soil carbon stocks; tropical montane cloud forest; puna; tree line

Global Change Biology (1999) 5, 157168

Elevated CO2 and temperature impacts on differentcomponents of soil CO2 efflux in Douglas-fir terracosms

G U A N G H U I L I N , * J A M E S R . E H L E R I N G E R , * PA U L T . R Y G I E W I C Z , M A R K G . J O H N S O N and D AV I D T . T I N G E Y *Stable Isotope Ratio Facility for Environmental Research, Department of Biology, University of Utah, Salt Lake City,UT 84112, USA, Present address: Biosphere 2 Center/Columbia University, PO Box 689, Oracle, AZ 85623, USA andLamont-Doherty Earth Observatory of Columbia University, Palisades, NY 10964, USA, U.S. Environmental ProtectionAgency, National Health and Environmental Effects Research Laboratory, Corvallis, OR 97333, USA

Abstract

Although numerous studies indicate that increasing atmospheric CO2 or temperaturestimulate soil CO2 efflux, few data are available on the responses of three majorcomponents of soil respiration [i.e. rhizosphere respiration (root and root exudates), litterdecomposition, and oxidation of soil organic matter] to different CO2 and temperatureconditions. In this study, we applied a dual stable isotope approach to investigate theimpact of elevated CO2 and elevated temperature on these components of soil CO2efflux in Douglas-fir terracosms. We measured both soil CO2 efflux rates and the 13Cand 18O isotopic compositions of soil CO2 efflux in 12 sun-lit and environmentallycontrolled terracosms with 4-year-old Douglas fir seedlings and reconstructed forestsoils under two CO2 concentrations (ambient and 200 ppmv above ambient) and twoair temperature regimes (ambient and 4 C above ambient). The stable isotope data wereused to estimate the relative contributions of different components to the overall soilCO2 efflux. In most cases, litter decomposition was the dominant component of soilCO2 efflux in this system, followed by rhizosphere respiration and soil organic matteroxidation. Both elevated atmospheric CO2 concentration and elevated temperaturestimulated rhizosphere respiration and litter decomposition. The oxidation of soilorganic matter was stimulated only by increasing temperature. Release of newly fixedcarbon as root respiration was the most responsive to elevated CO2, while soil organicmatter decomposition was most responsive to increasing temperature. Although someassumptions associated with this new method need to be further validated, applicationof this dual-isotope approach can provide new insights into the responses of soil carbondynamics in forest ecosystems to future climate changes.

Keywords: elevated CO2, forest ecosystem, global warming, soil respiration, stable isotopes

Received 10 October 1997; revised version received 26 January and accepted 17 February 1998

Introduction

Soils are the major reservoir of carbon in terrestrialecosystems, containing more than two-thirds of totalcarbon in the terrestrial part of the biosphere. A majorunknown in the response to anticipated climate changesis the extent to which forest ecosystems will become netsinks or sources of CO2. This uncertainty is in part drivenby our lack of knowledge on how much root respirationwill increase under elevated atmospheric CO2 and inpart on how elevated temperatures might accelerate

Correspondence: Guanghui Lin, fax 11/520-896-6214,e-mail [email protected]

1999 Blackwell Science Ltd. 157

the turnover of soil organic carbon. Understanding soilcarbon dynamics under elevated atmospheric CO2 andtemperature is thus critical for predicting future regionaland global carbon budgets (Schimel 1995).

Previous studies have suggested that increasing atmo-spheric [CO2] and temperature can stimulate soil CO2efflux (e.g. van Veen et al. 1991; Korner & Arnone 1992;Peterjohn et al. 1993, 1994; Johnson et al. 1994; Nakayamaet al. 1994; Pajari 1995; Vose et al. 1995; Hungate et al.1997). However, there is relatively little information onwhich components of the soil CO2 efflux are most sensit-ive to changes in atmospheric CO2 (see review by Paterson

158 G U A N G H U I L I N et al.

et al. 1997) or temperature. The components of soil CO2efflux can be partitioned into root respiration, microbialdecomposition of soil organic matter, and microbialdecomposition of surface-layer litter. Traditionally, radio-active materials have been used to partition componentsof the CO2 efflux, including continuous 14C-labelling (e.g.Whipps & Lynch 1983; Merckx et al. 1985) and 14C pulse-labelling (e.g. Cheng et al. 1996). These methods haveenvironmental health restrictions and thus are limited toshort-term experiments, lasting perhaps several monthsat most (Cheng et al. 1996). In ecosystems where thephotosynthetic pathway of the current vegetation (C3 orC4) is distinct from the vegetation responsible for thebulk of the soil organic matter accumulation, stableisotopes of carbon have been used to partition soil CO2efflux into old vs. recently formed soil carbon components(e.g. Schonwitz et al. 1986; Wedin et al. 1995; Cheng1996). However, such transition ecosystems are limitedin distribution. A recent variation on this approach hasbeen to add a C4 sugar substrate to the C3-dominatedsoil to quantify the microbial respiration component ofthe CO2 efflux (Hogberg & Ekblad 1996). Leavitt et al.(1996) and Nitschelm et al. (1997) demonstrated thatthe CO2 source in Free Air CO2 Enrichment (FACE)experiments was usually sufficiently different from atmo-spheric CO2 that this could also be used as a label topartition CO2 efflux from old vs. recently formed soilcarbon components. However, the amount of carbonadded to the soil carbon pool in a 1-or 2-year period wastoo small, restricting the utility of this approach.

We show here that by measuring changes in the isotopiccomposition of the soil CO2 efflux instead of changes inthe soil carbon pool in elevated CO2 studies, larger andmore reliable estimates of the soil carbon dynamics maybe obtained. Our approach is to use a combination ofanalyses of the stable isotope ratios of carbon and oxygenin CO2 efflux with soil water at different depths topartition soil CO2 efflux into three distinct components:rhizosphere respiration (including root respiration andmicrobial respiration resulting from consumption of rootexudates), microbial decomposition of surface litter, andmicrobial decomposition of soil organic matter (SOM).This approach overcomes potential concerns about theunnatural mixing of soils and plants, which would other-wise be a significant limitation for the study of naturalresponses by ecosystems. We apply this approach toanalyse the soil CO2 efflux responses in Douglas firterracosms with tree seedlings growing under elevatedCO2 and air temperature treatments.

Materials and methods

Elevated CO2 and temperature treatments

At the U.S. Environmental Protection Agencys TerrestrialEcophysiology Research Area (TERA), a study of eco-

1999 Blackwell Science Ltd., Global Change Biology, 5, 157168

Fig. 1 (a) Seasonal change in CO2 concentration in the ambient(n 5 6) and elevated CO2 treatment terracosms (n 5 6) and (b)seasonal change in air temperature in the ambient (n 5 6) andelevated temperature treatment terracosms (n 5 6) with Douglas-fir seedlings and reconstructed forest soils.

system response to elevated CO2 and temperature wasinitiated in July 1993, using 12 sun-lit controlled-environ-ment terracosm (see Tingey et al. 1996 for a detaileddescription of this facility). Each terracosm had a canopyvolume of 3.18 m3 (2 m wide, 1.5 m tall at the back, 1.2 mtall at the front, 1 m front-to-back) and a soil lysimetervolume of 2.0 m3 (2 m wide, 1 m front-to-back, 1 mdeep). The treatments imposed were: (i) ambient CO2and ambient temperature, (ii) elevated CO2 (ambient plus200 ppmv) and ambient temperature, (iii) ambient CO2and elevated temperature (ambient plus 40 C), and (iv)elevated CO2 and elevated temperature (Fig. 1). The soilwas a fine loam to loam texture with a medium granularstructure and was classified as a coarse-loamy, mixed,frigid Typic Hapludand. At the beginning of the experi-ment, soil carbon content averaged at 2.38% for A Horizonand 2.07% for B Horizon. Litter from an old-growthDouglas-fir forest was collected and added to the terra-cosm when the trees were planted (Tingey et al. 1996).Seedlings in each terracosm were irrigated with localwell water having the same isotopic compositions tomaintain similar soil moisture. However, the litter layertended to be drier in the teracosms than in naturalDouglas fir forest due to the control capability. As aresult, there were few roots growing in this layer.

The terracosms were maintained as closed systems

R E S P O N S E O F S O I L C O 2 E F F L U X T O C L I M A T E C H A N G E 159

Table 1 Carbon isotope ratios of the CO2 from the source tankand of the atmosphere outside but near the terracosms. Valuesare the mean of 4 replicates and standard error of the mean

13C ()

Sampling Month Tank CO2 Atmospheric CO2

April 1994 35.64 6 0.07 8.83 6 0.09June 1994 35.84 6 0.04 8.54 6 0.36August 1994 35.77 6 0.12 8.24 6 0.27October 1994 35.81 6 0.23 8.63 6 0.12Mean 35.77 6 0.05 8.55 6 0.13

most of time during the study (opened only wheneverneeded for maintenance, sampling or physiological meas-urements). Commercial tank CO2 and carbon dioxidescrubbers were used to maintain the desired CO2 concen-trations both in the ambient and elevated CO2 treatmentterracosms (Tingey et al. 1996). During day time, tankCO2 was added to compensate the photosynthetic uptakeand to maintain a specific CO2 concentration, while atnights the excessive CO2 from soil and plant respirationwas removed using Soda Lime scrubbers. AtmosphericCO2 concentration in each terracosm was monitoredcontinuously using a LI-6262 CO2/H2O analyser (LI-COR Inc., Lincoln, Nebraska, USA) which was calibratedregularly with CO2 standards. Efforts were made to usetank CO2 with relatively low and constant 13C contentthroughout the entire study (Table 1). As a consequence,new photosynthate produced during the experimentalperiod and allocated to the leaves and roots had adifferent 13C content than previously grown material.Repeated measurement of the outside atmospheric andtank CO2 confirmed that source atmospheric CO2 valuesremained constant throughout the growing season(Table 1). The mean carbon isotope ratio of the tank CO2was 35.77 6 0.09, compared with 8.55 6 0.13 forthe atmospheric CO2 at the TERA area. The temperaturein each terracosm was controlled at specific levels throughthe use of heat exchangers.

Soil CO2 efflux measurements

During four sampling periods in 1994 (1622 April, 2022 June, 1721 August and 1820 October), soil CO2 effluxin each terracosm was measured at two locations using aLI-600009 soil respiration chamber (LI-COR Inc., Lincoln,NB, USA) and PVC collars (10 cm diameter 3 7 cmheight). For each sampling location, a PVC collar wasinstalled in March 1994 to a depth of 5 cm in the soiland left undisturbed throughout the entire study. Thevolume of free space above the litter surface containedwithin each soil collar was measured prior to each soilCO2 efflux measurement. During each measurement,


change in the CO2 concentration in the closed chamberover 12 min was measured using the CO2 analyser of aLI-6200 portable photosynthesis system (LI-COR Inc.,Lincoln, NE, USA) and used to calculate soil CO2 effluxrate. In addition, the soil temperature at 5-cm depth wasmeasured with a thermocouple sensor attached to the LI-600009 soil respiration chamber.

Isotope sampling and analyses of soil CO2 efflux andorganic matter

After a soil CO2 efflux rate had been measured, CO2emitted from the soil was collected by inserting a watertrap and a 2-L air flask (which had been initially backfilled with dry N2) in series in a closed-loop with the LI-6200 system and LI-600009 soil respiration chamber.The terracosm atmospheric CO2 captured in the soilrespiration chamber was removed using Soda Lime scrub-ber as a by-path in the collection system. After filling theflask with effluxed CO2, the flask was moved to a portableextraction line and the CO2 was extracted cryogenically(Buchmann et al. 1997). Back at the University of Utah,N2O, which had been frozen out along with CO2 in thefield vacuum line, was removed using a gas chromato-graph 3-m Poroplot Q column before CO2 samples wereanalysed on an isotope ratio mass spectrometer (Delta S,Finnigan MAT, Breman, Germany) operated in the dualinlet mode (Ehleringer 1991). The precision and reliabilityof this sampling scheme was tested using two cham-berless controls (outdoor plots containing the same plants,soils, and litter) and three soil lysimeters containing thesame soils but without litter and without plants.

For each terracosm, a pooled sample of newly producedneedles from six seedlings was collected and frozenimmediately on the same day that soil efflux CO2 sampleswere collected. A pooled sample of litter from fourpositions surrounding the soil collar in each terracosmwas collected and sealed in 10-mL screw-cap vials andsecured with parafilm. Similarly, a pooled sample of topsoil from the A Horizon (05 cm below the bottom of thelitter layer) was collected from four positions around thesoil collars in each terracosm. The needle samples werefirst dried at 70 C for 48 h and then ground to passthrough a 20-mesh sieve. For the litter and soil samples,water was cryogenically extracted in a vacuum line.Following this, the dried litter and soil samples wereindividually ground to pass through a 20-mesh sieve forcarbon isotope ratio analyses. Soil samples were firstcleaned of obvious plant fractions and then treated with1N HCl to remove any carbonate.

The carbon isotope ratio of organic materials (needles,litter, roots, soil) was determined using an elementalanalyser in conjunction with the Delta S isotope ratiomass spectrometer. Results are expressed in delta notation


(13C) in (parts per thousand) relative to the Pee DeeBelemnite (PDB) standard:

() 5 (Rsam/Rstd 1) * 1000, (1)

where Rsam and Rstd are 13C/12C for the sample andstandard, respectively. The external precisionwas 6 0.11 based on repeated measurements of a labworking standard (Utah cabbage).

The oxygen isotope ratios (18O) of litter and soil waterwere determined with the CO2-H2O equilibrium methodmodified by Socki et al. (1992). All oxygen isotope ratios(soil CO2 efflux and water samples) are expressed sim-ilarly in units relative to the SMOW standard. Externalprecision for 18O measurements was 6 0.23, basedon repeated measurement of a lab working standard (SaltLake City ground water).

Partitioning of the soil CO2 efflux

Partitioning of the soil CO2 efflux into components wasmade using a 2-endmember linear model for interpretingthe 18O value of CO2 efflux and a 3-endmember triangu-lar model for the 13C value of the efflux. Before parti-tioning the soil CO2 efflux into three components,calculations were made of the expected carbon andoxygen isotope ratios of different carbon sources withinthe soil profile. The 13C value of CO2 released fromrhizosphere respiration was expected to be the same asthat of carbon in newly grown roots and root exudates,since there is no carbon isotopic fractionation duringheterotrophic respiration (Lin & Ehleringer 1997). How-ever, it was impractical to sample representative newlygrown roots and root exudates in the terracosms forcarbon isotopic analysis each time, so we measured thecarbon isotope ratios of the newly produced needles asan approximate estimation (they should come from thesame carbon source, i.e. the newly synthesized photo-synthate). In April and October 1994, we analysed the13C values for the fine roots (, 1 mm) from soil coringand found that they were , 0.3 more positive thanthose for the newly produced leaves in the ambient CO2chambers and , 1 more positive in the elevated CO2chambers (Lin et al. unpubl. data).

As mentioned earlier, tank CO2 with much lower 13Cvalues than atmospheric CO2 was used to maintain CO2concentration in all terracosms, so the new carbon in thenewly grown leaves and newly produced roots as wellas root exudates should have had much lower 13C valuesthan the old carbon in the litter and the soil organicmatter that started the terracosms. There is little carbonisotopic fractionation during the early decay of fallenplant materials (Balesdent et al. 1993), so we can assumethat the CO2 released from litter decomposition had asimilar carbon isotope ratio to that of litter. Because we


conducted this study in the second year of the treatments,there was little new litter formation and decompositionin our terracosms. Thus the carbon isotope ratio of CO2from litter decomposition should have been similar tothat of the bulk litter in each terracosm.

Soil organic carbon consists of several fractions withdifferent densities and turn-over times, and usuallyorganic matter with a lower density decomposes fasterthan that with higher density. Thus, the 13C of CO2 fromSOM oxidation will depend mainly on the carbon isotoperatio of low-density carbon in the soil. Bird et al. (1996)observed that the 13C value varied among different soilsize fractions, but the maximum difference for forest soilswas less than 0.5 in most cases. Variation in SOM13C can be attributed to differential discrimination ofbiochemical pathways in plants (lignin is relativelydepleted in 13C), but also substrate-dependent discrimina-tion during microbial mineralization of organic matter(Mary et al. 1992). Since there was no direct method formeasuring carbon isotope ratios of the CO2 from SOMdecomposition, we thus assumed that the 13C of CO2from SOM decomposition was similar to that of bulk soilorganic carbon (SOC) in the surface layer. We tested thisassumption by comparing the 13C value of SOC in thesurface layer with that of soil effluxed CO2 in the threesoil lysimeters (the same soils but without seedlings andlitter) and found that the difference between the two was0.20.3 at most.

We assumed that the 18O value of CO2 released fromdecomposition of litter was in equilibrium with the litterwater, because of both the high surface-to-volume ratioin litter and the ubiquity of carbonic anhydrase in soilmicrobial organisms. We assumed that CO2 efflux origin-ating from decomposition of litter was not subject to asignificant oxygen diffusion fractionation, since CO2 fromlitter decomposition would be turbulently transferredfrom this surface layer to the atmosphere. Thus, the 18Ovalue of CO2 from litter decomposition was calculatedfrom the 18O value of the litter water according to themodel of Bootinga & Craig (1969):

18OCO2 5 18Owater 1 ( 1) * 1000, (2)

where

5 (5.1120.214 t 1 0.00041t2 1 1000) * 0.00104075 (3)

and t is the water temperature in C.We assumed that CO2 evolved from rhizosphere

respiration and SOM decomposition (no matter at whatlayers they were produced) will reach isotopic equilib-rium with soil water in the top 05 cm layer (Ciais et al.1997; Tans 1998). Before escaping into the atmosphere,the CO2 from these two processes should have the 18Ovalues that can be estimated from the 18O of soil waterin the top layer using (2), (3). In addition, an 8.8


diffusion fractionation against the heavier CO2(13C18O16O) occurs as these CO2 efflux components dif-fused through to the uppermost soil layer to the litterlayer (Hesterberg & Siegenthaler 1991). Thus, CO2 effluxderived from either rhizosphere respiration or SOMdecomposition should be 8.8 more negative in 18Ovalue than that expected from the 18O of soil water inthe top layer.

Given these assumptions, we could calculate the relat-ive contributions of rhizosphere respiration, litter decom-position and SOM oxidation to the overall soil CO2 effluxrate. This could be algebraically partitioned and solvedby calculating the relative contribution factors [m for rootcarbon, n for litter carbon, and (1mn) for SOM carbon]from the measured isotopic compositions of the overallsoil CO2 effluxed from the soil surface as

13CR-soil 5 m*13CR-root 1 n*13CR-litter 1(1-m-n)*13CR-SOM (4)

18OR soil 5 n*18OR litter 1 (1 n)*18OR topsoil, (5)

where the subscripts for 13CR indicate the carbon sourceof CO2, i.e. R-soil for total soil CO2 efflux, R-root for root-derived CO2, R-litter for litter-derived CO2, and R-SOMfor SOM-derived CO2, and subscripts for 18OR indicatethe oxygen origin of soil-respired CO2, i.e. R-litter for litterlayer water and R-topsoil for the upper soil layer water.

Statistical analyses

The effects of elevated CO2 and temperature on soil CO2efflux rates over the four sampling dates were tested bya two-way ANOVA. The seasonal changes in soil CO2efflux rate, isotopic composition of soil-respired CO2,plant tissues and soils were tested with one-way ANOVA.The differences in the isotopic composition of effluxedsoil CO2, new needles, litter, soil organic matter, litterwater, and soil water among treatments were tested usingthe Tukey t-test. All statistical analyses were performedusing a PC SYSTAT 7.0 (SPSS Inc., Chicago, Illinois).

Results

Soil temperature and soil CO2 efflux rate

The soil temperature at the top layer of the mineral soilshowed a strong seasonal pattern, increasing from Aprilto June and then decreasing after August (all P , 0.001,Fig. 2a). There were no significant differences in soiltemperature between the ambient and elevated CO2treatments at either temperature treatment. However,there were consistent differences in soil temperaturebetween ambient and elevated temperature treatments(both P , 0.001). The absolute difference (2.12.6 C) in


Fig. 2 (a) Soil temperature at 5-cm depth and (b) soil CO2efflux rate in the terracosms under different CO2 concentrationand air temperature treatments at four sampling dates in 1994.Values are the mean and standard error of three replicates foreach treatment.

soil temperature between the two temperature treatmentswas slightly lower than the difference in air temperaturebetween treatments (3.7 C) (see also Tingey et al. 1996).

At ambient CO2 and ambient temperature, soil CO2efflux rates ranged from 2.4 to 4.4 mol m2 s1 betweenApril and October (Fig. 2b). There was a strong seasonaltrend in soil CO2 efflux (all P , 0.001). Relative to theambient CO2 and ambient temperature treatment, soilCO2 efflux was significantly increased by either elevatedCO2 or by elevated temperature (Fig. 2b), averaging 15%higher under elevated CO2 (130%, P , 0.05) and 50.3%higher under elevated temperature (1661%, P , 0.001).Under the combination treatment of elevated CO2 andelevated temperature, the response was 72.6% higher(54149%, P , 0.001).

Isotopic composition of soil CO2 efflux

Under both temperature treatments, the 13C of soilCO2 efflux was significantly lower for the elevated CO2treatment than for the ambient CO2 treatment (both


Fig. 3 Carbon and oxygen isotope ratios of soil CO2 effluxfrom the terracosms under different CO2 concentration andtemperature treatments at the four sampling dates in 1994.Values are the mean and standard error of three replicates foreach treatment.

P , 0.001, Fig. 3a). There was no significant differencein the seasonal mean 13C value between the temperaturetreatments at either CO2 treatment, although the 13Cdiffered significantly between these treatments in bothJune and October (both P , 0.001).

There was a strong seasonal trend in the 18O value ofthe soil CO2 efflux in all treatments (all P , 0.001), withan increase between April and June, and then a decreasefrom June to October (Fig. 3b). There was no significantdifference among the mean seasonal 18O values of thetreatments. However, the 18O values of any treatmentwith an elevated condition (i.e, elevated CO2 and ambienttemperature, ambient CO2 and elevated temperature,elevated CO2 and elevated temperature) were lower thanfor the control (ambient CO2 and ambient temperature)in June, August and October.


Isotopic composition of needles, litter, soil and water

The 13C values of newly grown needles ranged from 28.4 to 29.4 for the two ambient CO2 treatmentsand 34.1 to 35.7 for needles in the elevated CO2treatments (Fig. 4a). The differences between the twoCO2 treatments were highly significant (both P , 0.001).However, there was no significant difference in 13Cvalue for the newly grown needles between the twotemperature treatments under either ambient CO2 orelevated CO2. The 13C values for the litter and the SOMwere not significantly different among the four treatments(Fig. 4b,c). The 13C values for SOM were significantlyhigher than the values for the litter, which was alsosignificantly higher than that for the newly grown needlesin all treatments (all P , 0.001).

There was no significant difference in the 18O valuefor the litter water among treatments (Fig. 5a). However,there was a strong seasonal trend in the 18O value forthe litter water in all treatments (all P , 0.001), with ageneral increase from April to August, and then a decreasefrom August to October (Fig. 5a). The 18O value for soilwater in the top of the A Horizon was not significantlydifferent among the four treatments, and was relativelyconstant throughout the four sampling dates (Fig. 5b).The 18O value for soil water was also significantly lowerthan the value for litter water in all treatments and at allsampling times (all P , 0.001).

Relative contributions of rhizosphere respiration, litterand SOM decomposition

The CO2 originating from rhizosphere respiration, litterdecomposition, and SOM oxidation had distinct carbonand oxygen isotope ratios, and the total soil CO2 effluxhad the isotope ratios within the boundaries describedby the CO2 released from these three sources (Fig. 6).Using (4) and (5), we estimated the relative contributionof these carbon sources to the overall soil CO2 efflux forall four CO2 and temperature treatments.

In all treatments, litter carbon was the dominant carbonsource (mean of 6064%) of the soil CO2 efflux, followedby root carbon (2332%), and then SOM carbon (818%)(Table 2). However, the relative contribution of thesecarbon sources varied significantly among sampling dates(P , 0.001 for root carbon, P , 0.05 for litter carbonand P , 0.01 for SOM carbon). At ambient CO2, thetemperature treatment had no significant effect on therelative contributions of root, litter and SOM carbon.Elevated CO2 treatment at ambient temperature increasedthe contribution of root carbon in most cases, but hadlittle effect on the relative contributions of litter and SOMcarbon. Elevated CO2 and temperature together increasedthe contribution of root carbon, but had no effect on therelative contribution of either litter or SOM carbon.


Fig. 4 Carbon isotope ratios of the (a) newly grown needles, (b) litter, and (c) soil organic matter in the terracosms under differentCO2 concentration and air temperature treatments at four sampling dates in 1994. Values are the mean and standard error of threereplicates for each treatment.

Fig. 5 Oxygen isotope ratios of the waterin the (a) litter layer and (b) the top ofthe A horizon in the terracosms underdifferent CO2 concentration and airtemperature treatments at four samplingdates in 1994. Values are the mean andstandard error of three replicates for eachtreatment.

On average, elevated CO2 treatment increased rhizo-sphere respiration by 79% and litter decomposition by18%, but reduced oxidation of SOM by 14% (Table 3and Fig. 7). Elevated temperature under ambient CO2increased all three components of soil CO2 efflux, by60%, 44% and 189% for rhizosphere respiration, litterdecomposition and SOM oxidation, respectively. Theelevated CO2 and elevated temperature treatmentincreased rhizosphere respiration by 143%, litter decom-position by 69%, and SOM oxidation by 93% comparedwith the ambient-ambient treatment.

Discussion

Effect of elevated CO2 and temperature on overall soilrespiration

We observed substantial increases in the total soil CO2efflux by both elevated CO2 and elevated temperaturetreatments. The two treatments appeared to be additive,since the increase in soil CO2 efflux by the elevated CO2and elevated temperature treatment was similar to thesum of the two treatments. In this respect, our resultsare similar to those of many previous studies (e.g. van


Veen et al. 1991; Korner & Arnone 1992; Johnson et al.1994; Nakayama et al. 1994; Pajari 1995; Vose et al. 1995),although there are other studies showing no significanteffect of elevated CO2 on soil respiration (e.g. Oberbaueret al. 1986). As pointed out by Nakayama et al. (1994)and others, environmental conditions among the studieswere quite different, so it is difficult to generalize.

Isotopic partitioning of soil respiration components

The isotope ratio of soil CO2 efflux is influenced by boththe carbon sources and by the water in the soil and litterlayers. In the terracosms, the 13C of the soil carbonsources should reflect contributions from each of thethree primary sources: litter, roots, and SOM. The tankCO2 used for regulating atmospheric CO2 concentrationin all terracosms provided a much lower 13C value thantypical of atmospheric CO2, resulting in newly producedplant tissues (new needles, new roots) with distinctlymore negative 13C values than the older litter and SOMcarbon in the terracosms, which had been derived fromfield conditions. Additionally, litter in the terracosms hada lower carbon isotope ratio than SOM, which is typical(Nadelhoffer & Fry 1988; Buchmann et al. 1997). The


Fig. 6 The carbon and oxygen isotoperatios of total soil CO2 efflux (closedsymbols) and its three major carbonsources (open symbols) in the terracosmsunder different CO2 and temperaturetreatments.

Fig. 7 Seasonal dynamics of the relativecontributions of newly fixed carbon andpreviously formed litter and SOM carbonto total soil CO2 efflux in terracosmsunder different CO2 and temperaturetreatments.

combination of these three different possible 13C sourcesmeant that using only 13C analyses (eqn 4) would notallow us to partition the relative contributions of thenew and old carbon to the overall soil CO2 efflux.

The water in the litter layer had a substantially higher


18O value than the water in the top of the A horizon(Fig. 5), which resulted in distinct differences in 18Ovalues between the CO2 derived from the litter layer(via litter decomposition) and that from the soil layer(rhizosphere respiration and SOM oxidation) (Fig. 6).


Table 2 Relative contributions (%) of root carbon (RC), litter carbon (LC) and SOM carbon (SC) to the overall soil CO2 efflux in theterracosms under treatment combinations of ambient temperature, elevated temperature, ambient CO2, and elevated CO2 (mean 6 SE,n 5 3)

Ambient CO2 Elevated CO2

Sampling time Carbon source Ambient T Elevated T Ambient T Elevated T

April-94 RC 28 6 1 15 6 6 29 6 5 27 6 7LC 49 6 5 52 6 2 55 6 4 58 6 1SC 23 6 5 33 6 4 16 6 4 15 6 2

June-94 RC 12 6 1 22 6 5 22 6 4 33 6 2LC 73 6 1 67 6 2 72 6 4 64 6 1SC 15 6 2 11 6 4 6 6 1 3 6 2

August-94 RC 25 6 1 23 6 3 37 6 3 27 6 2LC 69 6 1 58 6 3 58 6 3 62 6 4SC 6 6 1 19 6 1 5 6 2 11 6 2

October-94 RC 30 6 2 31 6 1 40 6 1 31 6 4LC 64 6 1 62 6 1 54 6 1 59 6 3SC 6 6 2 7 6 1 6 6 2 10 6 4

Average RC 24 6 2 23 6 3 32 6 4 30 6 2LC 64 6 6 60 6 4 60 6 4 62 6 1SC 13 6 4 18 6 6 8 6 3 10 6 3

Table 3 Release rates of root, litter and SOM carbon in Douglas fir terracosms under ambient CO2 and temperature conditions (basevalue) and the percentage of increase (positive) or decrease (negative) by elevated CO2 and temperature treatments (Mean 6 SE, n 5 3)

% change byComponents of Sampling Base valuesoil respiration time (mol m2s1) Elevated CO2 Elevated T Elevated CO2 & T

Rhizosphere Apr-94 0.75 6 0.12 22 6 5 19 6 8 47 6 14respiration June-94 0.48 6 0.10 98 6 30 135 6 5 319 6 24

Aug-94 0.91 6 0.11 94 6 15 20 6 18 62 6 9Oct-94 0.70 6 0.05 101 6 5 106 6 3 142 6 29Average 0.71 6 0.09 79 6 19 60 6 36 143 6 62

Litter Apr-94 1.29 6 0.22 37 6 12 46 6 8 87 6 23decomposition June-94 3.07 6 0.15 5 6 9 4 6 6 28 6 11

Aug-94 2.47 6 0.19 11 6 8 25 6 12 43 6 11Oct-94 1.50 6 0.13 29 6 9 100 6 11 119 6 23Average 2.08 6 0.42 18 6 9 44 6 21 69 6 20

SOM oxidation Apr-94 0.63 6 0.19 17 6 17 88 6 23 4 6 9June-94 0.62 6 0.03 68 6 5 18 6 20 64 6 21Aug-94 0.23 6 0.06 3 6 2 563 6 37 180 6 25Oct-94 0.16 6 0.06 31 6 5 121 6 34 258 6 54Average 0.41 6 0.12 14 6 20 189 6 128 93 6 75

Using the two isotopic mass balance equations (eqns 4and 5) simultaneously, we could partition the threecomponents of soil CO2 efflux. For all treatments andsampling dates, the largest contributor to total soil CO2efflux was litter decomposition, followed by the contribu-tions of rhizosphere respiration and SOM oxidation. Thisdual-isotope approach also allowed us to examine theresponses of these three components of soil CO2 effluxto the CO2 and temperature treatments (see below), andalso should be applicable to field situations wherever


there are substantial isotopic variations in both carbonand oxygen isotope ratios for soil CO2 efflux.

Responses of soil CO2 efflux components to elevatedCO2 and temperature

Release of newly formed carbon (via rhizosphere respira-tion) responded the most to the combined elevated CO2and temperature treatment. The elevated CO2 treatmentalone significantly enhanced the release of newly fixed


carbon at all sampling dates. Elevated temperaturereduced the relative proportion of root carbon in Apriland August, but enhanced its release in June and October.Elevated CO2 and elevated temperature together stimu-lated release of newly formed carbon to a greater extentthroughout the year. It has been demonstrated previouslythat elevated CO2 increases carbon allocation to fine rootsand increases root exudation, and thus enhances soil CO2efflux (e.g. Norby et al. 1992; Johnson et al. 1994; Rogerset al. 1994).

Litter decomposition responded significantly to alltreatments involving elevated ambient CO2 or elevatedtemperature. We are unsure why litter decompositionwas so responsive to elevated CO2. Previous studies haveshown that elevated CO2 changes the quality of newlitter, probably affecting litter decomposition (Couteauxet al. 1991; van de Geijin & van Veen 1993). However,there was little new litter formation from the new needlesgrown under the CO2 and temperature treatments, sinceour study was conducted in the second year of thetreatments. Perhaps elevated CO2 increased plant carbonfixation and allocation to roots, enhancing root exudationand turnover processes, as has been suggested by others(Berntson & Woodward 1992; Norby et al. 1992; Rogerset al. 1994). The increased carbon likely stimulated micro-bial communities and nutrient cycling processes, therebyincreasing litter decomposition to meet their nitrogenrequirements (van Veen et al. 1991). This usually occurswhen inorganic nitrogen availability is low as is thesituation in this Douglas-fir system. Soil nitrogen waslow (, 0.1%), and NO3 or NH41 was never detected insoil solutions (detection limits for NO3 and NH41 were0.04 mg L1 and 0.10 mg L1, respectively).

Although SOM carbon contributed a relatively smallproportion to overall soil-respired CO2 (Table 2), releaseof this previously formed carbon was also increased byelevated temperature in most cases (Table 3, Fig. 7). Werecognize that the large relative increase in oxidation ofSOM may be an artifact stemming from the small basalvalues for the ambient CO2 and ambient temperaturetreatment. However, SOM is a large reservoir of globalcarbon, recently estimated at about 1600 Pg, which ismore than twice the atmospheric CO2-C pool (Schimel1995). Small changes in the size of the soil organiccarbon pool could significantly affect atmospheric CO2concentrations. The annual flux of CO2 from soils to theatmosphere is estimated at 76.5 Pg C y1, which is 3060% greater than terrestrial net primary productivity(Raich & Schlesinger 1992). If soil organic carbon isreduced under elevated atmospheric CO2 or elevatedtemperature conditions, then soils represent a significantcarbon source, increasing the amount of the carbon tothe atmosphere (Jenkinson et al. 1991). On the other hand,elevated CO2 usually stimulates photosynthetic carbon


uptake, so the overall effects of increasing atmosphericCO2 on carbon balance in the forest ecosystems of north-western USA will be determined by the responses inboth above-and below-ground processes.

Possible errors associated with assumptions

In this study, we had to make some assumptions for thecalculations of isotopic signals for different CO2 sources(see Materials and methods section). First, we assumedthat the CO2 from rhizosphere respiration has similar13C value to that of newly grown needles. If the 13C ofthe newly grown needles was more negative than thatof the active roots and root exudates, we would haveunderestimated the contribution of rhizosphere respira-tion to the overall soil CO2 efflux (see Fig. 6), especiallyin the ambient CO2 chambers where the differencebetween the new and old carbons were small (25).Our spot measurements indicate that the mean differencein the 13C between the newly grown needles and thefine roots (, 1 mm) was , 0.3 for the ambient CO2chambers and , 1.0 for the elevated CO2 chambers(Lin et al. unpubl. data). Thus, the possible errors associ-ated with this assumption may cause about 515% (0.3vs. 25 for the ambient CO2 chambers and 0.9 vs. 612 for the elevated CO2 chambers) deviation from theestimated values for the relative contributions, but willnot change the general patterns discussed in the abovesections.

A second assumption is that the 13C of CO2 fromSOM matter is similar to that of bulk SOC. As mentionedearlier, bulk SOC is made of several fractions which maydecompose at different rates and have different isotopiccompositions (Bird et al. 1997). Therefore, the actual 13Cof CO2 from SOM oxidation will be different from thatof bulk SOC. If the carbon contributed to the soil CO2efflux is lighter (more negative in 13C) than bulk SOC,we may then underestimate the relative contributionfrom SOM oxidation to the overall soil CO2 efflux (seeFig. 6), especially in the ambient CO2 chambers becauseof the relatively small difference in 13C between SOCand the CO2 from total soil respiration. Fortunately, thedifference 13C among SOC fractions was often less than0.5 in forest ecosystems (Bird et al. 1996), so the errorassociated with this assumption will be marginal andagain should not alter the patterns described in theprevious sections. Our test with the soil lysimeters alsoindicated that this was probably the case in our Douglasfir seedling systems (Lin et al. unpubl. data). It is clearthat further characterization of soil isotopic compositionsare needed to more accurately estimate the contributionof SOM oxidation.

In addition, the partitioning of the soil CO2 efflux intoits components depends in part on the extent to which


CO2 diffusing through a soil layer to the atmospheredoes or does not come into isotopic equilibrium with thesoil water in a particular layer (Tans 1998). This matteris not well understood at the moment and there arefew experimental data available. While Hesterberg &Siegenthaler (1991) assume full expression of the diffusionisotope fractionation factor for 12C18O16O between thesoils and the atmosphere (8.8) as did Farquhar et al.(1993) for the gradient between leaves and the atmo-sphere, Ciais et al. (1997) calculated that a value of 3.29was required in order to balance biosphere-atmospherefluxes at the global level. Given our assumption ofturbulent transfer from the litter layer and diffusivetransfer from the soil to the atmosphere, the effectiveoverall expression of the diffusion isotope fractionationfactor may be closer to that predicted by Ciais et al.(1997). However, experimental studies are required todetermine the extent to which CO2 fluxing out of soils isin equilibrium with particular soil and litter layers. Fromthe data presented in Fig. 6, it is clear that an exactinterpretation of the soil CO2 efflux data will require abetter understanding of the diffusion and equilibriumisotope fractionation factors.

It is worth mentioning that the difference in 13Cbetween the new carbon (i.e. root carbon) and oldcarbon (including both litter and SOC) in the ambientCO2 treatments was quite small (25) in relation tonatural variations within each carbon pool among replic-ate chambers (Fig. 6). Thus, the partitioning results basedon the 13C analysis alone would result in significanterrors (e.g. at the ambient CO2 treatments we wouldhave seen little contributions from rhizosphere respirationto total soil CO2 efflux). However, there was a largedifference in 18O between litter water and soil water,but there was no difference in 18O between the CO2from SOM oxidation and the CO2 from rhizosphererespiration (Fig. 6). Therefore, the partitioning based onthe 18O analysis alone would only allow separationof soil respiration into litter component and the othercomponent contributed from rhizosphere respiration andSOM oxidation. It is the combination of the 13C and 18Oanalysis that made it possible to partition all threecomponents of soil respiration (Fig. 6). Our trial applica-tion of this dual-isotope technique to the Douglas firterracosms under different CO2 and temperature condi-tions (Fig. 7, Tables 2,3) suggests that this novel methodcan provide new insights into the responses of carbonmetabolism in forest ecosystems to future climatechanges.

Acknowledgements

We thank R. Shimabuku, R. Waschmann, A. Fong, Y. Ke,N. Buchmann, C.F. Kitty, and C. Cook for their help during this


study. A special thanks to Dr. L. da L. Sternberg at the Universityof Miami for his discussion on applications of oxygen isotopesto the present study. We also thank three anonymous reviewersfor their constructive comments and suggestions. The researchdescribed in this article has been funded by the U.S. Environ-mental Protection Agency. This document has been subjected tothe Agencys peer and administrative review and approved forpublication.

References

Balesdent J, Girardin C, Mariotti A (1993) Site-related 13C oftree-leaves and soil organic matter in a temperate forest.Ecology, 74, 453462.

Berntson GM, Woodward FI (1992) The root system architectureand development of Senecio vulgaris in elevated CO2 anddrought. Functional Ecology, 6, 324333.

Bird M, Chivas AR, Head J (1996) A latitudinal gradient incarbon turnover times in forest soils. Nature, 381, 143146.

Bootinga Y, Craig H (1969) Oxygen isotope fractionation betweenCO2 and water, and the isotopic composition of marineatmospheric CO2. Earth and Planetary Science Letters, 5, 285295.

Buchmann N, Kao W, Ehleringer. JR (1997) Influence of standstructure on carbon-13 of vegetation, soils, and canopy airwithin deciduous and evergreen forests of Utah, UnitedStates. Oecologia, 110, 109119.

Cheng WX (1996) Measurement of rhizosphere respiration andorganic matter decomposition using natural C13. Plant andSoil, 183, 263268.

Cheng WX, Zhang QL, Coleman DC, Carroll CR, Hoffman CA(1996) Is available carbon limiting microbial respiration inthe rhizosphere? Soil Biology and Biochemistry, 28, 12831288.

Ciais P, Denning AS, Tans PP et al. (1997) A three-dimensionalsynthesis study of 18O in atmospheric CO2 1. Surface fluxes.Journal of Geophysical Research, 102, 58575872.

Couteaux M-M, Mousseau M, Celeier M-L, Bottner P (1991)Increased atmospheric CO2 and litter quality: decompositionof sweet chestnut leaf litter with animal food webs of differentcomplexities. Oikos, 61, 5464.

Ehleringer JR (1991) 13C/13C fractionation and its utility interrestrial plant studies. In: Carbon Isotope Techniques (edsColeman DC, Fry B), pp. 187200. Academic Press, SanDiego, CA.

Farquhar GD, Lloyd J, Taylor JA et al. (1993) Vegetation effectson the isotopic composition of oxygen in atmospheric CO2.Nature, 363, 439442.

van de Geijin SC, van Veen JA (1993) Implications of increasedcarbon dioxide levels for carbon input and turnover in soils.Vegetatio, 104/105, 283292.

Hesterberg R, Siegenthaler U (1991) Production and stableisotopic composition of CO2 in a soil near Bern, Switzerland.Tellus, 43B, 197205.

Hogberg P, Ekblad A (1996) Substrate-induced respirationmeasured in situ in a C3-plant ecosystem using additions ofC4-sucrose. Soil Biology and Biochemistry, 28, 11311138.

Hungate BA, Holland EA, Jackson RB, Chapin IIIFS, MooneyHA, Field CB (1997) The fate of carbon in grasslands undercarbon dioxide enrichment. Nature, 388, 576579.


Jenkinson DS, Adams DE, Wild A (1991) Model estimates ofCO2 emission from soils in response to global warming.Nature, 351, 304306.

Johnson DW, Geisinger DR, Walker RF, Newman J, Vose JM,Elliot K, Ball JT (1994) Soil pCO2, soil respiration, and rootactivity in CO2-fumigated and nitrogenfertilized ponderosapine. Plant and Soil, 165, 129138.

Korner C, Arnone JA (1992) Responses to elevated carbondioxide in artificial tropical ecosystems. Science, 257, 16721675.

Leavitt SW, Paul EA, Galadima A, Nakayama FS, Danzer SR,Johnson H, Kimball BA (1996) Carbon isotopes and carbonturnover in cotton and wheat FACE experiments. Plant andSoil, 187, 147155.

Lin G, Ehleringer JR (1997) Carbon isotopic fractionation doesnot occur during dark respiration in C3 and C4 plants. PlantPhysiology, 114, 391394.

Mary B, Mariotti A, Morrel JL (1992) Use of 13C variations atnatural abundance for studying the biodegradation of rootmucilage, roots and glucose in soil. Soil Biology andBiochemistry, 24, 10651072.

Merckx R, den Hortog A, van Veen JA (1985) Turnover of root-derived material and related microbial biomass formation insoils of different texture. Soil Biology and Biochemistry, 17,565569.

Nadelhoffer KJ, Fry B (1988) Controls on natural nitrogen-15and carbon-13 abundances in forest soil organic matter. SoilScience Society of America Journal, 52, 16331640.

Nakayama FS, Huluka G, Kimball BA, Lewin KF, Nagy J,Hendrey GR (1994) Soil carbon dioxide fluxes in natural andCO2-enriched systems. Agricultural and Forest Meteorology, 70,131140.

Nitschelm JJL, Uscher A, Hartwig UA, van Kessek C (1997)Using stable isotopes to determine soil carbon inputdifferences under ambient and elevated atmospheric CO2conditions. Global Change Biology, 3, 411416.

Norby RJ, Gunderson CA, Wullschleger SD, ONeill EG,McCracken MK (1992) Productivity and compensatoryresponses of yellow-poplar trees in elevated CO2. Nature,357, 322324.

Oberbauer SF, Oechel WC, Riechers GH (1986) Soil respirationof Alaskan tundra at elevated atmospheric carbon dioxideconcentrations. Plant and Soil, 96, 145148.


Pajari B (1995) Soil respiration in a poor upland site of scots pinestand subjected to elevated temperatures and atmosphericcarbon concentration. Plant and Soil, 169, 563570.

Paterson E, Hall JM, Rattray EAS, Griffiths BS, Ritz K, KillhamK (1997). Effects of elevated CO2 on rhizosphere carbon flowand soil microbial processes. Global Change Biology 3, 363377.

Peterjohn WT, Melillo JM, Bowles FP, Steudler PA (1993) Soilwarming and trace gas fluxes experimental design andpreliminary flux results. Oecologia, 93, 1824.

Peterjohn WT, Melillo JM, Steudler PA et al. (1994) Responses oftrace gas fluxes and N availability to experimentally elevatedsoil temperatures. Ecological Applications, 4, 617625.

Raich JW, Schlesinger WH (1992) The global carbon dioxide fluxin soil respiration and its relationship to vegetation andclimate. Tellus, 44B, 8199.

Rogers HH, Runion GB, Krupa SV (1994) Plant responses toatmospheric CO2 enrichment with emphasis on roots andthe rhizosphere. Environmental Pollution, 83, 155189.

Schimel DS (1995) Terrestrial ecosystems and the carbon cycle.Global Change Biology, 1, 7791.

Schonwitz R, Stichler W, Ziegler H (1986) 13 C values of CO2from soil respiration on sites with crops of C3 and C4 typeof photosynthesis. Oecologia, 69, 305308.

Socki RA, Karlsson HR, Gibson EK (1992) Extraction techniquefor the determination of oxygen-18 in water using pre-evacuated glass vials. Analytical Chemistry, 64, 829831.

Tans PP (1998) Oxygen isotopic equilibration between carbondioxide and water in soils. Tellus, 50, 163178.

Tingey DT, McVeety BD, Waschmann R, Johnson MG (1996) Aversatile sun-lit controlled-environment facility for studyingplant and soil processes. Journal of Environmental Quality, 25,614625.

van Veen JA, Liljeroth E, Lekkerkerk LJA, van de Geijin SC(1991) Carbon fluxes in plant-soil systems at elevatedatmospheric CO2 levels. Ecological Applications, 1, 175181.

Vose JM, Elliott KJ, Johnson DW, Walker RF et al. (1995) Effectsof elevated CO2 and N fertilization on soil respiration fromponderosa pine (Pinus ponderosa) in open-top chambers.Canadian Journal of Forest Research, 25, 12431251.

Wedin DA, Tieszen LL, Dewey B, Pastor J (1995) Carbon isotopedynamics during grass decomposition and soil organic matterformation. Ecology, 76, 13831392.

Whipps JM, Lynch JM (1983) Substrate flow and utilization inthe rhizosphere of cereals. New Phytologist, 95, 605623.

Date post:	06-Jan-2016
Category:	Documents
Upload:	jhon-del-aguila-pasquel
View:	223 times
Download:	0 times

soil carbon stocks; tropical montane cloud forest; puna; tree line

Documents