Plant and Soil 195: 221–232, 1997. 221c 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Estimating respiration of roots in soil: Interactions with soil CO2, soiltemperature and soil water content

Tjeerd J. Bouma�, Kai L. Nielsen, David M. Eissenstat and Jonathan P. LynchDepartment of Horticulture, The Pennsylvania State University, 103 Tyson Building, University Park,Pennsylvania 16802–4200, USA (�Author to whom correspondence should be addressed at the following address:Netherlands Institute of Ecology, Center for Estuarine and Coastal Ecology, P.O. Box 140, 4400 AC Yerseke, theNetherlands (fax: +31-113-573616; E-mail: tbouma@cemo.nioo.knaw.nl)

Received 9 January 1997. Accepted in revised form 24 June 1997

Key words: citrus, Citrus volkameriana Tan. & Pasq., CO2-diffusion gradient, root respiration, soil CO2 concen-tration, Volkamer lemon.

Abstract

Little information is available on the variability of the dynamics of the actual and observed root respiration rate inrelation to abiotic factors. In this study, we describe I) interactions between soil CO2 concentration, temperature,soil water content and root respiration, and II) the effect of short-term fluctuations of these three environmentalfactors on the relation between actual and observed root respiration rates. We designed an automated, open, gas-exchange system that allows continuous measurements on 12 chambers with intact roots in soil. By using threedistinct chamber designs with each a different path for the air flow, we were able to measure root respiration over a50-fold range of soil CO2 concentrations (400 to 25000 ppm) and to separate the effect of irrigation on observed vs.actual root respiration rate. All respiration measurements were made on one-year-old citrus seedlings in sterilizedsandy soil with minimal organic material.

Root respiration was strongly affected by diurnal fluctuations in temperature (Q10 = 2), which agrees well withthe literature. In contrast to earlier findings for Douglas-fir (Qi et al., 1994), root respiration rates of citrus werenot affected by soil CO2 concentrations (400 to 25000 ppm CO2; pH around 6). Soil CO2 was strongly affected bysoil water content but not by respiration measurements, unless the air flow for root respiration measurements wasdirected through the soil. The latter method of measuring root respiration reduced soil CO2 concentration to that ofincoming air. Irrigation caused a temporary reduction in CO2 diffusion, decreasing the observed respiration ratesobtained by techniques that depended on diffusion. This apparent drop in respiration rate did not occur if the airflow was directed through the soil. Our dynamic data are used to indicate the optimal method of measuring rootrespiration in soil, in relation to the objectives and limitations of the experimental conditions.

Introduction

More than 50% of plant photosynthates produced dai-ly may be respired by the roots, depending on relativegrowth rate and nutritional status of the plant (Lam-bers et al., 1996). Thus, quantitative information onroot respiration is as important to understanding plantgrowth as is photosynthesis. Root respiration can beaccurately measured in nutrient solutions (e.g., Bloomet al., 1992; Lambers et al., 1996; Veen, 1980). Mea-suring root respiration in soil is less precise due to vari-

able environmental conditions, but valuable as it moreclosely reflects natural conditions. One well-knownsource of variation is microbial activity (e.g., Rochetteet al., 1991). Here we focus on other factors that mayaffect root respiration measurements, using a sterilizedand sieved sandy soil with minimal organic matter tominimize microbial respiration of carbon not originat-ing from the plant.

Root respiration rate is affected by various abiot-ic factors. For example, effects of temperature (e.g.,Edwards, 1991) and soil water content (e.g., Palta and

code c1; PIPS: No.: 144843 BIO2KAP

*144843* plso7009.tex; 3/10/1997; 15:23; v.6; p.1

222

Nobel, 1989a) are well described. Soil temperatureand soil water content can also affect soil CO2 concen-tration (Johnson et al., 1994; Nobel and Palta, 1989)by altering soil diffusivity. Changes in soil CO2 con-centration and soil diffusivity potentially have signifi-cant implications for both the observed and actual rootrespiration rate. Firstly, discrepancies between actu-al and observed respiration rates may theoretically beexpected, if soil diffusivity changes faster than the timenecessary to reach a diffusion equilibrium. Second-ly, contradictory effects of soil CO2 concentration onroot respiration have been reported. Root respiration ofthree desert species was not affected by soil CO2 con-centrations ranging between 350 and 2000 ppm (Nobeland Palta, 1989; Palta and Nobel, 1989b), whereas rootrespiration rates of Douglas-fir was strongly inhibitedin the same range (Qi et al., 1994). This discrepancyhas important implications. Assuming that respiratorycarbon losses affects competitive strength, the habitatof a species may partly be determined by soil porosityif the respiratory response to soil CO2 concentrationsdiffers between species. Methodological errors mayoccur if respiration is measured at CO2 concentrationsdeviating from natural conditions, causing major devi-ations in estimated plant carbon balance.However, fewstudies describe the effect of soil atmosphere on root-respiration rate, especially compared to the large bodyof work showing a varied response of leaf and whole-plant respiration to elevated CO2 (reviews by Amthor,1991; Poorter et al., 1992; Wullschleger et al., 1994).The combination of possible direct and indirect effectsof soil temperature and soil water content make res-piration measurements in soils complex to interpret.Detailed information is needed on the dynamics ofthese processes.

Our objective is to analyze the short-term dynamicsand interactions among soil CO2 concentration, tem-perature, soil water content, and actual and observedroot respiration rates. In our first experiment, we test-ed the hypothesis that increased CO2 concentrationinhibits root respiration rates, as found by Qi et al.(1994). We used three different chamber designs toobtain a range of soil CO2 concentrations. Diurnal fluc-tuations in temperature were analyzed to describe therelation between root respiration and temperature. In asecond experiment, we examined how soil CO2 con-centration and root respiration (actual and observed)are affected by soil water content. In both experiments,soil CO2 was used as an indicator of changes in the soilatmosphere that may occur due to the air stream need-ed for respiration measurements. All respiration mea-

surements were made on one-year-old citrus seedlings,using a newly designed automated open gas-exchangesystem, allowing continuous respiration measurementson 12 chambers.

Materials and Methods

Experimental design

In the first experiment, plants were placed in threedifferent chamber designs (Figure 1; n = 4 per treat-ment). Soil CO2 concentrations were expected to varyamong chamber designs due to different paths of theair stream. The CO2 concentration in the air streamwas expected always to be in the same range, closeto that of the reference air, regardless of the chamberdesigns.

In the first design, a 550-ml Dee-pot (i.e., a plasticconical container; Stuewe and Sons, Corvallis, OR,USA) with plant was inserted into a 1400-ml root res-piration cuvette (described later), so that the Dee-potwas surrounded by the air being measured (‘surround-ing’ measurement; Figure 1A). For both the second andthird method, a plant was transplanted from a Dee-potinto a 1400-ml root respiration cuvette. In one case,air was blown through the soil (‘perfusive’ measure-ment; Figure 1B) whereas in the other method air wasblown over the soil surface, using a perforated tube asto obtain uniform air flow (‘headspace’ measurement;Figure 1C). In all cases, stem respiration was negligible(data not shown). Water was added frequently in smallquantities to minimize saturation of soil macropores,but in sufficient quantities to prevent signs of plantwaterstress. In one-year-old citrus seedlings, water-stress is indicated early by slight wilting of the smallestyoungest leaves on top of the shoot, long before anyof the other leaves show any sign of wilting.

In the second experiment plants were placed in thecuvettes for ‘surrounding’ and ‘headspace’ respirationmeasurements only (Figure 1). Soil CO2 and soil watercontent were measured both in pots (with and withouta plant) used for respiration measurements, as well asin pots (with and without a plant) not connected tothe automated gas-exchange system (overall design inTable 1). Pots without plants were used to determine ifour automated gas-exchange system affected soil CO2

and soil water content. Water was added approximatelyto water holding capacity of the soil at the first sign ofwilting of the smallest youngest leaves.

plso7009.tex; 3/10/1997; 15:23; v.6; p.2

223

Figure 1. Schematic representation of three different methods of measuring root respiration: A) ‘surrounding’ measurement with the air flowingaround a smaller pot inserted into a larger root respiration cuvette, B) ‘perfusive’ measurement with the air flowing through the growth media,and C) ‘headspace’ measurement with the air flowing over the soil surface. All methods used a gas-tight cuvette to obtain good flow recovery.Gas-sampling chambers (exp. 1 and 2) and TDR probes (exp. 2) were inserted to measure soil CO2 concentrations and soil water content,respectively.

Table 1. Design of experiment 2

pot type ‘surrounding’ ‘headspace’

use respiration control respiration control

content plant empty plant empty plant empty plant empty

replicates 3 2 3 2 4 3 4 3

To minimize microbial respiration, we used auto-claved Candler fine sandy soil (Typic quartzipsammentcollected from the Citrus Research and Education Cen-ter in Lake Alfred, FL, USA) from which most of theorganic matter had been removed by sieving. Althoughwe cannot distinguish if respiration is from roots ormicrobes, it is referred to as root respiration becausein sterilized sandy soils with negligible organic matterall respiration has to originate from roots (Palta andNobel, 1989a). Moreover, the majority of soil respi-ration is often most likely from roots (e.g., Chapman1979; Johnson et al.,1994).

Plant material

Scarified seeds of the citrus rootstock, Volkamer lemon(Citrus volkameriana Tan. & Pasq.), were germinatedin flats filled with vermiculite (18 July 1995).Seedlings

were transplanted into 125-ml conetainers with auto-claved Candler fine sandy soil (21 Sept. 1995) andthen into 550-ml Dee-pots (16 Nov. 1995). A weekbefore the first respiration measurements (24 May1996), a number of plants were transplanted into 1400-ml root respiration cuvettes, whereas the remainingplants were repotted in 550-ml Dee-pots. Control potswere also filled with sterile soil. All Dee-pots and rootrespiration cuvettes were equipped with gas-samplingdevices (described later) at 14 cm depth.Nutrients weresupplied to be non-limiting, by increasing the frequen-cy of addition of Hoagland’s solution (5 mM KNO3,5 mM Ca(NO3)2, 5 mM KH2PO4, 2 mM MgSO4,1 mM Fe as FeEDTA and micronutrients; Hoaglandand Arnon, 1939) with plant size. Salt accumulationwas prevented by flushing the soil between additionsof nutrient solution. Temperature was measured usingcopper-constantan thermocouples located between the

plso7009.tex; 3/10/1997; 15:23; v.6; p.3

224

pots at 15 cm depth. Greenhouse temperature fluctuat-ed between 20 to 35 �C depending on external weatherconditions. The respiration measurements were madein a greenhouse with temperatures controlled between20 and 30 �C.

Automated system for measuring root respiration

To enable continuous respiration measurements onseveral plants, we constructed a ‘open’ gas exchangesystem with an infrared gas analyzer (IRGA) for CO2

detection (methodological review by Layzell et al.,1989). Open systems allow operation under steady-state conditions, although this advantage may havesome limitations: i) steady-state conditions are notnecessarily identical to natural conditions of roots, ii)steady-state conditions often cannot be obtained for allenvironmental factors, and iii) natural fluctuations maybe important to understand root functioning. Our gasexchange system was automated to switch between 12different chambers every 4 minutes. The IRGA wasset up in a differential mode for sensitive detection ofsmall differences in root respiration rates (Figure 2;technical details in legend).

Air from an oilless compressor (SEARS depart-ment store) was scrubbed free of CO2 (Figure 2 - Sc)and humidified to a defined relative humidity (Figure2 - Hu). Two mass flow controllers were used to mixCO2-free air (Figure 2 - McA) with 10% CO2 (Figure 2- McC) to obtain the desired CO2 concentration. Fluc-tuations in the CO2 concentrations were limited to �2ppm h�1 by: (1) injecting 10% instead of pure CO2,(2) using a 1 mm ID injection needle (Figure 2 - Ne)to obtain a steady flow of CO2 into the air stream, and(3) blowing the air through four mixing tubes (Figure2 - Mv). The stable CO2 concentration allowed differ-ential measurements, despite small differences in pathlength between reference and sample air. This was amajor advantage for studying root respiration, as thepath length is not exactly known if air is blown throughsoil.

A manifold with 13 outlets with needle valves (Fig-ure 2 - Mf) regulated the air flow through each of the 12root respiration cuvettes (Figure 2 - Rc) and to the ref-erence cell of the IRGA (Figure 2 - Ia). All flows weremeasured with a mass flow meter (Figure 2 - Mm) andkept around 600 ml min�1. Both the reference air andthe air coming from the 12 root respiration cuvetteswere dried in condensation tubes (Figure 2 - Ct). Thedried air was blown into overflow tubes (Figure 2 -Ot), with a suction tube inserted into the wider outlet.

Figure 2. Flow diagram of automated system to measure root respi-ration on 12 different plants. The abbreviations in alphabetic orderare: Ct, condensation tubes with a waterlock at the bottom, to allowautomatic draining; Ia, infrared gas analyzer (LICOR 6252, Lin-coln, Nebraska); Hu, humidifier to regulate air humidity by flowingair over a temperature-controlled water surface; McA, mass flowcontroller for air (0 to 25 l min�1; Series 840, Sierra Instruments,Inc. Monterey, CA); McC, mass flow controller CO2 (0 to 200 mlmin�1; Series 1259C, MKS Instruments Inc., Andover, MD); Mf,manifold with 13 outlets with needle valves; Mm, mass flow meters(0 to 1 l min�1; Series 820, Sierra Instruments, Inc. Monterey,CA); Mv, mixing vessels made from 1.5-m-long PVC pipe (75 mmID), with a ventilator in the middle; Ne, 1 mm ID injection nee-dle; Ot, overflow tubes made from 5-ml disposable pipette tips; Pr,diaphragm pump for the reference air (LICOR 6200 part, Lincoln,NE); Pu, diaphragm pump (LICOR 6200 part, Lincoln, NE); Rc,root respiration cuvettes; Re, refrigerator; Ro, rotameter with a nee-dle valve; Rv, rotating valve (#S2564/1P-12T ScanCo Samplivalve,Scanivalve Corp., San Diego, CA); Sc, CO2 scrubber; Sv, three-waysolenoid valves (type 651068, KIP inc. Farmington, CT).

This construction enabled a constant rate of sampling(i.e., around 250 ml min�1), independent of the flowthrough the root respiration cuvettes.

The suction tubes coming from the 12 root respi-ration cuvettes led to a rotating valve (Figure 2 - Rv),moving one position every 4 minutes. The rotatingvalve, connected one outlet to a diaphragm pump (Fig-

plso7009.tex; 3/10/1997; 15:23; v.6; p.4

225

ure 2 - Pu) and the remaining 11 to open outlets. Ref-erence air coming from the overflow tube was directlysucked into a second pump (Figure 2 - Pr). The flowrate from both pumps to the IRGA was kept around250 ml min�1 by 2 rotameters, each with a needlevalve (Figure 2 - Ro). Two extra condensation tubes(Figure 2 - Ct) located in a refrigerator (Figure 2 - Re),were placed after the 2 rotameters, to ensure a constantrelative humidity of all air going to the IRGA.

Because CO2 differentials in a continuous-flow sys-tem may be very small, significant errors can occur ifcorrections for drift in the zero setting of the IRGA arenot made. Such corrections are especially important incase measurements are made at a different CO2 con-centration than that used for calibration (Garcia et al.,1994; pers. commun. R L Garcia). To correct for zerooffset, sample and reference cells of the IRGA wereswapped on a 2 minute basis, using two, three-waysolenoid valves (Figure 2 - Sv). With an air flow of250 ml min�1, stable readings were obtained in lessthan a minute, so that data were collected during the lastminute of each 2 minute interval. True inlet-outlet CO2

differential (∆Ct) was calculated from two sequentialmeasurements (∆C1 and ∆C2) which would typicallyhave opposite signs (Garcia et al., 1994):

∆Ct = ∆C1 � ((∆C1 + ∆C2)=2)

System control and data collection were done every10 seconds and stored as 1 minute averages (CR7,Campbell Scientific inc., Logan, UT).

Root respiration cuvettes

Root respiration cuvettes were gas-tight to obtaingood flow recovery. This was especially important forheadspace measurements (Figure 1C), because a smallleakage in the bottom of the pot will cause a gas flowopposite to the diffusion gradient. Gas-tight cuvetteswere constructed from PVC tubing (76 mm ID), witha flange glued to the top and an end-cap sealed to thebottom with silicone caulking. The flat surface of theflange allowed easy attachment of a solid PVC lid, witha small slot cut out for the stem. A gas-tight seal wasobtained by attaching foam (3.2 mm thick clip foamwith 58465 cohesive; H-O products, Winsted, CT) tothe bottom of each lid, and filling the slot plus thearea around the stem with a flexible sealant (Terostat).During ‘headspace’ respiration measurements (Figure1C), the drainage hole in the bottom of the pot wastemporary sealed with a solid stopper.

Soil CO2 concentrations and soil water content

Atmospheric soil CO2 concentrations were determinedin gas samples pulled from small chambers insertedinto the soil. These gas-sampling chambers were madefrom the top 3 ml of a 10-ml Nalgene syringe, witha 30-�m Nytex screen glued over the cut surface. Apiece of Tygon tubing (400 mm long; 1.6 mm ID) wasconnected to the tip where the needle would normallybe attached, allowing aboveground sampling. Sampleswere taken at 14 cm depth by filling a 3-ml syringe.The syringe was coated with Teflon tape to reduce CO2

losses. From each 3-ml sample, 0.5 ml was injected ina gas chromatograph (5840A, Hewlett-Packard comp.,Palo Alto, CA). The stainless steel column (length0.9 m, inner diameter 3.2 mm) contained Poropak OS(Supelco, Bellefonte, PA), the carrier gas was N2 andoven temperature was 70 �C. A Ni catalyst (350 �C)converted the CO2 to methane which was then detectedby a flame ionization detector (175 �C).

Soil water content was measured by time domainreflectometery (1502C metallic time domain reflec-tometer; Tektronix Inc., Beaverton, OR) (Topp andDavis, 1985; Topp, 1993). Probes (unbalanced design;11-cm length x 1.6 mm diameter stainless-steel rods)were inserted in the top of the soil, and consisted of3 stainless steel rods of 100 mm length, placed 10mm apart (Figure 1). Signal analysis was done usingthe software developed by R. Hubbard (Department ofSoil Physics, Utah State University).

Statistics

Correlation coefficients were tested for significanceat the 0.05 to 0.01 level (Rohlf and Sokal, 1981).Where correlations were significant, regression equa-tions were calculated. Error bars in figures are used toindicate the standard errors.

Results

General (experiment 1)

A complex pattern of root respiration was found in rela-tion to temperature, soil CO2 concentration, irrigationand cuvette designs (Figure 3). To facilitate interpreta-tion of the dynamics of the raw data (Figure 3), it wasnecessary to start with separating out effects of tem-perature (Figure 4) and soil CO2 concentration (Figure5). Subsequently, we analyzed the strong fluctuations

plso7009.tex; 3/10/1997; 15:23; v.6; p.5

226

Figure 3. Patterns of temperature (A), root respiration rate (B), soil CO2 concentration (C) combined with the irrigation frequency (verticaldashed lines), obtained for three different designs to measure root respiration (exp. 1). The circle triangle and square indicate the ‘surrounding’,‘perfusive’ and ‘headspace’ chamber designs, respectively (Figure 1). Vertical dashed lines (1 to 5) indicates irrigation with 50 ml per plant.Error bars indicate the standard errors of soil CO2 concentration (n = 4). Standard errors of respiration measurements were not shown toenhance clarity (n = four plants per treatment). The average values of the standard errors for the ‘surrounding’, ‘perfusive’ and ‘headspace’measurements were 0.71 (SD = 0.40; n = 53), 0.78 (SD = 0.36; n = 53) and 0.22 (SD = 0.35; n = 53), respectively. The standard deviations(SD) indicate the variation of the standard errors around their average values. After completing the respiration measurements (B), the lids wereremoved from the respiration cuvettes and the gas flow was stopped. We continued to measure soil CO2 concentration (C).

in root respiration at the beginning of the experimentand after each irrigation (Figure 3).

Effects of soil CO2 and temperature on rootrespiration (experiment 1)

The relatively stable respiration rates between irriga-tions exhibited parallel patterns for respiration rate

plso7009.tex; 3/10/1997; 15:23; v.6; p.6

227

Figure 4. Relationship of temperature (T; �C) with root respira-tion rate (r; nmol CO2 (g DW)�1 s�1) estimated by a exponentialregression model (r = 1.11(0:0739T) with R2 = 0.36; n = 141; p <0.01). Symbols (explained in Figure 3) represent the average of fourplants.

Figure 5. Relationship of soil CO2 concentration and root respirationrate standardized for a temperature of 25 �C (R2 = 0.15; n = 17; NS).All points indicated with the same number were measured at the sametime. Symbols (explained in Figure 3) represent the average of fourplants.

and temperature, indicating that root respiration ratesincreased with temperature (Figure 3A and 3B). Plot-ting respiration over temperature yielded a Q10 of 2.1(r2 = 0.36; n = 141; p < 0.01; Figure 4). By definition,the Q10 was fitted with an exponential regression mod-el (i.e., a Q10 of 2 means that the respiration doubles ifthe temperature increases by 10 �C).

During the respiration measurements, soil CO2

concentrations differed significantly between the three

types of respiration measurements (Figure 3C), asanticipated. The significantly reduced soil CO2 con-centration in the perfused cuvettes was expectedbecause ambient air has much lower CO2 concen-trations than air in the soil pores. Observed respira-tion rates were very unstable immediately after irriga-tion and at the beginning of the experiment. Startingapproximately 4 h after irrigation, all types of respi-ration measurements yielded similar rates despite thelarge variation in soil CO2 (Figure 3B). Combining alldata shows that for citrus, there is no short-term effectof soil CO2 concentration on root respiration (Figure5). This result was not due to temperature effects, asrespiration was standardized for temperature.

Estimating respiration rates in relation to irrigation(experiment 1)

Although estimates of root respiration were general-ly similar for the different cuvette designs, there wereimportant differences at the start of the measurementand after irrigation (Figure 3B). The initial observedrespiration rate obtained by ‘perfusive’ measurementswas relatively high compared to rates obtained by ‘sur-rounding’ and ‘headspace’ methods (Figure 3B). Thisrelatively high respiration rate coincided with a rapiddrop of soil CO2 in the perfused pots, until soil CO2

reached the CO2 concentration of incoming air (Figure3C). Thus, a flush of CO2 initially present in the soil,resulted in an overestimated value of initial respirationrate by the ‘perfusive’ method. Following irrigation,both the ‘surrounding’ and ‘headspace’ measurementsof respiration showed a temporarily reduced CO2 fluxfrom the soil (Figure 3B). This decrease was most like-ly due to increased resistance for CO2 diffusion in thesoil because of reduced air-filled soil pores. The forcedairflow in ‘perfusive’ measurements prevented diffu-sive resistance from becoming a limiting factor to CO2

flux (Figure 3C). Thus, during the first hours after irri-gation the observed respiration rate underestimated theactual respiration rate in measurements that dependedon CO2 diffusion in the soil.

Immediately after the second irrigation, ‘sur-rounding’ estimates of respiration were lower than‘headspace’ estimates (Figure 3B), but after approx-imately 1.3 days ‘surrounding’ estimates were high-er than ‘perfusive’ and ‘headspace’ estimates. Theincrease observed for the ‘surrounding’ method, wasaccompanied by a decrease in soil CO2 (Figure 3C).A direct effect of the air flow in a ‘surrounding’ mea-surement on soil CO2 concentration seems unlikely, as

plso7009.tex; 3/10/1997; 15:23; v.6; p.7

228

Figure 6. Relationship of irrigation (vertical lines) and temperature (A) on root respiration rate (B), soil CO2 concentration (C) and soil watercontent (D). Circles indicate plants used for respiration measurements, squares plants that were not connected to the gas-exchange system.Open and closed symbols represent the ‘surrounding’ and ‘headspace’ measurements, respectively. Vertical lines consisting of small dashes(number 3 and 5), large dashes (number 6) and the mix of small and large dashes (number 1, 2 and 4) indicate water addition to the plantsgrown in conetainers (i.e., ‘surrounding’ measurements), plants grown in root respiration cuvettes (i.e., ‘headspace’ measurements) and to bothset of plants, respectively. After irrigation, we indicated the soil moisture content to start at the highest level we measured. Error bars indicate� standard error (number of replicates, n, indicated in Table 1).

soil CO2 concentration did not increase after the ‘sur-rounding’ measurements were stopped (Figure 3C).Such an increase of soil CO2 concentration was foundafter terminating the ‘perfusive’ measurements (Figure3C). Therefore we hypothesize these differences weredue to the smaller pot used in the surrounding mea-

surements (i.e., 550 ml versus 1400 ml for ‘perfusive’and ‘headspace’ measurements). The smaller volumepresumably resulted in a relatively high soil moisturecontent (i.e., high resistance for diffusion) directly afterirrigation, but a relatively low soil water content (i.e.,low resistance for diffusion) after a prolonged period

plso7009.tex; 3/10/1997; 15:23; v.6; p.8

229

Figure 7. Relationship of soil moisture content (Sm; %) with soilCO2 concentration (Sc; ppm) estimated by a linear regression model(Sc = 3586 Sm + 1880; R2 = 0.657, n = 48; p < 0.01). Symbols(explained in Figure 6) represent the average of four plants subjectedto headspace measurements, or three plants subjected to surroundingmeasurements.

of transpiration by the plant. These predictions wereexamined in our second experiment where soil watercontent was monitored.

Effects of soil water content on soil CO2 and rootrespiration (experiment 2)

The parallel patterns in figure 6A and 6B, indicatedthat root respiration rates increased with temperature.As in experiment 1, the major exception was a tem-porary drop in CO2 flux from the soil after irrigation(Figure 6B). This drop was larger than observed in theprevious experiment (Figure 3B) because we irrigatedto pot capacity. Coinciding with the drop in respira-tion, soil CO2 increased gradually (Figure 6C) where-as soil water content increased immediately followedby a gradual decrease over time (Figure 6D). The rela-tion of soil CO2 to soil moisture can be more clearlydemonstrated by plotting both variables against eachother (Figure 7). Data measured when time = 0.5 day(Figure 6C and D) were excluded from regression anal-ysis, because the brief period after irrigation (70 min)was insufficient for soil CO2 concentration to build up.Soil CO2 was strongly affected by the soil water con-tent, presumably because of the effect of soil moistureon resistance to CO2 diffusion. However, apart from abrief reduction in the CO2 diffusion rate from the soilsurface (i.e., brief reduction of observed respirationrate), there was no clear effect of soil water content onactual root respiration for the short ‘drought’ periodsapplied in this experiment (Figure 8).

Estimating root respiration using ‘surrounding’ or‘headspace’ cuvettes had no effect on the soil CO2 con-

Figure 8. Relationship of soil moisture content and root respirationrate (R2 = 0.02, n = 30; NS). Symbols (explained in Figure 6) repre-sent the average of four plants subjected to headspace measurements,or three plants subjected to surrounding measurements.

centration; patterns of soil CO2 were the same for bothcuvettes whether or not they were connected to the gasexchange system (Figure 6C). Similarly, pots withoutplants also showed no effect of measuring backgroundrespiration on soil CO2 concentrations (stable, nonfluctuating values; data not shown). Whereas perfusivemeasurements altered soil CO2 concentrations to beclose to that of incoming air (Figure 3C), headspaceand surrounding techniques had a negligible effect onthe soil atmosphere.

Discussion

Effects of CO2 on respiration

A large body of work describes effects of elevatedatmospheric CO2 on photosynthesis, respiration andgrowth of the shoot. For root growth and root respi-ration, research on the effects of elevated atmosphericCO2 is increasing (reviewed by Rogers and Runion,1994; data for citrus in Idso and Kimball, 1992a,b).However, still little research focuses on how soil CO2

concentrations affect root processes, despite the factthat soil CO2 concentrations are generally a magnitudehigher than atmospheric CO2 concentration (Figures3, 6 and 7). Soil CO2 concentrations are affected byboth the amount of CO2-producing activity in the soilas well as soil diffusive resistance. This is seen as vari-ation of soil CO2 concentrations with depth (Duenaset al., 1995; Johnson et al., 1994), soil water content

plso7009.tex; 3/10/1997; 15:23; v.6; p.9

230

(Nobel and Palta 1989; our Figure 7), soil type (Duenaset al., 1995) and time of year (Johnson et al.,1994).

Because of the variable and high CO2 concentra-tions normally found in soils, we expected root respi-ration to be adapted to high CO2, as observed in bulkyfruits (Solomos, 1987, after Amthor, 1991). Reports ofCO2 effects on roots are still limited and controversial.Qi et al. (1994) observed that root respiration drops bya factor of 4 to 5 with a doubling of soil CO2 concen-trations. The largest effects were found in the rangefrom 100 to 2000 ppm CO2, even though soil CO2

concentrations up to 7000 ppm should be quite com-mon for Douglas-fir roots. In contrast, root respirationwas not affected by normal soil CO2 concentrationsfor three desert species; respiration was only inhibitedabove 3000 ppm CO2 (Nobel and Palta, 1989; Paltaand Nobel, 1989b). The latter finding is supported bythe present data, where respiration was unaffected atthe high soil CO2 concentrations natural to our pottedcitrus plants (up to 25000 ppm; Figure 5).

Qi et al. (1994) hypothesized that the lack of signif-icant growth respiration may explain why they found astronger CO2 response than observed by Nobel and Pal-ta (1989) and Palta and Nobel (1989b). This hypothesiswas based on earlier reports indicating high CO2 sen-sitivity of maintenance respiration of shoots (Reuveniand Gale, 1985; Wullschleger et al., 1992). Presentstudy was not designed to separate respiratory com-ponents for maintenance (g CO2 g�1 s�1) and growth(g CO2 g�1; to be multiplied with the relative growthrate to obtain overall costs for growth on a dry weightbasis). However, one-year-old citrus seedlings are rel-atively large, and have negligible growth compared tothe amount of biomass to be maintained. Despite thelow growth rate, we did not find a CO2 response ofroot respiration. Hence, present data do not to supportthe hypothesis of Qi et al. (1994) that lack of signif-icant growth respiration may result in a stronger CO2

response.An alternative explanation for ‘apparent’ different

effects of soil CO2 on root respiration may be causedby differences in soil pH (pers. commun. H Lambers).At relative high pH values in the soil, a larger fractionof the CO2 will be transformed in bi-carbonate. Thismay cause an apparent insensitivity of root respirationto CO2, which would disappear if the CO2 responsewould be studied at a lower pH. The pH in the presentstudy was around 6, whereas Douglas-fir generallygrows at soils with lower pH values.

Regarding the limited amount of literature on theeffect of CO2 on root respiration, it is not yet conclu-

sive if differences represent methodological artifactsor, species-specific adaptations. Although we foundno effect of CO2 on the respiration rate of citrus roots,there are potential artifacts involved in making mea-surements at CO2 concentrations and pH values deviat-ing from natural conditions when using other species.Such artifacts can have a major impact on modelsdescribing carbon budgets of whole-plant and ecosys-tems.

Effects of temperature and soil water content onrespiration

A parallel pattern between temperature (Figure 3A)and respiration rate (Figure 3B) is often observed (e.g.,Edwards, 1991). Temperature, however, may be afunction of solar radiation, especially if working withsmall pots in a greenhouse (data not shown). Althoughour study was not designed to separate these two fac-tors, fluctuations in root respiration were ascribed totemperature as our Q10 value agrees well with the rangereported for other species (Palta and Nobel, 1989aand after Frossard, 1985; Lawrence and Oechel, 1983;Rochette et al., 1991; Sowel and Spomer, 1986). Fur-thermore, experiments separating both factors foundno effect of light on root respiration (Palta and Nobel,1989a). Observed Q10 values should be used cautiouslyto predict short-term temperature responses at differentconditions, as Q10 response can acclimate to seasonaltemperature fluctuations (Edwards, 1991). Short-termresponse to temperature should also not be extrapolat-ed to predict effects of long-term temperature changes.David Bryla (pers. commun.) observed for citrus iden-tical root respiration rates at 25 and 35 �C, after oneweek of acclimation. Frossard (1985) showed differ-ent rates of acclimation between two maize genotypes,over a 48-h period. In general, existing literature indi-cates that the presence (Frossard, 1985; Smakman andHofstra, 1982; Zimmerman et al., 1989) or absence(Sowel and Spomer, 1986; Weger and Guy, 1991) ofacclimation of root respiration is species specific.

Root respiration generally decreases due to drought(Palta and Nobel, 1989a; Rochette et al., 1991), includ-ing citrus roots exposed to prolonged periods (> 3weeks) of drought (Kosola and Eissenstat, 1994). Dur-ing the periods between irrigation in present experi-ments, root respiration was not diminished by drought(Figure 8), even though the smallest youngest leavesof the shoots were wilted prior to the next irrigation.

plso7009.tex; 3/10/1997; 15:23; v.6; p.10

231

Implications for measuring root respiration in pots

Our results suggest that the best method to estimaterespiration of citrus roots is by blowing air through thesoil (‘perfusive’ method; Figure 1). This is the onlymethod where irrigation does not disrupt respirationmeasurements, so that only the initial reading shouldbe disregarded as opposed to having disturbed read-ings after each irrigation (Figure 3). A disadvantageof the perfusive (and also headspace) method is thatplants need to be grown in respiration cuvettes for atleast a short period. Therefore, surrounding measure-ments (Figure 1) might be preferred in situations whereroot respiration will be measured on a large numberof individual plants. Because the surrounding methoddoes not affect soil CO2 concentration, it might also bepreferred over the perfusive method if the relationshipbetween soil CO2 and root respiration is not known.Headspace method should be avoided because it ismore sensitive to leakage than both other methods. Incase of any leakage in the lower half of the respirationcuvette, air will flow through the soil opposite to thediffusion gradient when using the headspace method.A small airflow opposite the diffusion gradient can sig-nificantly affect respiration measurements (Kanemasuet al., 1974; unpublished data). In the perfused andsurrounding methods, a small leakage in the lower halfof the respiration cuvette will reduce the return flow,but does not affect the observed respiration rates.

Dynamics of respiration rates observed by methodsdepending on diffusion do not necessarily representplant processes, but may be a result of soil physicalprocesses. For example, irrigation can disrupt esti-mates of respiration made with the surrounding andheadspace methods (Figure 6). On the other hand, ifthe soil dries out slowly, soil CO2 will reduce gradually,causing a slight overestimation of observed root res-piration rate (surrounding method; Figure 3). Hence,frequent irrigation with small amounts of water willminimize problems when measuring root respirationby diffusive methods such as the headspace and sur-rounding method. Temporary effects of irrigation willnot affect estimates of respiration that are integratedover a longer period of time, because CO2 accumu-lated in the soil after irrigation will be released andmeasured as the soil dries. This is a main advantage ofusing a static approach (e.g., Edwards, 1982) if con-tinuous measurements are not possible. If respirationrates are only assessed by daily point measurements,time of measuring should be based on a response curveof observed root respiration over time after irrigation.

Implications for measuring root respiration in thefield

Effects of soil water content on respiration can also beseen in the field. Rochette et al. (1991) measured rootrespiration in a headspace on top of the soil, in oneof the few field studies measuring short-term dynam-ics of CO2 evolution. The hourly respiration measure-ments decreased below its previous level during a lightrainfall on an already wet soil, but came back to theearlier level after the rain stopped. As hypothesizedby Rochette et al. (1991) our CO2 and soil water con-tent measurements clearly show that observed respira-tion rates should decrease due to decreased soil CO2

diffusivity, causing increased soil CO2 concentrationsfollowing irrigation (Figures 6 and 7). Soil diffusivityneed not always be the main factor determining soilCO2 concentration (Johnson et al., 1994), but there aremany situations where it is (e.g., Duenas et al., 1995).

Conclusion

Respiration rates of citrus roots were affected by tem-perature, but not by soil CO2 concentration or soilwater content. The absence of a CO2 response ofroot respiration may not be generalized, and experi-ments should be designed to explain different respons-es between different species. Effects of soil pH shouldbe included in such experiments. Soil CO2 was strong-ly affected by soil water content, but not by rootrespiration measurements techniques that depend onCO2 diffusion. Respiration measurements that blowair through the soil typically change the soil CO2 con-centration. Irrigation was found to cause a temporaryreduction in CO2 diffusion. For that reason, root respi-ration measurements depending on CO2 diffusion willbe underestimated following irrigation. However, thiswill not affect estimates of root respiration if data areintegrated over a longer period. Such an underestima-tion can be prevented by ‘perfusing’ the air throughthe soil, which also makes the system less sensitive togas leaks.

Acknowledgments

We thank R De Visser, C S Pot, P H van Leeuwen(AB-DLO, Wageningen, the Netherlands) and R. Gar-cia (LICOR, Lincoln, NE) for their contribution indesigning the open gas-exchange system; several spe-

plso7009.tex; 3/10/1997; 15:23; v.6; p.11

232

cific parts and general principles were copied fromexisting systems they developed.David Bryla and Mar-ianne Resendes are gratefully acknowledged for thebeautiful set of 1-year-old citrus seedling as well asthe TDR probes. We thank Kathleen Brown for the useof the gas chromatograph. This research was supportedby grants from the National Science Foundation (IBN-9596050) and United State Department of Agriculture(NRI 94-37107-1024, 94-37100-0311).

References

Amthor J S 1991 Respiration in a future, higher-CO2 world. PlantCell. Environ. 14, 13–20.

Bloom A J, Sukrapanna S S and Warner R L 1992 Root respirationassociated with ammonium and nitrate absorption and assimila-tion by barley. Plant Physiol. 94, 85–90.

Chapman S B 1979 Some interrelationships between soil and rootrespiration in lowland Calluna heathland in southern England. J.Ecol. 67, 1–20.

Duenas C, Fernandez M C, Carretero J, Liger E and Perez M 1995Emission of CO2 from some soils. Chemosphere 30, 1875–1889.

Edwards N T 1982 The use of soda-lime for measuring root respira-tion rates in terrestical systems. Pedobiologia 23, 321–330.

Edwards N T 1991 Root and soil respiration responses to ozone inPinus taede L. seedlings. New Phytol. 118, 315–321.

Frossard J S 1985 Effect de la temperature des racines sur leurrespiration et sur la croissance de plantules de deux hybrides demais. Agronomie 5, 719–725.

Garcia R L, Idso S B, Wall G W and Kimball B A 1994 Changesin net photosynthesis and growth of Pinus eldarica seedlings inresponse to atmospheric CO2 enrichment. Plant Cell Environ. 17,971–978.

Hoagland D R and Arnon D I 1939 The water-culture method forgrowing plants without soil. University of California, Agricultur-al Experimental Station Circular 347, Berkeley, CA.

Idso S B and Kimball B A 1992a Effects of atmospheric enrichmenton photosynthesis, respiration and growth of sour orange trees.Plant Physiol. 99, 341–343.

Idso S B and Kimball B A 1992b Seasonal fine root biomass devel-opment of sour orange trees grown in atmospheres of ambient andelevated CO2 concentration. Plant Cell Environ. 15, 337–341.

Johnson D, Geisinger D, Walker R, Newman J, Vose J, Elliot K andBall T 1994 Soil pCO2, soil respiration, and root activity in CO2-fumigated and nitrogen-fertilized ponderosa pine. Plant Soil 165,129–138.

Kanemasu E T, Powers W L and Sij J W 1974 Field chamber mea-surements of CO2 flux from soil surface. Soil Science 118, 233–237.

Kosola K R and Eissenstat D M 1994 The fate of surface roots ofcitrus seedlings in dry soils. J. Exp. Bot. 45, 1639–1645.

Lambers H, Scheurwater I and Atkin O K 1996 Respiratory patternsin roots in relation to their functioning. In Plant Roots, the HiddenHalf. Eds. Y Waisel, A Eshel and K Kafkaki. pp 323–362. MarcelDekker, Inc., New York.

Layzell D B, Hunt S, King B J, Walsh K B and Weagle G E 1989A multichannel system for steady-state and continuous measure-ments of gas exchange from legume roots and nodules. In Appli-cation of continuous and steady-state methods to root biology.

Eds. J G Torrey and L J Winship. pp 1–28. Kluwer AcademicPublisher, Dordrecht.

Lawrence W T and Oechel W C 1983 Effects of soil temperature onthe carbon exchange of taiga seedlings. I. Root respiration. Can.J. For. Res. 13, 840–849.

Nobel P S and Palta J A 1989 Soil 02 and CO2 effects on rootrespiration of cacti. Plant Soil 120, 263–271.

Palta J A and Nobel P S 1989a Influence of water status, temperature,and root age on daily patterns of root respiration for two cactusspecies. Ann. Bot. 63, 651–662.

Palta J A and Nobel P S 1989b Influence of soil O2 and CO2 on rootrespiration for Agave deserti. Physiol. Plant. 76, 187–192.

Poorter H, Gifford R M, Kriedemann P E and Wong S C 1992 Aquantitative analysis of dark respiration and carbon content asfactors in the growth response of plants to elevated CO2. Aust. J.Bot. 40, 501–513.

Qi J, Marshall J D and Mattson K G 1994 High soil carbon dioxideconcentrations inhibit root respiration of Douglas fir. New Phytol.128, 435–442.

Reuveni J and Gale J 1985 The effect of high levels of carbon dioxideon dark respiration and growth of plants. Plant Cell Environ. 8,623–628.

Rochette P, Desjardins R L and Pattey E 1991 Spatial and temporalvariability of soil respiration in agricultural fields. Can. J. SoilSci. 71, 189–196.

Rogers H H and Runion G B 1994 Plant responses to atmosphericCO2 enrichment with emphasis on roots and the rhizosphere.Environ. Pollution 83, 155–189.

Rohlf F J and Sokal R R 1981 Statistical Tables, second edition.W.H. Freeman and Company, New York.

Smakman G and Hofstra R J J 1982 Energy metabolism of Plantagolanceolata, as affected by change in root temperature. Physiol.Plant. 56, 33–37.

Solomos T 1987 Principles of gas exchange in bulky plant tissues.Hort. Science 22, 766–771.

Sowell J B and Spomer G G 1986 Ecotypic variation in root respi-ration rate along elevational populations of Abies lasiocarpa andPicea engelmannii. Oecologia 68, 375–379.

Topp G C 1993 Soil water content. In Soil Sampling and Methodsof Analysis. Ed. M R Carter. pp 541–557. Can. Soc. Soil Sci.,Lewis Publishers.

Topp G C and Davis J L 1985 Measurement of soil water contentusing time-domain reflectometry (TDR): a field evaluation. SoilSci. Soc. Am. J. 49, 19–24.

Veen B W 1980 Energy cost of ion transport. In Genetic Engineer-ing of Osmoregulation. Impact on Plant Productivity for Food,Chemicals and Energy. Eds. D W Rains, R C Valentine and CHolaender. pp 187–195. Plentum Press, New York.

Weger H G and Guy R D 1991 Cytochrome and alternative pathwayrespiration in white spruce (Picea glauca) roots. Effects of growthand measurement temperature. Physiol. Plant. 83, 675–681.

Wullschleger S D, Norby R J and Gunderson C A 1992 Growth andmaintenance respiration in leaves of Liriodendron tulipifera L.exposed to long term carbon dioxide enrichment in the field. NewPhytol. 121, 515–523.

Wullschleger S D, Ziska L H and Bunce J A 1994 Respiratoryresponse of higher plants to atmospheric CO2 enrichment. Phys-iol. Plant. 90, 221–229.

Zimmerman R C, Smith R D and Alberte R S 1989 Thermal accli-mation and whole-plant carbon balance in Zostera marina L.(eelgrass). J. Exp. Mar. Biol. Ecol. 130, 93–109.

Section editor: H Lambers

plso7009.tex; 3/10/1997; 15:23; v.6; p.12