Traffic-induced soil compaction during manure spreading in spring in South-East Norway

Abstract

The objective of this study was to evaluate long-term effects of two tillage regimes (ploughing and minimum tillage) on the bearing capacity of a clay rich soil, by using two different slurry tankers (4.1 and 6.6 Mg wheel load) and contrasting wheeling frequencies (1 and 10 passes). The soil strength was assessed by laboratory measurements of the precompression stress (Pc) at −6 kPa in topsoil (20 cm) and subsoil (40 and 60 cm) samples. Stress propagation, elastic and plastic deformation during wheeling were measured in the field with combined stress-state-transducer and displacement transducer system. Results presented in this study show that minimum tilled soil had 74% higher Pc than ploughed soil in the upper soil layer, whilst differences were less distinct in subsoil. Wheeling increased Pc at all soil depths. Compared to ploughing, higher strength in the upper layer of minimum tilled soil led on average to 60% and 48% reductions in the major principal stress with the use of the light and heavy slurry tanker, respectively. The extent of the major principal stress was dependent on the ground pressure in the topsoil. The first pass of a wheel caused the greatest damage in some cases, but all wheelings led to accumulative plastic deformation in both vertical and horizontal directions. Wheeling with high intensity would have exceeded Pc in all cases when soil was at a matric potential of −6 kPa. The results show that soil water content is an important factor influencing bearing capacity. Drier soil (−100 kPa), in combination with minimum tillage, limited the occurrence of stresses exceeding Pc in the upper soil layer.

Keywords:

• long-term soil tillage
• soil inner strength
• subsoil
• major principal stress
• wheel load
• wheeling intensity

Introduction

Increasing costs of production factors (labour, fuel, fertilizer) lead to increasing economic pressure in modern farming. In the attempt to increase productivity and achieve more economical crop production, there is a continuing upward trend in the power and weight of farm machinery even on smaller farms (Soane et al. 1982; Flowers & Lal 1998). In Europe, both the weight of agricultural machinery and the number of wheelings during tillage, fertilizing, spraying and harvesting have increased at least threefold during the last 3–4 decades (Horn et al. 1995), with the overall wheeled area usually being several times as great as the field area (Håkansson & Reeder 1994).

The use of heavy farm machinery under unfavourable conditions, less rotation of crops and inappropriate soil management (e.g. working depth, tyre air pressure) increase the risk for soil compaction (Alakukku et al. 2003; Hamza & Anderson 2005) and reduce the productivity of arable land (Ehlers et al. 2000). Similar problems have been reported from all continents (Hamza & Anderson 2005), making soil compaction one of the most important factors in soil physical degradation (Pagliai et al. 2003), affecting about 68 million hectares worldwide, of which more than half is located in Europe. Most of this soil compaction is caused by traffic on agricultural land (Flowers & Lal 1998).

The use of animal slurry is important due to increasing costs of mineral fertilizer, and the aim of maintaining soil organic matter levels in the soil are leading to a growing interest in the use of animal slurry. However, vehicles used for slurry spreading are among the heaviest farm equipment, with wheel loads often exceeding 6 Mg. The latest machine under testing in Central Europe has a total weight in excess 60 Mg. Although there are none of the latter machines in Norway, due to small farm size, there is concern about the size and weight of machinery in use and larger machinery is becoming more common as the use of contract machinery operations increases. Large machinery is often in conflict with small field size in Norway in terms of the working width of machinery. This can result in inefficient traffic patterns on fields. Furthermore liquid manure is often transported long distances by road between farm and field, resulting in a conflict of adapting tyre size and inflation pressures to use on arable soil. In order to prevent nutrient leaching, and to satisfy the nutritional need of plants, slurry is often applied in spring when the subsoil is normally rather wet. In combination with high machinery weight, the risk for soil compaction is therefore severe. Because of the short growing season in northern latitudes, soils remain at field capacity for longer in spring and the return to field capacity in autumn occurs often earlier than in other regions. Field operations are therefore often performed under suboptimal circumstances, leading to a high risk of damaging soil structure due to low water tension and low effective soil strength (Soane et al. 1982).

Calculations based on Swedish experiments suggest that the costs of soil compaction due to manure spreading under wet circumstances may exceed the value of the nutrients supplied in the manure (Håkansson 1994). Such costs were mainly associated with compaction of the upper soil layer. Because of its persistence, subsoil compaction has a negative long-term effect on the production capacity of a soil. The cumulative effect of multiple wheeling may reduce the food supply for future generations, and this long-term threat to productivity implies that subsoil compaction also has an ethical aspect (Håkansson et al. 1987; Håkansson 1994).

The main object of this paper is to describe how slurry tankers of different weight (4.1 and 6.6 Mg wheel load) and contrasting wheeling frequency (1 and 10 passes) influence stress propagation in soil. The use of such heavy machinery has not been investigated under the conditions in south-eastern Norway, where the climate is characterized by long, cold winters and relatively short growing seasons with variable rainfall. The methods used to measure the effects of compaction include measurement of the precompression stress (Pc) to determine soil strength. In addition, a combined stress-state and displacement-stress transducer system (SST, DTS) was used to determine the major principal stresses and soil deformation in topsoil and subsoil that may occur during slurry spreading. Such technology has not been used in Norway before. Studies by Wiermann et al. (2000) and Zink et al. (2010) show that topsoils under reduced soil tillage may have a higher stability throughout the soil profile than soils under conventional tillage. Our study was therefore performed at a site where plots were available on which contrasting tillage operations had been practised for ten years.

Material and methods

Field site

Field measurements and sampling took place in spring 2011 at Øsaker, Sarpsborg (59°23′N and 11°02′E, 40 m above sea level), in south-eastern Norway. The soil is a Stagnosol derived from postglacial marine deposits, characterised as clay loam with 33% clay and 44% silt in the topsoil, overlying clay (46%). The topsoil contains around 4.3% humus (Table 1). Saturated porosity in the topsoil varies between 46% (minimum tillage) and 52% (ploughed plot), and the bulk density in the topsoil was approximately 1.4 g/cm³ on minimum tilled plot and 1.2 g/cm³ on ploughed plot (Table 2) The site was drained to about 80 cm depth with 8 m spacing.

Table 1. Soil properties of the reference profiles.

	Horizon	Depth (cm)	C org (%)	Sand	Silt (g/kg)	Clay
P	Ap	0–35	2.8	230	430	340
	Cl	35–80	1.0	230	450	320
	II Cl	80–150	0.7	100	440	460
M	rAp	0–30	2.2	240	440	320
	Cl	30–50	0.7	200	460	340
	II Cl	50–120	0.4	100	440	460

P, ploughing, M, minimum tillage, C org, organic carbon %.

Table 2. Bulk density (pB), total porosity and pore size distribution on ploughed (P) and minimum tilled (M) reference plots.

				Macropores, % (v/v)	Mesopores, % (v/v)		Fine pores, % (v/v)
	Depth (cm)	pB (g cm^‒3)	Total porosity	>30 μm	3–30 μm	0.2–3 μm	<0.2 μm
P	20	1.24	52	15	5	7	25
	40	1.38	46	7	4	5	30
	60	1.48	45	5	3	5	31
M	20	1.36	46	11	3	9	22
	40	1.48	46	6	2	5	32
	60	1.42	48	5	3	8	32

The measurements were performed on a soil tillage trial that was established in 2000 with oats (Avena sativa L.) and (winter) wheat (Triticum aestivum L.) grown each year in rotation. This trial included four tillage treatments of which the two extremes were used here: (P) autumn ploughing to both crops and (M) minimum tillage (direct drilling to wheat/shallow harrowing to oats). These treatments were performed up until the spring of 2010, thus giving the basis for investigating long-term tillage effects of contrasting tillage regimes on soil structure. No tillage was performed in autumn 2010 or in spring 2011 prior to the present soil investigations. Due to limited resources, there was no replication of the electronic measurements (SST and DTS). Soil samples were taken from one reference soil pit on both minimum tilled and ploughed plots at 20, 40 and 60 cm depth prior to wheeling (reference profile). After wheeling, soil samples were taken on one pit for each of the eight treatments compared (Figure 1).

Figure 1. Field layout: treatments: P = autumn ploughing, M = minimum tillage (direct drilling to winter wheat, shallow harrowing to spring oats). Black boxes mark sampling pits. The reference pits were situated close by.

Climate and soil water content at sampling

Whilst average rainfall in April is about 42 mm and average temperature is about 4.6°C for this region, it was both drier (no rainfall) and warmer (average 9.0°C) during the three weeks prior to soil sampling (Øsaker weather station). This resulted in drier conditions at the time of soil sampling, around 85% of assumed field capacity for Norwegian clay loam soils (−10 kPa) in the topsoil and slightly moister in deeper soil layers (88%; Table 3). This gave very trafficable field conditions while wheeling. All wheeling was done at similar water content.

Table 3. Soil moisture contents at sampling in 2011 expressed as percentages of field capacity measured at –10 kPa.

			20 cm		40 cm		60 cm
Profile	Treatment	Wheeling	%	SE	%	SE	%	SE
R	P		86	2	91	1	90	<1
	M		73	7	88	<1	87	<1
P	4.1 Mg	1×	93	6	89	1	88	1
		10×	87	1	89	<1	91	1
P	6.6 Mg	1×	81	3	83	3	84	3
		10×	83	3	78	6	88	1
M	4.1 Mg	1×	88	2	86	2	92	1
		10×	85	1	90	1	94	<1
M	6.6 Mg	1×	85	4	90	1	86	1
		10×	84	2	88	1	84	2

R, reference, P, ploughing, M, minimum tillage, SE, standard error.

Machinery

In both cases, single (1×) and multiple (10×) passes with the two tractor slurry tank combinations were performed (Figure 2). The smaller tractor/slurry tank combination had a total weight of 16 Mg, resulting in a wheel load of 4.1 Mg for the trailer (single axle), and is typical of the equipment commonly used today on small- and medium-sized farms in Norway. The larger tractor slurry tank combination had a laden weight of 36 Mg, resulting in a wheel load of 6.6 Mg for the trailer (tandem axles; Table 4), as is common in combination with the use of contract machinery operations in Norway.

Figure 2. Left: small slurry tanker, wheel load 4.1 Mg, Right: large slurry tanker, wheel load 6.6 Mg.

Table 4. Tyre dimension and inflation pressure for all wheels, contact area and ground pressure of trailer wheel (first axle) used in the compaction study.

			Contact area (cm^2)		Average ground pressure (kPa)
Wheel load	Tyre dimension	Inflation pressure (kPa)	M	P	M	P
4.1 Mg	Front: 650/65R28	80
	Rear: 650/65R38	80
	Trailer: 23/1 26	180	3143	3090	128	130
6.6 Mg	Front: 600/70R30	130
	Rear: 710/70R42	130
	Trailer: 800/45 30.5	180/250	5350	8402	121	77

M, minimum tillage, P, ploughing.

Both trailers were fully loaded, with tyre inflation pressures and driving speeds of 5 km/h, according to factory recommendations. The machinery was weighed prior to the field trial. Wheel load was measured on a portable scale, and the contact area of the wheels was determined by marking the tyre-print with flour. The latter was photographed from above, and the image was processed digitally (Gysi et al. 1999; Zink et al. 2010). To determine the average ground pressure, the total load was divided by the surface area.

Soil measurements

The precompression stress (Pc)

Pc is defined as the maximum major principal stress that a soil horizon can withstand against applied external forces without incurring irreversible damage (Horn & Fleige 2009). According to Horn and Fleige (2003), Pc can be classified as low (30–60 kPa), medium (60–90 kPa) and high (90–120 kPa). Pc is regarded as the stress limit (threshold value) at which soil deformation changes from elastic and reversible to plastic and irreversible (Peth et al. 2010). As long as the applied stress σ1 does not exceed Pc, the internal soil strength is sufficient to withstand the induced stress. In soils that are expected to deform and recover elastically, important soil physical properties, such as air permeability and saturated hydraulic conductivity, are not expected to be affected (Horn & Fleige 2003). The ratio between Pc and major principal stress can be classified as weak, additional plastic deformation (<0.8), labile (0.8–1.2), stable (1.2–1.5) and as very stable, elastic deformation (>1.5; Horn & Fleige 2003).

Compaction that exceeds the internal soil strength affects soil properties such as porosity. Macropores (>30 μm) are easily deformed due to their large size, leading to reduction of air capacity and air conductivity. Deformation of such pores often increases the amount of mesopores (0.2–30 μm) and plant-available water (Horn & Fleige 2003). Plastic, irreversible soil deformation increases strongly if the external soil stress applied to the soil exceeds the precompression (Wiermann et al. 2000). If the applied soil stress exceeds the Pc, soil structure may be damaged in two different ways. Firstly, the coarse inter-aggregate pores are destroyed, whilst the intra-aggregate pore system remains intact. Secondly, higher weight and continued wheeling leads to destruction also of the aggregates, resulting in complete homogenization (Wiermann et al. 2000). The latter changes are considered to be almost irreversible, at least at depth. Soil stress exceeding the Pc leads thus to increased soil strength. As not exceeding the Pc is the best protection against soil compaction, Pc may be used as a precautionary value in the avoidance of compaction (Horn & Fleige 2009). It can also be used as an indicator of soil structural stability. To characterize the compaction susceptibility of a soil, Pc can be compared with the applied stress (σ1; Zink et al. 2010). The ratio Pc/σ1 has been classified by Horn and Fleige (2003) in relation to the expected type of deformation.

The Pc may vary under the same soil use and identical climatic conditions, due to differences in soil structure, texture, matric potential and organic matter content (Horn et al. 1995). Pc values generally differ between topsoil and subsoil, with subsoil having higher values due to the overlaying weight of topsoil. Furthermore, values differ between soil tillage regimes. In this study, measurements in the field showed that the soil at sampling was drier (Table 3) than that used for laboratory Pc analyses (−6 kPa). Therefore, Pc was calculated for the different plots according to Terzaghi and Jelinek (1954) to get an estimate of Pc for drier soil (−100 kPa). Pc is a good parameter to illustrate changes in soil stability and to classify stress impacts. However, it should be accompanied by other measurements in order to classify its ecological relevance. Stress and shear strength data were measured by compression and frame shear tests on undisturbed soil core samples (volume 236 cm³) that were equilibrated at −6 kPa matric potential, which is the standard moisture content used for measuring Pc (Zink et al. 2010), often occurring in early spring and late autumn (Lebert & Horn 1991). Stress strain measurements to determine Pc were made using a pneumatic multistep oedometer (PMO). Compression curves were derived by successively applying loads of 10, 50, 100, 200 and 400 kPa for 10 hours each. These intervals were necessary to guarantee 95% compression of the samples. Eight samples per horizon (20, 40, 60 cm depth) were tested for the various treatments. The PMO consisted of 8 independent soil testing units that each comprised of a double-acting pneumatic cylinder and a proportional pressure regulator to apply the load (Peth et al. 2010). The applied load was measured and controlled according to the prescribed load. During the static loading test, ceramic microtensiometers were inserted into the bottom of the sample to measure matric suction (Peth & Horn 2006).

Soil settlement during testing was recorded at the top of the sample using potentiometric displacement sensors. The oedometer system was controlled by the LabView based computer software Pilatos (Peth et al. 2010). This gave the possibility to control different samples independently in each of the testing units. The Pc values were determined graphically according to Casagrande (1936), by drawing a tangent to the stress–strain curve at the point with greatest curvature (Arvidsson & Keller 2004). Before and after the static loading and multistep compression test, air permeability (k1) was measured by using a series of flow meters at a pressure head of 0.1 kPa. This value was chosen to avoid turbulent airflow (Peth et al. 2010). In order to determine soil properties and soil physical parameters (bulk density, total porosity and pore size distribution), unloaded reference plots and loaded wheel tracks were sampled (n = 6) by taking undisturbed cylinder core samples (100 cm³) at 20, 40 and 60 cm depth.

Stress-state and displacement stress transducer systems

In order to determine influence of the various wheel loads and wheeling intensities on soil structure, stress propagation during the wheeling was measured with a stress-state-transducer (SST) system (Figures 3 and 4). The sensor head used in this system had the form of a quarter sphere (diameter ca. 65 mm) and consisted of six pressure cells located at an angle of 54.73° to each other within a hollow aluminium body (Nichols et al. 1987). This allowed the stress state at one point to be calculated. The stress state is characterized by the mean normal stress (MNS) and the octahedral shear stress (OCTSS) at the peak of the major principal stress (Wiermann et al. 2000). In this publication, we only use the major principal stress (σ1); thus, data for MNS and OCTSS are not shown. Further details about stress theory and the mathematics behind the development and function of the transducer can be found in Nichols et al. (1987) and Zink et al. (2010). Soil compaction is caused by a combination of static forces and dynamic stresses that are caused by the weight and vibration of implements and/or the tractor engine and by wheel-slip of tractors and machinery (Horn et al. 1995). This results in both vertical and horizontal displacement of soil (Figure 3). The SST was connected to a displacement transducer system (DTS; Wiermann et al. 2000) which was located at 20 cm depth, thus measuring the amount of elastic and plastic displacement in both vertical and horizontal directions in the upper soil layer above 20 cm.

Figure 3. DTS: Sensor at 20 cm depth registers elastic and plastic deformation in vertical and horizontal directions in soil during each wheeling (modified after Zink 2009).

Figure 4. Experimental setup of the stress/strain measurement using the combined stress-state/displacement-stress transducer system (SST/DTS) (modified after Zink 2009).

In order to instal the measuring system, 1 × 1 m access trenches were dug parallel to the direction of wheeling. In our trial, sensors were installed successively on each side of the trenches used, thus allowing several treatments to be compared in each trench (see details of placement in Figures 1 and 4). The sensors were placed at 20, 40 and 60 cm depth beneath the centre of the wheel rut. The distance between sensor head and profile wall was about 50 cm. To avoid disturbing the original soil structure during installation, metal tubes were driven into the soil from each side of the trench, creating an access hole for inserting the measuring device. To ensure that the sensor head was fully in contact with the soil across its entire surface, the channel was made with a specially sharpened soil auger (Zink et al. 2010). To avoid collapse of the trench and damage to the sensor heads during wheeling, the side-walls of the trenches were reinforced with metal plates supported by mechanical jacks and wooden planks (Figure 4) and channels were refilled with the removed material and recompacted close to their original density. The sensor head consists of steel which has a different stiffness in contrast to the surrounding soil. In case of a low density of the soil (ex. ploughed topsoil), contact problems between the sensor head and the surrounding soil matrix may occur, which could cause a concentration of stress on the sensor head (Gräsle 1999). This problem declines with increasing topsoil density of the top soil (ex. due to compaction after wheeling). Therefore, all results for the first wheeling were omitted (Table 5) and results are presented as average values for all wheelings (Figure 5).

Table 5. Major principal stress (σ1) in kPa as registered by SST system.

	Ploughing						Minimum tillage
	4.1 Mg			6.6 Mg			4.1 Mg			6.6 Mg
Wheeling	20 cm	40 cm	60 cm	20 cm	40 cm	60 cm	20 cm	40 cm	60 cm	20 cm	40 cm	60 cm
1	[355]	[381]	md	[254]	[286]	[110]	[155]	[266]	[144]	[195]	[490]	[114]
2	417	369	md	342	291	122	133	[214]	145	202	[253]	77
3	424	339	md	319	190	116	124	198	149	309	[242]	33
4	414	284	md	374	194	132	92	117	99	289	166	34
5	419	226	md	316	143	117	130	137	93	272	112	38
6	410	189	md	256	134	105	155	132	96	250	64	38
7	416	247	md	310	131	104	165	106	75	213	39	36
8	400	270	md	285	121	95	172	98	76	370	61	23
9	397	234	md	279	112	93	173	96	76	263	38	28
10	386	229	md	273	86	89	166	87	79	278	20	27

Note: Contact problems between soil and sensor on the ploughed plot led to missing data (md) for 4.1 Mg 60 cm. Bracketed values are outliers caused by contact problems between the sensor head and the soil matrix. SST, stress-state-transducer system.

Figure 5. Average major principal stress (σ1) in kPa as measured by the SST system, P = ploughing, M = minimum tillage, ± standard error. Average values based on reduced number of passes (n = 7/9) due to huge variation in dataset. Outliers marked in Table 5 are omitted. Data based on the highest σ1 which occurred under the trailer wheels.

Statistical analyses

Statistical analyses (t tests) were performed with Minitab 15 software. Significance was assumed at an α level of 5%. The outliers of the SST measurements are marked in Table 5. These outliers are not included in average values used in t tests (Figure 5). All figures show the standard errors of the means.

Results

Stress-state-transducer measurements (SST)

Differences were found in the major principal stress (σ1), depending upon soil tillage, wheel load and soil depth. The data (Table 5) show some variation in σ1 between wheeling events. The smaller contact area found with the 4.1 Mg treatment led to higher σ1 at both 20 and 40 cm depth of the ploughed plot. The opposite was found on the minimum tilled plot at 20 cm depth. However, the stress was also higher with the 4.1 Mg wheel load than with the 6.6 Mg load at greater depth on the plot under minimum tillage (Table 6, Figure 5).

Table 6. Classification of ratio between major principal stress (σ1) and Pc.

			Ploughing		Minimum tillage
Depth (cm)	Wheel load (Mg)	Moisture (kPa)	Pc/σ1	classification	Pc/σ1	Classification
20	4.1	6	0.08	Weak	0.37	Weak
40			0.23	Weak	0.42	Weak
60			md	md	0.55	Weak
20	6.6	6	0.10	Weak	0.20	Weak
40			0.38	Weak	0.50	Weak
60			0.50	Weak	1.46	Stable
20	4.1	100	0.25	Weak	0.86	Labile
40			0.52	Weak	0.99	Labile
60			md	md	1.33	Stable
20	6.6	100	0.34	Weak	0.46	Weak
40			0.88	Labile	1.19	Labile
60			1.22	Stable	3.56	Very stable

Note: Data from Table 4 (σ1) and Table 5 (Pc). Values for −100 kPa calculated according to Terzaghi and Jelinek (1954).

On the minimum tillage plot, all wheeling treatments indicated lower σ1 than on the ploughed plot. Comparison between the two tillage treatments showed a reduction of stress following minimum tillage. At 20 cm depth, the stress impacts with the latter treatment were approximately 65% lower at 4.1 Mg wheel load (p = 0.001) and 11% lower at 6.6 Mg wheel load (n.s.) than on the ploughed plot. At 40 cm depth, stress following minimum tillage was 54% lower at both wheel loads (4.1 Mg p = 0.001, 6.6 Mg p = 0.005) than on the ploughed plot. At 60 cm depth, the stress was reduced by 78% with the 6.6 Mg wheel load (p = 0.001) relative to the ploughed plot.

Displacement stress transducer measurements (DTS)

DTS data are in relation to wheel load divided between a vertical and a horizontal dimension, caused by the movement of machinery (wheelslip + vibration of running gear). Both dimensions consist of elastic and plastic deformation.

Vertical deformation

Most of the measured soil deformation was found to be elastic, but wheeling also led to vertical plastic soil displacement in all cases (Figure 6). Higher wheel load led to increased plastic displacement. The minimum tillage plot showed less cumulative plastic displacement than the ploughed plot. In the case of the 4.1 Mg treatment, there was 41% less total plastic displacement following minimum tillage, whilst in the case of the 6.6 Mg treatment, there was a 19% reduction in total plastic displacement, compared to the ploughed plot.

Figure 6. Elastic and plastic vertical displacement (mm) in the upper soil layer of ploughed (P) and minimum tilled (M) plot for all ten passes, at both wheel loads (left side 4.1 Mg and right side 6.6 Mg). P = ploughing, M= minimum tillage. Percentages of elastic and plastic displacement for each pass are shown beneath the figures.

Each wheeling event led to additional plastic soil displacement, but there was an indication that additional plastic displacement decreased with wheeling intensity. The cumulative plastic displacement on the ploughed plot was approximately 16 mm at 4.1 Mg wheel load and 17 mm at 6.6 Mg wheel load. Differences in cumulative plastic displacement between 4.1 and 6.6 Mg wheel load were smaller on the ploughed plot than on the minimum tillage plot. Data show that ten wheelings on the minimum tillage plot led to approximately 9 mm and 14 mm cumulative plastic soil displacement at 4.1 and 6.6 Mg wheel load, respectively.

Horizontal deformation

The data show that all wheelings caused noticeable horizontal plastic displacement (Figure 7). Less soil deformation was registered on the minimum tillage plot than on the ploughed plot.

Figure 7. Elastic and plastic horizontal displacement (mm) in the upper soil layer of ploughed (P) and minimum tilled plot (M) all ten passes, at both wheel loads (left side 4.1 Mg and right side 6.6 Mg). Percentages of elastic and plastic displacement for each pass are shown beneath the figures.

Following minimum tillage, horizontal plastic displacement was 49% and 45% lower at 4.1 Mg and 6.6 Mg wheel loads, respectively, compared with the ploughed plot. There were only small differences between the different wheel loads. Wheeling with 4.1 Mg wheel load led to ca. 5 mm cumulative horizontal plastic deformation on the ploughed plot compared to ca. 7 mm with 6.6 Mg wheel load. There was, with both wheel loads, a tendency towards a reduction of soil displacement with increasing number of wheelings on the ploughed plot (Figure 7). Every wheeling led to plastic displacement also on the minimum tilled plot, with approximately 2 and 4 mm cumulative plastic displacement at 4.1 and 6.6 Mg respectively.

Precompression stress (Pc)

According to Pc theory, the compaction susceptibility of a soil varies according to soil tillage. Pc values increased with depth, with the highest value at 40 cm depth. All Pc values on reference plots can be classified as low. Multiple wheeling with 4.1 Mg wheel load led in the upper soil layer of the ploughed plot to increases in Pc of approximately 51% (p = 0.005) with 4.1 Mg wheel load and 112% with 6.6 Mg wheel load (p = 0.003), compared to reference plot. Pc following minimum tillage was 74% higher than that found after ploughing (p < 0.001; reference plot). In comparison to the reference plot, wheeling with 4.1 Mg wheel load led to increases in Pc of approximately 17%, (p = 0.030) at 4.1 Mg and 3% (n.s.) at 6.6 Mg (Figure 8).

Figure 8. Average (n = 8) Pc in soil profiles after multiple (10×) wheeling for ploughed (P) and minimum tilled plot (M). R = reference plot, 4.1 Mg wheel load, 6.6 Mg wheel load. Pc measured in laboratory (−6kPa), ± standard error. Pc can be classified as low (30–60 kPa), medium (60–90 kPa) and high (90–120 kPa) (Horn & Fleige 2003).

Figure 8 shows that the average of ten wheelings with 6.6 Mg wheel load on the ploughed plot led to a 33% increase (p = 0.002) in Pc value at 40 cm depth. Wheeling on the minimum tillage plot gave increases in Pc values of 88% at 4.1 Mg wheel load (p < 0.001) and 18% at 6.6 Mg wheel load (p = 0.015). The average increase in Pc in 60 cm depth was 7% (n.s.) and 20%, (p = 0.004) at 4.1 Mg and 6.6 Mg, respectively, on the ploughed plot, relative to the reference plot (Figure 8). Pc for the minimum tillage plot (60 cm) increased by 29% (p = 0.005) when wheeling with 4.1 Mg wheel load and by 4% (n.s.) when wheeling with higher wheel load (6.6 Mg) compared to the reference plot. A comparison of the reference plots in subsoil (40 + 60 cm) revealed no significant differences between tillage treatments in Pc.

Ratio of Pc to applied stress (Pc/σ1)

Soil structure was most stable when the moisture content was low (Figure 9). At high moisture content in soil, all wheeling on ploughed plot led to stresses that exceeded soil internal strength at all depths. The ratio Pc/σ1 was in most cases classified as weak (Table 6). Wheeling on drier soil might prevent exceeding Pc at 60 cm depth with 6.6 Mg wheel load, as the stress ratio was classified as stable.

Figure 9. Precompression values (Pc) in kPa and major principal stresses (σ1) in kPa on average for multiple wheeling (Table 7). Pc values for −100 kPa calculated according to Terzaghi and Jelinek (1954).

Reduced soil tillage led to higher and more evenly distributed Pc throughout the soil profile. In comparison to the ploughed plot, only a rudimentary plough pan could be found at 30 cm depth. When water potential was −6 kPa, all wheeling would have exceeded Pc in the 0–40 cm soil layers. At 60 cm depth, the stress ratio can be classified as stable and soil internal strength can withstand wheeling with 6.6 Mg wheel load (Table 6, Figure 9). This figure illustrates that the use of these two slurry tankers exceeded Pc in the upper soil layer in all cases. Especially for minimum tillage, the degree of damage to subsoil is dependent upon soil water content. At a water potential of −100 kPa, wheeling will probably only exceed Pc in the upper soil layer.

Discussion

Stress-state-transducer systems (SST)

The results measured with the SST equipment show some variation for the single wheeling events both between wheeling events at the same soil depth, but also between different depths (Table 5). In this study, we therefore focused on the average value of several wheeling events (Figure 5). Our data show that contact area influenced the major principal stresses in the upper soil layer of the ploughed plot. It is well known that tyre width may reduce the risk for compaction in the upper soil layer (Lebert et al. 1989). Our data illustrate that the contact area of the big tanker (6.6 Mg) was greater on the ploughed plot than on that with minimum tillage (Table 4). A similar result was also found by Alakukku et al. (2003), who stated that tyre contact area on a deformable surface (e.g. ploughed soil) is greater than on a more rigid surface (e.g. minimum tilled soil). This is due to the fact that on a deformable soil surface, both the tyre lugs and the area between the lugs are in contact with the soil, whilst on a rigid surface, it may be only the lugs that are in contact with the soil. Thereby, tyres with a high wheel load and small contact area may create deeper ruts on a deformable surface than on a rigid surface (Raper 2005), until soil strength increases enough to support the weight. This could explain why the small slurry tanker (4.1 Mg) caused higher major principal stress on ploughed plot than did the big slurry tanker (6.6 Mg).

Results from this study clearly mirror differences in Pc between the two different tillage methods. The major principal stress was less for all wheeling treatments on the minimum tillage plot (Figure 5). This corresponds with the findings of Wiermann et al. (2000), who found more pronounced soil displacement at conventionally tilled sites than on reduced tilled sites, which tend to have higher bulk density, increased amount of inter-particle contacts and lower stress propagation because of better horizontal support of soil particles (Zink 2009). On the minimum tilled plot, our findings show that the major principal stress was reduced on average by 60% and 48% with 4.1 and 6.6 Mg wheel load, respectively. This is in agreement with results found by other authors. Wiermann et al. (2000) and Zink et al. (2010) showed that reduced soil tillage is more effective than ploughing in terms of robustness against stresses during wheeling. Major principal stress in their trials on a reduced tillage plot was on average between 20% and 26% lower than on a ploughed plot. The level of the major principal stress measured in our trial (50–400 kPa) is in agreement with the findings of these authors.

The variation in major principal stress (Table 5) may be due to errors in measurement techniques caused by inserting sensors into the soil. This involves disturbing the soil and changing its surrounding properties (Dexter 1997; Figure 2). During wheeling, the soil particles surrounding the sensor are moved due to compaction, which may in some cases cause a lack of contact between the stiff metal-measuring device and surrounding soil. In addition to this, the angle and exact position of the tractor wheel in relation to the sensor head is important in order to get comparable results. With repeated wheeling under field conditions, it is difficult to hit the sensor at the same angle each time and there may thus be a variation in the angle between sensor head and tyre from one wheeling to another.

Studies by Alakukku et al. (2003) showed that the weight distribution during field operations may vary between axles and tyres due to vibration and speed. Furthermore, pressure is not uniformly distributed over the whole contact area and may be 4–10 times higher under lugs or stiff tyre walls (Burt et al. 1992). A 300% difference in major principal stress (σ1) between tyre lugs and the inter-lug area was described in Alakukku et al. (2003) and Horn (2003). This effect was found to be limited to the upper soil layer, but this does not correspond with the findings in our trial. Although soil settlement depends on the duration of loading (Horn & Hartge 1990) and despite the fact that higher velocity may reduce soil stress (Alakukku et al. 2003), the driving speed in this trial was in accordance with the recommendations of the slurry tanker producer. Under field conditions, the possibilities to increase speed are often limited, due to considerations of work quality, engine power and field size.

Displacement transducer systems (DTS)

Results from this study show that wheeling led to elastic and plastic displacement in all cases. Since elastic displacement is considered to be reversible, the plastic displacement has probably most influence on pore structure and pore function (Peth et al. 2010).

Our data show that stress induced by wheeling exceeded Pc and thereby led to plastic displacement in the upper soil layer. Plastic displacement will create a new soil structure with corresponding changes in soil properties and mechanical stability (Horn 2000 in Peth et al. (2006). This led to compaction of the upper soil layer which could be seen as ruts on the soil surface (Figure 4).

Zink (2009) described a reduction of soil displacement with increasing number of wheeling events, due to a more stable soil structure created by the progressive compaction of soil particles. This could explain our findings for the unstable upper soil layer of the ploughed plot. The results for the ploughed site showed a tendency towards a decreasing amount of plastic displacement with increasing number of wheeling events. The first wheeling might be most harmful in some cases, but our results show that in some cases, the highest major principal stress may nevertheless occur after several wheeling events. Every wheeling causes plastic displacement with a cumulative effect (Figures 6 and 7). This effect was less pronounced for minimum tillage, where results for the higher wheel load (6.6 Mg) showed a tendency towards a decreasing amount of plastic displacement with increasing number of wheelings. This is in agreement with the findings of other authors such as Bakker and Davis (1995) who found that the first pass of a wheel causes the major portion of soil compaction but that deformation increases with the number of passes. Effects of tyre lug pattern or lack of contact between sensor head and soil structure may also have influenced these results. These findings are important in practical agriculture, because they indicate that a few passes with a small tractor can do as much, or even greater, damage than one pass with a bigger, heavier machine, as was shown by Bakker and Davis (1995).

There are two major processes that cause soil displacement: compression and shear stress. The first causes a densification, and the second causes a displacement through rearrangement of soil particles (McGarry 2003). This is reflected in our data that show that, besides stress in the vertical direction, every wheeling also caused stress and plastic displacement in the horizontal direction. As expected, this horizontal component is less pronounced than the vertical displacement and shows the same pattern as described for the vertical displacement. Our data indicate that the Pc of the minimum tillage plot (Figure 8) gave greater robustness also in the horizontal direction. This corresponds with the findings of Alakukku et al. (2003), who stated that profiled, moving and deflecting tyres caused vertical and horizontal stresses and shear forces in soil. Kirby (1991) found that dynamic loadings, and their associated shear stresses, caused greater damage than did normal vertical stress alone.

Horn (2003) described how horizontal displacement caused an additional rearrangement of soil particles and a shearing of the pores in the upper soil layer. Thus, this horizontal displacement may cause a complete disturbance of the pore system, its pore continuity and ecological functioning. The maximum total horizontal plastic displacement up to 7 mm in our data (6.6 Mg ploughed plot, Figure 7) will cause a rearrangement of soil particles, and have important influences on both pore size and pore continuity. Keller et al. (2004) concluded, after comparing Pc measurements in laboratory and field, that even stresses below Pc can cause plastic displacement under field conditions.

Horn et al. (1995) showed that increasing wheeling intensity led to smaller vertical stresses in the upper soil layer due to an increase in bulk density, elasticity and shear strength. Continued wheeling may therefore result in further deformation of deeper soil horizons. It would be interesting to investigate elastic and plastic displacement at greater soil depth, in order to verify the influence of soil inner strength on soil deformation in subsoil, and whether deformation of the upper soil layer influences deformation of deeper soil layers.

Precompression stress (Pc)

All the Pc data are related to reference plots. The reference plots were not wheeled in this study, but had been wheeled during field-work in the previous years of the trial.

The comparison of the Pc of the two reference plots showed distinct differences between the tillage methods. The effect of soil tillage was most pronounced in the upper soil layer. Pc and thereby soil strength was about 74% higher in minimum-tilled soil than in ploughed soil in this layer. On the ploughed plot, we found a distinct dense layer at 40 cm depth with very high Pc, which was approximately 8% higher than on the minimum tilled plot at the same depth. The data show no differences in Pc in deeper soil layers (60 cm depth). The overall depth distribution of Pc describes the long-term effect of the different types of soil tillage on Pc. This was expected, since these findings correspond with studies of other authors, which have shown that long-term conservation tilled soils have higher Pc and a more equally distributed stress pattern than do conventionally tilled soils (Horn & Rostek 2000; Wiermann et al. 2000; Horn 2003; Peth et al. 2010). Thus, soils with reduced tillage are expected to be less sensitive to soil compaction (Wiermann & Horn 2000).

The results show that the wheel loads used in our trial had a significant effect on Pc at all depths. Loads exceeded the Pc of the reference plots in all cases down to 60 cm depth. The effect of wheel load was more pronounced on the ploughed plot at all three depths where wheeling with 6.6 Mg wheel load led in all cases to greater increases in Pc than wheeling with 4.1 Mg wheel load. This effect was reversed for minimum tillage at all depths, as 4.1 Mg wheel load led to higher increase in Pc on the minimum tillage plot than on the ploughed plot. This does not correspond with the findings of Horn et al. (1995), who concluded after a series of field trials including different wheeling intensities, that the greatest wheeling intensity always led to the highest Pc at all depths. This contradictory result could be explained by the higher contact pressure of the small slurry tanker (4.1 Mg; Table 4) that, in combination with the higher Pc of the minimum tilled plot (Figure 9), led to more concentrated stress underneath tractor wheels and thereby deeper stress transmission than underneath the wider tyres of the 6.6 Mg tanker. A general problem with measuring Pc is that it is measured in the laboratory at a soil matric potential of −6 kPa, which may differ from the soil moisture content under field conditions at the time of sampling, as was the case in our trial. Furthermore, loading times in the laboratory are much longer than under wheeling in agricultural practice (Peth et al. 2010). Lebert et.al. (1989) stated that stress depends on the time of loading and that Pc values increase with decreasing loading time (Zink 2009). This may be due to the fact that during dynamic stresses, soil strength is reduced more than under static forces (Keller et al. 2004). Pc increases with clay content. In clay soil, Pc may be doubled if stress is applied for a short time only (Keller et al. 2004). Classification values should therefore also take account of clay content (Horn & Fleige 2003). Pc is classified as a low indication factor (Horn & Fleige 2009), which depends strongly on soil water content and which does not give any information on how the soil strength should be classified in terms of ecological soil function. Nevertheless, Pc is a useful indicator to illustrate effects of induced stresses on soil inner strength, and it may thus be used as a precautionary value.

Ratio Pc/σ1

According to Pc theory, all soil stresses that exceed Pc will lead to plastic, irreversible changes in soil structure, soil function and ecological properties. If the ratio (Pc/σ1) is smaller than 0.8, a soil is classified as unstable and further changes in ecological properties can be expected (Horn & Fleige 2003).

Our results for high water content (matric potential −6 kPa when measuring Pc), suggest that there is an imbalance between Pc and applied stress. Wheeling with the chosen wheel loads was harmful on both plots, with nearly all ratio values smaller than 0.8, which is classified as weak (Table 6). Although the changes in Pc in the subsoil were less pronounced on the minimum tilled plot than for the ploughed plot, there is a risk for both wheel loads to increase Pc in the subsoil. This means that although minimum tillage gave higher Pc, it cannot completely protect the subsoil from damage at such wheeling intensities at this water content. At higher water content, the risk of damage would be even greater.

This result may be used as an indicator that more stable structure of minimum tilled soil is more effective in terms of compensation stress during wheeling than ploughed soil. Analogous to the findings of Zink et al. (2010), our study revealed higher Pc in the upper soil layer of the minimum tilled plot and simultaneously reduced σ1 at all depths for all loads, compared to the ploughed plot.

When soil is drier (−100 kPa), stress (σ1) still exceeded Pc at both 20 and 40 cm depth on the ploughed plot, whereas it only exceeded Pc in the upper soil layer on the minimum tilled plot. These findings show that soil water content has a very large influence on soil strength by limiting the plastic, irreversible soil deformation on the upper soil layer. This corresponds with findings of other authors who have suggested reducing soil moisture as one method to increase soil strength (Raper 2005). In practical terms, this means that permissible load of agricultural vehicles increases with decreasing soil moisture content (Hamza & Anderson 2005). In combination with the implementation of minimum soil tillage, low soil water content might help to limit harmful soil compaction to the upper soil layer. This would however involve a delay in spring sowing time, which is likely to reduce yield potential in this region. Furthermore, minimum soil tillage may under certain circumstances increase soil water content in the upper soil layer compared to ploughing. Since Pc in the upper soil layer was higher with minimum tillage than with ploughing (Figures 8 and 9), it can nevertheless be expected that minimum tilled soil may be more robust against induced stresses (Wiermann et al. 2000) (and thereby protect deeper soil layers) than ploughed soil despite its higher soil water content. This would be interesting to investigate in further field trials.

Peth et al. (2010) describe how stresses exceeding the Pc value cause irreversible changes in soil structure, soil function and ecological properties. A change in air permeability and plant available water capacity can be expected (Horn & Fleige 2003). According to our results, it can be expected that induced stresses have influence on soil functional parameters.

Summing up it can be stated, that minimum soil tillage is often aimed, among other things, towards obtaining a stable soil structure that may better protect the (sub-) soil from damage by compaction. Results from this trial show impressively that minimum soil tillage can provide higher soil strength, which is more robust against induced stresses under wheeling and may protect subsoil structure to a certain extent when comparing same machine on both plots. However, higher soil strength should not be misused by the use of heavier machinery, nor should compromises be made on tyre equipment etc., because continued wheeling with high wheel loads on clay soils, as in our trial, leads in all cases to plastic deformation and probably has negative effects on ecological soil functioning.

Furthermore, our results indicate that both tyre contact area and soil moisture content have important influences on soil strength and the risk for damage by compaction. In agricultural practice, it is important to keep in mind that it is not only the weight of machinery that is important, but also the number of wheelings, tyre inflation pressure and contact area. Whilst the soil sampling was done here under soil moisture conditions drier than field capacity, these conditions are not typical in Norway in early spring. It may be expected that wheeling with such slurry tankers is often done under less optimal conditions in practice, giving a much higher risk of compacting the soil, and even more distinct differences than we found between different tillage systems and axle loads. This is especially true for soils that are sensitive to compaction, such as the clay soils used in this study.

Especially against a background of increasing machinery size, increased economic pressure and climate change, it is important to choose the right type of machinery and the correct time of wheeling the field, in order to avoid irreversible plastic soil deformation. Large tyre contact area, use of minimum tillage and correct management in terms of both time and intensity of wheeling may help to reduce the risk of soil compaction. Further research should be performed over a longer time period to determine whether the described changes in inner soil strength negatively influence soil physical parameters and thereby yields.

Acknowledgements

Our thanks go also to Bjørn Inge Rostad and Per Ove Lindemark, Norwegian Agricultural Extension Service (South East), for technical support and permission to use the long-term field trial.

Funding

This project was financially supported by ‘The Norwegian fund for soil and peat investigations’.

Additional information

Funding

Funding: This project was financially supported by ‘The Norwegian fund for soil and peat investigations’.