lable at ScienceDirect

Soil Biology & Biochemistry 103 (2016) 201e212

Contents lists avai

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lb io

Is the fate of glucose-derived carbon more strongly driven by nutrientavailability, soil texture, or microbial biomass size?

Courtney A. Creamer a, *, Davey L. Jones b, Jeff A. Baldock a, Yichao Rui c,Daniel V. Murphy c, Frances C. Hoyle c, Mark Farrell a, c

a CSIRO Agriculture & Food, PMB2, Glen Osmond, SA 5064, Australiab Environment Centre Wales, Bangor University, Deiniol Road, Gwynedd, LL57 2UW, UKc School of Geography and Environmental Sciences, Institute of Agriculture, University of Western Australia, Crawley, WA 6009, Australia

a r t i c l e i n f o

Article history:Received 10 March 2016Received in revised form8 August 2016Accepted 22 August 2016Available online 31 August 2016

Keywords:14C tracerCarbon mineralizationMicrobial biomassNutrient stoichiometryOrgano-mineral associationsSoil texture

* Corresponding author. Current address: US GeoloRd MS 962, Menlo Park, CA 94025, USA.

E-mail address: courtneycreamer@gmail.com (C.A

http://dx.doi.org/10.1016/j.soilbio.2016.08.0250038-0717/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

Increasing organic matter (OM) in soil promotes the delivery of vital ecosystem services, such asimproving water retention, decreasing erosion, increasing plant productivity, and mitigating climatechange through terrestrial carbon (C) sequestration. The formation of organo-mineral associationsthrough microbial turnover of labile (i.e. easily decomposed) C is a potential pathway of soil C stabili-zation. However, association of added C with mineral surfaces may be impacted by soil clay content and/or by nutrient availability (due to higher microbial C use efficiency). We added 14C labeled glucose as amodel labile substrate together with either ion exchange resin beads (to induce nutrient limitation),water (no additional nutrients), or four increasing concentrations of nitrogen, phosphorus, and sulfur inconstant stoichiometric ratios to nine agricultural soils under the same climate and management butalong a texture gradient from 3 to 40% clay. The soils with 14C-glucose and a nutrient treatment wereincubated for 4 weeks during which the 14C was traced into CO2, microbial biomass, dissolved organic C(DOC), and soil organic C (SOC). Induced nutrient limitation (available C:N ratio around 300:1) reducedmineralization of glucose-derived C, particularly in soils with <15% clay. However, in soils with �15% clay,higher microbial biomass allowed for glucose-derived C mineralization despite nutrient limitation.Alleviating the nutrient limitation (available C:N < 50:1) allowed for greater transformation of added Cinto microbial biomass-C and SOC, particularly in soils with �21% clay, although further additions (downto C:N of 11:1) did not result in greater SOC or microbial biomass formation. Except under conditions ofnutrient limitation (where C:N > 50:1), soil texture and starting microbial biomass size, not nutrientavailability, were the drivers of SOC and microbial biomass formation during the incubation.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

The use of fertilizers containing (N), phosphorus (P), and sulfur(S) to increase food production has made humans a dominantdriver of global nutrient cycling, which will likely continue as weseek to sustainably intensify agricultural productivity (Vitouseket al., 1997; Foley et al., 2011). As the cycling of nutrients in soilsis inextricably linked to the turnover of soil organic carbon (C),adding nutrient fertilizers to soil may alter soil C cycling (G€arden€aset al., 2011). Soil organic matter (SOM) is a large global pool of C

gical Survey, 345 Middlefield

. Creamer).

(>1500 Pg C to depths of 1 m or more) that if mineralized can act asa C source, or if accumulated can act as a C sink, potentially off-setting anthropogenic greenhouse gas emissions (Lal, 2004). As aresult, a thorough understanding of the impact of nutrient avail-ability on soil C turnover is required to predict its potentialresponse to increases in inorganic fertilization.

Carbon is stabilized in soil by a variety of mechanisms. Histori-cally the chemical composition of SOM was thought to dictate itsturnover, with greater stabilization of more complex biopolymersdue to slower rates of decomposition (Derenne and Largeau, 2001).However, recent work emphasizes that microbial access to SOM,rather than chemical composition, controls its turnover (Dungaitet al., 2012). Microbial access to SOM is restricted by C associationwithmineral surfaces and by spatial isolationwithin soil aggregates(Jastrow et al., 2007). The importance of mineral surfaces for SOM

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212202

stabilization has been recognized for decades, with increases inreactive mineral phases corresponding to increases in the potentialfor carbon accumulation (Hassink, 1997; Baldock and Skjemstad,2000; Six et al., 2002).

Although OM sorption and co-precipitation play an importantrole in C protection on mineral surfaces (Kramer et al., 2010; Kleberet al., 2015), microorganisms, as the “eye of the needle” throughwhich SOM is processed (Jenkinson, 1978), are increasingly recog-nized as dominant drivers of SOM formation and longer-term sta-bilization. Improved analytical techniques have established thatmineral-bound OM is predominately derived from microbialproducts (Miltner et al., 2011), with a layering of OM on clay par-ticles where proteins and polysaccharides from microbial residuesaid in additional mineral-OM stabilization (Kleber et al., 2007).Therefore, organo-mineral associations can be formed during OMdecomposition as microbial residues associate with mineral sur-faces (Cotrufo et al., 2013). The quantity of microbial-derivedorganic matter (MOM) formed through this pathway is depen-dent on the amount of MOM produced during decomposition aswell as the capacity of mineral surfaces to protect MOM fromfurther decomposition (Six et al., 2006).

The quantity of MOM formed during decomposition depends onwhether decomposing OM is initially released as CO2 or used tobuild microbial biomass, also known as microbial carbon use effi-ciency (CUE) or microbial growth efficiency (MGE). There isincreasing recognition that changes in CUE due to C substratequality, elevated temperatures, or changes in microbial communitycomposition can alter C stocks in response to climate change(Allison et al., 2010; Frey et al., 2013;Wieder et al., 2014). Therefore,modification of these CUE drivers may result in enhanced soil Caccumulation, even with lower C inputs (Kallenbach et al., 2015).Interesting, the chemical complexity of C inputs appears to dictatethe pathway of soil C formation through this CUE control, withlabile C inputs (i.e. easily degradable C, metabolic plant compo-nents) forming greater quantities of mineral-bound C (i.e. pro-tected) than chemically complex inputs (i.e. structural plantcomponents), potentially due to higher CUE (i.e. greater allocationof substrate-C to microbial biomass-C than to CO2; Bradford et al.,2013; Cotrufo et al., 2015; Haddix et al., 2016).

In addition to substrate quality, nutrient addition can impactOM formation, with higher CUE as nutrient availability increases(Schimel andWeintraub, 2003; Hessen et al., 2004). The increase inCUE with nutrient availability is based upon the concept of nutrientstoichiometry, where microbial biomass has a constrained C:N:P:Sratio (Cleveland and Liptzin, 2007; Kirkby et al., 2011), and sotheoretically more microbial biomass is formed (and less Crespired) as the available C: nutrient ratios approach the ratiosrequired for optimum growth (around 20e25:1 for C:N in terres-trial microorganisms; Manzoni et al., 2012; Sinsabaugh et al., 2013).Operating under this mechanism, (due to high microbial contri-butions to mineral-associated OM) increases in nutrient availabilityshould increase stabilization of C in the finer textured fractions ofsoil (Neff et al., 2002; Cenini et al., 2015). However, the influence ofC complexity and nutrient availability on CUE could be confoundedby other potential drivers, such as changes in microbial communitystructure and function. For example, N deposition can decrease theabundance of C-use efficient oligotrophic bacteria and fungi (Leffet al., 2015) or cause microorganisms to adjust their nutrient useefficiency (Mooshammer et al., 2014), thereby impacting microbialbiomass formation. The timescale over which CUE is measured isalso of critical importance: over longer timescales (e.g. weeks) thisecosystem-level CUE includes the turnover of microbes and mi-crobial residues (Geyer et al., 2016), which may respond morestrongly to climate and nutrient availability (Hagerty et al., 2014;Kaiser et al., 2014) than shorter-term changes that exclude

microbial turnover. Importantly, whether microbial residuesformed from added C remain in soil depends on whether thatbiomass is protected from further decomposition, likely due toassociation with mineral surfaces (Throckmorton et al., 2014).

In this context, the purpose of our experiment was to determinewhether nutrient availability or edaphic properties (specificallyclay content and microbial biomass size) influenced the fate of anadded labile C substrate along a natural field-based gradient in soiltexture (3 to 40% clay). In our experiment, we define labile C asdissolved substrates that can be rapidly assimilated by a largeportion of the microbial community, and we use glucose as a modelcompound. We define stabilized soil C as that which is not miner-alized to CO2 and therefore remains in the soil, with the recognitionthat this definition is dependent upon the timescale of measure-ment (which for this experiment is four weeks). The concept was to(1) add a readily utilized, soluble, and 14C labeled substrate withlimited abiotic sorption potential (glucose) as a tracer at concen-trations high enough to increase microbial biomass, (2) determinewhether nutrient additions changed the allocation of substrate C tobiomass versus CO2 over four weeks, and then (3) assess whetherclay content impacted the quantity of glucose-derived C remainingin the soil as microbial biomass and soil C. We hypothesized thatsoils with higher clay contents would have a greater capacity toprotect formed glucose-derived SOC and microbial biomass fromfurther decomposition, thereby decreasing mineralization ofglucose-derived C as CO2 (independent of nutrient addition). Wealso hypothesized that higher nutrient availability would lead togreater formation of microbial biomass, and that the quantity ofMOM remaining in the soils after four weeks would increase withclay content.

2. Materials and methods

2.1. Soil sampling and site description

Soil samples were taken within a 1-m radius of previouslyidentified points (Murphy et al., 2009) from a field with a naturallyoccurring clay gradient near Dangin in Western Australia (32�50S,117�180E). The region has a semi-arid climate, with hot, dry sum-mers and cool, wet winters (when cropping occurs). Based on 15years of climate data (1997e2014) the area has a mean annualrainfall of 326.5 mm, andmean annual temperature of 25.4 �C. Soilsfrom nine sites across a texture gradient ranging from 3 to 40% claywere sampled when the soil was in a dry state (Austral autumn2014 prior to seeding) within a 10 ha area of the field. Afterremoving plant residues and surface debris, five soil cores of thesurface 0e10 cm layer (Ap horizon) were taken at each site using a10 cm diameter push-in auger and composited. After sampling,soils were stored at 4 �C, air-dried at 40 �C, sieved to <4 mm, andthen characterized and incubated as described below.

At one of the sampling sites (21% clay), 2.5 t ha�1 of gypsumwasapplied in 2008, but otherwise management was consistent acrossthe sites. Inorganic N application rates in this region to winterwheat typically vary from 0 to 100 kg N ha�1, based upon expectedyields (average yields of 1.9 t ha�1 for winter wheat) and antici-pated growing season rainfall (200e400 mm). Nitrogen fertiliza-tion contributes about 20% of total N to wheat cropping systems inAustralia, so that these low input farming systems are highly relianton biological N fixation and SOM mineralization to meet crop Ndemands (Angus, 2001). As a result, microbial turnover of SOM is animportant nutrient supply for plants in this system.

2.2. Soil characterization

Percentage distributions of sand (20e2000 mm), silt (2e20 mm),

Table 2Extractable organic C (DOC), organic N (DON), ammonium-N, nitrate-N, totalphosphorus, and total sulfur (mg kg�1 dry weight soil) from soils used in incubation.

Clay (%) DOCa f DONb f NH4þ- Nc f NO3

�-Nd f Pe Se

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212 203

and clay (<2 mm) were determined by particle size analysis(McKenzie et al., 2002). Variability between duplicate samples was<1%. The soils in this study are referred to by their clay content (%)provided in Table 1. Soil pH and electrical conductivity (EC) weremeasured in a 1:5 (w/v) slurry of soil: water after shaking for 1 hand settling for 20 min. Total soil C and N content of the soils weremeasured by high temperature combustion on a TruMAC CNSanalyzer (Leco Corp; Castle Hill, NSW Australia) after a negativeeffervescence test confirmed a lack of carbonates. Analytical pre-cision for C and N analysis was 0.011 mg N g�1 soil and 0.234 mg Cg�1 soil. Total elemental analysis (Ca, Mg, Na, K, S, P, Al, Cd, Co, Cr,Cu, Fe, Mn, Mo, Ni, Sb, Se, Zn) was conducted using SW-846 EPAMethod 3051A (1998) and analysed by inductively coupledplasma optical emission spectrometry (ICP-OES, Thermo ScientificiCAP 6000 Series). Soil particle size distribution, pH, and total OC, N,P, and S contents are shown in Table 1 (remaining elementalanalysis data shown in Suppl Table 1). Soils were fractionated intofine (�50 mm) and coarse (>50 mm) sized fraction following themethodology of Baldock et al. (2013) and total C and N measuredvia high temperature combustion as previously described (datashown in Suppl Table 2). Mass recovery of soil fractions varied from99.3 to 101.5%.

Soils were extractedwith cold (4 �C) 0.5MK2SO4 (1:5 soil: liquidratio) by shaking for 30 min at 180 rev min�1 followed by centri-fugation (3400 g, 10 min) and filtration (Whatman #42). Dissolvedorganic carbon (DOC) and total dissolved nitrogen (TDN) weremeasured on the extracts using a TOC-V þ TNM-1 analyzer (Shi-madzu, Nakagyoku, Kyoto, Japan). Inorganic N (NO3

� þ NO2�, NH4

þ)was quantified colorimetrically (Mulvaney, 1996; Miranda et al.,2001) using a microplate reader (Synergy MX Biotek; WinooskiVT, USA). Dissolved organic nitrogen (DON) was calculated as thedifference between TDN and inorganic N. Available P and S weremeasured by ICP-OES on refrigerated Mehlich 3 extracts of the soil(Mehlich, 1984; Rayment and Lyons, 2011). Extractable nutrients atthe start of the incubation are shown in Table 2.

3 139 16.8 1.16 1.68 43.9 18.04 170 18.7 1.58 0.96 70.0 35.36 106 14.4 0.27 4.18 135.2 16.98 164 18.4 1.04 3.88 74.9 21.211 119 13.0 1.24 3.47 65.6 20.015 217 21.5 4.40 2.90 86.8 23.221 324 33.1 3.98 3.42 36.1 124.030 156 15.2 4.31 2.47 54.2 20.240 179 20.9 3.82 5.64 29.9 30.7

a Precision 1.41 mg kg�1,b Precision 0.81 mg kg�1,c Precision 0.78 mg kg�1,d Precision 0.24 mg kg�1,e In Mehlich 3,

2.3. Incubation and destructive sampling

Glucose (1 mg glucose-C g�1 soil, [U-14C] at 0.15 kBq g�1 soil)was added along with either Hþ/OH� mixed cation/anion exchangeresin beads (PST-1 resin capsules, UNIBEST Inc., Bozeman, MT USA),water, or one of four nutrient solutions with constant N:P:S ratios(for a total of six nutrient treatments, see Table 3) to the ninetexture gradient soils (Table 1). Five grams of each of the nine air-dried soils were weighed into 50 ml polypropylene centrifugetubes. For the limiting nutrients treatment, the PST-1 resin capsuleswere opened and the resin beads were added at a rate of 0.30 g of

Table 1Properties of the nine agricultural soils used in the incubation study.

Soil Particle sizes (%)a pH ECb (mS cm�1) OCc

Clay Silt Sand

1 3 2 96 6.36 68 5.92 4 2 95 6.34 66 8.83 6 3 91 6.19 88 8.74 8 3 89 5.93 61 8.15 11 8 81 6.19 74 9.66 15 9 76 6.57 113 16.37 21 30 49 8.18 318 22.28 30 7 63 6.77 100 19.09 40 17 43 7.50 177 21.4

a Variability < 1%,b EC electrical conductivity,c OC organic carbon, precision 0.234 mg C g�1,d Precision 0.011 mg N g�1,e <is below limit of detection.

resin beads per 5 g of soil. This addition rate is consistent with themanufacturer's recommended ratio of one resin capsule per 50 g ofsoil. For all nutrient treatments, a stock glucose solution was pre-pared, at twice the concentration and 14C activity required for theexperiment, for each of the nine soils along the texture gradient.This stock glucose solution was diluted with 18 MU water andvarious amounts of a stock nutrient solution (containing N asNH4NO3, P as KH2PO4, and S as K2SO4) to achieve C addition rates of1 mg glucose-C g�1 soil (14C activity of 0.15 KBq g�1 soil) and N, P,and S addition rates relative to C as shown in Table 3. Nutrientswere added in stoichiometric ratios as required to form microbialbiomass (Creamer et al., 2014). Only nutrients released from the soilwere available in the water only treatment (i.e. no nutrient addi-tion), and the ion exchange resin beads were added to inducenutrient limitation by adsorbing any nutrients released from thesoil (Skogley and Dobermann, 1996). The glucose-C addition ratesare similar to the rates of glucose and amino acid additions fromJones and Murphy (2007) used to induce microbial growth insimilar Western Australian soils.

The air-dry soils were wet to 50% water holding capacity (WHC)with the glucose and nutrient solutions (pH ¼ 6.5). The soil WHCincreased with clay content from 0.26 g water g�1 soil in the 3% claysoil to 0.59 g water g�1 soil in the 40% clay soil (data are not shown).After allowing five minutes for the solutions to penetrate the soil, aCO2 trap containing 1ml of 1MNaOH in a 6ml scintillation vial wasplaced on the soil and the vials were capped and incubated at 20 �Cin the dark for four weeks. CO2 traps were exchanged after 2, 4, 12,24, 28, 32, 36, 48, 56, 72, 120, 144, 168, 240, 336, 504, and 672 h.After CO2 collection, 4 ml of Optiphase HiSafe 3 (Perkin Elmer Inc.,

(g kg�1) Nd (g kg�1) C:N Pe (mg kg�1) S (mg kg�1)

0.05 12.4 <100 <2000.07 12.0 104 2080.06 14.8 208 <2000.05 15.8 152 <2000.07 14.2 273 <2000.11 14.3 361 2070.18 12.4 411 3990.14 13.3 297 2560.16 13.5 339 321

f In 0.5 M K2SO4.

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212204

Waltham MA USA) scintillation fluid was added and the vials werecapped, vortexed, andmeasured for 14C activity using aWallac 1404liquid scintillation counter (LSC). Three CO2 traps were left open tothe atmosphere during sampling and measured at each time point;the 14C activity from these blanks was negligible. The soils were notpre-incubated in this study in order to increase the potential thatthe glucose and nutrient solutions would reach soil micropores,potentially allowing for microbial biomass-C stabilization on clayparticles. Although this may impact microbial community struc-ture, microbial biomass-C is more stable during the dry fallowperiod in this region (Gonzalez-Qui~nones et al., 2011), and as thisarea receives low rainfall (<20 mm) during hot summers (dailymaximums typically 30e40 �C), the community should be adaptedto dry-wet cycles (e.g. Lundquist et al., 1999). Moreover, soils fromthis region have consistent gross and net nitrification rates across awide range in water contents (Gleeson et al., 2010) further sug-gesting the community is adapted to dry-wet cycles.

At the end of the incubation, the 5 g of soil was homogenizedand separated for chloroform fumigation (2 g), extraction with0.5 M K2SO4 (2 g), or biological oxidation (1 g). Microbial biomassformed from the added glucose-C was determined with the chlo-roform fumigation-extraction method of Vance et al. (1987) fol-lowed by centrifugation (5 min at 20,800 g) and 14C quantification(LSC after addition of 4 ml HiSafe to 1 ml supernatant, as previouslydescribed). Microbial biomass was not measured on the ion resinbead treatment. The glucose-derived C remaining in solution at endof the incubationwas determined by shaking (1 h at 180 revmin�1)in 0.5M K2SO4 (1:5 soil: liquid) followed by centrifugation (5min at20,800 g) and 14C quantification (by LSC). Glucose-derived Cremaining in the soil was combusted in a biological oxidizer (OX-600; RJ Harvey Instruments, NY). The combusted CO2 was trappedin an Oxosol scintillation fluid (National Diagnostics, Chapel-Hill,NC, USA), and 14C activity was measured by LSC. Soil C derivedfrom glucose was then calculated by subtracting the 14C left insolution and in microbial biomass from 14C in the soil measured bycombustion. On average, 99% of added glucose-C was recovered asmineralized CO2, microbial biomass, and soil C.

2.4. C and nutrient adsorption

2.4.1. Nutrient sorption isothermsNutrient sorption isotherms were conducted to determine the

availability of added nutrients and were based upon a slightlymodified procedure presented in Strahm and Harrison (2007),where the form, concentration, and pH of the nutrient solutionsadded to the soil reflected the conditions of the experiment (NO3

�

and NH4þ at 0e30 mM, KH2PO4 at 0e4 mM and K2SO4 at 0e3 mM;

pH¼ 6.5). Briefly,1ml of the nutrient solutionwas added to 200mgof soil in a 1.5 ml microcentrifuge tube (1:5 w/v soil: liquid ratio).Tubes were shaken on their side at 200 rev min�1 at room

Table 3Addition rates of nitrogen (mg N-NO3

- and N-NH4þ), phosphorus (mg P-PO4

3-), andsulfur (mg S-SO4

2-) per 100 mg glucose-C. The nutrient treatments are given aslimiting (ion exchange resin bead), none (no nutrients), or relative to the C:N ratiosof the added substrates. These addition rates do not include soil-derived C ornutrients.

Nutrient treatment N P S C:N C:P C:S

Limiting 0.00 0.00 0.00 e e e

None 0.00 0.00 0.00 e e e

C:N 43:1 2.33 0.33 0.27 43 300 375C:N 21:1 4.67 0.67 0.53 21 150 188C:N 14:1 7.00 1.00 0.80 14 100 125C:N 11:1 9.33 1.33 1.07 11 75 94

temperature. After 1 h, tubes were centrifuged at 20,800 g for5 min. Nitrate and NH4

þ from the NO3NH4 sorption and in thestarting solutions and P from the KH2PO4 sorption and startingsolutions was quantified colorimetrically using a microplate reader(Murphy and Riley, 1962; Mulvaney, 1996; Miranda et al., 2001).Although nutrient sorption (in particular for P) can continue forweeks before reaching an equilibrium, the initial adsorption pro-cesses to soil are rapid with the majority becoming sorbed withinthe first hour (Barrow and Shaw, 1977; Bolan et al., 1985). For theK2SO4 isotherms a greater volume of soil and extractant (4 g soil:20 ml solution) were used to provide enough solution for S quan-tification by ICP-OES, but otherwise the procedure was identical. Allnutrient sorption treatments were performed in duplicate.

2.4.2. Nutrient sorption to resin beadsThe adsorption of soil available nutrients at the start of the

experiment was determined by weighing 5 g of soil into 50 mlcentrifuge tubes and adding 0.300 g of resin beads (as per the in-cubation described above). The soils and resin beads were mixedthoroughly, wet to 50% WHC, and incubated in the dark for 24 h at20 �C. Twenty-four hours was chosen for nutrient sorption as rapidaccumulation of ions on the resin bead occurs within the first day(Yang et al., 1991), and our aim was to quantify the adsorption ofavailable nutrients at the start (not throughout) the experiment forcomparison to the water only and nutrient addition treatments.After 24 h, the soils with beads were extracted with cold (4 �C)18 MUwater (1:5 soil: liquid ratio) by shaking for 30 min at 180 revmin�1 followed by centrifugation (3400 g, 10 min) and filtration(Whatman #42). Total inorganic nitrogen (NO3

�-N, NH4þ-N) was

quantified colorimetrially as previously described, and P and S werequantified by ICP-OES. Paired soils without beads were treatedsimilarly.

2.4.3. Glucose sorptionThe potential for the sorption of 14C labeled glucose to the soil

and ion exchange resin beads was determined for the concentra-tions used in the experiment (50e100 mM). One ml of a solutioncontaining 0.5 KBq ml�1 14C[U] labeled glucose and 50 mM of so-dium azide as a microbial inhibitor was added to 0.500 g soil in a2 ml microcentrifuge tube. The solution was shaken (60 min at 200rev min�1), centrifuged (20,800 g for 5 min), and the supernatantmeasured for 14C activity using LSC (Tri-Carb 3110 TR scintillationcounter, Perkin Elmer Inc.) after the addition HiSafe 3 andcompared to the 14C activity of the added solutions. A similarsorption experiment with the addition of 30 mg of resin beads to0.500 g soil was also conducted. Both sorption experiments wereconducted in duplicate. Similar to previously published studies (e.g.Kuzyakov and Jones, 2006), there was no significant sorption of 14Cglucose to any of the soils with or without beads, with between 92and 108% of the added glucose recovered (average 99% recoverywith beads and 101% recovery without beads; Suppl Table 3).

2.5. Microbial community

Microbial biomass yields and community structure was deter-mined on the air-dried soils using the phospholipid fatty acid(PLFA) procedure of White et al. (1979) as modified by Frostegårdet al. (1991). Briefly, 6 g of soil were extracted in a 1:2:0.8 ratio(v) of chloroform, methanol, and citrate buffer (0.15 M, pH 4.0) bysonication for 2 h. Phospholipids were eluted from the isolatedorganic phase with methanol through Supelclean™ silica columns(Sulpelco, Bellefonte PA USA). Fatty acid methyl esters (FAMES)formed from PLFA with mild alkaline methanolysis were re-suspended in an internal standard (methyl decanote) and quanti-fied relative to a 26-compontent standard (bacterial acid methyl

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212 205

esters [BAME] mix, Sulpelco) on an Agilent 7890B/5977A gaschromatograph/mass spectrometer equipped with a 30 m HP5-5MS column (Agilent Technologies Inc; Santa Clara, CA USA). TheGC program used a 1:10 split injection with a 1.2 ml min�1 columnflow and the following oven program: hold 50 �C for 1 min, ramp at10 �C min�1 to 150 �C, ramp at 5 �C min�1 to 240 �C, hold 240 �C for5 min (46 min runtime). Individual PLFA are named using standardnomenclature (Zelles, 1999). All PLFA extractions were performedin duplicate. Relative abundances (mol %) of individual PLFA werecalculated by converting measured PLFA yields (ng mL�1) into molaryields (nmol g�1 soil, taking into account FAME molecular weight,sample re-suspension volume, and the weight of soil extracted),and normalizing to total PLFA yields.

2.6. Calculations

If sorption to soil occurred, the NO3�, NH4

þ, PO43�, and SO4

2�

sorption isotherms were best described by a Freundlich equation:

q ¼ Kd � C1=n (1)

where q is the amount of adsorption (mmol adsorbate kg�1 soil), Kd

is the adsorption coefficient (or Freundlich equilibrium constant), Cis the equilibrium concentration of the adsorbate (in mmol adsor-bate L�1), and n is the correction factor. As n approaches 1, thecurves become increasingly linear. Nonlinear curve fitting for theFreundlich isotherms (Eq (1)) was performed in SigmaPlot (v12.3).

As the 14C glucose was added in concentrations to stimulatemicrobial growth, the C mineralization curves were predominatelysigmoidal, with a lag and exponential growth portion followed bytraditional mono- or biphasic exponential decay. A number ofmathematical models have been developed to separately modelboth the exponential growth (Anderson and Domsch, 1978;Blagodatsky et al., 2000) and biphasic exponential decay (withand without lag periods; Scow et al., 1986; Boddy et al., 2007)portions of CO2 mineralization that can be related to parameters ofmicrobial growth, biomass production, and biomass turnover.However, due to our discrete sampling and the large differences inC mineralization during microbial growth between treatments, wemodeled the entire mineralization curve with a combined logisticand biphasic first order exponential model (Gillis and Price, 2011),which calculates the transition (in h, tL) from exponential growth toexponential decay:

Cm ¼ CL�1þ e�kL�½t�tL��þ Cf �

�1� e½�kf�t��þ Cs �

�1� e½�ks�t�

�

(2)

where Cm is the cumulative % glucose-derived CO2 mineralized attime t (in h), C refers to pool sizes (% of total glucose-derived C), krefers to decay rates (in h�1), and the subscripts L, f, and s refer tologistic, fast, and slow portions of the glucose-derived C minerali-zation curve, respectively. For the logistic portions of the curve, halfof the curve (where t < tL) corresponds to exponential microbialgrowth, while the second half transitions into biphasic exponentialdecay. Therefore, where t < tL, kL is equivalent to the microbialgrowth rate, CL is proportional to the quantity of microbial biomassproduced during growth (assuming CO2 production was directlyproportional to growth), and tL indicates the duration of microbialgrowth (similar to parameters derived from exponential growthmodels). As a result, in this manuscript whenwe discuss the logisticphase of the curve, we will refer to it in terms of the microbialgrowth phase, although for the portions where t > tL, the logisticfunction includes part of exponential decay that would traditionally

be included in the fast pool in a simple two-pool biphasic miner-alization model. The fast and slow pools in our model are thereforesimilar (but not directly equivalent) to modeled pools sizes in cu-mulative mineralization without growth, due to inclusion of aportion of exponential decay within the logistic function.

Nonlinear curve fitting for cumulative glucose-derived C respi-ration was performed with the Solver Add-in in Microsoft Excelusing the GRG Nonlinear Algorithm and user-defined curves (Eq(2)). To determine whether a monophasic (where Cf and kf ¼ 0) orbiphasic exponential decay model better described the latter por-tions of the curve, the difference in the sum of squares of modeledversus measured values was used to calculate an F-statistic, andsubsequently a P-value, to determine if adding a second poolimproved the curve fit significantly. The biphasic model (where Cfand kf s 0) was used only if it significantly (P < 0.05) improvedcurve fit. The biphasic model was used for all treatments except inthe �11% clay soils with ion exchange resin bead addition. Duringcurve fitting, the size of Cs was set to 100-(CL þ Cf) and kf wasdefined as greater than ks. Linear and non-linear curve fitting of Cheld in the isolated soil fractions and glucose-derived C (%)measured as SOC relative to clay content were conducted in Sig-maPlot (v12.3).

2.7. Statistics

Univariate analyses of starting total microbial biomass yields(from PLFA), curve fit parameters from Equation (2) (CL, kL, t0, Cf, kf,Cs, ks) and glucose-derived C respired as CO2, measured inmicrobialbiomass, remaining in solution, and measured as SOM were con-ducted with two-way ANOVA in GenStat (16th edition) with thenine soils (3e40% clay) and six nutrient treatments (limiting, none,C:N 43:1, C:N 21:1, C:N 14:1, or C:N 11:1) as the between subjectfactors. As soils under the nutrient limiting treatment with �11%clay were better fit by the monophasic mineralization model(where Cf and kf ¼ 0 in Equation (2)), they were not included in thestatistical analysis of the biphasic curve fit parameters (i.e. themajority of the data). Datawere log-transformed prior to analysis tomeet assumptions of normality and equal variance where neces-sary. Differences between nutrient availability with and withoutresin beads (no nutrient addition) were analysed with two-wayANOVA in GenStat, with soil type and bead addition (yes or no)as between subject factors, after log transformation. Statisticalsignificance was set at a < 0.05, and main effects are discussed onlyif interactions were not significant. Linear regressions of glucose-derived C respired as CO2 relative to the log of soil clay contentwere conducted in GenStat. Principal components analysis (PCA) tocharacterize the measured initial soil parameters and microbialcommunity composition (PLFA relative abundance) were con-ducted in PRIMER with the PERMANOVA þ add-on (version 7). Agroup average cluster analysis in PRIMER was used to determinesignificant (P ¼ 0.05) groupings of the soils based upon initial soilparameters and microbial community composition.

3. Results

3.1. Soil characterization

The majority of measured initial soil properties were influencedby soil texture, although concentrations of K2SO4 or Mehlich-3extractable C, N, P, and S (excepting ammonium) showed noapparent relationship to clay content (Table 2). Increases in claycontent were concurrent with increases in the following soilproperties: pH, EC, % silt, total SOC, and total soil N, P, S (Table 1),NH4

þ-N (Table 2), Al, Ca, Co, Cu, Fe, K, Mg, Mn, Na, Ni, and Zn (SupplTable 1). For nearly all of these properties (except for Fe and Na,

Fig. 1. Relationship between the clay content of the soils and (a) the concentration of Cin the fine (�50 mm) and coarse (>50 mm) SOM fractions, and (b) whole soil C storagein the fine and coarse SOM fractions. Linear and non-linear curve fits are shown.

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212206

which increased linearly with % clay), the maximum levels werereached at 21% clay, and then stabilized. As a result, based upon themeasured initial soil properties the soils clustered significantly(P ¼ 0.05), with the lowest (3 and 4%), intermediate (6, 8, 11, 15%),highest (30 and 40%), and then the 21% clay content comprisingfour distinct groups (Suppl Fig. 1). The only properties to decreasewith clay content were the weight % (by mass) of the coarse SOMfraction, and fine SOM fraction C and N concentrations (SupplTable 2, Fig. 1a). Despite mass losses with clay content, total soil Cstorage in the coarse fraction increased due to increases in C con-centration (Fig. 1a,b). There were also increases in total soil C storedin the fine SOM fraction as clay content increased, due to the in-crease in fine SOM fraction mass (despite decreases in Cconcentration).

Total microbial biomass, calculated as the sum of all PLFA,increased with clay content, although the 21% clay soil (also the soilwith the legacy gypsum application) had the highest biomass yields(Fig. 2a). Microbial community composition (as determinedthrough changes in the relative abundance [mol %] of PLFA) wasalso strongly impacted by soil texture, with a relative decrease ingeneral PLFA (C16) and Gram-positive bacteria (i15, i16), and arelative increase in Gram-negative bacteria (18:1u7c, 16:1u7c,18:1u9c) with increasing clay content (Fig. 2b). Microbial com-munity composition was the most distinct in the 21% clay soil,while the 15, 30, and 40% clay soils formed a significant (P ¼ 0.05)group from the 4e11% and 3% clay soils.

3.2. Nitrate, ammonium, phosphate and sulfate adsorption

3.2.1. Nutrient adsorption to soilsTherewas sorption of NH4

þ-N and PO43--P, but not NO3

�-N or SO42--

S, to the soils (Suppl Fig. 2aed). For both NH4þ-N and PO4

3--P, withincreasing clay content the curves became more linear (n inEquation (1) approached a value of 1; increasing with NH4

þ-N anddecreasing with PO4

3--P) and the adsorption coefficient (Kd)increased, indicating stronger sorption at higher clay contents(Suppl Table 4).

3.2.2. Adsorption to ion exchange resin beadsIn all the soils, the ion exchange resin beads significantly

decreased the concentrations of water-extractable NO3�-N and S,

sorbing between 70 and 95% of NO3�-N and 76e96% of S after 24 h

(Table 4). In contrast, NH4þ-N and P decreased with bead addition

only at the lowest one (P) or two (NH4þ-N) clay contents, and P was

higher in the 15e30% clay soils with resin bead addition. Similar tothe patterns observed in the 0.5 M K2SO4 and Mehlich 3 extracts ofthe soils (Table 2), water-extractable NH4

þ-N and NO3�-N increased

with clay content, P was variable, and S was highest at 21% clay.Despite the significant decrease in available N, P, and S, there werestill significant increases with clay content for all the measurednutrients in the treatments with resin bead addition.

Due to the differences in (a) adsorption of added NH4þ-N and

PO43--P to the soils (Suppl Fig. 2), (b) differences in native water-

extractable (defined here as available) N, P, and S among the soils(Table 2), and (c) differences in the adsorption of native available N,P, and S with the ion exchange resin bead additions (Table 4), theratios of N, P, and S relative to added glucose-C were not identical tothe addition rates. Defining the water-extractable nutrients as“available” upon the start of the incubation, and any N, P, or Sadsorbed to the soils or to the beads as “unavailable”, we calculatedthe ratios of added glucose-C relative to available N (as NH4

þ-N andNO3

�-N), P, and S across the soil types and nutrient treatments(Table 5). Although there were trends of decreasing C:N and C:Sratios with clay content, the biggest differences in the C: nutrientratios were driven by the nutrient treatments. Similarly, C:P ratios

were also driven predominantly by nutrient treatment, but thetrends with clay content were minor.

3.3. Glucose-derived C transformations

On average, between 94 and 106% of glucose-derived C wasrecovered as CO2 or in the soil, microbial biomass, or soil solution.At the end of the four-week incubation, there was a significantinteraction (P < 0.001) between clay content and nutrient treat-ment on the glucose-derived C remaining in solution (i.e. extract-able with 0.5 M K2SO4). Excluding the soils that received the ionexchange resins, on average only 0.24% of the added glucose-derived C remained in solution at the end of the experiment, andthis concentration did not vary significantly with nutrients or claycontent (Suppl Fig. 3). However, with resin bead addition (i.e.limiting nutrient treatment) the glucose-derived C remaining insolution was significantly higher than the other nutrient treat-ments, and increased significantly from the 15e40% clay soils(3.10 ± 0.18% of added 14C) to the 3, 6, 8, and 11% clay soils(14.01 ± 1.00%) and finally to the 4% clay soil (22.74 ± 2.21%).

There was a significant interaction (P < 0.001) between nutrienttreatment and clay content on glucose-derived C respired duringthe incubation. Across all nutrient treatments (except the limitingnutrient treatment) the log of the soil clay content was a goodpredictor (R2 of 0.87e0.97) of cumulative glucose-derived Cmineralization, with progressively lower mineralization as claycontent increased (Fig. 3). However, in the limiting nutrient treat-ment, C mineralization increased with clay content, particularlyfrom the �11% to the �15% clay soils. Soils with the limiting

Fig. 2. Change in (a) total microbial biomass yields and (b) microbial community composition, as impacted by clay content. Symbols indicate significant (P ¼ 0.05) clustering. Thevariation explained by the principal components is shown in parentheses. PLFA with correlations >0.30 are shown; the length of the line is proportional to the strength of thecorrelation.

Table 4Concentrations of NH4

þ-N, NO3�-N, phosphorus (P,) and sulfur (S) in water extracts of the soils with and without ion exchange resin beads. Standard errors are shown in

parentheses. All interactions were significant. An asterisk indicates significant effects of the resin bead addition within a soil, while letters indicate significant differencesbetween soils with or without ion exchange resin beads. An asterisk by the legend for NO3

�-N and S indicates that the resin beads decreased concentrations in all soils.

Clay (%) NO3�-N mg kg soil�1 NH4

þ-N mg kg soil�1 P mg kg soil�1 S mg kg soil�1

No beads With beads* No beads With beads No beads With beads No beads With beads*

3 3.8 a(0.52)

0.17 a(0.01)

0.74 b(0.12)

0.18 a*(0.03)

9.0 de(0.82)

4.9 a*(0.54)

32.9 b(3.4)

3.21 ab(0.42)

4 3.0 a(0.09)

0.16 a(0.06)

0.60 ab(0.12)

0.04 a*(0.04)

5.6 bc(0.04)

2.1 a(2.10)

28.6 ab(0.5)

7.59 cd(5.87)

6 7.6 cd(0.72)

1.58 bc(0.33)

0.60 ab(0.24)

0.75 cd(0.01)

11.5 e(0.94)

9.3 c(0.29)

18.6 a(1.3)

2.70 ab(0.09)

8 5.4 b(0.21)

1.54 bc(0.48)

1.25 cd(0.34)

1.02 d(0.30)

4.8 ab(0.08)

5.7 a(2.09)

23.2 ab(1.2)

3.19 a(0.93)

11 7.8 cd(0.06)

2.25 c(0.64)

0.67 ab(0.04)

0.91 cd(0.30)

6.9 cd(0.03)

6.4 ab(0.13)

32.2 b(1.6)

2.41 a(0.27)

15 6.3 bc(0.01)

1.66 bc(0.88)

0.86 bc(0.11)

0.59 bc(0.24)

9.4 de(0.01)

14.3 d*(1.58)

31.3 b(1.5)

5.49 c(0.11)

21 15.7 e(0.33)

1.31 b(0.07)

0.42 a(0.17)

0.38 ab(0.01)

3.8 a(0.01)

9.1 c*(0.30)

234.0 d(0.7)

10.03 de(0.03)

30 10.1 d(0.24)

2.27 c(0.44)

1.37 d(0.14)

1.88 e(0.07)

4.4 ab(0.07)

7.9 bc*(0.49)

31.1 b(0.9)

5.01 bc(0.87)

40 16.9 e(1.13)

3.99 d(0.58)

2.80 e(0.31)

3.54 f(0.17)

8.3 d(0.11)

10.9 cd(2.67)

59.0 c(3.7)

14.15 e(1.49)

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212 207

nutrient treatment also had significantly less C respired than allother nutrient treatments. For some soils, the no nutrient treat-ment (C only) had less C respired than all other additions (at 30, 40%clay) or relative to the highest addition treatment only (C:N 11:1)with other treatments intermediate (at 6, 8, 11, 15% clay).

The temporal mineralization of glucose-derived C to CO2 by thesoil microbial community was well-described by the logistic and

Table 5Ratios of carbon (C) to nitrogen (N), phosphorus (P), and sulfur (S) in the soils in responnutrients at the start of the incubation. Nutrient treatments are given as limiting (ion exchain Table 3).

Clay (%) Limiting None C:N 43:1

N P S N P S N P

3 2873 204 314 222 112 31 36 834 5014 476 132 279 177 35 37 1146 429 107 370 121 87 54 34 808 392 176 315 150 208 43 34 13111 313 156 410 117 143 31 34 10615 439 69 180 139 106 32 40 8621 588 109 99 62 262 4 43 19330 243 127 202 88 227 32 44 16040 131 91 70 50 118 17 28 109

exponential decay equation for all soils (Equation (2); Suppl Fig. 4).For all soils and treatments, except the nutrient limiting treatmentin �11% clay soils, the biphasic model (with kf and Cf s 0) fit themineralization data significantly better than themonophasicmodel(with kf and Cf ¼ 0). For all optimized curve fit parameters, therewere significant interactions between nutrient treatment and claycontent on the growth portions of the curve (CL, kL, tL) but

se to nutrient treatments, calculated based upon added, extractable, and adsorbednge resin beads), none (C only), or by the C:N ratio of the added substrates (as shown

C:N 21:1 C:N 14:1 C:N 11:1

S N P S N P S N P S

28 20 65 26 14 54 25 11 46 2332 20 83 29 14 65 27 11 54 2547 20 70 42 14 61 37 11 54 3439 20 94 35 14 73 32 11 59 3028 20 82 26 14 67 25 11 56 2329 23 70 27 16 58 25 12 50 244 28 136 4 20 103 4 15 82 430 27 120 28 19 95 26 15 78 2416 21 93 15 17 81 15 14 70 14

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212208

significant main effects (of both nutrient treatment and clay con-tent) without interactions for latter portions of the curve (Cs, ks, kf)with only clay content significantly impacting Cf (Fig. 4; SupplTable 5). In general, nutrient addition increased the rate (kL) andduration (tL) of microbial growth, as well as the amount of CO2released (in soils with �15% clay; CL). Increasing clay contentdecreased CO2 release (CL) and the duration of microbial growth (tL)but progressively increased the growth rate (kL) from 3 to 15% clay,reaching maximum values at 21e40% clay.

As previously mentioned, the treatments with appreciableamounts of glucose-derived C left in solution (limiting nutrients in�11% clay; Suppl Fig. 3) had no modeled fast pool. Among thetreatments that did have a modeled fast pool, it comprised asignificantly smaller proportion of added glucose-derived C in the3% clay soils (25 ± 3.3% of added C) relative to all other soils(average of 30 ± 2.4% of added C), and in the limiting nutrientaddition (17 ± 2.1% of added C) relative to the other nutrienttreatments (average of 29 ± 2.4% of added C) (Suppl Table 5). Thefast pool decay rate (kf) increased progressively with clay contentand nutrient addition. For the slow pool, there was a progressiveincrease in the size of the slow pool (Cs) with clay content across allbut the liming nutrient addition treatment, where there was adecrease from lower (�11%) to higher (�15%) clay soils. The size (Cs)and decay rate (ks) of the slow pool also generally decreased withincreasing nutrient addition.

3.4. Glucose-derived C fate

The percentage of glucose-derived C measured as chloroform-extractable microbial biomass was significantly impacted by aninteraction between nutrient addition and clay content (P ¼ 0.013;Fig. 5a). Similar to starting microbial biomass size (Fig. 2a), acrossall nutrient addition levels, the 2e15% clay soils had significantlyless chloroform-extractable glucose-derived C than the 21e40%clay soils. If a KEC (correction factor for the extractable portion ofmicrobial biomass C) of 0.28 was applied to this data to account forextraction efficiency (Glanville et al., 2016), microbial biomassformed from glucose-derived C after 4 weeks of incubationincreased from approximately 10% of added C in the�15% clay soilsto 18% of added C in the�21% clay soils. In the 30 and 40% clay soils,the no nutrient addition treatment had significantly more glucose-

Fig. 3. Total glucose-derived C respired during the incubation, relative to the log ofclay content. Values are means ± standard errors. Nutrient treatments are given aslimiting (ion exchange resin bead addition), none (C only), or as the C:N ratio of addedsubstrates as shown in Table 3.

derived chloroform-extractable C than the other addition treat-ments (excluding the nutrient limiting treatment, which was notmeasured). No other soils had significant differences in glucose-derived chloroform-extractable C in response to nutrient addi-tion. At the end of the incubation, glucose-derived C measured inthe soil was significantly (P ¼ 0.008) influenced by an interactionbetween clay content and nutrient treatment (Fig. 5b). Similar tonative-C stored on fine particles (<50 mm; Fig. 1b), glucose-derivedSOC increased with clay content, with the highest clay contents (30and 40%) having the most glucose-derived SOC, although the sig-nificance of this relationship varied based upon nutrient treatment(Fig. 5b, Suppl Table 7). The differences between nutrient treat-ments were smaller and more variable, with the limiting treatmenttypically differing from the other nutrient addition treatments, butno other broad trends were significant.

4. Discussion

4.1. Glucose-derived C fate under nutrient limitation

With the addition of the ion exchange resin beads, potentiallyavailable N, P, and Swas adsorbed from solution (Table 4), creating anutrient limited state intended to inhibit microbial growth andglucose-derived C decomposition. The high starting C:N:P:S ratiofor the limiting nutrient treatment of approximately 300:1:3:2(Table 5) was much higher (particularly for N) than the ratio of75:8.75:1.25:1 of microbial biomass (Cleveland and Liptzin, 2007;Kirkby et al., 2011; Creamer et al., 2014), and likely prevented mi-crobial growth due to stoichiometric constraints (Commichau et al.,2006). This was confirmed by the slow growth rate and low mi-crobial biomass production during the long period of microbialgrowth (Fig. 4). By preventing microbial growth this nutrient lim-itation also prevented the mineralization of glucose-derived C(Fig. 3, Suppl Fig. 3), and resulted in a significant portion of added14C remaining in solution at the end of four weeks, particularly inthe soils with a smaller startingmicrobial population (i.e.�11% claysoils; Fig. 2a).

The no nutrient (C only) treatment also represented a nutrientlimited state. According to stoichiometric theory, microbes transi-tion from nutrient limitation to optimum growth (i.e. from Nimmobilization to N mineralization) at a critical substrate C:Nthreshold ratio of about 20e25 (Schimel and Weintraub, 2003;Sinsabaugh et al., 2013), which is reflected in the release ofmineralizable N when the C:N ratio of available substrates fallbelow this threshold (Manzoni et al., 2008; Murphy et al., 2011). Inall soils in this experiment, the C:N ratios of available substrates inthe no nutrient treatment were well above this threshold; greaterthan 100:1 in soils with �15% clay and greater than 50:1 in thehigher clay (�20%) soils. Similar to the limiting nutrient treatment,the no nutrient treatment had slow microbial growth and lowbiomass production during the longer growth phase, as well asslower glucose-derived C turnover in later stages of the incubationrelative to the nutrient additions (Fig. 4). Interestingly, in bothnutrient limited states the legacy of a larger starting microbialbiomass size, in soils with �15% clay, allowed for greater glucose-derived C mineralization. However, in the �11% clay soils nutrientlimitation decreased glucose-derived C turnover by preventingadditional microbial biomass formation.

Compared to the nutrient addition treatments (which representagricultural systems receiving periodic applications of fertilizer),the limiting and no nutrient addition treatments are more similarto natural systems, where productivity is frequently constrained bynutrient limitation (LeBauer and Treseder, 2008). The ion exchangeresins themselves are used to proxy plant-available nutrients, andare designed to mimic the action of a plant root (Skogley and

Fig. 4. Curve fit parameters of glucose-derived C mineralization in relation to clay content (x-axis) and nutrient treatments (lines). To improve figure clarity, significance wasomitted and statistically similar nutrient treatments were combined (see Suppl Table 5 for individual values and significance). Nutrient treatments are given as limiting (ionexchange resin beads), none (C only), or by the C:N ratio of the added substrates (Table 3). Values represent means ± standard errors.

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212 209

Dobermann, 1996). If the resin bead nutrient limitation is moreanalogous to conditions with plants, C and nutrient dynamicsobserved in soil-only incubations may be substantially differentthan in the field (Oburger and Jones, 2009), although these differ-ences are not always observed (Glanville et al., 2012). Relative toour laboratory experiments, the competition between microor-ganisms and plants for nutrients (Bardgett et al., 2003), as well astemporal and spatial variations in nutrient availability and/or mi-crobial biomass (Cambardella et al., 1994; Wardle, 1998), also mayalter microbial turnover of C with added nutrients in situ. Ourexperiment was unable to capture these more complex dynamics,but consistent with stoichiometric theory it did show that thetransition from a nutrient limiting to a nutrient replete system(although transitioning near a C:N threshold of roughly 50:1) was astrong driver of C mineralization.

Interestingly, we also observed higher glucose-derived micro-bial biomass in the high clay soils under nutrient limitation. Similardecreases in microbial biomass have been observed with Nenrichment in field studies (M€annist€o et al., 2016; Riggs andHobbie, 2016), although microbial biomass increases have alsobeen reported (Cusack et al., 2011). Lower microbial biomass undernutrient addition could result from changes in microbial commu-nity composition (Leff et al., 2015), decreases in pH from N addition(H€ogberg et al., 2006), or differences in living biomass turnoverrates (Strickland and Rousk, 2010). However, whether these trendswould continue beyond the four weeks of incubation is unclear,particularly due to the higher turnover rate of the slow pool undernutrient limitation (Fig. 4), suggesting the potential for furtherglucose-derived C mineralization.

4.2. Glucose-derived C fate under nutrient availability

While this experiment was designed to study the importance ofclay content, by no means was that the only measured soil propertythat was altered along the texture gradient. However, the C:N:P:Sratios of the treatments were more strongly impacted by nutrientaddition than clay content (Table 5), so that differences in glucose-derived C turnover between the soils can be predominatelyattributed to the combined effects of clay content and starting

microbial size and composition. Taking this approach, our resultsshow that except under conditions of nutrient limitation (i.e. C:Nratios >50:1), variations in these two properties altered partition-ing of glucose-derived C between respired and stabilized C (at theend of 4 weeks as microbial biomass and SOC), while nutrientavailability did not (Figs. 3e5).

With increasing clay content (in all but the nutrient limitingresin bead treatments, which have been discussed), we observedlower overall glucose-derived C mineralization and higher glucose-derived C transformation to microbial biomass and SOC at the endof the incubation (Figs. 3 and 5). As there was no abiotic sorption ofglucose to the soils (Suppl Table 3), lower glucose-derived Cmineralization with higher soil clay content resulted from greatertransformation and stabilization of MOM during the incubation.Lower mineralization of MOM formed from glucose-derived C withincreasing clay content is consistent with the concept that C (and inparticular MOM) is physically protected on clay particles (Baldockand Skjemstad, 2000; Miltner et al., 2011). In these soils, C con-tent of the fine fraction (�50 mm), increased with its proportionalabundance (Fig. 1b), similar to trends observed across other sites(Hassink, 1997; Six et al., 2002). The stabilization of added glucose-derived C, as microbial biomass and as SOC, was also positivelyrelated to the clay content of the soils (Fig. 5), confirming therelationship between fine particle C content and storage of MOM.

We also observed differences in the modeled growth phaseparameters in response to clay content and nutrient addition(Fig. 4). If CO2 mineralized during the growth phase was propor-tional to the size of formedmicrobial biomass (Wutzler et al., 2012),increases in nutrient availability increased microbial biomass for-mation (Schimel and Weintraub, 2003; Manzoni et al., 2012),particularly in the soils with a smaller initial biomass size (�15%clay). However, this biomass was mineralized to CO2 in the lowerclay soils during the incubation, and therefore was not stabilizedacross the timescale of the incubation. Alternatively, glucose-derived C mineralization could have been decoupled from growth(Blagodatsky et al., 2000), so that microbial biomass was notformed. Regardless of the mechanism, when not under nutrientlimitation (C:N < 50:1) the further addition of available nutrients(down to C:N of 11:1) did not lead to greater stabilization of added

Fig. 5. Relationship between the soil clay content (%) and glucose-derived C (%)measured (a) as chloroform-extractable microbial biomass (no KEC factor applied) and(b) as SOC at the end of the incubation. Values represent means ± standard errors(n ¼ 4). Nutrient treatments are given as limiting (ion exchange resin beads), none (Conly), or by the C:N ratio of the added substrates. Significant differences in glucose-derived chloroform-extractable C between clay contents for all nutrient additiontreatments are indicated by a capital letter, and significant differences betweennutrient treatments within clay content are indicated by symbols. Significant differ-ences in glucose-derived SOC are shown in Suppl Table 7 and curve fit parameters inSuppl Table 6.

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212210

glucose-derived C in soil C or microbial biomass during the incu-bation. Therefore the quantity and rate of glucose-derived Cmineralization was dependent upon the capacity of the soil toprotected glucose-derived C on fine particles, as well as startingmicrobial biomass size.

5. Summary

Recent work has emphasized the importance of including mi-croorganisms in soil C cycling models to accurately project theresponse of soil C to climate change (Wieder et al., 2013; Hararuket al., 2015). Although there is evidence that increasing C lability(Bradford et al., 2013; Frey et al., 2013) and nutrient availability(Manzoni et al., 2012) will increase microbial biomass formation atthe expense of CO2 production, there is less empirical evidence as tothe extent and necessary conditions for the stabilization of formedmicrobial-derived OM (Throckmorton et al., 2012, 2014). Theimportance of soil texture for C protection has long been recog-nized (Hassink, 1997; Baldock and Skjemstad, 2000) and ourshifting conceptual understanding of C stabilization in soils em-phasizes the formation of organo-mineral complexes through

microbial decomposition of easily degradable C (Kleber et al., 2007;Cotrufo et al., 2013). Correlations of soil C storage with mineralogy(Doetterl et al., 2015), the temperature sensitivity of organo-mineral complexes to climate change (Conant et al., 2011), andincorporation of organo-mineral associations into soil C models(Tang and Riley, 2015) all emphasize the importance of mineralogyand soil texture to global soil C cycling.

Our research has added to this understanding by examiningmicrobial transformation of added glucose (a proxy for labile C)along gradients of soil texture, microbial biomass size, and nutrientavailability. In particular, we show that alleviating nutrient limita-tion (at C:N ratios > 100) can increase microbial transformation ofglucose-derived C on short-term scales (4 weeks), potentiallyleading to greater downstream stabilization of soil C. However,further increasing nutrient availability (C:N ratios < 50) did not leadto greater C stabilization during the experiment, likely due togreater decomposition of formed microbial biomass. Edaphicproperties, in particular clay content and microbial biomass size,significantly altered the partitioning of added glucose-derived Cbetween respired and stabilized pools (i.e. SOC and microbialbiomass). As a result, newer microbially-explicit models thatincorporate metrics of changing CUE on soil C stocks should alsoinclude the dynamics of MOM stabilization as dictated by initialmicrobial biomass size and soil texture, particularly under non-limiting nutrient conditions.

Acknowledgements

Thanks to Janine McGowan (CSIRO) for assistance with soilfractionation and soil carbon and nitrogen measurements, andthanks to Kirsty Brooks (UWA) for help with soil sampling. Thiswork was funded by a CSIRO OCE post-doctoral fellowship and aCSIRO Land and Water Capability Development Fund award toC.A.C. M.F. was supported by a CSIRO OCE Julius Career Award.D.V.M. was supported by an Australian Research Council FutureFellowship (FT110100246).

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.soilbio.2016.08.025.

References

Allison, S.D., Wallenstein, M.D., Bradford, M.A., 2010. Soil-carbon response towarming dependent on microbial physiology. Nature Geoscience 3, 336e340.

Anderson, J.P.E., Domsch, K.H., 1978. A physiological method for the quantitativemeasurement of microbial biomass in soils. Soil Biology and Biochemistry 10,215e221.

Angus, J.F., 2001. Nitrogen supply and demand in Australian agriculture. AustralianJournal of Experimental Agriculture 41, 277e288.

Baldock, J.A., Sanderman, J., Macdonald, L.M., Puccini, A., Hawke, B., Szarvas, S.,McGowan, J., 2013. Quantifying the allocation of soil organic carbon to bio-logically significant fractions. Soil Research 51, 561e576.

Baldock, J.A., Skjemstad, J.O., 2000. Role of the soil matrix and minerals in pro-tecting natural organic materials against biological attack. Organic Geochem-istry 31, 697e710.

Bardgett, R.D., Streeter, T.C., Bol, R., 2003. Soil microbes compete effectively withplants for organic-nitrogen inputs to temperate grasslands. Ecology 84,1277e1287.

Barrow, N.J., Shaw, T.C., 1977. The slow reactions between soil and anions: 7. Effectof time and temperature of contact between an adsorbing soil and sulfate. SoilScience 124.

Blagodatsky, S.A., Heinemeyer, O., Richter, J., 2000. Estimating the active and totalsoil microbial biomass by kinetic respiration analysis. Biology and Fertility ofSoils 32, 73e81.

Boddy, E., Hill, P.W., Farrar, J., Jones, D.L., 2007. Fast turnover of low molecularweight components of the dissolved organic carbon pool of temperate grass-land field soils. Soil Biology and Biochemistry 39, 827e835.

Bolan, N.S., Barrow, N.J., Posner, A.M., 1985. Describing the effect of time on sorptionof phosphate by iron and aluminium hydroxides. Journal of Soil Science 36,

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212 211

187e197.Bradford, M.A., Keiser, A.D., Davies, C.A., Mersmann, C.A., Strickland, M.S., 2013.

Empirical evidence that soil carbon formation from plant inputs is positivelyrelated to microbial growth. Biogeochemistry 113, 271e281.

Cambardella, C.A., Moorman, T.B., Parkin, T.B., Karlen, D.L., Novak, J.M., Turco, R.F.,Konopka, A.E., 1994. Field-scale variability of soil properties in central Iowasoils. Soil Science Society of America Journal 58, 1501.

Cenini, V.L., Fornara, D.A., McMullan, G., Ternan, N., Lajtha, K., Crawley, M.J., 2015.Chronic nitrogen fertilization and carbon sequestration in grassland soils: evi-dence of a microbial enzyme link. Biogeochemistry 126, 301e313.

Cleveland, C.C., Liptzin, D., 2007. C: N:P stoichiometry in soil: is there a “Redfieldratio” for the microbial biomass? Biogeochemistry 85, 235e252.

Commichau, F.M., Forchhammer, K., Stülke, J., 2006. Regulatory links betweencarbon and nitrogen metabolism. Current Opinion in Microbiology Cell Regu-lation 9, 167e172. Edited by Werner Goebel and Stephen Lory.

Conant, R.T., Ryan, M.G., Ågren, G.I., Birge, H.E., Davidson, E.A., Eliasson, P.E.,Evans, S.E., Frey, S.D., Giardina, C.P., Hopkins, F.M., Hyv€onen, R.,Kirschbaum, M.U.F., Lavallee, J.M., Leifeld, J., Parton, W.J., Megan Steinweg, J.,Wallenstein, M.D., Wetterstedt, M.Å.J., Bradford, M.A., 2011. Temperature andsoil organic matter decomposition rates e synthesis of current knowledge and away forward. Global Change Biology 17, 3392e3404.

Cotrufo, M.F., Soong, J.L., Horton, A.J., Campbell, E.E., Haddix, M.L., Wall, D.H.,Parton, W.J., 2015. Formation of soil organic matter via biochemical and physicalpathways of litter mass loss. Nature Geoscience 8, 776e779.

Cotrufo, M.F., Wallenstein, M.D., Boot, C.M., Denef, K., Paul, E., 2013. The MicrobialEfficiency-Matrix Stabilization (MEMS) framework integrates plant litterdecomposition with soil organic matter stabilization: do labile plant inputsform stable soil organic matter? Global Change Biology 19, 988e995.

Creamer, C.A., Jones, D.L., Baldock, J.A., Farrell, M., 2014. Stoichiometric controlsupon low molecular weight carbon decomposition. Soil Biology and Biochem-istry 79, 50e56.

Cusack, D.F., Silver, W.L., Torn, M.S., Burton, S.D., Firestone, M.K., 2011. Changes inmicrobial community characteristics and soil organic matter with nitrogenadditions in two tropical forests. Ecology 92, 621e632.

Derenne, S., Largeau, C., 2001. A review of some important families of refractorymacromolecules: composition, origin, and fate in soils and sediments. SoilScience 166, 833e847.

Doetterl, S., Stevens, A., Six, J., Merckx, R., Van Oost, K., Casanova Pinto, M., Casa-nova-Katny, A., Mu~noz, C., Boudin, M., Zagal Venegas, E., Boeckx, P., 2015. Soilcarbon storage controlled by interactions between geochemistry and climate.Nature Geoscience 8, 780e783.

Dungait, J.A.J., Hopkins, D.W., Gregory, A.S., Whitmore, A.P., 2012. Soil organicmatter turnover is governed by accessibility not recalcitrance. Global ChangeBiology 18, 1781e1796.

Foley, J.A., Ramankutty, N., Brauman, K.A., Cassidy, E.S., Gerber, J.S., Johnston, M.,Mueller, N.D., O'Connell, C., Ray, D.K., West, P.C., Balzer, C., Bennett, E.M.,Carpenter, S.R., Hill, J., Monfreda, C., Polasky, S., Rockstr€om, J., Sheehan, J.,Siebert, S., Tilman, D., Zaks, D.P.M., 2011. Solutions for a cultivated planet. Na-ture 478, 337e342.

Frey, S.D., Lee, J., Melillo, J.M., Six, J., 2013. The temperature response of soil mi-crobial efficiency and its feedback to climate. Nature Climate Change 3,395e398.

Frostegård, Å., Tunlid, A., Bååth, E., 1991. Microbial biomass measured as total lipidphosphate in soils of different organic content. Journal of MicrobiologicalMethods 14, 151e163.

G€arden€as, A.I., Ågren, G.I., Bird, J.A., Clarholm, M., Hallin, S., Ineson, P., K€atterer, T.,Knicker, H., Nilsson, S.I., N€asholm, T., Ogle, S., Paustian, K., Persson, T.,Stendahl, J., 2011. Knowledge gaps in soil carbon and nitrogen interactions -from molecular to global scale. Soil Biology and Biochemistry 43, 702e717.

Geyer, K.M., Kyker-Snowman, E., Grandy, A.S., Frey, S.D., 2016. Microbial carbon useefficiency: accounting for population, community, and ecosystem-scale controlsover the fate of metabolized organic matter. Biogeochemistry 127, 173e188.

Gillis, J.D., Price, G.W., 2011. Comparison of a novel model to three conventionalmodels describing carbon mineralization from soil amended with organicresidues. Geoderma 160, 304e310.

Glanville, H., Rousk, J., Golyshin, P., Jones, D.L., 2012. Mineralization of low molec-ular weight carbon substrates in soil solution under laboratory and field con-ditions. Soil Biology and Biochemistry 48, 88e95.

Glanville, H.C., Hill, P.W., Schnepf, A., Oburger, E., Jones, D.L., 2016. Combined use ofempirical data and mathematical modelling to better estimate the microbialturnover of isotopically labelled carbon substrates in soil. Soil Biology andBiochemistry 94, 154e168.

Gleeson, D.B., Müller, C., Banerjee, S., Ma, W., Siciliano, S.D., Murphy, D.V., 2010.Response of ammonia oxidizing archaea and bacteria to changing water filledpore space. Soil Biology and Biochemistry 42, 1888e1891.

Gonzalez-Qui~nones, V., Stockdale, E.A., Banning, N.C., Hoyle, F.C., Sawada, Y.,Wherrett, A.D., Jones, D.L., Murphy, D.V., 2011. Soil microbial bio-massdInterpretation and consideration for soil monitoring. Soil Research 49,287e304.

Haddix, M.L., Paul, E.A., Cotrufo, M.F., 2016. Dual, differential isotope labeling showsthe preferential movement of labile plant constituents into mineral-bonded soilorganic matter. Global Change Biology 22, 2301e2312.

Hagerty, S.B., van Groenigen, K.J., Allison, S.D., Hungate, B.A., Schwartz, E.,Koch, G.W., Kolka, R.K., Dijkstra, P., 2014. Accelerated microbial turnover butconstant growth efficiency with warming in soil. Nature Climate Change 4,

903e906.Hararuk, O., Smith, M.J., Luo, Y., 2015. Microbial models with data-driven parame-

ters predict stronger soil carbon responses to climate change. Global ChangeBiology 21, 2439e2453.

Hassink, J., 1997. The capacity of soils to preserve organic C and N by their associ-ation with clay and silt particles. Plant and Soil 191, 77e87.

Hessen, D.O., Ågren, G.I., Anderson, T.R., Elser, J.J., Ruiter, P.C. de, 2004. Carbonsequestration in ecosystems: the role of stoichiometry. Ecology 85, 1179e1192.

H€ogberg, P., Fan, H., Quist, M., Binkley, D., Tamm, C.O., 2006. Tree growth and soilacidification in response to 30 years of experimental nitrogen loading on borealforest. Global Change Biology 12, 489e499.

Jastrow, J.D., Amonette, J.E., Bailey, V.L., 2007. Mechanisms controlling soil carbonturnover and their potential application for enhancing carbon sequestration.Climatic Change 80, 5e23.

Jenkinson, D.S., 1978. The Soil Biomass. NZ Soil News, pp. 213e218.Jones, D.L., Murphy, D.V., 2007. Microbial response time to sugar and amino acid

additions to soil. Soil Biology and Biochemistry 39, 2178e2182.Kaiser, C., Franklin, O., Dieckmann, U., Richter, A., 2014. Microbial community dy-

namics alleviate stoichiometric constraints during litter decay. Ecology Letters17, 680e690.

Kallenbach, C.M., Grandy, A.S., Frey, S.D., Diefendorf, A.F., 2015. Microbial physiologyand necromass regulate agricultural soil carbon accumulation. Soil Biology andBiochemistry 91, 279e290.

Kirkby, C.A., Kirkegaard, J.A., Richardson, A.E., Wade, L.J., Blanchard, C., Batten, G.,2011. Stable soil organic matter: a comparison of C: N:P: S ratios in Australianand other world soils. Geoderma 163, 197e208.

Kleber, M., Eusterhues, K., Keiluweit, M., Mikutta, C., Mikutta, R., Nico, P.S., 2015.Mineral-organic associations: formation, properties, and relevance in soil en-vironments. Advances in Agronomy 130, 1e140.

Kleber, M., Sollins, P., Sutton, R., 2007. A conceptual model of organo-mineral in-teractions in soils: self-assembly of organic molecular fragments into zonalstructures on mineral surfaces. Biogeochemistry 85, 9e24.

Kramer, C., Trumbore, S., Fr€oberg, M., Cisneros Dozal, L.M., Zhang, D., Xu, X.,Santos, G.M., Hanson, P.J., 2010. Recent (<4 year old) leaf litter is not a majorsource of microbial carbon in a temperate forest mineral soil. Soil Biology andBiochemistry 42, 1028e1037.

Kuzyakov, Y., Jones, D.L., 2006. Glucose uptake by maize roots and its trans-formation in the rhizosphere. Soil Biology and Biochemistry 38, 851e860.

Lal, R., 2004. Soil carbon sequestration impacts on global climate change and foodsecurity. Science 304, 1623e1627.

LeBauer, D.S., Treseder, K.K., 2008. Nitrogen limitation of net primary productivityin terrestrial ecosystems is globally distributed. Ecology 89, 371e379.

Leff, J.W., Jones, S.E., Prober, S.M., Barber�an, A., Borer, E.T., Firn, J.L., Harpole, W.S.,Hobbie, S.E., Hofmockel, K.S., Knops, J.M.H., McCulley, R.L., Pierre, K.L.,Risch, A.C., Seabloom, E.W., Schütz, M., Steenbock, C., Stevens, C.J., Fierer, N.,2015. Consistent responses of soil microbial communities to elevated nutrientinputs in grasslands across the globe. Proceedings of the National Academy ofSciences 112, 10967e10972.

Lundquist, E.J., Scow, K.M., Jackson, L.E., Uesugi, S.L., Johnson, C.R., 1999. Rapidresponse of soil microbial communities from conventional, low input, andorganic farming systems to a wet/dry cycle. Soil Biology and Biochemistry 31,1661e1675.

M€annist€o, M., Ganzert, L., Tiirola, M., H€aggblom, M.M., Stark, S., 2016. Do shifts inlife strategies explain microbial community responses to increasing nitrogen intundra soil? Soil Biology and Biochemistry 96, 216e228.

Manzoni, S., Jackson, R.B., Trofymow, J.A., Porporato, A., 2008. The global stoichi-ometry of litter nitrogen mineralization. Science 321, 684e686.

Manzoni, S., Taylor, P., Richter, A., Porporato, A., Ågren, G.I., 2012. Environmentaland stoichiometric controls on microbial carbon-use efficiency in soils. NewPhytologist 196, 79e91.

McKenzie, N.J., Coughlan, K.J., Cresswell, H.P., 2002. Soil Physical Measurement andInterpretation for Land Evaluation. In: Australian Soil and Land Survey Hand-book. CSIRO Publishing.

Mehlich, A., 1984. Mehlich 3 soil test extractant: a modification of Mehlich 2extractant. Communications in Soil Science and Plant Analysis 15, 1409e1416.

Miltner, A., Bombach, P., Schmidt-Brücken, B., K€astner, M., 2011. SOM genesis: mi-crobial biomass as a significant source. Biogeochemistry 111, 41e45.

Miranda, K.M., Espey, M.G., Wink, D.A., 2001. A rapid, simple spectrophotometricmethod for simultaneous detection of nitrate and nitrite. Nitric Oxide 5, 62e71.

Mooshammer, M., Wanek, W., H€ammerle, I., Fuchslueger, L., Hofhansl, F.,Knoltsch, A., Schnecker, J., Takriti, M., Watzka, M., Wild, B., Keiblinger, K.M.,Zechmeister-Boltenstern, S., Richter, A., 2014. Adjustment of microbial nitrogenuse efficiency to carbon:nitrogen imbalances regulates soil nitrogen cycling.Nature Communications 5.

Mulvaney, R.L., 1996. Nitrogendinorganic forms. In: Sparks, D.L., Page, A.L.,Helmke, P.A., Loeppert, R.H. (Eds.), Methods of Soil Analysis Part 3-chemicalMethods. Soil Science Society of America. American Society of Agronomy,Madison, WI, pp. 1123e1184.

Murphy, D.V., Cookson, W.R., Braimbridge, M., Marschner, P., Jones, D.L.,Stockdale, E.A., Abbott, L.K., 2011. Relationships between soil organic matter andthe soil microbial biomass (size, functional diversity, and community structure)in crop and pasture systems in a semi-arid environment. Soil Research 49,582e594.

Murphy, D.V., Osman, M., Russell, C.A., Darmawanto, S., Hoyle, F.C., 2009. Potentiallymineralisable nitrogen: relationship to crop production and spatial mapping

C.A. Creamer et al. / Soil Biology & Biochemistry 103 (2016) 201e212212

using infrared reflectance spectroscopy. Soil Research 47, 737e741.Murphy, J., Riley, J.P., 1962. A modified single solution method for the determination

of phosphate in natural waters. Analytica Chimica Acta 27, 31e36.Neff, J.C., Townsend, A.R., Gleixner, G., Lehman, S.J., Turnbull, J., Bowman, W.D.,

2002. Variable effects of nitrogen additions on the stability and turnover of soilcarbon. Nature 419, 915e917.

Oburger, E., Jones, D.L., 2009. Substrate mineralization studies in the laboratoryshow different microbial C partitioning dynamics than in the field. Soil Biologyand Biochemistry 41, 1951e1956.

Rayment, G.E., Lyons, D.J., 2011. 18. Miscellaneous extractable elements. In: soilchemical methods - Australasia. CSIRO Publishing, Collingwood VIC. Australia385e414.

Riggs, C.E., Hobbie, S.E., 2016. Mechanisms driving the soil organic matter decom-position response to nitrogen enrichment in grassland soils. Soil Biology andBiochemistry 99, 54e65.

Schimel, J.P., Weintraub, M.N., 2003. The implications of exoenzyme activity onmicrobial carbon and nitrogen limitation in soil: a theoretical model. SoilBiology and Biochemistry 35, 549e563.

Scow, K.M., Simkins, S., Alexander, M., 1986. Kinetics of mineralization of organiccompounds at low concentrations in soil. Applied and Environmental Micro-biology 51, 1028e1035.

Sinsabaugh, R.L., Manzoni, S., Moorhead, D.L., Richter, A., 2013. Carbon use effi-ciency of microbial communities: stoichiometry, methodology and modelling.Ecology Letters 16, 930e939.

Six, J., Conant, R.T., Paul, E.A., Paustian, K., 2002. Stabilization mechanisms of soilorganic matter: implications for C-saturation of soils. Plant and Soil 241,155e176.

Six, J., Frey, S.D., Thiet, R.K., Batten, K.M., 2006. Bacterial and fungal contributions tocarbon sequestration in agroecosystems. Soil Science Society of America Journal70, 555.

Skogley, E.O., Dobermann, A., 1996. Synthetic ion-exchange resins: soil and envi-ronmental studies. Journal of Environment Quality 25, 13.

Strahm, B.D., Harrison, R.B., 2007. Mineral and organic matter controls on thesorption of macronutrient anions in variable-charge soils. Soil Science Society ofAmerica Journal 71, 1926.

Strickland, M.S., Rousk, J., 2010. Considering fungal:bacterial dominance in soils e

methods, controls, and ecosystem implications. Soil Biology and Biochemistry42, 1385e1395.

SW-846 EPA Method 3051A, 1998. Microwave assisted digestion of sediments,sluges, soils and oils. In: Test Methods for Evaluating Solid Waste. U.S. Envi-ronmental Protection Agency, Washington, D.C.

Tang, J., Riley, W.J., 2015. Weaker soil carbon-climate feedbacks resulting frommicrobial and abiotic interactions. Nature Climate Change 5, 56e60.

Throckmorton, H.M., Bird, J.A., Dane, L., Firestone, M.K., Horwath, W.R., 2012. Thesource of microbial C has little impact on soil organic matter stabilisation inforest ecosystems. Ecology Letters 15, 1257e1265.

Throckmorton, H.M., Bird, J.A., Monte, N., Doane, T., Firestone, M.K., Horwath, W.R.,2014. The soil matrix increases microbial C stabilization in temperate andtropical forest soils. Biogeochemistry 122, 35e45.

Vance, E.D., Brookes, P.C., Jenkinson, D.S., 1987. An extraction method for measuringsoil microbial biomass C. Soil Biology and Biochemistry 19, 703e707.

Vitousek, P.M., Mooney, H.A., Lubchenco, J., Melillo, J.M., 1997. Human dominationof Earth's ecosystems. Science 277, 494e499.

Wardle, D.A., 1998. Controls of temporal variability of the soil microbial biomass: aglobal-scale synthesis. Soil Biology and Biochemistry 30, 1627e1637.

White, D.C., Davis, W.M., Nickels, J.S., King, J.D., Bobbie, R.J., 1979. Determination ofthe sedimentary microbial biomass by extractible lipid phosphate. Oecologia40, 51e62.

Wieder, W.R., Bonan, G.B., Allison, S.D., 2013. Global soil carbon projections areimproved by modelling microbial processes. Nature Climate Change 3,909e912.

Wieder, W.R., Grandy, A.S., Kallenbach, C.M., Bonan, G.B., 2014. Integrating microbialphysiology and physio-chemical principles in soils with the MIcrobial-MIneralCarbon Stabilization (MIMICS) model. Biogeosciences 11, 3899e3917.

Wutzler, T., Blagodatsky, S.A., Blagodatskaya, E., Kuzyakov, Y., 2012. Soil microbialbiomass and its activity estimated by kinetic respiration analysis e statisticalguidelines. Soil Biology and Biochemistry 45, 102e112.

Yang, J.E., Skogley, E.O., Schaff, B.E., 1991. Nutrient flux to mixed-bed ion-exchangeresin: temperature effects. Soil Science Society of America Journal 55, 762.

Zelles, L., 1999. Fatty acid patterns of phospholipids and lipopolysaccharides in thecharacterisation of microbial communities in soil: a review. Biology and Fertilityof Soils 29, 111e129.