Impact of Increasing CO₂, and Air Pollutants (NO_x, SO₂, O₃) on the Stable Isotope Ratios in Tree Rings

Abstract

Anthropogenic activities such as industrialization, land use change and intensification of agriculture strongly contribute to changes in the concentrations of atmospheric trace gases. Carbon dioxide (CO₂), oxidized N compounds (NOx), sulfur dioxide (SO₂) and ozone (O₃) have particularly significant impacts on plant physiology. CO₂, the substrate for plant photosynthesis, is in the focus of interest as the ambiguous effect of its increasing concentration is controversially discussed. Is its increase beneficial for plants or are plants non-responsive? NOx, a product of combustion and lightning, can have either fertilizing or toxic effects depending on the concentration and form. This is also the case for reduced forms of nitrogen (NHy), which are mostly emitted from agricultural and industrial activities. In combination CO₂ and N compounds can have a fertilizing effect. SO₂ and ground-level O₃ are mostly phytotoxic, depending on their concentrations, daily and seasonal exposure dynamics, and tree health condition. Elevated concentrations of both substances arise from industrial combustion processes and car emissions. All of the above-mentioned gaseous compounds affect plant metabolism in their specific ways and to different degrees. This impacts the isotope fractionation leaving specific fingerprints in the C, O, (H) and N isotope ratios of organic matter. In this chapter we will show how the impact of increasing CO₂ and air pollutants are reflected in the isotopic ratios of tree rings. Increasing CO₂ shows a considerable variation in responses of δ¹³C and to a minor degree in δ¹⁸O. Ozone and SO₂ exposure cause an overall increase of the δ¹³C values in tree rings and a slight decrease in δ¹⁸O, mimicking an increase in net photosynthesis (A_N) and to a minor degree in stomatal conductance (g_s). However, directly measured A_N and g_s values show the opposite, which does not always correspond with the isotope derived gas exchange data. NO₂ concentration as it is found near highly frequented freeways or industrial plants causes an increase of δ¹³C while δ¹⁸O decreases. This indicates an increase in both A_N and g_s, which corresponds well with directly measured gas exchange data. Thus the air quality situation must be taken in consideration for the interpretation of isotope values in tree rings.

1 Introduction

Stable C and O isotope ratios are indicative for changes at the leaf level CO₂ and H₂O exchange (Chap. 9; Farquhar et al. 1989; Chap. 10; Farquhar and Lloyd 1993). Via assimilates, the specific isotope signals are stored in tree rings making them an ideal recorder for information on past environmental changes. The goal of this chapter is to highlight the impact of changes in atmospheric CO₂ and air pollutants on stable isotope ratios in tree rings. Herein, we describe the pathways by which gaseous compounds are incorporated into plants with an emphasis on uptake via leaves through stomata. Impacts on tree physiology and potential phytotoxicity are discussed with respect to effects on isotopic fractionation that can lead to detectable isotopic variation in tree-rings. Overall, this chapter provides a broad review of the literature that demonstrates the effect of CO₂ and pollutants on tree-ring isotopes and how these influences need to be considered for the interpretation of stable isotopes in tree rings.

1.1 Reactive Pollutants and Trees

Reactive gaseous compounds, such as O_3, SO₂ and NOx, which are mainly products of high temperature combustion processes (e.g., ore smelters, oil refineries, chemical industrial plants, coal-fired power plants, and combustion driven mobility) were identified early on as the most relevant phytotoxic air pollutants (Ashmore 2005; Wellburn 1990; Martin and Sutherland 1990; Frank et al. 1979, Choi and Lee 2012). Over recent decades, SO₂ emissions have been substantially reduced through air quality regulations in most Western countries. Such a decrease for airborne N compounds is observed since the mid 1980s, with the introduction of catalyzers for car exhausts (Geng et al. 2014; Etzold et al. 2020). While a fraction of the N- and S-compounds are deposited on the soil surface and absorbed by microorganisms and roots (their ambiguous effects are described below), the gaseous forms are taken up via stomata through diffusional processes and directly metabolized. Before stomatal uptake, a portion of the reactive pollutant species are deposited on branch, leaf and soil surfaces (Fig. 24.1), which makes the quantification of leaf uptake rather difficult. Where possible, isotopic tracer studies (¹⁵N, ³⁴S) in the field (Ammann et al. 1999, Saurer et al. 2004) and in the laboratory (Chaparro-Suarez et al. 2011; Vallano and Sparks 2013) can lead to more realistic flux estimates including eddy covariance measurements. A broad overview about the impact of air pollutants on physiological responses as reflected in stable isotopes is given in Savard (2010).

Fig. 24.1

Pathways for the gaseous exchange between atmosphere and plant canopies and leaves (After Larcher, 2003). The arrows indicate the fluxes of the respective gas species. Diffusive resistances are shown as brown rectangles: r_aerodyn → Aerodynamic resistance; r_diff → diffusive resistance; r_canopy → diffusive resistances as the molecules diffuse though the canopy; r_soil → diffusive resistance as the molecules diffuse into or out of the soil; r_b → boundary layer resistance; r_s → physiologically controllable stomatal resistance; r_i → resistances in the intercellular system; r_w → cell wall resistance to CO₂ includes the transition of CO₂ in the liquid phase; r_mm → mitochondrial membrane resistance. r_c → chloroplast resistance to CO₂, includes the resistance of the chloroplast envelope and stroma. Mesophyll resistance is defined as r_m =r_w +r_c; CO₂ concentrations are shown for the ambient air (c_a), inside the leaf intercellular spaces (c_i), in the mesophyll cytosol (c_m) and in the chloroplast (c_c)

In the 1970’s and later, increasing concentrations of air pollutants were causing significant losses in agricultural crop yields and were regionally identified as causing significant forest decline (McLaughlin 1985; Heck et al. 1988; Slovik et al. 1996; Bytnerowicz and Fenn 1996; Muzika et al. (2004); Ashmore 2005; Matyssek et al. 2010, 2012). Climatic variation in combination with air pollutants can exacerbate the impact of stress modulating canopy-integrated leaf gas exchange, and thus tree growth. Indeed, one of the most toxic pollutants besides SO₂ (Darral 1989; Lee et al. 2017) is O₃ (Cailleret et al. 2018; Jolivet et al. 2016). Whereas SO₂ pollution is not as globally distributed as tropospheric O₃, and is mainly a regional problem near the vicinity of smelters and coal driven energy production, ozone has greater areal coverage as well as more spatial and temporal variation (Lefohn et al. 2018; Mills et al. 2018; Tarasick et al. 2019). In particular, photochemical reactions producing O₃ are enhanced during warm and sunny days when nitrogen oxides and volatile organic compounds are present (Royal Society 2008). Weather situations favoring ozone formation are often associated with dry and warm conditions that lead to low stomatal conductance (g_s) or stomatal closure, which reduces uptake and hence toxic dose effects. Thus, less injury will occur when stomata are closed even if O₃ or SO₂ concentrations are high. In contrast, conditions characterized by high air pollution concentrations and high g_s occur when the air is humid leading to higher uptake, which will exacerbate the effects of these pollutants (Jolivet et al. 2016). These include sluggish stomata and lower g_s. Sluggish stomata become unresponsive to sudden changes in environmental cues that affect stomatal aperture, such as light or humidity or even pollution levels. Generally, plants respond to the exposure of air pollutants and CO₂ with changes in g_s, net photosynthesis (A_N), and biochemical processes (e.g., Savard et al. 2020a), which reflect how O₃ and SO₂ can directly damage cells via oxidation of lipids in the plasma membrane (Jolivet et al. 2016; Duan et al. 2019) and impairment of other plant biochemical processes (decrease in Rubisco and potential increase in PEP carboxylase. Ainsworth et al. 2012).

While O₃ and SO₂ are predominantly phytotoxic, N compounds can be either beneficial or detrimental (Etzold et al. 2020), depending on their concentration and form. The ultimate impact of N can be dependent on its availability in the ecosystem. In N-limited forests, N deposition modifies the soil biogeochemical dynamics and can increase tree growth (Högberg 2007; Thomas et al. 2015), whereas in high N- systems, additional N-deposition can lead to soil acidification, and excessive foliar and root N, which can foster tree mortality (Aber et al. 1989; Vitousek et al. 1997; Galloway et al. 2004). The combined increase in CO₂ and N-deposition can often enhance C uptake and can result in an enhanced tree and forest growth (Norby 1998; Gruber and Galloway 2008; Etzold et al. 2020). However, where stem growth was either reduced or unchanged, access to nutrients was found to be restricted or the air contained other pollutants that were potentially phytotoxic (Giguère-Croteau et al. 2019; Savard et al. 2020a).

Air pollutants generally cause direct damages on the cell wall structure and impair biochemical processes (Jolivet et al. 2016), while CO₂ and N-compounds can be beneficial through various pathways, from the roots to the mesophyll (Fowler et al. 1998; 2001) (Fig. 24.1).

1.2 General Effects of Changes in the Atmosphere on Stable C, O and N Isotope Ratios

As shown in Chapters 9, 10 and 19 any environmental impact that alters the source isotopic signals or the foliar gas exchange of CO₂ and H₂O will result in specific changes of the stable C and O isotope ratios (δ¹³C and δ¹⁸O values) in plant organic matter. An increase in CO₂ concentration enhances photosynthesis and can cause a reduction in g_s. However, exposure to pollutants such as O₃ and SO₂ also induces stomatal closure, but in contrast is often accompanied by a concomitant decrease in net photosynthesis (A_N) (Matyssek et al. 1997; Linzon 1972; Ziegler 1972; Parry and Gutteridge 1984; Martin et al. 1988; Wedler et al. 1995). Exposure to NOx sometimes leads to an increase in both A_N and g_s (Gessler et al. 2000; Siegwolf et al. 2001). Changes in A_N and g_s alter the ratio of internal (leaf intercellular) [CO₂] (c_i) to external [CO₂] (c_a), i.e., c_i/c_a. When A_N increases without a change in g_s, c_i/c_a decreases. Since ¹²CO₂ is preferably assimilated relative to ¹³CO₂ the partial pressure of ¹²CO₂ decreases to a larger proportion than ¹³CO₂ augmenting the ratio of ¹³CO₂ / ¹²CO₂ in the favor of ¹³C. As a consequence the uptake of ¹³CO₂ rises, resulting in an increase of the δ¹³C values. Under conditions where there is an increase in g_s at a constant A_N, or a decrease in A_N at a constant g_s, the c_i/c_a will increase, leading to a decrease in δ¹³C values (Farquhar et al. 1989; Chap. 9). Chapter 10 describes the formation of δ¹⁸O values in plant organic matter, which is given by the δ¹⁸O ratio of source water (water before it enters the leaves, see Chaps. 10 and 18) and subsequently by the δ¹⁸O of leaf water, which is modified by the variability of g_s and water vapor in the air. The role of g_s is reflected in an inverse proportional relationship to δ¹⁸O values (an increase in g_s results in a decrease in δ¹⁸O and vice versa, Barbour et al. 2004; see also Chap. 10). As the ¹⁸O/¹⁶O ratio is independent of A_N, the combination of δ¹³C and δ¹⁸O from the same sample allows for the distinction of whether δ¹³C changes are a result of changes in A_N or g_s (Scheidegger et al. 2000). Here the application of the C and O dual isotope approach can be very useful for data interpretation (Chap. 16) and facilitates a more detailed analysis of the physiological plant responses to variable impacts of air pollutants, assuming that only intrinsic reactions (i.e., non-responsive to air pollutants) control the foliar isotopic signals. The specific isotopic ratio, formed in the leaves and imprinted on assimilates, are then transferred to stems (tree rings) and roots forming organic material (see Chap. 13).

Next we discuss the specific impact of gaseous N compounds taken up via stomata. Chapter 12 (δ¹⁵N values) and 23 (fertilization effects) provide a general discussion about N-absorption via roots and leaves. Various quantities for stomatal uptake of N compounds are provided in literature, i.e. values up to 30% of the total plant N uptake (Amman et al. 1999). In most cases, the N isotopic (δ¹⁵N) signature of the gaseous forms is reflected in the foliar organic matter (amino acids) and the N isotopic fractionation is low. However, the δ¹⁵N values of soil N compounds can depart largely from what is assimilated by the roots (see Chap. 12). An overview on the various N-fractionations during N incorporation in leaves is given in Tcherkez (2011) and Craine et al. (2015).

As the nitrogen content in tree rings is low and varies (on average 0.04–0.12%), the isotopic analysis of δ¹⁵N values in wood is quite challenging (Savard et al. 2020b). Where possible, the analysis should be done on single rings. In cases with very little material, samples can be pooled. Another challenge is the mobility of N compounds across several rings (Tomlinson et al. 2014). Yet numerous studies have made use of the δ¹⁵N values in tree rings and of the atmospheric sources to calculate the N incorporation of specific anthropogenic sources that have an isotopic signature distinct from natural δ¹⁵N values (Elhani et al. 2003; Amman et al. 1999; Saurer et al. 2004, Guerrieri et al. 2009, see below).

2 Sampling, Sample Preparation, Isotope Analysis and Flux Determination of Gaseous Pollutants

2.1 Sample Processing

In the following, we refer to Chaps. 4, 5 and 12 with regard to sampling and sample preparation. For the analysis of δ¹⁵N values in tree rings, whole wood is collected and milled the same way as for the ¹³C and ¹⁸O analyses. However, procedures for sample preparation to determine the ¹⁵N/¹⁴N ratio in wood are still under debate, in particular, the question of whether the mobile fraction of N compounds in tree rings should be removed. While some authors recommend removal to prevent false trends, other studies have shown no effect on the N concentration and δ¹⁵N values (Doucet et al. 2011; Tomlinson et al. 2014; Guerrieri et al. 2017; Chap. 12 for more details). Once the samples are ready for analysis, they are combusted in an elemental analyzer (analogous to the determination of δ¹³C), which is linked via an open split interface with a sector isotope ratio mass spectrometer to determine the C or N isotope ratios. A recent evaluation of the methodology for preparing spruce tree-ring δ¹⁵N series for the purpose of conducting environmental research used a total of 16 trees from two sites exposed to different types and levels of anthropogenic N emissions (Savard et al. 2020b). This study suggested that short-term δ¹⁵N changes (<7 years) are difficult to relate to long term environmental impacts, but that middle to long-term trends do. The authors also suggest that pooling tree rings of several trees from a site can provide reliable results provided that the sub-samples are of equal weight, but that the ideal method is to calculate the arithmetic mean of δ¹⁵N series from a minimum of three individual trees.

For δ¹⁸O determination samples are thermally decomposed in a helium stream under exclusion of oxygen (pyrolysis). The resulting CO is analyzed in the mass spectrometer (Farquhar et al. 1997; Woodley et al. 2012, Weigt et al. 2015).

2.2 Foliar Flux Quantification of Pollutants

Flux quantification of gaseous pollutants from the atmosphere into the leaf intercellular spaces follows the same principle as for CO₂ and H₂O gas exchange (Gessler et al. 2000; Siegwolf et al. 2001, Teklemariam and Sparks 2004). The pollutant fluxes are calculated by mutliplying g_s with the gaseous pollutant concentration difference between the ambient ([O_3a]) and intercellular ([O_3i]) for the respective gas molecule. The latter is mostly assumed to be zero (Gessler et al. 2000, 2002; Laisk et al. 1989) or where available, values from the literature are used (Yamulki et al. 1996; Thoene et al. 1996; Rennenberg and Geßler 1999). The stomatal conductance is either determined using gas exchange systems or derived from sap flux measurements (Köstner et al. 1996; Zweifel et al. 2007) or meteorological approaches (eddy covariance or energy balance, Wehr et al. 2017). A direct approach to determine gaseous pollutant uptake uses gas exchange chambers, where the pollutant concentration before and after the chamber is measured and the difference is multiplied with the gas flux though the chamber system (Neubert et al. 1993; Teklemariam and Sparks 2004; Grulke et al. 2005, 2007; Chaparro-Suarez et al. 2011). This approach is mostly applied in laboratory experiments or when leaves or twigs can be enclosed in cuvettes. In addition, the use of enriched labeled gases, e.g., ¹⁵NH₃, ¹⁵NO₂ or ³⁴SO₂ is an elegant and reliable method as the assimilated substance can be traced and quantified within the plants (Nussbaum et al. 1993; Segschneider et al. 1995; Vallano and Sparks 2007). In close vicinity of emission sources, the isotopic values of the pollutants are often different from the natural background values, e.g., heavily frequented roads (Amman et al. 1999; Saurer et al. 2004; Guerrieri et al. 2009) or industry plants (Guerrieri et al. 2009; Savard 2010; Savard et al. 2017) vs. clean air regions.

All these methods have their pros and cons. The most prominent disadvantage is that the accurate quantification of the pollutant fluxes is very challenging, since O₃ or NOx as well as NHy react with the surfaces of the leaf environment (i.e., the canopy, adjacent branches and leaf surfaces) and the instrument itself (walls of the tubing and measuring equipment resulting in a gas phase destruction), which can lead to an over-estimation of the real fluxes into the leaves.

3 Specific Impacts of Gaseous Compounds on Isotopic Ratios in Tree Rings

3.1 Increasing Atmospheric CO₂ Concentration

As atmospheric CO₂ is the primary C-source for terrestrial plants, its variations in concentration and δ¹³C values impact the isotopic ratios in organic matter via CO₂ and H₂O gas exchange. Since the initial stages of industrialization around 1850 the CO₂ concentration has increased from ~ 270 µmol/mol to 415 µmol/mol by 2020 (Belmecheri and Lavergne 2020). This 145 µmol/mol increase in c_a can have strong impacts on various aspects of plant physiology and lead to higher c_i as CO₂ diffuses through stomata and into leaves. During much of the last 20,000 years, long time periods have corresponded to increasing CO₂ (c_a). For plant carbon uptake, these increases in c_a caused increases in the difference of c_a−c_i as a consequence of an enhanced CO₂ fixation into sugars within the leaf by photosynthesis, but this also caused c_i to increase in absolute concentrations. Following Fick’s Law, the increasing difference of c_a−c_i is directly proportional to the stimulation of A_N, assuming g_s stays constant (Farquhar et al. 1989):

$${\text{A}}_{\text{N}} \, = \,g_{s \, } (c_a - c_i ).$$

(24.1)

The same relationship can be re-written as:

$${\text{A}}_{\text{N}} \,{ = }\,{\text{ g}}_{\text{s}} {\text{ (1 - (c}}_{\text{i/}} {\text{c}}_{\text{a}} {))}{\text{.}}$$

(24.2)

The term c_i/c_a in Eq. (24.2) is particularly important to understanding how CO₂ may affect tree-ring carbon isotope signatures because Farquhar et al. (1989) further showed that carbon isotope discrimination (Δ¹³C, see Chaps. 9 and 17 for details) can be described by:

$$\Delta^{13} {\text{C}}\,{ = }\,a - (b - a) \, (c_i /c_a ){,}$$

(24.3)

where a is the fractionation from diffusion through the stomata (4.4‰) and b is the fractionation by carboxylation by Rubisco (~27‰).

Equations 24.1–24.3 summarize the simplest version of a much more complex relationship provided by Farquhar et al. (1989) and these have been employed a large number of studies of terrestrial, marine and atmospheric carbon isotope signatures. However, reconsiderations of this approach have increasingly advocated modifying Eq. 24.3 as:

$$\Delta^{13} C\, = \,a - (b - a) \, (c_i /c_a ) - f\left( {{{\Gamma \ast } / {c_a }}} \right) \,$$

(24.4)

where f is the fractionation factor during photorespiration (‰) and Γ* is the CO₂ compensation point in the absence of CO₂ (Seibt et al. 2008; Keeling et al. 2017; Schubert and Jahren 2018; Lavergne et al. 2019).

The goal of this section is not to delve too deeply into the plant physiology and isotopic theory, and applications to Earth systems modeling, which many other papers have addressed within the past 10 years. Moreover, variation in mesophyll conductance is known to have a substantial effect on Δ¹³C values, but it will not be discussed in this chapter due to the great variation among species and a lack of consensus on the magnitude with which changing atmospheric CO₂ can affect this aspect of leaf gas exchange. Rather, this section serves to set the stage for understanding what CO₂-induced trends we should expect to observe in tree-ring Δ¹³C signatures, which is a function of c_i/c_a and f and Γ*, which in turn are directly related to intrinsic water-use efficiency (iWUE; see Chap. 17). Whether CO₂ may affect Δ¹³C values is a long-running and important question to dendrochronologists, plant physiologists, paleoecologists, and a broad array of scientists that use atmospheric δ¹³C or Δ¹³C as a tracer of past global-scale changes to climate and vegetation patterns. Therefore, relevant information for interpreting tree-ring isotope patterns has arisen in diverse fields of scientific inquiry that we will briefly summarize here.

The least intrusive and most realistic way to expose vegetation to elevated CO₂ (eCO₂) is to use chamberless exposure systems, also known as FACE systems (Free Air CO₂ Exposure). FACE studies undertaken in forests are logistically difficult and operationally expensive, so only a handful of these experiments have been undertaken or are still operating (Bader et al. 2013; Norby et al. 2016). Overall, these and other eCO₂ experiments have generally confirmed an expected fertilizing effect on A_N; whereas the expected reductions in g_s have often been absent or smaller than expected on mature trees (Körner et al. 2005; Bader et al. 2013; Keel et al. 2007, Streit et al. 2014; Klein et al. 2016; Gimeno et al. 2020; Jiang et al. 2020; Walker et al. 2020). In forest settings, g_s responses to eCO₂ tend to be a function of various other biological, physical and ecological factors mediating tree- or canopy-level whereas potted plants in the lab or in controlled environments show a strong responsiveness to changes in CO₂ (Jarvis and McNaughton 1986; Buckley et al. 2017; Sperry et al. 2019). Hence, if trees growing in natural environments respond similarly to what has been documented experimentally, with historical increases in A_N and little to no change in g_s (sensu Wieser et al. 2018), then we should expect that trees and forests should become more water-use efficient. Indeed, most tree-ring carbon isotope studies that focus on this issue have documented this shift in water-use efficiency. Most of these hundreds of studies have been recently digitized and summarized by Adams et al. (2020), confirming in a massive meta-analysis what has been demonstrated by a previous and somewhat smaller meta-analysis (Leonardi et al. 2012). Such increases in WUE can, but do not stringently demand, changes in Δ¹³C or c_i/c_a. Saurer et al. (2004) first noted how tree-ring carbon isotope records spanning 1861–1990 documented no change or moderate decreases in Δ¹³C or c_i/c_a despite substantial gains in iWUE. Since then, research interests have focused on whether CO₂ modifies Δ¹³C or c_i/c_a, because such knowledge could provide for more accurate determination of the magnitude with which Δ¹³C has responded to climate or other environmental variability from isotope signals in tree-rings and other plant and paleoecological or geological records.

As CO₂ increases it is often assumed that Δ¹³C or the c_i/c_a ratio remains constant. However, numerous studies show considerable variations in the isotopic responses (Fig. 24.2), which is either species specific or due to habitat properties or changes in the environment. Variations in isotopic response patterns as described by Saurer et al (2004) can even occur within the same tree during its life span (Wieser et al. 2018). To assess the potential impact of c_a on leaf gas exchange, a review of theory and modeling combined with estimates of past c_a and stomatal measurements over geological times concluded that plants maintain a constant c_i/c_a (Franks et al. 2013). Yet, tree-ring Δ¹³C records tended to indicate that Δ¹³C changed in response to increasing c_a, suggesting a correction to c_i/c_a or some other combination of factors not explicitly considered by previous tree-ring isotopes studies (i.e., f, Γ*, or mesophyll conductance) may be needed to accurately separate this effect from that of climate variability on tree-ring Δ¹³C values. However, before a set of mechanisms can be identified, it is first important to identify the magnitude of change. Based on tree-ring Δ¹³C data from four trees, Feng and Epstein (1995) first proposed a correction to Δ¹³C of 0.02‰/ppm CO₂. Soon after Kurschner (1996) proposed a rate of 0.0073‰/ppm CO₂ from a single eCO₂ experiment. Using an 1171-year record from a single species and region, Treydte et al. (2009) concluded that a shift of 0.012‰/ppm CO₂ best fit their data. An over two-fold range of responses among these individual studies clearly demonstrated that a synthesis effort was necessary to determine an overall response, and whether or not a large empirical assessment to determine whether c_i/c_a responds to c_a as indicated by tree-ring Δ¹³C or with the contention of Franks et al. (2013) that c_i/c_a does not respond to c_a. To do so, Δ¹³C responses to CO₂ from many paleo and eCO₂ studies and including all FACE studies that had been conducted in forests up to that time (Voelker et al. 2016; hereafter V16). The three earlier individual studies reporting changes in Δ¹³C/ppm CO₂ describe linear rates of change whereas the V16 data set and analyses predicted that c_i/c_a should increase with c_a but at a non-linear or diminishing rate as:

$$c_i /c_a \, \, = \, - 0.3974 \, + \, 0.5163(1 - e^{ - 0.0076\,ca} ).$$

(24.5)

Fig. 24.2

Responses of Δ¹³C/ppm CO₂ exemplifying the variation among selected individual tree-ring Δ¹³C studies that span at least 25 years and up to 1850 to 2017 (narrow black lines) as well as large-scale meta-analyses and modeling results (thick lines). The thin lines were calculated as the slope of a linear regression of Δ¹³C vs. atmospheric CO₂ concentration, and demonstrate the large potential range in variation that can be encountered for any given tree species and site combination, and thereby the danger in extrapolating a Δ¹³C/ppm CO₂ response from one or only a few such studies. The thick black line gives the mean value of these individual studies, weighted by the inverse of the range of CO₂ concentrations spanned by each study. In this figure, Γ* was assumed constant for all studies and was only formally incorporated into the calculation of Δ¹³C by Keeling et al. (2017) and Schubert and Jahren (2018)

In turn, Eq. (24.5), can be translated to a non-linear change in Δ¹³C per unit CO₂ as:

$$ {\permil / {ppm}}\, = \, \, 0.089 \, e^{ - 0.0076 \, ca} . $$

(24.6)

The negative exponential relationship given in Eq. (24.6) is plotted in Fig. 24.2 and clearly expands c_a values far lower and higher than all tree-ring studies. This relationship likely represented the most robust empirical assessment at that time. However, the intervening five years have provided a number of other relevant studies that can shed additional light on whether Eq. (24.6) is supported or refuted by independent analyses, and how trends in tree-ring Δ¹³C studies should be viewed in light of this body of work. As an example, Lavergne et al. (2019) analyzed > 100 tree-ring Δ¹³C studies constrained to the period 1950–2014. They found little detectable effect of c_a, converted pressure units on Δ¹³C after accounting for immense changes in climate and elevation among studies sites and over time. Ostensibly, this agrees qualitatively with the theory reviewed by Franks et al. (2013). However, the data set investigated by Lavergne et al. (2019) co-varied with a period of recent rapid climate change at many sites and spanned a relatively small gradient in c_a compared to some other studies, which would make it less likely to detect a response of Δ¹³C to c_a.

The Δ¹³C / ppm CO₂ response rate of the two paleo studies of conifer leaves or wood (Hare et al. 2018) and speleothems (Breeker 2017) both overlapped with that predicted by Eq. (24.6), which provided evidence of a consistent Δ¹³C / ppm CO₂ response rate prior to industrialization despite different data types and analytical approaches. Thereafter, Schubert and Jahren (2018) assessed Δ¹³C / ppm CO₂ across c_a of ca. 100 to 3000 ppm for Arabidopsis thaliana, grown in chambers. The model of change in Δ¹³C per ppm CO₂ resulting from this meta-analysis was given as:

$$ {\permil / {ppm}}\, = \, \, \left( f \right){{\left( {\Gamma^\ast } \right)} / {\left( {c_a^2 } \right)}}, $$

(24.7)

where Γ* was set to 40 ppm (Schubert and Jahren 2018).

When we re-applied this model to the sparse data compilation of woody plant responses to CO₂ provided in Schubert and Jahren (2018) (i.e., excluding the data of herbaceous plants grown in growth chambers), the response was nearly parallel to Eq. (24.5) but shifted substantially higher (Fig. 24.2, dark green line). Notably, the model described about 11% of the variation in Δ¹³C / ppm CO₂ whereas an empirical fit to the same data explained 11% of the variation and had a muted response at lower CO₂ concentrations. The value of f = 38.6‰ was found to maximize explained variation between the model and data. The value of 38.6 ‰ is substantially higher than the range of 9.1 to 22.0 ‰ that has been reported previously (Schubert andJahren 2018). These results suggest that their woody plant data set was not large enough for robust parameter estimation. To remedy this problem, we plotted Eq. (24.7) fitted to the much larger V16 data set (Fig. 24.2, thin brown line), which yielded a value of f = 19.4‰, which is at the high end of the previously reported range, but much more reasonable compared to f = 38.6‰. Overall, this analysis provides powerful independent confirmation of the shape and magnitude of the Δ¹³C response to CO₂ despite two very different analytical approaches. More specifically, V16 attributed CO₂-induced changes to Δ¹³C entirely to shifts in c_i/c_a, whereas Shubert and Jahren (2018) attributed the same changes entirely to fractionation associated with photorespiration. It is certainly possible that some combination of these two primary mechanisms, other morphological and biochemical contributing factors, as well as ecological and evolutionary feedbacks operating on long timescales, determine the overall Δ¹³C response to CO₂ over long time periods.

The measured record of ¹³CO₂ in the atmosphere provides yet another means for understanding how terrestrial Δ¹³C of the vegetation has been influenced by rising CO₂. When global models of ocean and terrestrial productivity and carbon cycling are constrained to match ¹³CO₂ records, the results required that Δ¹³C increase over the industrial period (Keeling et al. 2017). Although the positive value of this response rate of Δ¹³C / ppm CO₂ obtained by Keeling et al. (2017) is in broad agreement with many other studies (as noted in Fig. 24.2), it was substantially higher compared to most tree-ring Δ¹³C responses and meta-analyses for that same range in atmospheric CO₂. The apparent difference in the Δ¹³C / ppm CO₂ response rate between the ones reported by Keeling et al. (2017) and that plotted by using Eqs. (24.6) or (24.7) as applied to the V16 data set has most likely multiple, non-exclusive causes as detailed below. The difference may reflect (1) the difference between using woody plant Δ¹³C responses largely from temperate species as compared to global vegetation responses where C₄ grasses and other vegetation types play a significant role (2) the lack of a temperature response of Γ* being incorporated in the analyses of Eqs. (24.6) and (24.7) applied to the V16 data set, and (3) deficiencies in the global modeling approach of Keeling et al. (2017). With respect to point (1), reconciliation of the difference would require that C₄ plants and tropical forests, the vegetation types most conspicuously lacking from paleo and tree-ring Δ¹³C data sets, be characterized by stronger responses of Δ¹³C to rising CO₂ concentrations. The Δ values of C₄ plants are lower and far less sensitive to changes in CO₂ compared to C₃ plants, so the source of increased Δ¹³C responses to CO₂ would need an origin in tropical forests. This makes intuitive sense since C₃ net photosynthesis includes more dynamic responses to CO₂ at higher average temperatures characterizing tropical forests. Tropical rainforests would also have fewer and less severe moisture constraints, allowing greater sensitivity to CO₂ compared to many temperate forests. With respect to point (2), the incorporation of a temperature response of Γ* was not attempted because surface air temperatures are poorly constrained for paleo data collections spanning the Holocene and Last Glacial period. V16 did attempt to provide modern locations with similar climate conditions compared to the paleo data collections. The implicit assumption therein is that differences in temperature and Γ* would add noise to the data set but be unbiased. That analysis also eliminated the Δ¹³C difference between paleo and CO₂-enrichment studies at the approximate breakpoint between the two types of data sets at ca. 350 ppm that could have owed to differences in species composition, climates and the temperature dependence of Γ*. With respect to point (3), the findings of the simulation modeling by Keeling et al. (2017) spanned the years 1765–2014 and concluded that Δ¹³C of terrestrial vegetation must have changed significantly, and attributed this change to a steady increase in Δ¹³C of 0.014‰ / ppm CO₂. Although this change was attributed to the effects of atmospheric CO₂, this period encompassed increasingly widespread disturbance to forests at a global scale, which certainly decreased vegetation height and competition for soil water, which would have contributed to progressive reductions in globally integrated terrestrial Δ¹³C over this time period. Hence, other potential sources of bias outside of direct CO₂-effects on leaf gas exchange influenced ¹³CO₂ records and could have been mistakenly been attributed to impacts of rising atmospheric CO₂ on vegetation Δ¹³C by Keeling et al. (2017). Overall, the reconciliation of results provided here indicates that the Δ¹³C response to CO₂ in woody plants will generally be expected to follow either Eqs. (24.5) or (24.6) when parameterized with values of Γ* = 40 ppm and f = 19.4‰, but that the overall Δ¹³C response to CO₂ of global vegetation may be somewhat greater in magnitude according to the findings of Keeling et al. (2017), unless this attribution to CO₂ was biased by not including the effects of progressively greater disturbance on forest leaf gas exchange.

A number of readers of this text may be wondering what, in practice, should be done about expected increases in Δ¹³C in response to increasing atmospheric CO₂ concentrations. This topic is difficult to address because there is no one correct answer. An important aspect in assessing apparent Δ¹³C / ppm CO₂ trends at any given site or across sites is that there can be tremendous variation among individual tree-ring studies, as represented by the numerous narrow black lines displayed in Fig. 24.2. This site-to-site variability results from many interacting factors such as climate change, climate oscillations, tree development (i.e., increased tree height or stature), competition, pollution and other stressors that are superimposed on top of the relatively small effects of CO₂. Indeed, Leonardi et al. (2012) summarized a large number of tree-ring Δ¹³C studies and found a huge shift in responses between the periods 1860 and 1990 that included negative overall responses of Δ¹³C / ppm CO₂ during the 1952–1990 period (Fig. 24.2). Such a negative Δ¹³C response to CO₂ is probably not realistic based on first principles. At sites with low productivity there may be little to no response of Δ¹³C to CO₂ (Marchand et al. 2020) whereas in other regions, negative Δ¹³C trends have been associated with increased stomatal closure associated with warming- or competition-induced drought stress (Peñuelas et al. 2011; Saurer et al. 2014; Lévesque et al. 2014; Frank et al. 2015; Voelker et al. 2019). Indeed, some studies have utilized approaches that have determined differences from the expected trend in Δ¹³C / ppm CO₂ across a range of sites (Liu et al. 2018; Szejner et al. 2018; Voelker et al. 2019). The approach of these three papers was to remove the expected CO₂ responses with an aim of better isolating the impacts of changing climate or competition on ecophysiology. This approach may prove particularly useful for isolating spatio-temporal impacts of climate and climate change and other factors such as pollution that affect tree-ring carbon isotope signatures if future efforts adjust for the temperature dependence of Γ* and the partial pressures of CO₂ that differ with elevation (sensu Lavergne et al. 2019).

δ¹⁸O: In experiments with potted young trees, increases in CO₂ resulted in a distinct reduction of g_s. However, in some FACE experiments in temperate mixed forest little (Keel et al. 2007) or no stomatal response to changes in CO₂ was found (Bader et al. 2010, 2013; Klein et al. 2016). Even after 9 years of FACE treatment trees growing at the timberline at 2180 m above sea level (Stillberg, Davos, Switzerland) showed no stomatal response to increased CO₂ for either species, Larix decidua and Pinus mugo. Accordingly, the effect of eCO₂ was not reflected in δ¹⁸O values of organic matter, although g_s of the same plants was highly responsive to changes in VPD, which was clearly reflected in δ¹⁸O (Streit et al. 2014). As noted above, site-specific biological, physical and ecological factors cause tree- or canopy-level effects to be diminished compared to leaf-level responses to CO₂. So far, δ¹⁸O data of trees and tree rings from FACE experiments are rare. Before we derive CO₂ responses via stomata reflected in δ¹⁸O values from tree rings we must keep in mind that the oxygen isotope composition is modulated by a number of other factors, such as origin of precipitation water (Dansgaard 1953, 1964, Chap. 18), seasonality (Daansgard 1964; Rozanski et al. 1993) air humidity and δ¹⁸O of water vapor (Dongmann et al. 1974; Roden et al. 2005; Lehmann et al. 2018) soil properties and soil depth (Sprenger et al. 2018; Brooks et al. 2010). Voelker and Meinzer (2017) summarize further that post photosynthetic fractionation processes alter the δ¹⁸O values in wood during the transfer of assimilates in the phloem: an oxygen exchange occurs with ¹⁸O-depleted xylem water, modifying the δ¹⁸O of the assimilates (Roden and Ehleringer 1999; Barbour et al. 2004) especially during the synthesis of cellulose (Sternberg 2009, Gessler et al. 2014, Chap. 10). The variability of these processes can add signal noise, masking stomatal signals imprinted on the δ¹⁸O of tree rings to a degree that the stomatal signal is no longer detectable. This is especially the case if the modification of the oxygen isotope ratio by g_s is small (e.g. at high air humidity; Roden et al. 2005), resulting in an unfavorable signal to “noise” ratio. Nevertheless, we cannot conclude that changes in CO₂ have no effect on δ¹⁸O, as other studies for other species have found different responses (Battipaglia et al. 2013).

Albeit the reduced responsiveness of g_s to CO₂ in the field, using both, C and O isotope values widens the scope of interpretation; even more so if the δ¹⁸O of the source water and water vapor are known. The latter can also be estimated with specific models (Barbour et al. 2001). The application of various statistical tools or the dual isotope approach as described in Chap. 16 is valuable to evaluate the impact of long-term CO₂ dynamics on plant physiology, especially for the evaluation of the long-term intrinsic water-use efficiency (iWUE, see Chap. 17). Including tree-ring width and anatomical parameters (cell wall thickness, wood density, etc. see Sidorova et al. 2019) will help to identify, which factors impact tree growth and to what degree besides increasing [CO₂]. Such an enhanced data set is instrumental to correctly evaluate the carbon gain water loss relationship (see Chaps. 16 and 17).

To summarize: the wide variety in isotopic patterns in response to CO₂ changes reflects the dynamic variability of species-specific responses to elevated CO₂. This includes (a) stimulation of A_N under high CO₂ levels (e.g., Herrick and Thomas 2001; Bader et al. 2010; Klein et al. 2016), (b) down regulation of photosynthesis (Sage et al. 1989; Grams et al. 2007), (c) reduction in stomatal conductance (Gunderson et al. 2002), or (d) little (Keel et al. 2007; Bader et al. 2010) to no stomatal response to changes in CO₂ (Bader et al. 2013; Streit et al. 2014; Klein et al. 2016). These short-term responses on the leaf level leave their isotopic fingerprints in the assimilates, which are cumulated and integrated over the growth period and recorded in the biomass of tree ring (Chap. 14).

3.2 Ozone (O₃)

Ozone in the troposphere is considered to be one of the most phytotoxic air pollutants, impacting plant and ecosystem functions (Ashmore 2005; Ainsworth et al. 2012). A good description of its formation, propagation and fate is given in the report of the Royal Society (2008). The main process of O₃ removal from the troposphere is the dry deposition on land surfaces, with an important role for vegetation such as forest ecosystems by O₃ uptake through stomata (Cieslik 2004). Non-stomatal removal of O₃ occurs at soil and plant surfaces (e.g. cuticle deposition) and degradation reactions with soil or plant emitted NOx or biogenic VOCs (Lenhart et al. 2019; Fowler et al. 1998, 2001). These removal mechanisms complicate the quantification of the O₃ uptake via stomata; nevertheless, significant estimates of non-stomatal O₃ fluxes were reported (20–80%, Cieslik 2009; 26–44%, Rannik et al. 2012), depending on the type of vegetation cover and its surface characteristics. Albeit these uncertainties, flux estimates based on the combination of stomatal conductance and ozone concentration still yield important results (Karlsson et al. 2004, 2007; Grünhage et al. 2004) for the ozone dose effective on plant physiology. More sophisticated models aim at linking stomatal O₃ uptake to growth decline for a more mechanistic ozone risk assessment to forests (Matyssek et al. 2008, Anav et al. 2016, Feng et al. 2018, for crop plants, Hu et al. 2015, Xu et al. 2018).

3.2.1 Damages Caused by Ozone

Once O₃ has been taken up by the plants via the stomata it may cause numerous visible leaf injuries such as chlorosis and necroses, accelerated senescence and pre-mature leaf shedding (Günthardt-Goerg and Vollenweider 2007). Even before these symptoms become visible, a number of responses at the cellular and molecular level have already occurred. Upon uptake, O₃ rapidly reacts with wet surfaces in the intercellular cavity, producing a suite of reactive oxygen species (ROS) that challenge a plant’s detoxification mechanisms. Quite similar to defense reactions against pathogens and programmed cell death (for details see Kangasjarvi et al. 2005, Sandermann et al. 1998), the plant amplifies internal ROS production that cause the above-mentioned leaf symptoms in what is known as the hypersensitive response reaction. In the long run and often before visible symptoms develop, or even without them, stomatal functioning and Rubisco activity are affected by chronic O₃ stress. The stomatal response is not the same among species and O₃ concentrations. While short-term exposure to high concentrations typically result in a direct closing response of stomata, long-term exposure to moderately enhanced O₃ levels is also reported to reduce g_s (Hoshika et al. 2014, 2015; Wittig et al. 2007), and related to declining rates of CO₂ fixation by lower Rubisco protein amount and activity (Karnosky et al. 2005; Wittig et al. 2009; Dizengremel 2001). Other reports, however, describe impaired stomatal functioning by elevated O₃, causing delayed responses of stomata to environmental changes. This so called “stomatal sluggishness” increases response times of opening and closing movements, which can lead to increased water loss by transpiration under high VPD, respectively (Paoletti and Grulke 2010; Paoletti et al. (2020). At the whole tree level, elevated O₃ affects allocation processes of photoassimilates. In mature trees, long-term exposure to twice ambient O₃ concentrations significantly reduced allocation to stems (Ritter et al. 2011), which is well in line with the often-found reduction of stem growth in natural environments with high O₃ exposure (Karnosky et al. 2005). Below ground, O₃ impaired source-sink relations are believed to reduce belowground C allocation (reviewed by Andersen 2003, Andersen et al. 2010), typically resulting in lower root/shoot biomass ratios. However, a larger number of studies report on unchanged root/shoot biomass ratios (Agathokleous et al. 2016) and few experiments on field grown trees under long-term O₃ exposure even found fine root growth to be stimulated (Nikolova et al 2010; Pregitzer et al. 2008).

3.2.2 Consequences for Isotopic Fractionations

Since O₃ impairs stomatal regulation, photosynthesis and its biochemical processes, these impacts will be seen in the carbon and oxygen isotopic fractionations and thus in the variation of the ¹³C/¹²C and ¹⁸O/¹⁶O isotope ratios of different plant compartments. Species respond differently to elevated O₃ depending on O₃ uptake, mesophyll exposure, detoxification capacity, and plant age (Fuhrer and Booker 2003; Matyssek and Sandermann 2003).

In a 2-year phytotron study on juvenile Fagus sylvatica trees exposed to elevated O₃, a parallel increase of δ¹³C and δ¹⁸O was observed (Grams et al. 2007; Grams and Matyssek 2010), indicating a concurrent reduction of c_i as a response to stomatal closure. Likewise for mature trees of the same species, O₃ caused reductions in g_s and related increases in δ¹⁸O (Kitao et al. 2009; Gessler et al. 2009) as well as for different herbaceous plants (Jäggi and Fuhrer 2007). In accordance with reductions of g_s, the phytotron study found A_N to be reduced as often observed under elevated O₃ (Ainsworth et al. 2012) and not increased as one might have expected from higher δ¹³C values. Moreover, the increase in δ¹³C under O₃ compared to control trees was stronger than expected from measured c_i and corresponding calculations using the Farquhar model (Kozovits et al. 2005; Grams et al. 2007). This apparent mismatch between gas-exchange measurements and carbon isotope discrimination was also reported by Saurer et al (1991) for grain (Triticum aestivum), Patterson and Rundel (1993) for Pinus jeffreyi and by Elsik et al. (1993) for two Pinus species. Later studies found PEPC (Phosphoenolpyruvate carboxylase) to be increased in amount and activity under O₃ exposure, resulting in increased δ¹³C values for both leaves and stems (Saurer et al. 1995; Doubnerová and Ryslavá 2011). Regardless of the higher PEPC activity, A_N was reduced and g_s showed little response compared to control plants. The increased δ¹³C values were explained by the fact that C-fixation through PEPC does not discriminate as much against ¹³C compared to ¹²C resulting in an¹³C enrichment in the assimilates (Vogel 1993). Various detoxification mechanisms further the ¹³C enrichment along with an enhanced PEPC activity that are generally stimulated under stress (Doubnerová and Ryslavá 2011). An increase in δ¹³C under O₃ exposure, which would suggest an enhanced or constant A_N at reduced g_s is not in line with our understanding of ¹³C fractionation coherence according the fractionation model for C₃ plants of Farquhar et al. (1989; 1982). Based on direct CO₂ and H₂O gas-exchange measurements, we expect more negative δ¹³C values as A_N is reduced while g_s stays constant or is reduced to a minor degree (Grams et al. 2007). This example demonstrates once the assumptions of the well-accepted fractionation model are violated, the conclusions drawn from isotope measurements can lead to physiologically implausible results. As the ratio of Rubisco to PEPC activity decreases significantly under elevated O₃ (up to 20-fold, Dizengremel et al. 2009), the fractionation model must be adjusted, e.g., by increasing the parameter “b” representing the fractionation of all carboxylation processes as suggested by Saurer et al. (1995). Therefore, the consultation of the history of human impact in the past for the respective forests is of great value for the data interpretation.

3.3 Sulphur Dioxide (SO₂)

Ambient SO₂ concentrations increased considerably since the advent of industrialization and became an increasing problem across Asia, Europe and North America, particularly before the mid-1980s. This pollutant may have had a stronger impact on the isotope ratios in plants and tree rings than previously recognized. Sakata and Suzuki (2000) described a significant δ¹³C increase in Japanese fir trees that underwent diseases and insect pests when being weakened by SO₂. Savard et al. (2004) reported an unprecedented δ¹³C positive shift by up to 4‰ for spruce tree rings induced by SO₂ emissions from a large smelter in Canada. Wagner and Wagner (2006) and Boettger et al. (2014) described major increases in the long-term trend in pine and fir tree-ring δ¹³C series between 1945 and 1990 and a subsequent decrease after 1990 reflecting the trends of SO₂ concentration in German sites. In their study of English oak trees exposed to SO₂, Rinne et al. (2010) document a 2.5‰ δ¹³C rise with insignificant changes in the δ¹⁸O series. Many other examples of SO₂ effects in natural settings around the world could be cited, with the singular common impacts of rising tree-ring δ¹³C values during the time of exposure.

A possible explanation for this general δ¹³C increase is a secondary fractionation resulting from SO₂ phytotoxicity, which causes a decline in various physiological activities. Closure of stomata is generally induced by foliar SO₂ uptake, and acidification of solution often invoked to explain this closure. Recently, molecular biological research suggests instead that guard cell mortality cause stomatal closure (Ooi et al. 2019). However, plant response mechanisms are intricate as they depend on the SO₂ concentration (and presence of other co-pollutants), duration of exposure and the availability of water (e.g., Maier-Maercker and Koch 1995; Lee et al. 2017). Under acute controlled exposures (25 mg SO₂/m³), the foliar system of trees becomes photo inhibited and exhibits decreasing A, g_s and c_i (Duan et al. 2019). Overall, authors from diverse research fields ascribe the related tree-ring δ¹³C increase and growth reactions to various combinations of changes in g_s, A_N, starch production and priority of C allocation (e.g., Darral 1989; Meng et al. 1995; Kolb and Matyssek 2001; Grams et al. 2007).

In the Canadian example cited above, a reduction of g_s was not the only response caused by exposure because a concomitant δ²H decrease in tree-ring nitrated cellulose was significant, and stem growth did not notably change (Savard et al. 2005). Considering intrinsic factors only, decreases in δ²H and δ¹⁸O generally reflect stomatal opening (Dongmann et al. 1974; Farquhar and Lloyd 1993; Sensula and Wilczynski 2017). When considering extrinsic factors, this case of severe exposure of trees to emissions may have acidified the upper soil layers, damaging fine surface roots and shifting water uptake to large, deeper roots, where source water δ²H (δ¹⁸O) values are lower (Chap. 18). In the German examples, the combined physiological responses to high SO₂ pollution are expressed by the long-term positive δ¹³C series with no significant to moderate δ¹⁸O changes (Wagner and Wagner 2006; Boettger et al. 2014), and no effects on δ²H values (Boettger et al. 2014). This case reflects increased respiration rates that expel higher proportions of ¹²C and generate tissues with higher δ¹³C signals, without modifying the δ¹⁸O values. Otherwise, foliar gas-exchange responses inferred from increasing δ¹³C with moderately decreasing δ¹⁸O trends a priori should reflect increases in A_N and g_s. However, this scenario is in contrast with directly measured CO₂ and H₂O gas-exchange values under controlled SO₂ exposure, as Atkinson and Winner (1987), Kropff et al. (1990), Wedler et al. (1995) and Randewig et al. (2012) report a reduction in A_N and g_s, which should result in an overall decrease in δ¹³C values. The observed reduction of A_N better fits with the general observation that ring growth does not increase with the isotopic changes. Hence, aside from the possible extrinsic factors proposed above, the SO₂ induced enzymatic detoxification and reduced carboxylation rate represent other mechanisms that could explain strong increases in ¹³C uptake (Randewig et al. 2012), in some cases outweighing the prediction of the widely accepted C-isotope fractionation model by Farquhar et al. (1989, 1982).

As shown for ozone exposure, the impact of SO₂ detoxification and its effect on δ¹³C values are not considered by the C-isotope fractionation model for C₃ plants. Moreover, foliar response models do not take into account the extrinsic effects on tree-ring δ¹⁸O (δ²H) changes. Thus, the common data interpretation assuming that a δ¹³C increase with none to moderate δ¹⁸O (δ²H) changes would be caused by an increased A_N is not correct and does not correspond to real physiological responses to SO₂. Accordingly, paleoclimate reconstruction, the evaluation of the water-use efficiency (Chap. 17), or the use of the dual C and O isotope gas-exchange model (Chap. 16) could all be erroneous if based on tree-ring series impacted by airborne acidifying emissions. Therefore, the effects of these emissions must be considered for the evaluation of tree ring isotope chronologies originating from regions and periods with large SO₂ emissions.

3.4 Gaseous Reactive Nitrogen (N_r) Compounds: NO, NO₂, NH_X

As observed for ozone and SO₂, the emissions of biologically reactive N compounds (N_r) have greatly increased (NOx from combustion and NHy from agricultural fertilization) since the beginning of the industrialization. The distribution of N_r occurs via large-scale atmospheric transport and is deposited into terrestrial ecosystems. Although N deposition is identified as the most important growth driver in managed European forests (Etzold et al. 2020), the current often excessive N input can reduce forest growth (Waldner et al. 2014).

Vegetation demand for N is mostly met by root uptake of nitrate (NO₃), ammonium (NH₄+) and dissolved organic N-compounds from the pedosphere (Chap. 12). However, the direct uptake of gaseous N_r (NO, NO₂, NH_y) via stomata can be substantial and several studies report quantities obtained in this manner total between 10 and 37% of plant-N demand (Amman et al. 1999; Harrison et al. 2000; Millard and Grelet 2010; Vallano and Sparks 2013; Chap. 12).

3.4.1 Factors Impacting N_r-uptake by Leaves

The main factors influencing the assimilation of N by the foliar system are: (i) the species of trees, (ii) the proximity to anthropogenic emissions, (iii) the plant and canopy structure, and (iv) the type of N_r. (i) Chaparro-Suarez et al. (2011) reported a species dependent variability for NO₂ uptake, which is directly correlated with g_s confirming the regulation of stomatal uptake. Accordingly, they report Betula pendula had the highest g_s and highest NO₂ uptake, in contrast to Pinus sylvestris, where both values were the lowest. In all cases, they found no NO₂ compensation point (meaning that plants do not produce NO₂), or cuticular N transfer, which agrees with other studies (Teklemariam and Sparks (2004); Gessler et al. (2000). (ii) Savard et al. (2009) and Guerrieri et al. (2010) studied trees exposed to emissions in the vicinity of industrial complexes and reported tree-ring series carrying δ¹⁵N values apparently reflecting airborne industrial NO₂. For trees growing in the vicinity of highways, similar δ¹⁵N values were found in the tree rings as in the atmosphere (Amman et al. 1999; Saurer et al. 2004; Doucet et al. 2012). These authors found an increase in the δ¹⁵N values of the NO₂ by 5‰ to 8‰ relative to the δ¹⁵N of the background values. It is assumed that the exhaust treatment by car catalyzers results in an enrichment of the ¹⁵NO₂ due to the stronger bounding of the heavier¹⁵N . Thus, the lighter ¹⁴NO₂ reacts more readily on the catalyzer resulting in N₂ and O₂. (iii) Stand and canopy structure impact N-deposition considerably (Bettinger et al. 2017) as the N-compounds react with the surfaces of leaves, branches and stems, reducing their concentrations in air. Baldocchi and Wilson (2001) indicate a decreasing importance of surface deposition with decreasing foliage density, as indicated by reduced deposition velocities. (iv) For each of the different biological N_r molecules, different flux rates are found due to the different solubility of the respective molecules in an aqueous milieu. NH₃ shows a much higher water solubility than NO₂ and NO (Felix et al. 2017; Castro et al. 2005), and NO is less soluble than NO₂ by an order of magnitude (Neubert et al. 1993). Accordingly, this solubility hierarchy is reflected in the flux rates for all three molecules (Teklemariam and Sparks 2004, Gessler et al. 2000). Likewise, mesophyll resistance (r_m) is another critical parameter in the chain of incorporation for gaseous compounds, and is considerably higher for NO than for NO₂ (Van der Eerden et al. 1998), and higher for NO₂ than for NH₃ (Teklemariam and Sparks 2004; Castro et al. 2005; Gessler et al. 2000). In contrast to NO₂ and NH₃, most authors found stomatal NO fluxes negligibly small with little or no impact on plant physiology and in plant organic matter (Neubert et al. 1993). As NH₃ is highly water soluble and rapidly converted to NH₄⁺ in the aqueous milieu of the mesophyll cells, r_m for this gas is low and it is assimilated via the glutamine synthetase / glutamate synthase cycle (Sparks 2009; Castro et al. 2005; Pearson and Soares 1998). In contrast, the assimilation of NO₂ is somewhat more complex. Once this gas enters the intercellular cavity it is rapidly dissolved in the aqueous phase of the apoplast via disproportionation to nitrite and nitrate. NO₂ scavenging by ascorbate represents an alternative parallel pathway. The resulting nitrite and nitrate from both pathways are metabolized via nitrite and nitrate reductase for the ferredoxin dependent reduction of NO₂^– to NH₄⁺ and the glutamine synthetase driven production of amino acids (Ramge et al. 1993). A good overview is given by Sparks (2009).

3.4.2 Physiological Effects Caused by Gaseous N-compounds

The atmospheric δ¹⁵N values of N_r in the vicinity of emission sources are often distinctly different than those from regions without N_r-pollution (background values, either enriched or depleted in ¹⁵N). This allows the quantification of the direct uptake of N_r-gases via stomata and their incorporation into organic matter, as these products are transferred via phloem to branches, stems (tree rings) and roots. Since the δ¹⁵N variations are detectable in tree rings (although N is at lower concentrations than in leaves), the historical development of anthropogenic N emissions can be inferred retrospectively (Amman et al. 1999; Saurer et al. 2004; Choi et al. 2005; Guerrieri et al. 2009; Savard et al. 2009; Savard 2010; Sun et al. 2010; Mathias and Thomas 2018).

Assimilation and incorporation of N_r impact numerous metabolic processes (see above) reflected in A_N and g_s (e.g., Gessler et al. 2000; Siegwolf et al. 2001). However, few studies have investigated the impact of assimilating anthropogenic N on the C and O isotope fractionation in plant organic matter. In 1988, Martin et al. reported increasing δ¹³C values in leaves and wood of young trees, which were exposed to increased SO₂ + O₃ and SO₂+ O₃+ NO₂ concentrations for a four-week period. They argued that the δ¹³C increase was a result of stomatal limitation. But they did not examine the plant responses for each pollutant separately, thus it was not possible to clearly assign the impact of each pollutant on specific plant mechanisms due to the combined application of all three pollutants. As shown by Saurer et al. (1995), the increase in δ¹³C values for O₃ exposed plants was primarily caused by an increased PEPC activity.

In a comprehensive case study, Siegwolf et al. (2001) analyzed the impact of chronic NO₂ exposure on the C and O isotope ratios and CO₂ and H₂O gas exchange. Populus euramericana vars Dorskamp cuttings were grown in climate-controlled growth chambers under limiting and surplus soil-N regimes. One chamber was supplied with filtered air and in the other an average NO₂ concentration of 100 nmol/mol was maintained for 12 h per day for three months (details in Siegwolf et al. 2001). Irrespective of the soil-N regime, NO₂ exposure caused δ¹³C to increase in leaf organic matter relative to plants grown in clean air as reported in other studies, (Bukata and Kyser 2007, Vallano and Sparks 2013) whereas the δ¹⁸O values were reduced (Siegwolf et al. 2001; Fig. 24.3). Leaf uptake of N_r compounds allows for a direct and rapid N availability in plant metabolism and has a fertilizing effect. Since N_r uptake via leaves stimulates A_N and leads to higher g_s, this affects the C and O isotope fractionation (see Chaps. 9, 10 and 19). For both, limiting and surplus soil-N regimes, NO₂ exposure caused an increase in δ¹³C and reduction in δ¹⁸O values indicating higher A_N and g_s values, respectively, which is confirmed by CO₂ and H₂O gas-exchange measurements. In the absence of NO₂, the effect of soil-N surplus relative to soil-N limitation results in a decrease of δ¹³C and δ¹⁸O values, suggesting that A_N might be reduced or constant whereas g_s is considerably increased. Yet, an increase of the soil-N supply results in an increase of A_N and g_s, while g_s rises over-proportionally relative to A_N (Sage and Pearcy 1987; Dinh et al. 2017). The CO₂ and H₂O gas exchange measurements confirm an increase in g_s, while A_N is increased as well but not to the degree that δ¹³C would increase. However, care must be taken when generalizing this response to changes in soil N-supply. Chapter 23 describes different responses, mostly for coniferous species (see Figs. 23.5 and 23.6). A meta-analysis showed an opposite trend in response to soil-N fertilization (decrease in ∆¹³C, increase in δ¹³C) after the first few years. The impact of N-deposition can cause divergent isotopic response patterns, depending on the nutrient status, water supply, canopy traits and structure, species composition etc., which needs to be considered for the data interpretation.

Fig. 24.3

(adapted from Siegwolf et al. 2001). The solid orange arrows between data points indicate the δ¹³C and ¹⁸O shifts caused by NO₂ exposure, and the dashed green lines indicate the changes resulting from increased soil N addition. The fine lines with capital and small letters, respectively, represent the shifts caused by the N-treatments. The letters ‘b’ and ‘d’ indicate the shifts in δ¹³C (positive) and ¹⁸O (negative) for the low soil-N supplied plants, when exposed to NO₂. The capital letters ‘B’ and ‘D’ stand for the shifts of the high soil-N supplied plants, caused by NO₂ exposure. The letters ‘a’ and ‘c’ quantify both the δ¹³C and¹⁸O negative shifts caused by a high soil-N supply in NO₂ free air. ‘A’ and ‘C’ represent the shift of the NO₂ exposed plants, caused by high soil-N supply. The standard error is indicated by the horizontal and vertical bars in each data point (n = 6)

δ¹³C plotted against δ¹⁸O values from all four treatments (C stands for filtered air, C-low nitrogen (LN), C-high nitrogen (HN), NO₂-LN and NO₂-HN) in Populus × euramericana

It is noteworthy how the isotopic response patterns in deciduous trees differ between the cases of “soil N-changes only”, “NO₂ exposure only”, or “soil N change and NO₂ exposure” (Fig. 24.3). These isotopic patterns (Fig. 24.3) as a result of NO₂ exposure were also found in air pollution studies using tree-ring samples. Saurer and Siegwolf (2007) and Guerrieri et al. (2009) analyzed tree rings from Picea abies growing along a highly frequented freeway in Switzerland and found δ¹³C and δ¹⁸O patterns invoked by NO₂ exposure. However this pattern as shown in Fig. 24.3 might change with the occurrence of other more dominant impacts such as drought. In the same study of Guerrieri et al. (2009), the impact of NO₂ emissions from an oil refinery in Southern Italy were analyzed for changes in δ^{15 N} (as an indicator), δ¹³C and δ¹⁸O values from tree rings. Drought events at this Mediterranean site occur frequently, in contrast to the sites along the Swiss freeway where water supply is not limiting. Thus the trees near the oil refinery responded primarily to drought stress by reducing g_s to minimize water loss. Consequently, the protection against drought is more relevant for plant survival therefore the drought response outweighs the influence of NO₂, which was also reflected in the increased iWUE (Guerrieri et al. 2010). For trees, which are not subject to other dominant stressors (drought, high concentrations of pollutants, flooding or abrupt changes in the vegetation), this isotopic pattern as shown in Fig. 24.3 was seen in most NO₂-controlled studies, and its application can serve for diagnostic or monitoring purposes.

Regarding NH₃ or NO we have not found any literature that described the impact of NH₃ or NO on the isotopic fractionation in trees, which does not mean that we can exclude any isotopic effect. Based on gas-exchange data we assume the following isotopic response as a result of NH₃ exposure: Fangmeier et al. (1994) and Krupa (2003) report an increase in A_N and g_s with increasing NH₃ concentration. They hypothesized that g_s was indirectly impacted because an increase in NH₃ enhanced the carbon demand for building skeleton C compounds. This in turn generated an increase of A_N, reducing c_i, leading to an increase in g_s. Based on this mechanistic chain, we premise that ∂E/∂A (E is transpiration and A assimilation) is a constant (Cowan and Farqugar 1977). Therefore, we assume a decrease in δ¹⁸O (increase of g_s, Chap. 10), while δ¹³C stays constant, as a consequence of the proportional increase in A_N. However, this hypothesis must be verified experimentally. With regard to the small NO flux and its low aqueous solubility, we assume that its physiological impact is negligible at ambient concentrations.

3.4.3 Impact of Anthropogenic N_r on Root N Assimilation and Tree-Ring δ¹⁵N Series

The reader interested in the airborne N_r transformations in soils can find a relevant discussion in Chap. 12. A final point regarding the assimilation of anthropogenic N compounds as depicted by tree-ring series grown under natural conditions is that soil biogeochemical conditions can play a key role in the way coniferous trees respond to anthropogenic N deposition, especially in forests where soil is N limited. Indeed, depending on the N status of the forest and on soil conditions (pH, action exchange capacity, base cation saturation ratio, etc.), direct or ectomycorrhizal (EcM) root uptake may take place. The direct assimilation may not fractionate N isotopes prior to biological reactions within trees, but it is well known that EcM fungi generally are enriched in ¹⁵N in the process of fungal metabolism, and release the relatively light N that is transferred to the roots and stems of trees, with different degrees of fractionation depending on the abundance and diversity of the various community components (e.g., Hobbie and Högberg 2012). The proportions of direct and EcM N uptake modulate the final δ¹⁵N values recorded in tree-ring series. There might be a tipping point at which the bacterial and fungi communities in close association with the root system are modified so that trees exposed to similar enhancement of bioavailable N but growing on different physicochemical soil conditions will record opposite long-term δ¹⁵N trends (Savard et al. 2019). All things considered, it is no surprise that soil dynamics will impact the tree-ring series as the dominant part of N in coniferous stems originates from soil N compounds. It is clear however that research is needed on the ultimate modulation of tree-ring δ¹⁵N series by long-term changes of microbial dynamics in soils.

4 Concluding Remarks

Airborne pollutants may alter both the leaf and root assimilation paths for carbon and nutrients in trees. It is often difficult to uncover and identify the coherences between the impact of a single air pollutant or other environmental vectors and specific plant responses with their stable isotopic composition. In the real world, plants are often exposed to many pollutants, that together modify the atmospheric and soil environments. Therefore, the knowledge on how each pollutant specifically affects plant metabolism is instrumental for a better understanding of the combined impact under consideration of a given environmental situation. Furthermore, it facilitates the planning of combined experimental assessments and for modeling ecosystem responses to pollution exposure. The consideration of the vegetation and management history facilitates an accurate interpretation of the isotope values from long-term chronologies and the inclusion of other parameters (tree ring width, wood density etc.; Sidorova et al. 2019) strengthens the interpretations of tree-ring isotopic data.