Compaction of restored soil by heavy agricultural machinery—Soil physical and mechanical aspects

Compaction of restored soil by heavy agricultural machinery—Soil physical and mechanical aspects

Soil & Tillage Research 93 (2007) 28–43 Compaction of restored soil by heavy agricultural machinery—Soil physical and m...

1MB Sizes 1 Downloads 28 Views

Recommend Documents

No documents
Soil & Tillage Research 93 (2007) 28–43

Compaction of restored soil by heavy agricultural machinery—Soil physical and mechanical aspects B. Scha¨ffer *, W. Attinger, R. Schulin Institute of Terrestrial Ecology, ETH Zu¨rich, Universita¨tstrasse 16, 8092 Zu¨rich, Switzerland Received 19 August 2005; received in revised form 6 February 2006; accepted 21 March 2006

Abstract As construction and open-cast mining activities continue to expand on fertile agricultural land, the removal and subsequent restoration of soil to be re-used for plant growth has become an increasingly important issue in soil protection. A key factor for the success of soil restoration is that the soil is allowed to develop sufficient mechanical strength to withstand the stresses involved in the intended type of land use. The objective of this study was to investigate the effects of the first use of heavy agricultural machinery on the physical and mechanical properties of a restored soil after the period of restricted cultivation (as prescribed by current guidelines), when the soil is re-submitted to normal agricultural management. We performed two traffic experiments on a soil which had been restored according to current guidelines 4 years before the beginning of the study. In the first year of the study, a combine harvester passed two times across the wetted experimental area, and in the following year 10 times. Two passes along the same tracks caused only weak compaction effects, mainly reducing coarse porosity. In contrast, after 10 passes, deep ruts had formed, and coarse porosity was drastically reduced down to the subsoil. Confined uniaxial compression tests revealed an increase in precompression stress and a decrease in the slope of the virgin compression line, i.e. the compression index, after 10 passes. However, precompression stress was still much lower than the exerted soil stresses at the corresponding soil depths, indicating that due to the short duration of the wheel loadings equilibrium conditions were not reached in the traffic experiments and that further compaction would have occurred with additional passes. The decrease in compression index found after 10 passes may be due to the practice that samples are pre-conditioned to a specified water tension for the oedometer tests. The results show that loads may exceed precompression stress for short durations even in a restored soil which is still far from having re-gained normal strength without serious damage. Thus, the use of precompression stress as a criterion for traffickability was on the safe side in preventing damage to the ecological quality of the soil by compaction, even if the concept did not fully apply to the field reality of the mechanical stress conditions. # 2006 Elsevier B.V. All rights reserved. Keywords: Soil restoration; Soil compaction; Precompression stress; Compression index; Bolling probe pressure

1. Introduction Large areas of fertile land are temporarily used for construction and open-cast mining purposes or permanently destroyed by building. In the course of these activities, large quantities of soils are excavated. In

* Corresponding author. Tel.: +41 44 633 61 43; fax: +41 44 632 11 08. E-mail address: [email protected] (B. Scha¨ffer). 0167-1987/$ – see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.still.2006.03.007

many cases these soils are later restored at the original site as in many open-cast tunnelling, pipeline construction and gravel exploitation projects or used for soil restoration at a new site. In the Canton of Zurich (Switzerland), for instance, 50–60 ha of agricultural land are presently restored per year for re-cultivation of closed gravel pits and other landscaping activities. This corresponds to 0.07–0.08% of the total agricultural area of the canton (about 75 000 ha). Thus, restoration of soil to be re-used for plant growth has become an increasingly important issue in soil protection.

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

Excavation, transport and repacking disrupt the structure of a soil and cause a rearrangement of clods, aggregates and particles. This leads to mechanical destabilization and an increased risk of soil compaction. The risk depends on many factors, including soil wetness, texture and stabilization by plant roots, as well as on the care with which the soil is handled (Schro¨der et al., 1985). The increased risk of compaction does not end with the rebuilding of the soil. Freshly restored soils have a low degree of aggregation and are very susceptible to compaction (Lebert and Springob, 1994). Over the years, strength will redevelop (Schneider and Schro¨der, 1991) due to physical (Voorhees, 1983; Bullock et al., 1985), chemical (Dexter et al., 1988) and biological (Von Albertini et al., 1995) regeneration processes. The formation of aggregates is considered to play a key role in this process (Horn, 1983, 1988; Baumgartl and Horn, 1991; Lebert and Horn, 1991). It is important for the success of a soil restoration that the soil is allowed to regain sufficient mechanical stability before it is used again for agriculture. Undue handling during cultivation operations within the first years after restoration may easily damage or even completely destroy the weak soil structure and thus reduce water conductivity (Logsdon et al., 1992; Arvidsson, 2001) and air permeability (Horn, 1986; Gysi et al., 1999). This in turn may have negative ecological (Soane and van Ouwerkerk, 1995) and economical (e.g. yield losses: Ha˚kansson and Reeder, 1994) consequences. On the other hand, economic interest in the re-use of restored soil for crop production creates pressure to minimize the time of restricted cultivation. Responding to this conflict between economic interests and the imperative of sustainability, guidelines like those of Zwo¨lfer et al. (1991), VSS (2000), BUWAL (2001) and FSK (2001) have been issued. These guidelines only allow very restricted land use for at least three to four years after restoration in order to avoid compaction by excessive mechanical stresses. The guidelines are based on practical experience and to some degree represent a compromise between land use and soil protection interests. The high compaction risk of restored soils calls for preventive measures. The regeneration of mechanical stability after disturbance depends on soil properties and on the way how the soil is repacked and subsequently cultivated (Lebert, 1991; Davies and Younger, 1994). Therefore a stability criterion is needed that is directly related to the mechanical properties of restored soils. The concept of precompression stress has been applied to agricultural soil mechanics e.g. by Horn (1981) and Kirby (1991a). Precompression stress is


determined from compression curves (void ratio versus logarithm of normal stress) obtained by confined uniaxial compression tests (Koolen, 1974). Conceptually, it indicates the maximum stress a soil has been submitted to before under given conditions (Kirby, 1991a; Veenhof and McBride, 1996). According to the conventional concept of precompression stress, deformation is elastic (reversible) at stresses below and plastic (irreversible) above precompression stress. Compression above precompression stress occurs along the virgin compression line (VCL). The slope of the VCL, i.e. the ‘‘compression index’’ (CI) according to Larson et al. (1980), is inversely related to soil stiffness. Upon unloading, the maximum applied stress becomes the new precompression stress, and upon reloading, the soil is further compressed along the original VCL when the applied stress exceeds this new precompression stress. Thus, the concept implies that precompression stress increases during compaction, whereas CI remains unaffected. Its conceptual features make precompression stress attractive as a criterion for the limit up to which a soil may be loaded without irreversible damage to its ecological functions, i.e. as an indicator of ‘‘ecological’’ trafficability. The latter should be distinguished from what may be called ‘‘technical’’ trafficability, for which e.g. the empirical ‘‘California Bearing Ratio’’ (CBR, see e.g. Porter, 1950), which is an index of the shear strength of a soil (Turnbull, 1950), is widely used. Despite the advantage of the precompression stress concept that it directly relates soil stress to strength, there are a number of uncertainties when using precompression stress determined by laboratory tests as an indicator of the maximum stress experienced by the soil in situ. First, the concept presupposes that stress–strain conditions in soil samples during uniaxial compression tests are comparable to those in undisturbed soil subjected to mechanical load in the field. In particular, the concept relates to equilibrium conditions (i.e. sufficiently long exposure to a load, slow increase in load) and laterally confined expansion. These conditions may often not be sufficiently fulfilled in field situations. As a consequence, the (time-dependent) stress tensors describing the stress distribution below a moving wheel (e.g. Abu-Hamdeh and Reeder, 2003) and in an uniaxial compression test (Koolen, 1974) can be quite different. Second, precompression stress is conditional on the soil moisture status and the drainage conditions, which are usually not the same in laboratory tests and the field. Third, the concept refers to conditions with no changes in the structural arrangement of soil particles (i.e.


B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

turbation, kneading or crack formation) other than vertical displacement. Fourth, the procedure by Casagrande (1936), which is usually applied to determine precompression stress, is suited to define precompression stress when the change from purely elastic to plastic deformation occurs rather abruptly. However, in unsaturated soils, the transition is often very gradual (e.g. Arvidsson and Keller, 2004), and plastic deformation is likely to take place even below the stress defined by the Casagrande procedure (Keller et al., 2004). Finally, sampling, sample trimming (Muhs and Kany, 1954), sample dimensions (Muhs and Kany, 1954; Koolen, 1974), as well as the experimental testing method (e.g. Keller et al., 2004) will influence the results. The concept of precompression stress was tested for validity in several field studies on land with undisrupted subsoil. Gysi et al. (1999) found that after a single vehicle pass, precompression stress was increased to approximately the maximum soil pressure where soil stability was exceeded. Semmel and Horn (1995) reported that intensive trafficking during several years increased precompression stress in comparison to soil under conservation tillage. Keller et al. (2002) observed distinct permanent soil displacement at locations where the vertical stress exerted by the machine exceeded the precompression stress and much less, but still measurable, residual displacement under conditions where the applied load did not exceed the precompression stress. Arvidsson et al. (2001) found a good agreement between the calculated depth of compaction and the depth to which vertical soil displacement was observed. Also Berli et al. (2004) showed that expected and measured compaction after trafficking with tracked machinery corresponded well. However, we are not aware of any field studies in which the precompression stress concept was tested for restored soils. Thus, the objective of this study was to investigate the effects of the first use of heavy agricultural machinery on the physical and mechanical properties of a restored soil and to determine whether compaction effects were in agreement with the concept of precompression stress as outlined above. We performed two traffic experiments on a restored soil at the end of the period of restricted cultivation prescribed by guidelines, after which the soil is submitted to normal agricultural management for the first time. During the first experiment a fully loaded combine harvester made two passes and during the second experiment 10 passes over wetted areas of restored land. We measured soil pressure during the passages of the machine and determined mechanical

and physical soil properties of samples taken from trafficked and non-trafficked locations. 2. Material and methods 2.1. Test site The traffic experiments were performed at a restored site in the north-western part of Switzerland near Solothurn (78330 E, 478120 N). The soil of the site had been restored in autumn 1999 for agricultural cultivation above a highway tunnel which had been constructed in open-cast fashion. The original soil was a Eutric Cambisol. Restoration operations had been allowed only under conditions where the water potential of the soil was below 6 kPa. The soil had been repacked in stripes of 6 m width using a crawler excavator which never trafficked the freshly deposited soil (Rohr and Schuler, 2004). The restored soil consisted of a 40 cm layer of former topsoil material above a 70 cm layer of former subsoil material. The subsoil was built on a 40 cm thick drainage layer, consisting of gravel-rich sand, which in turn was built on top of an impermeable bottom layer (artificially compacted excavation material consisting of clay and sandstone), separating the restored soil from the tunnel roof. 2.2. Management and preparation of the test site A few days after the soil had been put in place in 1999, radish (Brassica oleara) was sown manually. In the following spring (2000), after the topsoil had been mulched, grubbed and harrowed, a grass mixture (UFA 323 Gold) including clover and alfalfa (Medicago sativa) was sown. In late spring the grass was mown and left on site. All these operations were conducted using a Steyr 968 tractor with a total weight of 3985 kg. The maximum additional weight of tillage equipment on the tractor was approximately 650 kg. This corresponds to mean ground contact pressures of 28–33 kPa (static load). In the following 3 years, the site was mowed two times (2001) to three times (2002 and 2003) a year, using the same Steyr 968 equipped with a front mower. The mean contact pressure of the tractor during these operations was 33 kPa. The grass was harvested either for silage by means of a Jaguar crop chopper of 6830 kg with a mean contact pressure of 48 kPa or as hay after drying on-site. For transportation a trailer was used that weighed 4000 or 6000 kg when fully loaded with hay or silage, respectively. The resulting mean ground contact pressures varied between 113 and 170 kPa under the

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43


Fig. 1. Layout of the two traffic experiments (TE1: 2 passes; TE2: 10 passes) with sampling and measuring locations of TE1 (N1: non-trafficked; T1: trafficked; T1a: trafficked, additional sampling under the second track; B1: Bolling probes; TM1: tensiometers) and of TE2 (N2: non-trafficked; T2: trafficked; B2: Bolling probes; TM2: tensiometers).

trailer and between 37 and 42 kPa under the tractor. From summer 2003 until the end of our experiments in autumn 2004, manual mowing was the only operation performed on the test site. During the whole period soil water potentials were monitored using tensiometers. Operations on the land were only allowed when the soil water pressure was below 15 kPa in the topsoil and below 10 kPa in the subsoil (Rohr, 2002). Equipment with ground contact pressures higher than 60 kPa was allowed to operate only when the soil was definitely drier than that, i.e. when the respective tensiometer readings gave more negative values than these limits. Thus, cultivation fully conformed with Swiss guidelines (VSS, 2000; BUWAL, 2001; FSK, 2001). In autumn 2004, after the traffic experiments had been completed, the soil was ploughed to a depth of 0.20 m, and cultivation was changed from grassland to crop rotation. Adopting conventional terminology, we refer to the upper 0.20 m as ‘‘topsoil’’ and to the soil below as ‘‘subsoil’’. 2.3. Traffic experiments The first of the two traffic experiments (denoted as TE1) was conducted in summer 2003 and the second (denoted as TE2) in summer 2004. Their design is depicted in Fig. 1. TE1 consisted of two passes, representing a moderate loading as it may occur within a field during normal cultivation. TE2 consisted of 10 passes. It represents the situation of high loading as it may occur at the edge of a field, where traffic concentrates and manoeuvres are carried out. Because we were interested in the effects of trafficking under realistic worst case conditions the two sub-plots were irrigated several times in the days

before the experiments. After the last irrigation, the water was left to drain and to redistribute in the profile for 1 day before the soil was trafficked. Tensiometer readings were taken immediately before the passages with the combine harvester. Soil water potentials were similar in both traffic experiments (Table 1). They ranged between 3 and 4 kPa in the topsoil and between 2 and 3 kPa in the subsoil. 2.4. Machine properties For both traffic experiments we used the same combine harvester, model Claas Dominator 76, fully loaded and equipped with a mower bar (Fig. 2). The total weight of the machine was 9730 kg. The combine harvester had V-shaped Goodyear Traction Sure Grip front tyres (dimensions 23.1–26) and Dunlop rear tyres with longitudinal ribs (dimensions 12.5/80–18). Tyre inflation pressure was around 120 kPa in both experiments. As both the front and rear tyres were slightly flattened on the ground, the tyre–soil contact areas were approximately of rectangular shape. Length and width of the contact areas were determined on hard ground using a measuring tape. Mean ground contact pressure, Table 1 Soil water potentials at time of trafficking (averaged values; standard errors in parentheses) Traffic experiment



Soil water potential (kPa) 0.15–0.20 m depth

0.30–0.35 m depth

1 2

2.9 (0.2) 4.2 (0.4)

2.8 (0.3) 3.2 (0.3)

1 2

3.0 (0.1) 2.8 (0.1)

2.4 (0.1) 2.1 (0.1)

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43


Fig. 2. Combine harvester used for the traffic experiments.

i.e. the ratio between wheel load and contact area (length  width of the respective tyres), was around 120 kPa under the front tyres and 200 kPa under the rear tyres (Table 2). Shape and dimensions of the contact areas did not vary visibly in the course of the traffic experiments. The rear axle of the machine was slightly narrower than the front axle. Consequently, the tracks of the front and rear tyres were slightly displaced relative to each other and the total track width was larger than the width of the individual tyres (Fig. 1 and Table 2). The speed of the vehicle was approximately 0.2 m s1 in both experiments. 2.5. Measurements and parameter determinations Soil pressure during the vehicle passes was measured by means of Bolling probes (Bolling, 1987) installed at a depth of 0.32 m directly under as well as 0.13 m to the left and to the right of the centre line of the front tyre track (denoted as centre, inner and outer probes with regards to the centre of the vehicle axle). Pressure readings were taken at a frequency of 32 s1 by means of a data logger. Samples were collected within 1–4 days after trafficking from two sub-plots at sampling locations below and beside the tracks, as depicted in Fig. 1. Samples were taken from the topsoil at 0.12–0.22 m depth and from the subsoil at 0.27–0.37 m depth (with respect to the soil surface of the respective sampling locations after trafficking) using sharpened steel cylinders with a volume of approximately 1 l (diameter

10.8 cm, height 11.0 cm). Due to the formation of ruts, it was not possible to take the samples from the trafficked locations at exactly the depths corresponding to the specified sampling depths before displacement by wheeling. However, topsoil and subsoil of the reconstituted profile were quite homogeneous and the differences between the sampled layers were small, indicating that vertical gradients were almost negligible. As the results show, the differences between the two also remained small after trafficking. Thus, any error that may have originated from taking all samples at the same depths with reference to the ‘‘modified’’ soil surface (i.e. also in the ruts after trafficking) was considered tolerable. Twelve samples were taken per sub-plot, depth and treatment, resulting in 96 soil samples per experiment. After TE1 we also sampled five cores per sub-plot and depth under the second track (total of 20 samples from T1a locations in Fig. 1) in order to check if the effects under the two tracks were comparable. These cores were sampled 11 weeks after the first traffic experiment. Therefore, they were not included in the evaluation and statistical analysis of the trafficking effects. The soil samples were stored at a temperature of 4 8C until they were analyzed for coarse porosity, fine-tointermediate porosity, bulk density, precompression stress and CI. The samples were water-saturated, weighed, and subsequently desorbed to a water potential of 6 kPa (value with respect to the centre of the sample) by applying a hanging water column. Then they were weighed again. The drained pore volume was interpreted as the coarse porosity that is drained at field capacity (AG-Boden, 1982; Berli et al., 2004). Thereafter we used the samples conditioned to 6 kPa water potential to conduct confined uniaxial compression tests as described by Berli et al. (2004). The cylinders with the samples were built into the compression cells of oedometers. The samples were subjected to 16 steps of increasing pressure from 4 to 600 kPa. Each compression step lasted for 1800 s. In general, this time allowed for drained compression conditions. Only at very high pressures, some samples were not sufficiently drained. This resulted in a decrease

Table 2 Characteristics of the fully loaded and equipped combine harvester used for the traffic experiments

Front tyres Rear tyres

Wheel load (kg)

Width of contact area (m)

Length of contact area (m)

Contact area (m2)

Ground contact pressure (kPa)

3455 1410

0.52 0.27

0.55 0.25

0.2860 0.0675

119 205

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

of the slope of the VCL. When this occurred, the respective measurements were not taken into account in the analysis. Precompression stress was determined numerically from the resulting compression curves according to the procedure of Casagrande (1936) using an Excel spreadsheet. The bisector of the horizontal and tangent through the point of maximum curvature was calculated, and the VCL was determined by fitting a regression line to the linear part of the compression curve at high stresses. The relevant points were chosen by visual inspection. Precompression stress was calculated from the intersection of the bisector and the VCL, and the CI from the slope of the VCL. After compression, the samples were dried at 105 8C for at least 6 days and weighed again. The pore volume desorbed between 6 kPa water potential and oven-dry state was interpreted as fine-to-intermediate porosity. The oven-dry weight was also used to calculate bulk density (Blake and Hartge, 1986). Finally, the dried samples were taken out of the steel cylinders and wetsieved, using a 2 mm sieve. The coarse fraction was dried at 105 8C and weighed. We used the mean gravel density of the samples to calculate the volumetric gravel content. To avoid that heterogeneities in the spatial distribution of the coarse fraction within the test site would add additional variance to the data, we subtracted its volume and mass from the respective values of the bulk soil and calculated the coarse porosity CPfe, the fineto-intermediate porosity FIPfe, and the bulk density BDfe with respect to the residual mass and volume, i.e. to the fine earth fraction: CPfe ¼ VCP =ðVSample  Vcf Þ


FIPfe ¼ VFIP =ðVSample  Vcf Þ


BDfe ¼ ðmSample  mcf Þ=ðVSample  Vcf Þ


where the subscript fe denotes the fine earth fraction, VCP and VFIP are the volumes of the coarse pores and fine-to-intermediate pores, VSample and msample are the sample volume and mass, and Vcf and mcf are the volume and mass of the coarse fraction. (We performed the same analyses as we did with these fine-earth related


measurements also with the respective measurements including the coarse fraction, and obtained very similar results, but with less statistical significance.) 2.6. Statistical analysis For the evaluation and statistical analysis of the trafficking effects, arithmetic means of the measured values were determined, separately for each sampling location and depth (i.e. N1, T1, N2, T2 locations, cf. Fig. 1), resulting in four replicates of non-trafficked reference locations (two sub-plots each in 2003 and 2004) and in two replicates each of the moderately (2 passes) and strongly trafficked (10 passes) soil per depth (cf. Fig. 1). Compaction effects due to trafficking were analysed by two-way analysis of variance with depth and number of passes as independent variables, followed by a Fisher’s least significant difference post hoc analysis to compare differences between group means of the dependent variables. Main effects of the independent variables and their interaction as well as the differences between the group means of the dependent variables were considered to be significant if the probability level was 0.05 or less. 3. Results 3.1. Soil properties Table 3 gives the soil texture, organic matter and gravel content of the experimental site. The texture was a loam according to US soil taxonomy (USDA, 1997). The organic matter content of the test site was around 0.02 kg kg1. It was slightly lower in the subsoil. The average gravel content was about 10% by volume, the mean gravel densities were 2.31 g cm3 in the sampling locations of TE1 and 2.28 g cm3 in the sampling locations of TE2. Whereas gravel content was highly variable within the test site, texture, organic matter content and also the soil physical and mechanical properties of the non-trafficked locations were quite homogeneous within the test site. Nonetheless, some

Table 3 Texture, organic matter and gravel content of the soil material (averaged values of the sampling locations, cf. Fig. 1; standard errors in parentheses)

Topsoil Subsoil a b

Sampling locations

Sanda (kg kg1)

Silta (kg kg1)

Claya (kg kg1)

Organic matterb (kg kg1)

Gravel (m3 m3)

N1, T1, N2, T2 N1, T1, N2,T2

0.498 (0.022) 0.514 (0.016)

0.319 (0.011) 0.314 (0.009)

0.183 (0.011) 0.173 (0.007)

0.022 (0.003) 0.019 (0.004)

0.097 (0.010) 0.094 (0.019)

Determined with the pipette method. Measured as weight loss after oxidation by H2O2.


B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

Fig. 3. Tracks after 2 passes (left) and 10 passes (right).

sampling locations differed significantly from the others, in the topsoil with respect to texture (P < 0.05), gravel content (P < 0.001), coarse porosity (P < 0.01), fine-to-intermediate porosity (P < 0.001), bulk density (P = 0.001) and precompression stress (P < 0.01), and in the subsoil with respect to texture (P < 0.05), gravel content (P < 0.001), coarse porosity (P < 0.05), fine-to-intermediate porosity (P < 0.001), bulk density (P < 0.01) and precompression stress (P < 0.01) (Kruskal–Wallis test). These differences indicate a spatial variability which is typical for restored soils.

from the tyre lugs. The 10 passes of the second experiment produced ruts of up to 10 cm depth, and at certain locations they were even deeper. The soil structure of the trafficked sub-plots was not visibly altered by the two passes of TE1. It remained crumbly in the upper 10 cm and blocky below. The 10 passes of TE2 destroyed these aggregates and led to a dense cloddy or even massive, coherent structure down to the bottom of the opened profiles, i.e. to at least 40 cm depth below the ruts.

3.2. Visible compaction effects in the field

To reduce random scatter, the Bolling probe pressure data were smoothed by taking the running median of five subsequent values for each point in time. Fig. 4 shows an example of the smoothed time course of the Bolling probe pressure measurements, BPPi(t). Index i denotes the number of the pass (i = 1, 2 for TE1 and

The passages of the combine harvester left clearly visible ruts in both experiments (Fig. 3). The ruts of the two passes of the first experiment were of moderate depth (2–6 cm), consisting mainly of the impressions

3.3. Soil pressure during trafficking

Fig. 4. Example of the time course of the smoothed Bolling probe pressure, BPPi(t) (- - -), its first derivative ( ) used to define the beginning and the end of the pass by the front tyre, t1 and t2, and by the rear tyre, t3 and t4, and the additional soil pressure exerted by the front tyre, SPi,f(t) (for t1  t  t2), and the rear tyre, SPi,r(t) (for t3  t  t4), of the combine harvester (—). P¯ i;f (for t < t1), P¯ i;r (for t2 < t < t3) and P¯ i;e (for t4 < t) are the baselines of BPPi(t) before the passes of the front and rear tyre and at the end of the vehicle pass, respectively. Index i denotes the number of the pass (i = 1, 2 for TE1 and i = 1, 2, . . ., 10 for TE2).

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43


Fig. 5. Peak pressures exerted by the combine harvester at 0.32 m depth vs. number of passes under the front tyre (left) and rear tyre (right) during TE1 (2 passes) and TE2 (10 passes). Averaged values of the measurements of the centre (*, ~), inner ( , ) and outer probes (*, ~) (with regard to the centre of the vehicle axle). Triangles represent the measurements in TE1, circles the measurements in TE2.

i = 1,2, . . ., 10 for TE2). All time courses of the recorded Bolling probe pressure during wheeling were very similar in shape: narrow, high peaks with a steep increase and an equally steep decrease in pressure. The precise beginning and end of the passes were determined, separately for the front and rear tyre, as the points in time where the first derivative of BPPi(t) in time started and ceased to differ from zero, respectively. In the example given in Fig. 4 the pass of the front tyre started at t1 and ended at t2, while the pass of the rear tyre started at t3 and ended at t4. The additional soil pressure exerted by the combine harvester was calculated, separately for the front tyre and rear tyre (Fig. 4), as SPi;f ðtÞ ¼ BPPi ðtÞ  P¯ i;f ;

displaced relative to the front tyres and, consequently, also relative to the probes. Therefore the outer probes recorded much lower values than the other probes. Despite these variations there is a clear trend of increasing peak values with number of passes under both tyres. In TE2, the peak pressures increased by a factor of about 1.5–1.6 from the first to the last pass of the front tyre and even more (1.6–1.9) under the rear tyre. During TE1 no increase was found under the front tyre. Under the rear tyre the peak values of the second pass were between 1.1 and 2.9 times higher than those of the first pass. Fig. 6 shows how the baselines of the Bolling probe pressure, recorded directly before and after pass i,


for pass i of the front tyre ðt1  t  t2 Þ SPi;r ðtÞ ¼ BPPi ðtÞ  P¯ i;r ;


for pass i of the rear tyre ðt3  t  t4 Þ where BPPi(t) is the smoothed Bolling probe pressure during pass i, and P¯ i;f and P¯ i;r are the baselines of BPPi(t) for the two tyres (Fig. 4). Due to drifts in BPPi(t) between the passes of the front and rear tyres, P¯ i;f and P¯ i;r were usually different. Fig. 5 shows the changes in peak pressures, i.e. the maximum values of SPi,f(t) and SPi,r(t) with the number of passes. Soil pressure was generally higher in the centre line of the tracks than on the outer and inner sides of them during the passes of the front tyres. The values recorded during the passes of the rear tyres varied greatly, as the rear tyres were not only narrower, but also slightly

Fig. 6. Relative changes in the baselines of BPPi(t) directly before ðDP¯ i;f Þ and after ðDP¯ i;e Þ pass i, vs. number of passes during TE1 and TE2. +: increase; : decrease. Averaged values of the measurements of the centre (*, ~), inner ( , ) and outer probes (*, ~) (with regard to the centre of the vehicle axle). Triangles represent the measurements in TE1, circles the measurements in TE2.

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43


differed from the hydraulic pressure in the Bolling probes before the first passage (in %): DP¯ i;f ¼ ðP¯ i;f  P¯ 1;f Þ=P¯ 1;f  100%;


before pass i of the front tyre DP¯ i;e ¼ ðP¯ i;e  P¯ 1;f Þ=P¯ 1;f  100%;


after pass i of the rear tyre where P¯ 1;f , P¯ i;f and P¯ i;e are the baselines of BPPi(t) of pass 1 and i, respectively (Fig. 4). After a large initial increase of the baselines in TE1, they decreased in the probes of the centre line, but did not distinctly change in the other probes. In TE2, except for the initial increase in the inner probes, the baselines of BPPi(t) decreased steadily with the number of passes. After the 10 passes in TE2, they were 13–21% lower than at the beginning of the experiment.

3.4. Traffic-induced changes in soil physical properties Multivariate two-way analysis of variance revealed that the investigated soil physical and mechanical (see next chapter) properties were significantly altered by trafficking (Wilks’ Lambda, P < 0.005), but did not significantly vary with depth. There was no significant interaction between trafficking and depth. This means that compaction effects did not significantly differ between the two sampling depths. As already indicated by the ruts and the deformation of soil structure below the tracks, the 10 passes of the second experiment caused much stronger effects on bulk density and porosity than the two passes of the first experiment (Fig. 7 and Tables 4 and 5). Two passes reduced coarse porosity by 12–13% (P < 0.05), 10 passes even by up to almost 50% (P < 0.001) (Fig. 7 and Tables 4 and 5). A strong effect

Fig. 7. (a) Coarse porosity, (b) fine-to-intermediate porosity and (c) bulk density of non-trafficked (*) and trafficked locations after 2 ( , ) and 10 passes (*). Each line represents a separate sampling location (cf. Fig. 1). Grey triangles refer to samples taken after TE1 under the second track in the T1a locations.

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43


Table 4 Coarse porosity, fine-to-intermediate porosity, bulk density, precompression stress and compression index of the non-trafficked and trafficked sampling locations (2 and 10 passes) and differences in percentage between trafficked and non-trafficked sampling locations (relative to the values of the non-trafficked locations) Number of passes

Sampling locations

Fine earth fraction Coarse porosity (m3 m3)

Samples including gravel Fine-to-intermediate porosity (m3 m3)

Averaged values of the sampling locations (standard errors in parentheses) Topsoil 0 N1, N2 0.168 (0.007) 0.329 (0.004) 2 T1 0.146 (0.005) 0.347 (0.002) 10 T2 0.086 (0.008) 0.383 (0.003) Subsoil

0 2 10

N1, N2 T1 T2

0.158 (0.006) 0.139 (0.005) 0.098 (0.009)

0.330 (0.002) 0.339 (0.004) 0.371 (0.004)

Differences between trafficked and non-trafficked sampling locationsa (%) Topsoil 2 T1, N1, N2 13 5 10 T2, N1, N2 49 16 Subsoil a

2 10

T1, N1, N2 T2, N1, N2

12 38

3 12

Bulk density (g cm3)

Precompression stress (kPa)

Compression index

1.33 (0.01) 1.33 (0.01) 1.42 (0.01)

43 (3) 43 (3) 46 (3)

0.222 (0.009) 0.221 (0.006) 0.148 (0.006)

1.35 (0.01) 1.38 (0.02) 1.43 (0.01)

43 (2) 42 (3) 50 (4)

0.217 (0.008) 0.206 (0.010) 0.153 (0.005)

0 7

1 6

1 34

2 6

2 16

5 30

+: higher values in the trafficked sampling locations, : lower values in the trafficked sampling locations.

on the coarse pores was expected, as they are generally reported to be more sensitive to compaction than finer pores (Horn et al., 1995). The decrease in coarse porosity was partially compensated by an increase in the fine-to-intermediate porosity (Fig. 7 and Table 4). A similar conversion of coarse pores into fine-to-intermediate pores due to compaction has also been reported by Gupta et al. (1989), Alakukku (1996) and Richard et al. (2001). After the two passes in 2003, the fine-to-intermediate porosity in the trafficked locations was 3 and 5% higher than in the non-trafficked locations (not significant), and after the 10 passes in 2004, it was 12–16% higher in the trafficked than in the non-trafficked locations (P < 0.001, Table 5). As coarse porosity was decreased more than fine-tointermediate porosity increased, the net effect on bulk density was also an increase (6–7%) after 10 passes (P < 0.01) (Fig. 7 and Tables 4 and 5). Similar effects

on bulk density were reported e.g. by Gameda et al. (1984), Schjønning and Rasmussen (1994) and Arvidsson (2001). The effects of the two passes of TE1 on bulk density were weak. In the topsoil, no difference between trafficked and non-trafficked locations was found, while in the subsoil, bulk density was increased by 2%, but this increase was not significant (Tables 4 and 5). The additional samples taken after the two passes of TE1 in the T1a locations (cf. Fig. 1) showed the same effects as the other samples taken after TE1 (Fig. 7). 3.5. Traffic-induced changes in soil mechanical properties The 10 passes of the second experiment also had much more pronounced effects on the soil mechanical properties than the two passes of the first experiment (Figs. 8 and 9 and Tables 4 and 5).

Table 5 Statistical significances of traffic-induced changes in the soil properties (Fisher’s least significant difference post hoc analysis) Compared groups

Sampling locations

Fine earth fraction Coarse porosity

Fine-to-intermediate porosity

Bulk density

0 vs. 2 passes 0 vs. 10 passes 2 vs. 10 passes

T1, N1, N2 T2, N1, N2 T1, T2

P < 0.05 P < 0.001 P < 0.001

P < 0.001 P < 0.05

P < 0.01 P < 0.05

Blank fields indicate P > 0.05.

Samples including gravel Precompression stress

Compression index P < 0.001 P < 0.001


B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

Fig. 8. Mean compression curves in the topsoil (left) and in the subsoil (right) of non-trafficked (*) and trafficked locations after 2 ( ) and 10 passes (*).

Fig. 8 shows the averaged compression curves for the eight sampling locations of the two experiments. No distinct effects on the curves were observed after two passes. However, the 10-pass experiment strongly reduced the initial void ratio (i.e. the void ratio at 1 kPa normal stress of the compression curves) and changed the shape of the compression curves. The transition zone between elastic and plastic behaviour was slightly shifted to higher normal stresses in the compression curves of the samples from the 10-time trafficked locations, as compared to those from the non-trafficked locations. The VCLs of both the topsoil and the subsoil converged with increasing normal stress. This was particularly distinct in the subsoil, where all curves had very similar void ratios at 600 kPa normal stress. They seemed to coalesce into a single point at a void ratio of about 0.45 m3 m3. Similar convergence was also reported

by Veenhof and McBride (1996) and Culley and Larson (1987). The corresponding effects on precompression stress and CI are shown in Fig. 9. After two passes, the precompression stress was not changed; due to spatial variability of the test site it was even slightly lower in the trafficked than in the non-trafficked locations. Ten passes increased precompression stress by 6% in the topsoil and by 16% in the subsoil (Table 4). However, the changes were not significant (Table 5). Two passes decreased CI slightly by 1–5%, whereas 10 passes decreased CI highly significantly (P < 0.001, Table 5) by about one-third (Fig. 9 and Table 4). The additional samples taken from the T1a locations (cf. Fig. 1) again revealed the same effects as the samples taken in the T1 locations (Fig. 9). In summary, heavy trafficking (10 passes) led to a strong alteration in the soil pore space and the

Fig. 9. (a) Precompression stress and (b) compression index of non-trafficked (*) and trafficked locations after 2 ( , ) and 10 passes (*). Each line represents a separate sampling location (cf. Fig. 1). Grey triangles refer to samples taken after TE1 under the second track in the T1a locations.

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43


Fig. 10. Relationship between soil physical and mechanical parameters of all samples (n = 187) of non-trafficked (*) and trafficked locations after 2 ( ) and 10 passes (*). (a) Precompression stress vs. initial void ratio, (b) compression index vs. initial void ratio, and (c) precompression stress vs. compression index, and linear regressions of all samples (—).

mechanical properties, while after ‘‘light’’ trafficking (2 passes), the effects were much less pronounced, and mainly coarse porosity was reduced (Table 4). This highlights the value of coarse porosity as an indicator for soil compaction.

initial void ratio, or in other words, the soil became stiffer with increasing initial mass density. At high stresses the VCLs converged to approximately the same points (Fig. 8). 4. Discussion and conclusions

3.6. Relationships between mechanical properties and void ratio

4.1. Soil pressure during trafficking

Fig. 10 shows relationships between the mechanical properties and initial void ratio and the results of linear regression analysis. Here the initial void ratio refers to the total volume of the samples including gravel, as the latter has an important effect on the mechanical properties and cannot be removed without destroying the structure of the samples. The relationships between the mechanical properties and the initial void ratio were not affected by trafficking (Fig. 10). Precompression stress decreased (R2 = 0.21) and CI increased (R2 = 0.43) as void ratio increased. Accordingly precompression stress decreased with increasing CI. The relationship was weak though (R2 = 0.12). The positive correlation between CI and initial void ratio means that the VCL became steeper with increasing

The peak pressures recorded under the centreline of the track during the first pass by the front tyre (60 kPa in TE1 and 76 kPa in TE2) were similar to the pressure of about 54 kPa found by Gysi et al. (1999) during the pass of a sugar beet harvester of 132 kPa mean contact pressure under moist conditions. One reason for the progressive increase in peak pressure (Fig. 5) might have been the formation of the ruts, because this decreased the distance between the load (i.e. tyres) and the Bolling probes, so that increased pressures were recorded. This explanation was found to be in good agreement with calculated values using the stress propagation equation of Fro¨hlich (1934). In addition, also a modification in stress propagation with number of passes might have occurred. Horn et al.


B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

(1994), Semmel and Horn (1995) and Horn (2003) reported that repeated wheeling during a short time period at constant water content induced an increase in the ratio between vertical and horizontal stress, which is equivalent to an increase in the concentration factor. This effect may have been due to the increase in wetness, as the degree of water saturation increases with compaction and the concentration factor increases with the water content of a soil (So¨hne, 1953, 1958; Horn, 1983). The increase might also have been due to the structural deterioration below the ruts. Unstructured soils generally have a higher concentration factor than structured soils (Horn, 1988). Similar to our findings, Horn et al. (1994) found that the absolute magnitudes of the mean normal stress and the vertical stress increased with the number of passes, whereas Semmel and Horn (1995) reported that the absolute magnitudes decreased with the number of passes. The fact that the baseline of BPPi(t) returned to lower values after a pass than where it had been before (Fig. 6) means that the probes expanded to slightly larger volumes after the load was removed from pass to pass. It is possible that due to increased compactness with each subsequent pass the soil around the probes carried more of the weight of the overlying soil, so that the probes were slightly unloaded. An increase in water saturation during compaction would have caused a decrease in water suction. This may have decreased the resistance of the soil against expansion of the probes, allowing the probes to expand. 4.2. Traffic-induced changes in soil properties At the time of the traffic experiments in the fourth and fifth year after restoration, the precompression stress averaged 43 kPa on the test site (Table 4). Compared to literature values of 30–125 kPa for upper layers of restored soils of different ages (Schneider and Schro¨der, 1991; Lebert and Springob, 1994) and of 10– 250 kPa for upper layers of undisrupted agricultural soils (Lebert, 1989; Quasem et al., 2000), the soil at the considered site was still rather weak. The moderate effects on soil porosity and bulk density after two passes in our first experiment are similar to those that have been found in corresponding depths in agricultural soils of similar stability after a single pass by sugar beet harvesters (Gysi et al., 1999; Arvidsson, 2001). The two passes had similar effects on bulk density and stability, and distinctly stronger effects on soil porosity than several passes by tracked construction machinery with low ground contact pressures on agricultural soils of higher stability (Berli et al., 2004). The restored soil of

our study behaved similar to an undisturbed soil of low stability. It could withstand moderate loads, but was strongly compacted when heavily trafficked. To check if soil pressure exerted by the machine during the experiments could have exceeded the precompression stress of the test site, we calculated the stresses expected under the centre of the tyre, separately for the front and rear tyre, using the stress propagation equation given by Fro¨hlich (1934). As explained above, the tyre–soil contact area was assumed to be rectangular. Furthermore, we assumed that the vertical stress was evenly distributed over this area. The resulting average ground contact stresses are given in Table 2. We are aware that in contrast to our assumption, stress is usually unevenly distributed below tyres (e.g. Gysi et al., 2001) and maximum stresses were estimated to be as high as 1.5–2 times the mean contact stress (Koolen and Kuipers, 1983). However, as we had no information on the stress distribution over the tyre–soil contact areas of the combine harvester used in our experiments, we chose the assumption of even distribution for several reasons: (i) it was the least arbitrary choice, (ii) with increasing depth, propagated stresses depend less and less on deviations of the assumed from the actual contact stress distribution, and (iii) we could assume that if the calculated stress exceeded the precompression stress, then the latter would have been exceeded quite certainly also in the experiments. A concentration factor of 5 was chosen to account for the wet soil conditions. For the depth of 0.17 m, the calculated vertical stress was 118 kPa under the front tyre and 156 kPa under the rear tyre (including 3 kPa vertical stress of the overlying soil). At 0.32 m depth, the respective values were 99 and 83 kPa (including 5 kPa vertical stress of the overlying soil). The magnitude of these values is in reasonably good agreement with the peak values of the measured pressures at 0.32 m depth (Fig. 5). The calculated maximum stresses all clearly exceed the precompression stress found in the non-trafficked locations (Table 4). Under equilibrium conditions, thus strong soil compaction should have occurred, given that the soil was actually wetter in the traffic experiments (Table 1) and hence weaker than that in the confined uniaxial compression test, and that our calculations will have underestimated rather than overestimated the ‘‘true’’ vertical soil stresses. However, distinct compaction was observed only in the second experiment with 10 passes, and also in this experiment precompression stress increased only to 46–50 kPa (Table 4), which is considerably less than the maximum stress applied to the soil.

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

We believe that the main reason for this discrepancy is that conditions were far from equilibrium. In our experiment, trafficking stress lasted for 6–8 s under the front tyre (from t1 to t2 in Fig. 4) and for 5–6 s under the rear tyre (from t3 to t4) during a single pass. Soils exhibit higher strength and are less compacted if a load is applied for a short time only (Ghezzehei and Or, 2001; Fazekas and Horn, 2005). This explains why Horn and Hartge (1990) found that precompression stress increased less at high than at low wheeling speeds and always remained below the values of ground contact pressure. The hypothesis that equilibrium conditions were not reached is further supported by the cumulative nature of the observed compaction effects. Each pass produced some additional deformation, from which the soil did not recover until the next pass. In the course of 10 passes, these effects accumulated to a significant compaction effect. Cumulative effects due to repetitive trafficking have been found in various other studies, as reviewed by Ha˚kansson and Reeder (1994). Finally, also the progressive increase of peak stresses indicates that equilibrium conditions were not reached even after 10 passes. Thus, we conclude that further compaction would have occurred with additional passes. The discrepancy between the observed changes in precompression stress and the (conservatively) calculated stresses may in addition be due to the differences in the stress situation experienced by a sample volume of soil under a moving wheel in the field and in a confined uniaxial compression test in the laboratory, as pointed out already before (see Section 1). According to critical state soil mechanics, however, we should have expected the soil to yield under a bulk stress that was even lower than precompression stress if the samples were subjected to larger shear stresses in the field experiment than generated during oedometer tests (cf. Hettiaratchi, 1987; Kirby, 1989). An unexpected effect at first glance was the decrease in CI. According to the concept of precompression stress CI should not be affected by compaction, i.e. exceedance of the precompression stress. Indeed, various authors have shown experimentally in laboratory compression tests that CI remains unaffected if samples are subjected to subsequent cycles of loading and unloading at constant water content (e.g. Stone and Larson, 1980; Kirby, 1991a), and Kirby (1989) found that artificial soil samples, which had been reconstituted at the same water content, but to different initial void ratios, tended to have the same VCL (and thus CI) upon compression. Most of the field work on compaction by trafficking only looked into effects on precompression


stress, but not on CI (e.g. Semmel and Horn, 1995; Gysi et al., 1999; Berli et al., 2004). One exception is the study of Culley and Larson (1987). In accordance to our findings, these authors found that CI was lower in trafficked than in non-trafficked locations and that it decreased with increasing bulk density irrespective of the experimental trafficking. The effect may be at least partially due to the commonly adopted practice that samples are conditioned to a specific initial water tension for water-unsaturated compression tests (e.g. Horn, 1981; Culley and Larson, 1987) and not to a specific water content. If soil is compacted and then conditioned to the same water potential as samples of the uncompacted soil, as it is the case in field experiments like ours, this means that the compression tests are not run at the same water content for compacted and control samples, because water retention characteristics are changed by the compaction. Whether this fully explains the observed decrease in CI remains to be shown. The literature on the dependence of CI on water content is conflicting. Whereas Larson et al. (1980) and Kirby (1991b) did not find any changes in CI due to variation in water content, Leeson and Campbell (1983), Hettiaratchi (1987) and O’Sullivan and Robertson (1996) found that CI strongly depended on water content. In summary, our results show that loads may strongly exceed precompression stress for short durations without causing serious damage even on a rather weak soil with undeveloped structure, but that sub-critical compaction may rapidly accumulate to substantial effects if such loadings are repeated frequently. This means that the concept of precompression stress was not adequate to precisely predict the resistance of the restored soil investigated in this study against compaction under the applied loads and experimental conditions. Nonetheless, we believe that precompression stress provides a useful parameter on which to base limit values in regulations designed to prevent compaction of restored soil, because it was on the safe side with a large margin of security. Acknowledgments We are very grateful to the following persons for the permission to conduct the experiments on the test site, for the great technical support in the field and the information about the restoration and the subsequent cultivation operations of the experimental site (in alphabetical order): Franz Borer, Dieter Fux, Martin Keller, Christian Ledermann, Werner Rohr, Peter Schuler, Gaby von Rohr, Urs Zuber. The work was


B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43

partially funded by the Swiss Agency for the Environment, Forests and Landscape (BUWAL/SAEFL) and by the Environment Protection Agency of the Canton of Solothurn, Switzerland. References Abu-Hamdeh, N.H., Reeder, R.C., 2003. Measuring and predicting stress distribution under tractive devices in undisturbed soils. Biosyst. Eng. 85, 493–502. AG-Boden, 1982. Bodenkundliche Kartieranleitung. 3rd ed. Schweizerbart’sche Verlagsbuchhandlung (Na¨gele und Obermiller), Stuttgart. Alakukku, L., 1996. Persistence of soil compaction due to high axle load traffic. I. Short-term effects on the properties of clay and organic soils. Soil Till. Res. 37, 211–222. Arvidsson, J., 2001. Subsoil compaction caused by sugarbeet harvesters in southern Sweden. I. Soil physical properties and crop yield in six field experiments. Soil Till. Res. 60, 67–78. Arvidsson, J., Keller, T., 2004. Soil precompression stress. I. A survey of Swedish arable soils. Soil Till. Res. 77, 85–95. Arvidsson, J., Trautner, A., van den Akker, J.J.H., Schjønning, P., 2001. Subsoil compaction caused by heavy sugarbeet harvesters in southern Sweden. II. Soil displacement during wheeling and model computations of compaction. Soil Till. Res. 60, 79–89. Baumgartl, T., Horn, R., 1991. Effect of aggregate stability on soil compaction. Soil Till. Res. 19, 203–213. Berli, M., Kulli, B., Attinger, W., Keller, M., Leuenberger, J., Flu¨hler, H., Springman, S.M., Schulin, R., 2004. Compaction of agricultural and forest subsoils by tracked heavy construction machinery. Soil Till. Res. 75, 37–52. Blake, G.R., Hartge, K.H., 1986. Bulk density. In: Klute, A. (Ed.), Methods of Soil Analysis. Part 1. Physical and Mineralogical Methods. 2nd ed. Am. Soc. Agr., Soil Sci. Soc. Am., Madison, Wisconsin, pp. 363–375. Bolling, I., 1987. Bodenverdichtung und Triebkraftverhalten bei Reifen—Neue Mess- und Rechenmethoden. Ph.D. Thesis. Forschungsbericht Agrartechnik des Arbeitskreises Forschung und Lehre der Max-Eyth-Gesellschaft (MEG), Technische Universita¨t Mu¨nchen, Mu¨nchen. Bullock, P., Newman, A.C.D., Thomasson, A.J., 1985. Porosity aspects of the regeneration of soil structure after compaction. Soil Till. Res. 5, 325–341. BUWAL, 2001. Bodenschutz beim Bauen, Leitfaden Umwelt, vol. 10. Bundesamt fu¨r Umwelt, Wald und Landschaft (BUWAL/SAEFL), Bern. Casagrande, A., 1936. The determination of the pre-consolidation load and its practical significance. In: Proceedings of the International Conference on Soil Mechanics and Foundation Engineering, vol. III. Harvard University, Cambridge, pp. 60–64. Culley, J.L.B., Larson, W.E., 1987. Susceptibility to compression of a clay loam Haplaquoll. Soil Sci. Soc. Am. J. 51, 562–567. Davies, R., Younger, A., 1994. The effect of different post-restoration cropping regimes on some physical properties of a restored soil. Soil Use Manage. 10, 55–60. Dexter, A.R., Horn, R., Kemper, W.D., 1988. Two mechanisms for age-hardening of soil. J. Soil Sci. 39, 163–175. Fazekas, O., Horn, R., 2005. Zusammenhang zwischen hydraulischer und mechanischer Bodenstabilita¨t in Abha¨ngigkeit von der Belastungsdauer. Z. Pflanzenerna¨hr. Bodenk. (J. Plant Nutr. Soil Sci.) 168, 60–67.

Fro¨hlich, O.K., 1934. Druckverteilung im Baugrunde mit besonderer Beru¨cksichtigung der plastischen Erscheinungen. Verlag von Julius Springer, Wien. FSK, 2001. Schweizerischer Fachverband fu¨r Sand und Kies (FSK), FSK-Rekultivierungsrichtlinie, Richtlinie fu¨r den fachgerechten Umgang mit Bo¨den. Haller und Jenzer AG, Burgdorf. Gameda, S., Raghavan, G.S.V., McKyes, E., Theriault, R., 1984. Soil and crop response to high axle load subsoil compaction—recovery and accumulation. Am. Soc. Agric. Engineers, St. Joseph, Michigan, Paper No. 84-1542, 24 pp. Ghezzehei, T.A., Or, D., 2001. Rheological properties of wet soils and clays under steady and oscillatory stresses. Soil Sci. Soc. Am. J. 65, 624–637. Gupta, S.C., Sharma, P.P., DeFranchi, S.A., 1989. Compaction effects on soil structure. Adv. Agron. 42, 311–338. Gysi, M., Ott, A., Flu¨hler, H., 1999. Influence of single passes with high wheel load on a structured, unploughed sandy loam soil. Soil Till. Res. 52, 141–151. Gysi, M., Maeder, V., Weisskopf, P., 2001. Pressure distribution underneath tires of agricultural vehicles. Trans. ASAE 44, 1385–1389. Ha˚kansson, I., Reeder, R.C., 1994. Subsoil compaction by vehicles with high axle load—extent, persistence and crop response. Soil Till. Res. 29, 277–304. Hettiaratchi, D.R.P., 1987. A critical state soil mechanics model for agricultural soils. Soil Use Manage. 3, 94–105. Horn, R., 1981. Eine Methode zur Ermittlung der Druckbelastung von Bo¨den anhand von Drucksetzungsversuchen. Zeitschrift fu¨r Kulturtechnik und Flurbereinigung 22, 20–26. Horn, R., 1983. Die Bedeutung der Aggregierung fu¨r die Druckfortpflanzung im Boden. Zeitschrift fu¨r Kulturtechnik und Flurbereinigung 24, 238–243. Horn, R., 1986. Auswirkung unterschiedlicher Bodenbearbeitung auf die mechanische Belastbarkeit von Ackerbo¨den. Z. Pflanzenerna¨hr. Bodenk. (J. Plant Nutr. Soil Sci.) 149, 9–18. Horn, R., 1988. Compressibility of arable land. In: Drescher, J., Horn, R., De Boodt, M. (Eds.), Impact of Water and External Forces on Soil Structure. Catena Supplement 11. Catena, CremlingenDestedt, pp. 53–71. Horn, R., 2003. Stress–strain effects in structured unsaturated soils on coupled mechanical and hydraulic processes. Geoderma 116, 77– 88. Horn, R., Hartge, K.H., 1990. Effects of short-time loading on soil deformation and strength of an ameliorated Typic Paleustalf. Soil Till. Res. 15, 247–256. Horn, R., Werner, D., Baumgartl, T., Winterot, C., 1994. Wirkungen technogener Druckbelastung auf die Spannungsverteilung und das Bodengefu¨ge einer Schwarzerde aus Lo¨ss. Z. Pflanzenerna¨hr. Bodenk. (J. Plant Nutr. Soil Sci.) 157, 433–440. Horn, R., Domz˙ał, H., Słowin´ska-Jurkiewicz, A., van Ouwerkerk, C., 1995. Soil compaction processes and their effects on the structure of arable soils and the environment. Soil Till. Res. 35, 23–36. Keller, T., Trautner, A., Arvidsson, J., 2002. Stress distribution and soil displacement under a rubber-tracked and a wheeled tractor during ploughing, both on-land and within furrows. Soil Till. Res. 68, 39–47. Keller, T., Arvidsson, J., Dawidowski, J.B., Koolen, A.J., 2004. Soil precompression stress. II. A comparison of different compaction tests and stress–displacement behaviour of the soil during wheeling. Soil Till. Res. 77, 97–108. Kirby, J.M., 1989. Measurements of the yield surfaces and critical state of some unsaturated agricultural soils. J. Soil Sci. 40, 167–182.

B. Scha¨ffer et al. / Soil & Tillage Research 93 (2007) 28–43 Kirby, J.M., 1991a. Strength and deformation of agricultural soil: measurement and practical significance. Soil Use Manage. 7, 223– 229. Kirby, J.M., 1991b. Critical-state soil mechanics parameters and their variation for Vertisols in eastern Australia. J. Soil Sci. 42, 487– 499. Koolen, A.J., 1974. A method for soil compactibility determination. J. Agric. Eng. Res. 19, 271–278. Koolen, A.J., Kuipers, H., 1983. Agricultural soil mechanics. Advanced Series in Agricultural Sciences, vol. 13. Springer, Berlin. Larson, W.E., Gupta, S.C., Useche, R.A., 1980. Compression of agricultural soils from eight soil orders. Soil Sci. Soc. Am. J. 44, 450–457. Lebert, M., 1989. Beurteilung und Vorhersage der mechanischen Belastbarkeit von Ackerbo¨den. Ph.D. Thesis. Bayreuther Bodenkundliche Berichte, vol. 12, Universita¨t Bayreuth, Bayreuth. Lebert, M., 1991. Stabilita¨tseigenschaften rekultivierter Bo¨den aus Lo¨ss im Rheinischen Braunkohlen-Tagebaurevier. Mitt. Dtsch. Bodenkd. Ges. 66, 165–168. Lebert, M., Horn, R., 1991. A method to predict the mechanical strength of agricultural soils. Soil Till. Res. 19, 275–286. Lebert, M., Springob, G., 1994. Zur Rekultivierung von Lo¨ssbo¨den (I): Wirkungen verschiedener Planierverfahren auf die Bodenstruktur. Z. f. Kulturtechnik und Landentwicklung 35, 112–121. Leeson, J.J., Campbell, D.J., 1983. The variation of soil critical state parameters with water content and its relevance to the compaction of two agricultural soils. J. Soil Sci. 34, 33–44. Logsdon, S.D., Allmaras, R.R., Nelson, W.W., Voorhees, W.B., 1992. Persistence of subsoil compaction from heavy axle loads. Soil Till. Res. 23, 95–110. Muhs, H., Kany, M., 1954. Einfluss von Fehlerquellen bei Kompressionsversuchen. In: Grundbau. Teil 1. Fortschritte und Forschungen im Bauwesen, vol. 17. Franckh’sche Verlagshandlung, Stuttgart, pp. 125–152. O’Sullivan, M.F., Robertson, E.A.G., 1996. Critical state parameters from intact samples of two agricultural topsoils. Soil Till. Res. 39, 161–173. Porter, O.J., 1950. Development of the original method for highway design. Trans. ASCE 115, 461–467. Quasem, H., Attinger, W., Berli, M., Schulin, R., 2000. Zusammenha¨nge zwischen o¨dometrischer Verdichtungsempfindlichkeit und morphologischer Merkmale von Bo¨den, untersucht an ausgewa¨hlten Standorten auf den Trassee der Transitgasleitungen Da¨niken— Grimsel (TRG 2) und Rodersdorf—Lostorf (TRG 3). Schlussbericht. Institute of Terrestrial Ecology, Swiss Federal Institute of Technology Zu¨rich (ETHZ), Zu¨rich. Richard, G., Cousin, I., Sillon, J.F., Bruand, A., Gue´rif, J., 2001. Effect of compaction on the porosity of a silty soil: influence on unsaturated hydraulic properties. Eur. J. Soil Sci. 52, 49–58. Rohr, W., 2002. Vorgaben fu¨r die Folgebewirtschaftung von Fruchtfolgefla¨chen beim Autobahnbau N5, Kanton Solothurn. GEOTEST AG, Zollikofen, Switzerland. Rohr, W., Schuler, P., 2004. Tempora¨r beanspruchte Fla¨chen, Kulturerdearbeiten und Bodenschutz beim Autobahnbau N5, Kanton


Solothurn, Schlussbericht. GEOTEST AG, Zollikofen, Switzerland. Schjønning, P., Rasmussen, K.J., 1994. Danish experiments on subsoil compaction by vehicles with high axle load. Soil Till. Res. 29, 215–227. Schneider, R., Schro¨der, D., 1991. Bodenmechanische Untersuchungen an Neulandbo¨den aus Lo¨ss und Hafenschlick. Mitt. Dtsch. Bodenkd. Ges. 66, 221–224. Schro¨der, D., Stephan, S., Schulte-Karring, H., 1985. Eigenschaften, Entwicklung und Wert rekultivierter Bo¨den aus Lo¨ss im Gebiet des Rheinischen Braunkohlen-Tagebaues. Z. Pflanzenerna¨hr. Bodenk. (J. Plant Nutr. Soil Sci.) 148, 131–146. Semmel, H., Horn ,R., 1995. Mo¨glichkeiten zur Bestimmung der mechanischen Belastbarkeit und der Druckfortpflanzung im Boden im Hinblick auf die Ableitung von bodentyp- und maschinenspezifischen Grenzwerten. In: Bodenverdichtung, Kuratorium fu¨r Technik und Bauwesen in der Landwirtschaft (KTBL), vol. 362. Landwirtschaftsverlag GmbH, Mu¨nster-Hiltrup, Germany, pp. 61–92. Soane, B.D., van Ouwerkerk, C., 1995. Implications of soil compaction in crop production for the quality of the environment. Soil Till. Res. 35, 5–22. So¨hne, W., 1953. Druckverteilung im Boden und Bodenverformung unter Schlepperreifen. Grundlagen der Landtechnik 5, 49– 63. So¨hne, W., 1958. Fundamentals of pressure distribution and soil compaction under tractor tires. Agric. Eng. 39, 276–291. Stone, J.A., Larson, W.E., 1980. Rebound of five one-dimensionally compressed unsaturated granular soils. Soil Sci. Soc. Am. J. 44, 819–822. Turnbull, W.J., 1950. Appraisal of the CBR method. Trans. ASCE 115, 547–554. USDA, 1997. Keys to Soil Taxonomy by the Soil Survey Staff, Soil Cons. Serv., U.S. Dept. Agr., 7th ed. Pocahontas Press, Blacksburg, Virginia. Veenhof, D.W., McBride, R.A., 1996. Overconsolidation in agricultural soils: I. Compression and consolidation behavior of remolded and structured soils. Soil Sci. Soc. Am. J. 60, 362–373. Von Albertini, N., Leuenberger, J., La¨ser, H.P., Flu¨hler, H., 1995. Regeneration der Bodenstruktur eines verdichteten Ackerbodens unter Kunstwiese. In: Bodenkundliche Gesellschaft der Schweiz, Aktuelle Bodenforschung in der Schweiz, Dokument 7. Juris Druck und Verlag, Dietikon, pp. 10–16. Voorhees, W.B., 1983. Relative effectiveness of tillage and natural forces in alleviating wheel-induced soil compaction. Soil Sci. Soc. Am. J. 47, 129–133. VSS, 2000. Schweizer Norm SN 640 583. Erdbau, Boden. Eingriff in den Boden, Zwischenlagerung, Schutzmassnahmen, Wiederherstellung und Abnahme. Vereinung Schweizerischer Strassenfachleute (VSS), Zu¨rich. Zwo¨lfer, F., Geiß, M., Adam, P., Heinrichsmeier, K., Herrmann, H., 1991. Erhaltung fruchtbaren und kulturfa¨higen Bodens bei Fla¨cheninanspruchnahmen. Luft, Boden, Abfall, vol. 10. Ministerium fu¨r Umwelt Baden-Wu¨rttemberg, Stuttgart.