Diurnal and seasonal variation in the carbon isotope composition of leaf dark-respired CO2 in velvet mesquite (Prosopis velutina)

Authors


W. Sun. Fax: +1 307 766 6403; e-mail: wsun@uwyo.edu

ABSTRACT

We evaluated diurnal and seasonal patterns of carbon isotope composition of leaf dark-respired CO2 (δ13Cl) in the C3 perennial shrub velvet mesquite (Prosopis velutina) across flood plain and upland savanna ecosystems in the south-western USA. δ13Cl of darkened leaves increased to maximum values late during daytime periods and declined gradually over night-time periods to minimum values at pre-dawn. The magnitude of the diurnal shift in δ13Cl was strongly influenced by seasonal and habitat-related differences in soil water availability and leaf surface vapour pressure deficit. δ13Cl and the cumulative flux-weighted δ13C value of photosynthates were positively correlated, suggesting that progressive 13C enrichment of the CO2 evolved by darkened leaves during the daytime mainly resulted from short-term changes in photosynthetic 13C discrimination and associated shifts in the δ13C signature of primary respiratory substrates. The 13C enrichment of dark-respired CO2 relative to photosynthates across habitats and seasons was 4 to 6‰ at the end of the daytime period (1800 h), but progressively declined to 0‰ by pre-dawn (0300 h). The origin of night-time and daytime variations in δ13Cl is discussed in terms of the carbon source(s) feeding respiration and the drought-induced changes in carbon metabolism.

INTRODUCTION

Carbon isotope ratio (δ13C) measurements have emerged as an important tool for probing metabolic processes and tracing carbon fluxes at ecosystem and larger scales. The δ13C value of CO2 exchanged between the land surface and atmosphere is useful for partitioning net CO2 exchange into autotrophic and heterotrophic component fluxes and investigating terrestrial carbon cycle responses to environmental change (Lloyd et al. 1996; Bowling, Tans & Monson 2001; Scartazza et al. 2004; Knohl & Buchmann 2005). Autotrophic respiration is one of the major components of biosphere–atmosphere CO2 exchange; approximately 50% of plant-assimilated carbon is quickly released back to the atmosphere through this pathway (Ryan 1991). Estimation of autotrophic respiration fluxes and sources based on δ13C measurements requires that isotopic signatures of these fluxes be modelled or measured at the appropriate spatial and temporal scales. This is challenging over short-time scales because leaf dark-respired CO2 is enriched in 13C relative to primary respiratory substrates (Duranceau et al. 1999; Ghashghaie, Duranceau & Badeck 2001) and δ13C values of released CO2 can potentially shift by up to 8‰ over a diurnal period (Hymus et al. 2005; Prater, Behzad & Jeffery 2006; Werner et al. 2007; Priault, Wegener & Werner 2009).

Highly 13C-enriched values of leaf dark-respired CO2 compared with primary respiratory substrates have been observed in a number of species and across a range of environmental conditions (Duranceau et al. 1999; Ghashghaie et al. 2001). Changes in 13C enrichment occur over very short (minute) (Barbour et al. 2007; Werner et al. 2007) and diurnal time scales (Duranceau, Ghashghaie & Brugnoli 2001; Ghashghaie et al. 2001; Tcherkez et al. 2003; Hymus et al. 2005; Prater et al. 2006; Werner et al. 2007; Gessler et al. 2009; Priault et al. 2009). At the leaf scale, the apparent 13C/12C fractionation between respiratory substrates and CO2 is partly attributed to the non-statistical 13C distribution within hexose molecules (Rossmann, Butzenlechner & Schmidt 1991; Duranceau et al. 1999; Ghashghaie et al. 2001; Tcherkez et al. 2003) resulting from isotope effects of aldolase involved in the formation of fructose-1,6-bisphosphate from triose phosphates (Gleixner & Schmidt 1997; Schmidt 2003). Incomplete oxidation of hexoses associated with C allocation to biosynthesis can cause the δ13C value of leaf dark-respired CO2 (δ13Cl) to be enriched by up to 6‰ (Ghashghaie, Badeck & Lanigan 2003; Hobbie & Werner 2004). This respiratory apparent 13C/12C fractionation is believed to be partly responsible for diurnal variation in δ13Cl, which varies among plant functional groups and is highly influenced by plant resource availability (Hymus et al. 2005; Prater et al. 2006; Werner et al. 2007; Priault et al. 2009). Decarboxylation of 13C-enriched phosphoenolpyruvate carboxylase-derived malate associated with light-enhanced dark respiration (LEDR) (Azcon-Bieto 1986; Atkin, Evans & Siebke 1998) is also partly responsible for the observed 13C enrichment in CO2 evolved from darkened light-acclimated leaves and the daytime increase in δ13Cl (Barbour et al. 2007; Gessler et al. 2009).

Diurnal variation of δ13Cl is likely to be affected also by changes in the δ13C signature of primary respiratory substrates associated with short-term changes in photosynthetic 13C discrimination (ΔP). Leaf dark respiration is fuelled by two distinct pools: a fast pool that is exchanged within hours and a slow pool with a mean residence time of a few days (Schnyder et al. 2003). Recently assimilated photosynthates account for approximately 50% of the carbon lost by leaf dark respiration (Nogués et al. 2004). The isotopic signatures of the slow pools (soluble sugars, starch, lipids and cellulose) are expected to be relatively constant over diurnal time periods (Hymus et al. 2005; Göttlicher et al. 2006), whereas the δ13C signature of the fast pool is likely to shift quickly with changes in environmental and physiological conditions that alter ΔP. The impact of short-term environmental and physiological changes on the cumulative flux-weighted δ13C value of photosynthates (δ13Cpw) can be modelled from gas exchange measurements (Farquhar, O'Leary & Berry 1982). δ13Cpw integrates effects of current and previous environmental conditions on ΔP over the day, and is likely to provide a reasonable estimation of the δ13C value of respiratory substrates.

Environmental changes over diurnal and seasonal time periods can alter the degree of 13C enrichment in leaf dark-respired CO2 relative to respiratory substrates (Ghashghaie et al. 2003; Hymus et al. 2005). Drought, for example, can affect δ13Cl through its impact on ΔP and associated changes in the δ13C value of photosynthates, and by altering the size and limitation of metabolite pools and the activities of key biosynthetic pathways affecting respiratory apparent 13C/12C fractionation. In general, drought stress is expected to reduce respiratory apparent fractionation as key substrates would be limiting. Soil drying alters δ13Cl under controlled conditions (Duranceau et al. 1999; Ghashghaie et al. 2001), but the effects of drought on diurnal and seasonal variation in δ13Cl under field conditions have not been explored. Drought-stressed plants are predicted to show less diurnal variation in ΔP because water limitation reduces the diurnal range of stomatal conductance and Ci/Ca (Valladares & Pearcy 1997). Understanding relationships between drought and diurnal patterns of δ13Cl is required to properly interpret the isotope signature of CO2 exchange in terrestrial ecosystems where water plays a key role in constraining plant physiological processes.

We measured δ13Cl in velvet mesquite (Prosopis velutina Woot.), a widespread and invasive leguminous shrub in the semiarid south-western USA, in xeric upland and mesic riparian habitats to evaluate how drought influences diurnal and seasonal patterns of this important isotopic signal. This work was motivated by our interest in how woody plant encroachment into grasslands alters coupling of carbon and water cycles in different landscape positions (Williams et al. 2006). This deep-rooted shrub can access ground water in riparian settings, reducing its reliance on precipitation, whereas it is completely dependent on limited rainfall supplies in upland habitats (Cable 1980; Snyder & Williams 2000; Fravolini et al. 2005). Our research objectives were to characterize diurnal variation in δ13Cl in P. velutina and to examine the effects of soil and atmospheric drought on the diurnal pattern and amplitude of δ13Cl. We expected to observe large diurnal variation in δ13Cl in both riparian and upland sites and through the dry pre-monsoon and wet monsoon seasons, but the magnitude of apparent respiratory 13C/12C fractionation would be lowest in the xeric upland setting and during the dry pre-monsoon period. We also expected to observe less diurnal shift in δ13Cl in P. velutina during the pre-monsoon dry period compared to the wetter part of the growing season, as severe soil and atmospheric drought would reduce the magnitude of diurnal variation in the δ13C value of recently fixed photosynthates.

MATERIALS AND METHODS

Study sites

Two velvet mesquite (P. velutina) populations, one in an upland xeric habitat and another in a riparian mesic habitat, were selected for this study. The riparian population was located on a remnant flood plain terrace of the San Pedro River, approximately 12 km east of Sierra Vista, Arizona, USA (31°40′N, 110°10′W, 1200 m elevation). Long-term (1982–2007) mean maximum and minimum temperatures are 25.1 and 9.5 °C, respectively, and mean annual precipitation is 368 mm in Sierra Vista (1982–2007; Western Regional Climate Center; http://www.wrcc.dri.edu). Approximately 60% of the total annual precipitation occurs during the summer monsoon period from July through September. The plant community at this site was dominated by 3 to 4 m tall P. velutina. Sacaton grass (Sporobolus wrightii Munro ex Scribn.) and smaller shrubs were abundant in scattered patches in the tree interspaces. Depth to groundwater at this site is about 6.5 m (Scott et al. 2006). Deep-rooted P. velutina plants on alluvial terraces along the San Pedro River obtain a large fraction of their water from the near-surface water table and experience only mild water stress over the growing season compared with plants in upland areas (Snyder & Williams 2000; Fravolini et al. 2005). Soils at this site consist mainly of gravelly sandy loam layers interspersed with clay and gravel lenses (Scott et al. 2006).

The upland P. velutina population was located in the Santa Rita Experimental Range 35 km south of Tucson, AZ, USA (31°47′N, 110°50′W, 1190 m elevation). The mean maximum and minimum temperatures are 24.7 and 11.1 °C, respectively, and mean annual rainfall is 563 mm (1950–2007; Western Regional Climate Center; http://www.wrcc.dri.edu). More than 51% of the total annual precipitation occurs during the summer monsoon. The plant community at this upland site was dominated by P. velutina, perennial grasses and sub-shrubs. Although P. velutina roots extend to great depths, roots do not access stable groundwater in these upland sites and winter and summer precipitation rarely penetrate below 2 m (Cable 1980; Frasier & Cox 1994; Fravolini et al. 2005). The Holocene-aged alluvial soils at this site have a sandy loam texture (Fravolini et al. 2005).

Field sampling of leaf dark-respired CO2

Sampling of leaf dark-respired CO2 was conducted during the dry pre-monsoon (June 5 and 7 at upland and riparian sites, respectively) and rainy monsoon (August 25 and 27 at riparian and upland sites, respectively) seasons to capture differences in water stress conditions. Large-bore (60 ml) syringes were used to collect dark-respired CO2 samples from P. velutina leaves using a method modified from Werner et al. (2007). Several young, fully expanded leaves were collected from upper parts of P. velutina canopies that were fully illuminated during daytime periods. For each plant (n = 5), approximately 10 leaves were composited into a single sample, placed inside the syringe barrel and immediately flushed completely with CO2-free air five times by actuating the syringe plunger. Syringe barrels were completely opaque to prevent photosynthesis. CO2 from leaf respiration in the barrel was allowed to build for 15 min and then 5 ml of air in the barrel was injected into 12 ml helium-flushed vials fitted with septum caps. This process was repeated nine times for each of the five plants over a 24-h period (0600, 0900, 1200, 1500, 1800, 2100, 0000, 0300 and 0600 h). Vials were stored in the field in dry, insulated coolers and shipped overnight to the University of Wyoming Stable Isotope Facility for isotopic analysis.

Water potential and leaf gas exchange measurements

Leaf gas exchange rates were measured concurrently with the collection of leaf dark-respired CO2 using a LI-6400 portable photosynthesis system (LI-COR Biosciences, Lincoln, NE, USA) on three replicate leaves on each of the five P. velutina trees used for respiration sampling. Young, fully expanded leaves from near the top of the canopy in fully illuminated locations were used for these measurements. For each measurement, environmental conditions inside of the leaf chamber (i.e. photosynthetically active radiation, chamber block temperature, relative humidity and CO2 concentration) were set to match ambient conditions. Vapour pressure deficit at the leaf surface (Dl) was calculated as the difference between the saturation vapour pressure in the sub-stomatal cavity and the vapour pressure in the leaf chamber, which closely matched that of ambient conditions. Dark respiration rate (R) during the daytime was measured by setting photosynthetically active radiation to zero. Leaf area was determined by scanning the leaflets used for gas exchange measurements. Pre-dawn leaf water potential (Ψpd) was measured on three P. velutina twigs for each of the five trees at each site between 0300 and 0500 h using a pressure chamber (PMS Instruments, Corvallis, OR, USA).

Starch and soluble carbohydrate extractions

Leaves comparable with those used for CO2 sampling and gas exchange measurements (same canopy position) were collected at each sampling interval over the 24 h measurement periods. Leaves were immediately flash frozen in liquid nitrogen, stored on dry ice and then freeze-dried using a Labconco freeze dryer (Labconco, Kansas City, MO, USA). Freeze-dried P. velutina leaves were ground into fine powder using a ball mill. Starch and soluble carbohydrates were then extracted following the procedures described by Göttlicher et al. (2006). Briefly, 100 mg powdered dry leaf material was extracted with 1.5 mL methanol/chloroform/water (MCW, 12:5:3, v/v/v) for 30 min at 70 °C. After centrifuging at 10 000 g for 2 min, 800 µL of the supernatant was transferred into a new vial for the extraction of soluble carbohydrates. The pellet was washed with deionized water and oven-dried after re-extraction with MCW three times. Chloroform was added to the supernatant to induce phase separation. The upper phase containing soluble carbohydrates was transferred to a self-made ion exchange column for the separation of the neutral fraction from the total low molecular weight compounds. The column was filled with a mixture of anion-exchange resin (Dowex 1 × 8, Sigma-Aldrich, St. Louis, MO, USA) and cation-exchange resin (Dowex 50W × 8, Sigma-Aldrich) with deionized water as mobile phase. The eluted neutral fraction was collected, oven-dried and re-dissolved in 1 mL water. Fifty microlitres of this sample was placed into a tin capsule, oven-dried at 65 °C, and then analysed on an isotope ratio mass spectrometer (described below).

Starch in the oven-dried pellet was gelatinized at 100 °C in a water bath for 15 min. Gelatinized starch was hydrolysed by heat stable α-amylase (Sigma-Aldrich) at 85 °C for 120 min. The aqueous phase, containing sugars derived from starch, was separated from the pellet through centrifugation at 10 000 g. The aqueous phase was purified in a centrifugal filter device (Microcon YM-10, Millipore, Billerica, MA, USA). Fifty microlitres of the filtrate was transferred to a tin capsule and oven-dried at 65 °C for stable isotope analysis. We extracted starch and sucrose standards with the methods described above and measured their stable carbon isotope composition. No significant isotopic fractionation was detected during both starch and soluble carbohydrate extractions.

Isotope ratio mass spectrometry

All isotope analyses were performed at the University of Wyoming Stable Isotope Facility. Leaf dark-respired CO2 was analysed within 5 d of field collection using a GasBench II attached to a Finnigan Delta + XP continuous flow inlet isotope ratio mass spectrometer (Thermo Finnigan, Bremen, Germany). Repeated measurements with laboratory CO2-in-air working standards had a precision of <0.1‰. The δ13C of CO2 in respiration samples were corrected to the international Vienna Pee Dee Belemnite (VPDB) standard using CO2-in-air working standards calibrated with CO2-in-air standards obtained from National Oceanic and Atmospheric Administration, Climate Monitoring and Diagnostic Laboratory. The δ13C of leaf biomass, starch and soluble carbohydrates were determined by a Micromass IsoPrime continuous flow isotope ratio mass spectrometer (Micromass, Manchester, UK). Precision of repeated measurements of laboratory standards was <0.1‰. Carbon isotope ratios are reported in parts per thousand relative to VPDB as: δ13C (‰) = (Rsample/Rstandard − 1) × 1000.

δ13C value estimates of the leaf photosynthate pool

δ13C of the recently fixed photosynthate pool was estimated as:

image(1)

where δ13Cpwi is the assimilation-weighted, cumulative carbon isotopic value of the recently fixed photosynthates at time i (0600, 0900, 1200, 1500 and 1800 h); Ai and δ13Cpi are the instantaneous net assimilation rate and the δ13C value of photosynthates at time i, respectively. δ13Cpi was solved from:

image(2)

where δ13Ca represents the δ13C value of atmospheric CO2 (assumed to be −8‰).

Δp, the discrimination against 13C during photosynthesis, was estimated from the simplified linear model described by Farquhar et al. (1982), which assumes that mesophyll conductance is constant and relatively large.

image(3)

The constant a represents the fractionation occurring during diffusion through air (4.4‰); b is the net fractionation associated with carboxylation (27‰); Ci and Ca are the concentrations of CO2 in the sub-stomatal cavity and the atmosphere, respectively. Although b could vary slightly with temperature (O'Leary, Madhavan & Paneth 1992), the range of daytime leaf temperatures expected across sites and from pre-monsoon to monsoon seasons in this study would not be expected to alter this term significantly.

Statistical analysis

We conducted repeated measures analysis of variance to test the effects of site, season and sampling time as well as their interactions on the δ13C of leaf dark-respired CO2, leaf starch, leaf carbohydrates and bulk leaf biomass. Since we sampled leaves repeatedly from the same trees over 24 h field campaigns during both pre-monsoon and monsoon seasons within each site, time and season were treated as within subjects factors and site (riparian and upland) was treated as a between subjects factor in the analysis. Simple linear regression analysis was conducted to evaluate the dependence of δ13Cl on δ13C of photosynthates, leaf starch, leaf-soluble carbohydrates and bulk leaf biomass. All statistical analyses were carried out using SAS version 9.0 (SAS Institute Inc., Cary, NC, USA). Average values are reported as the arithmetic mean ± 1 standard deviation.

RESULTS

Water potential and leaf gas exchange

Prosopis velutina twigs at the riparian site during the pre-dawn period equilibrated to higher water potentials (Ψpd) than did those at the upland site during both pre-monsoon and monsoon sampling periods. Ψpd increased from −3.7 ± 0.2 to −1.3 ± 0.1 MPa from the pre-monsoon to monsoon sampling periods in P. velutina at the upland site, whereas only a slight increase in Ψpd, from −1.1 ± 0.2 to −0.5 ± 0.1 MPa, was observed at the riparian site over this same period.

Net carbon assimilation rate (A) of P. velutina trees growing at both riparian and upland sites reached a maximum between 0600 and 0900 h, and between 0900 and 1200 h, respectively, for pre-monsoon and monsoon sampling periods, and then gradually declined thereafter over the remaining daylight period (Fig. 1e). Surprisingly, no differences were observed in the mean A values for plants at the riparian and upland sites during either the pre-monsoon or monsoon, although those at the riparian site had higher pre-dawn water potential values. Stomatal conductance (gs), vapour pressure deficit at the leaf surface (Dl) and Ci/Ca also did not differ among plants at the two sites during the pre-monsoon period (Fig. 1c, d & g), suggesting that soil water availability was not the only factor constraining plant carbon assimilation. High Dl (Fig. 1d) during the pre-monsoon season at both sites likely had important impacts on A, gs and Ci/Ca.

Figure 1.

Diurnal patterns of environmental and physiological parameters for Prosopis velutina trees growing at riparian (circle) and upland (triangle) sites during pre-monsoon (solid) and monsoon (open) sampling periods. Diurnal trends in (a) leaf temperature (°C), (b) photosynthetically active radiation (PAR, µmol m−2 s−1), (c) stomatal conductance (gs, mol·m−2 s−1), (d) leaf vapour pressure deficit (Dl, kPa), (e) net assimilation rate (A, µmol m−2 s−1), (f) dark respiration rate (R, µmol m−2 s−1) and (g) Ci/Ca ratio. Data are the means of five replicate trees (standard deviations are shown when larger than symbols). The shaded area denotes the dark period.

A increased after the onset of monsoonal rainfall at both riparian and upland sites. Increased A during the monsoon season was likely due to increased soil water supply and reduced Dl and leaf temperature, as photosynthetically active radiation was not appreciably different between the pre-monsoon and monsoon sampling dates (Fig. 1a, b & d).

Leaf respiration rate (R) gradually increased over the daytime period, reaching maximum values at 1500 h (Fig. 1f). R rapidly decreased thereafter to minimum values at the beginning of the night-time period and fluctuated near minimum values for the remainder of the night-time (Fig. 1f). R at the upland site, in general, was higher than that at the riparian site, but only during the monsoon period.

δ13C of biomass, soluble carbohydrates, starch and leaf dark-respired CO2

No diurnal variation was detected for carbon isotope composition of leaf biomass (δ13Cb), but δ13Cb was lower in leaves collected during the monsoon compared with those collected during the pre-monsoon period at both upland and riparian sites (Fig. 2; Table 1). Significant site differences (P < 0.001) in δ13Cb were also detected; δ13Cb values for the upland P. velutina plants were 2.0 ± 0.3 and 1.2 ± 0.3‰ more positive than for plants at the riparian site during pre-monsoon and monsoon sampling periods, respectively (Table 1). The carbon isotope composition of starch (δ13Cs) and total soluble carbohydrates (δ13Csc) showed no clear diurnal shift (Fig. 2). Compared with bulk leaf biomass, soluble carbohydrates were enriched in 13C by 2–3‰. No diurnal shifts or site differences were observed in δ13Csc, but δ13Csc values did decrease slightly from pre-monsoon to the monsoon period (Table 1). Soluble carbohydrates were depleted in 13C by 7.0 ± 0.6 and 5.8 ± 0.6‰ relative to bulk leaf biomass during pre-monsoon and monsoon periods, respectively, at the upland site and by 5.7 ± 0.5 and 4.4 ± 0.4‰ at the riparian site.

Figure 2.

Diurnal patterns of carbon isotope signatures (‰) of leaf dark-respired CO2 (solid circle), bulk leaf biomass (open circle), soluble carbohydrates (solid triangle) and starch (open triangle) from pre-monsoon and monsoon sampling periods for Prosopis velutina trees growing at upland and riparian sites. Carbon isotope ratios of leaf dark-respired CO2 are the means of five replications. Carbon isotope ratios of bulk leaf biomass, soluble carbohydrates and starch are the means of three replications (standard deviations are shown when larger than symbols). The shaded area denotes the dark period.

Table 1.  Degrees of freedom (d.f.) and P values from the repeated measures anova of carbon isotope composition of leaf dark-respired CO2 (δ13Cl), soluble carbohydrates (δ13Csc), starch (δ13Cs) and bulk leaf biomass (δ13Cb)
FactorsSiteSeasonTimeSite × seasonSite × timeSeason × timeSite × season × time
  1. Site is the between subjects factor and season and time are within subjects factors.

  2. anova, analysis of variance.

d.f.1181888
δ13ClP<0.01<0.01<0.010.120.520.020.06
δ13CscP0.570.020.670.440.410.610.35
δ13CsP<0.01<0.010.700.780.710.520.65
δ13CbP<0.01<0.010.99<0.010.720.900.89

The carbon isotope ratio of leaf dark-respired CO2 (δ13Cl) demonstrated systematic diurnal variation (Fig. 2). Similar diurnal patterns in δ13Cl were observed between P. velutina plants at riparian and upland sites; δ13Cl values became progressively more positive during the daytime period reaching maximum values at 1800 h, and gradually declined thereafter at the end of the photoperiod. δ13Cl values were significantly different between pre-monsoon and monsoon sampling periods within each site (Table 1). δ13Cl values during the monsoon sampling period were 3.0 ± 0.9 and 2.6 ± 0.9‰ more negative than those observed during the pre-monsoon period at riparian and upland sites, respectively. δ13Cl values during the pre-monsoon and monsoon periods were 1.3 ± 0.8 and 1.7 ± 0.6‰ more positive, respectively, for plants at the upland compared with those at the riparian site.

Leaf dark-respired CO2 in P. velutina was enriched in 13C relative to the soluble carbohydrate pool at riparian and upland sites and during pre-monsoon and monsoon periods. The magnitude of this enrichment varied with the time of the day. Compared with starch, leaf dark-respired CO2 was 13C-enriched for most of the daytime during the pre-monsoon sampling period at the riparian site. 13C enrichment in leaf dark-respired CO2 relative to starch was observed only between 1500 and 2100 h at the upland site during both pre-monsoon and monsoon periods and at the riparian site during the monsoon period (Fig. 2). Leaf dark-respired CO2 was either 13C-depleted relative to starch or had similar δ13C values compared with the starch pool during the night-time period. Leaf dark-respired CO2 was 13C-enriched relative to the bulk leaf material, except late in the night-time through the early morning hours.

Significant positive correlations were detected between δ13Cl and δ13Cb (Fig. 3a) and δ13Csc (Fig. 3b) and δ13Cs (Fig. 3c) across all sampling periods and sites, but the coefficients of determination were low and the distribution of data was highly clustered by season and site. The slopes of the relationship between δ13Cl and δ13Cb (1.17) and δ13Cs (1.10) were lower than that of δ13Cl and δ13Csc (1.95). Isotopic differences between δ13Cl and δ13Cb and δ13Csc and δ13Cs were highly influenced by diurnal and seasonal variation in δ13Cl.

Figure 3.

Relationships between carbon isotope signatures (‰) of leaf dark-respired CO2 (δ13Cl) and carbon isotope values of (a) bulk leaf biomass (δ13Cb), (b) soluble carbohydrates (δ13Csc) and (c) starch (δ13Cs) for riparian (circle) and upland (triangle) habitats and pre-monsoon (solid) and monsoon (open) seasons. Dashed lines represent the 1:1 line. Linear regression functions, r2 and P values are provided.

δ13C of the leaf photosynthate pool

The carbon isotope values of photosynthates (δ13Cp) in P. velutina at both riparian and upland sites modelled from gas exchange parameters varied widely during the light period (Table 2). During the pre-monsoon period, δ13Cp varied by 5.5 ± 0.5 and 7.4 ± 0.4‰ for riparian and upland sites, respectively. The diurnal range in δ13Cp increased during the monsoon period, with 8.8 ± 0.7 and 8.9 ± 0.5‰ for riparian and upland sites, respectively. We compared the modelled δ13Cp with the observed δ13Cl at five time points (0600, 0900, 1200, 1500 and 1800 h) over daytime periods. Significant correlation was detected between δ13Cl and instantaneous values of δ13Cp (Fig. 4a). However, δ13Cl was more strongly correlated with the δ13C values of the cumulative, assimilation-weighed photosynthates (δ13Cpw; Fig. 4b).

Table 2.  Diurnal variation in the δ13C value (‰) of photosynthates for P. velutina at riparian and upland sites during pre-monsoon and monsoon sampling periods
SiteSeason0600 h0900 h1200 h1500 h1800 h
  1. Values are modelled from gas exchange parameters. See text for details. Data are the means (±1 standard deviation) of five trees at each site sampled multiple times through the daytime period.

RiparianPre-monsoon−24.7 ± 1.8−21.0 ± 2.3−21.7 ± 2.8−25.0 ± 1.8−26.3 ± 2.3
Monsoon−33.1 ± 0.1−27.1 ± 0.7−24.9 ± 0.8−24.8 ± 0.9−30.1 ± 2.1
UplandPre-monsoon−23.4 ± 1.1−21.3 ± 1.6−20.9 ± 0.7−22.6 ± 1.5−28.0 ± 2.1
Monsoon−30.5 ± 1.2−25.6 ± 0.2−22.1 ± 1.2−22.0 ± 0.6−24.6 ± 1.7
Figure 4.

Relationships between carbon isotope signatures (‰) of leaf dark-respired CO2 (δ13Cl) and predicted carbon isotope ratios of (a) photosynthates (δ13Cp) and (b) predicted carbon isotope ratios of cumulative photosynthates (δ13Cpw) for riparian (circle) and upland (triangle) habitats and pre-monsoon (solid) and monsoon (open) seasons at 0600, 0900, 1200, 1500 and 1800 h. Dashed lines represent the 1:1 line. Linear regression functions, r2 and P values are provided.

DISCUSSION

Large diurnal changes in the carbon isotope ratio value of leaf dark-respired CO2 (δ13Cl) were observed in P. velutina, with δ13Cl increasing to maximum values late during daytime periods and declining gradually over night-time periods to minimum values at pre-dawn. These findings are similar to those for other drought-tolerant trees (Hymus et al. 2005; Prater et al. 2006; Priault et al. 2009). This general pattern of diurnal variation in δ13Cl in P. velutina held across upland and riparian habitats and from the dry pre-monsoon to the wet monsoon seasons, but the mean value and magnitude of diurnal variation in δ13Cl were strongly impacted by soil and atmospheric drought. Diurnal changes in δ13Cl have important implications for using δ13C measurements to constrain estimates of CO2 fluxes between ecosystems and the atmosphere (Bowling et al. 2001). The pronounced night-time shift in δ13Cl observed in P. velutina is especially relevant for studies employing the Keeling-plot approach (Keeling 1958) to partition night-time ecosystem respiration into its component fluxes. This approach assumes that the δ13C signatures of respiration sources within the ecosystem are constant over short time periods and can be accurately estimated or measured directly (Pataki et al. 2003). Accounting for diurnal and seasonal shifts in δ13Cl within this context requires better understanding of the effects of environmental conditions such as drought on δ13Cl and the mechanisms underlying these patterns.

Photosynthetic 13C/12C discrimination and primary respiratory substrates

An important starting point for understanding the δ13C signature of leaf dark-respired CO2 is to examine the δ13C signature of putative respiratory substrates. However, in P. velutina, diurnal variation in δ13Cl was not explained by changes in the carbon isotope ratio of leaf starch (δ13Cs) or total soluble carbohydrates (δ13Csc) (Fig. 3). Hymus et al. (2005) also did not observe co-variation over short time periods between δ13Cl and either δ13Cs or δ13Csc. Rather, isotopic analysis of sugars directly associated with leaf respiratory metabolism, such as glucose and recently fixed photosynthates, are likely to better reflect the controls on δ13Cl (Ghashghaie et al. 2001). Indeed, recently fixed photosynthates can account for about 50% of leaf dark respiration (Nogués et al. 2004). We estimated the carbon isotope composition of recently fixed photosynthates from gas exchange measurements and the simplified version of the photosynthetic discrimination model of Farquhar et al. (1982). During the daytime, δ13Cl in P. velutina was strongly correlated with the carbon isotope ratio value of the cumulative photosynthate pool (δ13Cpw) associated with short-term changes in photosynthetic carbon isotope discrimination (ΔP) (Fig. 4b). Increases in daytime δ13Cl values observed in other studies were correlated with cumulative photosynthesis, but not the instantaneous δ13C value of photosynthates estimated from gas exchange parameters (Hymus et al. 2005; Priault et al. 2009). However, these studies did not report values for δ13Cpw. Our results suggest that short-term changes in the δ13C value of photosynthates play an important role in determining daytime increases of δ13Cl in P. velutina growing under field conditions.

We observed progressive decreases of δ13Cl in P. velutina at the end of the photoperiod and through the early night-time period under natural field conditions, similar to patterns observed in leaves of other tree species (Hymus et al. 2005; Prater et al. 2006). Other studies have noted similar decreases in δ13Cl during extended dark periods in controlled settings (Nogués et al. 2004; Barbour et al. 2007; Werner et al. 2007; Priault et al. 2009). We plotted δ13Cl at 1800, 2100, 0000 and 0300 h against δ13Cpw recorded at the end of the daytime (1800 h) (Fig. 5) to investigate the role of recently fixed photosynthates in controlling night-time δ13Cl. δ13Cpw was strongly correlated with night-time δ13Cl across upland and riparian sites and the wet and dry seasons. But important offsets were observed between δ13Cl and δ13Cpw and the magnitude of the offset shifted over the night-time period. If we assume that the flux-weighted photosynthates were the primary respiratory substrates at night, then 13C enrichment of respired CO2 relative to these substrates decreased from about 6‰ at 1800 h to approximately 0‰ at 0300 h. The convergence of δ13Cl on δ13Cpw at 0300 h suggests that the night-time δ13C value of CO2 respired by the plant canopy in ecosystem scale studies can be accurately estimated from daytime gas exchange characteristics.

Figure 5.

Relationships between predicted carbon isotope ratios of the cumulative photosynthate pool (δ13Cpw) at 1800 h with carbon isotope signatures of leaf dark-respired CO2 (δ13Cl) at 1800 h (solid circle and dash line), 2100 h (open circle and dash-dot line), 0000 h (solid triangle and dash-dot-dot line) and 0300 h (open triangle and dot line), respectively, for both sites and both seasons. The solid line represents the 1:1 line.

Drought effects on δ13Cl

Leaf dark-respired CO2 in P. velutina during the dry pre-monsoon season was significantly 13C enriched compared to that during the wet monsoon season (Table 1). Water limitation-induced 13C enrichment in leaf dark-respired CO2 has also been observed in Nicotiana sylvestris and Helianthus annuus growing in controlled environment settings (Duranceau et al. 1999; Ghashghaie et al. 2001). Seasonal differences in δ13Cp values in P. velutina (Table 2) illustrate how soil water limitation and vapour pressure deficit at the leaf surface (Dl) influenced δ13Cl values through impacts on photosynthetic carbon isotope discrimination (ΔP). Drought-induced shifts in δ13Cl, therefore, can be an important component of seasonal variation in the carbon isotope composition of ecosystem respiration (δ13CR) in savanna ecosystems where physiological processes are strongly controlled by water availability.

The amplitude of the diurnal variation in δ13Cl in P. velutina was significantly affected by differences in soil water availability and Dl across riparian and upland sites and between the pre-monsoon and monsoon seasons (Fig. 2; Table 1). The diurnal amplitude in δ13Cl (calculated as δ13Cl at 1800 h minus δ13Cl at 0600 h) was greater during the monsoon period at riparian (5.1 ± 1.1‰) and upland (5.1 ± 0.9‰) sites than during the pre-monsoon period (3.3 ± 0.8‰ and 2.8 ± 0.7‰, respectively, at riparian and upland sites). Changes in the magnitude of diurnal variation in δ13Cl could potentially influence patterns of δ13CR, if foliar respiration fluxes are large, relative to other component fluxes. Indeed, Bowling et al. (2003) observed that δ13CR varied significantly over a single night in a grassland ecosystem, which may have been due to night-time shifts in δ13Cl.

A positive correlation between δ13Cl and the δ13C value of photosynthates modelled from gas exchange measurements suggests that the magnitude of daytime increases in δ13Cl in P. velutina were controlled primarily by daytime changes in ΔP (Fig. 4b). Progressive changes in daytime ΔP can also affect the pattern of night-time decreases in δ13Cl. Transitory starch synthesis during the daytime and remobilization at night may induce progressive night-time shifts in the δ13C value of primary respiratory substrates. The remobilization of transitory starch follows the ‘first in, last out’ rule, which means that the first glucose to be polymerized into starch will be the last one to be broken down into maltose and exported out of the chloroplast (Zeeman, Smith & Smith 2007). Hence, if photosynthates that were used to synthesize starch were gradually 13C enriched during the daytime, then the δ13C value of respiratory substrates should become progressively reduced over the night-time period. This hypothesis is consistent with observations of the gradual increase in δ13C of photosynthates from 0600 to 1500 h modelled from gas exchange and the gradual decrease in δ13Cl over the night-time period. Although the instantaneous modelled δ13C value of photosynthates declined at 1800 h (Table 2), its effect on the δ13C value of starch is likely to be negligible considering the low assimilation rates at that time point (Fig. 1e).

Although we demonstrate that drought can alter the diurnal amplitude of δ13Cl potentially through its impact on ΔP and resultant δ13C values of respiratory substrates and transitory starch pools, other processes impacted by drought stress may also affect short-term variation in δ13Cl. For example, the pool size of malate fuelling respiration in light-acclimated darkened leaves (i.e. LEDR), and the partitioning of intermediate respiratory metabolites between biosynthesis and adenosine triphosphate (ATP) production in mitochondria are also likely impacted by drought stress. Positional 13C labelling experiments suggest that variation in the pyruvate dehydrogenase (PDH) to acetyl coenzyme A (acetyl-CoA) oxidation ratio governs the magnitude of the daytime increase in δ13Cl (Priault et al. 2009). We hypothesize that reductions in the ratio of PDH:acetyl-CoA oxidation leading to decreases in δ13Cl are also likely to occur over the night-time period. The ratio of PDH:acetyl-CoA oxidation is sensitive to substrate availability (Priault et al. 2009), which is likely to be affected by drought. Reduced substrate availability during drought is therefore likely to dampen the diurnal variation in δ13Cl by decreasing the ratio of PDH:acetyl-CoA oxidation.

A gradual increase in the malate pool size during the day could partially contribute to the observed daytime increases in δ13Cl (Gessler et al. 2009). We do not know the degree to which the decarboxylation of malate associated with LEDR influenced the daytime increase of δ13Cl in P. velutina. Drought-stressed plants are likely to have a reduced and less variable pool size of malate compared to non-stressed plants, which would limit daytime increases in δ13Cl. However, the night-time decrease of δ13Cl observed in P. velutina was unlikely a result of changes in malate decarboxylation associated with LEDR.

Finally, drought may affect diurnal variation in δ13Cl through its impact on 13C/12C fractionations associated with respiratory enzymatic reactions (Deniro & Epstein 1977; Tcherkez & Farquhar 2005). In general, the range of diurnal variation in isotope effects associated with respiratory enzymatic reactions is determined by substrate availability and the magnitude of diurnal shifts in leaf temperature. Large diurnal shifts in leaf temperature during the dry pre-monsoon season (Fig. 1a) are expected to produce greater diurnal changes in the isotope effects of respiratory enzymes. However, we observed less diurnal variation in δ13Cl during the dry pre-monsoon season than during the wet monsoon season, suggesting that temperature-dependent isotope effects associated with respiratory enzymes did not play a dominant role in controlling variation in δ13Cl.

Diurnal and seasonal changes in the δ13C value of respired CO2 from autotrophic and heterotrophic tissues can provide unique insight into metabolic activities in plants and ecosystems. The observed large diurnal amplitude of δ13Cl in P. velutina and in other trees and shrubs (Hymus et al. 2005; Prater et al. 2006; Priault et al. 2009) is apparently common for woody C3 species. Resolving mechanisms responsible for short-term changes in δ13Cl would assist in the interpretation of carbon isotope fluxes between ecosystems and atmosphere and provide further insight into plant metabolic responses to environmental change.

ACKNOWLEDGMENTS

This research work was supported by NSF (EPS-0447081) and NSF (DEB-0414680). We acknowledge S. Göttlicher for offering instructions for leaf-soluble carbohydrates and starch extraction. S. Sharma and M. Larson helped with carbon isotope analysis. R. Scott, T. Huxman, G. Barron-Gafford and N. Pierce supported the field work.

Ancillary