Anda di halaman 1dari 14

Geoderma 247248 (2015) 2437

Contents lists available at ScienceDirect

journal homepage:

Linkages between aggregate formation, porosity and soil

chemical properties
Inge C. Regelink a, Cathelijne R. Stoof a,b,c, Svetla Rousseva d, Liping Weng a,, Georg J. Lair e,f, Pavel Kram g,
Nikolaos P. Nikolaidis h, Milena Kercheva f, Steve Banwart g,i, Rob N.J. Comans a
Department of Soil Quality, Wageningen University, P.O. Box 47, 6700 AA Wageningen, The Netherlands
Department of Biological and Environmental Engineering, Riley Robb Hall, Cornell University, Ithaca, NY 1485. USA
Soil Geography and Landscape Group, Wageningen University, P.O. Box 47, 6700 AA Wageningen, The Netherlands
Institute of Soil Science, Agrotechnology and Plant Protection, N. Poushkarov, Shosse Bankya Str. 7, 1080 Soa, Bulgaria
Institute of Soil Research, University of Natural Resources and Applied Life Sciences, Peter-Jordan-Str. 82, 1190 Vienna, Austria
Institute of Ecology, University of Innsbruck, Sternwartestr. 15, 6020 Innsbruck, Austria
Department of Environmental Geochemistry and Biogeochemistry, Czech Geological Survey, Klarov 3, 11821 Prague 1, Czech Republic
School of Environmental Engineering, Technical University of Crete, University campus, 73100 Chania, Greece
Kroto Research Institute, North Campus, University of Shefeld, Shefeld S3 7HQ, UK

a r t i c l e i n f o a b s t r a c t

Article history: Linkages between soil structure and physicalchemical soil properties are still poorly understood due to the wide
Received 25 August 2014 size-range at which aggregation occurs and the variety of aggregation factors involved. To improve understanding
Received in revised form 18 December 2014 of these processes, we collected data on aggregate fractions, soil porosity, texture and chemical soil properties of
Accepted 30 January 2015
127 soil samples from three European Critical Zone Observatories. First, we assessed mechanistic linkages
Available online 25 February 2015
between porosity and aggregates. There was no correlation between the fractions of dry-sieved aggregates
(N1 mm, DSA) and water-stable aggregates (N0.25 m, WSA). Soil microporosity and micro + mesoporosity
Aggregate stability increased with increasing abundance of aggregates, though this correlation was only signicant for the WSA frac-
Critical Zone Observatories tion. The fraction of DSA did not affect the overall porosity of the soil, but affected the ratio between micro- and
Iron-(hydr)oxides mesopores (30 kPa/0.25 kPa), suggesting that micropores are dominantly located within DSA whereas mesopores
Porosity are located in between DSA and loose particles. Second, we studied the relations between the physical and chemical
Soil organic matter soil properties and soil structure. Soil texture had only a minor effect on the fractions of WSA and DSA whereas Fe-
Water retention (hydr)oxide content was correlated positively with both WSA fraction and porosity. This may be attributed to Fe-
(hydr)oxides providing adsorption sites for organic substances on larger minerals, thereby enabling poorly reactive
mineral particles to be taken up in the network of organic substances. The fraction of WSA increased with an in-
crease in the soil organic carbon (SOC) and Fe-(hydr)oxides content and with a decrease in pH. This pH-effect
can be explained by the enhanced coagulation of organically-coated particles at a lower pH. Overall, this study in-
dicates that mechanistic linkages exist between soil chemical properties, aggregate formation and soil porosity.
2015 Elsevier B.V. All rights reserved.

1. Introduction limits our ability to improve soil structural properties through manage-
ment. Therefore, there is a need for a mechanistic understanding of the
Soils provide major ecosystem services such as food production, physical and chemical key factors controlling the formation of soil
water retention and sequestration of soil organic carbon (SOC). The aggregates from loose mineral particles.
physical structure of the soil plays a crucial role in the processes that Soil structure has been dened as the size, shape and characteristics
facilitate these soil functions. Yet, intensive land use and poor agricul- of soil particles, aggregates, and pores across the size-range from nano-
tural management practices have led to declines in soil structure, lead- meters to centimeters (Bronick and Lal, 2005; Oades, 1993). This deni-
ing to a decrease in the agricultural productivity and SOC stabilization tion implies that soil structure is the product of various structural
capacity of soils worldwide (Banwart, 2011; Montgomery, 2007). Soil entities that cover a size range of several orders of magnitude. According
structure is the result of multiple physical, chemical and biological inter- to the widely used aggregate hierarchy theory (Oades and Waters,
actions in the soil. However, the complex interactions between chemi- 1991), soil aggregates and soil pores are formed in a hierarchical man-
cal properties and soil structure are still poorly understood, which ner meaning that microaggregates (b0.25 mm) are the building blocks
for macro-aggregates (N0.25 mm) and that aggregation is controlled
Corresponding author. by different aggregation factors at each size scale. In order to verify
E-mail address: (L. Weng). the aggregate hierarchy theory, we aimed to quantify soil structure at
0016-7061/ 2015 Elsevier B.V. All rights reserved.
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 25

the various hierarchical levels by a combination of (i) the size distribu- and that SOC content and pH are dominant at the larger size scales.
tion and stability of the aggregates and (ii) the volume and size distribu- These hypotheses are tested with a dataset of physical and chemical
tion of the pores. Soil aggregates and soil porosity reect different soil properties collected from 12 different sites in three diverse
aspects of the structure, with micropores reecting much smaller struc- European Critical Zone Observatories (CZO's).
tural entities than mesopores, and water-stable aggregates (WSA)
reecting much stronger and smaller aggregates than dry-sieved aggre- 2. Material and methods
gates (DSA). Because pores are the voids within and between aggre-
gates, the number of aggregates and the pore volume are related. 2.1. Sites
However, experimental evidence for the relation between aggregate
fractions and porosity for various soil types is still scarce and limited The soil samples for this study were collected from 12 sampling sites
to a small number of soil samples collected for a narrow range of soil at three European Critical Zone Observatories (CZO's), which are de-
types (Boix Fayos, 1997; Nyamangara et al., 2001). Insight into these re- scribed in detail by Banwart et al. (2012). The CZO's are characterized
lationships can help to understand how the various hierarchal levels of by different parent material, land use, and climate and are located in
aggregate formation are related. Austria (Marchfeld CZO), in the Czech Republic (Slavkov Forest CZO),
The interplay between the various physical and chemical processes and on the island of Crete, Greece (Koiliaris CZO):
involved in aggregate formation is highly complex and affected by a The Marchfeld CZO is located on the former oodplains of the River
range of soil properties. Various factors have been reported in literature Danube downstream of Vienna in Austria. This CZO includes six sites
to affect or control aggregation (Bronick and Lal, 2005), including (i) soil that are part of a chronosequence (2020,000 years old) (Lair et al.,
texture, (ii) the amount of metal-(hydr)oxides, (iii) the amount and 2009a). Climate is temperate with dry and hot summers, and soils are
characteristics of organic binding agents and (iv) pH and the concentra- covered by deciduous foresty and annual crops. The soils, classied as
tion of multivalent cations. Clay minerals affect aggregation because of Fluvisols and Chernozems, are calcareous and the clay mineralogy is
their interactions with organic matter (OM), leading to the formation dominated by illite and chlorite (Banwart et al., 2012). This CZO has
of organo-mineral assemblages (Angers, 1998; Fernndez-Ugalde been extensively studied, in particular in terms of soil age (Lair et al.,
et al., 2013; Six et al., 2000a). Similar, Fe-(hydr)oxides play an important 2009a), phosphorus speciation and adsorption (Lair et al., 2009b;
role in the formation of organic-mineral assemblages due to their strong Zehetner et al., 2008) and carbon sequestration (Zehetner et al., 2009).
interactions with humic acids (HA) (Asano and Wagai, 2014; Lehtinen The Koiliaris CZO is located on Crete, east of the city of Chania. Cli-
et al., 2014; Pronk et al., 2011). However, the role of Fe-(hydr)oxides mate is Mediterranean and soils are covered by natural vegetation
is mainly studied in oxide-rich soils (Barths et al., 2008; Duiker et al., (shrubs and herbaceous species). The three sites have been formed at
2003; Igwe et al., 2009); their role in soils with low Fe-(hydr)oxide con- different sedimentary deposits: uvial, unconsolidated and consoli-
tents is still poorly understood. OM plays a crucial role in the aggrega- dated bedrock. Soils are classied as Fluvisols (site K1,2) and
tion of mineral particles and is by far the most intensively studied Cambisols (site K3), respectively, and are degraded due to deforesta-
aggregation factor in literature (Angers, 1998; Bronick and Lal, 2005; tion for cropping and grazing. Organic matter contents and aggregate
Dilkova et al., 1998; Pulleman and Marinissen, 2004; Six et al., 2004). stability are therefore low (Stamati et al., 2013) and soils are suscep-
OM is a complex mixture of multiple organic substances such as tible to erosion and desertication. Further information about this
humic-like substances, polysaccharides, proteins, lignin and poorly sol- CZO can be found in Stamati et al. (2013a) and Moraetis et al. (2014).
uble aliphatic compounds (Calabi-Floody et al., 2011; Wattel-Koekkoek The Slavkov Forest CZO, in western Czech Republic, is located in a na-
et al., 2001) and this range of properties enables OM to affect aggrega- tional protected area. Climate is temperate and the soils are used for for-
tion via different mechanisms and at different scales (Tisdall and estry (mainly Norway spruce). This CZO includes three sites, which are
Oades, 1982). Another factor that affects aggregation is the amount of located on the slopes of three rst-order catchments (Lysina, Pluhuv
exchangeable multi-valent cations, such as Ca2+, Al3+ and Fe3+, since Bor, Na Zelenem). The catchments are characterized by different parent
the coagulation behaviour of particles and the strength of the interac- material (granite, serpentinite and amphibolite) and therefore by three
tions between OM and mineral particles strongly depends on the con- distinct soil types (Podzol, Stagnosol and Cambisol, respectively). They
centrations of multi-valent cations (Curtin et al., 1994; Kaiser et al., therefore show strong differences with respect to their mineralogy
2012). Soil pH has a strong effect on the concentration of multivalent and chemistry (Table 1). Detailed information about the chemistry
cations in the soil solution because, in acid soils, the concentration of and hydrology of the three catchments can be found in Kram et al.
metal ions is often controlled by the solubility of their respective (2012) and Banwart et al. (2012). Site coordinates are given in Table 1.
metal-(hydr)oxides and in calcareous soils, the solubility of CaCO3 con-
trols the concentration of Ca (Weng et al., 2011). 2.2. Soil sampling
Despite the extensive knowledge on aggregation mechanisms
(Bronick and Lal, 2005; Six et al., 2004), a quantitative understanding Soils samples were taken from all 12 sampling sites in spring 2010.
of the relationships between physical/chemical properties and soil Three soil prole pits were dug at each site, yielding a total of 36 pits.
structure is still lacking. Therefore, our aim was to assess the contribu- For the rst soil pit, soil samples were taken from each distinct soil
tion of the different aggregation factors on formation of soil structure horizon down to the C-horizon, which resulted in 36 soil samples per
at the various hierarchical levels using porosity, water-stable aggregates pit. The C-horizon was always within 1.2 m of the surface. For the sec-
(WSA) and dry-sieved aggregates (DSA) as indicators for aggregation at ond and third pit, only the most prominent soil horizons were sampled
these different levels. We hypothesize that the volume of micropores is which resulted in 13 soil samples per pit. Soil samples taken from the
related to the abundance of small aggregates and that the volume of upper 10 cm were considered as topsoils whereas samples from deeper
mesopores is related to the abundance of larger aggregates. We further- horizons were considered as subsoils. For the 36 pits together, this sam-
more hypothesized soil structure to be the product of multiple physical pling design yielded a total of 126 samples: 39 top- and 87 subsoil sam-
and chemical interactions, and thus that aggregation is best predicted ples. Because of the heterogeneity of the eld sites and the differences in
by a combination of texture, Fe-(hydr)oxide content, organic matter sampling depth of the various samples taken from the different pits, all
content and pH. Finally, we hypothesized that the dominant aggrega- soil samples are regarded as individual samples in the data analyses.
tion factor differs for the different hierarchical levels of aggregate for- Bulk soil samples were taken for the analysis of soil texture, aggregates,
mation as are in this study reected by the various soil physical and soil chemical properties. The eld-moist samples were gently bro-
indicators, in which we expect that Fe-(hydr)oxides and clay particles ken up by hand so that the larger clods broke along natural ssures,
are the dominant aggregation mechanisms at the smallest size scales, and roots and any rock fragments were removed by hand. Undisturbed
26 I.C. Regelink et al. / Geoderma 247248 (2015) 2437

Table 1
General description of three European Critical Zone Observatories (CZO's) and their 12 sites.

Critical Zone Sitea Coordinates Elevationb Soil type Horizons Land Descriptionc
Observatory (CZO)a use/cover

Marchfeld CZO, M1 4807.688 157 Fluvisol AC, C Soft wood Island in Danube river-bed, soil age: b50 y. Calcareous,
Austria (n = 8) 1643.611 forest high Feox content (N3 g/kg), low clay content (b10%), SOC
(n = 64) content: 1525 g/kg. CEC dominated by Ca for all
Six sites with varying stage of Marchfeld soils.
soil development M2 4808.028 156 Fluvisol A1, A2, AC, Soft wood Forest site inside a ood control dam, soil age: 50400.
(n = 12) 1639.693 C forest Calcareous. SOC content: 1036 g/kg.
M3 4808.692 157 Chernozem A1, A2, AC, Hard-wood Outside ood control dike, soil age: 400800.
(n = 14) 1641.634 C forest Calcareous, SOC content: 1045 g/kg.
M4 4809.483 155 Chernozem A, AC1, Cropland Old cropland site, soil age: N2500 y. Calcareous, low Feox
(n = 9) 1641.135 AC2, C content (b1 g/kg), low SOC content (b20 g/kg)
M5 4811.709 148 Chernozem A1, A2, AC, Cropland Old cropland site, soil age: N2500 y. Calcareous, low Feox
(n = 13) 1644.712 C content (b1 g/kg), low SOC content (b20 g/kg)
M6 4817.005 165 Chernozem A, AE, Ab, Hard-wood Old forest site, soil age: N15,000 y. pH: 6.07.7, clay
(n = 8) 1639.834 BC, C forest content: 15%, Feox content: 1 g/kg, SOC
content b 24 g/kg.
Koiliaris CZO, K1 3526.900 10 Fluvisol A, AC, C1, Shrubland Degraded soil, Elevation: 10 m, parent material: uvial
Crete (n = 11) 2408.253 C2, C3 sediments. calcareous, clay content: 1523%, Feox content:
(n = 29) 0.21.5 g/kg
Three sites, varying in parent K2 3526.666 9 Fluvisol AC, C1, C2, Shrubland Degraded soil, parent material: unconsolidated
material and elevation (n = 9) 2408.209 C3 sedimentary rock, calcareous, low clay content (b6%), low
Feox content (b0.6 g/kg),
K3 3521.653 1062 Cambisol A, AC1 AC2 Shrubland Degraded soil, abandoned terrace. parent material:
(n = 9) 2404.308 AC3 consolidated sedimentary rock, pH 5.56.2, clay content:
1530%, Feox content: 2.24.2 g/kg
Slavkov Forest CZO, S1 502.085 890 Podzol AE, E, Bh, Norway Lysina catchment. Parent material: granite. pH: 3.34.0,
Czech Republic (n = 16) 1239.316 Bs, C spruce CEC dominated by Al. Translocation of Fe and Al. High
(n = 33) gravel content (2035%)
Three sites, varying in parent S2 503.840 773 Stagnosol AE, BE, Bt, Norway Pluhuv Bor catchment. Parent material: Mg-rich
material (n = 11) 12 46.908 Cr spruce serpentinite. pH 4.06.6, CEC dominated by Al and Mg.
High Feox content (26 g/kg)
S3 4959.785 771 Cambisol A, B Norway Na Zelenem catchment. Parent material: amphibolite. pH
(n = 6) 1242.530 spruce 3.55.0, CEC dominated by Al and Ca. High Feox content
(28 g/kg)
The number of soil samples taken per CZO and site are given between brackets. For each site, soil samples were taken from each distinct soil horizon from three different soil pits.
Elevation in metre above mean sea level.
SOC: soil organic carbon; Feox: oxalate-extractable Fe; CEC: cation-exchange-capacity.

soil cores (5 cm high, 5 cm diameter, 100 cm3) were taken for the soil pH (Pansu and Gautheyrou, 2006). The Fed content and CEC were
analysis of soil water retention and bulk density. Organic matter-rich measured for a selection of soils only and these data were solely used
horizons (O, AO, Ah; only present at the Slavkov Forest CZO) were for site description and not for statistical analysis.
2.4. Physical soil characterization
2.3. Chemical soil characterization
2.4.1. Soil texture
Homogenized soils samples were oven dried at 40 C and sieved 10 g air-dried soil was chemically dispersed in 25 ml 0.4 M
over 2 mm. The pH was measured potentiometrically in a soilwater Na-pyrophosphate (Na 4 P2 O7 ). The gravel (N2 mm) and sand
suspension at a soil/solution ratio (SSR) of 0.4 kg/l. The total C (TC) (0.052.0 mm) fractions were determined by wet sieving and the
and N (TN) content were determined by a CN-analyzer (Carlo Erba silt (0.0020.05 mm) and clay fraction (b 0.002 mm) were determined
Nitrogen analyzer 1500). For calcareous soils, the carbonate content by the pipette method (Pansu and Gautheyrou, 2006).
was measured gas-volumetrically after acidication (Pansu and
Gautheyrou, 2006) and the SOC content was calculated as the differ- 2.4.2. Dry-sieved aggregates
ence between TC and carbonate-C. The C:N ratio of the OM was dened The size-distribution of the dry aggregates was determined by
as the ratio between SOC and TN. The amount of amorphous Fe- and manual dry-sieving of air-dried bulk soil (500 g) for aggregate fractions
Al-(hydr)oxides were determined by an ammonium-oxalate extraction N10, 105, 53, 31 and b1 mm, respectively. The results of the dry-
at pH 3.0 (Pansu and Gautheyrou, 2006; Schwertmann, 1991) for which sieving were expressed as the mass of the aggregates larger than
suspensions (0.05 kg l 1) were shaken for 2 h (horizontal shaker, 1 mm calculated according to Eq. (1):
85 rpm) in the dark and the oxalate-extractable Fe (Feox) and Al
(Alox) concentrations were analysed in the 0.45-m ltered extracts DSA 1
by inductively coupled plasmaatomic emission spectrometry (ICP
AES; Varian Vista Pro). For the Marchfeld site, Alox was not determined.
where DSA is the fraction of dry sieved aggregates (w/w), wN1 mm is the
The total amount of Fe-(hydr)oxides (i.e., the amorphous and crystal-
mass of soil larger than 1 mm, and wtotal is the total mass of the soil.
line oxides) was determined by a dithionitecitratebicarbonate
Additionally, the dry sieving results were expressed as the mean
extraction (Fed). For the Marchfeld CZO, Fed was determined for all
weight diameter (MWDdry), calculated according to Eq. (2):
soil samples following the procedure of Mehra and Jackson (1960).
For the Koiliaris and Slavkov Forest CZO, Fed was determined following X
n w
the procedure of Holmgren (1967). The cation exchange capacity (CEC) MWDdry i
di 2
and exchangeable cations were determined using the BaCl2 method at i1
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 27

where MWD is the mean weight diameter (mm), n is the number of 3. Results
aggregate fractions (ve), di is the mean diameter of the ith fraction
and wi is the weight of soil in the fraction i expressed as a percentage For each CZO, mean, minimum and maximum values of the soil
of the dry soil mass. chemical and physical parameters are given in Table A1 and the R2
values from linear regression analysis are given in Tables A24. First,
2.4.3. Water-stable aggregates the results of the soil characterization of the three CZO's will be given
The fraction of water-stable aggregates was determined following followed by the relations between the physical and chemical properties.
the wet-sieving method similar to the method of Haynes and Swift
(1990) using 20 g of 13 mm sized dry-sieved aggregates. Kercheva 3.1. Soil characteristics
et al. (2011) found a very good relationship between the water stability
determined on a single aggregate fraction (31 mm) and a sample com- 3.1.1. Marchfeld CZO
posed of different aggregate fractions. The aggregates were immersed in The six sites of the Marchfeld CZO were calcareous (2386 g CaCO3/kg)
water and sieved manually over two sieves with mesh-sizes of 1 and and characterized by a neutral to alkaline pH (5.98.4) (Table A1). Tex-
0.25 mm by moving the sieves ten times up and down for 10 cm tural composition strongly differed among the six sites and clay content
while the aggregates remained under water. The mass of soil in each varied between 5 and 34%. The Feox content (0.3 and 1.5 g/kg) was low
fraction (13 mm, 0.251 mm, and b 0.25 mm) was determined after compared to the Fed content (2.5 and 8.1 g/kg). The Feox/Fed ratio
oven-drying (105 C). Subsequently, for the 13 mm fraction, the frac- (0.10.6) decreased with increasing soil age as previously described
tion of sand particles larger than 1 mm was determined by dispersion by Lair et al. (2009a, 2009b). As expected, SOC content was generally
of the aggregates in 0.4 M Na-pyrophosphate and subsequent sieving higher for topsoils (1545 g/kg) than for subsoils (525 g/kg). For the
over 1 mm. The fraction of water-stable aggregates (N0.25 mm), topsoils, the SOC content was lower in the arable elds than in the forest
corrected for the sand particles larger than 1 mm, was calculated elds, which may be explained by lower carbon inputs and enhanced
using Eq. (3): biodegradation due to mechanical disturbance of soil aggregates (Six
et al., 2002; Zehetner et al., 2009). SOC content correlated poorly with
wN0:25 wsandN1 Feox content (R2 = 0.31, Fig. A1) but was neither correlated with clay
WSA 3 (Fig. A1), nor with Fed content (Table A2).
wdry wsandN1

3.1.2. Koiliaris CZO

where wN0.25 is the mass of soil in the fraction larger than 0.25 mm after The three sites of the Koiliaris CZO distinctly differed with respect
wet-sieving, wsand N 1 is the mass of soil particles larger than 1 mm and to their textural and chemical composition (Table A1). Site K1
wdry is the mass of the initial dry-sieved aggregates (13 mm). For most (Fluvisol) was characterized by a high pH (7.48.1), high CaCO3 con-
soils, the coarse fraction amounted to less than 3% of the recovered soil tent (38160 g/kg), low Feox content (b0.5 g/kg) and high clay content
mass, except for the gravel-rich site S1 of the Slavkov Forest where (1523%). Site K2 (Fluvisol) was also calcareous but characterized by a
1050% of the soil mass was larger than 1 mm. The wet sieving proce- very low clay content (b6%) and low Feox content (b 0.5 g/kg). In con-
dure was performed in triplicate and the average of three replicates trast, site K5 (Cambisol) was characterized by a rather high Feox content
was taken. The standard deviations of the mass fractions calculated (2.04.2 g/kg), high clay content (1330%) and a slightly acid pH
from the triplicates was lower than 0.2 for all samples. (5.56.0). Despite the rather low Feox content, Fed content was high
for all sites (1025 g/kg, Table A1) and Fe ox/Fed ratios were conse-
2.4.4. Water retention and bulk density quently very low (0.050.15) indicating that the majority of the Fe-
Water retention and bulk density were measured in triplicate on (hydr)oxides in these soils is crystalline. Feox was strongly correlated
undisturbed soil cores (100 cm3). Bulk density was determined on with Alox (R2 = 0.89, Table A3). The SOC content (2.950 g/kg) was
oven-dried soils. For water retention, samples were brought to near- higher in topsoils than in subsoils (Fig. 1) and correlated positively
saturation (0.25 kPa) over the course of 20 days, after which the drying with the Feox content (R2 = 0.47, Fig. A1) but not with the clay fraction
branch of the water retention characteristic was determined at six (Fig. A1).
tensions (0.25, 1, 5, 10, 30, 1500 kPa) using a suction plate apparatus
(0.2530 kPa) and a membrane pressure plate device (1500 kPa) 3.1.3. Slavkov Forest CZO
(Dane and Topp, 2002). For each water potential, equilibrium was The three sites of the Slavkov Forest CZO covered a wide range in soil
established within 57 days. Standard deviations of the water contents, textural and chemical properties (Table A1). The gravel content was
calculated from the three lab replicates, were below 0.1 for all samples higher at site S1 (Podzol, 2030%) than at the other two sites (Stagnosol
and mean values were used in data analysis. Soil porosity was calculated and Cambisol, b 20%). Clay content varied between 3 and 25% and the
from the water retention curves. Following Luxmoore (1980), we de- pH values ranged between 3.3 and 6.6. Most soil samples were charac-
ned microporosity as the volume of water stored at 30 kPa (equivalent terized by a high Feox (1.513 g/kg) and Alox (0.59 g/kg) content,
to a pore size b 9 m) and mesoporosity as the volume of water stored except for the E-horizons in which the Feox and Alox content were typi-
between 0.25 and 30 kPa (equivalent to a pore size between 9 and cally low. The SOC content (217 g/kg) was higher in the topsoils and
1000 m, Table 2). B-horizons than in the E- and C-horizons. The SOC content showed a
positive correlation with the Feox content (R2 = 0.62) however, only
2.5. Statistical analyses when the three topsoils of the Podzol site were excluded from the cor-
relation (Fig. A1). In contrast, the SOC content did not correlate with the
Relationships between soil properties were assessed using single re- clay content (Fig. A1).
gression analysis, and stepwise multiple regression (MR) analysis and
partial least square (PLS) analysis. The PLS regression is used in a similar 3.2. Soil aggregates
way to multiple regression analysis, but PLS can handle correlated vari-
ables. The PLS regression creates a model of latent variables in which the 3.2.1. Dry-sieved aggregates
information from soil properties is combined in such a way that it max- Results from the dry-sieving were expressed as the MWDdry and as
imizes the explanation of variation in soil structure. Prior to the PLS the fraction of DSA larger than 1 mm. Both parameters were strongly
modelling, all variables were scaled to unit variance. All statistical calcu- correlated (R2 = 0.61, Fig. A2) and therefore, the results presented
lations were performed with Genstat 13th (VSN International Ltd, UK) here are restricted to the fraction of DSA. The fraction of DSA showed
28 I.C. Regelink et al. / Geoderma 247248 (2015) 2437

Table 2
Overview of the soil physical parameters.

Parameter Pore size/aggregate size Method Comment

Dry-sieved aggregates Aggregate size: 110 mm Dry sieving of air-dried soil Macro-aggregates
Water-stable aggregates Aggregate size: 0.253 mm Wet-sieving of 13 mm sized dry-sieved aggregates. Macro-aggregates with a high mechanical stability
Micro + mesoporosity Pore size: b1000 m Volumetric water content at 0.25 kPa Sum of micro- and mesoporosity
Mesoporosity Pore size: 91000 m Volumetric water content stored between 0.2530 kPa Gravitational water
Microporosity Pore size: b9 m Volumetric water content at 30 kPa Plant-available and unavailable water
Indicator for micro-aggregates and micro-structure

a poor correlation with the SOC content for the Marchfeld (R2 = 0.26), The Feox content was the best single predictor for WSA in the
Fig. 1d, whereas this correlation was not signicant for the other two Marchfeld (R2 = 0.56, Fig. 2e) and the combination of Feox, SOC and
CZO's (Fig. 1e,f). For the Koiliaris CZO, the fraction of DSA correlated pos- pH explained 72% of the variance in WSA in soils from this CZO
itively with the clay content (R2 = 0.65, Fig. 1b), however, this correla- (Table 3). Feox was also correlated with the fraction of WSA in the
tion was dominantly caused by the low fraction of DSA in the clay-poor Koiliaris CZO (R2 = 0.45, Fig. 2f). In contrast, the fraction of WSA did
soils of site K2 (Fluvisol). For the other CZO's, the clay content did not not correlate with Feox for the Slavkov Forest due to the high fraction
correlate with the fraction of DSA (Fig. 1a,c). Neither was there a corre- of WSA in the iron-poor soils of the Podzol site, meaning that Fe-
lation with the Feox content or any other soil chemical parameter (hydr)oxides were not a preriquisite to form stable aggregates in this
(Tables A2-4). acid forest soil (Fig. 2g). PLS analysis showed that the Feox + Alox con-
tents explain a similar percentage of the variance as the Feox content
(Fig. A4). Furthermore, PLS analysis showed that the combination of
3.2.2. Water-stable aggregates Feox, SOC and pH explains most of the variance in the WSA fraction.
The fraction of WSA was generally higher in the topsoils than in the For the Marchfeld CZO, Feox is the dominant explaining factor whereas
subsoils and correlated positively with the SOC content for all CZO's SOC is the dominant explaining factor in the Koiliaris and Slavkov Forest
(Marchfeld R2 = 0.42, Koiliaris R2 = 0.85 and Slavkov Forest R2 = CZO. However, due to the co-correlation between Feox and SOC, this
0.76, Fig. 2ac). However, the correlation between WSA and SOC was analysis does not indicate that SOC is the only explaining factor. Instead,
poor when the data of the three CZO's were combined because a similar Fe-(hydr)oxides likely also contribute to the formation of WSA. Correla-
SOC content corresponded to a higher fraction of WSA for acid soils than tions between the fraction of WSA and soil texture were not signicant
for neutral or alkaline soils (Fig. 2d, 3). The role of pH was further con- for any of the CZO's (Tables A24).
rmed by multiple regression analysis, which showed that the fraction
of WSA in the three CZO's was best predicted from the combination of 3.3. Soil porosity
SOC and pH. This combination explained 63% of the observed variance
in WSA whereas SOC alone explained only 31% of the observed variance 3.3.1. Microporosity
(Table 3). The coefcient for the pH is negative meaning that the frac- For the Marchfeld CZO, microporosity correlated well with silt
tion of WSA decreased with increasing pH. content (R2 = 0.57, Fig. 3e) whereas this correlation was poor in the

a b c

d e f

Fig. 1. The fraction of dry-sieved aggregates larger than 1 mm (DSA, w/w) as a function of the clay content (ac) and the soil organic carbon (SOC) content (df) in the Marchfeld, Koiliaris
and Slavkov Forest CZO. Different colours represent the various sites and symbols with crosses represent the topsoils (010 cm). Given regression lines are signicant at p b 0.0001
(Marchfeld) or p b 0.001 (Koiliaris, Slavkov Forest).
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 29

a b c d

e f g h

Fig. 2. The fraction of water-stable aggregates larger than 0.25 mm (WSA, w/w) as a function of the soil organic carbon (SOC) content (ad) and the amorphous Fe-(hydr)oxide (Feox)
content (eh) in the Marchfeld, Koiliaris, Slavkov Forest CZO. Given regression lines are signicant at p b 0.0001 (Marchfeld) or p b 0.001 (Koiliaris, Slavkov Forest). Fig. 2d and h show
the data for the three CZO's grouped as a function of the pH. The lines in Fig. 2d show the model predictions from the multiple regression model given in Table 3 (WSA = 1.2 + 0.18
ln(SOC) 0.11 pH).

Koiliaris CZO (R2 = 0.32) and not signicant in the Slavkov Forest (Fig. 5c). Correlations with the fraction of WSA were stronger for
(Fig. 3f, g). Silt was the best single predictor for microporosity when micro + mesoporosity than for microporosity alone (Fig. 4, 5). As
the data of the three CZO's were combined (R2 = 0.45, Fig. 3h) and for microporosity, micro + mesoporosity correlated strongly with
was a better predictor for microporosity than clay (Fig. S4). Feox showed Feox content for all three CZO's (Fig. 5mo) as well as with SOC con-
a strong correlation with microporosity for the three CZO's (Fig. 3ac). tent for the Marchfeld and Koiliaris CZO (Fig. 5i,j). Furthermore,
However, Feox was not a good predictor for microporosity when the micro + mesoporosity correlated strongly with the bulk density for
data of the three CZO's were lumped because the slope of the regression all three CZO's together (Fig. A5).
line was different for the Slavkov Forest than for the other two CZO's
(Fig. 4d). Multiple regression analysis showed that the microporosity 3.3.4. The ratio between microporosity and micro + mesoporosity
in the Marchfeld and Koiliaris CZO was best predicted from the combi- (30 kPa/0.25 kPa)
nation of texture and WSA, or from the combination of Feox and texture The ratio between microporosity and micro + mesoporosity
(Table 4, Fig. A4). The importance of Feox was conrmed by the PLS (30 kPa/0.25 kPa ) showed a positive correlation with the fraction of
results which show that Feox is the dominant explaining variable with DSA for the Marchfeld (R2 = 0.69), Koiliaris (R2 = 0.92) and Slavkov
smaller contributions of texture and SOC for all CZO's (Fig. A4). Forest CZO (R2 = 0.33) (Fig. 6ac). For the three CZO's together, DSA
explained 70% of the observed variance in the 30 kPa/0.25 kPa ratio
3.3.2. Mesoporosity (Fig. 6d). In contrast, correlations between DSA and the microporosity
Mesoporosity was generally lower than microporosity and did not itself were poor or not signicant (Tables A24). Thus, DSA did not
vary much between top- and subsoils (Fig. A3). For the Koiliaris CZO, affect the overall volume of pores but an increase in DSA led to an in-
mesoporosity showed a poor correlation with silt content and this cor- crease in the fraction of the micropores and a consequent decrease in
relation was not signicant for the Koiliaris and Slavkov Forest CZO the fraction of mesopores.
(Fig. A3df). Feox content correlated poorly with mesoporosity in the
Marchfeld and Slavkov Forest (Fig. A3g, A3i). For the Koiliaris CZO, 4. Discussion
Feox correlated strongly with mesoporosity when the data are split up
for the sandy Fluvisol (site K2, N 64% sand) and the silty Fluvisol and 4.1. Linkages between aggregate fractions and porosity
Cambisol (site K1,3 N 45% silt) indicating that an increase in Feox content
corresponds to an increase in mesoporosity in soils with a similar tex- The WSA fraction, which reects the small and water-stable
ture (Fig. A3h). In contrast, correlations between the SOC content and aggregates, is affected by soil depth and land use (Fig. 2) while the
mesoporosity were not signicant for any of the CZO's (Fig. A3jl). DSA fraction, which reects large and loose aggregates, is not (Fig. 1).
Previous studies also found WSA to be a more sensitive indicator for dif-
3.3.3. Micro + mesoporosity ferences in soil structure among land use types (Pulleman et al., 2003;
Micro + mesoporosity correlated well with the fraction of WSA for Six et al., 2000b) and therefore, wet-sieving has become a popular
the Marchfeld (R2 = 0.50) and Koiliaris CZO (R2 = 0.75) (Fig. 5a and approach to characterize aggregates in soils (Amzketa, 1999;
b), but this correlation was not signicant for the Slavkov Forest CZO Kercheva et al., 2011). In the present study, wet-sieving was performed
30 I.C. Regelink et al. / Geoderma 247248 (2015) 2437

a b c d

e f g h

i j k l

m n o p

Fig. 3. Microporosity (micro) as a function of the fraction of water-stable aggregates (WSA, ad) the silt content (eh), the soil organic carbon (SOC) content (il) and the amorphous
Fe-(hydr)oxide (Feox) content (mp) in the Marchfeld, Koiliaris, and Slavkov Forest CZO and in the three CZO's together. Microporosity was dened as the volumetric water-content
stored at 30 kPa, which is equivalent to the volume of pores smaller than 9 m. Given regression lines are signicant at p b 0.0001 (Marchfeld) or p b 0.001 (Koiliaris, Slavkov Forest).
Symbols with crosses represent top-soils (010 cm).

on the 13 mm sized dry-sieved aggregates. Therefore, the results of the Since DSA and WSA reect two different types of aggregates, they
wet-sieving reects the breakdown of large DSA into small and stable also differ in their relation with soil porosity. The DSA fraction did not
WSA (N0.25 mm) and dispersible particles or aggregates b 0.25 mm. correlate with microporosity or micro + mesoporosity (Tables A24).
The fraction of WSA covered a wide range (195%, Fig. 2), meaning In contrast, the WSA fraction correlated well with microporosity and
that DSA can be built-up from WSA and dispersible particles in any micro + mesoporosity (Figs. 4, 5) for the Koiliaris and Marchfeld sites
ratio. As such, our results show that there is a strong variation in (Fig. 4). Multiple regression analysis showed that soil porosity is the
strength or stability of DSA. This, and the lack of a correlation between sum of the inherent porosity, determined by its texture, and the addi-
DSA and WSA (Fig. A2), show that these two aggregate fractions reect tional porosity due to the presence of WSA (Table 4). The Slavkov Forest
a different type of aggregate and are controlled by different aggregation site deviated from these trends (Figs. 4, 5) and the poor correlations be-
mechanisms. tween the WSA and porosity in this CZO can be explained by the large
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 31

Table 3 any larger size. The limited correlation between texture and the fraction
Multiple-regression relations between soil properties and the fraction of water-stable of DSA and WSA (Figs. 1, A4 and Table A2) was therefore not expected.
Only for the Koiliaris site, an effect of texture was observed since the
R2 formation of DSA was inhibited in the sandy Fluvisol (K2, Fig. 1)
Marchfeld CZO (n = 64) which has a very low clay (b6%) and Feox content (b0.5 g/kg) compared
WSA = 2.2 + 0.16 Feox 0.25 pH 0.67 to other sites. For the other CZO's, the size and stability of the aggregates
WSA = 1.80 + 0.13 Feox + 0.086 SOC 0.22 pH 0.72 were not correlated with the clay, silt or sand fractions (Tables A24)
Koiliaris CZO (n = 36) despite the wide range in texture among the studied soils. The absence
WSA = 0.11 + 0.28 ln(SOC) 0.85 of a correlation between the clay content and the fraction of WSA is in
Slavkov Forest CZO (n = 33)
contrast with previous observations that argued that clay minerals are
WSA = 0.89 + 012 ln(SOC) 0.09 pH 0.83 important for aggregation because of their interactions with OM
(Angers, 1998). However, for the three CZO's used in this study, the
All (n = 127)
clay fraction is not the dominant surface for sorption of OM, since
WSA = 1.2 + 0.18 ln(SOC) 0.11 pH 0.66
there was no correlation between the clay fraction and the SOC content
Best ts obtained by stepwise multiple-regression using the soil physicalchemical
(Fig. A1). Instead, in our soils, Fe-(hydr)oxides seem to dominate the
properties as possible parameters. WSA = fraction of water-stable aggregates (w/w);
SOC = soil organic carbon (%); Feox = oxalate-extractable Fe (g/kg). Multiple regression sorption capacity for OM since the Feox content correlated moderately
models are signicant at p b 0.001 and the coefcients of the included variables are signif- with the SOC content (Fig. A1). The role of the clay fraction in aggregate
icant at p b 0.05. formation also depends on the mineralogy of the clay fraction (Six et al.,
2000a). In the studied CZO's, the clay fractions are dominated by non-
swelling clay minerals such as illite, kaolinite and chlorite (Banwart
et al., 2012). Swelling clay minerals, such as montmorillonite, can have
amount of gravel (N 2 mm) in the soils from this CZO, which results in a much stronger effect on dispersion of aggregates than non-swelling
discrepancies between water retention data, measured on undisturbed clay minerals (Brubaker et al., 1993). For the Marchfeld CZO, the miner-
soils, and aggregate fractions, measured on the 2 mm sieved soils. Both alogy of the clay fraction is similar among the studied sites (Lair et al.,
microporosity and micro + mesoporosity are related to the fraction of 2009a, 2009b), which means that the lack of correlation between the
WSA, meaning that WSA affects porosity over wide range of pore sizes clay fraction and soil aggregate fractions cannot be attributed to miner-
(01000 mm). It is therefore unlikely that the relation between WSA alogical differences. Overall, soil texture affected the inherent porosity
and porosity is solely due to the formation of aggregates with a specic of the soil however, with respect to aggregation, correlations between
size. Instead, we suggest that an increase in the fraction of WSA corre- texture and the various aggregate fractions were absent or weak.
sponds to an increase in aggregation over a wide range of size scales
and that, as a result, the WSA fraction corresponds to an increase in 4.2.2. Iron-(hydr)oxides
porosity over a wide range of pore sizes. Iron-(hydr)oxides play a crucial role in the binding of particles to
The greater abundance of DSA affects the distribution of aggregates as is shown by the strong correlations between the Feox con-
soil pores between micropores and mesopores. This distribu- tent and the micro- and micro + mesoporosity in the three CZO's
tion, expressed as the ratio between microporosity and micro + (Figs. 4, 5) as well as by the correlations between the Feox content and
mesoporosity (30 kPa/0.25 kPa), correlates strongly with the fraction of the fraction of WSA in the Marchfeld and Koiliaris site (Fig. 2). To our
DSA when the data of the three CZO's are lumped together (Fig. 6). knowledge, a strong, positive correlation between the Feox content
Thus, an increase in the fraction of DSA corresponds to an increase in and soil porosity has not been shown before in literature. The contribu-
the relative contribution of micropores. This suggests that micropores tion of Fe-(hydr)oxides to aggregation is strongest at the smallest size
are dominantly located within DSA whereas mesopores are dominantly scales of aggregation since the correlations between Feox and the soil
located in between loose particles, which may also include aggregates physical properties are strong for the porosity, moderate for the WSA
smaller than 1 mm (Fig. 6). The slope of the regression line between and absent for the DSA. Thus, Fe-(hydr)oxides play an important role
DSA and the 30 kPa/0.25 kPa ratio was very similar for the three CZO's in aggregation, despite the small contribution of the Feox content to
(Fig. 6), which shows that this relation appears to occur among a the soil mass (b1%). This can be explained by the small size (b10 nm)
range of soil types having different textural compositions. The relation and consequently large surface-to-mass ratio (1001000 m2g 1) of
between aggregate formation and porosity has been described in previ- Fe-(hydr)oxides in soils (Eusterhues et al., 2005; Hiemstra et al., 2010;
ous conceptual models (Currie, 1966; Dexter, 1988; Wsten et al., Mikutta et al., 2006; Regelink et al., 2014). As a result, Fe-(hydr)oxides
2001) however, experimental evidence for the linkages between aggre- generally dominate the surface area available for sorption of SOC
gate fractions of varying mechanical strength and soil porosity are and consequently play an important role in the stabilization of SOC
scarce and have not been shown before on a data set that includes a and in the formation of organo-mineral assemblages within soils
range of soil types, such as the dataset assessed in this study. Overall, (Kgel-Knabner et al., 2008; Kaiser et al., 2011; Regelink et al., 2013).
our results show the WSA fraction correlates with the porosity of the The interaction between Feox and SOC is also evident from the correla-
soil, irrespective of pore size, whereas the DSA fraction correlates with tion between these two parameters, and PLS analysis suggests
the distribution of soil pores between micro- and mesopores. that both Fe ox and SOC contribute to formation of pores (Fig. A4).
Fe-(hydr)oxides have the ability to form strong bindings with clay,
4.2. Linkages between soil physical and chemical properties, aggregates and silt and sand particles, meaning that Fe-(hydr)oxides can provide
porosity sorption sites for organic substances on the surfaces of these other-
wise low to moderately reactive particles (Arias et al., 1996; Sei
The relations between the physical and chemical properties, aggre- et al., 2002). Some studies have suggested that these poorly reactive
gates, and porosity are discussed for four aggregation factors: (i) soil minerals, which are dominantly present in the silt and sand fraction,
texture, (ii) Fe-(hydr)oxides, (iii) organic binding agents and (iv) coag- are not actively involved in the aggregation process (Elliott et al.,
ulation conditions (pH). 1991; Six et al., 2000b). However, fractionation studies have shown
that silt and sand particles are the dominant constituents of water-
4.2.1. Soil texture stable macro-aggregates (Barberis et al., 1991; Puget et al., 2000).
Soil texture is a measure for the size-distribution of the primary Furthermore, a substantial part of the Fe-(hydr)oxides in soils are
mineral particles which are the building blocks for all aggregates at associated with silt and sand particles. For example, Barberis et al.
32 I.C. Regelink et al. / Geoderma 247248 (2015) 2437

a b c d

e f g h

i j k l

m n o p

Fig. 4. Micro + mesoporosity (micro + meso) as a function of the fraction of water-stable aggregates (WSA ad), the silt content (eh), the soil organic carbon (SOC) content (il) and the
amorphous Fe-(hydr)oxide (Feox) content (mp) in the Marchfeld, Koiliaris and Slavkov Forest site, and the three CZO's together. Micro + mesoporosity was dened as the volumetric
water-content stored at 0.25 kPa, which is equivalent to the volume of pores smaller than 1 mm. Given regression lines are signicant at p b 0.0001 (Marchfeld) or p b 0.001 (Koiliaris,
Slavkov Forest). Symbols with crosses represent top-soils (010 cm).

(1991) showed, for a range of soils, that about 50% of the Feox content Thus, Fe-(hydr)oxides prevent dispersion of these soil particles and
of the soil is associated with the silt and sand fraction. A small mass thereby enhance the formation of small-scale aggregates and soil pores.
fraction of Fe-(hdyr)oxides can already signicantly increase In addition to the amorphous Fe-(hydr)oxides, other metal-
the surface area of particles in the silt/sand range due to the extremely (hydr)oxides may also enhance aggregation. Considering the role
small size of the Fe-(hydr)oxides compared to the silt/sand particles. of crystalline Fe-(hydr)oxides rst, the lack of a correlation between
Therefore, we suggest that this coating of Fe-(hydr)oxides facilitates the Fed content and the porosity and fraction of WSA in the Marchfeld
the interaction between larger mineral particles and organic sub- CZO (Table A2) shows that the crystalline Fe-(hydr)oxides (Fed minus
stances and we argue that this Fe-(hydr)oxide coating is in particu- Feox) are of less importance than the amorphous Fe-(hydr)oxides
lar effective for quartz and other poorly reactive minerals that (Feox), which is in agreement with previous studies (Duiker et al.,
otherwise lack the necessary surface sites to interact with organic sub- 2003). Considering the role of Al-(hydr)oxides, PLS results show that
stances and to be taken up within the network of organic substances. the sum of the Feox and Alox content explains a smaller or a similar
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 33

Table 4 (Fig. 3). For acids soils, the same amount of SOC corresponds to a higher
Multiple-regression relations between soil properties and microporosity.a fraction of WSA compared to soils with a neutral to slightly alkaline pH
R2 (Figs. 2, 3). This pH-effect reects the role of the coagulation conditions,
Marchfeld + Koiliaris (n = 100)
which will be discussed in detail in Section 4.2.4. In addition to SOC, Feox
micro = 0.1 + 0.004 silt + 0.21 WSA 0.72 also correlates with WSA and porosity and both aggregation factors
micro = 0.03 + 0.07 Feox 0.58 seem to contribute to the formation of aggregates (Fig. A4). The fact
micro = 0.02 + 0.05 Feox + 0.002 silt 0.64 that the fraction of WSA is controlled by the combination of these
micro = 0.006 + 0.04 Feox + 0.002 silt 0.003 clay 0.70
factors rather than by the SOC content only, shows that, in addition
Slavkov Forest (n = 33) to the total SOC content, also the interactions between SOC and the
micro = 0.06 + 0.02 Feox 0.72 Fe-(hydr)oxide coatings as well as the interactions between SOC itself
All (n = 127) (coagulation) are of importance for aggregation. The importance of the
micro = 0.07 + 0.003 silt + 0.03 SOC 0.62 Fe-(hydr)oxide coating in facilitating the adsorption of OM on mineral
Best ts obtained by stepwise multiple-regression. micro = microporosity (m /m3);
3 surfaces has been shown in spectroscopic studies, in which OM on silt
WSA = water-stable aggregates (w/w) SOC = soil organic carbon (%); Feox = oxalate- and sand particles was shown to be present as patches associated
extractable Fe (g/kg). Multiple regression models are signicant at p b 0.001 and the with Fe-rich particles rather than as a homogenous layer (Chenu and
coefcients of the included variables are signicant at p b 0.05. Plante, 2006; Mikutta et al., 2006; Vogel et al., 2014). OM, in particular
HA, has a high afnity for adsorption to Fe-(hydr)oxides and will likely
form a monolayer on the Fe-(hydr)oxide surfaces with a thickness
of about 12 nm (Weng et al., 2008). Although this monolayer of
part of the variance in soil structure compared to the Feox content alone HA may enhance coagulation of clay and Fe-(hydr)oxide particles
(Fig. A4). This suggests that Fe-(hydr)oxides are more important for ag- (Baalousha, 2009; Majzik and Tombacz, 2007), a 12 nm thick layer
gregation than Al-(hydr)oxides, which is in line with the results of of HA is too thin to overcome the pore spaces in between silt- and
Duiker et al. (2003). Nevertheless, in other soil types, amorphous sand-sized particles. Thus, a thicker OM layer is needed in order to
Al-(hydr)oxides or allophane may contribute substantially to aggregate incorporate silt- and sand-sized particles within the aggregates.
formation, depending on the amounts of the respective minerals pres- The OM layer can be extended by the adsorption of other larger
ent in the soil (Asano and Wagai, 2014; Penn et al., 2001). Furthermore, organic compounds, in particular carbohydrates and glomalin. Car-
clay minerals may also acts as sorbents for humic substances, however, bohydrates and glomalin have a much lower afnity for adsorption
the surface area of Fe-(hydr)oxides in soils is usually much larger than to Fe-(hydr)oxides than humic acids (Kaiser, 2003) and their associ-
the surface area of clay minerals (Eusterhues et al., 2005). As such, ation with soil minerals can be explained by the zonal model of
Fe-(hydr)oxides seem to be the dominant reactive mineral involved in Kleber et al. (2007) which describes that organic molecules adsorb
aggregation of particles. in a layered structure on the mineral surfaces. In the contact zone
at the Fe-(hydr)oxide surface, HA is strongly bound via ligand
4.2.3. Organic binding agents exchange and electrostatic attraction. The hydrophobic parts of the
SOC correlates strongly with the fraction of WSA (Fig. 2) as well as adsorbed HA extend into the solution and act as a sorbent for other
with the porosity of the soil (Figs. 4, 5). Multiple regression analysis organic substances, such as carbohydrates and glomalin, that bind
showed that SOC, in combination with the pH, best predicts the forma- to the rst layer through hydrophobic interactions and form the sec-
tion of WSA for the data of the three CZO's together (Table 3). In con- ond layer. In the outer layer, organic substances are loosely bound
trast, our results do not show any effect of SOC on the formation of via cation bridging and hydrophobic interactions (Kaiser et al.,
DSA (Fig. 2). Thus, the contribution of SOC to the formation of aggre- 2012). Following this conceptual model (Kleber et al., 2007), the
gates seems to be most prominent at the level of WSA, i.e., the stable importance of carbohydrates and glomalin in aggregate formation
micro- to millimetre sized aggregates. The amount of SOC needed to (Rillig, 2004; Spohn and Giani, 2010) can be explained by their
form a certain number of WSA strongly differs among the different amphiphilic properties, which facilitates the formation of multiple
study sites (Fig. 2). Previous studies have suggested that these differ- OM layers. We suggest that this process leads to the formation of
ences are related to differences in the clay content (Asano and Wagai, an OM layer on the Fe-(hydr)oxide surfaces, which is sufciently
2014). In contrast, our results show that the pH partly explains the dif- thick to enable the formation of OM bridges between particles and
ferences in the number of aggregates formed at a certain SOC content to allow these particles to be taken-up within aggregates.

a b c d

Fig. 5. The ratio between microporosity and micro + mesoporosity (30/0.25) plotted against the fraction of dry-sieved aggregates (DSA, w/w) for the Marchfeld, Koiliaris, Slavkov Forest
CZO (ac), and three CZO's together (d). Microporosity was dened as the volumetric water-content stored at 30 kPa and the micro + mesoporosity was dened as the volumetric water-
content at 0.25 kPa. Regression lines are signicant at p b 0.0001 (Marchfeld) or p b 0.001 (Koiliaris, Slavkov Forest).
34 I.C. Regelink et al. / Geoderma 247248 (2015) 2437

Particle-particle interactions
clay/silt/sand particle
1 m

Fe-(hydr)oxide surface coating

Layer of humic acid

polysacharides, proteins 10 m

Water-stable aggregates

Mineral particles
100 m

Soil organic carbon

Water-stable aggregates and dry-sieved aggregates

mesopores, 9-1000 m
(gravitational water)
1 mm

micropores, <9 m
(plantavailable water)

dry-sieved aggregates
(1-10 mm)

loose particles (<1 mm)

water-stable aggregate
(>0.25 mm)
10 mm

Fig. 6. Conceptual diagram showing the different size-scales at which aggregate formation occurs and the factors controlling aggregate formation at each size scale, as explained in the
main text.

4.2.4. Coagulation conditions impact of pH coated particles. The effect of pH on aggregation cannot be solely attrib-
The strong increase in the WSA fraction with decreasing pH (Fig. 3, uted to the proton concentration itself since the pH also affects the
Table 3) can be explained by two pH-dependent processes. First, solubility of multi-valent cations, such as Ca2 +, Al3 + and Fe3 +. For
adsorption of HA to Fe-(hydr)oxides and clay minerals increases with example, the concentrations of Al3 + and Fe3+ strongly increase with
decreasing pH (Feng et al., 2005; Weng et al., 2008) meaning that a decreasing pH due to dissolution of Fe- and Al-(hydr)oxide minerals.
lower pH corresponds to a higher loading of HA on the mineral surfaces. In calcareous soils, the Ca2+ concentration is controlled by the solubility
Secondly, the tendency of organic and mineral particles to coagulate in- of CaCO3 and the Ca2 + concentration decreases with increasing pH
creases with decreasing pH, which can be explained by a decrease in the (Weng et al., 2011). Aggregate formation increases with increasing con-
electrostatic repulsive forces among negatively charged substances centrations of Ca2+, Al3+ and Fe3+ in the soil solution due to formation
(Nguetnkam and Dultz, 2011; Oste et al., 2002; Tombacz and Szekeres, of cation bridges between organically-coated particles and clay minerals
2004). The surface potential of organically-coated particles is controlled (Bronick and Lal, 2005). As such, the pH-dependent solubility of multi-
by the surface potential of the adsorbed OM (Weng et al., 2011) and was valent cations in soils amplies the above-described pH effect. However,
therefore predicted by the NICA-Donnan model (Fig. 3). For soils with a pH is only a rough indicator for the concentration of multi-valent
SOC content below 15 g/kg, the fraction of WSA followed the predicted cations in the soil solution. Additionally, CEC and composition of the
surface potential of HA indicating that the stability of the aggregates is exchangeable cations, in particular the contribution of Na and Mg, will
controlled by the tendency of these particles to coagulate. Thus, the in- affect aggregate strength (Bronick and Lal, 2005; Dontsova and Norton,
crease in aggregate stability with a decrease in pH can be explained by 2002). The wide variation in the WSA fraction (00.95, Fig. 2) for soils
the (i) increase in the thickness of the OM layer on the mineral surfaces with a pH between 7 and 8 might be explained by the strong differences
and by (ii) the increase in the binding strength between the organically in the concentrations and composition of the cations within this pH
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 35

range. Therefore, determining the ionic composition of the soil and soil with quantitative relationships that may contribute to strategies
solution in relation to aggregate stability may further improve predic- towards conserving or improving soil physical properties upon changes
tions of stable aggregate fractions in soils. in soil management or land use.

4.3. Conceptual model 4.4. Conclusions

We propose a conceptual aggregation model (Fig. 6) based on the Analysis of linkages between soil physical and chemical properties
results and discussion presented above. No convincing evidence was for a dataset collected from three diverse European CZO's indicate that:
found for a relation between texture and aggregate formation (Fig. A4).
Instead, our results suggest that the presence of Fe-(hydr)oxides are of
The WSA fraction, which are the small and stable aggregates, and
importance for formation of small aggregates. In this conceptual
the DSA fraction, which are the larger (N1 mm) and less stable ag-
model, we propose that these Fe-(hydr)oxides form a coating on the
gregates, are not correlated and these two aggregate fractions reect a
surfaces of clay, silt and sand-sized particles, and as such provide
different type of soil aggregate.
adsorption sites for OM on the otherwise low to moderately reactive
An increase in the fraction of WSA leads to an increase in micropo-
minerals. The role of Fe-(hydr)oxides is not restricted to a certain size-
rosity (b9 m) and micro + mesoporosity (b 1000 m). There was no
scale of aggregates, instead Fe-(hydr)oxides improve aggregation over
convincing evidence that this relation is stronger for micropores than
the whole range from micro- to mesopores and WSA (Fig. 2, 4, 5). Our
for mesopores (91000 m) meaning that the positive effect of WSA
results suggest that the role of Fe-(hydr)oxides is somewhat more im-
on porosity is not related to a specic pore size.
portant at the smallest size-scale of aggregate formation, reected by
The DSA fraction controls the distribution of soil pores between
the microporosity, than at the size-scale of water-stable aggregates
micropores and mesopores, as was reected by the 30 kPa/0.25 kPa
above 0.25 mm (Fig. A4). Secondly, we propose that OM adsorbs to the
ratio. This suggests that micropores are dominantly located within
Fe-(hydr)oxides in the form of a multilayer composed of humic sub-
dry-sieved aggregates whereas mesopores are dominantly located
stances and larger polysaccharides and proteins. This multilayer of OM
between loose soil particles.
is formed due to hydrophobic interactions between the various organic
Aggregation at the lowest hierarchical level, which we assumed
compounds, and is thick compared to a monolayer of OM and can there-
to be represented by microporosity, is mostly controlled by Fe-
fore act as a bridging agent between multiple particles. The adsorption of
(hydr)oxides with a smaller contribution of SOC. We suggest that
OM on Fe-(hydr)oxides as well the coagulation of organically-coated
Fe-(hydr)oxides are attached to clay and silt particles and provide
particles and clay particles is enhanced towards a lower pH. As such,
adsorption sites for organic substances on these otherwise moder-
the combination of the SOC content, Feox content and pH controls the
ately reactive particles.
fraction of WSA in soils (reference or indication of Fig/table). At the
Aggregation of small and stable aggregates, as measured by wet-
size-scale of the WSA (N0.25 mm), the contribution of SOC seems
sieving, is dominantly controlled by the combination of SOC and
more important than the contribution of Fe-(hydr)oxides and no effect
pH, with a smaller contribution of Fe-(hydr)oxides. We suggest
of texture was observed (Figs. 2, A4). Next, the large and less stable
that the pH effect can be explained by enhanced coagulation and
DSA are built-up from loose particles and WSA. There was no correlation
stronger organic-mineral interactions with decreasing pH. Despite
between the WSA and DSA fraction (Fig. A2), suggesting that the
the fact that the texture of the soils considered widely varied, an ef-
contribution of WSA and loose soil particles to the formation of
fect of texture on the formation of water-stable aggregates was not
DSA can cover a wide range. The DSA fraction was not correlated with
any of the measured soil properties (Fig. 1), suggesting that other aggre-
Aggregation of large aggregates, as measured by dry-sieving, is not
gation factors such as cementing agents, fungal hyphae or roots control
controlled by any of the studied soil properties.
the formation of DSA. Additionally, short-term physical processes like
shrinking and swelling may play a major role for the formation of DSA
(Horn, 1990). Acknowledgement
Our results show that different aggregation factors can be limiting in
the different CZO's. Considering the Marchfeld CZO rst, the near- The EU Commission is greatly acknowledged for nancing the
alkaline pH of the soils reduces the coagulation potential, which SoilTrEC (Soil Transformations in European Catchments) project in its
means that a higher SOC and Feox content are needed to form the 7th Framework Program (grant number 244118).
same amount of WSA as compared to soils with a lower pH. The arable
sites in the Marchfeld CZO were characterized by a low porosity, low Appendix A. Supplementary data
fraction of WSA and low SOC content compared to the forest sites. How-
ever, these differences cannot be solely attributed to differences in land Supplementary data to this article can be found online at http://dx.
use since the Feox content was also somewhat lower in the arable sites
than in the forest sites. Nevertheless, transitions from forestry to arable
farming often coincide with a loss of SOC due to ploughing, lower References
carbon inputs and high phosphate fertilization levels changing
Amzketa, E., 1999. Soil aggregate stability: a review. J. Sustain. Agric. 14 (23), 83151.
chemical and physical properties of soil (Six et al., 2002; Regelink Angers, D.A., 1998. Water-stable aggregation of Qubec silty clay soils: some factors
et al., 2014; Zehetner et al., 2009) and should therefore be regarded as controlling its dynamics. Soil Tillage Res. 47, 9196.
a threat for the physical quality of soils. In the Koiliaris CZO, the Fluvisols Arias, M., Barral, M.T., Diaz-Fierros, F., 1996. Effects of associations between humic acids
and iron or aluminium on the occulation and aggregation of kaolin and quartz.
(sites K1, K2) are strongly weathered and characterized by a low Feox Eur. J. Soil Sci. 47, 335343.
content (amorphous Fe-(hydr)oxides) despite their high Fed content Asano, M., Wagai, R., 2014. Evidence of aggregate hierarchy at micro- to submicron scales
(1025 g/kg, Table A1). As such, these sites are characterized by a rather in an allophanic Andisol. Geoderma 216, 6274.
Baalousha, M., 2009. Aggregation and disaggregation of iron oxide nanoparticles; inu-
low porosity and aggregate stability and can be expected to be suscep-
ence of particle concentration, pH and natural organic matter. Sci. Total Environ.
tible to erosion. In the Slavkov Forest CZO, conditions are favourable for 407 (6), 20932101.
aggregate formation due to the combination of a high SOC content, low Banwart, S., 2011. Save our soils. Nature 474, 151152.
pH and high Feox content. Overall, our results have identied amor- Banwart, S., Menon, M., Bernasconi, S.M., Bloem, J., Blum, W.E., Souza, D.M., Davidsdotir,
B., Duffy, C., Lair, G.J., Kram, P., 2012. Soil processes and functions across an interna-
phous Fe-(hydr)oxides, SOC and pH as the major soil properties that tional network of critical zone observatories: introduction to experimental methods
control aggregate formation and porosity in a wide range of soil types, and initial results. C. R. Geosci. 344, 758772.
36 I.C. Regelink et al. / Geoderma 247248 (2015) 2437

Barberis, E., Ajmore-Marsan, F., Boero, V., Arduino, E., 1991. Aggregation of soil particles Majzik, A., Tombacz, E., 2007. Interaction between humic acid and montmorillonite in the
by iron oxides in various size fractions of soil B horizons. J. Soil Sci. 42, 525542. presence of calcium ions II. Colloidal interactions: charge state, dispersing and/or
Barths, B.G., Kouakoua, E., Larr-Larrouy, M.C., Razambelo, T.M., de Luca, E.F., aggregation of particles in suspension. Org. Geochem. 38 (8), 13301340.
Azontonde, A., Neves, C.S.V.J., de Freitas, P.L., Feller, C.L., 2008. Texture and Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays by a
sesquioxide effects on water-stable aggregates and organic matter in some tropical dithionitecitrate system buffered with sodium bicarbonate. Proc. 7th nat. conf.
soils. Geoderma 143 (12), 1425. clays 5, pp. 317327.
Boix Fayos, C., 1997. The roles of texture and structure in the water retention capacity of Mikutta, R., Kleber, M., Torn, M., Jahn, R., 2006. Stabilization of soil organic matter: asso-
burnt Mediterranean soils with varying rainfall. Catena 31, 219236. ciation with minerals or chemical recalcitrance? Biogeochemistry 77 (1), 2556.
Bronick, C.J., Lal, R., 2005. Soil structure and management: a review. Geoderma 124 (12), Montgomery, D.R., 2007. Soil erosion and agricultural sustainability. Proc. Natl. Acad. Sci.
322. U. S. A. 104, 1326813272.
Brubaker, S.C., Holzhey, C.S., Brasher, B.R., 1993. Estimating the water-dispersible clay Moraetis, M., Paranychianakis, N.V., Nikolaidis, .P., Banwart, S., Rousseva, S., Kercheva,
content of soils. Soil Sci. Soc. Am. J. 56, 12271232. M., Nenov, M., Shishkov, T., de Ruiter, P., Bloem, J., Blum, W.E.H., Lair, G.J., van
Calabi-Floody, M., Bendall, J.S., Jara, A.A., Welland, M.E., Theng, B.K.G., Rumpel, C., de la Luz Gaans, P., Verheul, M., 2014. Sediment provenance, soil development, and carbon
Mora, M., 2011. Nanoclays from an Andisol: extraction, properties and carbon stabi- content in uvial and manmade terraces at Koiliaris River Critical Zone Observatory.
lization. Geoderma 161 (34), 159167. J. Soils Sediments 118.
Chenu, C., Plante, A.F., 2006. Clay-sized organo-mineral complexes in a cultivation Nguetnkam, J.P., Dultz, S., 2011. Soil degradation in Central North Cameroon: water-
chronosequence: revisiting the concept of the primary organo-mineral complex. dispersible clay in relation to surface charge in Oxisol A and B horizons. Soil Tillage
Eur. J. Soil Sci. 57 (4), 596607. Res. 113, 3847.
Currie, J.A., 1966. The volume and porosity of soil crumbs. J. Soil Sci. 17 (1), 24. Nyamangara, J., Gotosa, J., Mpofu, S., 2001. Cattle manure effects on structural stability
Curtin, D., Steppuhn, H., Selles, F., 1994. Effect of magnesium on cation selectivity and and water retention capacity of a granitic sandy soil in Zimbabwe. Soil Tillage Res.
structural stability of sodic soils. Soil Sci. Soc. Am. J. 57 (5), 12771283. 62 (3), 157162.
Dane, J.H., Topp, G.C., 2002. Methods of soil analysis. Part 4 physical methods. SSA Book Oades, J.M., 1993. The role of biology in the formation, stabilization and degradation of
Series. Soil Science Society of America, Madison. soil structure. Geoderma 56, 377400.
Dexter, A.R., 1988. Advances in characterization of soil structure. Soil Tillage Res. 11, Oades, J.M., Waters, A.G., 1991. Aggregate hierarchy in soils. Soil Res. 29, 815828.
199238. Oste, L.A., Temminghoff, E.J.M., Van Riemsdijk, W.H., 2002. Solid-solution partitioning of
Dilkova, R., Kerchev, G., Kercheva, M., 1998. Evaluating and grouping of soils according to organic matter in soils as inuenced by an increase in pH or Ca concentration. Envi-
their susceptibility to anthropogenic degradation. Advances in GeoEcology 31. Catena ron. Sci. Technol. 36, 208214.
Verlag, Reiskirchen, pp. 125131. Pansu, M., Gautheyrou, J., 2006. Handbook of Soil Analysis: Mineralogical, Organic and
Dontsova, K.M., Norton, L.D., 2002. Clay dispersion, inltration, and erosion as inuenced Inorganic Methods. Springer Berlin Heidelberg New York.
by exchangeable Ca and Mg. Soil Sci. 167, 184193. Penn, R.L., Zhu, C., Xu, H., Velben, D.R., 2001. Iron-oxide coatings on sand grains from the
Duiker, S.W., Rhoton, F.E., Torrent, J., Smeck, N.E., Lal, R., 2003. Iron (hydr)oxide crystallin- Atlantic coastal plain: high-resolution transmission electron microscopy characteri-
ity effects on soil aggregation. Soil Sci. Soc. Am. J. 67 (2), 606611. zation. Geology 29 (9), 843846.
Elliott, E.T., Palm, C.A., Reuss, D.E., Monz, C.A., 1991. Organic matter contained in soil ag- Pronk, G.J., Heister, K., Kgel-Knabner, I., 2011. Iron oxides as major available interface
gregates from a tropical chronosequence: correction for sand and light fraction. component in loamy arable topsoils. Soil Sci. Soc. Am. J. 75 (6), 21582168.
Agric. Ecosyst. Environ. 34, 443451. Puget, P., Chenu, C., Balesdent, J., 2000. Dynamics of soil organic matter associated
Eusterhues, K., Rumpel, C., Kogel-Knabner, I., 2005. Organo-mineral associations in sandy with particle-size fractions of water-stable aggregates. Eur. J. Soil Sci. 51 (4),
acid forest soils: importance of specic surface area, iron oxides and micropores. Eur. 595605.
J. Soil Sci. 56, 753763. Pulleman, M., Marinissen, J., 2004. Physical protection of mineralizable C in aggregates
Feng, X., Simpson, A.J., Simpson, M.J., 2005. Chemical and mineralogical controls on humic from long-term pasture and arable soil. Geoderma 120 (3), 273282.
acid sorption to clay mineral surfaces. Org. Geochem. 36, 15531566. Pulleman, M., Jongmans, A., Marinissen, J., Bouma, L., 2003. Effects of organic versus
Fernndez-Ugalde, O., Barr, P., Hubert, F., Virto, I., Girardin, C., Ferrage, E., Caner, L., conventional arable farming on soil structure and organic matter dynamics in a ma-
Chenu, C., 2013. Clay mineralogy differs qualitatively in aggregate-size classes: rine loam in the Netherlands. Soil Use Manag. 19, 147165.
clay-mineral-based evidence for aggregate hierarchy in temperate soils. Eur. J. Soil Regelink, I.C., Weng, L., Koopmans, G.F., Van Riemsdijk, W.H., 2013. Asymmetric ow
Sci. 64 (4), 410422. eld-ow fractionation as a new approach to analyse iron-(hydr) oxide nanoparticles
Haynes, R., Swift, R., 1990. Stability of soil aggregates in relation to organic constituents in soil extracts. Geoderma 202, 134141.
and soil water content. J. Soil Sci. 41 (1), 7383. Regelink, I.C., Voegelin, A., Weng, L., Koopmans, G.F., Comans, R.N.J., 2014. Character-
Hiemstra, T., Antelo, J., Rahnemaie, R., Riemsdijk, W.H.v., 2010. Nanoparticles in natural ization of colloidal Fe from soils using eld-ow fractionation and Fe K-edge X-ray
systems I: the effective reactive surface area of the natural oxide fraction in eld sam- absorption spectroscopy. Environ. Sci. Technol. 48, 43074316.
ples. Geochim. Cosmochim. Acta 74 (1), 4158. Rillig, M.C., 2004. Arbuscular mycorrhizae, glomalin, and soil aggregation. Can. J. Soil Sci.
Holmgren, G.G.S., 1967. A rapid citrateditionite extractable iron procedure. Soil Sci. Soc. 84 (4), 355363.
Am. J. 31 (2), 210211. Schwertmann, U., 1991. Solubility and dissolution of iron oxides. Plant Soil 130, 125.
Horn, R., 1990. Aggregate characterization as compared to soil bulk properties. Soil Tillage Sei, J., Jumas, J.C., Olivier-Fourcade, J., Quiquampoix, H., Staunton, S., 2002. Role of iron
Res. 17, 265289. oxides in the phosphate adsorption properties of kaolinites from the ivory coast.
Igwe, C.A., Zarei, M., Stahr, K., 2009. Colloidal stability in some tropical soils of southeast- Clay Clay Miner. 50 (2), 217222.
ern Nigeria as affected by iron and aluminium oxides. Catena 77 (3), 232237. Six, J., Elliott, E.T., Paustian, K., 2000a. Soil structure and soil organic matter: II. A
Kaiser, K., 2003. Sorption of natural organic matter fractions to goethite (-FeOOH): ef- normalized stability index and the effect of mineralogy. Soil Sci. Soc. Am. J. 64,
fect of chemical composition as revealed by liquid-state 13C NMR and wet-chemical 10421049.
analysis. Org. Geochem. 34, 15691579. Six, J., Paustian, K., Elliott, E.T., Combrink, C., 2000b. Soil structure and organic matter: I.
Kaiser, M., Walter, K., Ellerbrock, R.H., Sommer, M., 2011. Effects of land use and mineral Distribution of aggregate-size classes and aggregate-associated carbon. Soil Sci. Soc.
characteristics on the organic carbon content, and the amount and composition of Am. J. 64, 681689.
Na-pyrophosphate-soluble organic matter, in subsurface soils. Eur. J. Soil Sci. 62, Six, J., Callewaert, P., Lenders, S., De Gryze, S., Morris, S.J., Gregorich, E.G., Paul, E.A.,
226236. Paustian, K., 2002. Measuring and understanding carbon storage in afforested soils
Kaiser, M., Ellerbrock, R.H., Wulf, M., Dultz, S., Hierath, C., Sommer, M., 2012. The inuence by physical fractionation. Soil Sci. Soc. Am. J. 66 (6), 19811987.
of mineral characteristics on organic matter content, composition, and stability of Six, J., Bossuyt, H., Degryze, S., Denef, K., 2004. A history of research on the link be-
topsoils under long-term arable and forest land use. J. Geophys. Res. 117, 116. tween (micro)aggregates, soil biota, and soil organic matter dynamics. Soil Tillage
Kercheva, M., Rousseva, S., Dimitrov, E., Nenov, M., Shishkov, T., 2011. Soil aggregation Res. 79, 731.
estimates in CZO-Fuchsenbigl. Appl. Geochem. 26, S57S59 (Supplement). Spohn, M., Giani, L., 2010. Water-stable aggregates, glomalin-related soil protein, and car-
Kleber, M., Sollins, P., Sutton, R., 2007. A conceptual model of organo-mineral interactions bohydrates in a chronosequence of sandy hydromorphic soils. Soil Biol. Biochem. 42
in soils: self-assembly of organic molecular fragments into zonal structures on miner- (9), 15051511.
al surfaces. Biogeochemistry 85, 924. Stamati, F.E., Nikolaidis, N.P., Schnoor, J.L., 2013. Modeling topsoil carbon sequestration in
Kgel-Knabner, I., Guggenberger, G., Kleber, M., Kandeler, E., Kalbitz, K., Scheu, S., Eusterhues, two contrasting crop production to set-aside conversions with RothC calibration
K., Leinweber, P., 2008. Organo-mineral associations in temperate soils: integrating issues and uncertainty analysis. Agric. Ecosyst. Environ. 165, 190200.
biology, mineralogy, and organic matter chemistry. J. Plant Nutr. Soil Sci. 171 (1), 6182. Tisdall, J.M., Oades, J.M., 1982. Organic matter and water-stable aggregates in soils. J. Soil
Kram, P., Hruska, J., Shanley, J.B., 2012. Streamwater chemistry in three contrasting Sci. 33 (2), 141163.
monolithologic Czech catchments. Appl. Geochem. 27, 18541863. Tombacz, E., Szekeres, M., 2004. Colloidal behavior of aqueous montmorillonite suspen-
Lair, G.J., Zehetner, F., Hrachowitz, M., Franz, N., Maringer, F.J., Gerzabek, M.H., 2009a. Dating sions: the specic role of pH in the presence of indifferent electrolytes. Appl. Clay
of soil layers in a young oodplain using iron oxide crystallinity. Quat. Geochronol. 4 (3), Sci. 27, 7594.
260266. Vogel, C., Mueller, C.W., Hschen, C., Buegger, F., Heister, K., Schulz, S., Schloter, M., Kgel-
Lair, G.J., Zehetner, F., Khan, Z.H., Gerzabek, M.H., 2009b. Phosphorus sorptiondesorption in Knaber, I., 2014. Submicron structures provide preferential spots for carbon and
alluvial soils of a young weathering sequence at the Danube River. Geoderma 149 (2), nitrogen sequestration in soils. Nat. Commun. 5.
3944. Wattel-Koekkoek, E.J.W., Van Genuchten, P.P.L., Buurman, P., Van Lagen, B., 2001. Amount
Lehtinen, T., Lair, G.J., Mentler, A., Gisladttir, G., Ragnarsdttir, K.V., Blum, W.E.H., 2014. and composition of clay-associated soil organic matter in a range of kaolinitic and
Soil aggregate stability in different soil orders quantied by low dispersive ultrasonic smectitic soils. Geoderma 99, 2749.
energy levels. Soil Sci. Soc. Am. J. 78, 713723. Weng, L., Van Riemsdijk, W.H., Hiemstra, T., 2008. Humic nanoparticles at the oxide
Luxmoore, R.J., 1980. Micro-, meso-, and macroporosity of the soil. Soil Sci. Soc. Am. J. 45, water interface: interactions with phosphate ion adsorption. Environ. Sci. Technol.
671672. 42, 87478752.
I.C. Regelink et al. / Geoderma 247248 (2015) 2437 37

Weng, L., Vega, F.A., Van Riemsdijk, W.H., 2011. Competitive and synergistic effects in pH Zehetner, F., Lair, G.J., Maringer, F.J., Gerzabek, M.H., Hein, T., 2008. From sediment to soil:
dependent phosphate adsorption in soils: LCD modeling. Environ. Sci. Technol. 45 oodplain phosphorus transformations at the Danube River. Biogeochemistry 88 (2),
(19), 84208428. 117126.
Wsten, J., Pachepsky, Y.A., Rawls, W., 2001. Pedotransfer functions: bridging the gap Zehetner, F., Lair, G.J., Gerzabek, M.H., 2009. Rapid carbon accretion and organic matter
between available basic soil data and missing soil hydraulic characteristics. pool stabilization in riverine oodplain soils. Glob. Biogeochem. Cycles 23 (4).
J. Hydrol. 251 (3), 123150.