Keywords

1 Introduction What Are Plant Volatiles?

Plant-released organic volatiles constitute a vast spectrum of compounds, more than 30,000 different compounds with a certain capacity to escape into the gas phase from a liquid or solid (Niinemets et al. 2004). In common with the compounds characteristically called volatiles is that they have normal pressure boiling points between ca. 30 and 250 °C and, thus, support a relatively high vapor partial pressure at ambient temperatures (between ca. 101 and 105 Pa at 25 °C) (Fuentes et al. 2000; Copolovici and Niinemets 2005; Kosina et al. 2013). In addition, studies on plant volatiles also often consider semivolatiles that support a much lower vapor pressure (partial pressure ca. 10−6 and 100 at 25 °C) (Helmig et al. 2003; Widegren and Bruno 2010; Kosina et al. 2013). Semivolatiles have a large capacity for partitioning into liquid and solid phases and, once released from plants, play a major role in atmospheric particle formation (Ehn et al. 2014).

All plants emit a plethora of volatiles that are synthesized in different subcellular compartments involving multiple biochemical pathways (Fig. 1), and the emissions can be further tissue- and organ-specific. The volatiles emitted can be intermediates of normal plant metabolic activity and are released from plant tissues because the metabolic pathways are “leaky.” Emissions of such compounds can be enhanced under certain periods of plant life. For instance, plants emit methanol as the result of demethylation of cell wall pectins (Nemecek-Marshall et al. 1995; Fall and Benson 1996). Methanol emissions occur at low level from all physiologically active plant tissues, but the emissions are strongly enhanced in growing tissues due to relaxation and rigidification of cell walls during tissue expansion growth (Harley et al. 2007; Hüve et al. 2007). Analogously, root zone hypoxia during flooding leads to ethanol formation in the roots and its transportation to the leaves with the transpiration stream (Bracho Nunez et al. 2009; Kreuzwieser and Rennenberg 2013). Ethanol that reaches the leaves can be further enzymatically oxidized to acetic acid via acetaldehyde, and enter into the primary metabolism, but some ethanol and acetaldehyde escape leaves due to limited alcohol and aldehyde dehydrogenase activities (Kreuzwieser et al. 2000, 2001; Rottenberger et al. 2008; Kreuzwieser and Rennenberg 2013).

Fig. 1
figure 1

Main volatiles emitted by plants are formed in different subcellular compartments and their synthesis involves a variety of biochemical pathways . Methanol and lipoxygenase (LOX) pathway volatiles (also called green leaf volatiles, dominated by various C6 aldehydes and alcohols) are ubiquitous volatiles that can be released from all plant tissues as the result of constitutive activity of key enzymes of their synthesis pathways, while volatile isoprenoids (hemiterpenes like isoprene, monoterpenes, and sesquiterpenes) and phenylpropanoids and benzenoids are specialized volatiles and are emitted as the result of induction of expression of genes coding for specific rate-limiting synthases, although in several species, certain specialized volatiles can be emitted constitutively. Methanol is released as the result of demethylation of pectins in cell walls in growing tissues or upon different biotic and abiotic stresses, whereas green leaf volatiles are formed from free polyunsaturated fatty acids released from membrane lipids upon membrane damage characteristic to exposure to severe stresses (Liavonchanka and Feussner 2006; Andreou and Feussner 2009). Emissions of phenolic compounds typically originate from cytosol, but isoprenoid emissions can originate from plastids or cytosol, depending on the compound emitted (Niinemets et al. 2013; Rosenkranz and Schnitzler 2013; Pazouki and Niinemets 2016). There is evidence that some terpenoids can be potentially also released from mitochondria (not shown in the figure, see Tholl and Lee 2011; Dong et al. 2016), but the possible mitochondrial release of volatiles is much less studied

Full size image

These examples demonstrate how normal physiological processes in plant life, and the ecosystem services they provide, can be associated with major leakage of organic compounds due to relatively high vapor pressure of these compounds. Apart from metabolic intermediates, a large number of organic compounds are specifically made to be volatile, including several volatile benzenoids and phenylpropanoids and volatile isoprenoids such as the hemiterpene isoprene (C5), monoterpenes (C10), and sesquiterpenes (C15) (Fig. 1, Peñuelas and Llusià 2004; Fineschi et al. 2013; Guenther 2013a; Portillo-Estrada et al. 2015). These specialized volatiles constitute the plant “talk,” fulfilling a plethora of biological and ecological functions from communication among plant organs, other plants, and other organisms to altering the plant stress resistance. Furthermore, all volatiles participate in multiple landscape- to global-scale processes, altering ambient air oxidative status, atmospheric particle condensation, and cloud cover (Peñuelas and Staudt 2010; Kulmala et al. 2013). Both the specific biological and broad-scale atmospheric roles of volatiles result in a number of key ecosystem services beneficial to humans. As discussed in this chapter, these services can be local to regional scale services such as preservation of ecosystem integrity under biotic and abiotic stresses and, thus, the preservation of the capacity to provide the “standard,” well-perceived, ecosystem services to humans, e.g., wood production of forest stands. These services also include regional to global-scale services such as environmental cooling and dampening the global climate change .

In the current chapter, I first provide a short overview of key types of biological volatile emission and of the environmental controls on volatile emission and modification of emissions by abiotic and biotic stresses. Then I focus on the biological roles of volatiles, on the roles of volatiles in large-scale biosphere-atmosphere processes , and on ecological services provided by plant volatiles. I demonstrate that the trace gas release is a key vegetation characteristic that contributes a number of unique ecosystem services that alter the performance of ecosystems in current and future climates with major implications for human life. Quantitative significance of plant volatiles in Earth system processes is still poorly understood, and the role of plant volatiles in climate has been largely neglected in the last report of the Intergovernmental Panel on Climate Change (IPCC, Field et al. 2014; Stocker et al. 2014). In this chapter, I show that the evidence is accumulating that plant trace gas exchange participates in multiple feedback loops that can potentially play major roles in Earth system processes and argue that biosphere-atmosphere interactions mediated by plant volatiles need to be included in Earth system models intending to predict future climate.

2 Plant Volatile Diversity, Environmental Controls on Emission and Emission Capacities

2.1 Sites of Volatile Synthesis and Diversity

Synthesis of volatile phenolics typically occurs in cytosol, while volatile isoprenoids can be synthesized in plastids or cytosol, depending on the compound emitted (Niinemets et al. 2013; Pazouki and Niinemets 2016). Hemiterpene and monoterpene synthesis is considered to occur in the plastids where corresponding C5 and C10 terminal enzymes responsible for terpene synthesis, terpene synthases, are located, while sesquiterpene synthases are located in the cytosol (Fig. 1, Chen et al. 2011). However, there is also recent evidence of mixed substrate specificity of some terpene synthases (Pazouki and Niinemets 2016), suggesting that product profiles can potentially vary depending on substrate availability in different subcellular compartments. Furthermore, there is evidence that some mono- and sesquiterpenes can be synthesized in mitochondria (Tholl and Lee 2011; Dong et al. 2016), further complicating the picture.

The diversity of emitted volatiles varies for different volatile compound classes. In the case of hemiterpenes, in addition to isoprene, plants also emit the oxygenated hemiterpene 2-methyl-3-buten-2-ol (Gray et al. 2006, 2011), but mono- and sesquiterpene-emitters typically release a wide spectrum of compounds. Often more than 20 different terpenoids are observed in the emissions from a single species (e.g., Niinemets et al. 2002a; Winters et al. 2009). Such a high diversity reflects the presence of multiple terpene synthases in the given emitting species (e.g., Falara et al. 2011; Jiang et al. 2016a) but also the specific reaction mechanism of terpene synthesis. In particular, terpene synthesis involves formation of a highly reactive carbocation intermediate, and depending on the extent to which the carbocation can be stabilized, the product specificity of terpene synthases strongly varies (e.g., Christianson 2008). In fact, most terpene synthases catalyze formation of multiple products, and only some terpene synthases form single terpenes, explaining the huge chemical diversity of emitted volatile terpenoids.

2.2 Constitutive and Stress-Induced Volatile Emissions

2.2.1 Constitutive Emissions of Specialized Volatiles

While all plant species can emit metabolic intermediates, only some species emit specialized volatiles, in particular, volatile isoprenoids under typical non-stressed physiological conditions, being, thus, constitutive emitters (Peñuelas and Llusià 2004; Fineschi et al. 2013). The capacity for constitutive emission of certain volatiles requires that the specific synthesis pathways are constitutively active, although the degree of activation can vary with environmental conditions, sometimes several-fold (e.g., Niinemets et al. 2010b). The constitutive emissions can result from emission of compounds stored in specialized storage structures such as oil glands in Citrus species, resin ducts in conifers, and glandular trichomes in species from Labiatae or Solanaceae. Typically, volatiles stored in these structures are mono- and sesquiterpenes, but sometimes benzenoids are also stored (Loreto et al. 2000; Jiang et al. 2016a). Filling up the storage structures takes typically multiple days to months, and thus, the release of the volatiles from the storage is uncoupled from the synthesis of these compounds. Thus, the rate of constitutive emissions from storage structures depends on the rate of compound evaporation and diffusion and therefore scales exponentially with temperature (Niinemets et al. 2010b; Grote et al. 2013).

The hemiterpenes, isoprene and 2-methyl-3-buten-2-ol, cannot be stored due to high volatility. In addition, several constitutive monoterpene emitters such as the Mediterranean evergreen oaks Quercus ilex and Q. suber and the broad-leaved deciduous temperate tree Fagus sylvatica lack specialized storage structures (Staudt et al. 2004; Dindorf et al. 2006). In hemiterpene and non-storage monoterpene emitters, the volatile emissions result from de novo compound synthesis, and the emissions respond to environmental drivers similarly to photosynthesis, increasing asymptotically with light intensity and scaling positively with temperature up to an optimum temperature and decreasing thereafter (Niinemets et al. 2010b; Monson et al. 2012). Emissions of de novo synthesized volatiles also respond to immediate changes in ambient CO2 concentration. In particular, isoprene emissions decrease with increasing CO2 concentration (e.g., Wilkinson et al. 2009), but the CO2 response is less clear for monoterpene emissions (Sun et al. 2012 for a discussion). However, the CO2 sensitivity of isoprene is gradually lost at higher temperatures (Rasulov et al. 2010; Li and Sharkey 2013), and the emission response to longer-term changes in ambient CO2 concentration can be different from the immediate response due to acclimation of isoprene synthesis pathway to long-term ambient CO2 concentration (Sun et al. 2012).

2.2.2 Induction of Volatiles Upon Abiotic and Biotic Stresses

In field environments, plants are often exposed to various abiotic and biotic stresses . Although only some plant species can emit volatiles constitutively, all species can be triggered to release volatiles upon biotic and abiotic stresses (Fig. 2). Among the volatiles triggered are ubiquitous stress compounds such as lipoxygenase (LOX) pathway volatiles (also called green leaf volatiles). Green leaf volatiles, typically dominated by various C6 aldehydes and alcohols, are rapidly formed when free polyunsaturated fatty acids are released from membrane lipids (Beauchamp et al. 2005; Copolovici et al. 2011). This typically occurs upon membrane-level damage characteristic to exposure to severe stresses such as mechanic wounding during biotic stresses but also upon exposure to severe heat, frost, and ozone stress (Beauchamp et al. 2005; Copolovici et al. 2011). Methanol , which can be released from non-stressed tissues constitutively, also serves as another ubiquitous stress volatile; major rapid methanol bursts are associated with both severe biotic and abiotic stresses (Beauchamp et al. 2005; von Dahl et al. 2006).

Fig. 2
figure 2

Induction of monoterpene emissions in broad-leaved temperate deciduous tree Alnus incana upon feeding by larvae of the moth Cabera pusaria (a), and correlation of the degree of elicitation of emissions with the amount of leaf area consumed on the fourth day since the start of feeding (b). Alnus incana does not significantly emit volatiles in non-stressed conditions and is therefore considered a non-emitter species, but exposure to different stresses results in significant emissions of green leaf volatiles, mono-, sesqui-, and homoterpenes (Modified from Copolovici et al. 2011)

Full size image

Apart from the ubiquitous compounds emitted from virtually all plant species, induction of synthesis of specialized compounds requires the presence of given stress-elicited synthase genes in plant genomes. For instance, glucosinolate pathway volatiles (short-chained S- and/or N-containing volatiles resulting from breakdown of glucosylated amino acids and their derivatives) are only released upon stress from the species in the order Brassicales (Kask et al. 2016). On the other hand, all plants do include terpene synthase genes in their genomes, and expression of these genes is typically activated upon stress. Studies have demonstrated elicitation of emissions of mono- and sesquiterpenes upon biotic stresses such as herbivory (Fig. 2, Blande et al. 2007; Copolovici et al. 2014a, 2011; Farré-Armengol et al. 2015); pathogen infections such as powdery mildew (Fig. 3), leaf rust (Fig. 4), and canker fungus (Achotegui-Castells et al. 2015) infections; and upon abiotic stresses such as ozone stress (Beauchamp et al. 2005) and heat stress (Copolovici et al. 2012; Kask et al. 2016). However, the number of terpenoid synthase genes strongly varies among species from as few as only one synthase gene to more than 80 genes (Rajabi Memari et al. 2013), implying that the diversity of induced terpenoid emission responses can also be variable. Furthermore, different stresses can trigger emissions of different volatiles (Dicke et al. 2009; Zhang et al. 2009), but so far, understanding of the overall stress-dependent emission diversity and variation of emission profiles under different stresses is very limited.

Fig. 3
figure 3

Oak powdery mildew (Erysiphe alphitoides) is a major pathogen infesting pedunculate oak (Quercus robur) all over Europe. The visual damage symptoms can be detected through the growing season in young oak trees (a), and in late summer and autumn in old plants (b). Erysiphe alphitoides infections are associated with reduction of constitutive emissions of isoprene (c) and elicitation of emissions of monoterpenes (d), sesqui- and homoterpenes (data not shown) and green leaf volatiles (e) ((c–e) Modified from Copolovici et al. 2014b). Open symbols denote non-infected leaves

Full size image
Fig. 4
figure 4

Poplar (Populus spp.) is the telial host for the widespread rust fungus Melampsora larici-populina, infections of which are initially associated with diffusely spread yellow spots (a), followed by extensive leaf yellowing and necrosis and premature leaf senescence encompassing large parts of the canopy (encircled area in b). Melampsora larici-populina infection reduces constitutive emissions of isoprene (c) and induces emissions of monoterpenes (d), sesqui- and homoterpenes (data not shown) and green leaf volatiles (e) in an infection-dependent manner (Jiang et al. 2016b). Open symbols refer to emissions from leaves without visible signs of infection. Pictures (a) and (b) are for P. laurifolia and data (c–e) for P. balsamifera var. suaveolens

Full size image

Differently from LOX volatiles and methanol , emissions of which are triggered rapidly due to the presence of a certain constitutive activity of lipoxygenases (Liavonchanka and Feussner 2006) and pectin methylesterases (Micheli 2001), elicitation of terpenoid emissions is more time-consuming because it requires gene expression and protein synthesis to reach a certain terpene synthase activity. Typically, emissions of terpenoids can be detected hours after the stress impact and the emissions peak 24–48 h after the impact (Fig. 2, Pazouki et al. 2016). On the other hand, when the stress is relieved, the emissions gradually decrease, reaching the initial non-induced level in a few days after the stress relief (Fig. 2a). This reduction is a characteristic feature to induced emissions and contrasts to constitutive emissions. Although the rate of constitutive emissions can also be affected by stress, typically negatively the level of constitutive emissions almost never reaches zero.

2.3 Relationships Among Constitutive and Stress-Induced Emissions

As noted above, constitutive emissions are present only in several plant species, while induced emissions can be elicited in all species, including the constitutive emitters . Typically, the composition of induced emissions is different from constitutive emissions. For instance, in constitutively isoprene-emitting deciduous oak Quercus robur, infection by oak powdery mildew (Erysiphe alphitoides) results in emissions of mono- and sesquiterpenes (Fig. 3). Analogously, infection of the constitutively isoprene-emitting poplar (Populus spp.) with the rust fungus Melampsora larici-populina, results in mono- and sesquiterpene emissions (Fig. 4).

In the case of constitutive monoterpene emitters , an environmental or a biotic stress often results in elicitation of emissions of monoterpenes different from constitutively emitted monoterpenes. In particular, typical stress-elicited monoterpenes are ocimenes, linalool, and 1,8-cineole, while constitutive emissions are characteristically dominated by limonene and pinenes (Staudt and Bertin 1998; Niinemets et al. 2002b). Importantly, in constitutive storage emitters, stress-induced monoterpene emissions reflect de novo synthesis of volatiles and scale similarly with light and temperature as the emissions in non-storage emitters (Niinemets et al. 2010a, b). In addition to monoterpenes, stress often results in elicitation of emissions of sesquiterpenes and homoterpenes 4,8-dimethylnona-1,3,7-triene (DMNT) and 4,8,12-trimethyltrideca-1,3,7,11-tetraene (TMTT) that are not observed in constitutive emissions (e.g., Niinemets et al. 2010b; Staudt and Lhoutellier 2011; Tholl et al. 2011).

In both of the biotic stress case studies highlighted here, the constitutive isoprene emissions decreased in pathogen-infected leaves (Figs. 3 and 4), and analogous negative relationships between induced and constitutive emissions have been demonstrated in the case of other stresses as well (e.g., Kleist et al. 2012 for heat stress). Overall, positive stress dose vs. induced emission relationships have been observed for several abiotic stresses such as frost and heat stress (Copolovici et al. 2012) and ozone stress (Beauchamp et al. 2005). Although biotic impacts have been considered to be hard to quantify (Niinemets et al. 2013), quantitative stress dose vs. induced emission responses have been observed for several biotic stresses such as herbivory (Fig. 2) and fungal pathogen infections (Figs. 3 and 4). The key issue with the biotic stresses seems to be how to quantify the severity of biotic stress (Copolovici and Niinemets 2016), but once the biotic stress severity has been properly characterized, it becomes clear that the rate of emission of volatile organic compounds scales with stress severity similarly to abiotic stresses (Niinemets et al. 2013; Copolovici and Niinemets 2016).

3 Ecosystem Services and Impacts of Plant Volatiles

3.1 Biological Role of Plant Volatiles in Ecological Processes

The biological role of constitutive emissions of isoprene and monoterpenes is not yet fully clear, but there is increasing evidence that these compounds have antioxidative and/or membrane-stabilizing properties (Sharkey et al. 2001; Loreto et al. 2004; Vickers et al. 2009; Peñuelas and Staudt 2010), and it has been postulated that they play an important role in enhancing the abiotic stress resistance in species emitting them (Vickers et al. 2009; Possell and Loreto 2013). Furthermore, volatiles released from constitutively emitting species can be taken up from the ambient air by neighboring non-emitting species (Copolovici et al. 2005; Noe et al. 2008; Himanen et al. 2010). This can enhance the stress resistance of non-emitters, resulting in an overall increase in ecosystem stress resistance. Such an increase of ecosystem stress resistance is the key ecological service which allows for maintenance of all other well-recognized ecosystem services provided by healthy ecosystems.

While the constitutive emissions can provide certain protection against chronic mild abiotic stresses that do not lead to induction of stress volatiles, once induced, chemically similar induced volatiles can also directly participate in stress protection during short-term severe stresses that trigger their emission. However, more importantly, stress-elicited volatiles play various functions in communication among plants and plants and other organisms (Dicke and Baldwin 2010; Holopainen et al. 2013; Blande et al. 2014). In plants, stress-induced volatiles serve as infochemicals eliciting stress response pathways, leading to plant acclimation to the altered environmental conditions and priming defenses against herbivore and pathogen attacks in leaves and neighboring not yet stressed plants (Dicke and Baldwin 2010; Peng et al. 2011). Such a defense priming can again augment the whole ecosystem resistance to both abiotic and biotic stresses. Furthermore, herbivory-induced volatiles serve as infochemicals for the enemies of herbivores, and thus, the release of attractants to predatory and parasitic insects can importantly reduce the spread of herbivores (D’Alessandro et al. 2006; Dicke and Baldwin 2010) and, thus, provides a further important means for enhancing the resilience of ecosystems.

3.2 Plant Volatiles in Broad-Scale Ecological Processes

Apart from the biological role of plant-emitted volatiles, plant-generated volatiles play important roles in large-scale regional and global processes. The global amount of emitted biogenic volatiles has been estimated to be roughly 1.1 Pg yr−1 (equivalent to ca. 0.84 Pg C yr−1) (Guenther et al. 2012). Thus, the biogenic release of trace gases exceeds the anthropogenic release by more than a factor of ten (Guenther et al. 2012). Despite only certain species are capable of constitutive isoprenoid emissions , many of these species are widely distributed, often being the dominating plant species in given ecosystems. In fact, global plant emissions are currently derived using species-specific emission potentials obtained by extensive screening studies (e.g., Karlik and Winer 2001; Simon et al. 2006; Keenan et al. 2009; Llusià et al. 2010, 2014). The emphasis in these screening studies has been on constitutive emissions, and thus, global emission estimates are mainly based on constitutive emitters (Guenther et al. 2012).

Indeed, constitutively emitted isoprene is the most important plant-generated volatile compound with the global source strength predicted to be ca. 550 Tg yr−1 by different models (Arneth et al. 2008; Guenther et al. 2012; Guenther 2013b). While different global emission models based on profoundly different emission algorithms converge to a similar value of global isoprene emission, due to uncertainties in the share of storage vs. de novo emissions and constitutive vs. stress-induced emissions , the model estimates are more variable for mono- and sesquiterpenes than for isoprene (Arneth et al. 2008). Based on empirical model approaches, total emissions of ca. 160 Tg yr−1 for mono- and ca. 30 Tg yr−1 for sesquiterpenes have been estimated (Guenther et al. 2012). The rest of global BVOC emission source strength of ca. 0.4 Pg yr−1 is mainly made up of oxygenated compounds, dominated by methanol , ethanol, acetone, and acetaldehyde (Guenther et al. 2012).

The largest uncertainty in the global volatile emission estimates seems to be the lack of proper consideration of stress-induced emissions. As shown above, exposure to stress conditions can turn a constitutively non-emitter or moderate emitter species into a strong emitter of mono- and sesquiterpenes. This might mean that the overall capacity of vegetation to emit volatiles has been strongly underestimated. Of course, the stress-elicited emissions occur only when there is a stress and emissions return to background levels when the stress is relieved (e.g., Fig. 2), but it is relevant to consider that these relatively short-termed emission peaks might not only importantly contribute to the total emission amount, but alter the timing of peak atmospheric concentrations of volatiles with major consequences for large-scale physiological and atmospheric processes.

3.2.1 Role of Volatiles in Altering Atmospheric Reactivity

Due to the large emissions, biogenic volatiles play major roles in biosphere-atmosphere processes . The chemical reactivity of non-oxygenated non-saturated compounds, in particular, non-oxygenated isoprenoids, is much larger than the reactivity of oxygenated volatiles, and thus, the role of different compound classes, oxygenated vs. non-oxygenated non-saturated compounds, in large-scale processes is different. Highly reactive compounds play a major role in ozone formation and quenching reactions in the troposphere. In particular, in a human-polluted air enriched with NOx (sum of NO and NO2), plant-generated volatiles contribute to photochemical ozone production and, in fact, control the rate of ozone formation in the atmosphere (Chameides et al. 1992; Fehsenfeld et al. 1992; Fuentes et al. 2000). In contrast, in non-polluted air with low NOx, reactive hydrocarbons contribute to a reduction of ozone concentrations (Lerdau and Slobodkin 2002; Atkinson and Arey 2003; Loreto and Fares 2007).

The influence of elevated ozone concentration on photosynthetic productivity and volatile emissions can be different over the short- and the long-term. Increases in ozone concentration can strongly reduce plant photosynthetic productivity, but constitutive volatile emission itself can initially increase under moderately elevated atmospheric ozone concentration (Calfapietra et al. 2013 for a review). This might have global consequences as a reduction in photosynthetic CO2 fixation can speed up elevation of atmospheric CO2 concentration, and this in turn can increase the rate of temperature increase (Sitch et al. 2007). Due to the positive effect of temperature on volatile emissions, this is further expected to enhance volatile release and ozone formation and reduce carbon gain even more (Lerdau 2007; Sitch et al. 2007). Furthermore, the concentration of ozone, significantly driven by the concentration of reactive volatiles, is itself an important greenhouse gas that can contribute to global warming (Shindell et al. 2006), amplifying these patterns.

A severe ozone stress results in a reduction of constitutive volatile emissions as well (Calfapietra et al. 2013 for a review), but it also leads to elicitation of induced emissions in both constitutive emitters and non-emitters (Beauchamp et al. 2005; Hartikainen et al. 2009). Among the induced volatiles, mono- and sesquiterpenes are typically more reactive in ozone formation reactions than isoprene (Calogirou et al. 1996) and, thus, could temporarily even speed up ozone formation, especially because these emissions are induced in all species in vegetation. So far, the understanding of ozone-dependent modifications in constitutive and induced emissions is only rudimentary, limiting quantitative assessment of species physiological responses in ozone formation potential of vegetation . Nevertheless, the available evidence suggests that the use of static emission factors estimated from screening studies that have considered only constitutive emitters and omission of physiological modifications in volatile emissions driven by ozone can lead to major uncertainties in predicting tropospheric ozone formation and quenching.

3.2.2 Volatiles in Altering Solar Radiation Scattering and Penetration and Ambient Temperature

Upon oxidation, the volatility of isoprenoids dramatically decreases, implying that they partition much more strongly to the liquid and solid phases than to the gas phase, creating secondary organic aerosols (SOA) (Kulmala et al. 2004a; Chen and Hopke 2009; Mentel et al. 2009; Kirkby et al. 2016). The presence of SOA decreases atmospheric clearness, thereby potentially reducing solar radiation penetration, but also increasing light scattering and, thus, the diffuse to total solar radiation ratio (Fig. 5, Malm et al. 1994; Farquhar and Roderick 2003; Misson et al. 2005). Because diffuse radiation penetrates deeper into the plant canopies and results in a more uniform distribution of solar radiation (Cescatti and Niinemets 2004), increases in diffuse to total solar radiation ratio enhance vegetation productivity (Gu et al. 2002, 2003; Mercado et al. 2009). Due to strong light effects on constitutive de novo volatile emissions , greater canopy light interception is expected to directly enhance these emissions (Fig. 5). In fact, the emissions are expected to increase even more than the rate of photosynthesis, because de novo volatile emissions are more light-limited than photosynthetic carbon fixation, typically saturating at higher light intensities than photosynthesis (Niinemets et al. 1999, 2015). Although the connection between photosynthesis and storage emissions is less straightforward, enhanced carbon availability can also enhance the emissions in the storage emitters by increasing the size of the storage pools (Blanch et al. 2007, 2009, 2011).

Fig. 5
figure 5

Regional and global feedback relationships between constitutive biogenic volatile organic compound (BVOC) emissions and climatic drivers . The rates of BVOC emission are controlled by environmental drivers, whereas environmental drivers, in turn, are modified by BVOC emission generating feedback loops. There are two main types of BVOC emitters, the storage emitters and the de novo emitters (Grote et al. 2013; Copolovici and Niinemets 2016). In the storage emitters, BVOC is released from large storage pools, and the emissions depend only on temperature, increasing exponentially with increasing temperature (Niinemets et al. 2010b; Grote et al. 2013). In the case of de novo emitters, the emissions rely on immediately synthesized BVOCs and depend both on light intensity and temperature. These emissions increase asymptotically with increasing light intensity and increase exponentially with temperature up to an optimum temperature and decline thereafter (Niinemets et al. 2010b; Grote et al. 2013). For simplicity, the figure shows only the environmental controls on BVOC emission for de novo emitters. In the case of light, light interception by plant canopies depends on total solar radiation flux and the distribution of solar radiation between diffuse and direct components. Diffuse light drives canopy photosynthesis more efficiently because of its penetration to deeper canopy layers and resulting greater uniformity of radiation field and total light interception (Gu et al. 2003). Atmospheric volatiles enhance concentration of secondary organic aerosols (SOA) that increase the radiation scattering, but SOA also moderately reduce total radiation penetration through atmosphere (Spracklen et al. 2008; Chen and Hopke 2009; Kulmala et al. 2013). Increasing concentrations of SOA, in turn, enhance the concentration of cloud condensation nuclei (CCN) thereby contributing to enhanced cloudiness and thus, increasing radiation scattering too, but CNN more strongly reduce total solar radiation flux penetrating through the atmosphere to the vegetation (Roderick et al. 2001; Huff Hartz et al. 2005; Spracklen et al. 2008; Kulmala et al. 2013). The other key environmental driver, temperature, depends on total solar radiation flux, but can also directly depend on CCN as the result of reduced thermal radiation losses to the space, but this effect is more relevant at night and is therefore not shown in the figure. In addition to the feedbacks shown, more complex feedbacks can operate through modifications in atmospheric CO2 and water vapor concentrations (e.g., Kulmala et al. 2013) as well as through induction of BVOC emissions upon abiotic and biotic stresses, severity of which depends on environmental drivers (section “Constitutive and Stress-Induced Volatile Emissions”)

Full size image

Once formed in the atmosphere, the size of SOA particles increases in time due to condensation of atmospheric organics on particle surface (Kulmala et al. 2004a; Kirkby et al. 2016). These bigger particles can also serve as cloud condensation nuclei (CCN) , especially if their hygroscopicity also increases as the result of further compound oxidations on particle surface or as more hydrophilic compounds condense onto the particle surface (Engelhart et al. 2008; Kulmala et al. 2013). Greater CCN concentrations imply a greater condensation sink and potentially higher cloudiness that can dramatically increase the diffuse to total solar radiation ratio but also strongly reduce the total solar radiation penetration (Spracklen et al. 2008; Still et al. 2009; Kulmala et al. 2013). Thus, although radiation penetration into deeper canopy layers is increased by enhanced cloudiness, the reduction in total radiation intensity reaching to the top of the vegetation is the dominating factor, ultimately reducing the vegetation productivity and the rate of volatile emission (Fig. 5). Furthermore, cooling due to increased cloudiness also directly reduces both de novo and storage volatile emissions (Fig. 5).

As the result of volatile effects on SOA and CCN concentrations , multiple feedback loops operate between solar radiation, temperature, volatile emission, and productivity at regional and global scales, and the overall effect of volatiles on climate depends on the relative significance of these loops (Fig. 5, Kulmala et al. 2013). In particular, both rising SOA concentrations and greater cloudiness can reduce the surface temperature, and this can directly reduce volatile formation due to the physiological controls on the emission rates, and this, in turn, is expected to inhibit further SOA and CCN formation (Fig. 5, Kulmala et al. 2013). On the other hand, enhanced SOA concentrations increase the fraction of diffuse radiation, thereby increasing the rate of volatile release and further enhancing SOA and CCN formation (Fig. 5, Kulmala et al. 2013). In contrast, dramatic reductions in total solar radiation by enhanced cloudiness are expected to lead to decreased volatile formation, thereby feedback-inhibiting SOA and CCN formation . Testing the quantitative significance of these feedback loops requires a combination of long-term data with regional- and global-scale modeling. The first such modeling exercise based on 15 years of measurements of vegetation carbon fixation fluxes and 6 years of measurements of emission fluxes of volatile organic compounds in a boreal forest ecosystem suggests that such feedback loops do indeed operate in nature (Kulmala et al. 2013). Due to both physiological and acclimation responses of volatile emissions to changes in environmental drivers, quantitative prediction of the feedback responses is complex, and clearly, more such case studies are needed to scale up to other regions and the globe and to quantitatively evaluate the role of volatile emissions on solar radiation penetration and surface temperature.

3.2.3 Volatiles as Modifiers of the Lifetime of Greenhouse Gases

As demonstrated above, reactive plant volatiles in polluted atmospheres can exacerbate the plant abiotic stress due to enhancing the key atmospheric pollutant, ozone, concentrations, while reactive volatiles in non-polluted atmospheres can reduce ozone concentrations. Thus reactive plant hydrocarbons can alter the rate of global warming by changes in vegetation CO2 fixation capacity and, thus, by long-term modifications in atmospheric CO2 concentration. Furthermore, ozone itself is a strong greenhouse gas (section “Constitutive Emissions of Specialized Volatiles”, Shindell et al. 2006). In addition, alteration in solar radiation intensity, diffuse to total solar radiation ratio, and ambient temperature due to modifications in SOA and CCN concentrations directly affect the vegetation carbon fixation capacity and, thus, can strongly alter atmospheric CO2 concentration as well, creating further major feedback loops (Kulmala et al. 2004b, 2013).

Apart from the highly reactive compounds, less reactive oxygenated hydrophilic compounds such as aldehydes, alcohols, and ketones, including, for instance, lipoxygenase pathway volatiles, but also saturated oxygenated volatiles, can also partition to particle phase and participate in SOA and CCN formation (Mentel et al. 2009). In addition, several of these volatiles can react with atmospheric OH radicals, reducing atmospheric OH radical concentration (Fall 2003; Sinha et al. 2010; Nölscher et al. 2012). Given that the reaction with OH radicals is the primary process reducing the atmospheric concentrations of the key greenhouse gas methane , a reduction of OH concentration due to biogenic volatiles increases the methane lifetime in the atmosphere (Jacob et al. 2005; Ashworth et al. 2013; Voulgarakis et al. 2013), thereby significantly contributing to global warming. In particular, plant-generated emissions of methanol , the oxygenated volatile with the greatest global source strength (Stavrakou et al. 2011; Guenther et al. 2012), can potentially contribute to the greatest degree to the increases in methane lifetime.

3.3 Trace-Gas-Driven Ecosystem Services

The vegetation capacity to emit volatiles has not generally been considered as an ecosystem service, and overall, the atmosphere is often not considered as part of ecosystem services (Cooter et al. 2013). In fact, due to the contribution of volatiles to ozone formation in NOx-polluted atmospheres, volatile emission has even been considered an ecosystem “disservice” (Russo et al. 2016). However, from a biological perspective, plants can provide several key ecosystem services due to their trace gas emission. Although these ecosystem services are little-recognized in the community, they can have profound impacts on ecosystem performance. Among these biological services are:

  • Direct enhancement of plant abiotic stress resistance by directly quenching reactive oxidative species generated in plant membranes upon abiotic stress (Vickers et al. 2009) or more specifically, improving heat stress tolerance by increasing membrane stability (Velikova et al. 2012; Sun et al. 2013)

  • Increases in ecosystem resilience through communication among plants and with other organisms (e.g., plants and herbivore enemies), thereby ameliorating the biotic and abiotic stress impacts

  • Enhancement of ecosystem capacity to uphold diversity, in particular, to maintain the integrity of multitrophic interactions

Although often “hidden,” these services are crucial for stability and performance of ecosystems, and impairing some of these services can result in drastic deterioration of other ecosystem services. For instance, the carbon gain of chemically less diverse ecosystems such as monospecific tree plantations or crop fields can be much more vulnerable to deleterious pest attacks than that of more diverse ecosystems (Lerdau and Slobodkin 2002; Altieri and Nicholls 2004; Tooker and Frank 2012). This is highly relevant from the perspective of the “traditionally” perceived ecosystem services as pulpwood production of tree plantations or yield of crop fields, both of which are directly dependent on vegetation carbon gain.

From the perspective of atmospheric chemistry and large-scale biosphere-atmosphere processes , the balance between ecosystem service and “disservice” of volatile emissions importantly depends on the relative extent to which different atmospheric processes are affected by volatiles. The balance between different processes strongly varies regionally and depends on human effects on atmosphere. While in urban NOx-polluted atmospheres, volatiles emitted by vegetation contribute to elevated ozone concentration and photochemical smog, in remote non-polluted atmospheres with low NOx levels, plant-generated volatiles are expected to reduce atmospheric oxidant concentrations, including ozone concentrations (Lerdau and Slobodkin 2002). Analogously, reduction in atmospheric clearness and alteration of ambient air particle concentration by volatile contribution to SOA formation could be considered a disservice in urban habitats (Cooter et al. 2013), although it might also contribute to cooling of urban environments (Arneth et al. 2009).

From a global perspective, plant volatiles can provide three key ecosystem services:

  • Improvement of ecosystem capacity to fix carbon by altering diffuse/direct radiation due to SOA formation

  • Cooling of environment through production of SOA and CCN

  • Reduction of the rate of global climate change through improved carbon gain and reduced transmission of solar radiation

The ultimate significance of these services depends on the relative magnitude of different environmental changes and volatile emissions as connected through multiple feedback loops (Fig. 5). It is, furthermore, highly likely that globally changing environmental drivers and CO2 concentration alter the quantitative significance of the feedback loops due to the modifications in plant stress status, carbon fixation, and trace gas release as discussed in the next section.

3.4 Plant Stress, Volatile Emissions, and Trace-Gas-Dependent Ecosystem Services in Changing Climates

Climate change is expected to result in more severe heat stress worldwide (Kirtman et al. 2013; Field et al. 2014). In addition, climate change alters the distribution of precipitation with some areas predicted to become drier and some areas wetter (Kirtman et al. 2013; Field et al. 2014). Thus, the overall abiotic stress level is expected to increase in the future, but prediction of how enhanced stress level modifies volatile emission and ecosystem services is complicated. Increases in temperature can initially result in enhanced emissions of constitutive de novo emissions until the physiological optimum is exceeded (Loreto et al. 1998; Staudt and Bertin 1998; Rasulov et al. 2015). Beyond the physiological optimum, the rate of constitutive emissions decreases, but the stress volatile emissions are induced (Staudt and Bertin 1998; Kleist et al. 2012). These emissions under heat stress have been detected at the ecosystem scale (Karl et al. 2008), demonstrating that the stress emissions do contribute to large-scale atmospheric processes. Given that the stress emissions are induced in all plant species, more frequent heat waves in future climates can strongly enhance the overall release of volatiles. However, the quantitative information on the kinetics of elicitation of emissions under heat stress and interspecific variability of the capacity for heat-dependent release of stress volatiles is currently very limited, hindering scaling up from case studies to whole ecosystems, regions, and globe.

Global changes in environmental drivers have to be tempered in light of simultaneous modifications of ambient CO2 concentrations. Elevation of atmospheric CO2 concentration itself can improve plant carbon gain in drier climates due to reduction of the diffusion limitations on the CO2 pathway from the ambient air through stomata and mesophyll to the chloroplasts where photosynthesis takes place (Niinemets et al. 2011; Flexas et al. 2016). In addition, elevated CO2 can protect leaves from the heat stress, possibly by increasing leaf sugar concentrations that enhance the heat stability of cell and chloroplast membranes (Darbah et al. 2010; Sun et al. 2013). In the case of constitutive isoprene emissions , however, several studies have demonstrated that the capacity for isoprene emission decreases with a long-term increase in ambient CO2 concentration similarly to the response of isoprene emissions to rapid changes in CO2 concentration (Niinemets et al. 2010a; Possell and Hewitt 2011). Such an acclimatory response would mean that the elevation of CO2 concentration impairs the heat stress protection by isoprene. However, in other studies, plants grown under elevated CO2 had greater isoprene emission potential and improved heat stress resistance (Peñuelas and Staudt 2010; Sun et al. 2012, 2013). These controversial responses are evident when comparing model projections of emissions, which diverge greatly between models under future climate change (Keenan et al. 2009). This implies that prediction of constitutive emissions in future atmospheres is subject to large uncertainties and calls for more work on acclimation responses of constitutive emissions.

Apart from abiotic stresses, global environmental change is predicted to result in increasingly more frequent and severe pest and pathogen attacks (DeLucia et al. 2008; Chakraborty 2013). Such a greater biotic stress pressure can result from a reduction of constitutive defenses of vegetation due to reduced photosynthetic carbon availability under more severe abiotic stresses, shorter life cycles of current pests, and pathogens in hotter climates as well as invasion of new pests and pathogens facilitated by global trade and travel (Fig. 6, Vanhanen 2008; Huang et al. 2010; Gutierrez and Ponti 2014). Although biotic stress itself typically elicits volatile emissions for a relatively short period of time (Fig. 2), as the result of greater biotic stress pressure, the frequency of multiple sequential and simultaneous biotic attacks is likely to increase (Fig. 6). Thus, in the future climates , the emissions from biotic stresses are expected to continue longer and contribute to a larger extent to the overall emission of plant-produced volatiles.

Fig. 6
figure 6

Illustration of single and multiple biotic infections of A. incana leaves in the field. Mass infestations by the alder leaf beetle (Agelastica alni, a) and alder rust (Melampsoridium hiratsukanum, b, c) are frequently observed in European alder stands, and one might often also encounter combined infestations by both A. alni and M. hiratsukanum (d). In particular, the eastern Asian rust M. hiratsukanum that was first observed to lead to mass alder infestation in the mid-1990s in northern Europe (Põldmaa 1997; Hantula et al. 2009) has spread over much of the Europe by now, and infestations involving all alder trees in a given stand are common. Typically, the signs of infestations, orange urediniospores on the lower leaf surface, are observed in late summer and ultimately result in premature leaf senescence, extensive necrosis, and the leaf drying out with characteristic inward rolling of leaf margins (c)

Full size image

The available evidence collectively suggests that global change enhances emissions induced by both abiotic and biotic stresses and might reduce constitutive emissions. Given that induced emissions occur in all plant species, vegetation trace gas emissions are expected to increase. However, due to stochastic nature of stress events, the emission kinetics of induced emissions is inherently much less predictable than the kinetics of constitutive emissions (Arneth and Niinemets 2010). Such stochasticity poses a challenge for model approaches trying to evaluate the influence of climate change on the feedback loops between constitutive and stress-induced volatile emissions, SOA , CCN , plant photosynthetic production, diffuse/direct radiation ratio, and warming and elevated CO2 (Fig. 5). Future experimental work should fill the gaps in quantitative understanding of how stress-elicited volatile emissions scale with the severity of different abiotic and biotic stresses, how stress and altered atmospheric CO2 modify constitutive emissions, and what is the biological variability in these responses. Armed with this knowledge, the community can start targeting the key research questions on the extent to which plant trace gas release can reduce the effects of climate change on vegetation and the extent to which it can reduce the rate of climate change.

4 Concluding Perspectives

Several widespread plant species are strong constitutive emitters of volatile isoprenoids such as isoprene and monoterpenes, and all plants can be induced to release volatiles upon abiotic and biotic stresses. These emissions play a variety of biological and biogeochemical roles, overall improving directly or/and indirectly the stress resistance of vegetation and altering the ambient environment at local, regional, and global scales. From a local perspective, release of volatiles can be considered both ecosystem service or ecosystem disservice depending on the human impact on atmospheric composition. In atmospheres polluted with nitrogen mono-oxides, plant volatiles contribute to ozone formation in the atmosphere, and thus, volatile release adversely affects the environment. In clean atmospheres, however, plant volatiles reduce atmospheric ozone levels and thus contribute to atmospheric cleansing. Furthermore, by enhancing vegetation stress resistance, the volatiles contribute to the maintenance of ecosystem integrity and, thus, contribute to all the well-perceived ecosystem services such as the capacity of forest ecosystems to provide wood and agroecosystems to provide agricultural products. In addition to controlling atmospheric ozone levels, plant volatiles provide a number of other key regulating services of local to global importance. In particular, plant volatiles alter atmospheric clearness due to the effects of volatiles on atmospheric particle concentrations and cloudiness. Modifications in atmospheric clearness in turn alter the ratio of diffuse to total solar radiation and atmospheric temperature with ultimate impacts on global vegetation productivity, rate of change in atmospheric CO2 concentration, and rate of global change. In future hotter more stressful environments, stress-induced volatile emissions can be particularly relevant in driving the global feedbacks between volatile production, modifications in atmospheric oxidative status, clearness, and global change. Given that the role of volatiles in global coupled vegetation-climate models is still largely unaccounted, I argue that the plant/atmosphere interface should be a high priority research target in future climate change studies.