Soil & Tillage Research 163 (2016) 130–140

Two decades of no-till in the Oberacker long-term field experiment:Part II. Soil porosity and gas transport parameters

Ingrid Martíneza, Andreas Chervetb, Peter Weisskopfa, Wolfgang G. Sturnyb, Jan Reka,Thomas Kellera,c,*aAgroscope, Department of Natural Resources & Agriculture, Reckenholzstrasse 191, CH-8046 Zürich, SwitzerlandbBern Office of Agriculture & Nature, Soil Conservation Service, Ruetti, CH-3052 Zollikofen, Switzerlandc Swedish University of Agricultural Sciences, Department of Soil & Environment, Box 7014, SE-75007 Uppsala, Sweden

A R T I C L E I N F O

Article history:Received 19 October 2015Received in revised form 23 May 2016Accepted 23 May 2016Available online xxx

Keywords:Air-filled porosityAir permeabilityGas diffusivityDirect drillingMouldboard ploughing

A B S T R A C T

This is the second in a series of papers describing the impact of two decades of no-till in the Oberackerlong-term field experiment in Switzerland. The first focused on crop yields, soil organic carbon andnutrient distributions in the soil, while this study investigated the impact on soil gas transport properties.The Oberacker long-term field experiment was established in 1994 on a sandy loam and compares twotillage systems, mouldboard ploughing (MP) and no-till (NT). Undisturbed soil cores were collected at 0.1(topsoil) and 0.4 m depth (subsoil) from both tillage treatments. Permanent grass (PG) strips locatedbetween the experimental plots were also sampled, as references. The soil cores were used formeasurements of air-filled porosity (ea), air permeability (ka) and relative gas diffusion coefficient (Dp/D0) at five matric potentials ranging from �30 to �500 hPa. The results show that the soil pore systemand gas transport properties of the NT soil were similar to those under PG for both topsoil and subsoil. Incontrast, the soil under MP showed a clear stratification: ea, ka and Dp/D0were higher than in NT and PG inthe topsoil, but lower in the subsoil. The Dp/D0 vs. ea relationship was accurately described by a modelderived from percolation theory. A linear relationship between log ka and log ea was found, but the slopewas smaller than could be expected from percolation theory, in particular for subsoil, possibly because ofthe anisotropy of macropores. The pore system showed slightly higher specific diffusivity and muchhigher specific air permeability in the topsoil of MP than in NT or PG, but the relations were reversed inthe subsoil. Thus it is highly important to consider both the topsoil and subsoil when tillage systems areevaluated.

ã 2016 Elsevier B.V. All rights reserved.

Contents lists available at ScienceDirect

Soil & Tillage Research

journal homepage: www.else vie r .com/locate /s t i l l

1. Introduction

No-till or direct drilling means that crops are sown without anyprior loosening of the soil, where the only mechanical soildisturbance arises from the disc openers of the drill (Soaneet al., 2012). No-till is practised for various reasons, includingerosion protection, lower fuel consumption or higher work rates(Soane et al., 2012). Omitting soil loosening affects soil structureand various soil properties and processes and therefore hassignificant implications for soil functions, including crop growth.

* Corresponding author at: Agroscope, Department of Natural Resources &Agriculture, Reckenholzstrasse 191, CH-8046 Zürich, Switzerland.

E-mail addresses: thomas.keller@agroscope.admin.ch, thomas.keller@slu.se(T. Keller).

http://dx.doi.org/10.1016/j.still.2016.05.0200167-1987/ã 2016 Elsevier B.V. All rights reserved.

Soil structure, i.e. the spatial arrangement of voids and soilconstituents, controls important soil functions and key processes,such as water and gas transport and storage, chemical reactions,biological activity and mechanical strength, which in turn affectfluxes to roots, plant growth and crop yields (Berisso et al., 2013;Chen et al., 2014). The present study focused on soil gas transportparameters (air permeability, ka, and relative gas diffusioncoefficient, Dp/D0), because (i) they control soil aeration and arethus a key factor for microbial activity, root growth and cropproductivity (Deubel et al., 2011; Moldrup et al., 2013) and (ii) theycan be used to derive information on the characteristics of the soilpore system (Gradwell, 1961; Ball, 1981; Groenevelt et al., 1984;Schjønning et al., 2002). Subsurface transport of gases through air-filled pore space predominantly occurs by diffusion (controlled byDp/D0), but near-surface pressure fluctuations induce gas transportby advection (controlled by ka) (e.g. Deepagoda et al., 2011).Diffusion and advection are affected differently by pore size.

I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140 131

Diffusion is independent of pore size unless pores are very small, inwhich case gas molecules collide with the pore walls morefrequently than with each other (Knudsen diffusion), while thevolumetric flow rate due to transport by advection is proportionalto the fourth power of the pore radius as described by the Hagen-Poiseuille law (Alaoui at al., 2011). Consequently, Dp/D0 is afunction of air-filled porosity, while ka is controlled by the largestsoil pores (Ball et al., 1988). Hence, soil pore characteristics can beobtained by combining Dp/D0, ka and their relationships with ea,revealing aspects of pore structure such as continuity andtortuosity (Groenevelt et al., 1984; Ball et al., 1988).

Omitting soil tillage in no-till may promote the creation ofcontinuous pores, in particular biopores, with positive effects onsoil transport functions (Hartmann et al., 2012). However, this mayenhance preferential flow and leaching of agrochemicals (Jarvis,2007). Moreover, no-till may create dense topsoils, with negativeimpacts on soil aeration and root penetration resistance (Kay andVandenBygaart, 2002; Schjønning and Thomsen, 2013; Nuneset al., 2015). Conventional tillage using mouldboard ploughing mayresult in reduced earthworm populations (Díaz-Zorita and Grove,2002; Kay and VandenBygaart, 2002), interruption of pores and ahorizontally orientated pore system at the plough pan (Dörner andHorn, 2009; Alaoui et al., 2011; Berisso et al., 2013). It may alsocause subsoil compaction by in-furrow ploughing (Lipiec andHatano, 2003). The depth of tillage also affects the distribution ofsoil organic carbon in the soil profile (Blanco-Canqui and Lal,2007), which in turn affects soil structure (Kautz et al., 2013).

The objective of this study was to investigate the long-termimpact (two decades) of soil tillage (no-till versus conventionaltillage including mouldboard ploughing) on the soil pore systemand gas transport properties in the topsoil and subsoil in a long-term field experiment in Switzerland.

2. Materials and methods

2.1. Site description and soil sampling

The Oberacker long-term field experiment at the INFORAMARuetti in Zollikofen near Berne, Switzerland (47.0�N, 7.5�E; 557 m a.s.l.) was established in 1994 and compares two tillage systems:mouldboard ploughing (MP; in-furrow ploughing to about 0.25 mdepth before 2002 and on-land ploughing to approximately 0.15 mdepth since 2003) and no-till (NT). The experiment has a split-plotdesign, with six experimental plots (9 m � 80 m) per treatment.Between the experimental plots there are permanent grass (PG)strips of 3 m width. Mean annual temperature at the site is 9.3 �Cand mean annual precipitation is 1109 mm (Martínez et al., 2016).The soil is classified as a Eutric Cambisol with a sandy loam texture.Further details on the Oberacker long-term field experiment can befound in Martínez et al. (2016). The experiment is run with a six-year crop rotation, and the current crop rotation is: peas (Pisumsativum L.)—winter wheat (Triticum aestivum L.)—field beans(Phaseolus vulgaris L.)—winter barley (Hordeum vulgare L.)—sugarbeet (Beta vulgaris L.)—silage maize (Zea mays L.), representing atypical Swiss arable crop rotation (Prasuhn, 2012). The croprotation has received some modifications throughout the years (fordetails, see Martínez et al., 2016). For example, potatoes (Solanumtuberosum L.) were included until 1999 but discarded due tounsatisfactory planting technique (using direct mulch planting)and poor tuber quality in NT.

Three NT plots, three MP plots and two PG strips were sampledin spring 2013 (more than half a year after the last tillage/seedingoperation in MP), i.e. 19 years after the start of the experiment.Penetrometer measurements performed prior to sampling (Mar-tínez et al., 2016) identified two critical soil layers with substantialdifferences between MP and NT, namely in the topsoil at 0.05–

0.2 m depth (higher resistance in NT than in MP) and in the subsoilat 0.3–0.4 m depth (lower resistance in NT than in MP) (see Fig. 4bin Martínez et al., 2016). Therefore these two layers were selectedfor soil sampling in the present study. Soil cores (height: 0.06 m;diameter: 0.1 m) were sampled at 0.08–0.14 m (topsoil) and 0.35–0.41 m depth (subsoil); for simplicity, theses depths are referred toas 0.1 (topsoil) and 0.4 m (subsoil) in the remainder of this paper. Ateach sampling location and depth, five undisturbed soil cores werecollected.

2.2. Measurement procedure for gas transport parameters

In the laboratory, the undisturbed soil core samples were slowlysaturated from below and then drained stepwise to five differentmatric potentials, h: �30 (corresponding to pF 1.5; where pF = log(�h) and h is in hPa), �60 (pF 1.8), �100 (pF 2.0), �200 (pF 2.3) and�500 hPa (pF 2.7). All measurements described below wereperformed at these five matric potentials on each sample. Fromthe water retention measurements, the fraction of soil pores wasderived and the pore diameter, d (mm), was estimated as d = 3000/(�h) (e.g. Schjønning et al., 2002).

The soil cores were weighed at each matric potential prior to themeasurements of air permeability and gas diffusivity and afteroven-drying (at least 24 h at 105 �C) to determine the soil watercontent at the different matric potentials. Air-filled porosity (ea)was calculated from the volumetric soil water content and totalporosity, which was derived from soil bulk density and particledensity.

2.2.1. Air permeabilityAir permeability was measured using a steady-state method

similar to that described by Iversen et al. (2001). The soil at the veryedge of the core was carefully pressed to the cylinder walls in orderto minimise leakage of air between the soil and the cylinder wall(Ball and Schjønning, 2002). The airflow was recorded when thevolumetric flow rate was stabilised at a pressure of 2 hPa. Airpermeability, ka, was then calculated from the volumetric flow rateand the applied pressure head using Darcy’s equation:

ka ¼ �QlshDpAs

ð1Þ

where Q is the volumetric flow rate (m3 s�1), ls is the height of thesoil sample (m), h is the dynamic air viscosity (Pa s), Dp is thedifference in air pressure (Pa) and As is the cross-sectional area ofthe soil sample (m2).

2.2.2. Gas diffusivityGas diffusivity was measured in a one-chamber apparatus

similar to that described by Schjønning et al. (2013), using O2 as thediffusing gas. The apparatus is described in detail in theSupplementary methods. The O2 diffusion constant, Dp

*, wasderived from the relationship between time and the logarithm ofthe O2 concentration difference between chamber and ambient airby assuming steady state diffusion as described in Schjønning et al.(2013):

D�p ¼ � hS � hC � AC

AS

� � ln DCtDC0

� �t

24

35 ð2Þ

where DCt is the difference in O2 concentration between the twoends of the sample (i.e. between the O2 concentration in thechamber and ambient O2 concentration), DC0 is the initialdifference in O2 concentration between the two ends of thesample, hs is the height of the cylindrical sample, hc is the height ofthe cylindrical chamber, As is the cross-sectional area of the

132 I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140

sample, Ac is the cross-sectional area of the diffusion chamber and tis time. This solution does not take into account gas storage in thesample. However, with the dimensions of the diffusion apparatusand considering that our interest was primarily in diffusion undermoist soil conditions, the error should be small (Schjønning et al.,2013). The measurements were corrected to a standard tempera-ture, T0, of 293 K (20 �C) and a standard air pressure, p0, of1000 mbar using the equation (Bakker and Hidding, 1970):

Dp ¼ D�p

T0

Ta

� �1:75 p0pa

� �ð3Þ

where Dp* is the actual measured Dp (obtained from Eq. (2)) at

actual measured temperature (Ta) and air pressure (pa). Therelative gas diffusion coefficient, Dp/D0, was then obtained bydividing Dp by the gas diffusion coefficient of O2 in free air, D0 (e.g.Schjønning et al., 2011).

2.3. Calculations of soil pore characteristics

From the measurements of Dp/D0 and ea, the ‘specific diffusivity’CD also referred to as ‘diffusion efficiency’ (Kühne et al., 2012) wascalculated (Gradwell, 1961) as:

CD ¼ Dp=D0

eað4Þ

The specific diffusivity of a pore system consisting of parallelstraight tubes (capillary tubes model) is CD = (Dp/D0)/ea = 1, while0 < CD<1 for real soils. From the measurements of ka and ea, the‘specific air permeability’ CA (Groenevelt et al., 1984), (also referredto as ‘pore organisation’ (Blackwell et al., 1990) or ‘C2’ (Ball et al.,1988)) was obtained as:

CA ¼ kaea

ð5Þ

The terms ‘specific diffusivity’ and ‘specific air permeability’ areused here, because the terms ‘efficiency’ or ‘pore organisation’could be misleading � the highest possible efficiency or poreorganisation implies a ‘pipe-like’ pore system, which may not bedesirable from a soil structure and soil quality point of viewbecause large parts of the soil matrix could be bypassed(preferential flow). On the other hand, pipe-like pores ensureadequate aeration of the subsoil. Consequently, optimum con-ditions may be obtained at an intermediate specific diffusivity andspecific air permeability. Finally, we calculated C3 (Groeneveltet al., 1984; Ball et al., 1988):

C3 ¼ kae2a

ð6Þ

and P (Kawamoto et al., 2006a; Eden et al., 2011):

P ¼ kaDp=D0

ð7Þ

which have been suggested as soil structure indices. C3 and Preveal similar information and, in the case of Dp/D0 = ea2

(corresponding to the Buckingham (1904) model), C3 = P.

2.4. Modelling gas transport properties as a function of air-filledporosity

In this study we investigated the applicability of equationsderived from percolation theory that describe how the relative gasdiffusivity coefficient and air permeability evolve as a function ofair-filled porosity, by fitting these equations to our experimentaldata. Percolation theory can be used to describe properties atmacroscopic scale (e.g. permeability) that are determined by the

connectivity of the elements (e.g. bonds, which could representpores) of a system (e.g. porous media, such as soil). A centralfeature of percolation theory is that the volumetric fluid flow, Q, isdetermined as (Berkowitz and Ewing, 1998):

Q / N � Ncð Þk ð8Þwhere N is the number of bonds, Nc is the critical number of bondsrequired for fluid flow to occur and k is an exponent. Thus thenetwork (e.g. a soil pore network) can only conduct a fluid if N > Nc,while there is no flow through the system (e.g. from the bottom tothe top of a soil sample) at N < Nc. This property is known as thepercolation threshold. See Berkowitz and Ewing (1998) for furtherreading on percolation theory and its application in soil physics.

Gas diffusivity in porous media can be modelled as (Ghanbarianand Hunt, 2014):

Dp=D0 ¼ ea � et1 � et

� �mð9Þ

where et is the percolation threshold and m is the scaling exponent.Eq. (9) is identical to the empirical equation proposed by Troehet al. (1982). The percolation threshold can be interpreted as theminimum air-filled porosity required to allow for gas transport bydiffusion. The percolation threshold depends on the pore structure,pore size distribution and pore continuity (Ghanbarian and Hunt,2014). We fitted Eq. (9) to our experimental data by non-linearleast squares fitting.

The change in air permeability, ka, with air-filled porosity, ea,can be described based on percolation theory as (Ghanbarian-Alavijeh and Hunt, 2012):

ka eað Þ ¼ ka ea ¼ fð Þ ea � ecf � ec

� �t; ð10Þ

where ec is the percolation threshold, t is the scaling exponent andf is the porosity. Note that ec may differ from et in Eq. (9).According to Ghanbarian-Alavijeh and Hunt (2012), any measuredka can be used as a reference point in Eq. (10) if ka(F) is notavailable. Since ka(F) was not measured for our samples, we fitted(Ghanbarian-Alavijeh and Hunt, 2012):

ka eað Þ ¼ ka ea ¼ e500ð Þ ea � ece500 � ec

� �tð11Þ

to our data, where e500 is the air-filled porosity at �500 hPa (i.e.the driest condition used in this work) and ka(e500) is our referencepoint.

An interesting aspect of percolation theory is that the exponent(k in Eq. (8), m in Eq. (9), t in Eqs. (10)–(11)) has the same value formany systems, which is known as universality (Berkowitz andEwing, 1998). In theory, the exponent takes values of 1.3 and 2 intwo and three dimensions (two and three-dimensional flow),respectively (Ghanbarian-Alavijeh and Hunt, 2012; Ghanbarianand Hunt, 2014). Different values for the exponent characterisenon-universal behaviour (Berkowitz and Ewing, 1998). One of thebasic assumptions in percolation theory is that the medium isinfinite, stationary and random (Berkowitz and Ewing, 1998).

2.5. Statistical analysis

Soil porosity and gas transport parameters were analysed usingthe InfoStat statistical analysis software (Di Rienzo et al., 2009). Airpermeability data were log-transformed before analysis. A split-plot linear mixed model was used, with fixed effects of tillagetreatment, depth, pF (=log (�h), where h is the matric potential)and tillage treatment-by-depth interaction, and random effects ofplots. Means were tested using the Tukey method at significancelevel p < 0.05.

Table 1Characteristics of the topsoil (0.1 m) and subsoil (0.4 m) at the Oberacker site in plots with mouldboard ploughing (MP), no-till (NT) and permanent grass (PG).Clay <0.002 mm; silt 0.002–0.05 mm; sand 0.05–0.2 mm; OM: organic matter content. Numbers in brackets indicate standard error .

Soil characteristics Topsoil, 0.1 m Subsoil, 0.4 m

MP NT PG MP NT PG

Clay (% by weight) 18.2 (1.2) 19.0 (1.7) 17.4 (1.1) 16.1 (3.5) 15.4 (4.3) 15.4 (6.0)Silt (% by weight) 22.7 (0.5) 22.6 (1.8) 23.4 (1.5) 22.0 (1.1) 19.6 (4.7) 20.5 (0.0)Sand (% by weight) 59.1 (0.5) 58.4 (3.3) 59.2 (0.9) 61.9 (2.9) 65.0 (8.5) 64.1 (6.0)OM (% by weight) 2.7 (0.2) 2.3 (0.3) 2.6 (0.6) 1.0 (0.2) 1.0 (0.3) 1.0 (0.0)Dry bulk density (Mg m�3) 1.35 (0.02) 1.47 (0.02) 1.46 (0.02) 1.54 (0.02) 1.49 (0.02) 1.48 (0.03)Particle density (Mg m�3) 2.59 (0.03) 2.60 (0.03) 2.59 (0.02) 2.64 (0.02) 2.63 (0.01) 2.64 (0.03)

I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140 133

3. Results

3.1. Bulk density, porosity and pore size distribution

In the topsoil, bulk density (BD) was significantly lower(p < 0.05) in MP (1.35 Mg m�3) than in NT (1.47 Mg m�3) andPG (1.46 Mg m�3). In the subsoil, BD was higher in MP (1.54 Mgm�3) than in NT and PG (1.49 and 1.48 Mg m�3, respectively),although the differences were not significant (Table 1). Twoaspects of this are interesting; first, NT and PG showed similar BDand second, there was no significant difference in BD betweentopsoil and subsoil in PG and NT, while there was a clearstratification in MP with lower BD in the topsoil.

The fractions of pores derived from water retention measure-ments are presented in Fig. 1. Interestingly, the volume ofpores >100 mm was larger in the subsoil than in the topsoil,especially for NT and PG (Fig. 1). Pores >100 mm could be biopores(root and earthworm channels) or inter-aggregate pores such asvoids between fragments formed during tillage. In the topsoil, thelower BD in MP compared with NT and PG was mainly due to thelarger volume of pores >100 mm, as there was little differencebetween treatments for the other pore size classes (Fig. 1). Hence,plant-available water (i.e. water stored in pores with equivalentdiameter between 0.2 and 30 mm, corresponding to a pF range of2–4.2; e.g. Reynolds and Clarke Topp, 2008) was similar in alltreatments. The volume of the smallest pore fraction measured(d <6 mm) was much smaller in the subsoil compared with thetopsoil, which is associated with the lower soil organic carboncontent of the subsoil (Dexter et al., 2008). Pores smaller than 3 mmstore water and constitute protective pores where smallerorganisms (bacteria, protozoa) take refuge from predation (e.g.Brussaard and van Faassen, 1994; Schjønning et al., 2002).

Fig. 1. Fraction of pores derived from water retention measurements influenced byno-till (NT), mouldboard ploughing (MP) and permanent grassland (PG) at 0.1 and0.4 m depth.

In agreement with Fig. 1, MP showed significantly higher air-filled porosity, ea, than PG and NT in the topsoil at any given matricpotential, but in the subsoil PG and NT showed the highest values(Fig. 2a–b). The line in Fig. 2a–b is ea = 0.1 m3m�3 which is a criticallower limit for satisfactory root growth (Grable and Siemer, 1968),as discussed in Section 4.3.

3.2. Gas transport properties: relative gas diffusion coefficient and airpermeability

The relative gas diffusion coefficient (Dp/D0) for the five matricpotentials in the topsoil and subsoil was similar for PG and NT(Fig. 2c–d). In the topsoil, Dp/D0was significantly higher in MP thanin PG and NT, but the relationship was reversed (i.e. lower Dp/D0 inMP) in the subsoil (Fig. 2c–d). The horizontal lines in Fig. 2c–dindicate critical thresholds for soil aeration according toStepniewski (1980, 1981) (Dp/D0 = 0.005) and (Schjønning et al.,2003) (Dp/D0 = 0.02), as discussed in Section 4.3.

The data in Fig. 2e–f show log ka as a function of pF. According toFish and Koppi (1994), soils that are below log ka<1.3 (corre-sponding to ka = 20 mm2), indicated by horizontal lines in Fig. 2e–f,are considered poorly aerated. Air permeability in the topsoil washigher in MP than in NT and PG at all matric potentials. The lowerair permeability for PG and NT than MP in the topsoil is consistentwith the smaller proportion of large pores (Fig. 1). As for Dp/D0, therelationship was also reversed for ka in the subsoil, i.e. higher ka inNT and PG than in MP (Fig. 2e–f). It is interesting that in NT and PG,Dp/D0 and ka were higher in the subsoil than in the topsoil. Incontrast, in MP Dp/D0 and ka were higher in the topsoil than in thesubsoil.

3.3. Soil structure indices derived from gas transport properties

The different soil pore characteristics/soil structure indices(Eqs. (4)–(7)) obtained from our data, as average values for all fivematric potentials, are presented in Fig. 3. The indices did not varygreatly with matric potential and the relationships between thetreatments did not change with matric potential (not shown). InMP, both the specific air permeability (Fig. 3a) and the specificdiffusivity (Fig. 3b) were significantly larger in the topsoil, butsignificantly smaller in the subsoil, compared with NT and PG. Thedifference between MP and NT or PG was larger for CA (Fig. 3a) thanfor CD (Fig. 3b), especially in the topsoil.

As Fig. 3c and d show, there were differences betweentreatments for the topsoil (higher C3 and P in MP), but not forthe subsoil (but values for MP were lowest). The values for C3 and Pwere similar in the subsoil, while the values for C3 in the topsoilwere higher than those for P by a factor of 1.4-2.3. As mentionedelsewhere, C3 = P for Dp/D0 = ea2. The higher values for C3 than P areexplained by disproportionately high values of ka in relation to ea,caused by single structural pores (inter-aggregate pores, biopores)in the topsoil.

Fig. 2. (a,b) Air-filled porosity, ea, (c,d) relative gas diffusion coefficient, Dp/D0, and (e,f) air permeability, ka, for the three soil management systems (NT: no-till; MP:mouldboard ploughing; PG: permanent grassland) at 0.1 m (a, c, e) and 0.4 m depth (b, d, f) as a function of pF (=log (�h), where h is the matric potential)). Note that NT and PGare slightly displaced for better readability, and that the legends of the top panels apply to all data shown. The dashed lines represent critical limits for ea (Grable and Siemer,1968), Dp/D0 (Stepniewski, 1980, 1981; Schjønning et al., 2003), and ka (Fish and Koppi, 1994). Error bars indicate �1 standard error.

134 I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140

Fig. 3. (a) Specific air permeability, CA, (i.e. air permeability, ka, divided by air-filled porosity, ea), (b) specific diffusivity, CD, (i.e. relative gas diffusivity, Dp/D0, divided by ea), (c)C3 = ka ea�2 and (d) P = ka (Dp/D0)�1 for no-till (NT; open symbols), mouldboard ploughing (MP; dark grey symbols) and permanent grassland (PG; light grey symbols) at 0.1 and0.4 m depth. Note that MP and PG are slightly displaced for better readability, and that the legend of the top right panel applies to all data shown. Error bars indicate �1standard error.

I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140 135

3.4. Modelling gas transport properties as a function of air-filledporosity: percolation threshold and scaling exponent

The relative gas diffusion coefficient as a function of air-filledporosity is presented in Fig. 4. The line in Fig. 4 shows the fit ofEq. (9) to the data from all treatments and depths. The averagevalues for et and m (Eq. (9)) per soil management system and soillayer are given in Table 2. There were no significant differences in mbetween the different soil management systems, but a trend forhigher (closer to 2) values in the subsoil. The Dp/D0-associatedpercolation threshold was found to be smallest for PG and largestfor MP, and et was generally higher in the subsoil than in thetopsoil. On average for all treatments and layers, we found a valuefor et of 0.006, i.e. the average et was close to zero and the average mclose to 2. For et = 0 and m = 2, Eq. (9) equals the Buckingham (1904)

model (i.e. Dp/D0 = ea2). Consequently, the Buckingham (1904)relation fitted our data almost equally well (not shown). It isinteresting that the rather simple Eq. (9) fitted our data well,irrespective of soil management or soil layer (Fig. 4; Table 2).

Air permeability as a function of air-filled porosity is shown inFig. 5 in log–log scale. The graph reveals similar slopes of the log kavs log ea relationship for all subsoils, while the slopes for thetopsoils vary (larger slope for MP and PG, smaller slope for NT). Theka-associated percolation thresholds and scaling exponents for thedifferent soil managements and soil layers are summarised inTable 2. The percolation threshold was smaller in the topsoil thanin the subsoil. In the topsoil, ec was similar in MP and NT. In thesubsoil, the lowest ec was found for MP and the highest ec for NT.For all treatments and layers, we obtained an average value for ec of0.031. The ka-associated scaling exponents were close to 2 for the

Fig. 4. Relative gas diffusivity, Dp/D0, as a function of air-filled porosity, ea, for thetopsoil (circles) and the subsoil (triangles) of the no-till (NT, open symbols),mouldboard ploughing (MP, dark grey symbols) and permanent grass (PG, light greysymbols) treatments. Each data point is the treatment average at a certain matricpotential. The curve represents the universal scaling law for gas diffusion in porousmedia (Eq. (9)) with a value for the percolation threshold, et, of 0.006 and a scalingexponent, m, of 1.92, corresponding to the average values for et and m found for ourdata (Table 2).

136 I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140

topsoil of MP and PG, while they were <1 for the topsoil of NT andfor all subsoils. On average (all treatments, both depths), weobtained a value for t of 1.1 (Table 2).

4. Discussion

4.1. Air-filled pore space, soil gas transport capability and soil structureindices

The measurements of ea, ka, and Dp/D0 revealed consistentlyhigher values for MP than for NT or PG in the topsoil, butconsistently lower values for MP than for NT or PG in the subsoil(Fig. 2). Similar values of ea, ka and Dp/D0 were observed for NT andPG (Fig. 2). A distinct difference in soil pore properties betweentopsoil and subsoil was found for MP (higher values in the topsoilthan in the subsoil), while there was a smaller difference betweentopsoil and subsoil for NT and PG (similar values or higher values inthe subsoil than in the topsoil) (Figs. 2 and 3).

Differences in CA and CD (Fig. 3a–b) indicated (i) higherconnectivity, higher continuity and lower tortuosity (or anycombination), especially for macropores because the differences

Table 2Percolation threshold, et, and scaling exponent, m, for gas diffusion (Eq. (9)) and percolatiomanagement systems investigated. Mean values, with standard deviation in brackets.

Soil management system and depth et

TopsoilMouldboard ploughing 0.009 (0.013)

No-till 0.003 (0.005)

Permanent grass 0.000 (0.000)

Mean Topsoil 0.004

SubsoilMouldboard ploughing 0.014 (0.024)

No-till 0.008 (0.013)

Permanent grass 0.000 (0.000)

Mean Subsoil 0.007

Mean Mouldboard ploughing 0.012

Mean No-till 0.005

Mean Permanent grass 0.000

Mean 0.006

were larger for CA in the topsoil of MP compared with the topsoil ofeither NT or PG, and (ii) lower connectivity, lower continuity andhigher tortuosity (or any combination) in the subsoil of MPcompared with the subsoil of NT and PG. The pore structures of NTand PG were similar for both depths. In MP, the volume ofmacropores was not lower in the subsoil than in the topsoil (Fig. 1),despite the lower ka in the subsoil. Hence, the MP macroporesystem had higher connectivity, higher continuity or lowertortuosity in the topsoil compared with the subsoil. NT and PGhad lower macroporosity in the topsoil than in the subsoil (Fig. 1),which correlates with the lower ka in the topsoil (Fig. 2e–f).

The values of P recorded here (Fig. 3d) were 3- to 5-fold higherthan the P = 700 reported by Kawamoto et al. (2006a) for Danishloamy soils. The soils used by Kawamoto et al. (2006a), which aredescribed in Kawamoto et al. (2006b), had on average a slightlylower content of silt and a correspondingly higher content of sand(59–96% sand compared with 57–65% in our study), while the claycontent was similar. Their soil also had a lower soil organic mattercontent. The higher values of P found in the Oberacker soil indicatea higher proportion of continuous macropores, e.g. biopores, thanin the soils used by Kawamoto et al. (2006a). Thirteen out of the 22soils in their dataset had a very high sand content (70–96%), andmacropores may be poorly connected in sandy soils (Håkansson,2005).

4.2. Percolation threshold and scaling exponent

The percolation threshold for advective air flow (ec = 0.031) inthe present study was almost one order of magnitude larger thanthe percolation threshold for gas diffusion (et = 0.006) (Table 2).The average et found here (et = 0.006) was about 10-fold smallerthan the average value found by Ghanbarian and Hunt (2014), whoanalysed a large dataset including intact and remoulded samples.Our samples had a smaller total porosity, F, as well as smaller emin

and emax (minimum and maximum ea, respectively) than most soilsincluded in the study by Ghanbarian and Hunt (2014). Further-more, et in our study was similar to values obtained by Ghanbarianand Hunt (2014) for certain soils. The average ec value of 0.031found for our data was similar to the values presented inGhanbarian-Alavijeh and Hunt (2012).

The Dp/D0-associated scaling factor, m (Eq. (9)), was close to thetheoretical value of m = 2 for three-dimensional flow (Ghanbarianand Hunt, 2014), with little variation among treatments and soildepths (Table 2). In contrast, the ka-associated scaling factor, t (Eqs.(10)–(11)), varied between 0.35 and 2.08, with an overall average of

n threshold, ec, and scaling exponent, t, for air permeability (Eq. (11)) of the three soil

m ec t

1.80 (0.00) 0.031 (0.053) 2.08 (0.96)1.87 (0.08) 0.032 (0.031) 0.64 (0.47)1.90 (0.14) 0.000 (0.000) 2.05 (1.40)1.86 0.021 1.59

1.98 (0.16) 0.000 (0.000) 0.87 (0.26)1.90 (0.18) 0.088 (0.020) 0.35 (0.30)2.06 (0.08) 0.037 (0.046) 0.55 (0.57)1.98 0.042 0.59

1.89 0.015 1.481.89 0.060 0.501.98 0.019 1.30

1.92 0.031 1.09

Fig. 5. The logarithm of air permeability, ka, as a function of the logarithm of air-filled porosity, ea, for the topsoil (circles) and the subsoil (triangles) of the no-till(NT, open symbols), mouldboard ploughing (MP, dark grey symbols) and permanentgrass (PG, light grey symbols) treatments. Each data point is the treatment averageat a certain matric potential. The lines represent fits of Eq. (11) for the topsoil(dotted lines) and subsoil (dashed lines) of each treatment.

I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140 137

t = 1.1 (Table 2). Only the topsoil of MP and PG had a value of t closeto 2, while all subsoils and the topsoil of NT had t < 1. Ghanbarian-Alavijeh and Hunt (2012) found values for t close to 2 for theirdatasets, although small values of t in a similar range to our valueswere obtained for some soils studied by Tang et al. (2011). The lowexponents in our case may be due to non-randomness of the poresize distribution and anisotropy. Macropores (which govern ka)may predominantly be vertically orientated (biopores, desiccationcracks), especially in the subsoil, and such a pore system governingadvective gas transport may be rather two-dimensional, resultingin values of t < 2. Biopores are mostly vertically orientated in thesubsoil, where anecic earthworms create continuous verticalburrows (Bouché, 1977; Lamandé et al., 2011), and roots have apredominantly vertical orientation (Soane et al., 2012). In contrast,endogeic earthworms in the topsoil produce burrows with nopreferential orientation (Bouché, 1977; Lamandé et al., 2011), andthe root system in the topsoil includes more lateral roots (Kautzet al., 2013). A vertically orientated macropore system is thereforecharacteristic for arable subsoils, which would explain the lowvalues of t found for subsoils in our study. The values of t obtainedfor the topsoils are more difficult to explain and this issue could beaddressed in future research (e.g. using computed tomographyimaging). MP may have a more random structure due to annual soilfragmentation by tillage, hence the value of t close to 2. The reasonsfor the differences between NT and PG are unclear, but they may berelated to different root systems (arable crops in NT and legumeand grass species in PG) or differences in field traffic (frequency oftraffic and size of machinery). Gas diffusion is much less affected bythe anisotropy of the macropore system because diffusion iseffectively pore-size independent in the range of matric potentialsconsidered here, which explains the generally higher values of mthan t. Berisso et al. (2013) analysed the anisotropy of ka and Dp/D0

and found higher anisotropy for ka than for Dp/D0. Their resultssupport the higher direction-dependency (and hence lowerdimensionality and lower scaling exponent) of ka than Dp/D0.Our results support findings by Arthur et al. (2012), who presentedvalues for the slope of log Dp/D0 vs. log ea (i.e. m, termed Hd byArthur et al. (2012)) and the slope of log ka vs. log ea (i.e. t, termedHa by Arthur et al. (2012)) for soil samples from two Danish arablefields. They found that the exponents were (i) smaller for ka thanfor Dp/D0, and (ii) more variable for ka than for Dp/D0. We obtaineda median value for Ha of about 1.5 and Hd of about 2.1 by re-analysing the data presented in Fig. 4 of Arthur et al. (2012).

Similarly, Kawamoto et al. (2006a) found smaller exponents (calledh in their publication) for ka than for Dp/D0. For their Gorslev soil,which has a similar texture to our soil (see Kawamoto et al.,2006b), they obtained h = 2.02 for gas diffusion and h = 1.34 for airpermeability (Kawamoto et al., 2006a).

4.3. Soil aeration and implications for plant growth

Soil aeration, i.e. gas exchange between the atmosphere and thesoil, which supports aerobic respiration by soil biota (plant roots,microorganisms) is a dynamic process that depends on variousfactors, including soil structure and soil moisture (and thereby thevolume of air-filled pores and the gas transport capability of soil,which is governed by ka and Dp/D0), temperature, the compositionof soil air (e.g. the concentrations of O2 and CO2) and O2

consumption (Boone and Veen, 1994; Stepniewski et al., 1994).The air-filled porosity and gas transport properties (ka and Dp/D0)are key properties that control the potential soil aeration, as theyimpose the soil physical constraints on soil aeration. Generally,high Dp/D0 and high ka would facilitate soil-atmosphere gasexchange (i.e. replenishment of O2 from the atmosphere into thesoil and removal of toxic gases from the root zone), while low Dp/D0

and low ka would increase the risk of anoxic conditions withnegative impacts on root growth and a potential increase ingreenhouse gas emissions.

Various lower limits for adequate soil aeration for soil physicalproperties have been proposed. The critical thresholds that havebeen established state that soil aeration is deficient if ea<0.1 m3

m�3 (Grable and Siemer, 1968) and if Dp/D0<0.005 (Stepniewski(1980, 1981) to 0.02 (Schjønning et al., 2003). We note thatea = 0.1 m3m�3 corresponds to Dp/D0 = 0.01 on applying theBuckingham (1904) model, viz. Dp/D0 = ea2. Boone (1986) estab-lished a lower and upper critical limit for oxygen diffusion toensure sufficient aeration for low and high O2 consumption,corresponding to Dp = 7.5 �10�8m2 s�1 for the lower limit andDp = 7.5 �10�8m2 s�1 (Stepniewski et al., 1994), resulting in Dp/D0 = 0.0037 for the lower limit and Dp/D0 = 0.073 for the upperlimit. The upper limit seems rather high, and was not reached evenat the driest condition (-500 hPa) in our soil (Fig. 2). The lower limitis only slightly lower than the critical limit proposed byStepniewski (1980, 1981). The upper and lower critical limitssuggested by Boone (1986) clearly show that soil physicalproperties, such as Dp/D0, are a measure of potential aeration,while the actual soil aeration depends on the actual O2 demand (i.e.the O2 consumption), among other factors. Similarly, drawingconclusions on actual greenhouse gas emissions (e.g. N2O, CO2)from values of Dp/D0 is not straight-forward, because additionalfactors such as temperature, microbial communities and N and Cavailability control aeration (Ball, 2013). Since gas transport in soilprimarily occurs by diffusion (Stepniewski et al., 1994), lowerlimits for ka are less well established. However, a soil withka = 1 mm2 is considered effectively impermeable (Ball et al., 1988),and ka<20 mm2 is suggested to limit soil aeration (Fish and Koppi,1994).

As regards the critical threshold for ea, NT was below the criticalthreshold in the topsoil when the soil was wetter than pF 2, i.e.wetter than field capacity, and PG was at critically low levels at pF�2.3 (Fig. 2a–b). In contrast, ea in MP was �0.1 m3m�3 at all matricpotentials studied (Fig. 2a–b). Based on the critical threshold of Dp/D0 = 0.02 for soil aeration proposed by Schjønning et al. (2003), ourmeasurements indicated that critically low levels of Dp/D0 werereached in the topsoil of NT and PG at nearly all matric potentialsand in the subsoil of MP at pF �2 (Fig. 2c–d), although Dp/D0 wasalways above the critical threshold of Dp/D0 = 0.005 suggested byStepniewski (1980,1981) in all treatments and at both depths. Notethat the critical limit suggested by Schjønning et al. (2003) was

138 I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140

obtained on sandy soil, while our soil was a sandy loam. In thetopsoil, PG and NT displayed ka values below log ka<1.3 at pF �2(NT) and at pF �2.3 (PG) (Fig. 2e–f), which may limit soil aeration(Fish and Koppi, 1994). However, the cited critical limits werederived from conventionally tilled soils and Reichert et al. (2009)suggest that different critical limits of soil physical conditions forroot growth may apply to no-till soils.

Nevertheless, at a given matric potential, our results clearlyindicate that the gas transport capability was: (i) highest under MPin the topsoil, but (ii) lowest under MP in the subsoil. The resultsalso suggest that Dp/D0 and ka of the topsoil were at critically lowlevels in NT and PG, which we attributed to field traffic incombination with lack of soil loosening (no-till). In contrast, Dp/D0

and ka were low in the subsoil of MP, which we attributed to theactions of mouldboard ploughing (high mechanical subsoilstresses during in-furrow ploughing, soil smearing at thetopsoil-subsoil interface, killing of earthworms due to ploughing).Tillage, in particular mouldboard ploughing, is known to have anegative impact on pore continuity (especially on the continuity ofbiopores) at the topsoil-subsoil interface, and this discontinuitymay impair root growth (Munkholm et al., 2005; Olesen andMunkholm, 2007). In contrast, reduced tillage may not be asdetrimental to earthworm populations, and reduced tillage mayresult in higher macropore volumes in the subsoil than in thetopsoil (Kautz et al., 2013).

Biological activity, root density and nutrient concentrations aregreater in the topsoil than in the subsoil, which reflects a greaterimpact of the topsoil than the subsoil on crop performance (Kautzet al., 2013; Uteau et al., 2013). Hence, the higher soil porosity andhigher gas transport capability (higher ka and higher Dp/D0) of thetopsoil in MP than in NT would indicate more favourable cropgrowth conditions in MP, especially during early crop develop-ment. However, the topsoil is typically drier (i.e. more negativematric potential or higher pF value) than the subsoil due to higherroot water uptake in the topsoil, and so the critically low gastransport capability at moist conditions in NT and PG may berelevant only under limited time periods, while the subsoil in MPmay more often be at critical levels in terms of gas transport.Furthermore, the soil may be more prone to slaking andsubsequent surface sealing in MP due to soil destabilisationduring fragmentation by tillage, which could drastically reduce soilaeration even under dry conditions (Håkansson et al., 2012). Gutet al. (2015) reported no difference in the long-term average soilmatric potential at 0.3 m depth between NT and MP. However,Chervet et al. (2006) found on average slightly more moistconditions in NT than in MP in maize during the summer months(June to August) based on tensiometer and TDR measurementsduring the period 1998–2005.

The contribution of subsoil to crop growth is not negligible. Forexample, subsoil compaction is known to result in permanent yielddecline (e.g. Håkansson and Reeder, 1994) and crops are known toacquire nutrients from the subsoil (Kautz et al., 2013). Differentcrops may be more or less dependent on the subsoil, and thedependency is affected by the climate and the weather conditions.In the Oberacker long-term field experiment, cereals (winterbarley and winter wheat) and legumes (field beans and peas)produce higher yields in NT, while tuber and root crops (potatoesand sugar beet) yield more in MP (Martínez et al., 2016). The loweryield of tuber and root crops may be related to the higher topsoilcompactness (higher BD) in NT than in MP. Cereals and legumesseem to not have suffered from the dense topsoil in NT and theirroots may have been able to better explore the subsoil in NT than inMP. The higher density in the topsoil of NT may also have beencompensated for with regard to crop productivity by the highernutrient and organic carbon concentrations in the uppermost soillayers in NT (Martínez et al., 2016). The potential specific uptake

rate (per unit root length) of nutrients is unlikely to be affected bythe tillage system (Boone and Veen, 1994).

The higher macroporosity and associated higher Dp/D0 and ka inthe subsoil under NT and PG compared with MP may be associatedwith greater earthworm abundances. Results after 10 experimentalyears of the Oberacker long-term field experiment (Maurer-Troxleret al., 2005) showed that the earthworm biomass in MP was only49.5% of that in NT. Unpublished data (Maurer-Troxler et al., 2015)from 2014, i.e. from the 19th experimental year, show similartrends. These results are in line with findings in other studies thatthe earthworm population in conventional tillage is about half thatin no-till (e.g. Lagerlöf et al., 2012). In particular, anecic earth-worms such as Lumbricus terrestris are present at much higherabundance in NT than in MP in the Oberacker experiment (Maurer-Troxler et al., 2005), resulting in a higher number of intact,continuous bio-channels in NT. This correlates with the higher gastransport capability (i.e. higher Dp/D0 and higher ka) found in thesubsoil in NT compared with MP. Earthworm bioturbationincreases the access to subsoil for crop roots in arable soils (Kautzet al., 2013), and may therefore contribute to the higher (althoughnot significantly) overall average yield under NT found for theOberacker long-term field experiment (Martínez et al., 2016).

5. Conclusions

Soil pore structure and gas transport properties (air permeabil-ity, ka, and relative gas diffusion coefficient, Dp/D0) 19 years afterinitiation of the Oberacker long-term field experiment were foundto be similar in the no-till (NT) plots and the grass strips betweenthe experimental plots (PG), while the soil pore structure wasmarkedly different in the mouldboard ploughed soil (MP). Thissuggests that the soil structure in NT evolves towards that in anatural grassland. There was clear stratification in gas transportproperties in MP, caused by a relatively loose topsoil and a compactsubsoil, while the differences between the topsoil and subsoil weresmall for NT and PG. In the topsoil, there were indications of higherspecific diffusivity and much higher specific air permeability in MPthan in NT and PG, reflecting higher continuity, higher connectivityand lower tortuosity, especially of macropores in MP. In the subsoil,the gas transport capability was lower in MP than in NT and PG,presumably due to fewer continuous biopores. The gas transportproperties reached low values under moist conditions in thetopsoil of NT and PG and in the subsoil of MP, which may restrictsoil aeration. Thus it is highly important to include the propertiesof the subsoil when evaluating tillage systems, as the results can bediametrically opposed to those in the topsoil and can affectdifferent crops in different ways.

Acknowledgements

The stay at Agroscope (Zürich, Switzerland) of the first author(Ingrid Martínez) was made possible through a Becas-Chilescholarship from the Comisión Nacional de Investigación Científicay Tecnológica (CONICYT), which is gratefully acknowledged.Marlies Sommer and Samuel Gut (Agroscope, Zurich, Switzerland)are thanked for help with the soil core sampling and support in thelaboratory. We wish to thank Per Schjønning (Aarhus University,Denmark) for valuable support with the design of the diffusionapparatus and for providing a sample of the autoclaved aeratedconcrete. The three anonymous reviewers are thanked for valuablecomments that helped to improve the manuscript.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.still.2016.05.020.

I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140 139

References

Alaoui, A., Lipiec, J., Gerke, H.H., 2011. A review of the changes in the soil pore systemdue to soil deformation: a hydrodynamic perspective. Soil Tillage Res. 115–116,1–15.

Arthur, E., Moldrup, P., Schjønning, P., de Jonge, L.W., 2012. Linking particle and poresize distribution parameters to soil-gas transport properties. Soil Sci. Soc. Am. J.76, 18–27.

Bakker, J.W., Hidding, A.P., 1970. The influence of soil structure and air content ongas diffusion in soils. Neth. J. Agric. Sci. 18, 37–48.

Ball, B.C., Schjønning, P., 2002. Air permeability. In: Dane J, H., Topp, G.C. (Eds.),Methods of Soil Analysis, Part 4. SSSA, Madison, pp. 1141–1158 (Al).

Ball, B.C., O’sullivan, M.F., Hunter, R., 1988. Gas diffusion, fluid flow and derived porecontinuity indices in relation to vehicle traffic and tillage. J. Soil Sci. 39, 327–339.

Ball, B.C., 1981. Modelling of soil pores as tubes using gas permeabilities, gasdiffusivities and water release. J. Soil Sci. 32, 465–481.

Ball, B.C., 2013. Soil structure and greenhouse gas emissions: a synthesis of 20 yearsof experimentation. Eur. J. Soil Sci. 64, 357–373.

Berisso, F.E., Schjønning, P., Keller, T., Lamandé, M., Simojoki, A., Iversen, B.V.,Alakukku, L., Forkman, J., 2013. Gas transport and subsoil pore characteristics:anisotropy and long-term effects of compaction. Geoderma 195-196 184–191.

Berkowitz, B., Ewing, R.P., 1998. Percolation theory and network modelingapplications in soil physics. Surv. Geophys. 19, 23–72.

Blackwell, P.S., Ringrose-Voase, A.J., Jayawardane, N.S., Olsson, K.A., McKenzie, D.C.,Mason, W.K., 1990. The use of air-filled porosity and intrinsic permeability to airto characterize structure of macropore space and saturated hydraulicconductivity of clay soils. J. Soil Sci. 41, 215–228.

Blanco-Canqui, H., Lal, R., 2007. Soil structure and organic carbon relationshipsfollowing 10 years of wheat straw management on no-till. Soil Tillage Res. 95,240–254.

Boone, F.R., Veen, B.W., 1994. Mechanisms of crop responses to soil compaction. In:Soane, van Ouwerkerk (Eds.), Soil Compaction in Crop Production. Elsevier,Amsterdam, pp. 237–264.

Boone, F.R., 1986. Towards compaction limits for crop growth. Neth. J. Agric. Sci. 34,349–360.

Bouché, M., 1977. Stratégies lombriciennes. Ecol. Bull. 25, 122–132.Brussaard, L., van Faassen, H.G., 1994. Effects of compaction on soil biota and siol

biological processes. In: Soane, van Ouwerkerk (Eds.), Soil Compaction in CropProduction. Elsevier, Amsterdam, pp. 215–235.

Buckingham, E., 1904. Contributions to our knowledge of the aeration of soils. USBur. Soils Bulletin, 25, U.S. Gov. Print. Off., Washington, D.C

Chen, G., Weil, R.R., Hill, R., 2014. Effects of compaction and cover crops on soil leastlimiting water range and air permeability. Soil Tillage Res. 136, 61–69.

Chervet, A., Ramseier, L., Sturny, W.G., Weisskopf, P., Zihlmann, U., Müller, M.,Schafflützel, R., 2006. Bodenwasser bei Direktsaat und Pflug. Agrarforschung 13,162–169.

Díaz-Zorita, M., Grove, J.H., 2002. Duration of tillage management affects carbonand phosphorus stratification in phosphatic Paleudalfs. Soil Tillage Res. 66, 165–174.

Dörner, J., Horn, R., 2009. Direction-dependent behaviour of hydraulic andmechanical properties in structured soils under conventional and conservationtillage. Soil Tillage Res. 102, 225–232.

Deepagoda, T.K.K.C., Moldrup, P., Schjønning, P., de Jonge, L.W., Kawamoto, K.,Komatsu, T., 2011. Density-corrected models for gas diffusivity and airpermeability in unsaturated soil. Vadose Zone J. 10, 226–238.

Deubel, A., Hormann, B., Orzessek, D., 2011. Long-term effects of tillage onstratification and plant availability of phosphate and potassium in a loesschernozem. Soil Tillage Res. 117, 85–92.

Dexter, A.R., Richard, G., Arrouays, D., Czyz, E.A., Jolivet, C., Duval, O., 2008.Complexed organic matter controls soil physical properties. Geoderma 144,620–627.

Di Rienzo, J.A., Casanoves, F., Balzarini, M.G., Gonzales, L., Tablada, M., Robledo, C.W.,2009. InfoStat Version. Grupo InfoStat, FCA, Universidad Nacional de Córdoba,Argentina.

Eden, M., Schjønning, P., Moldrup, P., de Jonge, L.W., 2011. Compaction androtovation effects on soil pore characteristics of a loamy sand soil withcontrasting organic matter content. Soil Use Manag. 27, 340–349.

Fish, A.N., Koppi, A.J., 1994. The use of a simple field air permeameter as a rapidindicator of functional soil pore space. Geoderma 63, 255–264.

Ghanbarian, B., Hunt, A.G., 2014. Universal scaling of gas diffusion in porous media.Water Resour. Res. 50, 2242–2256.

Ghanbarian-Alavijeh, B., Hunt, A.G., 2012. Comparison of the predictions ofuniversal scaling of the saturation dependence of the air permeability withexperiment. Water Resour. Res. 48 doi:http://dx.doi.org/10.1029/2011WR011758 (W08513).

Grable, A.R., Siemer, E.G., 1968. Effect of bulk density, aggregate size, and soil watersuction on oxygen diffusion redox potentials, and elongation of corn roots. SoilSci. Soc. Am. J. 32, 180–186.

Gradwell, M.W., 1961. A laboratory study of the diffusion of oxygen through pasturetopsoils. N. Z. J. Sci. 4, 250–270.

Groenevelt, P.H., Kay, B.D., Grant, C.D., 1984. Physical assessment of a soil withrespect to rooting potential. Geoderma 34, 101–114.

Gut, S., Chervet, A., Stettler, M., Weisskopf, P., Sturny, W.G., Lamandé, M., Schjønning,P., Keller, T., 2015. Seasonal dynamics in wheel load-carrying capacity of a loamin the Swiss Plateau. Soil Use Manag. 131, 132–141.

Håkansson, I., Reeder, R.C., 1994. Subsoil compaction by vehicles with high axle loadextent: persistence and crop response. Soil Tillage Res. 29, 277–304.

Håkansson, I., Keller, T., Arvidsson, J., Rydberg, T., 2012. Effects of seedbed propertieson crop emergence: 4. Inhibitory effects of oxygen deficiency. Acta Agric. Scand.B 62, 166–171.

Håkansson, I., 2005. Machinery-induced compaction of arable soils. Reports fromthe division of soil management, Uppsala, no. 109, 153 pp.

Hartmann, P., Zink, A., Fleige, H., Horn, R., 2012. Effect of compaction, tillage andclimate change on soil water balance of Arable Luvisols in Northwest Germany.Soil Tillage Res. 124, 211–218.

Iversen, B.V., Moldrup, P., Schjønning, P., Loll, P., 2001. Air and water permeability indifferently textures soils at two measurements scales. Soil Sci. 166, 643–659.

Jarvis, N.J., 2007. A review of non-equilibrium water flow and solute transport in soilmacropores: principles, controlling factors and consequences for water quality.Eur. J. Soil Sci. 58, 523–546.

Kühne, A., Schack-Kirchner, H., Hildebrand, E.E., 2012. Gas diffusivity in soilscompared to ideal isotropic porous media. J. Plant Nutr. Soil Sci. 175, 34–45.

Kautz, T., Amelung, W., Ewert, F., Gaiser, T., Horn, R., Jahn, R., Javaux, M., Kemna, A.,Kuzyakov, Y., Munch, J., Pätzold, S., Peth, S., Scherer, H.W., Schloter, M.,Schneider, H., Vanderborght, J., Vetterlein, D., Walter, A., Wiesenberg, G.L.B.,Köpke, U., 2013. Nutrient acquisition from arable susoils in temperate climates Areview. Soil Biol. Biochem. 57, 1003–1022.

Kawamoto, K., Moldrup, P., Schjønning, P., Iversen, B.V., Komatsu, T., Rolston, D.E.,2006a. Gas transport parameters in the vadose zone: Development and tests ofpower-law models for air permeability. Vadose Zone J. 5, 1205–1215.

Kawamoto, K., Moldrup, P., Schjønning, P., Iversen, B.V., Rolston, D.E., Komatsu, T.,2006b. Gas transport parameters in the vadose zone: gas diffusivity in field andlysimeter soil profiles. Vadose Zone J. 5, 1194–1204.

Kay, B.D., VandenBygaart, A.J., 2002. Conservation tillage and depth stratification ofporosity and soil organic matter. Soil Tillage Res. 66, 107–118.

Lagerlöf, L., Pålsson, O., Arvidsson, J., 2012. Earthworms influenced by reducedtillage, conventional tillage and energy forest in Swedish agricultural fieldexperiments. Acta Agric. Scand. B 62, 235–244.

Lamandé, M., Labouriau, R., Holmstrup, M., Torp, S.B., Greve, M.H., Heckrath, G.,Iversen, B.V., de Jonge, L.W., Moldrup, P., Jacobsen, O.H., 2011. Density ofmacropores as related to soil and earthworm community parameters incultivated grasslands. Geoderma 162, 319–326.

Lipiec, J., Hatano, R., 2003. Quantification of compaction effects on soil physicalproperties and crop growth. Geoderma 116, 107–136.

Martínez, I., Chervet, A., Weisskopf, P., Sturny, W.G., Etana, A., Stettler, M., Forkmann,J., Keller, T., 2016. Two decades of no-till in the Oberacker long-term fieldexperiment: Part 1. Crop yield, soil organic carbon and nutrient distribution inthe soil. Soil Till. Res. doi:http://dx.doi.org/10.1016/j.still.2016.05.021 in press.

Maurer-Troxler, C., Chervet, A., Ramseier, L., Sturny, W.G., Oberholzer, H.-R., 2005.Bodenbiologie nach zehn Jahren Direktsaat und Pflug. Agrarforschung 12, 460–465.

Moldrup, P., Deepagoda, T.K.K.C., Hamamoto, S., Komatsu, T., Kawamoto, K., Rolston,D.E., de Jonge, L.W., 2013. Structure-dependent water-Induced linear reductionmodel for predicting gas diffusivity and tortuosity in repacked and intact soil.Vadose Zone J. 12 doi:http://dx.doi.org/10.2136/vzj2013.01.0026.

Munkholm, L.J., Schjønning, P., Rüegg, K., 2005. Mitigation of subsoil recompactionby light traffic and on-land ploughing I. Soil response. Soil Tillage Res. 80, 149–158.

Nunes, M.R., Denardin, J.E., Pauletto, E.A., Faganello, A., Spinelli Pinto, L.F., 2015.Mitigation of clayey soil compaction managed under no-tillage. Soil Tillage Res.148, 119–126.

Olesen, J.E., Munkholm, L.J., 2007. Subsoil loosening in a crop rotation for organicfarming eliminated plough pan with mixed effects on crop yield. Soil Tillage Res.94, 376–385.

Prasuhn, V., 2012. On-farm effects of tillage and crops on soil erosion measured over10 years in Switzerland. Soil Tillage Res. 120, 137–146.

Reichert, J.M., Suzuki, L.E.A.S., Reinert, D.J., Horn, R., Håkansson, I., 2009. Referencebulk density and critical degree-of-compactness for no-till crop production insubtropical highly weathered soils. Soil Tillage Res. 102, 242–254.

Reynolds, W.D., Clarke Topp, G., 2008. Soil water analyses: principles andparameters, In: Carter, M.R., Gregorich, E.G. (Eds.), Soil Sampling and Methods ofAnalysis. 2nd edition CRC Press, Taylor & Francis Group, Boca Raton, FL, USA, pp.913–937.

Schjønning, P., Thomsen, I.K., 2013. Shallow tillage effects on soil properties fortemperate-region hard-setting soils. Soil Tillage Res. 132, 12–20.

Schjønning, P., Munkholm, L.J., Moldrup, P., Jacobsen, O.H., 2002. Modelling soil porecharacteristics from measurements of air exchange: the long-term effects offertilization and crop rotation. Eur. J. Soil Sci. 53, 331–339.

Schjønning, P., Thomsen, I.K., Moldrup, P., Christensen, B.T., 2003. Linking soilmicrobial activity to water and air-phase contents and diffusivities. Soil Sci. Soc.Am. J. 67 (156.165).

Schjønning, P., Thomsen, I.K., Petersen, S.O., Kristensen, K., Christensen, B.T., 2011.Relating soil microbial activity to water content and tillage-induced differencesin soil structure. Geoderma 163, 256–264.

Schjønning, P., Eden, M., Moldrup, P., de Jonge, L.W., 2013. Two-chamber, two-gasand one-chamber, one-gas methods for measuring the soil-gas diffusioncoefficient: validation and inter-calibration. Soil Sci. Soc. Am. J. 77, 729–740.

Soane, B.D., Ball, B.C., Arvidson, J., Basch, G., Moreno, F., Roger-Estrade, J., 2012. No-till in northern, western and south-western Europe: a review of problems andopportunities for crop production and the environment. Soil Tillage Res. 118,66–87.

140 I. Martínez et al. / Soil & Tillage Research 163 (2016) 130–140

Stepniewski, W., Gli�nski, J., Ball, B.C., 1994. Effects of compaction on soil aerationproperties. In: Soane, van Ouwerkerk (Eds.), Soil Compaction in CropProduction. Elsevier, Amsterdam, pp. 167–189.

Stepniewski, W., 1980. Oxygen diffusion and strength as related to soil compaction:I. ODR. Pol. J. Soil Sci. 13, 3–13.

Stepniewski, W., 1981. Oxygen diffusion and strength as related to soil compactionII. Oxygen diffusion coefficient. Pol. J. Soil Sci. 14, 3–13.

Tang, A.M., Cui, Y.-J., Richard, G., Défossez, P., 2011. A study on the air permeability asaffected by compression of three French soils. Geoderma 162, 171–181.

Troeh, F.R., Jabro, J.D., Kirkham, D., 1982. Gaseous diffusion equations for porousmaterials. Geoderma 37, 239–253.

Uteau, D., Pagenkemper, S.K., Peth, S., Horn, R., 2013. Root and time dependent soilstructure formation and its influence on gas transport in the subsoil. Soil TillageRes. 132, 69–76.