8/11/2019 Ann Bot-2012-Graham-253-8

1/6

PART OF A SPECIAL ISSUE ON ROOT BIOLOGY

Roots affect the response of heterotrophic soil respiration to temperaturein tussock grass microcosms

Scott L. Graham 1,2,*, Peter Millard 2,3, John E. Hunt 2, Graeme N. D. Rogers 2 and David Whitehead 21 School of Biological Sciences, University of Canterbury, Private Bag 4800, Christchurch 8140, New Zealand, 2 Landcare

Research, PO Box 40, Lincoln 7640, New Zealand and 3

The James Hutton Institute, Craigiebuckler, Aberdeen AB15 8QH, UK * For correspondence. E-mail scott.graham@pg.canterbury.ac.nz

Received: 28 November 2011 Returned for revision: 25 January 2012 Accepted: 22 February 2012 Published electronically: 6 April 2012

Aims and Background While the temperature response of soil respiration ( RS ) has been well studied, the parti-tioning of heterotrophic respiration ( RH ) by soil microbes from autotrophic respiration ( RA ) by roots, known tohave distinct temperature sensitivities, has been problematic. Further complexity stems from the presence of roots affecting RH , the rhizosphere priming effect. In this study the short-term temperature responses of RAand RH in relation to rhizosphere priming are investigated.

Methods Temperature responses of RA , RH and rhizosphere priming were assessed in microcosms of Poa cita

using a natural abundance d13

C discrimination approach. Results The temperature response of RS was found to be regulated primarily by RA , which accounted for 70 %of total soil respiration. Heterotrophic respiration was less sensitive to temperature in the presence of plant roots,resulting in negative priming effects with increasing temperature.

Conclusions The results emphasize the importance of roots in regulating the temperature response of RS , and aframework is presented for further investigation into temperature effects on heterotrophic respiration and rhizo-sphere priming, which could be applied to other soil and vegetation types to improve models of soil carbonturnover.

Key words: Soil respiration, heterotrophic, rhizosphere priming, temperature sensitivity, partitioning respiration,Poa cita .

INTRODUCTION

Soil respiration ( RS ) represents an important source of CO 2from the terrestrial biosphere to the atmosphere, approximatelyten times greater in magnitude than anthropogenic CO 2 emis-sions ( Raich et al. , 2002; Bond-Lamberty and Thomson,2010 ). It is widely accepted that RS is likely to increase in re-sponse to increasing soil temperature, resulting in a positivefeedback to rising atmospheric CO 2 concentration and result-ant global warming. However, the exact nature of the relation-ship between RS and temperature is still poorly understood.Much of this uncertainty arises from confounding effects of the other important drivers of soil respiration, which includesoil water content and carbon substrate supply ( Davidsonet al. , 2006). As a result, soil-driven positive feedbacks toclimate change remain a critical source of uncertainty in

coupled climate models ( Sitch et al. , 2008 ).Soil respiration is a combination of CO 2 uxes from mul-tiple, distinct carbon sources, primarily autotrophic respiration( RA ), originating from roots and closely associated rhizospheremicrobes, and heterotrophic respiration ( RH ) from microbialdecomposition of soil organic matter (SOM) ( Hanson et al. ,2000 ). Separation of these components remains a challengeto evaluating the direct effects of temperature on RS , as RAand RH are likely to have different drivers and distinct tem-perature sensitivities. Much disagreement exists over the rela-tive temperature sensitivities of RA and RH with some studiesshowing that RA is more sensitive than RH to temperature

(Boone et al. , 1998 ; Wan and Luo, 2003 ), while others showthat RA and RH are similarly sensitive to changes in tempera-

ture (Baath and Wallander, 2003 ). However, photosynthetic al-location to roots has also been shown as an important driver,particularly of RA , and a potential confounding factor in deter-mining the temperature sensitivity of RS (Hogberg et al. , 2001 ;Bhupinderpal et al. , 2003 ).

While both RA and RH could potentially increase underclimate warming scenarios, it is the breakdown of SOM bysoil microbes (i.e. RH ) that is of particular importance, asSOM represents a large pool of stored carbon rather than thecarbon recently assimilated by plants. Syntheses of theoryand experimental data have led to identication of processesinvolved in the regulation of SOM dynamics and their likelyresponses to temperature ( Davidson and Janssens, 2006 ; vonLutzow and Ko gel-Knabner, 2009 ; Conant et al. , 2011).

However uncertainty remains over the net response of decom-position of SOM to temperature. Central to this uncertainty isthe existence of multiple pools of carbon in the soil, whichhave varying turnover times and degrees of recalcitrance(Trumbore, 2000 ). A small fraction of carbon residing insoils is rapidly turned over by microbial mineralization,while the majority of SOM remains relatively inert, with turn-over times ranging from decades to millennia. Kinetic theorypredicts that the more recalcitrant carbon will have a highertemperature sensitivity ( Bosatta and A gren, 1999 ). However,some studies show that only the small, labile carbon compo-nent of SOM will increase in turnover under warmer

# The Author 2012. Published by Oxford University Press on behalf of the Annals of Botany Company. All rights reserved.

For Permissions, please email: journals.permissions@oup.com

Annals of Botany 110 : 253258, 2012doi:10.1093/aob/mcs073, available online at www.aob.oxfordjournals.org

mailto:scott.graham@pg.canterbury.ac.nzmailto:scott.graham@pg.canterbury.ac.nz

8/11/2019 Ann Bot-2012-Graham-253-8

2/6

conditions, while the majority of soil carbon will be insensitiveto temperature ( Liski et al. , 1999 ; Giardina and Ryan, 2000 ).This is supported by the nding of many soil- warming experi-ments that RS is enhanced initially by warming, but this effectis reduced over time ( Rustad, 2001 ; Melillo et al. , 2002 ). Thisapparent acclimation of RS is hypothesized to be related, in

part, to the depletion of a small active carbon pool that is vul-nerable to temperature increase, while the majority of soilcarbon remains unaffected. However, this observed acclima-tion is not exclusive of a temperature response of recalcitrantSOM (Kirschbaum, 2004 ; Larionova et al. , 2007 ).

Another source of uncertainty in evaluating the temperatureresponse of RS is that the rate of RH is known to be inuencedby roots, the so called rhizosphere priming effect ( Kuzyakov,2002 ). The rhizosphere priming effect refers to the inuencethat living roots exert on SOM turnover due to their impacton the physical and chemical environment within surroundingsoil resulting in an increase or decrease in RH relative to root-free soil. Priming effects have been shown to range from a 50% decrease in RH to a 380 % increase in response to the pres-

ence of roots ( Gardena s et al. , 2011), varying with plantspecies, plant phenology and soil fertility ( Cheng et al. ,2003 ; Dijkstra et al. , 2006; Phillips and Fahey, 2008 ).Despite the potentially large effect of roots on rates of RH ,the effects of rhizosphere priming on the temperature sensitiv-ity of RH have been largely unexplored.

Stable isotope techniques provide a powerful tool for evalu-ating rates of RH in undisturbed systems ( Hanson et al. , 2000 ).These techniques utilize distinct d13 C signatures of CO 2respired by roots and associated rhizosphere microbes andthe respiration of microbes involved in heterotrophic SOM de-composition to partition RS into RA and RH . Some recentstudies have used such techniques to evaluate temperatureeffects on RH and rhizosphere priming ( Bader and Cheng,

2007 ; Uchida et al. , 2010 ; Zhu and Cheng, 2011 b). Whilethese studies have utilized C 3 /C4 shifts and continuous 13 C

labelling techniques to provide greater contrast in the d13 Csignatures of RA and RH , Millard et al. (2010) have demon-strated successful partitioning of RS in native C 3 systems.SOM matter is typically enriched in d13 C, compared withplant biomass, and microbial biomass is still further enrichedcompared with the d13 C signature of bulk soil ( Ehleringeret al. , 2000 ; Bowling et al. , 2008 ). This, combined with thefact that RA is typically more depleted than root biomass(Zhu and Cheng, 2011 a ), allows for the possibility of partition-ing RS using a d

13 C approach, without the use of labelling orC3 /C4 vegetation transitions.

In this study, we use a natural abundance d13 C approach to

partition RS into RA and RH in microcosms of the C 3 tussock grass Poa cita and native tussock grassland soil to investigatethe short-term responses of RA , RH and rhizosphere priming tochanges in soil temperature. This will resolve importantsources of uncertainty in temperature effects on RS by allowingfor direct comparison of the temperature sensitivities of root-derived and SOM-derived components of RS in an undisturbedsoil plant system. As a model system, our tussock grassmicrocosms are representative of grasslands in New Zealand,which allocate a large proportion of their carbon belowground and are an important store of soil carbon nationally(Trotter et al. , 2004 ). As well, globally, grasslands represent

an important below-ground carbon sink ( Scurlock and Hall,1998 ). We present an experimental framework which couldbe applied to other soil and vegetation types to constraintemperature responses of SOM and reduce uncertainty insoil-driven feedbacks in coupled-climate models.

MATERIALS AND METHODS

Soil description

The soil used in this experiment was a silt loam collected froma tussock grassland in central South Island, New Zealand(43.034 8S, 171 .758 8E, 590 m a.s.l.). Soils are classied asacidic allophanic brown ( Hewitt, 2010 ), with a total carboncontent of 4 .2 %, total nitrogen content of 0 .30 % and anaverage microbial biomass carbon of 592 mg kg soil

2 1 . Thetop 300 mm of the mineral horizon was excavated, sievedthrough an 8-mm sieve and well mixed. Field-moist conditionswere maintained to preserve microbial biomass.

Microcosm design

Twenty-eight polyvinyl chloride (PVC) pots, 200 mm indiameter and 300 mm deep, were each lled with 8 .5 kg of eld-moist soil. The bottom of each microcosm was coveredwith 80 % shade cloth to ensure adequate drainage. A100-mm-diameter PVC measurement collar was inserted to adepth of 50 mm in the centre of each microcosm. In 20 of the microcosms, three plants of the tussock grass Poa cita ,100 mm in height, were planted around the periphery of thepot. Remaining microcosms were left unplanted as root-freecontrols. Microcosms were left for 5 months in a shadehouse to allow the soil to settle and roots to proliferatebefore being moved into two controlled environment chambers

set to 15 8C with a 14-h day length and 1000 mmol m2 2

s2 1

irradiance (400700 nm). Microcosms were maintained ingrowth cabinets for 5 months prior to measurement. At thetime of measurement, average root biomass, as estimatedfrom a soil core taken from the centre of each measurementcollar at the conclusion of the experiment, was 12 .2 + 3.9 gpot

2 1 dry weight (approx. 7 .8 L soil volume).Microcosm soil was maintained at approx. 35 % volumetric

water content by wetting soils to eld capacity 2 d prior tomeasurement. When measurements were conducted on con-secutive days, changes in mass were used to determine waterloss. Additional water was added after each days measure-ments to maintain constant soil water content over the entireperiod of measurement.

Temperature control

Soil temperature ( T S ) was manipulated by wrapping eachmicrocosm with a 100-W resistance heating cable (ArgusHeating Ltd, Christchurch, New Zealand) covered with insula-tion. Soil temperature was measured by a thermocouple(Type-T; Omega Engineering, Inc., Stamford, CT, USA),placed at 100-mm depth in the centre of each microcosm,and recorded on a datalogger (CR-5000; Campbell Scientic,Logan, UT, USA). Heating was applied by pulse width modu-lation with maximum heating times of 4060 s with off times

Graham et al. Roots affect temperature response of heterotrophic respiration254

8/11/2019 Ann Bot-2012-Graham-253-8

3/6

of 40120 s to achieve a constant rate of warming (2 8C h2 1 )

and hold a steady soil temperature once measurement tempera-ture was reached. Though some lateral temperature gradientwas inevitable, by applying heat in pulses, the temperature dif-ference between the edge of the pot and centre could be main-tained at , 5 8C and thermocouple placement was selected to

represent the temperature of the soil most relevant to the mea-sured respiration rates, directly beneath the measurementcollars. Heating was started 8 h prior to measurement tobring soil temperature to a measurement temperature of 15(no heating), 20 or 25 8C.

Isotope measurements

RS was partitioned into RA and RH using a natural-abun-dance d13 C approach ( Midwood et al. , 2008 ; Millard et al. ,2010 ). Soil surface efux was collected using a dynamicchamber system described by Midwood and Millard (2011) .A PVC chamber was sealed on the measurement ring of a

microcosm with a foam gasket. Chamber CO 2 concentrationwas maintained at 50 100 mmol mol 2 1 above ambient by sup-plying CO 2 -free air at a rate of 20120 mL min

2 1 dependingon the rate of RS . Soil surface efux was sampled from thechamber at a rate 510 mL min

2 1 lower than the CO 2 -freeair supply to prevent incursion of atmospheric air. Ambientpressure was maintained by allowing excess air to escapethrough a vent tube. CO 2 -free air ow was maintained over aperiod of 150 min, the maximum time required to achievesteady-state conditions within the chamber and purge all at-mospheric air. CO 2 was then collected into an evacuatedTedlar w bag for 2030 min and analysed for d13 C () on atunable diode laser (TGA-100A; Campbell Scientic Inc.).This method has been tested by Midwood et al. (2008) on

columns of sand with CO 2 ow of a known d13

C signatureand shown to provide an accurate sample of soil surfaceefux, without fractionation. Immediately following gas col-lection, chambers were removed and, following a 10-minequilibration period, the rate of RS was measured with a port-able respiration system (SRC-1 and EGM-4; PP Systems,Hitchin, UK). The proportion of RS contributed by hetero-trophic respiration ( f RH ) was determined using the two endmember mixing model ( Robinson and Scrimgeour, 1995 ):

f RH = 1 d 13C RS d 13C RHd 13 C RA d 13C RH

(1)

where d13 C RS is the d13 C signature of soil surface efux and

d13

C RA and d13

C RH are the d13

C signatures of the autotrophicand heterotrophic respiration. Soil surface efux was collectedfrom each microcosm at measurement temperatures of 15, 20and 25 8C. The order of measurements was randomized withthe limitation that, on the nal day of measurement, half of the microcosms were measured at 15 8C and half at 25 8C.Likewise, time of measurement was randomized, as half of the measurements were conducted in the morning and theremainder in the afternoon.

Microcosms containing no plants were used as an independ-ent measure of RH in the absence of roots, RHF . Only respir-ation rate was measured in these microcosms, CO 2 was

not collected for isotopic analysis as no partitioning wasnecessary.

The d13 C signature of RH was obtained by taking a65-mm-diameter soil core from the centre of the measurementcollar immediately following the nal collection of soilsurface efux. Roots were rapidly removed within 35 min

and root-free soil was then sealed in a Tedlarw

bag. The bagwas evacuated using a suction pump and repeatedly ushedwith nitrogen. Soil microbes are known to rapidly shift theirsubstrate utilization following disruption of the soil structure(Crow et al. , 2006 ; Pendall and King, 2007 ), so nitrogen wasused to remove oxygen and slow soil microbial activitywhile atmospheric air was purged from the soil. Following23 min of ushing with nitrogen, CO 2 -free air was addedto the bag and newly respired CO 2 was allowed to accumulate.When CO 2 concentration reached a value within the calibratedrange of the tunable diode laser (TDL; 300500 mmol mol

2 1 ),CO2 was analysed for d

13 C. This value, obtained within 1015 min of taking the core, is the most consistent estimate of the isotopic signature of RH and has been validated in the

eld through comparison with soil surface efux collectedfrom root-exclusion plots.The d13 C signature of RA was obtained by rinsing roots

removed from the soil core, sealing them in a Tedlar w bag,evacuating the air with a suction pump and ushing repeatedlywith CO 2 -free air. Roots were then incubated in CO 2 -free airuntil the concentration of root respired CO 2 was within thecalibrated range of the TDL.

Statistical analyses

An Arrhenius-type equation was used to model the effect of temperature on respiration rate ( Lloyd and Taylor, 1994 ):

R = RE 0

156 02

1T S 227 13

10 (2)

where R10 is the basal rate of respiration at 10 8C, E 0 is relatedto the energy of activation and T S is soil temperature (K).Equation (2) was tted to measurements of RS , RA , RH and RHF using non-linear mixed-effects models conducted in thenlme package ( Pinheiro and Bates, 2000 ) for R v.2 .12.1 (RDevelopment Core Team, 2010 ). Each sample in the analysisconsisted of one measurement of respiration ( RS , RA , RH or RHF ), at a given temperature on a single day. Microcosmswere included as random effects to account for repeated meas-urement of the same microcosm at different temperatures. A

model including the presence of roots as a xed effect on R10 and E 0 was compared with a model that did not includethe xed effect of roots using a likelihood ratio test.

The effect of temperature on d13 C RA was evaluated by con-ducting a Students t -test between d13 C signatures of root-respired CO 2 at 15 and 25 8C. This was repeated ford

13C RH . The effect of temperature on d

13C RH was modelled

using least squares regression.To include the error associated with the isotopic partitioning

method into data analyses, we used a simulation-based ap-proach to generate a distribution of RH values against whichto compare our observed values. Mean f RH and standard

Graham et al. Roots affect temperature response of heterotrophic respiration 255

8/11/2019 Ann Bot-2012-Graham-253-8

4/6

error were calculated according to Phillips and Gregg (2001) at15 and 25 8C using the subset of microcosms for which eachd

13 C RS , d13 C RA and d

13 C RH were collected at the same timeand at the same temperature ( n 10 at each 15 and 25 8C).A population of 1000 samples of f RH was simulated usingmeans and standard errors calculated as described above.

Simulated populations of 1000 samples each were also pro-duced using means and standard errors of RS calculated frommeasurements. Simulated values of f RH were then used to par-tition simulated values of RS to produce 1000 samples of RH ,which reect variation resulting from both measurements of RSand the partitioning method. This population, reecting allsources of variation, was then used to produce 95 % condenceintervals for RH at each 15 and 25 8C.

RESULTS

Average ( + s.e.) d13 C RA and d13 C RH were 29 .73 + 0.08 and

22.94 + 0.17 , respectively. A Students t -test determinedthat d13 C RA was not signicantly different between 15 and 25

8C (P 0.40, Table 1), and thus the value of d13

C RA that wasmeasured for a given microcosm was used for partitioning thatmicrocosm at each of the three measurement temperatures. Thed

13 C signature of RH was signicantly affected by temperature(P , 0.002), becoming depleted by 1 at 25 8C comparedwith 15 8C. As it was only possible to measure d13 C RA andd

13 C RH at 15 and 25 8C, a d13 C RH () value was modelled

from the linear regression equation:

d 13 C RH = 0 08T S 21 26 (3)

where T S is soil temperature ( 8C). This soil end-member agreeswell with the d13 C RS of 22.1 + 0.2 (n 6) measured forroot exclusion plots in the tussock grassland from which the

microcosm soil was sampled. Average d13

C RS was 27 .73 +0.10. Similar to d13 C RH , d

13 C RS became more depleted astemperature increased, though to a greater extent, reectingboth the more depleted signature of RH as well as a greatercontribution of roots to RS (Table 1). Given the differencesbetween d13 C RS , d

13C RA and d

13C RH , all microcosms were

successfully partitioned at all temperatures.As well as isotopic signature, the rate of RS was strongly

inuenced by the presence of roots, with RHF equivalent toapprox. 37 % of RS at all temperature levels (Fig. 1).Likewise, partitioned RH was a small proportion of RS , with f RH ranging from 0 .38 at 15 8C to 0.23 at 25 8C (Table 2).

The temperature sensitivity of RH was also affected by thepresence of roots (Fig. 1). With roots present, RH had amuch lower temperature sensitivity, i.e. a lower E 0 , than RS ,while the temperature sensitivity of RHF was very similar tothat of RS (Table 3). The likelihood ratio test indicated that amodel of RH which included the presence of roots as a xed

effect on E 0 provided a signicantly better t than a modelwhich did not include the effect of presence of roots on RH(P 0.0002).

Rhizosphere priming effects, calculated as the differencebetween RH and RHF , as a percentage of RHF , were 0, 32and 38 % at 15, 20 and 25 8C, respectively, indicating nopriming effects at 15 8C and negative priming effects at 20and 25 8C.

Calculation of errors in f RH associated with the isotopic par-titioning method following Phillips and Gregg (2001) resultedin average values ( + s.e.) for f RH of 0.40 + 0.04 at 15 8C and0.25 + 0.04 at 25 8C. These values differ slightly from those inTable 2 as they represent a subset of the data for which d13 C RS ,d

13C RA and d

13C RH were measured in the same microcosm, at

the same temperature, while the mixed-effects models utilizedthe entire dataset by using a single d13 C RA for each microcosmand a value of d13 C RH modelled from T S . The simulation of RH

TABLE 1. Average ( + s.e.) d13

C signatures of soil surface efux(RS ), root respiration (R A ) and heterotrophic respiration (R H ) bymeasurement temperature

Respiration component n Soil temperature ( 8C) d13 C ()

RS 20 14.4 + 0.1 26 .97 + 0.1320 20 .6 + 0.1 27 .98 + 0.1220 25 .6 + 0.2 28 .25 + 0.14

RH 10 14.5 + 0.1 22 .46 + 0.2410 25 .8 + 0.1 23 .42 + 0.13

RA 10 14.5 + 0.1 29 .81 + 0.1110 25 .8 + 0.1 29 .65 + 0.13

4

3

2

R ( m m o

l m

2 s

1 )

1

015 20

Soil temperature (C)25

R S

R HF

R H

F IG . 1. Relationship between temperature and rate of total soil respiration( RS ), heterotrophic respiration in the presence of roots ( RH ) and heterotrophic

respiration in root-free soil ( RHF ) with curves tted using eqn (2).

TABLE 2. Average ( + s.e.) rate of total soil respiration (R S ) and heterotrophic respiration in the presence (R H ) and absence of roots (R HF ) at three different measurement temperatures and the proportion of soil respiration constituted by heterotrophic

respiration ( f RH )

Respirationcomponent n

Soiltemperature

(8C)

Respiration rate(mmol CO 2 m

2 2

s2 1 ) f RH

RS 20 14.4 + 0.1 1.71 + 0.07 0.38 + 0.0220 20 .6 + 0.1 2.52 + 0.09 0.26 + 0.0120 25 .6 + 0.2 3.34 + 0.11 0 .23 + 0.02

RH 20 14.4 + 0.1 0.64 + 0.03 20 20 .6 + 0.1 0.65 + 0.04 20 25 .6 + 0.2 0.78 + 0.07

RHF 8 14.4 + 0.1 0.64 + 0.05 8 20.6 + 0.1 0.96 + 0.10 8 25.8 + 0.1 1.26 + 0.11

Graham et al. Roots affect temperature response of heterotrophic respiration256

8/11/2019 Ann Bot-2012-Graham-253-8

5/6

resulted in 95 % condence intervals between 0 .54 and0.83 mmol m

2 2 s2 1 at 15 8C, which overlap completely with

the condence intervals of RHF which occurred between 0 .54and 0.73 mmol m

2 2 s2 1 , suggesting no signicant priming

effects at 15 8C. At 25 8C, condence intervals of RH occurredbetween 0 .57 and 1 .09 mmol m

2 2 s2 1 and condence intervals

of RHF were between 1 .02 and 1 .49 mmol m2 2 s

2 1 . The con-dence intervals of RH and RHF overlap slightly; however, in ananalysis where 1000 values of each RH and RHF , treated aspairs, were simulated, 99 .5 % of simulations resulted invalues of RH lower than those of RHF , supporting the resultsof the non-linear mixed-effects model which indicatednegative priming effects at higher temperatures.

DISCUSSION

In our system, RA is primarily responsible for driving both the

rate and temperature sensitivity of RS . On average, RH contrib-uted , 30 % of total RS and, in the presence of roots, RH wasinsensitive to a temperature increase of 10 8C. In the absenceof roots, RH ( RHF here) exhibited a similar temperature re-sponse to RS , suggesting that roots play a critical role, notonly in contributing to overall rate of RS , but also in regulatingthe temperature response of RH in this system.

The priming effects documented here, ranging from noeffect at 15 8C to a 38 % decrease in RH (negative primingeffects) in the presence of roots at 25 8C, fall well within therange of published values for priming effects reviewed byGardena s et al. (2011) . Mechanistic explanations for negativepriming effects include competition for mineral nutrients (par-ticularly nitrogen) between the rhizosphere and heterotrophic

soil microbes and preferential substrate utilization(Kuzyakov, 2002 ). The preferential substrate use hypothesisstates that soil microbes preferentially use labile, root-derivedsubstrates over more recalcitrant SOM resulting in decreasedSOM decomposition in the presence of roots.

In our study, low labile substrate availability is a particularlylikely scenario due to the fact that harvest and preparation of our microcosm soils constituted a signicant soil disturbance.Average total carbon concentration ( + s.e.) measured for un-disturbed tussock grassland adjacent to the area where micro-cosm soils were sampled was 6 .2 + 0.2 %, 30 % higher thanthat for microcosm soil. This suggests that soil carbon was

lost during the disturbance, likely from the most labile fraction.This supports preferential substrate utilization as a possible ex-planation for suppressed rates of RH with increasing tempera-ture in rooted soil, as it is likely there was a large difference inthe quality of root-derived and soil-derived carbon.

While preferential substrate use serves to explain negative

priming effects at a single temperature, it does not providean explicit mechanism for the observed temperature responseof rhizosphere priming. A previous study showed similar dam-pening of the temperature response of RH , which was alsoattributed to preferential use of root-derived carbon atwarmer temperatures ( Uchida et al. , 2010). However, thisstudy involved warming of both plant and soil, and a corre-sponding measured increase in photosynthesis was hypothe-sized to fuel the increasing carbon demand by therhizosphere. In our study, only T S was manipulated, thuswithout an increase in leaf temperature, it is unlikely thatphotosynthetic rates were strongly affected by soil warming.Long-term temperature increases have been known to increasecarbon exudation by roots ( Uselman et al. , 2000 ). However, it

is uncertain how root exudation responds to short-termwarming, especially in the absence of increased photosyn-thesis, as in our study. Without further measurements of carbon substrate availability and microbial biomass and activ-ity at the different temperatures, it is impossible to propose adenitive explanation for the observed temperature responseof rhizosphere priming. However, our results indicate a shifttoward microbial use of root-derived carbon with increasingtemperature.

Regardless of the mechanism behind the observed decreasein temperature sensitivity of RH , these results have importantimplications for the use of temperature responses of RH mea-sured using root exclusion methods or incubations of root-freesoils. In our study, parameterization of models of RH using

data from our root-free soils would result in a substantial over-estimate of RH. Of particular relevance are diurnal temperatureresponses of RS . In the tussock grassland eld site from whichour microcosm soils were sampled, during the 5 months whenplants are most active, diurnal temperature variation at100-mm soil depth is 6 8C on average, but can be as great as13 8C. Our results suggest that temperature-related variationin RS can mainly be attributed to RA at this time scale.

While it is certain that these results should not be general-ized to other soil and vegetation types, and represent onlyshort-term temperature responses of RH , it is also clear thattemperature responses of RH that do not include the effectsof roots should be applied with caution, as they may produceerroneous estimates of SOM turnover when applied in

models of carbon cycling. We present here an experimental ap-proach, which could be applied to other soil types, vegetationtypes and temperature regimes to produce more precise tem-perature responses of RH and include the effects of roots inmodels of soil carbon cycling.

ACKNOWLEDGEMENTS

We are grateful to the Ministry of Science and Innovation,Landcare Research and the Miss E. L. Hellaby IndigenousGrassland Research Trust for funding this research, JasonTylianakis for helpful comments on this manuscript, Elena

TABLE 3. Average ( + s.e.) parameter values of E 0 (kJ mol2 1 )

and R 10 (mmol m2 2 s

2 1 ) for total soil respiration (R S ),autotrophic respiration (R A ) and heterotrophic respiration in the

presence (R H ) and absence (R HF ) of roots

Respirationcomponent n E 0 P-value R10 P-value

Soil and roots RS 60 272 + 12 , 0.0001 1 .17 + 0.06 , 0.0001 RH 60 80 + 37 , 0.0001 0 .53 + 0.05 0.0377 RA 60 347 + 19 , 0.0001 0 .67 + 0.05 , 0.0001Root-free soil RHF 64 279 + 27 , 0.0001 0 .43 + 0.03 0.0001

Parameters were tted by non-linear mixed-effects model using eqn (2).

Graham et al. Roots affect temperature response of heterotrophic respiration 257

8/11/2019 Ann Bot-2012-Graham-253-8

6/6

Moltchanova for statistical advice and Jennifer Peters andTony McSeveny for assistance in the laboratory.

LITERATURE CITEDBa a th E, Wallander H. 2003. Soil and rhizosphere microorganisms have the

same Q 10 for respiration in a model system. Global Change Biology 9:17881791.

Bader NE, Cheng W. 2007. Rhizosphere priming effect of Populus fremontiiobscures the temperature sensitivity of soil organic carbon respiration.Soil Biology and Biochemistry 39: 600606.

Bhupinderpal S, Nordgren A, Ottosson Lo fvenius M, Ho gberg MN,Mellander PE, Ho gberg P. 2003. Tree root and soil heterotrophic respir-ation as revealed by girdling of boreal Scots pine forest: extending obser-vations beyond the rst year. Plant, Cell & Environment 26: 12871296.

Bond-Lamberty B, Thomson A. 2010. Temperature-associated increases inthe global soil respiration record. Nature 464 : 579582.

Boone RD, Nadelhoffer KJ, Canary JD, Kaye JP. 1998. Roots exert a stronginuence on the temperature sensitivityof soil respiration. Nature 396:570572.

Bosatta E, A gren GI. 1999. Soil organic matter quality interpreted thermo-dynamically. Soil Biology and Biochemistry 31: 18891891.

Bowling DR, Pataki DE, Randerson JT. 2008. Carbon isotopes in terrestrialecosystem pools and CO 2 uxes. New Phytologist 178 : 2440.

Cheng W, Johnson DW, Fu S. 2003. Rhizosphere effects on decomposition:controls of plant species, phenology, and fertilization. Soil ScienceSociety of America Journal 67: 14181427.

Conant RT, Ryan MG, A gren GI, et al . 2011. Temperature and soil organicmatter decomposition rates: synthesis of current knowledge and a wayforward. Global Change Biology 17: 33923404.

Crow SE, Sulzman EW, Rugh WD, Bowden RD, Lajtha K. 2006. Isotopicanalysis of respired CO 2 during decomposition of separated soil organicmatter pools. Soil Biology and Biochemistry 38: 32793291.

Davidson EA, Janssens IA. 2006. Temperature sensitivity of soil carbondecomposition and feedbacks to climate change. Nature 440 : 165173.

Davidson EA, Janssens IA, Luo Y. 2006. On the variability of respiration interrestrial ecosystems: moving beyond Q 10 . Global Change Biology 12:154164.

Dijkstra FA, Cheng W, Johnson DW. 2006. Plant biomass inuences rhizo-sphere priming effects on soil organic matter decomposition in two differ-ently managed soils. Soil Biology and Biochemistry 38: 25192526.

Ehleringer JR, Buchmann N, Flanagan LB. 2000. Carbon isotope ratios inbelowground carbon cycle processes. Ecological Applications 10:412422.

Ga rdena s AI, Agren GI, Bird JA, et al . 2011. Knowledge gaps in soil carbonand nitrogen interactions: from molecular to global scale. Soil Biologyand Biochemistry 43: 702717.

Giardina CP, Ryan MG. 2000. Evidence that decomposition rates of organiccarbon in mineral soil do not vary with temperature. Nature 404:858861.

Hanson PJ, Edwards NT, Garten CT, Andrews JA. 2000. Separating rootand soil microbial contributions to soil respiration: a review of methodsand observations. Biogeochemistry 48: 115146.

Hewitt A. 2010. New Zealand soil classication . Lincoln, New Zealand:Mannaki Whenua Press.

Ho gberg P, Nordgren A, Buchmann N, et al . 2001. Large-scale forest gird-ling shows that current photosynthesis drives soil respiration. Nature 411 :789792.

Kirschbaum MUF. 2004. Soil respiration under prolonged soil warming: arerate reductions caused by acclimation or substrate loss? Global Change Biology 10: 18701877.

Kuzyakov Y. 2002. Review: factors affecting rhizosphere priming effects. Journal of Plant Nutrition and Soil Science 165 : 382396.

Larionova AA, Yevdokimov IV, Bykhovets SS. 2007. Temperature responseof soil respiration is dependent on concentration of readily decomposableC. Biogeosciences 4: 10731081.

Liski J, Ilvesniemi H, Ma kela A, Westman CJ. 1999. CO 2 emissions fromsoil in response to climatic warming are overestimated: the decompositionof old soil organic matter is tolerant of temperature. Ambio 28: 171174.

Lloyd J, Taylor JA. 1994. On the temperature dependence of soil respiration.Functional Ecology 8: 315323.

von Lu tzow M, Ko gel-Knabner I. 2009. Temperature sensitivity of soilorganic matter decomposition what do we know? Biology and Fertility of Soils 46: 115.

Melillo JM, Steudler PA, Aber JD, et al . 2002. Soil warming and carbon-cycle feedbacks to the climate system. Science 298 : 21732176.

Midwood AJ, Millard P. 2011. Challenges in measuring the d13 C of the soilsurface CO 2 efux. Rapid Communications in Mass Spectrometry 25:232242.

Midwood AJ, Thornton B, Millard P. 2008. Measuring the 13 C content of soil-respired CO 2 using a novel open chamber system. Rapid Communications in Mass Spectrometry 22: 20732081.

Millard P, Midwood AJ, Hunt JE, Barbour MM, Whitehead D. 2010.Quantifying the contribution of soil organic matter turnover to forestsoil respiration, using natural abundance d13 C. Soil Biology and Biochemistry 42: 935943.

Pendall E, King JY. 2007. Soil organic matter dynamics in grassland soilsunder elevated CO 2 : insights from long-term incubations and stable iso-topes. Soil Biology and Biochemistry 39: 26282639.

Phillips DL, Gregg JW. 2001. Uncertainty in source partitioning using stableisotopes. Oecologia 127 : 171179.

Phillips RP, Fahey TJ. 2008. The inuence of soil fertility on rhizosphereeffects in northern hardwood forest soils. Soil Science Society of America Journal 72: 453461.

Pinheiro J, Bates D. 2000. Mixed-effects models in S and S-PLUS . New York,NY: Springer-Verlag.

R Development Core Team. 2010. The R project for statistical computing .http://www.r-project.org/ .

Raich JW, Potter CS, Bhagawati D. 2002. Interannual variability in globalsoil respiration, 198094. Global Change Biology 8: 800812.

Robinson D, Scrimgeour CM. 1995. The contribution of plant C to soil CO 2measured using d13 C. Soil Biology and Biochemistry 27: 16531656.

Rustad LE. 2001. A Meta-Analysis of the response of soil respiration, net ni-trogen mineralization, and aboveground plant growth to experimental eco-system warming. Oecologia 126 : 543562.

Scurlock JMO, Hall DO. 1998. The global carbon sink: a grassland perspec-tive. Global Change Biology 4: 229233.

Sitch S, Huntingford C, Gedney N, et al . 2008. Evaluation of the terrestrialcarbon cycle, future plant geography and climate-carbon cycle feedbacksusing ve Dynamic Global Vegetation Models (DGVMs). Global Change Biology 14: 20152039.

Trotter CM, Tate KR, Saggar S, Scott NA, Sutherland MA. 2004. A multi-scale analysis of a national terrestrial carbon budget: uncertainty reduc-tion and the effects of land-use change. In: Shiyomi M, Kawahata H,Tsuda A, Away Y. eds. Global environmental change in the ocean and on land. Tokyo: Terrapub.

Trumbore S. 2000. Age of soil organic matter and soil respiration: radiocar-bon constraints on belowground dynamics. Ecological Applications 10:399411.

Uchida Y, Hunt J, Barbour M, Clough T, Kelliher F, Sherlock R. 2010.Soil properties and presence of plants affect the temperature sensitivityof carbon dioxide production by soils. Plant and Soil 337 : 375387.

Uselman S, Qualls R, Thomas R. 2000. Effects of increased atmosphericCO 2 , temperature, and soil N availability on root exudation of dissolvedorganic carbon by a N-xing tree ( Robinia pseudoacacia L. ). Plant and Soil 222 : 191202.

Wan S, Luo Y. 2003. Substrateregulationof soil respiration in a tallgrass prairie:resultsof a clipping andshadingexperiment. GlobalBiogeochemical Cycles17: 1054.

Zhu B, Cheng W. 2011 a . 13 C isotope fractionation during rhizosphere respir-ation of C 3 and C 4 plants. Plant and Soil 342 : 277287.

Zhu B, Cheng W. 2011 b . Rhizosphere priming effect increases the tempera-ture sensitivity of soil organic matter decomposition. Global Change Biology 17: 21722183.

Graham et al. Roots affect temperature response of heterotrophic respiration258

http://www.r-project.org/http://www.r-project.org/http://www.r-project.org/