1 Introduction

Water is the most important matter for plant growth and productivity development for terrestrial ecosystems in the world (Prasolova et al. 2000, 2001; Xu and Chen 2006; Xu et al. 2000). Therefore, water availability and drought stress are intimately linked with the terrestrial carbon (C) cycling (van der Molen et al. 2011). Global climate change is likely to increase drought stress in many terrestrial ecosystems, which could alter the global patterns of organic matter production and decomposition (Feyen and Dankers 2009; Xu et al. 2009). Drought stress is expected to influence the C cycling more strongly in the future (Knorr et al. 2007). Ecosystem respiration is one of the largest gross fluxes in the global C cycle and integrates quite different processes of plant and microbial functions (Trumbore 2006; Casals et al. 2011). Soil CO2 emission, combined plant–root (R r), and heterotrophic (R h) and mycorrhizal respiration (R m) exceed anthropogenic CO2 emissions by about 10 times (Hanson et al. 2000; Kuzyakov 2006). Because auto- and heterotrophic activity belowground is largely controlled by substrate availability, soil respiration is strongly linked to plant metabolism, photosynthesis, microbes, extracellular enzymes, and litterfall (Ryan and Law 2005). Therefore, the coupling of the biological processes of above- and belowground regulates the response of R s to drought stress.

Climate-change scenarios indicate that many areas of the world will have much variation in precipitation patterns, with more frequent extreme rainfall events and longer drought stress periods for tropical and subtropical regions (Alley 2007; Allan and Soden 2008). Recently, many experiments have focused on the research meteorological drought stress affecting the soil CO2 emission across the different biomes. These studies suggested that drought stress could alter the relationship between soil CO2 efflux and its (environmental and biological) driving factors (Lavigen et al. 2004; Ruehr et al. 2009; van der Molen et al. 2011; Walter et al. 2011; Ford et al. 2012; Selsted et al. 2012; Signarbieux and Feller 2012). However, global climate models predict increased temperatures, and CO2 concentration would interact with changes in precipitation patterns to affect R s. In humid environments, increased temperatures may enhance R s by stimulating microbial activity, whereas in dry soils, they coincide with or even contribute to dryer conditions, resulting in an apparent negative R s response to temperature (Wan et al. 2007). Elevated CO2 concentration and extended drought stress showed a positive interaction, and elevated [CO2] can reduce the effect of drought stress on R s. Thus, the questions are how drought stress is coupled with soil respiration and how increased temperature and CO2 concentration regulate the relationship between drought stress and soil CO2 efflux. Here, we review the state of understanding of the relationship between drought stress and the soil respiration, and make some suggestions for further study in the context of global climate change. The paper is organized around a few aspects relevant to drought stress and the soil respiration and its components (Fig. 1). These include the following: (1) How does drought stress directly and indirectly influence soil respiration-observed results? (2) What are the mechanisms for the responses of soil respiration to drought stress? (3) What is the effect of drought stress on soil respiration in the context of climate change?

Fig. 1
figure 1

Conceptual framework outlining the distinctions among direct effects of soil water availability (SWA) on soil respiration in situ under climate change conditions (solid line), the effects of SWA on soil respiration that occur via changes in microbial activity, community composition and extracellular enzymes dynamics and the indirect effects of climate on soil microorganisms (dotted lines), and shifts in relationship between the aboveground and belowground processes. The interaction of global change drivers (atmospheric N deposition, elevated atmospheric CO2, and warming) and SWA affects indirectly on the soil respiration

Full size image

2 How drought stress influence soil respiration-observed results

2.1 The direct effect of drought stress on soil respiration and its components

In this paper, the drought stress is defined as precipitation deficiency deviated from normal conditions with respect to a percentage of normal rainfall (Joos et al. 2009; Brilli et al. 2011; Signarbieux and Feller 2012) and a period of more than some particular number of days with precipitation less than some specified small amount (Walter et al. 2011; Bloor and Bardgett 2012). We reviewed the studies of effects of soil water availability on R s, to which drought stress applied. According to C pools recognized as sources of CO2 efflux, agents of CO2 production, and locations of CO2 production, Kuzyakov (2006) suggested that the five sources are the main contributors to total soil CO2 efflux: (1) basal respiration derived from decomposition of soil organic matter (SOM), (2) priming effect induced microbial decomposition of SOM, (3) microbial decomposition of dead plant remains, (4) rhizomicrobial respiration produced by microbial decomposition of rhizodeposits of living roots, and (5) root respiration. Because it was difficult to identify the process of items 2–4, in this study, we partition R s into autotrophic respiration (R a) including root respiration (R r) and mycorrhizal respiration (R m) and heterotrophic respiration (R h) including items 1–4.

Components of soil respiration followed the different seasonal trends but all were affected by drought stress (Rey et al. 2002; Risk et al. 2012). It is stated that drought stress can reduce R s ratio in the mesic and xeric ecosystems (Borken et al. 2006; Joos et al. 2009) (Fig. 2a), while it can stimulate R s in the hydric ecosystem (Savage and Davidson 2001; Knorr et al. 2007; Sowerby et al. 2008) (Fig. 2b). For example, soil respiration may decline by 25–50% in response to modest water stress (minimum Ψs of −0.6 to −0.2 MPa) in a balsam fir ecosystem (Lavigen et al. 2004). The degree of R s reduced depends on community composition (Talmon et al. 2011; Breulmann et al. 2012), species richness (Correâ Dias et al. 2010), and the intensity and frequency of drought stress.

Fig. 2
figure 2

Hypothetical variation in soil respiration (R s) and its components (root respiration, R r; mycorrhizal respiration, R m; and heterotrophic respiration, R h along drought stress gradients in a mesic and xeric ecosystem and b a hydric ecosystem. The vertically dashed lines in the top panel means the R h approximate 0 when Ψs ≈ −14 MPa

Full size image

Since components of soil respiration have the potential to respond quite differently to climatological controls, microbial community composition, and climate change, the impacts of drought stress on R r, R h, and R m are distinctive in different ecosystems (Fig. 2). The autotrophic component of soil CO2 efflux is dominant and mainly from maintenance respiration of root, whereas the heterotrophic respiration derived from microbial respiration is minimized or even ceased during the drought stress period (Casals et al. 2011; Bulter et al. 2012). Overall, R h is more sensitive to drought stress than R a (Fig. 2a). At the initial stage of drought stress, R h begins to decline at the upper soil layers, whereas R r remains stable and thus increases its relative contribution to total soil efflux (Unger et al. 2010). However, an extreme drought stress causes greater reductions in the heterotrophic component of soil respiration compared with the rhizospheric component (Scott-Denton et al. 2006). Some studies suggested that the microbial respiration is more strongly affected than root respiration by intensive and long-term drought stress events(Scott-Denton et al. 2006; Williams 2007; Suseela et al. 2012), while other researchers found that root respiration was more sensitive than microbial respiration to water stress (Lavigen et al. 2004; Carbone et al. 2011). The reasons for these conflicting experimental results may contribute to that were run at the different ecosystems and different temporal measuring scale. Because a thinner film of water coats the soil particles, diffusion of labile substrates is slowed and the activity of exoenzymes needed for the decomposition of organic matter in drier soils is reduced (Suseela et al. 2012). Therefore, the mycorrhizal respiration rate decreases during a period of drought stress and is more sensitive to drought stress than R h (Fenn et al. 2010). In hydric ecosystems, R s has been observed to increase as water tables drop and peat begins to dry out (Brown 1998; Alm et al. 1999; Aurela et al. 2001). Drying raised air-filled porosity in the soil and increase R h by stimulating the microbial activity. However, root associated respiration (R a) was concentrated in the uppermost peat layer and was insensitive to drought stress compared to R h (Knorr et al. 2008) (Fig. 2b).

The responses of components of soil respiration to drought stress intensity are different. Suseela et al. (2012) found R h decreased sharply when volumetric soil moisture dropped below 15% or exceeded 26%, but R h increased more gradually when soil moisture rose from the lower threshold. In mineral soil, microbial respiration is near zero when Ψs ≈ −14 MPa (Manzoni et al. 2012). The responses of R r, R m, and R h to the time of drought stress and duration are inconsistent (Lavigen et al. 2004; Carbone et al. 2011; Gomez-Casanovas et al. 2012). Therefore, it is urgent to quantify the variation in rates of soil respiration and its components under different intensity and frequency of drought stress in different terrestrial ecosystems. In addition, the effect of drought stress, especially extreme drought stress, on soil respiration can persist for 1–9 years. However, recent studies have always focused on 1–3 years. Therefore, long-term and continuous measurements are also necessary.

2.2 Drought stress indirect affects through the coupling of temperature and soil respiration

Respiration of both microbial communities and plant roots can be sensitive to variations in soil temperature. The sensitivity of soil respiration to soil water content and soil temperature may be a function of the proportions of auto- and heterotrophic components of soil respiration at any given time (Gomez-Casanovas et al. 2012). Exponential relationships, especially the Q 10—the increase in respiration rates with a 10 °C increase in temperature—have been commonly used to estimate R s rates with the temperature (Epron et al. 1999; Buchmann 2000; Rey et al. 2002). The coupling of temperature and soil respiration would be broken by drought stress since drought stress reduced the diffusion of soluble C substrate and the extracellular enzymes, consequently, limited the microbial activity (Manzoni et al. 2012). Soil–water deficit substantially reduced R s sensitivity to temperature, resulting in a lower Q 10 (Sampson et al. 2007; Jassal et al. 2008; Manzoni et al. 2012; Rey et al. 2002). The responses of temperature sensitivity of its components of soil respiration to drought stress are different, depending on the specific controlling factors. R r is largely associated with the physiological response of root to drought stress. The seasonal variation in root growth with temperature may explain the difference of Q 10 of R r (Boone et al. 1998; Rey et al. 2002). The growth and death of roots are very sensitive to drought stress, depending on plant functional types. C4 plants required the lower root biomass than C3 plants under the same water stress and have lower Q 10 of root-respiration as well than C3 during drought stress periods (Zhou et al. 2012). Moisture limitation can suppress microbial activity regardless of temperature, which should decrease the temperature sensitivity of R h (Davidson and Janssens 2006). The influence of drought stress on response of R h to temperature depends on seasonal dynamics and soil properties (Craine and Gelderman 2011).The sensitivity to temperature of R m under drought stress conditions is not as significant as those of R r and R h. The variation in Q 10 of R m depends on the presence of mycorrhizal and rhizospheric respiration sources, as well as on plant development (Moyano et al. 2007; Fenn et al. 2010).

3 What are the mechanisms for the responses of soil respiration to drought stress?

3.1 Direct effect of drought stress through soil microorganisms and extracellular enzymes on soil respiration

Soil organic matter decomposition ultimately contributes very significantly to soil CO2 efflux and thus to global CO2 emissions (Bond-Lamberty et al. 2004). The role of microbial population in how SOM decomposition responds to drought stress with different intensities and durations could be of paramount importance. The relative contributions of the microbial functional groups, such as fungi and bacteria, to SOM decomposition are different, and these groups directly influence terrestrial C cycling. The factors controlling ecological aspects of the soil microbial community directly limit the SOM decomposition. Drought stress not only limits physiological performance, diffusion of microbes, and the diffusion of nutrients to microbes in the soil pore space but also submits the microbial community to an important adaptive force. Furthermore, the limited substrate supply under drought stress reduces the activity of microbes and diminishes R s (Bardgett et al. 2005; Suseela et al. 2012). Apart from the effect of drought stress on R s through limiting microbial activity, the change of microbial community composition also may influence R s. The response of R h to water availability change depends on the relative importance of fungi in the decomposer community. As the soil dries, conditions shift from favorable to soil fauna and bacteria to more favorable to fungi and actinomycetes (Freckman 1986), and the microbial community structure may become dominated by drought stress-tolerant species (Williams and Hallsworth 2009; Wallenstein and HE 2011; Manzoni et al. 2012). Shifting from microbial communities to fungal dominant communities improves the C stability of the physical environment and generates a more protected and stable C (Bardgett et al. 2005; Manzoni et al. 2012). Alternatively, Chowdhury et al. (2011) found that drought stress did not significantly affect the microbial community abundance (bacteria and fungi), suggesting that large differences in soil respiration were due to modulation of the activity per cell. In addition, mobility of microbes can be limited by drought stress due to the lack of soil water. The SOM decomposition rate of mineral soils containing the bulk of organic matter is limited by microbes’ ability to access due to drought stress (John et al. 2005).

Extracellular enzymes are the “proximate agents of organic matter decomposition,” and its activity is generally used to infer shifts in microbial demand for C (Henry 2012). Microbial extracellular enzyme production is related to microbial activity as well as soil physico-chemical properties (Sowerby et al. 2005). While an ecosystem encounters drought stress, two steps are required for microbes to process the material wrapped by a thinner film of water coats—first a physical step of aggregate disruption, and then possibly a second step in which exoenzymes break up the polymers (Navarro-García et al. 2011). Thus, the slowed diffusion of labile substrates and the reducing activity of extracellular enzymes would limit the decomposition of SOM. During the period of drought stress, the activity of enzymes associated with C cycling, such as xylanse, β-cellobiosidase, and β-glucosidase, would reduce and thus slow the decomposition of litters (Kardol et al. 2010; Sanaullah et al. 2011). Drought stress also may affect the production of extracellular polymeric materials, which bridge between microbes and their substrates. It allows microbes to survive in dry soils and can either promote (Chenu and Roberson 1996) or constrain C diffusion to microbes (Holden et al. 1997). The activity of enzymes under drought stress conditions is also affected by plant successional stages in which the composition and activity of microbial communities change and evolve as a consequence of changes in the physico-chemical status of soils. Less enzyme activities and soil respiration reduction were observed in soils with mature plant covered than in degraded soils under drought stress (Fioretto et al. 2009). However, Freeman et al. (1996) found the elevated enzyme activities without an associated increase in microbial respiratory activity in a peatland ecosystem, suggesting that drought stress conditions influence peatland mineralization rates through a direct stimulation of existing enzymes, rather than through a generalized stimulation of microbial metabolism. Although these studies of in situ microbial communities can partially account for the mechanisms in physiological and community ecology of soil respiration under drought stress, molecular phylogenetics and functional genomics should be applied to link microbial ecology to the process of R s. Combinations of new molecular methodology and genomics may link microbial phylogeny with function in R s studies and provide significant insights into plant–microbe interactions in the R s.

3.2 The regulation of plant growth and composition and belowground C allocation to soil respiration under drought stress

Seasonal variations in soil respiration and its auto- and heterotrophic components have often been accounted for by changes in soil temperature and moisture (Davidson and Holbrook 2009; Phillips et al. 2011; Taneva and Gonzalez-Meler 2011). However, evidence from recent studies suggested that recent and current plant photosynthesis influences soil respiration and its components at different time scales in both forests and grasslands (Rey et al. 2002; Ruehr et al. 2009; Flanagan 2009; Fenn et al. 2010; Kuzyakov and Gavrichkova 2010; Moyano et al. 2012). Because the responses of the components of soil respiration to photosynthesis are different, the effect of drought stress on soil respiration through reduced plant growth, productivity and variation in vegetation composition would be distinctive. The response of R m to photosynthetic activity is more sensitive than R r and R h (Moyano et al. 2007). Photosynthesis supplies C substrate for root metabolism and growth, and a decrease in substrate supply can decrease R r within days (Bowling et al. 2002; Bardgett et al. 2005; Ruehr et al. 2009). Under drought stress condition, low-productivity grassland ecosystem has lower decomposition rate of root litters, influencing the microbial communities, as well as the ratio between fungal and bacterial biomass and patterns of soil enzyme activity (Breulmann et al. 2012). During drought stress periods, photosynthesis declines and stored nonstructural carbohydrates are used to maintain living tissues so that R r is temporarily decoupled from the canopy photosynthesis (Högberg et al. 2001).

Apart from the reduced coupling between canopy photosynthesis and the belowground process, the life types of plants, vegetation cover, and structure and community composition also influence the response of soil respiration to drought stress. Soil respiration rate decreases from wet to dry seasons in the shallow-rooted ecosystem, whereas deep-rooted forests maintain substantial soil respiration during the dry seasons (Davidson et al. 2000). Talmon et al. (2011) suggested that drought stress decreased the rates of soil respiration not only by lowering biological activity but also by drastically reducing shrub cover and in heterogeneous ecosystems, such as Mediterranean and deserts shrublands. Thus, there would be a major impact on R s by feedbacks through change in vegetation structure. Long-term drought stress may affect soil respiration by altering the community composition, particularly the presence of legume species, which affect soil organic C (SOC) dynamics and below-ground microbial processes, especially via roots. However, the regulating role of plant diversity in the response of R s to drought stress is still poorly understood. In general, higher plant diversity with higher productivity may partly counteract the impact of drought stress (Craine et al. 2001). The reduction of the N concentration of organic matter in more diverse plant community may negatively affect both R a and R h (Correâ Dias et al. 2010). Therefore, how changes in plant community diversity can affect soil respiration needs to be studied. In addition, drought stress also has effects on soil respiration through feedback influence of soil nitrogen (N) pools since the microbial processes that regulate soil N availability are sensitive to short-term variations in soil moisture. The increased duration and intensity of drought stress are usually linked to decreasing N mineralization and inorganic N fluxes (Borken and Matzner 2009).

In general, drought stress leads to an increase in C allocation to roots, although the magnitude and duration of drought stress responses differ among the studies (Gilgen and Buchmann 2009; St. Clair et al. 2009; Jentsch et al. 2011). However, some studies found that drought stress may decrease C flow (that is, C assimilate by photosynthesis, and carbon input into soil by litter and root excretion, and decomposed by microbes) from the roots to the soil compartments. Symbiotic fungi and many free-living microorganisms depend on the C transfers from roots, and microbial CO2 production is consequently reduced (Palta and Gregory 1997; Gorissen et al. 2004). In addition, about half of the soil CO2 efflux is thought to originate from recent (i.e., several days old) photosynthates (Högberg et al. 2008) and is rapidly transferred specifically to mycorrhizal fungi and other decomposers. However, drought stress decreases C assimilation and increases the residence time of recently assimilated C in leaf biomass, and thereby prolonging the time course of the recent C allocation from the plants to the soil (Högberg et al. 2008; Ruehr et al. 2009). Therefore, R s probably have a temporary decoupling with C allocation under drought stress conditions.

3.3 Substrate supply and litters affect soil respiration

Microbial-mediated decomposition of SOM ultimately makes a considerable contribution to soil respiration, which is typically the main source of CO2 arising from terrestrial ecosystems (Yuste et al. 2011). The effect of drought stress on the rates of SOM respiration and the sensitivity to temperature has been discussed above. Here, we focus on the issue about how drought stress indirectly influences soil respiration through shaping the supply and availability of substrate. Although drought stress can alter R r because of reduced photosynthates supply (Domec and Gartner 2003; Carbone et al. 2011; Gomez-Casanovas et al. 2012), it is not known to what extent this substrate for microbial respiration is derived from younger soil organic C, such as dead and/or older labile SOC free from the physical breakup of soil aggregates. In general, with water contents between desiccation stress (usually < −1.5 MPa matric potential) and field capacity (about −0.1 MPa matric potential), variation in drought stress is thought to affect microbial respiration primarily through reducing the diffusion of organic solutes (Grant and Rochette 1994; Sowerby et al. 2005; Davidson and Janssens 2006). As substrate availability largely depends on soil water content, this would explain why drought stress dampened the increase in respiration at higher temperatures. In the case of microbial respiration, substrate levels decreased during the drought stress periods as litters from the plants declined. The much weaker response or more resistance to drought stress of rhizospheric respiration is also likely related to their dependence on substrate supply. Among components of R s, the mostly affected one by substrate supply under drought stress conditions is R m, which exhibited a strong relationship with the photosynthetic activity of the previous day (Bardgett et al. 2005; Ryan and Law 2005; Unger et al. 2010, 2012; Heinemeyer et al. 2012).

Litter decomposition not only supplied the substrates for microbial activity but also emitted CO2 to the atmosphere. Drought stress slowed leaf litter C and N decomposition by more than 50% compared to regular weather conditions, mainly by strongly decreasing the decomposition rate constants and also depending on litter quality (Sanaullah et al. 2012). The CO2 derived from the litter decomposition and contributed approximately 3–4% of the total CO2 efflux during drought stress periods may decline by 74%, indicating that drought stress would have a stronger effect on the CO2 release from litter than on the belowground-derived CO2 efflux (i.e., SOM and root respiration) (Joos et al. 2009; van Straaten et al. 2010). Theis et al. (2007) found in an alpine grassland that during the drought stress period the CO2 efflux from litters and top soil horizons were close to zero due to the desiccation of these layers. Soil respiration was obviously originating from the deeper soil horizons with different temperature and moisture regimes. However, recent radiocarbon data suggested that respiration of young C substrates, such as those respired by living roots, would be less affected by drought stress than the microbial decomposition of older substrates in the litter layer in the forests (Borken et al. 2006) and grassland ecosystems (Xiang et al. 2008). In addition, plant litter quality also controls litter decomposition, apart from water availability. How the influence of the interaction of drought stress and quality on plant litter decomposition contributes to the R s is not well known. During the period of drought stress, some fluctuations in soil respiration occurred independently of microbial growth and abiotic interactions, such as desorption in combination with solvation status and conformational changes of enzymes (Alarcón-Gutiérrez et al. 2010).

4 Effect of drought stress on soil respiration in the context of climate change

4.1 The effect of the interaction of elevated temperature, CO2 concentration, and drought stress on soil respiration

Soil respiration and gross primary productivity (GPP) are the most important flux between the atmosphere and terrestrial ecosystems in terms of C quantities (Schimel 1995; Schlesinger and Andrews 2000). Much evidence suggested that temperature (Davidson and Janssens 2006; Aanderud et al. 2013), soil water content (Davidson et al. 1998; Metcalfe et al. 2011), and GPP (Davidson and Holbrook 2009; Janssene et al. 2001; Martin et al. 2012) are three common drivers for respiration. As atmospheric [CO2] continues to rise due to fossil fuel combustion and land use change, global mean surface temperature is projected to concurrently increase by 1.1–6.4 °C towards the end of this century (Solomon et al. 2007). Increased temperature may directly accelerate R h and R a by stimulating the activities of soil microbes and plant roots (Rustad et al. 2001; Melillo et al. 2002) and indirectly stimulate R h by stimulating N mineralization, primary production, litter production, and therefore substrate and N availability (Stromgren and Linder 2002; Pendall et al. 2004). Recent model simulation suggested the magnified R a responses to temperature increasing with the depressed R h responses across the whole range of soil temperature anomaly (0–8 °C) under extended drought stress conditions (Shen et al. 2009). In general, under drought stress, warming can cause further drying of the ecosystem and hence act to further reduce R s (Selsted et al. 2012). However, the study manipulated on a temperate spruce suggested that R h and R a were affected to the same extent by soil drought stress and warming, and if summer drought stress become more server in the future, warming reduced C loses will likely to offset by reduced soil efflux during summer drought stress (Schindlbacher et al. 2012).

Elevated CO2 often accelerates R s and R a through direct stimulation on photosynthetic activity (Nowak et al. 2004; Sarah Bachman et al. 2010), root growth, litter production, and rhizospheric deposition (Zak et al. 2000; Norby and Luo 2004) or indirect impacts on litter chemistry, and microbial community composition (Pendall et al. 2004; Wan et al. 2007; Pregitzer et al. 2008) and enzyme activity (Kandeler et al. 2006). Importantly, an elevated CO2 environment may also increase plant water use efficiency and thus soil water contents due to the reduction in stomatal conductance and alleviate the impact of drought stress (Dermody et al. 2007; Leuzinger and Körner 2010). Elevated CO2 may reduce the impact of summer drought stress on R s, depending on seasonal variation (Selsted et al. 2012). However, the studies about interactive effects of multiple factors on R s and its components under the scenarios of upcoming drought stress are lacking, especially the studies about underlying mechanisms. Therefore, furthermore work is needed to establish the comprehensive research and evaluate the integrated impact of multiple factors on soil respiration under drought stress in different ecosystems.

4.2 The impact of drought stress legacy on soil respiration

The legacy of drought stress can regulate the responses of soil respiration rates to rewetting events in terrestrial ecosystems through changes in abiotic drivers and in microbial community structure (Gordon et al. 2008; Bapiri et al. 2010; Evans and Wallenstein 2012; Göransson et al. 2013). Upon rewetting dry soils, within minutes, the respiration rates increase to a magnitude of 10–20-fold that in the continually moist soil. Wood et al. (2001) attributed this drastic response of R s during rewetting to the stress response of microbes to suddenly increase in moisture and soil water potential: Microbes must expend energy to regulate osmotic pressure to their microenvironment and release solutes before osmotic pressure bursts cells. However, after reaching peak respiration rates, the respiration rate in continuous moist soils then remain at rates higher than those in the drought stress-treated soils (Göransson et al. 2013). The soil respiration rates would change less in response to dry-rewetting pulses in the experienced extreme drought stress patterns, due to the adaptation of soil microbial communities, than those that experienced ambient rainfall events (Evans and Wallenstein 2012). However, these studies was operated generally with a coupled field-laboratory experiment and persisted in short term. It is unknown how the drought stress history with long-term modifications of the timing and intensity of drought stress events affect soil respiration after rewetting events through altering microbial community composition and function. We also do not understand whether microbial adaptation to precipitation regimes can affect soil respiration. In general, long-term or severe drought stress may cause the extensive root mortality. Root mortality reduces mycorrhizal fungal abundance by severing their access to root C (Högberg et al. 2001; Landesman and Dighton 2011) and the quantity of root exudates (Clarholm 1981), which can suppress microbial growth following rainfall, resulting in declines of R h. However, we have no direct evidence to prove that root respiration will change less in the experienced drought stress in response to rewetting events than that experienced ambient rainfall events. During the peak drought stress, the plants with mycorrhizas had greatest resistance to drought stress (Davies et al. 1993; Hughes et al. 2008). Therefore, we assume that the soil respiration with mycorrhizas will change less in the experienced drought stress in response to rewetting events than that experienced ambient rainfall events.

5 Conclusions

There are a lot of evidence that drought stress, characterized by the severity, duration, and frequency, are expected to influence the C cycling more strongly in the future. Soil respiration is one of the most important biological processes in C cycling. Drought stress reduces rates of soil respiration in mesic and xeric ecosystems through decoupling between the above- and belowground processes and increases soil respiration in hydric ecosystem through stimulating soil microbial activity. However, intensity and frequency, duration of drought stress impacts on the soil respiration are not well understood, particularly in terms of the synergetic controlling of driver factors. In addition, how the responses of R r, R m, and R h to drought stress are also not well understood, due to the difficulty of partitioning the components of soil respiration. Therefore, further work should focus on how the responses of soil respiration and its components to different drought stress scenarios and how the controlling factors of soil respiration function on the same temporal scale. The next challenge is to establish the relationship between soil respiration and global change parameters (such as warming and elevated [CO2] under drought stress. It is worth pointing out that the models of simulating the rates of soil respiration and its components under global climate change and drought stress should be developed.