Biogenic volatile organic compounds in the Earth system



This article is corrected by:

  1. Errata: Corrigendum Volume 184, Issue 1, 276, Article first published online: 2 September 2009

Author for correspondence:
C. N. Hewitt
Tel:+44 (0) 1524 593 931



  • Summary 27

  • I. Introduction 27
  • II. Regulation of BVOC emission 30
  • III. Roles of BVOCs in the Earth system 32
  • IV. BVOCs in a changing global environment 36
  • V. Synthesis 44
  • Acknowledgements 44

  • References 44


Biogenic volatile organic compounds produced by plants are involved in plant growth, development, reproduction and defence. They also function as communication media within plant communities, between plants and between plants and insects. Because of the high chemical reactivity of many of these compounds, coupled with their large mass emission rates from vegetation into the atmosphere, they have significant effects on the chemical composition and physical characteristics of the atmosphere. Hence, biogenic volatile organic compounds mediate the relationship between the biosphere and the atmosphere. Alteration of this relationship by anthropogenically driven changes to the environment, including global climate change, may perturb these interactions and may lead to adverse and hard-to-predict consequences for the Earth system.


1-aminocyclopropane-1-carboxylic acid


biogenic volatile organic compound


free-air CO2 enrichment


Intergovernmental Panel on Climate Change


jasmonic acid


methyl jasmonate


peroxymethacrylic nitric anhydride


methyl salicylate




salicylic acid


secondary organic aerosol

I. Introduction

The Earth is a single and partially self-regulating system that consists of interlinked physical, chemical and biological components. The terrestrial biosphere is one subsystem of this and, by acting as a source of biogenic volatile organic compounds (BVOCs) to the atmosphere, provides a strong link between the Earth's surface, atmosphere and climate. Most of these BVOCs are synthesized by one of three major biochemical routes: the isoprenoid, the lipoxygenase or the shikimic acid pathways (Feussner & Wasternack, 2002; Dudareva et al., 2006; Matsui, 2006; Xiang et al., 2007; Qualley & Dudareva, 2008). A number of low-molecular-weight (C < 5) BVOCs are also emitted by plants, for example methanol, ethylene, formaldehyde, ethanol, acetone and acetaldehyde (Kreuzwieser et al., 1999; Fall, 2003; Argueso et al., 2007). These pathways have been relatively well studied and the routes of formation are now well understood (Fig. 1). However, the biochemical regulation and function of most of these compounds are not clearly known

Figure 1.

Simplified description of the metabolic pathways of biogenic volatile organic compound (BVOC) biosynthesis in leaves and/or flowers and roots. Volatile compounds are shown inside oval-shaped areas and the enzymes responsible for BVOC synthesis are shown in boxes (adapted from Dudareva et al., 2006). ACC, 1-aminocyclopropane-1-carboxylic acid; ADH, alcohol dehydrogenase; AdoMet, S-adenosyl-l-methionine; ALDH, aldehyde dehydrogenase; AOC, allene oxide cyclase; AOS, allene oxide synthase; B2H, benzoic acid-2-hydroxylase; BSMT, S-adenosyl-l-methionine:benzoic acid/salicylic acid carboxyl methyltransferase; CoA, coenzyme A; DAHP, 3-deoxy-d-arabino-heptulosonate; DMAPP, dimethylallyl diphosphate; FPP, farnesyl diphosphate; FPPS, FPP synthase; F6P, fructose-6-phosphate; GA-3P, glyceraldehyde-3-phosphate; GGPP, geranylgeranyl diphosphate; GGPPS, GGPP synthase; GLU, β-glucosidase; GPP, geranyl diphosphate; GPPS, GPP synthase; HCN, hydrogen cyanide; HG, homogalacturonic acid; HNL, hydroxynitrile lyase; HPL, fatty acid hydroperoxide lyase; IGL, indole-3GP lyase; indole-3GP, indole 3-glycerol phosphate; IPP, isopentenyl diphosphate; IspS, isoprene synthase; JA, jasmonic acid; JMT, jasmonic acid carboxyl methyltransferase; LOX, lipoxygenase; MeBA, methyl benzoate; MeJA, methyl jasmonate; MEP, 2-C-methyl-d-erythritol 4-phosphate; MeSA, methyl salicylate; Met, methionine; MOX, methanol oxidase; MVA, mevalonate; OPDA, 12-oxo-phytodienoic acid; PAL, phenylalanine ammonia lyase; PDC, pyruvate decarboxylase; PEP, phosphoenolpyruvate; Phe, phenylalanine; PMEs, wall-localized pectin methylesterases; SA, salicylic acid; SAMT, S-adenosyl-l-methionine:salicylic acid carboxyl methyltransferase; TCA, tricarboxylic acid or citric acid; TPSs, terpene synthases.

BVOCs are released from above- and below-ground plant organs. In general, flowers and fruits release the widest variety of BVOCs, with emission rates peaking on maturation (Dixon & Hewett, 2000; Knudsen et al., 2006; Knudsen & Gershenzon, 2006; Soares et al., 2007), but leaves have the greatest mass emission rates. The vegetative parts of woody plants are more likely to release diverse mixtures of terpenoids, including isoprene, monoterpenes, sesquiterpenes and some diterpenes (Owen et al., 2001; Keeling & Bohlmann, 2006), whereas grass species emit relatively large amounts of oxygenated BVOCs and some monoterpenes (Kirstine et al., 1998; Fukui & Doskey, 2000). When plants are damaged, the emissions of these compounds may be increased and other, so-called, green leaf volatiles (C6 aldehydes and ketones) may also be produced (Fall et al., 1999; Laothawornkitkul et al., 2008a). Biotic and abiotic stresses may also induce the production of some BVOCs, such as terpenes, methyl jasmonate (MeJA) and methyl salicylate (MeSA), from leaves, the magnitude and quality of which depend on the type of damage (Takabayashi et al., 1994; Seo et al., 2001; Mithofer et al., 2005; Laothawornkitkul et al., 2008a). The major classes of BVOCs, the major groups of BVOC-emitting plants and estimates of current and future fluxes into the atmosphere are shown in Table 1.

Table 1.  The major classes of biogenic volatile organic compounds (BVOCs), the major groups of BVOC-emitting plants and estimates of current and future BVOC fluxes into the atmosphere
BOVC speciesPresent estimated annual global emission (1012 g C)Future estimated annual global emission (1012 g C)Atmospheric lifetime (d)ExampleMajor emitting plants
  1. Source: (Guenther et al., 1995; Kirstine et al., 1998; Fall, 1999; Fukui & Doskey, 2000; Lathiere et al., 2005; Arneth et al., 2008a; Davison et al., 2008), and

Isoprene412–601638–6890.2 Populus, Salix, Platanus, Cocos, Elaeis, Casuarina, Picea and Eucalyptus
Monoterpene33–480265–3160.1–0.2β-Pinene, α-pinene, limoneneLycopersicon, Quercus, Cistus, Malus, Pinus and Trichostema
Other reactive BVOCs~260~56–159 (only for acetaldehyde and formaldehyde)< 1Acetaldehyde, 2-methyl-3-buten-2-ol and hexenal familyGrassland (mix of C3 plants), Vitis, Brassica, Secale and Betula
Other less reactive BVOCs~260~292–514 (only for methanol, acetone, formic acid and acetic acid)> 1Methanol, ethanol, formic acid, acetic acid and acetoneGrassland (mix of C3 plants), Vitis, Brassica, Secale and Betula
Ethylene8–25 1.9  

The single most important BVOC in the Earth system is probably isoprene (C5H8, 2-methyl-1,3-butadiene). Its production and emission by plants were first described by Sanadze (1956), and its effect on the physics and chemistry of the atmosphere was first described by Went (1960) (Table 2). Notwithstanding the dominance of isoprene, the biosphere produces and emits hundreds, if not thousands, of reactive BVOCs into the atmosphere. Of these, probably a few tens to a hundred specific species have significant and discernible effects in the atmosphere. Since the 1960s, over 1000 peer-reviewed papers have been published on the biosynthesis, role and function of BVOCs in the biosphere and atmosphere. It is now clear that these compounds have important effects within plants, between plants, between plants and other organisms and in the atmosphere at the local, regional and global scales.

Table 2.  Keystone publications on the role of biogenic volatile organic compounds (BVOCs) in the atmosphere (1) and biosphere (2)
Isoprene is emitted from plantsSanadze (1956)1
BVOC emissions from forests can lead to aerosol formation and have environmental effectsWent (1960)1
An airborne cue from herbivore-damaged plants induces chemical defence in neighbouring undamaged plantsBaldwin & Schultz (1983)2
BVOCs emitted from damaged plantsDicke (1986)2
BVOCs contribute to photochemical smog and should be considered when developing air pollutant control strategiesChameides et al. (1988)1
Airborne methyl jasmonate induces plant defence and interplant communication occurs between plants from different speciesFarmer & Ryan (1990)2
First review of role of BVOCs in the atmosphereFehsenfeld et al. (1992)1
Global emission of BVOCs from terrestrial plants is > 1000 Tg yr−1Guenther et al. (1995)1
Isoprene emission protects photosynthesis from transient heat stressSharkey & Singsaas (1995)1,2
Plants can actively produce BVOCs in response to herbivoryParé & Tumlinson (1995)2
Induced BVOCs repel herbivores and are produced at nightDe Moraes et al. (1997)2
BVOCs play a role in indirect defence against herbivory in natureKessler & Baldwin (2001)2
BVOCs can protect plants from oxidative stressLoreto et al. (2001b)1,2
Isoprene oxidation leads to secondary organic aerosol formationClaeys et al. (2004)1
BVOCs prime neighbouring plants against herbivore attacksEngelberth et al. (2004)2
Isoprene synthesis can be under circadian controlWilkinson et al. (2006)1
Isoprene influences plant–herbivore interactions and tritrophic interactionsLaothawornkitkul et al. (2008c), Loivamaki et al. (2008)1,2

How and why plants synthesize BVOCs, and what are their effects or functions, are of interest to at least three distinct scientific communities. Atmospheric chemists are interested in BVOC emissions in terms of their effects on atmospheric composition and on the atmosphere's chemistry–climate system. Plant biologists are interested in the functions of BVOCs in the biosphere, i.e. their roles in plant biology and ecology. Entomologists are interested in their role as signalling agents.

Several lines of current evidence have demonstrated the tight interconnections that exist between the roles of BVOCs in the biosphere and the atmosphere, but there has been little communication between these research areas to date. This review therefore aims to summarize and identify gaps in our current knowledge of BVOCs in the Earth system, with particular emphasis on their functions. It also highlights the strong interlinkages between the roles of BVOCs in the biosphere and the atmosphere, and hence demonstrates how an integration of knowledge and resources between the biological and atmospheric chemistry research fields is necessary to advance our understanding of the Earth system.

As noted above, an enormously wide range of BVOCs are synthesized and emitted into the atmosphere by plants. Compounds which may be described as BVOCs, but which are specifically excluded from this review, include dimethyl sulphide and methane. Dimethyl sulphide is known to be very important in the Earth's climate system (Charlson et al., 1987), but is produced by oceanic, not terrestrial, plants. Methane is similarly important in the climate system, but reports of its direct biosynthesis by terrestrial plants (Keppler et al., 2006) remain controversial. We therefore focus on nonmethane volatile organic compounds produced by terrestrial plants.

II. Regulation of BVOC emission

Little is known about the regulation of BVOC synthesis rates, with probably more than 90% of the genes involved in their biosynthesis still unidentified. There is evidence to suggest that BVOC biosynthesis is largely controlled at the level of gene expression: microarray analyses show that BVOC biosynthesis genes are upregulated following herbivory via jasmonic acid (JA), salicylic acid (SA) and ethylene signalling pathways (Hermsmeier et al., 2001; Kant et al., 2004; Ralph et al., 2006). The changes in expression of the genes involved in BVOC synthesis positively correlate with their emission rates, and this control leads to the spatial (local and systemic) and temporal pattern of their emissions (Dudareva et al., 2003; Arimura et al., 2004; Underwood et al., 2005). However, emissions of many BVOCs are also strongly correlated with enzyme activities under both optimum and stress conditions (Kuzma & Fall, 1993; Loreto et al., 2001a; Fischbach et al., 2002). This indicates that transcriptional regulation may not be the only controlling factor, and hence post-transcriptional, post-translational and enzyme regulatory mechanisms leading to changes in protein levels or enzyme activities remain to be explored as a further means of control.

The availability of substrate for the final reaction leading to BVOC synthesis is also a crucial rate-limiting factor. Some enzymes with broad substrate specificities can generate different types of product, depending on the level of supplied substrates (Negre et al., 2003; Boatright et al., 2004; Pott et al., 2004). Genetic manipulation resulting in the redirection of cytosolic or plastidic isoprenoid precursors elevates BVOC production in transgenic tobacco plants (Wu et al., 2006). These studies highlight the importance of precursor fluxes through the entire biosynthetic pathway in the regulation of BVOC production and emission.

The emissions of BVOCs from flowers, and from undamaged and herbivore-damaged leaves, often show distinct diurnal or nocturnal patterns (Dudareva et al., 2005; Wilkinson et al., 2006; Loivamaki et al., 2007). This may be the result of circadian regulation of substrate availability, transcription or enzyme activity (Yakir et al., 2007). As yet, there is little information on the molecular mechanisms of circadian control of BVOC emissions. As different BVOCs may result from different biosynthetic pathways, it is not yet clear how the controls of these pathways are co-ordinated to give rise to a specific mixture of BVOCs.

The emission rates of all BVOCs also depend, at least in part, on leaf temperature, which may influence the availability of substrate and the activity of rate-limiting enzymes. However, emission rates from leaves are not only limited by physiological factors, but also by physicochemical constraints caused by temperature, stomatal conductance and leaf structure (Niinemets et al., 2004). These limit volatility (determined by gas phase partial pressure, and aqueous and lipid phase concentrations), diffusion through the gas, aqueous and lipid phases within the leaves and diffusion from the leaf surface. Gas phase diffusion at the leaf–air interface, determined by stomatal conductance, can influence significantly the synthesis and emission of BVOCs with low Henry's law constants, such as formic acid, formaldehyde and methanol. This does not apply to the less water-soluble compounds, such as isoprene and the nonoxygenated terpenes (Niinemets et al., 2004), the emission rates of which are independent of stomatal conductance. Soil moisture availability, carbon dioxide (CO2) concentration and other environmental stresses, including ozone (O3) concentration, may therefore affect the production and emission of some BVOCs through their effects on stomatal conductance.

The photon flux density determines the emission rates of some BVOCs. This largely depends on the presence of storage compartments in leaves. Some plants, such as Pinus, Abies, Eucalyptus and those in the family Rutaceae, store BVOCs in specialized storage compartments (for example, resin ducts, cavities, oil glands or glandular trichomes), whereas others, such as some oaks (Quercus spp.), do not (Loreto et al., 1998a). In the absence of such storage compartments, only small and temporary pools of BVOCs can be nonspecifically stored in plant tissue in the lipid phase (nonoxygenated lipophilic BVOCs) or in the aqueous phase (oxygenated lipophobic BVOCs). The absence of these compartments results in emission rates being closely coupled to incident light intensity (Staudt & Bertin, 1998). In plants with BVOC storage compartments, the emissions are mostly light independent and are closely coupled to leaf temperature, because BVOC volatilization comes from large stored pools (Tingey et al., 1980). Some compounds, for example, isoprene, are not stored at all and are highly volatile: their emission rate depends on temperature and light. The relationships between light and temperature control of biosynthesis rates, intraplant storage capacity and light and temperature control of emission rates are shown in Fig. 2. These relationships are the basis of recently developed models of BVOC emission rates (for example, Grote & Niinemets, 2008).

Figure 2.

Schematic representation of the relationships between light and temperature controls of biogenic volatile organic compound (BVOC) synthesis rates (I), intraplant storage capacity and light and temperature controls of BVOC emission rates (E) (modified with permission from Grote & Niinemets, 2008). BVOC synthesis rates response to light and temperature, based on enzyme kinetic expressions, whereas BVOC evaporation from storage pools depends on diffusion resistances and compound physicochemical characteristics. Therefore, BVOC emissions from large stores (e.g. most monoterpenes) are not dependent directly on the synthesis rate (I ≠ E), but on diffusion resistances and compound physicochemical characteristics. Hence, the photon flux density does not influence the emission rate. However, BVOC emissions from small storage pools (e.g. isoprene) depend directly on the synthesis rate (I = E). The cross-sections are for a representative needle of monoterpene-emitting species Pinus radiata (the magnifications shows the resin duct) and for a leaf of isoprene-emitting species Populus tremula (the magnification outlines the chloroplasts where isoprene is synthesized). The scale bars are 0.1 mm in allcases, except for chloroplasts (µm). The images were non-stained (P. tremula) or stained with toluidine blue (P. radiata).

III. Roles of BVOCs in the Earth system

BVOCs play numerous roles in the Earth system and provide interlinkages between its biological, chemical and physical compartments, as shown schematically in Fig. 3.

Figure 3.

Schematic diagram summarizing the current understanding of the roles of biogenic volatile organic compounds (BVOCs) in the Earth system. BVOCs exert their roles in the biological, chemical and physical components of the Earth system, and hence provide a connection between the biosphere and the atmosphere. The use of three compartments labelled biology, chemistry and physics is not intended to imply that chemical and physical processes do not occur within the biosphere. SOA, secondary organic aerosol; inline image, night-time.

1. Roles of BVOCs in the biosphere

BVOCs as signalling compounds within plants  The roles of MeJA, ethylene and MeSA in plants are very diverse and have been reviewed extensively (Raskin, 1992; Creelman & Mullet, 1997; Bleecker & Kende, 2000). Here, we focus on their roles in the regulation of BVOC production.

MeJA and related compounds  MeJA and JA are ubiquitously distributed throughout the plant kingdom and are collectively called jasmonates (Creelman & Mullet, 1997). They are involved in inducing the production of fruit ripening-related BVOCs, including ethylene (Kondo et al., 2007; Ziosi et al., 2008). Jasmonate treatment induces the expression of the 1-aminocyclopropane-1-carboxylic acid (ACC) synthase gene (Kondo et al., 2007), whereas the internal ethylene concentration influences the production of MeJA-mediated volatiles (Kondo et al., 2005). This suggests that jasmonate and ethylene signalling pathways may interact and modulate BVOC production in a range of fruits.

JA-dependent signalling also mediates the synthesis of BVOCs from vegetative plant parts (van Poecke & Dicke, 2002; Ament et al., 2004; Girling et al., 2008), some of which can attract parasitoids/predators of herbivores (Thaler, 1999; Thaler et al., 2002a; van Poecke & Dicke, 2002). When damaged by herbivory, some plants also release the volatile cis-jasmone, a compound related to JA and MeJA (Loughrin et al., 1995; Lou & Baldwin, 2003; Röse & Tumlinson, 2004). Cis-jasmone may be another plant regulator, as its exogenous application increases plant resistance to aphids (Bruce et al., 2003a) and elevates plant BVOC production and attraction to the parasitoid Aphidius ervi (Birkett et al., 2000; Bruce et al., 2003b; Bruce et al., 2008). It induces the expression of defence genes which are independent from those induced by MeJA, suggesting that these two compounds may produce distinct signalling cascades (Bruce et al., 2008).

Although various studies have shown that JA-dependent signalling plays a central role in the induction of BVOC emission (Ament et al., 2004; Girling et al., 2008; Herde et al., 2008), it is unclear which cell types are responsible in mediating this pathway, and in what form and how far the JA-dependent signals can travel in plants. However, more recent experiments have suggested that amino acid conjugates of JA, especially jasmonoyl-isoleucine, are essential in JA-dependent signalling (Staswick, 2008).

Ethylene  Ethylene can diffuse freely from cell to cell across membranes and is a potent regulator in plants. Both exogenous and post-pollination-derived ethylene downregulate floral volatile production by mediating the expression and activity of enzymes involved in BVOC synthesis (Negre et al., 2003; Underwood et al., 2005). This may help plants to modulate their resource allocation, because, once flowers are pollinated, floral scents have accomplished their role. In contrast with its role in flowers, ethylene upregulates volatile production in ripening fruits and positively regulates the expression of various enzymes involved in aroma formation (Yahyaoui et al., 2002; Manríquez et al., 2006). Transgenic fruit with impaired ethylene production produces much less ripening-related volatiles (Bauchot et al., 1998), indicating that such processes are regulated by developmental factors that must be coordinated with ethylene synthesis and perception.

Vegetative plant parts may also release ethylene as part of a herbivore wounding response (Arimura et al., 2002). In general, ethylene enhances BVOC production and emission, but this is dependent on the type of BVOC (Horiuchi et al., 2001; Schmelz et al., 2003a,b; Arimura et al., 2008). Several lines of evidence have indicated that ethylene and JA synergistically regulate BVOC synthesis (Horiuchi et al., 2001; Schmelz et al., 2003a,b; Arimura et al., 2008). However, the interplay between JA- and ethylene-dependent signals is not yet clear. Staswick & Tiryaki (2004) have suggested that an unknown enzyme might be responsible for conjugation between JA and ACC, leading to an inactive JA–ACC conjugate, with subsequent hydrolysis of such a conjugate yielding JA and ACC available for the corresponding signalling routes. Ethylene may also regulate the JA pathway by influencing the expression of allene oxide synthase involved in JA biosynthesis (O'Donnell et al., 1996; Laudert & Weiler, 1998; Sivasankar et al., 2000).

MeSA  MeSA is the volatile counterpart of SA. The SA signalling cascade is involved in the induction of both local and systemic defences (systemic acquired resistance) to a broad range of pathogens and some insects (Bostock, 1999; Dempsey et al., 1999; Vasyukova & Ozeretskovskaya, 2007). The most recent grafting study using tobacco plants with different genetic backgrounds has provided unambiguous evidence that MeSA is the mobile signal that is required for systemic resistance induction in tobacco (Nicotiana tabacum) (Park et al., 2007).

SA- and JA-dependent signalling are required for defence activation against herbivores and pathogens, and are generally known to function antagonistically (Thaler et al., 2002b,c). Although JA plays a central role in the production of induced BVOCs and mediates MeSA production (Ament et al., 2004), the presence of SA or SA-derived signals is also required for the production of herbivore-induced volatiles that mediate an indirect defence response (see below) (van Poecke & Dicke, 2002; Girling et al., 2008). The balance between the JA, ethylene and SA signalling cascades seems to help plants to discriminate the quality and quantity of tissue damage and thus control specific blends of herbivore-induced volatiles (Ozawa et al., 2000; Engelberth et al., 2001; Girling et al., 2008).

Roles of BVOCs in plant reproduction  To ensure reproductive success, flowering plants release a myriad of BVOCs from their flowers in order to attract pollinators (Wright et al., 2005) and to assist them to identify conspecific flowers whilst foraging (Andersson et al., 2002). The different BVOC mixtures and their relative abundances make the scent bouquet released by a particular flower characteristic of that bloom (Knudsen & Tollsten, 1993; Knudsen et al., 2006). This specificity may therefore be used by pollinators to distinguish a particular flower within a single species and across plant species and lead them to specific food sources (Andersson et al., 2002; Schiestl & Ayasse, 2002; Wright et al., 2005).

There is strong evidence indicating that flowers compete for pollinator visitors (Basra, 2006). Exogenous application of isoprene promotes early flowering of barley, oilseed rape and Arabidopsis (Terry et al., 1995). These observations have led to the hypothesis that isoprene emission may disrupt pollination in competing plants and so confer competitive advantage to isoprene emitters. Further studies are required to test this in experimental and natural systems.

Following pollination, fruits also produce a range of BVOCs that change according to their developmental and ripening stages (Goff & Klee, 2006). Fruit odour can attract seed dispersers and allows them to locate and discriminate between ripe and unripe fruits even within the same plant species (Luft et al., 2003; Hodgkison et al., 2007). BVOCs therefore play a role at all stages of plant reproduction and development.

Roles of BVOCs in plant defence against biotic stresses  Some BVOCs released from flowers, leaves and roots may protect plant organs from pathogens by their antimicrobial or antifungal activity (Croft et al., 1993; Shiojiri et al., 2006). They can also directly affect the physiology and behaviour of herbivores through their toxic, repellent and deterrent properties (De Moraes et al., 2001; Vancanneyt et al., 2001; Aharoni et al., 2003; Laothawornkitkul et al., 2008c). Some, such as 4,8,12-trimethyl-1,3(E),7(E),11-tridecatetraene and 4,8-dimethyl-1,3(E),7-nonatriene, serve as information conveyors that can provide communication between and within trophic levels. Foliage may emit blends of herbivore-induced BVOCs that attract insect or acarid predators and parasitoids, as first demonstrated by Dicke (1986). Since then, it has been shown that BVOCs serve several functions in plant ecology (Table 2). Recently, it has been demonstrated that isoprene influences plant–herbivore interactions by deterring herbivores from feeding (Laothawornkitkul et al., 2008c) and by interfering in tritrophic interactions (Loivamaki et al., 2008).

Tritrophic communication is not restricted only to above-ground plant parts, but may also occur below ground. For example, insect attack on maize roots triggers the release of a sesquiterpene, (E)-β-caryophyllene, which attracts nematodes that prey on insect larvae (Rasmann et al., 2005). However, little is known at present about the role of BVOCs in the rhizosphere and in soil ecology. This is, at least in part, a result of the difficulty of conducting experiments and field observations on soil without disturbing soil structure and root systems (Hayward et al., 2001; Owen et al., 2007).

Some BVOCs, for example MeJA (Farmer & Ryan, 1990), MeSA (Shulaev et al., 1997), some green leaf volatiles (Engelberth et al., 2004; Farag et al., 2005) and some terpenes (Arimura et al., 2002), can serve as airborne signals between plants (Engelberth et al., 2004; Kessler et al., 2006; Ton et al., 2007) and between organs within the same plant (Karban et al., 2006; Frost et al., 2007; Heil & Silva Bueno, 2007). This communication can occur between neighbours of the same or different species (Dolch & Tscharntke, 2000; Kessler et al., 2006). On perception by receiver plants, these BVOC signals can directly activate herbivore defence mechanisms or may prime a subset of defence-related genes for earlier and/or stronger induction on subsequent defence elicitation (Arimura et al., 2000; Engelberth et al., 2004; Kessler et al., 2006; Frost et al., 2007; Ton et al., 2007).

Molecular, chemical and behavioural assays show that VOC-induced priming, which targets a specific subset of JA-inducible genes, leads to improvements in both direct and indirect defences (Ton et al., 2007). However, the reliability of this mechanism varies. For example, the BVOCs released by Manduca sexta-infested wild tobacco plants (Nicotiana attenuata) fail to prime neighbouring N. attenuata for defence (Paschold et al., 2006), but BVOCs emitted by mechanically damaged sagebrush (Artemesia tridentata tridentata) can prime N. attenuata against subsequent attack by M. sexta (Kessler et al., 2006). By contrast, communication among silver sagebrush (Artemesia cana) individuals does not lead to increased resistance to herbivory in receiver plants (Shiojiri & Karban, 2008). What causes this variability requires further explanation; there would seem to be no benefit for damaged plants to warn their neighbours when they are competing for limited resources in a local environment. One possible explanation is that plants might have evolved such communication for their own use, namely for communication within an individual plant, as BVOC concentrations in air decrease rapidly with distance from source (Karban et al., 2006).

Plant resistance mechanisms can be induced or primed by BVOCs released from mechanically damaged neighbouring plants (Kessler et al., 2006; Shiojiri & Karban, 2006) or by such damage within the same plant (Karban et al., 2006). This raises several questions: (i) can plants distinguish mechanical damage caused by biotic factors (e.g. pathogens or herbivores) vs abiotic factors (e.g. hail and strong wind) and, if so, how?; and (ii) how do plants discriminate a ‘stress’ signal from background BVOCs in heterogeneous and changing environments? A mechanistic understanding of the nature of BVOC receptors and the cells responsible for mediating the signal transduction pathways requires further investigation, as do the ecological consequences of BVOC-induced resistance and priming. Such knowledge could have potential in the future development of sustainable agricultural practices.

Roles of BVOCs in plant defence against abiotic stresses Isoprene emission might serve as a metabolic safety valve to dissipate excess energy (Sanadze, 2004) and metabolites (Rosenstiel et al., 2004). However, Sharkey et al. (2007) argued that this does not explain the random distribution of the isoprene emission trait across the plant kingdom or differences in isoprene emission capacity at the canopy level. In addition, there are probably other energy-consuming mechanisms in plants that are more effective than isoprene synthesis.

Isoprene and monoterpenes can protect the photosynthetic apparatus of plants from damage caused by transient high-temperature episodes, and may prevent a progressive reduction in photosynthetic capacity (Singsaas et al., 1997; Loreto et al., 1998b; Behnke et al., 2007) (Fig. 3). Several mechanistic explanations of this phenomenon have been proposed (Sharkey & Yeh, 2001). When thylakoid membranes become leaky at high temperature, isoprene may enhance hydrophobic interactions and so strengthen the thylakoid membrane. It might also help more generally to enhance the integrity of membranes and protein complexes. Recent mechanistic evidence supports this hypothesis by showing that isoprene can directly protect a model phospholipid membrane from heat spikes (Siwko et al., 2007).

Despite early work, which suggested that isoprene–O3 interactions may damage plant tissue (Hewitt et al., 1990), it is now known that isoprenoids function as antioxidants in leaves and confer protection against O3-induced oxidative stress and singlet oxygen accumulation during photosynthesis (Loreto et al., 2001b, 2004; Affek & Yakir, 2002; Vickers et al., 2009). Isoprenoids may perhaps exert their protective action at the membrane level by quenching hydrogen peroxide formed in leaves and by reducing lipid peroxidation of cellular membranes caused by oxidants (Loreto & Velikova, 2001), and may interfere with the molecular signalling that leads to programmed cell death (Velikova et al., 2005). This process might counteract the hypersensitive response (for example, rapid cell death in response to pathogen infection) that requires initiation by reactive oxygen species. This suggests possible antagonistic interactions between the hypersensitive response and the antioxidant capacity of BVOCs. Clearly, how plants are able to balance their defence strategies in response to both abiotic and biotic stresses is complicated and the role played by BVOCs remains to be determined.

2. Roles of BVOCs in the atmosphere

Estimates of the global flux of BVOCs from the biosphere to the atmosphere are rather uncertain, but may be 700–1000 × 1012 g (C) per year (Table 1). There are large uncertainties associated with these estimates, although the remotely sensed concentrations of BVOC oxidation products in the atmosphere, inverted and modelled using an atmospheric chemistry transport model, are now beginning to constrain these estimates (for example, the use of formaldehyde observations to constrain isoprene emission estimates; Guenther et al., 2006). In any event, the BVOC flux far exceeds the global anthropogenic VOC flux. Although very many BVOC species have been identified from plants, as mentioned above, much of the global flux and subsequent effect on atmospheric chemistry is probably caused by a relativity small number of compounds. Isoprene makes the largest contribution, followed by the monoterpene family (Levis et al., 2003). Some oxygenated compounds, such as methanol, acetone and acetaldehyde, may also be important in the atmosphere (Guenther et al., 1995; Kesselmeier & Staudt, 1999; Fuentes et al., 2000). Estimating the emission rates of C15 sesquiterpenes and related compounds is difficult as they present particular analytical challenges because of their reactivity and low vapour pressures; they are important precursors to secondary organic aerosols (SOAs) (Hoffmann et al., 1997; Bonn & Moortgat, 2003).

Oxidation of BVOCs in the atmosphere  When reactive BVOCs are released into the atmosphere, they are subject to oxidation reactions, potentially leading to the ultimate products of CO2 and water (Fig. 3). Many of their intermediate, partially oxidized, products are water soluble and hence may be removed from the atmosphere by wet deposition (Fehsenfeld et al., 1992), or may have lower vapour pressures than the primary compounds and hence enter the particle (solid or aerosol) phase and be removed from the atmosphere by wet and dry deposition, thereby removing reactive carbon from the atmosphere. The relative importance of this process is not currently possible to quantify, but requires a better understanding of the yield of SOAs from BVOCs.

Hydroxyl radicals (OH) dominate the daytime chemistry of the troposphere and the oxidation of VOCs is primarily initiated by reaction with them. OH is itself produced in part by the photolysis of tropospheric O3 and the subsequent reaction of electronically excited atomic oxygen, O(1D), with water vapour. The initial products of the VOC–OH reaction can be further oxidized to form peroxy radicals (RO2). In the presence of sufficient oxides of nitrogen (NOx = NO plus NO2), for example in polluted air, these RO2 species may oxidize NO to NO2, which can, in turn, be photodissociated, leading to the formation of O3 and the regeneration of OH (Fig. 3). In clean air with low NOx concentrations, RO2 may recombine or react with HO2 to form less reactive peroxides, which may be removed from the atmosphere by deposition processes (Fehsenfeld et al., 1992), which lead to the net consumption of O3. Recent field observations of OH and BVOC concentrations, supported by laboratory experiments, have suggested that our understanding of BVOC oxidation processes may in fact be inadequate, and that, in low-NOx conditions, more regeneration of OH by these reactions may occur than previously thought (Lelieveld et al., 2008). This has significant implications for the understanding of the oxidant budget of air receiving large BVOC inputs, for example in the boundary layer above tropical and boreal forests. However, this important result has yet to be verified, and further field, laboratory and modelling studies are required to test it.

As well as OH, O3 can itself act as an oxidant for unsaturated BVOCs. The addition of O3 to carbon–carbon double bonds leads to the formation of ozonides, which are unstable and undergo rapid decomposition. This can generate organic free radicals that can form OH and RO2, so mediating the O3 budget of the troposphere as outlined above.

At night, when OH concentrations are effectively zero, BVOC oxidation may be driven by reaction with the nitrate radical (NO3) (Wayne, 2000) (Fig. 3). Because of its rapid reaction with NO and its short lifetime (∼5 s) in sunlight as a result of photolysis, NO3 concentrations are low during the day but can increase substantially at night. This may lead to the removal of BVOCs that would otherwise be available for daytime O3 formation. However, the reaction rates of NO3 with most BVOCs are quite low (one-fifth of that with OH in the case of isoprene), and so reaction with OH is normally the dominant route of oxidation.

Although the details of BVOC oxidation reactions are not yet known with complete certainty, it is clear that BVOC oxidation may affect the oxidative capacity of the troposphere and hence influence the rate of oxidation, formation and concentration of other trace gases (see below) (Fehsenfeld et al., 1992; Wayne, 2000; Atkinson & Arey, 2003; Lelieveld et al., 2008).

Gas phase chemistry of BVOCs  As mentioned above, the oxidation of BVOCs by OH can, in the presence of sufficient NOx, lead to the formation of O3 in the troposphere by disruption of the photochemical steady state of O3 (i.e. allow the oxidation of NO to NO2 without removal of an O3 molecule), and so cause elevated O3 concentrations (Fig. 3). NOx emissions may result from fossil fuel combustion, fertilizer application and biomass burning, as well as natural production by lightning. As tropospheric photochemistry is highly nonlinear with respect to the emissions of O3 precursors, modelling is required to determine the effects of BVOC emissions on O3 concentrations in the troposphere (Fowler et al., 2008).

Since the seminal work of Chameides et al. (1988), it has been recognized that BVOC emissions may be important precursors of photochemical smog and regional-scale O3 production. Furthermore, because OH is the principal oxidant of methane, the third most important greenhouse gas in the atmosphere (after water vapour and CO2), emissions of BVOCs may increase the atmospheric lifetime of methane and so indirectly influence the Earth's radiation balance (Wuebbles et al., 1989). The resulting changes in climate may, in turn, directly and indirectly affect BVOC emission rates, potentially establishing a positive feedback in the climate system. The development of next-generation coupled BVOC emission–atmospheric chemistry–climate models is required before the magnitude of this effect can be constrained.

Although carbon monoxide (CO) is emitted directly by living, senescing and dead leaves (Tarr et al., 1995), the oxidation of BVOCs also contributes significant amounts of CO to the atmosphere (Hatakeyama et al., 1991; Fehsenfeld et al., 1992; Bergamaschi et al., 2000; Griffin et al., 2007). CO influences the oxidative capacity of the atmosphere in the same way as isoprene by functioning as a sink for OH (Logan et al., 1981). Hence, the oxidation of CO can act as a source or sink of O3, depending on the availability of NOx. Once generated, CO can be transported over large distances because of its relatively long atmospheric lifetime of several months, and hence BVOCs can, in this way, influence atmospheric chemistry on the global scale (Fehsenfeld et al., 1992; Lerdau et al., 1997; Lerdau & Slobodkin, 2002).

Atmospheric oxidation of BVOCs and their primary oxidation products (e.g. methyl vinyl ketone and methacrolein in the case of isoprene) can, in the presence of NOx, result in the formation of organic nitrates, including peroxyacetylnitrates (PANs) and peroxymethacrylic nitric anhydrides (MPANs) (Fehsenfeld et al., 1992). PANs and MPANs have longer atmospheric lifetimes than NOx (days to months) and hence can be transported over greater distances, allowing them to act as carriers of reactive nitrogen (Fig. 3). Once thermally decomposed in warmer air, they release NOx (Fehsenfeld et al., 1992; Poisson et al., 2000), resulting in an increase in NOx concentrations in areas without local NOx sources. This process may markedly alter atmospheric composition and chemistry and lead to O3 formation in remote areas. PANs, MPANs and other organic nitrates may be lost by wet deposition (Neff et al., 2002), removing reactive nitrogen from the atmosphere.

Influence of BVOCs on aerosol formation  BVOCs not only influence gas phase atmospheric chemistry, but can also lead to the formation of SOAs (Fig. 3). The mechanisms by which BVOC oxidation may lead to SOAs in clean air are still not fully understood (Kulmala, 2003), but it is clear that BVOC oxidation products generally have lower vapour pressures than the primary compounds, and so may more readily condense on pre-existing molecular clusters (Joutsensaari et al., 2005). Laboratory studies and field observations suggest that terpenes and sesquiterpenes emitted by vegetation may be significant sources of SOAs (Leaitch et al., 1999; Joutsensaari et al., 2005), with yields as high as 80% (Hoffmann et al., 1997). Oxidation of isoprene also produces SOAs (Claeys et al., 2004; Meskhidze & Nenes, 2006). However, recent field observations over tropical forests have not always found significant SOA production to the degree expected (Rizzo et al., 2006), indicating that further work is needed in this area.

Aerosols directly affect climate by scattering solar radiation. They also indirectly alter the Earth's radiative balance by acting as cloud condensation nuclei, changing cloud albedo and the degree of cloud cover, so potentially leading to net cooling of the Earth's surface during the day. Although it is known that a substantial fraction of the aerosol particles in remote regions is organic material, and that the oxidation of BVOCs may lead to the formation of SOAs, it is not yet clear how important is SOA formation in altering the climate system. Increased cloud cover may also reduce the occurrence of low night-time surface temperatures, which can damage plants (Hayden, 1998). The possibility that SOA formation from BVOC emissions cools the Earth and so moderates temperature-dependent BVOC emission from plants – and other similar feedbacks in the Earth system – is the focus of much current research. Hence, there is the potential for feedback between BVOC emissions, SOA and climate.

IV. BVOCs in a changing global environment

In the sections above, we have described the impact of BVOCs on the Earth's environment. We now turn to addressing how changes in environmental conditions may affect BVOC production. As the Earth's biosphere and atmosphere change, as a result of both natural processes and human activities, BVOC emissions from the terrestrial biosphere to the atmosphere will change, with the potential to cause feedbacks, so potentially exacerbating the effects of change on the environment. Understanding how BVOC emissions respond to future environmental change will help us to predict the future impacts of BVOCs. The ultimate goal of this research is to build comprehensive predictive models of the Earth system.

1. Atmospheric CO2 concentration and BVOC emissions

The CO2 concentration in the atmosphere has risen by approximately 35% from pre-industrial times to the present and is predicted to double within the 21st century [Intergovernmental Panel on Climate Change (IPCC), 2007]. Elevated CO2 concentrations have been shown to increase (Sharkey et al., 1991; Staudt et al., 2001), decrease (Sharkey et al., 1991; Loreto et al., 2001a; Rosenstiel et al., 2003; Possell et al., 2004; Vuorinen et al., 2004c; Wilkinson et al., 2008) or have no significant effects (Penuelas & Llusia, 1997; Constable et al., 1999; Buckley, 2001; Centritto et al., 2004) on BVOC production and emission at the whole plant, shoot or leaf levels. Various factors, including plant species, age, experimental duration and CO2 concentration, may explain these contrasting results. Limitations in experimental design and implementation may also cause confounding results. Glasshouses (Penuelas & Llusia, 1997; Staudt et al., 2001; Possell et al., 2004), artificially illuminated controlled environment chambers (Vuorinen et al., 2004c; Wilkinson et al., 2008), open-top and closed solar domes (Buckley, 2001; Loreto et al., 2001a) and free-air CO2 enrichment (FACE) facilities (Centritto et al., 2004) have all been used to study the effect of elevated CO2 on BVOC emissions. The size limitation of most experimental facilities (except FACE) means that young, small pot-grown plants are usually used. The resulting limited rooting volume may diminish plant responses to elevated CO2 by both nutrient exhaustion (Korner, 2003) and root compaction (Thomas & Strain, 1991). Solar domes and other chambers may influence vegetation growth by causing differences in aerial microclimate inside the chamber (Murray et al., 1996). Despite these problems, on balance, it seems that increasing CO2 causes a decrease in isoprene emissions on a leaf surface area basis, but that this might be offset by increases in emissions as a result of increasing vegetation productivity and leaf area growth caused by elevated CO2 (Possell et al., 2005; Arneth et al., 2007).

Although growth under elevated CO2 concentrations increases leaf foliar density, BVOC emissions from most plant canopies are limited by light intensity (Sharkey et al., 1996; Guenther et al., 2006) and temperature (Monson et al., 1992; Sharkey et al., 1996). Thus, the increase in shading associated with increased leaf area index might also directly affect canopy-scale emission rates (Possell et al., 2005; Guenther et al., 2006). This should be taken into account when enclosure experiments are extrapolated to the canopy scale.

2. Global warming and BVOC emissions

Climate models suggest that, during the 21st century, the mean global temperature will increase by 1–6°C (with a best estimate of 2–3°C) (IPCC, 2007). This increase in temperature will directly affect plant biochemical activity and the length of the active growing season (Myneni et al., 1997). Emissions of BVOCs are strongly temperature dependent because higher temperatures increase chemical reaction rates, increase cellular diffusion rates and increase the vapour pressures of volatile compounds (Tingey et al., 1991; Lerdau et al., 1994; Fuentes et al., 2000; Sharkey & Yeh, 2001). Various attempts have been made to estimate how an increase in temperature will enhance BVOC emission rates. For example, Penuelas & Llusia (2003) have suggested that increasing mean global temperatures by 2–3°C could enhance global BVOC emissions by 25–45%. At the regional scale, using Great Britain as a case study, it was predicted that an increase in temperature of 1°C would increase isoprene emissions by 14% in the summertime, whereas a 3°C increase would increase emissions by 50% (Stewart et al., 2003). At very high temperatures (above approximately 40°C), isoprene emissions decline dramatically, and it is possible that extreme temperature rises will, eventually, cause a decrease in isoprene emissions, first in the tropics, irrespective of other changes to ecosystems.

Climate warming can also indirectly influence global- and regional-scale BVOC emissions by altering vegetation species composition and vegetation characteristics (Starfield & Chapin, 1996; Wilmking et al., 2004). Warming can also alter latitudinal and altitudinal treelines (Starfield & Chapin, 1996; Lerdau & Slobodkin, 2002; Wilmking et al., 2004). Simulation models predict forest dieback at lower latitudes (Cox et al., 2004), especially in Amazonia, but show the upward and northward expansion of boreal forests under climate warming (Chapin et al., 2000; Kittel et al., 2000), as confirmed by field observations (Luckman & Kavanagh, 2000; Kullman, 2001; Penuelas & Boada, 2003). The expansion of boreal forests may increase BVOC emissions through the spread of high-BVOC-emitting taxa, i.e. Populus sp. and Picea spp. (Lerdau & Slobodkin, 2002), but degradation of lower latitude forests, such as in the Amazonian area, may diminish the increase in BVOC production at the global scale.

3. Land use change and BVOC emissions

The Earth is experiencing massive land use and land cover changes at unprecedented rates, not only as a result of climate change, but also because of urbanization, agriculture and agroindustrialization. These pressures are altering plant species distributions and characteristics, and may dramatically influence BVOC emissions as a result of their biome- and species-specific characteristics. Inventories and spatial analysis suggest a global increase in crop area of 455% in the past 300 yr (1700–1990) and a more than six-fold increase in pasture area (Goldewijk, 2001). Grasses and cereals are not generally major isoprene emitters (Table 1), although they do emit oxygenated BVOCs, particularly during harvesting (König et al., 1995; Kirstine et al., 1998; Davison et al., 2008). Hence, the conversion of forest to crops is predicted to decrease BVOC emissions over large geographical areas. For example, in Amazonia, the isoprene emission flux may decrease by as much as 90% following deforestation (Ganzeveld & Lelieveld, 2004) and, in East Asia, annual isoprene and monoterpene emissions may decrease by 30% and 40%, respectively, because of the expansion of cropland (Steiner et al., 2002). However, forest restoration by the planting of higher isoprene-emitting species (Table 1) will have major effects on BVOC emission rates, especially at the local and regional scales (Lathiere et al., 2006). The large-scale expansion in the cultivation of Elaeis (oil palm, Table 1) that is currently occurring in the tropics for the production of biofuel and other applications may be having a significant impact on BVOC emissions in these regions.

4. Drought stress and BVOC emissions

Precipitation frequency and intensity are predicted to change in the future in response to increasing surface temperature (IPCC, 2007). Drought stress already affects vegetation in many areas (Le Houérou, 1996). Empirical data, summarized in Table 3, indicate that moderate drought can decrease, enhance or have no effect on isoprene and monoterpene emissions, but severe, long-lasting water stress, leading to gross wilting or complete inhibition of photosynthesis, significantly reduces BVOC emissions. However, for sesquiterpenes, the effects of drought are more consistent in the four plant species studied, causing a significant reduction in emissions (Ormeno et al., 2007).

Table 3.  Effect of drought on biogenic volatile organic compound (BVOC) emissions (D, decrease; I, increase; sig., significant; soil moisture reduced by % of field capacity; mono, monoterpenes; iso, isoprene; ses, sesquiterpenes)
SourcesSubdescriptionApproachBVOC measurement scalePlant speciesReplication (n)Effect on BVOC emissions
Bertin & Staudt (1996)Laboratory observation18 d of drought period (severe drought)Branch chamberQuercus ilex L.2D 100% (mono)
Pot-grown plants
Young plants (age not specified)
Soil moisture reduced by ~54%    
Pegoraro et al. (2004)Laboratory observation10–12 d of drought period (severe drought)Leaf enclosureQuercus virginiana Mill.6D 64% (iso)
Pot-grown plants
2-yr-old plants
Soil moisture reduced by ~80%    
Plaza et al. (2005)Field observation
Mediterranean oak forest
30-yr-old plants
Two growing seasons (2000–01)
Natural drought (measured diurnal courses of emission rate)Branch enclosureQuercus ilex spp. rotundifolia1 or 2Inconsistent monoterpene emission over the 2 yr
Pegoraro et al. (2006)Closed biospheres36 d of drought period (mild drought)Ecosystem level gas exchange measurementMixed isoprene-emitting and nonisoprene-emitting species with deep roots No sig. effect (iso)
(Biosphere 2 tropical rain forest)
~15-yr-old plants
Soil moisture reduced by ~50% from field capacity    
Llusia et al. (2006)Field observationSliding plastic curtain (mild drought)Solvent extraction from leavesPinus halepensis L.2–4Contrasting results depending on seasons, plant species, year and type of BVOC
Mediterranean scrubland (2002–04)Soil moisture reduced by 19% from field capacity Globularia alypum L.
Rosmarinus officinalis L.
Erica multiflora L.
Ormeno et al. (2007)Laboratory observation11 d of drought period (severe drought)Branch enclosureRosmarinus officinalis L.6D ~ 20% (total mono + ses)
Pot-grown plantsSoil moisture reduced by ~82% from field capacity   No sig. effect (total mono)
3-yr-old plants    D ~ 70% (total ses)
   Pinus halepensis L. I ~ 290% (total mono + ses)
     I ~ 270% (total mono)
     D ~ 28% (total ses)
   Cistus albidus L. I ~ 107% (total mono + ses)
     I ~ 285% (total mono)
     D ~ 13% (total ses)
   Quercus coccifera L. No sig. except day 7 I ~ 265% (total mono + ses)
     No sig. effect (total mono)
     D ~ 1% (total ses)
Llusia et al. (2008)Field observationSliding plastic curtainBranch enclosurePinus halepensis L.3I ~ 166.5% (selected mono)
Mediterranean scrublandLong-term drought (mild drought) Globularia alypum L. I 75% (selected mono)
Two growing seasons (2003–05) (protect all rain events)Soil moisture reduced by ~16% from field capacity Erica multiflora L. D 19% (iso), I 26.4% (selected mono)
Fortunati et al. (2008)Laboratory observation35 d of drought period (severe drought)Leaf enclosurePopulus nigra L. 9D ~ 71% (iso)
Pot-grown plantsSoil moisture reduced by ~65% from field capacity    
1-yr-old plants     

The varying responses of BVOC emissions to moderate drought may be a result of differences in leaf physiology, BVOC biochemistry and experimental protocol. One important difference in leaf physiology across plant species is the presence or absence of terpene reservoirs (see above). Plants that possess specific monoterpene storage compartments are able to maintain their emissions of monoterpenes, even when they experience a decrease in photosynthesis rate (Llusia & Penuelas, 1998; Pegoraro et al., 2004; Fortunati et al., 2008). Drought can also increase the accumulation of plant secondary metabolites by decreasing carbon allocation to plant growth, as a result of a trade-off between growth and defence (Turtola et al., 2003). It may be that extra-chloroplastic carbon sources temporarily compensate for a reduction in carbon from the choroplastic, photosynthesis-dependent, 2-C-methyl-d-erythritol 4-phosphate pathway (Funk et al., 2004; Fortunati et al., 2008).

As for air pollutant exposure experiments, variations in experimental design across studies may explain the contrasting results seen for water stress. Although field experiments using natural plants are preferable to laboratory experiments using potted plants, the field manipulation of drought is difficult, in part because of the deep rooting of field-grown plants (Pegoraro et al., 2006). Table 3 suggests that drought period and soil moisture content are not necessarily correlated, causing difficulties in the comparison of laboratory and field studies (Pegoraro et al., 2004, 2006).

Neither Quercus coccifera L. nor Quercus ilex have monoterpene storage compartments, yet they exhibit a different response to drought. Quercus coccifera maintained its emission when the soil moisture content was reduced by 82%, but the monoterpene emission of Q. ilex was inhibited when the soil moisture content was reduced by only 54% (Table 3). This may result from the better water-use efficiency of Q. coccifera (Vilagrosa et al., 2003).

These examples highlight the importance of the measurement of leaf water potential and soil moisture to allow better comparison of results across different experimental protocols. Pegoraro et al. (2004) have also suggested that pre-dawn leaf water potential could be used to parameterize drought impact on isoprene emissions.

5. Elevated atmospheric O3 concentration and BVOC emissions

It is highly likely that the concentrations of ground-level O3 will change in the future. The emission rates of the precursors to O3 formation will change over time, and changes to the Earth's climate will cause changes in atmospheric circulation, both of which will directly affect O3 concentrations. Ground-level O3 is already a serious regional-scale air pollutant in many areas of the world, but the prediction of future trends depends critically on assumptions made about precursor emissions. It may be that ground-level O3 pollution will be reduced in some regions, where strict emission controls are implemented, but worsened in other less-developed regions (Fowler et al., 2008).

As both short-term O3 episodes and long-term elevated concentrations have adverse effects on plant growth, species composition and ecosystem functioning (Ashmore, 2005), it is likely that changes in O3‘climatology’ will change BVOC emissions over time. These changes may be the result of the direct effects of O3 on plants, or may be caused by the indirect effects of species composition. Experimental observations have probed the former, and next-generation Earth systems models will, before long, be able to make predictions about the latter.

Experimental evidence on the direct effects of O3 on BVOC emissions is, as for other abiotic stresses, not clear cut, showing that elevated O3 can increase, decrease or have no effect on BVOC emission rates (Table 4). These differences depend on the plant species (Heiden et al., 1999; Peñuelas et al., 1999), the season (Llusià et al., 2002) and the BVOC species (Llusia et al., 2002). Recent work by Ryan et al. (2009) has shown that two genotypes of hybrid poplar, with differing sensitivities to O3, have different VOC responses when exposed to O3. The O3-tolerant genotype was able to maintain its isoprene emission rate when exposed to 120 ppb O3 for 6 h d−1 for 8 d, whereas the O3-sensitive genotype could not; its isoprene emission rate fell on exposure to O3. A different effect has been seen in tobacco, where elevated O3 significantly increases BVOC emissions from the O3-sensitive clone (Heiden et al., 1999), but not from the tolerant clone. However, in both cases, the maintenance of BVOC emissions from the tolerant clone may be because these plants have a higher ability to detoxify reactive oxygen species that occur after O3 uptake through the stomata, possibly because they have a higher carotenoid content, which allows O3 quenching inside O3-tolerant leaves (Ryan et al., 2009; Calfapietra et al., 2008). This could lead to lower cell membrane damage in O3-tolerant plants. This hypothesis is supported by the low C6 emission rates of O3-tolerant plants, compared with those from O3-sensitive plants. It should also be noted that elevated O3 may induce the production of BVOCs that are not present in unexposed plants (Heiden et al., 1999; Vuorinen et al., 2004a).

Table 4.  Effect of ozone on biogenic volatile organic compound (BVOC) emissions (D, decrease; I, increase; sig., significant; OTCs, open top chambers; mono, monoterpenes; iso, isoprene; ses, sesquiterpenes)
SourcesSubdescriptionFumigation methodO3 level (ppbv)Fumigation periodBVOC measurement scalePlant speciesReplication (n)BVOC measure atEffect on BVOC emissions
Peñuelas et al. (1999)Field observationOTCsAmbient + 408 hWhole plantsPinus halepensis L.3Not specifiedNo sig. effect (total BVOCs)
    Leaf enclosureSolanum lycopersicum L. var. Tiny Tim  I ~ 74% (total BVOCs)
Heiden et al. (1999)Laboratory observationsCylindrical glass chamber120–1705 hShoot enclosureNicotiana tabacum L. cv. Bel B (O3-tolerant)2–324 h after fumigationNo sig effect (total BVOCs)
Laboratory observationsCylindrical glass chamber120–1705 hShoot enclosureNicotiana tabacum L. cv. Bel W3 (O3-sensitive)2–324 h after fumigationI ~ 270% (total BVOCs)
Sig. presence of C6 VOCs
Field observationsOTCs508 h d−1 for 2 yrNot specifiedPinus sylvestris L.4 I 40% (mono)
Llusia et al. (2002)Field observationOTCsAmbient + 408 hLeaf enclosureCeratonia siliqua L.3 I ~ 65% (total BVOCs of the four species)
Pot-grown plants    Olea europaea L.   
3-yr-old plants    Quercus ilex spp. ilex L.
Quercus ilex spp. rotundifolia L.
Loreto et al. (2004)Laboratory observationGrowth chamber100–2004 h d−1 for 5 dLeaf enclosureQuercus ilex L.42 d after fumigationI ~ 182% (mono)
Pot-grown plants
3-yr-old plants
Whole-plant fumigation       
Loreto et al. (2004)Laboratory observationGas exchange cuvette2504 hExcised leaf enclosureQuercus ilex L.44 h after fumigationI ~ 60% (mono)
Pot-grown plants
3-yr-old plants
Single-leaf fumigation       
Vuorinen et al. (2004a)Laboratory observationGrowth chamber150–4008 h for 1st dayShoot enclosurePhaseolus lunatus cv. Sieva6Soon after fumigationI ~ 36% (total BVOCs)
Pot-grown plants
5–7 d-old plants
Whole-plant fumigation Unspecified hours for 2nd day     
Velikova et al. (2005)Laboratory observationGas exchange cuvette3003 hExcised leaf enclosurePhragmites australis L.6Immediately after fumigationI ~ 55% (iso)
Pot-grown plantsSingle-leaf fumigation       
Calfapietra et al. (2008)Field observationFACE65Long-termLeaf enclosurePopulus tremuloides (271 O3-tolerant)3Measurements of both clones performed at O3 concentration at which plants were growingNo significant effect (iso)
10-yr-old plants    Populus tremuloides (42 O3-sensitive)3 D ~ 20% (iso)
Ryan et al. (2009)Laboratory observationGrowth chamber1206 h for 8 dLeaf enclosurePopulus deltoides × P. trichocarpa (O3-tolerant)3–4Soon after fumigationNo significant effect (iso)
Pot-grown plants    P. deltoides × P. trichocarpa (O3-sensitive)  D ~ 18% (iso)

As summarized in Table 4, considerable differences in experimental design have been used and may be responsible for some of the observed differences in response to O3. O3 concentrations above 200–300 ppb do not have environmental relevance, and future experiments should use realistic exposures.

6. Interactions and feedback

Future increases in global temperature will occur simultaneously with other drivers and effects of global change (IPCC, 2007). Concern has already been expressed about how the relationship between plants and biotic stresses mediated by BVOCs may be altered in response to global change – future climatic conditions might strengthen or weaken the performance of herbivores and pathogens, depending on their traits (Manning & Vontiedemann, 1995; Ward & Masters, 2007). Similarly, global change may affect plant performance and hence may alter their defences against biotic stresses. As BVOCs have been shown to exhibit direct and indirect functions in plant defences (see above), alteration of BVOC emissions as a result of environmental changes may affect these defence mechanisms.

Although evidence of the influence of environmental change on the direct role of BVOCs in plant–herbivore interactions is lacking, much work has been carried out to investigate changes in indirect plant defences. O3 may interfere with parasitoid olfactory responses and damage their searching efficiency (Gate et al., 1995). Importantly, however, the rapid reaction of O3 with some BVOCs in the gas phase may degrade the BVOC signal from herbivore-infested plants. As noted above, exposure to O3 may suppress or enhance BVOC emission rates. Hence, elevated O3 may disrupt the plant–herbivore–predator/parasitoid system. However, some recent experiments have indicated that O3 does not affect the orientation of a predatory mite (Phytoseiulus persimilis) or parasitoid (Cotesia plutellae) (Pinto et al., 2007, 2008). It may be that natural enemies learn to exploit degraded BVOC products rather than the primary (emitted) BVOCs, or that long-distance signals between plants and predators or parasitoids could be provided by the more stable herbivore-induced volatile compounds, such as MeSA, methanol and benzyl cyanide (Pinto et al., 2007).

By contrast, elevated CO2 concentrations may disturb BVOC signals to the third trophic level by weakening the plant response induced by insect herbivores. However, this may vary with specific combinations of plants and herbivore enemies (Vuorinen et al., 2004b). Field studies have shown that interactions in a tree–herbivore–parasitoid system may be modified by O3 and elevated CO2 concentrations, and that the degree of modification is dependent on plant genotype (Holton et al., 2003).

Other abiotic factors, including water stress, light intensity, temperature and nutrient availability, are also important in determining the intensity and variability of induced plant volatiles. Water-stressed corn plants (Zea mays) produced larger amounts of induced plant volatiles than did nonstressed plants, although the former did not show any symptoms of desiccation (Gouinguene & Turlings, 2002). When grown under high light, undamaged Lima beans released larger relative amounts of volatile synomones and were more attractive to predatory mites than those grown under low light (Takabayashi et al., 1994). Changes in climatic factors can therefore alter significantly the relative ratios of the emitted BVOCs, and hence influence the quality of the induced odour blends. These studies have been undertaken on annual plants, and there is still a need to investigate such effects on perennial or woody plants which are abundant in forest ecosystems.

Although trends in BVOC emission rates as the Earth's climate changes are still uncertain, reactive BVOCs, especially isoprene, are of obvious concern, as they may give rise to species-specific feedbacks between plants and the atmosphere (Shallcross & Monks, 2000; Fuentes et al., 2001; Lerdau, 2007; Arneth et al., 2008b). Simplistically, it may be expected that climate warming will increase BVOC emissions, because of their strong temperature dependence, and so increase atmospheric concentrations, causing a decrease in the concentration of OH, and so leading to a reduction in the capacity of the atmosphere to remove tropospheric methane and CO, resulting in even further global warming. Enhancement of isoprene emissions in response to rising temperature may also have the dual effect of promoting tropospheric O3 production in NOx-polluted air, whilst contributing to reduced O3 damage to leaves in isoprene-emitting species (Loreto et al., 2001b; Velikova et al., 2005).

However, such simplistic models require considerable elaboration as many BVOCs serve to protect plants against biotic and abiotic stresses (see above). It is also possible that isoprene may serve multiple purposes in plants (Laothawornkitkul et al., 2008b), and therefore changes to BVOC emission rates caused by stresses may render the plants more susceptible to other stresses. Ultimately, these effects might be indirectly amplified by other consequences of global change, such as regional shifts in precipitation amount and pattern, the geographical redistribution of biomass/plant species, lengthening of the growing season and increases in invasive herbivore/pathogen species.

Present models are unable to adequately predict these possible interactions and feedbacks, partly because the combined effects of global warming with other global environmental drivers on BVOC emissions may not always give straightforward outcomes. Drought episodes, for example, may remove the positive effect of warming on isoprene emission (Fortunati et al., 2008), whereas enhanced UVB radiation, together with warming, may increase emissions (Tiiva et al., 2007). Changes in cloudiness driven by BVOC emissions and subsequent SOA formation will change the intensity of photosynthetically active radiation, so changing the emission rates of some light-dependent BVOCs. Although many experiments have explored the effects of global change parameters (e.g. temperature, CO2 and O3 concentrations, water stress, etc.) on BVOC emissions and possible disruption to their functions in and between plants, multivariate laboratory and field studies are needed to provide further understanding of possible interactions and feedbacks between environmental change and BVOC emissions.

V. Synthesis

It is clear that BVOCs emitted by the terrestrial biosphere have effects on the biological, chemical and physical components of the Earth system, providing connections between the biosphere and atmosphere, and between plants, insects and animal communities. However, the unprecedented pressure that humans are now exerting on the Earth system, and the impact that this is having on the global environment, may change the existing relationships mediated by BVOCs and lead to unforeseen consequences. Although our understanding of the sources, controls and effects of BVOCs has increased significantly over the past few decades, and now allows us to make informed (but still uncertain) predictions of their current emissions and of their responses to future global environmental changes, it is clear that there is still much more to be explored about the roles of BVOCs in the Earth system. In the near future, it seems likely that societal pressures around food security and more sustainable agricultural practises will promote further research into the role of BVOCs in tritrophic interactions and their use and development, through conventional breeding or genetic engineering, for crop protection (Poppy & Sutherland, 2004; Kappers et al., 2005). Similarly, increasing societal concern over air quality will inevitably drive further research into BVOC emissions and atmospheric chemistry. Concern over the Earth's climate system will also drive the development of coupled and interactive models of the Earth system which will better allow the role of BVOCs to be explored.

The exchange of resources and knowledge between atmospheric chemists and plant biologists, especially chemical ecologists, has greatly enhanced our understanding of the roles and impacts of BVOCs. The recent development of fast-response, highly sensitive (at the pptv level) analytical tools commonly used in atmospheric chemistry research, such as the proton transfer reaction mass spectrometer (Hewitt et al., 2003; Canagaratna et al., 2007), now allows rapid (Hz) BVOC concentration and flux measurements to be made. The application of such tools in plant ecology can, for example, allow the response time of stress application to be explored.

Although it is possible to factor several parameters into experiments or models to simulate the effects of global change on BVOCs, the incorporation of all the dimensions of global change into an experiment to mimic real conditions is not currently feasible. At present, it is therefore necessary to continue to probe this topic by, for example, combining experimental results, gradient studies, simulation modelling and remote sensing. Using these integrated approaches, it should be possible to make substantial progress in the mechanistic understanding of the effects of the important interactions mediated by BVOCs, and their potential to generate positive and negative feedbacks in response to future global change and climate warming. However, the interactive incorporation of all of these variables into a comprehensive model of the Earth system is still many years away.


The authors thank Alistair Hetherington for inviting us to write this review, the Engineering and Physical Sciences Research Council (EPSRC)/Royal Society Dorothy Hodgkin Postgraduate Awards to J.L., the European Science Foundation ‘VOCBAS’ programme and the EC FP6 ‘ISONET’ Marie Curie Research Training Network for financial support, and Malcolm Possell and Michael Wilkinson for stimulating discussions.