhanced with a higher photosynthetically active radiation and a longer growing season length (Loisel et al., 2012) but is also influenced by local ecohydrological factors (Gunnarsson, 2005; Loisel and Yu, 2013). The capitula turnover rate might be proportional to the stem increment rate as the capitula density is lower when stem growth is faster (Breeuwer et al., 2008; Loisel et al., 2012). Aldous (2002) used the ratio of number of branches per capitulum to number of branches per 1-cm of stem to normalize the turnover of capitula to stem increment. This method showed that capitula could replace themselves completely with every 1.65-cm stem increment for a case study of S. capillifolium from peat bogs in the northeastern United States (Aldous, 2002). Nevertheless, the exact time spans for the capitula replacement and the growth of newest 1-cm stem in the dataset are unknown. Other than the physical replacement, capitula δ¹⁸O_c renewal rate may also depend on abiotic factors such as the residence time of metabolic leaf water δ¹⁸O, which in principle should be governed by both the size of acrotelm water reservoir and the rate of precipitation and evaporation. The role of residence time has been proposed to explain the smaller range of along-stem δ¹⁸O_c variation compared to δ¹⁸O_p variation in a case study of S. magellanicum collected from a southern Patagonian peat bog (Xia et al., 2020). The δ¹⁸O_c renewal not only influences how many months of precipitation input are relevant for Sphagnum cellulose synthesis but also the average growth temperature, which are two important but potentially site-specific metrics for our analysis on the temperature-dependence of ε_app.

In our analysis of δ¹⁸O_c data measured from the recent growth of Sphagnum, we used the δ¹⁸O_p2 as the precipitation input to derive the ε_app that was paired with the T₁ as the relevant growth temperature. This choice assumed that the sampling month temperature was the most relevant for capitula growth and the newest increment of stem, while metabolic leaf water not only derived from the precipitation of the sampling month but also partly inherited the precipitation input in the previous month. We notice that changing the definition of these time spans does not affect the overall trend between ε_appmin and temperature, except in the case that the relationship between ε_appmin and temperature became more like linear when the recent 3-month amount-weighted δ¹⁸O_p was paired with the recent 3-month average temperature (Supplementary Figure 1). Another possible complexity is that Sphagnum from the colder regions could have a slower capitulum turnover or stem increment rate (Loisel et al., 2012). If so, a longer time span of both δ¹⁸O_p and temperature should be used specifically for these data. However, we argue that the site-specific local ecohydrological factors are equally important to climate factors to affect Sphagnum growth. For example, the growth rate of Sphagnum in cold arctic regions with very short summer could be remarkably high (Sonesson et al., 1980). That being said, when the recent 3-month amount-weighted δ¹⁸O_p and the recent 2-month average temperature were used instead for samples with T₁ lower than 6°C, we found that the overall trend between ε_appmin and temperature was not affected (Supplementary Figure 3).

The Effect of Temperature on Biochemical Enrichment Factor

The demonstrated relationship between Sphagnum ε_appmin and growth temperature supports the previous finding that the ε_bio is temperature-dependent in wheat seedling experiment by Sternberg and Ellsworth (2011). The relationship between ε_bio and growth temperature in that experiment was established by assuming a known value of substrate δ¹⁸O and a constant of p_ex, while the second order polynomial regression based on their experimental data was almost identical to that based on a compilation of aquatic plant δ¹⁸O_c data (Sternberg and Ellsworth, 2011). From both their and our studies, the influence of temperature on ε_bio is mainly manifested over the range of low temperature. Another previous study showing that the ε_bio is insensitive to temperature (Roden and Ehleringer, 2000) could be reconciled by the fact that the growth temperature for their plants was above 20°C, at which the temperature sensitivity of ε_bio is too weak to be convincingly detected.

Our results imply that the temperature-dependence of ε_bio shown in heterotrophic generation of cellulose is not an artifact from a variable p_ex as previously suggested (Zech et al., 2014a, b), because we used the site-specific minimum Sphagnum δ¹⁸O_c data by which the term Δ_leaf was minimized and temperature-driven variation in p_ex, if any, was too small to introduce variation in δ¹⁸O_c in Eq. (1). The Δ_leaf is usually a large quantity for vascular plants with stomata and increases substantially with decreasing relative humidity (Tuthorn et al., 2014; Zech et al., 2014a). Due to the large contrast between evaporated leaf water δ¹⁸O and unevaporated stem water δ¹⁸O in vascular plants, the p_ex that could vary with plant types and in response to different growth conditions could play an important role in shaping the tree-ring δ¹⁸O_c (Barbour and Farquhar, 2000; Ellsworth and Sternberg, 2014; Song et al., 2014a). Controversy has arisen from determining whether the p_ex is dependent on temperature or not in laboratory experiments (Sternberg, 2014; Zech et al., 2014b), which in turn affects the determination of the corresponding ε_bio. However, our study focusing on the field-collected Sphagnum mosses has avoided the complexity brought by the p_ex. Finally, any variation in p_ex could be further dampened by the factor of p_x, which is less than unity for leaf tissue.

The second order polynomial regression from our dataset is slightly different from that by Sternberg and Ellsworth (2011) in that our regression line is flatter when temperature is above 10°C and steeper when temperature is below 5°C than their extrapolated line (Figure 4C). We recognize that the growth temperature we used is essentially lower than the true temperature for moss growth that only occurs in warmer daytime. By comparison, both the temperatures for seedling growth and of aquarium medium water were controlled as experimental invariables and are more comparable to the daytime temperature for terrestrial plants in field conditions. Future studies should further investigate the discrepancy between mean temperature and photosynthetic temperature for naturally growing mosses and vascular plants.

Implications for Paleoclimate and Plant Physiology Studies

Plant δ¹⁸O_c is a valuable paleoclimate proxy in a range of terrestrial climate archives, but studies employing plant δ¹⁸O_c data should incorporate the influence of temperature on ε_bio into the process-based proxy system model framework for interpreting proxy data (Dee et al., 2015; Keel et al., 2016; Okazaki and Yoshimura, 2019). The temperature effect might be small for Holocene-scale paleoclimate reconstruction, as temperature variability is in general less than 1°C (Liu et al., 2014). However, this effect would be important for high-latitude and alpine regions as the ε_bio exhibited the strongest sensitivity at low temperature and long-term temperature changes at these locations could be amplified due to various feedback mechanisms (Ballantyne et al., 2010; Feng et al., 2019). Furthermore, this temperature effect could be important in cold season, potentially resulting in varying biochemical enrichments in intra-annual tree-ring δ¹⁸O_c records that could be used to reconstruct past climate seasonality (Schubert and Jahren, 2015; Zeng et al., 2016).

Plant δ¹⁸O_c has also been used to understand the ecohydrological processes of plants such as tracing the water source for plants (Sargeant and Singer, 2016) or inversely modeling their growth temperature (Helliker and Richter, 2008; Song et al., 2011). Assuming a constant value of ε_bio, Helliker and Richter (2008) showed that boreal tree-canopy temperature must be substantially higher than ambient temperature. Further analysis by Helliker et al. (2018) found that their previously inferred high tree-canopy temperature for boreal biome is essentially the result of a similarly warm daytime air temperature during growing season but tree-canopy itself could not reach such a high temperature. In fact, the temperature-dependence of ε_bio is more important for subalpine and alpine regions where trees have a cool growing-season temperature (below 10°C) and thus a considerably elevated ε_bio. Indeed, existing tree-ring data from the Swiss Alps (Keel et al., 2016) and the Rocky Mountains in the western United States (Helliker et al., 2018) have shown that incorporating a temperature-dependent ε_bio into mechanistic modeling would produce a better fit for observed δ¹⁸O_c. Furthermore, in a study from the trans-Himalayan region (Chhumuja at an elevation of 4400 m above sea level) where growing season temperature is lower than 10°C, a tree-ring δ¹⁸O_c record was found to be best fitted with a relative humidity of 64%, despite the nearby station-observed values between 70% and 100% during the monsoon season (Brunello et al., 2019). This discrepancy could be alternatively explained by a higher ε_bio at a low growing-season temperature.

Summary

This study focused on Sphagnum mosses as a model plant and used an observational dataset of Sphagnum oxygen isotope measurements as a “natural experiment” to provide insights into the relationship between ε_bio and growth temperature. After a careful consideration of Sphagnum physiology relevant for understanding the degree of ¹⁸O enrichment in leaf water and the time span of precipitation input and plant growth, we demonstrated the temperature-dependence of ε_bio for Sphagnum mosses comparable to what was determined experimentally in heterotrophic generation of cellulose. Although the relationship between photosynthetic temperature for field-growing mosses and mean air temperature remains to be refined and perhaps explains the discrepancy in two regression equations, both approaches suggest that the factor of temperature becomes increasingly important to increase the ε_bio from the often assumed 27‰ when growth temperature is lower and thus should be considered in future paleoclimate and plant physiology studies wherever δ¹⁸O_c data are measured, in particular for trees in high-elevation areas. Additionally, but not discussed here, our data compilation further constrained the within-site variability in Sphagnum δ¹⁸O_c and implied that their δ¹⁸O_c variations in peat core might reflect past changes in evaporative enrichment possibly linked to peatland surface patterning rather than δ¹⁸O_p.

Data Availability Statement

The datasets presented in this study can be found in online Supplementary Material.

Author Contributions

ZX conceived the research, performed the laboratory and data analyses, and wrote the manuscript with input from ZY. ZY provided some samples and contributed to the discussion. Both authors contributed to the article and approved the submitted version.

Funding

This study was supported by United States National Science Foundation (ERA-1502891 and DEB-1802810) and National Natural Science Foundation of China (41877458).

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

We thank Yuwen Fu, Jingjing Sun, Zuo Wang, Yingfan Xia, and other for providing samples from China, and two reviewers for constructive comments.

Supplementary Material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/feart.2020.00307/full#supplementary-material

References

Aldous, A. R. (2002). Nitrogen translocation in Sphagnum mosses: effects of atmospheric nitrogen deposition. New Phytol. 156, 241–253. doi: 10.1046/j.1469-8137.2002.00518.x

CrossRef Full Text | Google Scholar

Amesbury, M. J., Charman, D. J., Newnham, R. M., Loader, N. J., Goodrich, J., Royles, J., et al. (2015). Can oxygen stable isotopes be used to track precipitation moisture source in vascular plant-dominated peatlands? Earth Planet. Sci. Lett. 430, 149–159. doi: 10.1016/j.epsl.2015.08.015

CrossRef Full Text | Google Scholar

Aravena, R., and Warner, B. G. (1992). Oxygen-18 composition of Sphagnum, and microenvironmental water relations. Bryologist 95, 445–448. doi: 10.2307/3243570

CrossRef Full Text | Google Scholar

Aucour, A. M., Hillaire-Marcel, C., and Bonnefille, R. (1996). Oxygen isotopes in cellulose from modern and quaternary intertropical peatbogs: implications for palaeohydrology. Chem. Geol. 129, 341–359. doi: 10.1016/0009-2541(95)00179-4

CrossRef Full Text | Google Scholar

Ballantyne, A. P., Greenwood, D. R., Sinninghe Damsté, J. S., Csank, A. Z., Eberle, J. J., and Rybczynski, N. (2010). Significantly warmer Arctic surface temperatures during the Pliocene indicated by multiple independent proxies. Geology 38, 603–606. doi: 10.1130/G30815.1

CrossRef Full Text | Google Scholar

Barbour, M. M. (2007). Stable oxygen isotope composition of plant tissue: a review. Funct. Plant Biol. 34, 83–94. doi: 10.1071/FP06228

CrossRef Full Text | Google Scholar

Barbour, M. M., and Farquhar, G. D. (2000). Relative humidity- and ABA-induced variation in carbon and oxygen isotope ratios of cotton leaves. Plant Cell Environ. 23, 473–485. doi: 10.1046/j.1365-3040.2000.00575.x

CrossRef Full Text | Google Scholar

Bowen, G. J. (2020). ). The Online Isotopes in Precipitation Calculator, version 3.1. Available: http://wateriso.utah.edu/waterisotopes/pages/data_access/oipc.html (accessed January 2020).

Google Scholar

Bowen, G. J., and Revenaugh, J. (2003). Interpolating the isotopic composition of modern meteoric precipitation. Water Resour. Res. 39:1299. doi: 10.1029/2003wr002086

CrossRef Full Text | Google Scholar

Bowen, G. J., Wassenaar, L. I., and Hobson, K. A. (2005). Global application of stable hydrogen and oxygen isotopes to wildlife forensics. Oecologia 143, 337–348. doi: 10.1007/s00442-004-1813-y

PubMed Abstract | CrossRef Full Text | Google Scholar

Brader, A. V. (2013). Precipitation in Peatlands: The Stable Isotope Record of Sphagnum Mosses as a Proxy for Environmental Change. Doctoral dissertation, Netherlands: Vrije Universiteit, Amsterdam.

Google Scholar

Brader, A. V., van Winden, J. F., Bohncke, S. J. P., Beets, C. J., Reichart, G.-J., and de Leeuw, J. W. (2010). Fractionation of hydrogen, oxygen and carbon isotopes in n-alkanes and cellulose of three Sphagnum species. Org. Geochem. 41, 1277–1284. doi: 10.1016/j.orggeochem.2010.09.006

CrossRef Full Text | Google Scholar

Breeuwer, A., Heijmans, M. M. P. D., Robroek, B. J. M., and Berendse, F. (2008). The effect of temperature on growth and competition between Sphagnum species. Oecologia 156, 155–167. doi: 10.1007/s00442-008-0963-8

PubMed Abstract | CrossRef Full Text | Google Scholar

Brenninkmeijer, C. A. M., van Geel, B., and Mook, W. G. (1982). Variations in the D/H and ¹⁸O/¹⁶O ratios in cellulose extracted from a peat bog core. Earth Planet. Sci. Lett. 61, 283–290. doi: 10.1016/0012-821X(82)90059-0

CrossRef Full Text | Google Scholar

Brunello, C. F., Andermann, C., Helle, G., Comiti, F., Tonon, G., Tiwari, A., et al. (2019). Hydroclimatic seasonality recorded by tree ring δ¹⁸O signature across a Himalayan altitudinal transect. Earth Planet. Sci. Lett. 518, 148–159. doi: 10.1016/j.epsl.2019.04.030

CrossRef Full Text | Google Scholar

Cernusak, L. A., Farquhar, G. D., and Pate, J. S. (2005). Environmental and physiological controls over oxygen and carbon isotope composition of Tasmanian blue gum, Eucalyptus globulus. Tree Physiol. 25, 129–146. doi: 10.1093/treephys/25.2.129

PubMed Abstract | CrossRef Full Text | Google Scholar

Charman, D. J., Beilman, D. W., Blaauw, M., Booth, R. K., Brewer, S., Chambers, F. M., et al. (2013). Climate-related changes in peatland carbon accumulation during the last millennium. Biogeosciences 10, 929–944. doi: 10.5194/bg-10-929-2013

CrossRef Full Text | Google Scholar

Clymo, R. S. (1970). The growth of Sphagnum: methods of measurement. J. Ecol. 58, 13–49. doi: 10.2307/2258168

CrossRef Full Text | Google Scholar

Clymo, R. S., and Hayward, P. M. (1982). “The Ecology of Sphagnum,” in Bryophyte Ecology, ed. A. J. E. Smith (Dordrecht: Springer Netherlands), 229–289. doi: 10.1007/978-94-009-5891-3_8

CrossRef Full Text | Google Scholar

Daley, T. J. (2007). Tracking Holocene Climate Change Using Peat Bog Stable Isotopes. Doctoral dissertation, University of Southampton, Southampton.

Google Scholar

Daley, T. J., Barber, K. E., Street-Perrott, F. A., Loader, N. J., Marshall, J. D., Crowley, S. F., et al. (2010). Holocene climate variability revealed by oxygen isotope analysis of Sphagnum cellulose from Walton Moss, northern England. Quat. Sci. Rev. 29, 1590–1601. doi: 10.1016/j.quascirev.2009.09.017

CrossRef Full Text | Google Scholar

Dee, S., Emile-Geay, J., Evans, M. N., Allam, A., Steig, E. J., and Thompson, D. M. (2015). PRYSM: an open-source framework for PRoxY System Modeling, with applications to oxygen-isotope systems. J. Adv. Model Earth Syst. 7, 1220–1247. doi: 10.1002/2015ms000447

CrossRef Full Text | Google Scholar

DeNiro, M. J., and Epstein, S. (1981). Isotopic composition of cellulose from aquatic organisms. Geochim. Cosmochim. Acta 45, 1885–1894. doi: 10.1016/0016-7037(81)90018-1

CrossRef Full Text | Google Scholar

Dilks, T. J. K., and Proctor, M. C. F. (1979). Photosynthesis, respiration and water content in bryophytes. New Phytol. 82, 97–114. doi: 10.1111/j.1469-8137.1979.tb07564.x

CrossRef Full Text | Google Scholar

Dorrepaal, E., Aerts, R., Cornelissen, J. H. C., Callaghan, T. V., and Van Logtestijn, R. S. P. (2004). Summer warming and increased winter snow cover affect Sphagnum fuscum growth, structure and production in a sub-arctic bog. Glob. Change Biol. 10, 93–104. doi: 10.1111/j.1365-2486.2003.00718.x

CrossRef Full Text | Google Scholar

Edwards, T. W. D., and McAndrews, J. H. (1989). Paleohydrology of a Canadian Shield lake inferred from ¹⁸O in sediment cellulose. Can. J. Earth. Sci. 26, 1850–1859. doi: 10.1139/e89-158

CrossRef Full Text | Google Scholar

El Bilali, H., and Patterson, R. T. (2012). Influence of cellulose oxygen isotope variability in sub-fossil Sphagnum and plant macrofossil components on the reliability of paleoclimate records at the Mer Bleue Bog. Ottawa, Ontario, Canada. Org. Geochem. 43, 39–49. doi: 10.1016/j.orggeochem.2011.11.003

CrossRef Full Text | Google Scholar

Ellsworth, P. V., and Sternberg, L. S. L. (2014). Biochemical effects of salinity on oxygen isotope fractionation during cellulose synthesis. New Phytol. 202, 784–789. doi: 10.1111/nph.12696

PubMed Abstract | CrossRef Full Text | Google Scholar

Feng, X., Zhao, C., D’Andrea, W. J., Liang, J., Zhou, A., and Shen, J. (2019). Temperature fluctuations during the Common Era in subtropical southwestern China inferred from brGDGTs in a remote alpine lake. Earth Planet. Sci. Lett. 510, 26–36. doi: 10.1016/j.epsl.2018.12.028

CrossRef Full Text | Google Scholar

Finsinger, W., Schoning, K., Hicks, S., Lücke, A., Goslar, T., Wagner-cremer, F., et al. (2013). Climate change during the past 1000 years: a high-temporal-resolution multiproxy record from a mire in northern Finland. J. Quatern. Sci. 28, 152–164. doi: 10.1002/jqs.2598

CrossRef Full Text | Google Scholar

Gallego-Sala, A. V., Clark, J. M., House, J. I., Orr, H. G., Prentice, I. C., Smith, P., et al. (2010). Bioclimatic envelope model of climate change impacts on blanket peatland distribution in Great Britain. Clim. Res. 45, 151–162. doi: 10.3354/cr00911

CrossRef Full Text | Google Scholar

Genet, H., Oberbauer, S. F., Colby, S. J., Staudhammer, C. L., and Starr, G. (2013). Growth responses of Sphagnum hollows to a growing season lengthening manipulation in Alaskan Arctic tundra. Polar Biol. 36, 41–50. doi: 10.1007/s00300-012-1236-x

CrossRef Full Text | Google Scholar

GMAO. (2015a). MERRA-2 Const_2d_asm_Nx: 2d, Constants V5.12.4. Greenbelt, MD: GES DISC.

Google Scholar

GMAO (2015b). MERRA-2 instM_2d_asm_Nx: 2d, Monthly Mean, Single-Level, Assimilation, Single-Level Diagnostics V5.12.4. Greenbelt, MD: GES DISC.

Google Scholar

GMAO (2015c). MERRA-2 tavgM_2d_flx_Nx: 2d, Monthly Mean, Time-Averaged, Single-Level, Assimilation, Surface Flux Diagnostics V5.12.4. Greenbelt, MD: GES DISC.

Google Scholar

Granath, G., Rydin, H., Baltzer, J. L., Bengtsson, F., Boncek, N., Bragazza, L., et al. (2018). Environmental and taxonomic controls of carbon and oxygen stable isotope composition in Sphagnum across broad climatic and geographic ranges. Biogeosciences 15, 5189–5202. doi: 10.5194/bg-15-5189-2018

CrossRef Full Text | Google Scholar

Gunnarsson, U. (2005). Global patterns of Sphagnum productivity. J. Bryol. 27, 269–279. doi: 10.1179/174328205X70029

CrossRef Full Text | Google Scholar

Helliker, B. R., and Ehleringer, J. R. (2002). Differential ¹⁸O enrichment of leaf cellulose in C₃ versus C₄ grasses. Funct. Plant Biol. 29, 435–442. doi: 10.1071/PP01122

CrossRef Full Text | Google Scholar

Helliker, B. R., and Richter, S. L. (2008). Subtropical to boreal convergence of tree-leaf temperatures. Nature 454, 511–514. doi: 10.1038/nature07031

PubMed Abstract | CrossRef Full Text | Google Scholar

Helliker, B. R., Song, X., Goulden, M. L., Clark, K., Bolstad, P., Munger, J. W., et al. (2018). Assessing the interplay between canopy energy balance and photosynthesis with cellulose δ¹⁸O: large-scale patterns and independent ground-truthing. Oecologia 187, 995–1007. doi: 10.1007/s00442-018-4198-z

PubMed Abstract | CrossRef Full Text | Google Scholar

Hong, B., Liu, C., Lin, Q., Yasuyuki, S., Leng, X., Wang, Y., et al. (2009). Temperature evolution from the δ¹⁸O record of Hani peat, Northeast China, in the last 14000 years. Sci. China Earth Sci. 52, 952–964. doi: 10.1007/s11430-009-0086-z

CrossRef Full Text | Google Scholar

Hong, Y. T., Jiang, H. B., Liu, T. S., Zhou, L. P., Beer, J., Li, H. D., et al. (2000). Response of climate to solar forcing recorded in a 6000-year δ¹⁸O time-series of Chinese peat cellulose. Holocene 10, 1–7. doi: 10.1191/095968300669856361

CrossRef Full Text | Google Scholar

Jones, M. C., Anderson, L., Keller, K., Nash, B., Littell, V., Wooller, M., et al. (2019). An assessment of plant species differences on cellulose oxygen isotopes from two Kenai Peninsula, Alaska peatlands: implications for hydroclimatic reconstructions. Front. Earth Sci. 7:25. doi: 10.3389/feart.2019.00025

CrossRef Full Text | Google Scholar

Kahmen, A., Sachse, D., Arndt, S. K., Tu, K. P., Farrington, H., Vitousek, P. M., et al. (2011). Cellulose δ¹⁸O is an index of leaf-to-air vapor pressure difference (VPD) in tropical plants. Proc. Natl. Acad. Sci. U.S.A. 108, 1981–1986. doi: 10.1073/pnas.1018906108

PubMed Abstract | CrossRef Full Text | Google Scholar

Kaislahti Tillman, P., Holzkämper, S., Kuhry, P., Sannel, A. B. K., Loader, N. J., and Robertson, I. (2010). Stable carbon and oxygen isotopes in Sphagnum fuscum peat from subarctic Canada: implications for palaeoclimate studies. Chem. Geol. 270, 216–226. doi: 10.1016/j.chemgeo.2009.12.001

CrossRef Full Text | Google Scholar

Keel, S. G., Joos, F., Spahni, R., Saurer, M., Weigt, R. B., and Klesse, S. (2016). Simulating oxygen isotope ratios in tree ring cellulose using a dynamic global vegetation model. Biogeosciences 13, 3869–3886. doi: 10.5194/bg-13-3869-2016

CrossRef Full Text | Google Scholar

Krebs, M., Gaudig, G., and Joosten, H. (2016). Record growth of Sphagnum papillosum in Georgia (Transcaucasus): rain frequency, temperature and microhabitat as key drivers in natural bogs. Mires Peat 18, 1–16. doi: 10.19189/MaP.2015.OMB.190

CrossRef Full Text | Google Scholar

Küttim, M., Küttim, L., Ilomets, M., and Laine, A. M. (2020). Controls of Sphagnum growth and the role of winter. Ecol. Res. 35, 219–234. doi: 10.1111/1440-1703.12074

CrossRef Full Text | Google Scholar

Libby, L. M., Pandolfi, L. J., Payton, P. H., Marshall, J., Becker, B., and Giertz-Sienbenlist, V. (1976). Isotopic tree thermometers. Nature 261, 284–288. doi: 10.1038/261284a0

CrossRef Full Text | Google Scholar

Lin, Q., Hong, Y., Zhu, Y., Dong, L., Wang, Y., Leng, X., et al. (2001). The carbon and oxygen isotope composition of modern plants from typical peat bogs in China and its significance on the palaeoclimatic study. Bull. Mineral. Petrol. Geochem. 20, 93–97.

Google Scholar

Lindholm, T. (1990). Growth dynamics of the peat moss Sphagnum fuscum on hummocks on a raised bog in southern Finland. Ann. Bot. Fenn. 27, 67–78.

Google Scholar

Liu, H. T., Schäufele, R., Gong, X. Y., and Schnyder, H. (2017). The δ¹⁸O and δ²H of water in the leaf growth-and-differentiation zone of grasses is close to source water in both humid and dry atmospheres. New Phytol. 214, 1423–1431. doi: 10.1111/nph.14549

PubMed Abstract | CrossRef Full Text | Google Scholar

Liu, J. (2017). The Oxygen Isotope Composition of Modern Plants in Dajiuhu Peat Bog and its Climate Significance. Master’s thesis, Nanjing Normal University, Nanjing.

Google Scholar

Liu, Z., Zhu, J., Rosenthal, Y., Zhang, X., Otto-Bliesner, B. L., Timmermann, A., et al. (2014). The Holocene temperature conundrum. Proc. Natl. Acad. Sci. U.S.A. 111, E3501–E3505. doi: 10.1073/pnas.1407229111

PubMed Abstract | CrossRef Full Text | Google Scholar

Loader, N. J., Street-Perrott, F. A., Mauquoy, D., Roland, T. P., van Bellen, S., Daley, T. J., et al. (2016). Measurements of hydrogen, oxygen and carbon isotope variability in Sphagnum moss along a micro-topographical gradient in a southern Patagonian peatland. J. Quatern. Sci. 31, 426–435. doi: 10.1002/jqs.2871

CrossRef Full Text | Google Scholar

Loisel, J., Gallego-Sala, A. V., and Yu, Z. (2012). Global-scale pattern of peatland Sphagnum growth driven by photosynthetically active radiation and growing season length. Biogeosciences 9, 2737–2746. doi: 10.5194/bg-9-2737-2012

CrossRef Full Text | Google Scholar

Loisel, J., Garneau, M., and Hélie, J. F. (2009). Modern Sphagnum δ¹³C signatures follow a surface moisture gradient in two boreal peat bogs, James Bay lowlands, Québec. J. Quatern. Sci. 24, 209–214. doi: 10.1002/jqs.1221

CrossRef Full Text | Google Scholar

Loisel, J., and Yu, Z. (2013). Surface vegetation patterning controls carbon accumulation in peatlands. Geophys. Res. Lett. 40, 5508–5513. doi: 10.1002/grl.50744

CrossRef Full Text | Google Scholar

Managave, S. R., Sheshshayee, M. S., Borgaonkar, H. P., and Ramesh, R. (2010). Past break-monsoon conditions detectable by high resolution intra-annual δ¹⁸O analysis of teak rings. Geophys. Res. Lett. 37:L05702. doi: 10.1029/2009GL041172

CrossRef Full Text | Google Scholar

McCarter, C. P. R., and Price, J. S. (2014). Ecohydrology of Sphagnum moss hummocks: mechanisms of capitula water supply and simulated effects of evaporation. Ecohydrology 7, 33–44. doi: 10.1002/eco.1313

CrossRef Full Text | Google Scholar

Ménot-Combes, G., Burns, S. J., and Leuenberger, M. (2002). Variations of ¹⁸O/¹⁶O in plants from temperate peat bogs (Switzerland): implications for paleoclimatic studies. Earth Planet. Sci. Lett. 202, 419–434. doi: 10.1016/S0012-821X(02)00794-X

CrossRef Full Text | Google Scholar

Mora, G., and Zanazzi, A. (2017). Hydrogen isotope ratios of moss cellulose and source water in wetlands of Lake Superior, United States reveal their potential for quantitative paleoclimatic reconstructions. Chem. Geol. 468, 75–83. doi: 10.1016/j.chemgeo.2017.08.017

CrossRef Full Text | Google Scholar

Moschen, R., Kühl, N., Rehberger, I., and Lücke, A. (2009). Stable carbon and oxygen isotopes in sub-fossil Sphagnum: assessment of their applicability for palaeoclimatology. Chem. Geol. 259, 262–272. doi: 10.1016/j.chemgeo.2008.11.009

CrossRef Full Text | Google Scholar

Murray, K. J., Harley, P. C., Beyers, J., Walz, H., and Tenhunen, J. D. (1989). Water content effects on photosynthetic response of Sphagnum mosses from the foothills of the Philip Smith Mountains, Alaska. Oecologia 79, 244–250. doi: 10.1007/bf00388484

PubMed Abstract | CrossRef Full Text | Google Scholar

Okazaki, A., and Yoshimura, K. (2019). Global evaluation of proxy system models for stable water isotopes with realistic atmospheric forcing. J. Geophys. Res. Atmos. 124, 8972–8993. doi: 10.1029/2018jd029463

CrossRef Full Text | Google Scholar

Ooki, S., Akagi, T., Jinno, H., Franzén, L. G., and Newton, J. (2018). Hydrological study of Lyngmossen bog, Sweden: isotopic tracers (³H, δ²H and δ¹⁸O) imply three waters with different mobilities. Quat. Sci. Rev. 199, 97–107. doi: 10.1016/j.quascirev.2018.09.014

CrossRef Full Text | Google Scholar

Price, J. S., Edwards, T. W. D., Yi, Y., and Whittington, P. N. (2009). Physical and isotopic characterization of evaporation from Sphagnum moss. J. Hydrol. 369, 175–182. doi: 10.1016/j.jhydrol.2009.02.044

CrossRef Full Text | Google Scholar

Richter, S. L., Johnson, A. H., Dranoff, M. M., and Taylor, K. D. (2008). Continental-scale patterns in modern wood cellulose δ¹⁸O: implications for interpreting paleo-wood cellulose δ¹⁸O. Geochim. Cosmochim. Acta 72, 2735–2743. doi: 10.1016/j.gca.2008.01.030

CrossRef Full Text | Google Scholar

Risi, C., Noone, D., Worden, J., Frankenberg, C., Stiller, G., Kiefer, M., et al. (2012). Process-evaluation of tropospheric humidity simulated by general circulation models using water vapor isotopologues: 1. Comparison between models and observations. J. Geophys. Res. Atmos. 117:D05303. doi: 10.1029/2011jd016621

CrossRef Full Text | Google Scholar

Roden, J. S., and Ehleringer, J. R. (2000). There is no temperature dependence of net biochemical fractionation of hydrogen and oxygen isotopes in tree-ring cellulose. Isot. Environ. Healt. S. 36, 303–317. doi: 10.1080/10256010008036389

PubMed Abstract | CrossRef Full Text | Google Scholar

Roden, J. S., Lin, G., and Ehleringer, J. R. (2000). A mechanistic model for interpretation of hydrogen and oxygen isotope ratios in tree-ring cellulose. Geochim. Cosmochim. Acta 64, 21–35. doi: 10.1016/S0016-7037(99)00195-7

CrossRef Full Text | Google Scholar

Royles, J., Amesbury, M. J., Roland, T. P., Jones, G. D., Convey, P., Griffiths, H., et al. (2016). Moss stable isotopes (carbon-13, oxygen-18) and testate amoebae reflect environmental inputs and microclimate along a latitudinal gradient on the Antarctic Peninsula. Oecologia 181, 931–945. doi: 10.1007/s00442-016-3608-3

PubMed Abstract | CrossRef Full Text | Google Scholar

Royles, J., Sime, L. C., Hodgson, D. A., Convey, P., and Griffiths, H. (2013). Differing source water inputs, moderated by evaporative enrichment, determine the contrasting δ¹⁸O_CELLULOSE signals in maritime Antarctic moss peat banks. J. Geophys. Res. Biogeo. 118, 184–194. doi: 10.1002/jgrg.20021

CrossRef Full Text | Google Scholar

Rydin, H., and Jeglum, J. K. (2013). The Biology of Peatlands. Oxford: Oxford University Press.

Google Scholar

Sargeant, C. I., and Singer, M. B. (2016). Sub-annual variability in historical water source use by Mediterranean riparian trees. Ecohydrology 9, 1328–1345. doi: 10.1002/eco.1730

CrossRef Full Text | Google Scholar

Schubert, B. A., and Jahren, A. H. (2015). Seasonal temperature and precipitation recorded in the intra-annual oxygen isotope pattern of meteoric water and tree-ring cellulose. Quat. Sci. Rev. 125, 1–14. doi: 10.1016/j.quascirev.2015.07.024

CrossRef Full Text | Google Scholar

Sonesson, M., Persson, S., Basilier, K., and Stenström, T. A. (1980). Growth of Sphagnum riparium Ångstr. in relation to some environmental factors in the Stordalen mire. Ecol. Bull. 30, 191–207.

Google Scholar

Song, X., Barbour, M. M., Saurer, M., and Helliker, B. R. (2011). Examining the large-scale convergence of photosynthesis-weighted tree leaf temperatures through stable oxygen isotope analysis of multiple data sets. New Phytol. 192, 912–924. doi: 10.1111/j.1469-8137.2011.03851.x

PubMed Abstract | CrossRef Full Text | Google Scholar

Song, X., Clark, K. S., and Helliker, B. R. (2014a). Interpreting species-specific variation in tree-ring oxygen isotope ratios among three temperate forest trees. Plant Cell Environ. 37, 2169–2182. doi: 10.1111/pce.12317

PubMed Abstract | CrossRef Full Text | Google Scholar

Song, X., Farquhar, G. D., Gessler, A., and Barbour, M. M. (2014b). Turnover time of the non-structural carbohydrate pool influences δ¹⁸O of leaf cellulose. Plant Cell Environ. 37, 2500–2507. doi: 10.1111/pce.12309

PubMed Abstract | CrossRef Full Text | Google Scholar

Sternberg, L., and Ellsworth, P. F. V. (2011). Divergent biochemical fractionation, not convergent temperature, explains cellulose oxygen isotope enrichment across latitudes. PLoS One 6:e28040. doi: 10.1371/journal.pone.0028040

PubMed Abstract | CrossRef Full Text | Google Scholar

Sternberg, L. D. S., and DeNiro, M. J. D. (1983). Biogeochemical implications of the isotopic equilibrium fractionation factor between the oxygen atoms of acetone and water. Geochim. Cosmochim. Acta 47, 2271–2274. doi: 10.1016/0016-7037(83)90049-2

CrossRef Full Text | Google Scholar

Sternberg, L. D. S. L. (2014). Comment on “Oxygen isotope ratios (¹⁸O/¹⁶O) of hemicellulose-derived sugar biomarkers in plants, soils and sediments as paleoclimate proxy I: insight from a climate chamber experiment” by Zech et al. (2014). Geochim. Cosmochim. Acta 141, 677–679. doi: 10.1016/j.gca.2014.04.051

CrossRef Full Text | Google Scholar

Sternberg, L. D. S. L., Deniro, M. J., and Savidge, R. A. (1986). Oxygen isotope exchange between metabolites and water during biochemical reactions leading to cellulose synthesis. Plant Physiol. 82, 423–427. doi: 10.1104/pp.82.2.423

PubMed Abstract | CrossRef Full Text | Google Scholar

Sternberg, L. S. L. (2009). Oxygen stable isotope ratios of tree-ring cellulose: the next phase of understanding. New Phytol. 181, 553–562. doi: 10.1111/j.1469-8137.2008.02661.x

PubMed Abstract | CrossRef Full Text | Google Scholar

Taylor, M. (2008). Continental-Scale Validation of the Temperature Signal in Oxygen Isotopes of Sphagnum Cellulose and Its Application as a Paleoclimate Proxy. Master’s thesis, University of Wyoming, Laramie, WY.

Google Scholar

Treydte, K. S., Schleser, G. H., Helle, G., Frank, D. C., Winiger, M., Haug, G. H., et al. (2006). The twentieth century was the wettest period in northern Pakistan over the past millennium. Nature 440, 1179–1182. doi: 10.1038/nature04743

PubMed Abstract | CrossRef Full Text | Google Scholar

Turetsky, M. R., Crow, S. E., Evans, R. J., Vitt, D. H., and Wieder, R. K. (2008). Trade-offs in resource allocation among moss species control decomposition in boreal peatlands. J. Ecol. 96, 1297–1305. doi: 10.1111/j.1365-2745.2008.01438.x

CrossRef Full Text | Google Scholar

Tuthorn, M., Zech, M., Ruppenthal, M., Oelmann, Y., Kahmen, A., Valle, H. F. D., et al. (2014). Oxygen isotope ratios (¹⁸O/¹⁶O) of hemicellulose-derived sugar biomarkers in plants, soils and sediments as paleoclimate proxy II: insight from a climate transect study. Geochim. Cosmochim. Acta 126, 624–634. doi: 10.1016/j.gca.2013.11.002

CrossRef Full Text | Google Scholar

Waterhouse, J. S., Cheng, S., Juchelka, D., Loader, N. J., McCarroll, D., Switsur, V. R., et al. (2013). Position-specific measurement of oxygen isotope ratios in cellulose: isotopic exchange during heterotrophic cellulose synthesis. Geochim. Cosmochim. Acta 112, 178–191. doi: 10.1016/j.gca.2013.02.021

CrossRef Full Text | Google Scholar

Xia, Z., Zheng, Y., Stelling, J. M., Loisel, J., Huang, Y., and Yu, Z. (2020). Environmental controls on the carbon and water (H and O) isotopes in peatland Sphagnum mosses. Geochim. Cosmochim. Acta 277, 265–284. doi: 10.1016/j.gca.2020.03.034

CrossRef Full Text | Google Scholar

Yakir, D., and DeNiro, M. J. (1990). Oxygen and hydrogen isotope fractionation during cellulose metabolism in Lemna gibba L. Plant Physiol. 93, 325–332. doi: 10.1104/pp.93.1.325

PubMed Abstract | CrossRef Full Text | Google Scholar

Yu, Z., Loisel, J., Brosseau, D. P., Beilman, D. W., and Hunt, S. J. (2010). Global peatland dynamics since the Last Glacial Maximum. Geophys. Res. Lett. 37:L13402. doi: 10.1029/2010GL043584

CrossRef Full Text | Google Scholar

Zanazzi, A., and Mora, G. (2005). Paleoclimatic implications of the relationship between oxygen isotope ratios of moss cellulose and source water in wetlands of Lake Superior. Chem. Geol. 222, 281–291. doi: 10.1016/j.chemgeo.2005.08.006

CrossRef Full Text | Google Scholar

Zech, M., Mayr, C., Tuthorn, M., Leiber-Sauheitl, K., and Glaser, B. (2014a). Oxygen isotope ratios (¹⁸O/¹⁶O) of hemicellulose-derived sugar biomarkers in plants, soils and sediments as paleoclimate proxy I: insight from a climate chamber experiment. Geochim. Cosmochim. Acta 126, 614–623. doi: 10.1016/j.gca.2013.10.048

CrossRef Full Text | Google Scholar

Zech, M., Mayr, C., Tuthorn, M., Leiber-Sauheitl, K., and Glaser, B. (2014b). Reply to the comment of Sternberg on “Zech et al. (2014) Oxygen isotope ratios (¹⁸O/¹⁶O) of hemicellulose-derived sugar biomarkers in plants, soils and sediments as paleoclimate proxy I: insight from a climate chamber experiment. GCA 126, 614–623.”. Geochim. Cosmochim. Acta 141, 680–682. doi: 10.1016/j.gca.2014.05.048

CrossRef Full Text | Google Scholar

Zeng, X., Liu, X., Evans, M. N., Wang, W., An, W., Xu, G., et al. (2016). Seasonal incursion of Indian Monsoon humidity and precipitation into the southeastern Qinghai–Tibetan plateau inferred from tree ring δ¹⁸O values with intra-seasonal resolution. Earth Planet. Sci. Lett. 443, 9–19. doi: 10.1016/j.epsl.2016.03.011

CrossRef Full Text | Google Scholar