can intra-aggregate pore structures affect the aggregate’s effectiveness in protecting carbon?

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

http://www.elsevier.com/copyright

Author's personal copy

Can intra-aggregate pore structures affect the aggregate’s effectiveness inprotecting carbon?

K. Ananyeva, W. Wang, A.J.M. Smucker, M.L. Rivers, A.N. Kravchenko*

Dept. Crop and Soil Sci., Michigan State University, East Lansing, MI 48824-1325, United States

a r t i c l e i n f o

Article history:Received 13 June 2012Received in revised form16 October 2012Accepted 18 October 2012Available online 15 November 2012

Keywords:X-ray computed micro-tomographySpatial variabilityNative vegetation successionConventionally tilled row crop agriculturalsystem

a b s t r a c t

Aggregates are known to provide physical protection to soil organic matter shielding it from rapiddecomposition. Spatial arrangement and size distribution of intra-aggregate pores play an important rolein this process. This study examined relationships between intra-aggregate pores measured using X-raycomputed micro-tomography images and concentrations of total C in 4e6 mm macro-aggregates fromtwo contrasting land use and management practices, namely, conventionally tilled and managed rowcrop agricultural system (CT) and native succession vegetation converted from tilled agricultural land in1989 (NS). Previous analyses of these aggregates indicated that small (<15 mm) and large (>100 mm)pores prevail in NS aggregates while medium (30e90 mm) pores are more abundant in CT aggregates(Kravchenko et al., 2011; Wang et al., 2012). We hypothesized that these differences in pore sizedistributions affect the ability of macro-aggregates to protect C. The results of this study supported thishypothesis. Consistent with greater heterogeneity of pore distributions within NS aggregates weobserved higher total C and greater intra-aggregate C variability in NS as compared with CT aggregates.Total C concentrations and intra-aggregate C standard deviations were negatively correlated with frac-tions of medium sized pores, indicating that presence of such pores was associated with lower but morehomogeneously distributed total C. While total C was positively correlated with presence of small andlarge pores. The results suggest that because of their pore structure NS macro-aggregates provide moreeffective physical protection to C than CT aggregates.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Macro-aggregates are known to provide physical protection tosoil organic matter shielding it from rapid decomposition and thusare regarded among the key elements enabling soil C sequestration(Beare et al., 1994; Paustian et al., 1997; Bossuyt et al., 2002; vonLützow et al., 2006). Intra-aggregate physical protection is theleading driver of C sequestration occurring when land underintensive agricultural management is converted to conservationalland use practices (Jastrow, 1996; Grandy and Robertson, 2007). Insoils under conservational land use, e.g., grasslands or soils aban-doned from agriculture, macro-aggregates tend to have higher Cconcentrations and are richer in newer C than other soil fractions(Jastrow, 1996; De Gryze et al., 2004). The newly added C oftenserves as a binding agent holding the macro-aggregates together.When intra-aggregate physical protection is eliminated by crushingmacro-aggregates, the intra-aggregate C accumulated by conser-vational management is easily mineralized (Beare et al., 1994;

Hassink, 1997). However, when macro-aggregates stay intact forprolonged time periods in undisturbed soils of conservationalmanagement, decomposition of organic binding agents is suffi-ciently slow to allow for formation of micro-aggregates wherephysical protection is enhanced by physicochemical and chemicalprotection processes (Six et al., 2000; Denef et al., 2001; Chenu andPlante, 2006).

One of the mechanisms of organic matter protection is hetero-geneity of soil microenvironment which limits the access ofdecomposing microorganisms and their enzymes to organicmaterial (Schmidt et al., 2011). Macro-aggregate formationincreases such heterogeneity, and thus organic matter protection,by increasing complexity in spatial arrangement of soil matrixpores (Baldock and Skjemstad, 2000). Pores affect water and airfluxes, spatial distributions of soil nutrients, and movement ofmicroorganisms through soil on scales ranging from a soil profile toa micro-aggregate (Or et al., 2007). For example, it has been shownthat large pores of preferential flow paths contain younger soilorganic C than their surroundings and serve as “hotspots” for bio-logical activity (Bundt, 2001). Greater microbial activity and organicmatter decomposition was observed in larger pores or were

* Corresponding author.E-mail address: [email protected] (A.N. Kravchenko).

Contents lists available at SciVerse ScienceDirect

Soil Biology & Biochemistry

journal homepage: www.elsevier .com/locate/soi lbio

0038-0717/$ e see front matter � 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.soilbio.2012.10.019

Soil Biology & Biochemistry 57 (2013) 868e875


associatedwith greater numbers of large pores (Killham et al., 1993;Yoo et al., 2006; Ruamps et al., 2011). Soil nutrients were reportedto accumulate around pores approximately 100e200 mm in diam-eter within soil aggregates (Jassogne, 2008). Pore arrangementsmay restrict microbial access to organic material and reducemicrobial activity by limiting oxygen supply (Sexstone et al., 1985;Golchin et al., 1994; von Lützow et al., 2006).

Aggregate boundaries serve as dividers between smaller intra-aggregate pores inside aggregates and larger inter-aggregate poressurrounding them. Majority of air and water fluxes occur via inter-aggregate pores from which air, water, chemicals, and microor-ganisms can enter the aggregates. Comparisons of the intra-aggregate soil properties at different distances from aggregateboundaries supply ample evidence of the importance of pore sizesfor intra-aggregate biogeochemical processes including thoserelated to C. Within the aggregates, gradients are often observed interms of a variety of soil properties reflecting the proximity toaggregate boundaries, that is, to inter-aggregate pores. Among thegradients that have been observed are gradients in oxygen levels(Sexstone et al.,1985), Ca,Mg, K, Na,Mn, K, Al, and Fe concentrations(Santos et al., 1997; Jasinska et al., 2006), and in composition oforganic matter (Ellerbrock and Gerke, 2004; Urbanek et al., 2007).Differences have been found between external and internal aggre-gate layers in terms of microbial activities (Jasinska et al., 2006) andmicrobial community compositions (Blackwood et al., 2006). Thesegradients appear to be soil specific. For example, Chenu et al. (2001)found that microbial decomposition occurs throughout the wholeaggregates of a sandy soil, while only near the surface of a clay soilaggregates. Especially diverse are results obtained for total C wheresome studies reported no intra-aggregate patterns (Ellerbrock andGerke, 2004; Urbanek et al., 2007), while others found such gradi-ents in some but not other soils (Santos et al., 1997; Jasinska et al.,2006). It is likely that not only proximity to the inter-aggregatepore space but also abundance, size, and connectivity of intra-aggregate pores play a substantial role in intra-aggregate C-related processes and resulting intra-aggregate patterns in spatialdistributions of C.

When land is abandoned from an intensive agricultural use andis either converted into grassland or allowed to be taken by nativevegetation, multiple factors contributing to C sequestration becomeactivated. Among them are year-around presence of vegetationcover supplying fresh organic inputs different in quality from thoseof intensive agriculture, increased microbial and faunal activity,elimination of soil disturbance, and an increase in quantities andstabilities of soil aggregates. Recently it has been shown that suchsoils are not only different from conventionally managed agricul-tural soils in terms of aggregate size distributions or aggregatestability but also in terms of intra-aggregate pore characteristics.For example, Peth et al. (2008) reported prevalence of longcontinuous pores in a grassland aggregate while short thin inter-connected pores were more abundant in a CT aggregate.Kravchenko et al. (2011) observed greater heterogeneity in intra-aggregate pore distributions in macro-aggregates from soil undernative succession vegetation as compared to conventional agri-culture. Wang et al. (2012) reported greater fractions of 30e60 mmcrack-like pores in macro-aggregates from conventionally tilled soilwhile greater fractions of >90 mm pores of biological origin inaggregates from soil abandoned from agriculture for past 18 years.We hypothesize that in addition to the listed above factorscontributing to C sequestration in soils converted to native vege-tation the changes that take place in intra-aggregate pore struc-tures constitute yet another factor that enables enhanced Csequestration to take place in such soils.

The goal of this study is to determine whether the intra-aggregate pore characteristics of macro-aggregates from

a conservational land use system make such aggregates moreeffective in C protection as compared to those of conventionalintensive agriculture. We examined soils from two highly con-trasting land uses: a heavily disturbed soil in a conventionallyplowed row crop agricultural system and an undisturbed soil undernative vegetation. During past 18 years the first systemwas losing C(Senthilkumar et al., 2009) while the second system has beenshown to rapidly accumulate it (Grandy and Robertson, 2007). Theobjectives of the study are to assess spatial patterns of C distribu-tion within the aggregates from the two systems and to examinethe relationships between intra-aggregate C and spatial distribu-tions of intra-aggregate pores obtained via X-ray computed micro-tomography.

2. Materials and methods

2.1. Soil sampling

Soil samples were collected from Long Term Ecological Research(LTER) site at the W. Kellogg Biological Station in southwestMichigan. The soil is Kalamazoo loam (fine-loamy, mixed, mesic,Typic Hapludalf), developed on glacial outwash. The LTER experi-ment was established in 1988 (for details on site description,experimental design and research protocols see http://lter.kbs.msu.edu) (KBS, 2011). Prior to 1988 the entire experimental site was inconventionally plowed row crop agricultural management for atleast past 100 years. The two most contrasting treatments in termsof soil management were used in this study, namely, conventionaltillage (chisel-plowed) corn-soybean-wheat rotation with conven-tional chemical inputs (CT) and native succession grassland,abandoned from agricultural use after spring plowing in 1989 (NS).

Sampling was conducted in 2008 from two adjacent 1 ha plots,one plot per treatment. Within each plot three randomly selectedsampling locations were identified and at each location a soil blockapproximately 15 � 15 � 15 cm in size was extracted by spade fromthe soil surface. Soil blocks were packed into closed-lid plasticcontainers and transported to the lab, where they were gentlymanually crushed to ensure breakage along the natural planes ofweakness. Then crushed soil was air-dried and dry-sieved to obtainmacro-aggregates 4e6.3 mm in size. In the rest of the manuscriptwe will refer to them simply as aggregates. Grandy and Robertson(2007) reported that based on wet-sieving results 2e8 mmaggregates constituted around 15% and 50% of the aggregates inCT and NS treatments, respectively. The aggregate size was chosenas a compromise between the need to have aggregates largeenough to provide sufficient amount of soil material for multipleintra-aggregatemeasurements of soil C and the need for aggregatesto be sufficiently small to enable fine resolution in X-ray computedmicro-tomography scanning. Air-dried aggregates were stored inairtight containers at the room temperature until used for analyses.Six replicated aggregates from each treatment, i.e., two aggregatesfrom each sampling site, were used in further analyses.

2.2. Image collection and analyses

The aggregate images were obtained at the Advanced PhotonSource of Argonne National Laboratory (station BMD-13) using X-ray micro-tomography. Image resolution was equal to approxi-mately 15 mm in x, y, and z directions. To obtain information onintra-aggregate pores, the images were segmented by classifyingeach image voxels as either a pore or a solid material. Segmentationwas conducted using indicator kriging method (Oh and Lindquist,1999; Wang et al., 2011). The pore characteristics that werestudied are image-based porosity, that is the percent of poresvisible at image resolution (>15 mm), and size distributions of such

K. Ananyeva et al. / Soil Biology & Biochemistry 57 (2013) 868e875 869


pores. Pore size distributions were assessed using burn numberdistribution approach implemented in 3DMA-Rock software(Lindquist et al., 2000). A burn number represents the shortestdistance from the center of the pore to the pore’s wall and can beregarded as the measure of pore diameter. For clarity purposes, inthis study wewill refer to the pores with different burn numbers byconverting the burn number values into ranges of pore diametervalues. That is, pores with burn number equal to 1 will be referredto as 15e37.5 mm pores, those with burn number of 2 will be called37.5e67.5 mm pores, those with burn number of 3 67.5e97.5 mmpores, etc. The aggregates used in this study are a subset of theaggregates analyzed for pore spatial patterns by Kravchenko et al.(2011) and Wang et al. (2012) where detailed descriptions of theimage collection, pre-processing, and processing procedures aresupplied. Here we only provide a summary of the pore character-istics for the 12 aggregates that were used for C measurements ofthis study to facilitate the discussion of C results.

2.3. Aggregate cutting and intra-aggregate C measurements

To assess intra-aggregate spatial variability of soil C and to relateC content to intra-aggregate pore characteristics, each aggregatewas cut into 11e20 sections. In order to facilitate cutting, theaggregates were wetted to fill approximately 30% of their porevolume with distilled water. Then cutting was performed usinga sharp #11 scalpel and a 24� magnifying glass. The relative posi-tion of each aggregate section was recorded. Depending on the sizeand shape of the aggregate we were able to cut 6e12 exterior

sections and 4e8 interior sections. After cutting, each section of theaggregate was oven-dried for 24 h at 104 �C, weighed, and sub-jected to C measurement. Total C measurements were performedon soil aggregate sections using the elemental combustion systemECS 4010 (Costech Analytical, USA).

The variables that were analyzed in this study consisted ofindividual C values from each section; the C values from sectionscombined into aggregate exterior and interior groups; andsummary statistics for each aggregate calculated from individualsection values including aggregate’s minimum and maximum Cvalues, and aggregate’s C standard deviations.

Virtual cutting of aggregate images was performed to matchphysical cutting (Fig. 1a). It yielded regions in the 3D tomographicimages that corresponded to the physically cut sections. Image-based porosity and pore size distributions were determined ineach virtual section of each aggregate.

2.4. Statistical analysis

The influence of land management and position within theaggregate, i.e., exterior vs. interior, on total C values was analyzedusing individual aggregate section data in PROC MIXED in SAS 9.2(SAS Institute, 2009). The statistical model included land manage-ment, position, and interaction between them as fixed factors.Multiple runs of C measurements on ECS 4010 systems wereincluded as a random blocking factor. Aggregates nested withinland use and the interaction between aggregates and positionswere the random factors used as error terms for testing land use

Fig. 1. 3D X-ray computed micro-tomography image of one of the studied aggregates with a schematic representation of the sections that the aggregate was cut into (a) and 2Dimages of representative NS (b) and CT (c) aggregates. Examples of the areas with no pores visible at the studied resolution, i.e., pores >15 mm, are marked by rectangles. Arrows arepointing to large (>100 mm) pores of biological origin on the NS image and to medium (30e90 mm) pores on the CT image.

K. Ananyeva et al. / Soil Biology & Biochemistry 57 (2013) 868e875870


and position effects, respectively. Analysis of land use effects onminimum and maximum intra-aggregate C values and on intra-aggregate standard deviations was conducted using the statisticalmodel with land use as the fixed factor. In all analyses the normalityassumption was checked using normal probability plots and thedata were log-transformed in cases of right-skewed distributions.Equality of variances was checked using visual examination of boxplots and, where applicable, analysis with unequal variances wasconducted. The differences were declared to be statistically signif-icant at 0.05 level.

Analyses of the relationships of C values in aggregate sectionsalong with C aggregate means, minimums, maximums and stan-dard deviations with pore characteristics in the aggregate sectionswere performed in SAS 9.2 using PROC REG. Regression analyseswere conducted for the entire data set and for individual data setsof each land use.

3. Results

3.1. Intra-aggregate C variability

Overall, the intra-aggregate C variability in NS aggregates wasmuch higher than that in CT. Standard deviations of C in sections ofNS aggregates were equal to 2.2 g/kg, which was twice as high asthat in CT aggregates, 1.1 g/kg (p < 0.05). In all but one CT aggregatethe range of C values among the aggregate sections was substan-tially lower than that in NS aggregates (Fig. 2). In CT aggregates theintra-aggregate range of C values wasw3.7 g/kg, while in NS it wasw7.7 g/kg and in some NS aggregates the difference between thelowest and the highest C values was as high as w9 g/kg.

Across both CT and NS aggregates the lowest C concentrationsobserved within the aggregate sections were equal to 5e6 g/kgwhile the highest were around 20 g/kg. The minimum C valuesin aggregate sections were equal to 7.8 and 8.7 g/kg in CT and NStreatments, respectively, and were not significantly different fromeach other (p < 0.05) (Fig. 3). A much larger difference wasobserved between maximum C values which were 1.5 timeshigher in NS than in CT aggregates (Fig. 3). Along with sectionsrelatively rich in C, NS aggregates contained sections with Cconcentrations as low as 5e7 g/kg, that is, the values similar tothose of CT aggregates.

In both CT and NS the aggregate sections located in the aggre-gate interiors tended to have lower C levels than the aggregate

exteriors (Fig. 4). The difference between interior and exteriorpositions was statistically significant in NS aggregates (p < 0.05).

Overall the NS aggregates had higher mean C than CT aggre-gates, equal to 12.3 and 9.3 g/kg in NS and CT aggregates, respec-tively (Fig. 3). However, the results of comparisons between CT andNS were different in aggregate exteriors vs. interiors (Fig. 4). Inaggregate exteriors the difference between NS and CT was equal to2.7 g/kg; it was significantly greater than zero (p < 0.05). However,in the aggregate interiors the difference was smaller, 1.7 g/kg, andnot significantly different from zero.

3.2. Relationships between C and intra-aggregate porecharacteristics

Across all aggregates from both treatments therewas aweak butstatistically significant negative correlation between total C andimage-based porosity, i.e., percent of pores>15 mm (Fig. 5a). That is,the aggregate sections with higher percent of pores visible on theimages tended to have lower C levels. However, as can be observedfrom Fig. 5a, the variability in C values from individual sections was

CT aggregates

CT6 CT7 CT8 CT9 CT10 CT12 NS6 NS7 NS8 NS9 NS10 NS11

To

ta

l C

, g

/kg

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

20.0

22.0

NS aggregates

Fig. 2. Total C concentrations in the sections of the studied aggregates. Aggregate IDnumbers are shown on the x-axis.

*

minimum mean maximum

To

ta

l C

, g

/kg

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

CTNS

*

Summary statistics of intra-aggregate sections

Fig. 3. Minimum, mean, and maximum total C values in sections of the studiedaggregates from CT and NS treatments. Bars represent standard errors. The differencebetween CT and NS aggregates was statistically significant for means and maximums(p < 0.05) (marked with *).

*

CT NS

To

ta

l C

, g

/kg

6.0

8.0

10.0

12.0

14.0

Exterior

Interior

Fig. 4. Total C concentrations in interior and exterior sections of the CT and NSaggregates. Bars represent standard errors. The difference between the positions wasstatistically significant in NS (p < 0.05) (marked with *).



very high, especially in the sections with lower image-basedporosity.

The relationships between C and pores differed between thetwo studied treatments. In NS there was a statistically significantpositive correlation between total C and presence of the smallest

image visible pores, i.e., pores 15e37.5 mm in size (Fig. 5b), anda significant negative correlation between C and medium sized(37.5e67.5 mm) pores (Fig. 5c). Regression equation predicted totalC to be around only 8 g/kg when 15e37.5 mmpores constituted 30%of all the image-based pores, while it increased to 13 g/kg when15e37.5 mm pores constituted 90% of the image based pores.Regression equation predicted total C to be around only 9 g/kg insections with >25% of 37.5e67.5 mm pores, while it was as high as15 g/kg in sections with only w5% of 37.5e67.5 mm pores. Weaknegative correlation was also observed between C and 67.5e97.5 mm pores. In NS no significant correlation was observedbetween C and larger size pores, while in CT intra-aggregate total Cwas not correlated with pores of any sizes.

4. Discussion

Image analysis indicated substantial differences in intra-aggregate pore size distributions of NS and CT systems(Kravchenko et al., 2011; Wang et al., 2012). Kravchenko et al.(2011) observed greater uniformity of pore distributions withinCT as opposed to NS aggregates, while Wang et al. (2012) reportedthat NS aggregates had more large pores (>97.5 mm) and moresmall pores (<15 mm) than CT aggregates; however, medium sizepores (37.5e97.5 mm) were more abundant in CT aggregates. Anexample of two representative 2D images from CT and NS aggre-gates is shown in Fig. 1b and c. NS aggregates contained multipleareas of soil material with very few pores visible at the imageresolution, that is > 15 mm pores (Fig. 1b marked by rectangles). Atthe same time, there were multiple large round pores (>100 mm) inNS aggregates that likely originated from either roots or animalborrowing activities (Fig. 1b marked by arrows). For the aggregatesused in this study such pores were 5 times more abundant in NSthan in CT soil (Fig. 6). Conversely, aggregates from CT possesseda uniformly distributed highly interconnected network of moder-ately sized pores (37.5e67.5 mm) (Fig. 1c marked by arrows). Suchpores were present in NS aggregates as well, though in fewernumbers. In the studied aggregates the fraction of such pores wasequal to 0.23 in CT as opposed to 0.16 in the NS aggregates (Fig. 6).As has been noted by Wang et al. (2012) these pores were oftencracks of non-biological origin observed to evenly permeate mostCT aggregates.

We hypothesized that these differences in intra-aggregate porecharacteristics could lead to more effective physical protection of

R2=0.10

Fraction of 15-37.5 m pores

0.2 0.4 0.6 0.8 1.0

To

ta

l C

, g

/kg

0.0

5.0

10.0

15.0

20.0

25.0b

R2=0.17

Fraction of 37.5-67.5 m pores

0.00 0.05 0.10 0.15 0.20 0.25 0.30

To

ta

l C

, g

/k

g

0.0

5.0

10.0

15.0

20.0

25.0c

R2=0.11

Image-based porosity, %

0 5 10 15 20

To

ta

l C

, g

/kg

0

5

10

15

20

25a

Fig. 5. Correlations between total C in aggregate sections with a) image-basedporosity, i.e., pores > 15 mm for the entire data set; and with fractions of 15e37.5 mm pores (b) and 37.5e67.5 mm pores (c) in NS aggregates.

Pore size, m

37.5-67.5 >127.5

Fra

ctio

n (

po

re

s 3

7.5

-6

7.5

m

)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Fra

ctio

n (

po

re

s >

12

7.5

m

)

0.00

0.01

0.02

0.03

0.04

0.05

CTNS

Fig. 6. Fractions of 37.5e67.5 mm and >127.5 mm pores in the studied CT and NSaggregates. Bars represent standard errors. Left y-axis corresponds to pores 37.5e67.5 mm; right y-axis corresponds to pores >127.5 mm. In both pore size classes thedifference between CT and NS treatments was statistically significant (p < 0.05).



soil C within NS as compared to CT aggregates. Note that here weare specifically referring to 4e6 mm macro-aggregates, whileaggregates of different size classes can possess different porecharacteristics. It is likely that absence of sizeable pathways mightrestrict access of microorganisms andmicrobial enzymes into areasof NS aggregates with low abundance of >15 mm pores, whilelimited air flow might reduce microbial activity, thus protectingorganic matter there from decomposition. Small pores (<4 mm)indeed have been reported to be positively correlated with Caccumulation (Strong et al., 2004).

On the other hand, root-originated large pores could serve assources of incoming C within the aggregates by either exudatesfrom roots and microorganisms located within them (Alami et al.,1999; Hinsinger et al., 2005; Watteau et al., 2006) or by dissolvedorganic matter transported through such pores (Park et al., 2007).Air regime in such pores is beneficial for aerobic bacterial activity.Other conditions however might render these pores to be a harsherenvironment for bacterial survival. For example, bacteria theremight bemore susceptible to predation, while special water regimeof such pores, e.g., fast convective water flow for short periods oftime and lack of water for most of the time, might limit movementof bacteria and transport of microbial enzymes (Or et al., 2007).Nunan et al. (2003) observed lower bacterial densities in directproximity of larger pores while the densities increased at >20 mmdistances from pore boundaries.

Consistent with the above considerations, the aggregates of thisstudy where small (15e37.5 mm) and large (>127.5 mm) pores weremore abundant indeed appeared to be more effective in C protec-tion by exhibiting higher C levels and greater intra-aggregate Cvariability. In the NS aggregates the intra-aggregate sections withhigher percent of 15e37.5 mm pores tended to have higher C(Fig. 5b). Across both treatments the aggregates with high fractionof >127.5 mm pores tended to have higher C levels (Fig. 7a). Greaterintra-aggregate C variability can be interpreted as a potentialpresence of intra-aggregate areas where conditions are especiallybeneficial for C protection as opposed to the main body of theaggregate. Majority of aggregates fromNS, that is, from the land usethat intensively sequesters C (Grandy and Robertson, 2007),demonstrated such variability with almost every aggregate con-taining sections with very low as well as very high C levels (Fig. 2).The overall higher mean C levels of NS aggregates appear to be notdue to a uniform increase in C but due to presence of such intra-aggregate maximum C spots (Fig. 3). The importance of largepores in this process is further highlighted by positive correlationbetween >127.5 mm pores and standard deviations of intra-aggregate C (Fig. 7b). The same statistically significant trend waspresent when the relationship between>127.5 mmpores and intra-aggregate C standard deviations was examined in the CT and NSaggregates separately (data not shown).

These results are consistent with those reported in literature.Strong et al. (2004) reported faster decomposition of plant residuesin 15e60 mm pores, but lower C decomposition in soils with higherpercent of very small and very large pores. Chenu et al. (2001)showed that bacterial growth and decomposition of addedsubstrate was limited mainly to the surface of clayey aggregateswith fine pores, while sandy aggregates with frequent medium andlarge pores permitted easy transport of substrate and bacteria intotheir interiors, and active substrate decomposition there. Stronget al. (2004) reported a negative correlation between total C and15e60 mm pores as well as greater microbial activity in such pores.

Our results also point to the negative role of medium sized poresin C protection. A negative correlation between C in the aggregatesections and the fractions of 37.5e67.5 mm pores was observed forthe entire data set (data not shown) and for the NS aggregates(Fig. 5c). Greater presence of such pores not only resulted in overall

ln(Maximum intra-aggregate relative frequency of >127.5 µm pores)

-5 -4 -3 -2

Sta

ndar

d de

viatio

n fo

r int

ra-a

ggre

gate

tota

l C, g

/kg

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

b

Average intra-aggregate relative frequency of 37.5-67.5 µm pores

0.10 0.15 0.20 0.25 0.30

Sta

ndar

d de

viat

ion

for i

ntra

-agg

rega

te to

tal C

, g/k

g

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

c

R2=0.45

R2=0.42

R2=0.26

ln(Maximum intra-aggregate fraction of >127.5 µm pores)

-5 -4 -3 -2

Ave

rage

intra

-agg

rega

te to

tal C

, g/k

g

4.0

6.0

8.0

10.0

12.0

14.0

16.0

18.0

a

Fig. 7. Relationships between maximum intra-aggregate fractions of >127.5 mm pores(log-transformed) and total C concentration within the aggregates (a) and standarddeviation of total C concentration within the aggregates (b); and (c) between standarddeviation of total C concentration within the aggregates and the intra-aggregateaverages in fractions of 37.5e67.5 mm pores.



lower intra-aggregate C levels but also made the intra-aggregate Cdistribution more uniform. Standard deviation of intra-aggregate Cwas negatively correlated to the abundance of 37.5e67.5 mm pores(Fig. 7c). As mentioned earlier, such pores were found to be moreuniformly distributed through CT than through NS aggregates(Kravchenko et al., 2011) and, consistently, the distribution of Cthrough CT aggregates was much more uniform than that in NSaggregates (Figs. 2 and 3).

Another likely cause for the presence of more uniform Cdistributions within CT aggregates is frequent aggregate turnoverdue to tillage (Balesdent et al., 2000; Six et al., 2000) when soilmaterial is recurrently mixed during aggregate destruction andreformation. However, analysis of the relationship between thepresence of intra-aggregate medium sized pores and intra-aggregate C variability indicated that not only in CT but also in NSaggregates, higher abundance of such pores was associated withlower variability in C values (Fig. 7c).

The advantage of this study is that the aggregate samples werecollected in close spatial proximity (w25 m apart), that is, from thesoil of the same mineralogy and texture. Also until 1989 they hadsimilar vegetation and management history. This allows us tounequivocally point to the lack of soil disturbance and differencesin vegetation, that is, in timings, quantities and qualities of organicinputs, as the key factors driving the differences in both intra-aggregate pore distributions and their relationships with soil Clevels. A recent study conducted at KBS-LTER demonstrated that NShad more than twice higher microbial biomass, short-term evolvedCO2eC, and mineralizable C than those of CT treatment (Jangidet al., 2011).

As an indication of vegetation’s contribution to aggregateformation, consider that in most of the aggregates with high frac-tion of >127.5 mm pores, we observed root residues inside the largepores. This suggests that these were the aggregates of root origin,that is, the aggregates formed around relatively large living roots.Aggregate formation around plant roots could have led tocompaction of the soil material not only directly adjacent to theroot but also throughout the entire aggregate, thus eliminatingmedium sized pores (Hinsinger et al., 2005). Such compactedregions could be the areas with very few pores visible at imageresolution (pores >15 mm) marked on Fig. 1b for the NS aggregate.Roots are probably the leading aggregate forming factor in NSwhere root systems from a variety of plant species are active duringmost of the year. There were only a few aggregates of root origin,i.e., the aggregates with clearly visible root remnants in largeround pores, among CT aggregates (Kravchenko et al., in press). Themajority of CT aggregates could have formed not around live rootsbut around dead POM nuclei material or as a result of physicaleffects of wetting/drying or fungal enmeshment (Park andSmucker, 2005; Urbanek et al., 2011). We can speculate that theinflux of root originated C and the lack of medium pores are thedriving causes of the enhanced C protection in the aggregates ofroot origin as compared to other aggregates. Furthermore, the localcompaction occurring within root-originated aggregates might beparticularly beneficial for formation of micro-aggregates thusfurther enhancing C protection (Six et al., 2000, 2002; Chenu andPlante, 2006). However, further experimentation with aggregatespossessing different pore size distributions but of the samemanagement treatment would be necessary to test this hypothesis.

Another source of C influx into aggregates is from inter-aggregate pores that resulted in the higher C in aggregate exte-riors observed in this study (Fig. 4). These observations areconsistent with the hypothesis by Park and Smucker (2005) andPark et al. (2007) regarding movement of dissolved organic matterfrom inter-aggregate pores into the aggregates as one of themechanisms of increasing C contents. Park and Smucker (2005)

reported that newer C is located more in the aggregate exteriors,and diffuses into aggregate interiors during subsequent wettingand drying. Continuous influx of dissolved organic C from outsidemight be reinforcing and strengthening the existing intra-aggregate bonds derived from interactions of metals and organicligands with mineral surfaces (Horn and Smucker, 2005). Slowermacro-aggregate turnover in NS (Six et al., 1999) would allow fora longer duration for enrichment of C exteriors with inter-aggregateC influxes. Further studies are necessary to address this point.

5. Conclusions

Previously reported differences in intra-aggregate pore sizedistributions between macro-aggregates of an Alfisol undera conventionally tilled agricultural system and under a long termnative succession vegetation system were associated with intra-aggregate patterns in distribution of total C. These results wereconsistent with the hypothesis that NS aggregates provide greater Cprotection benefits than CT aggregates. However, further studieswith direct measurements of C decompositions and microbialactivities in aggregates of different treatments and with differentpore size distributions are necessary to enable its full testing.Overall higher total C and higher intra-aggregate variability in Cwere observed in the aggregates with greater numbers of verysmall and very large pores, while lower total C and lower variabilitywere found in aggregates with greater abundance of medium sizedpores. It appears that greater heterogeneity in spatial distributionof intra-aggregate pores, characterized by multiple areas with veryfew if any pores>15 mm and more biologically originated >100 mmpores, was advantageous for C accumulation.

It can be speculated that the differences between pore struc-tures within soil aggregates translate into different modes of bio-logical activity, and thus different rates of organic matterdecomposition. In undisturbed soils, macro-aggregate interiorsmay provide both the environments conducive to enhancedmicrobial activity, e.g., vicinities of large pores, and the environ-ments with limited oxygen/bacterial access where microbialactivity will be limited. Conversely, an increased intra-aggregateaeration can promote rapid and uniform decomposition of Cthrough whole bodies of macro-aggregates in highly disturbedsoils. Further measurements are necessary to explore thesehypotheses.

Acknowledgements

The project was supported in part by the National ResearchInitiative of the USDA Cooperative State Research, Education andExtension Service, grant number 32008-35102-04567. Support forthis research was also provided by the NSF Long-Term EcologicalResearch Program at the Kellogg Biological Station and byMichiganState University AgBioResearch. Authors would like to thank H.-C.Chun, A. Worth, M. Ladoni, J-D Munoz, and M. Mazher for theirhelp in sampling, CMT scanning, and aggregate processing.

References

Alami, Y., Champolivier, L., Merrien, A., Heulin, T., 1999. The role of Rhizobium sp.,a rhizobacterium that produces exopolysaccharide in the aggregation of therhizospherical soil of the sunflower, effects on plant growth and resistance tohydric constraints. OCL-Oleagineux Corps Gras Lipides 6, 524e528.

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.

Balesdent, J., Chenu, C., Balabane, M., 2000. Relationship of soil organic matterdynamics to physical protection and tillage. Soil & Tillage Research 53, 215e230.



Beare, M.H., Hendrix, P.F., Coleman, D.C., 1994. Water-stable aggregates and organicmatter fractions in conventional- and no-tillage soils. Soil Science SocietyAmerica Journal 58, 777e786.

Blackwood, C.B., Dell, C.J., Smucker, A.J.M., Paul, E.A., 2006. Eubacterial communitiesin different soil macroaggregate environments and cropping systems. SoilBiology & Biochemistry 38, 720e728.

Bossuyt, H., Six, J., Hendrix, P.F., 2002. Aggregate-protected carbon in no-tillage andconventional tillage agroecosystems using carbon-14 labeled plant residue. SoilScience Society America Journal 66, 1965e1973.

Bundt, M., Jaggi, M., Blaser, P., Siegwolf, R., Hagedorn, F., 2001. Carbon and nitrogendynamics in preferential flow paths and matrix of a forest soil. Soil ScienceSociety America Journal 65, 1529e1538.

Chenu, C., Hassink, J., Bloem, J., 2001. Short-term changes in the spatial distributionof microorganisms in soil aggregates as affected by glucose addition. Biologyand Fertility of Soils 34, 349e356.

Chenu, C., Plante, A.F., 2006. Clay-sized organo-mineral complexes in a cultivationchronosequence: revisiting the concept of the primary organo-mineralcomplex. European Journal of Soil Science 57, 596e607.

Denef, K., Six, J., Bossuyt, H., Frey, S.D., Elliott, E.T., Merckx, R., Paustian, K., 2001.Influence of dryewet cycles on the interrelationship between aggregate,particulate organic matter, and microbial community dynamics. Soil Biology &Biochemistry 33, 1599e1611.

De Gryze, S., Six, J., Paustian, K., Morris, S.J., Paul, E.A., Merckx, R., 2004. Soil organiccarbon pool changes following land-use conversions. Global Change Biology 10,1120e1132.

Ellerbrock, R.H., Gerke, H.H., 2004. Characterizing organicmatter of soil aggregatecoatings and biopores by Fourier transform infrared spectroscopy. EuropeanJournal of Soil Science 55, 219e228.

Golchin, A., Oades, M., Skjemstad, J.O., Clarke, P., 1994. Soil structure and carboncycling. Australian Journal Soil Research 32, 1043e1068.

Grandy, A.S., Robertson, G.P., 2007. Land-use intensity effects on soil organic carbonaccumulation rates and mechanisms. Ecosystems 10, 58e73.

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

Hinsinger, P., Gobran, G.R., Gregory, P.J., Wenzel, W.W., 2005. Rhizosphere geometryand heterogeneity arising from root mediated physical and chemical processes.New Phytologist 168, 293e303.

Horn, R., Smucker, A.J.M., 2005. Structure formation and its consequences for gasand water transport in unsaturated arable and forest soils. Soil Tillage Research82, 5e14.

Jangid, K., Williams, M.A., Franzluebbers, A.J., Schmidt, T.M., Coleman, D.C.,Whitman, W.B., 2011. Land-use history has a stronger impact on soil microbialcommunity composition than aboveground vegetation and soil properties. SoilBiology & Biochemistry 43, 2184e2193.

Jasinska, E., Baumgartl, T., Wetzel, H., Horn, R., 2006. Heterogeneity of physico-chemical properties in structured soils and its consequences. Pedosphere 16,284e296.

Jassogne, L., Ganga, H., Chittleborough, D., McNeill, A., 2008. Distribution andspeciation of nutrient elements around micropores. Soil Science SocietyAmerica Journal 73, 1319e1326.

Jastrow, J.D., 1996. Soil aggregate formation and the accrual of particulate andmineral-associated organic matter. Soil Biology & Biochemistry 28, 665e676.

Kellogg Biological Station, 2011. Experimental Design. KBS, Hickory Coner, MI[online]. Available at: http://lter.kbs.msu.edu/about/experimental_design.php(accessed 31.07.11).

Killham, K., Amato, M., Ladd, J., 1993. Effect of substrate location in soil and soilpore-water regime on carbon turnover. Soil Biology & Biochemistry 25, 57e62.

Kravchenko, A.N., Wang, W., Smucker, A.J.M., Rivers, M.L., 2011. Long-term differ-ences in tillage and land use affect intra-aggregate pore heterogeneity. SoilScience Society America Journal 75, 1658e1666.

Kravchenko, A., Chun, H.-C., Mazer, M., Wang,W., Rose, J.B., Smucker, A.J.M., Rivers, M.Relationships between intra-aggregate pore structures and distributions ofEscherichia coli within soil macro-aggregates. Applied Soil Ecology, in press.

Lindquist, W.B., Venkatarangan, A., Dunsmuir, J., Wong, T.F., 2000. Pore and throatsize distributions measured from synchrotron X-ray tomographic images ofFontainebleau sandstones. Journal Geophysical Research 105B, 21509e21527.

Nunan, N., Wu, K., Young, I.M., Crawford, J.W., Ritz, K., 2003. Spatial distribution ofbacterial communities and their relationship with the micro-architecture ofsoil. FEMS Microbiology Ecology 1496, 1e12.

Oh, W., Lindquist, W.B., 1999. Image thresholding by indicator kriging. IEEE Trans-actions Pattern Analysis and Machine Intelligence 21, 590e602.

Or, D., Smets, B.F., Wraith, J.M., Dechesne, A., Friedman, S.P., 2007. Physicalconstraints affecting bacterial habitats and activity in unsaturated porousmedia e a review. Advances in Water Resources 30, 1505e1527.

Park, E.J., Smucker, A.J.M., 2005. Saturated hydraulic conductivity and porositywithin macroaggregates modified by tillage. Soil Science Society AmericaJournal 69, 38e45.

Park, E.J., Sul, W.J., Smucker, A.J.M., 2007. Glucose additions to aggregates subjectedto drying/wetting cycles promote carbon sequestration and aggregate stability.Soil Biology & Biochemistry 39, 2758e2768.

Paustian, K., Collins, H.P., Paul, E.A., 1997. Management controls on soil carbon. In:Paul, E.A., Paustian, K., Elliott, E.T., Cole, C.V. (Eds.), Soil Organic Matter inTemperate Agroecosystems. CRC Press, Boca Raton, pp. 15e49.

Peth, S., Horn, R., Beckmann, F., Donath, T., Fischer, J., Smucker, A.J.M., 2008. Three-dimensional quantification of intra-aggregate pore-space features usingsynchrotron-radiation-based microtomography. Soil Science Society AmericaJournal 72, 897e907.

Ruamps, L.S., Nunan, N., Chenu, C., 2011. Microbial biogeography at the soil porescale. Soil Biology & Biochemistry 43, 280e286.

SAS Institute, 2009. SAS User’s Guide. Version 9.2. SAS Inst., Cary, NC.Santos, D., Murphy, S.L.S., Taubner, H., Smucker, A.J.M., Horn, R., 1997. Uniform

separation of concentric surface layers from aggregates. Soil Science SocietyAmerica Journal 61, 720e724.

Schmidt, M.W.I., Torn, M.S., Abiven, S., Dittmar, T., Guggenberger, G., Janssens, I.A.,Kleber, M., Koügel-Knabner, I., Lehmann, J., Manning, D.A.C., Nannipieri, P.,Rasse, D.P., Weiner, S., Trumbore, S.E., 2011. Persistence of soil organic matter asan ecosystem property. Nature 478, 49e56.

Senthilkumar, S., Basso, B., Kravchenko, A.N., Robertson, G.P., 2009. Contemporaryevidence for soil carbon loss under different crop management systems andnever tilled grassland in the US corn belt. Soil Science Society America Journal73, 2078e2086.

Sexstone, A.J., Revsbech, N.P., Parkin, T.B., Tiedje, J.M., 1985. Direct measurement ofoxygen profiles and denitrification rates in soil aggregates. Soil Science SocietyAmerica Journal 49, 645e651.

Six, J., Elliott, E.T., Paustian, K., 1999. Aggregate and soil organic matter dynamicsunder conventional and no-tillage systems. Soil Science Society America Journal63, 1350e1358.

Six, J., Elliott, E.T., Paustian, K., 2000. Soil macroaggregate turnover and micro-aggregate formation: a mechanism for C sequestration under no-tillage agri-culture. Soil Biology & Biochemistry 32, 2099e2103.

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

Strong, D.T., De Wever, H., Merckx, R., Recous, S., 2004. Spatial location of carbondecomposition in the soil pore system. European Journal Soil Science 55,739e750.

Urbanek, E., Hallett, P., Feeney, D., Horn, R., 2007. Water repellency and distributionof hydrophilic and hydrophobic compounds in soil aggregates from differenttillage systems. Geoderma 140, 147e155.

Urbanek, E., Smucker, A.J.M., Horn, R., 2011. Total and fresh organic carbon distri-bution in aggregate size classes and single aggregate regions using natural 13C/12C tracer. Geoderma 164, 164e171.

von Lützow, M., Kögel-Knabner, I., Ekschmitt, K., Matzner, E., Guggenberger, G.,Marschner, B., Flessa, H., 2006. Stabilization of organic matter in temperatesoils: mechanisms and their relevance under different soil conditions-a review.European Journal of Soil Science 57, 426e445.

Wang, W., Kravchenko, A.N., Smucker, A.J.M., Rivers, M.L., 2011. Comparison ofimage segmentation methods in simulated 2D and 3D microtomographicimages of soil aggregates. Geoderma 162, 231e241.

Wang, W., Kravchenko, A.N., Smucker, A.J.M., Liang, W., Rivers, M.L., 2012. Intra-aggregate pore characteristics: x-ray computed microtomography analysis. SoilScience Society America Journal 76, 1159e1171.

Watteau, F., Villemin, G., Burtin, G., Jocteur-Monrozier, L., 2006. Root impact on thestability and types of microaggregates in silty soil under maize. EuropeanJournal of Soil Science 57, 247e257.

Yoo, G., Spomer, L.A., Wander, M.M., 2006. Regulation of carbon mineralization ratesby soil structure and water in an agricultural field and a prairie-like soil. Geo-derma 135, 16e25.


can intra-aggregate pore structures affect the aggregate’s effectiveness in protecting carbon?

Documents