This study was carried out on an experimental field of the Agricultural Research and Application Centre of Suleyman Demirel University from 2010 to 2013. Chemical and physical properties of the soil are given in Table 1.

Organic vegetables were cultivated with farmyard manure (FYM), green manure (GM), and conventional tillage (CT) without any organic matter treatments. The field experiment was set up with a completely randomised design with three replications. FYM application was executed between 21 May and 7 June on 35 t ha-1 (FYM consist of 45 % dry matter), and the FYM was thoroughly mixed into the soil surface layer (10–15 cm) using a rototiller. Common vetch (Vicia sativa L.) was sowed in the second week of March and allowed to grow until the flowering stage between 21 May and 7 June. It was then incorporated into the soil first by chisel (20–25 cm tillage depth) and followed by rototillers (10–15 cm tillage depth). Both GM and FYM treatments were performed at the same time. The field was tilled by means of disc-harrowing at 10–15 cm depth (16 September 2013), and the field was then sprinkler irrigated on 27 September 2013. In order to increase the susceptibility of soils to compaction, field traffic was simulated at a soil water content just below the field capacity (0.23 g g-1) on 5 November 2013. A tractor (80–66 s Fiat, manufactured in 1998) with 85 horse power (HP) weighing 3460 kg, including the operator, was used to resemble field traffic. Tractor passes (one, three, and five times) were performed at 5 km h-1 speed on the same track. One disturbed and three undisturbed soil cores per plot were taken from 0–10 and 10–20 cm depths before each tractor pass; only undisturbed core samples were taken after each treatment pass. Penetration resistance, which is an indication of soil compaction, was measured with a penetrometer (Eijkelkamp penetrograph) equipped with a 1 cm2 cone attachment. Penetration resistance was performed in 15 replications per treatment at 2 m intervals and 0–20 cm depth. The averages of resistance values obtained for the 0–10 and 10–20 cm depths were evaluated. Physical and chemical properties of experimental soil, FYM, and plots before passing are given in Tables 1 and 2. Particle size distribution was determined by means of the Bouyoucos hydrometer method (Bouyoucos, 1962) and bulk density was determined using undisturbed soil cores (Blake and Hartge, 1986). For this analysis, soil cores were oven-dried, weighed, and then calculated by dividing the dry weight by the core volume. Total porosity was calculated from the ratio of water weight at saturation to total volume of a 100 cm3 undisturbed soil core (Danielson and Sutherland, 1986). Microporosity was accounted for as the volume of water at field capacity, measured by using a pressure membrane apparatus at 0.033 MPa suction. Macroporosity was calculated from the difference between total porosity and microporosity (Danielson and Sutherland, 1986). A modified Walkley–Black wet oxidation method and dry ashing method (at 400 ∘C for 16 h in an oven) were used for determining organic matter content of soil and FYM respectively (Burt, 2004). Soil pH and electrical conductivity (EC) were measured in a 1:1 soil-to-distilled water suspension, whereas pH and EC of FYM were determined in a 1:2.5 FYM-to-distilled water suspension (Burt, 2004). Carbonate equivalent was determined by a volumetric method using a Scheibler calcimeter (Kacar, 2009).

The data were subjected to descriptive analyses in order to check normal distribution. Aside from compaction data sets, all parameters measured showed typical normal distribution. The compaction data were log-transformed before analysis of variance (ANOVA) using Minitab 16 statistical package programme (Minitab, 2010). The mean separation between the treatments was performed using a least significant difference (LSD) test at 95 % confidence level.

The application of FYM and GM for 4 subsequent years significantly reduced penetration resistance in both depths; however, the effect of FYM treatment was higher than GM treatment (Fig. 1, Table 3). This finding is in accordance with the previous studies (Celik et al., 2010; Gulser and Candemir, 2012; Xin et al., 2016). Incorporation of organic matter in clay-textured soils can strengthen the aggregates by weakening cohesion forces and interfering with the formation of crust and large aggregate (Aksakal et al., 2012). The larger amounts of added organic matter may mediate the formation of clay–organic matter complexes, which in fact reduces the penetration resistance on the one hand and protects organic matter against microbial decay on the other hand. In this respect, Blanco-Moure et al. (2016) investigated the effect of soil texture on carbon and organic matter distribution among different fractions under different tillage and management practices. They found that soil clay had a critical role in the chemical stabilisation of organic matter through clay–organic complexes in the soils. Czyz and Dexter (2016) pointed out the relation between the magnitude of clay–soil complex and the porous and open nature of the structure. Thus, stable organic matter sources such as FYM resulted in desirable penetration resistance (< 2 MPa) for plant growth under changing field traffic.

The increasing number of passes, irrespective of the organic matter treatments and soil depth, increased soil compaction measured by penetration resistance (Table 3). The effect of field traffic on penetration resistance, as expected, was more negative in 0–10 cm depths (Table 3). Accordingly, Carman (1994) and Seker and Isildar (2000) determined a higher compaction ratio in the 0–10 and 0–15 cm surface layers respectively. A penetration resistance value as high as 3.60 MPa in the surface layer caused by five passes in the control treatment (Fig. 1) with no organic matter may have significant inverse effects on infiltration, percolation, and run-off-induced erosion under intensive precipitation events in slopy lands (Kozlowski, 1999; Seker and Isildar, 2000). In this study, we also determined the well-known manner in which surface soil becomes more compact with field traffic, and the severity of the problem may be overcome by adding organic matter to soil or by adopting soil management systems with decreased annual traffic. The penetration resistance value at 10–20 cm depth obtained for CT and GM treatments after five passes was over 2 MPa, which is considered the limit value by the USDA (1993) as a critical physical quality parameter in conventional agricultural practices. This critical value can change depending on the soil tillage systems. For example, with minimum tillage practices where a chisel is used for soil tillage it is 3 MPa and with no-till practices it is 3.5 MPa (De Moraes et al., 2014). The critical penetration resistance value that inhibits root development is accepted as 3 MPa (Busscher and Sojka, 1987; Hakansson and Lipiec, 2000; Aksakal et al., 2011). Soil management systems can change soil organic carbon contents (Munoz-Rojas et al., 2015) and field traffic density, which ultimately degrade soil physical traits for optimal plant growth such as water-stable aggregates and bulk density (Gelaw et al., 2015). In fact, these tendencies of soil physical traits can lead to more compaction in both the surface and subsurface soil layers, as in our case.

The main effect of organic matter treatments on bulk density was statistically significant (P < 0.01). Both FYM and GM incorporation into soil were distinctly different than the control (Table 3). FYM amendment reduced the bulk density to as low as 1.40 g cm-3. Similar to our findings, decreases induced by adding organic matter (Haynes and Naudi, 1998; Chaudhari et al., 2013; Gulser and Candemir, 2012) and increases in bulk density related to field traffic (Seker and Isildar, 2000; Patel and Mani, 2011) have been frequently reported in the literature. The magnitude of the above-mentioned changes were depth dependent. In terms of plant growth, Aksakal et al. (2016) reported the enhancing effect on bulk density of increasing vermicompost treatment rate for three soils with differing clay content. In general, mixing of soil with less-dense organic material results in decreased particle density in soils amended with organic manures (Haynes and Naidu, 1998). However, their efficiency in improving bulk density for plant growth is related to the magnitude and quality of organic residues (Aksakal et al., 2016). Green manure, which is largely decomposed by leaving a smaller amount of organic matter, has a limited influence on soil physical attributes (Sauerbeck, 1982). Since FYM is more stable than the GM in terms of decomposition resistance, more organic compounds accumulated in soils (Table 2) treated with FYM. This in fact mediated the formation of aggregates resistant to soil traffic and therefore minimum bulk density was obtained in FYM plots. The main effect of soil depth irrespective of number of passes and organic matter addition was significant (Table 3). Surface layer had a smaller bulk density. The minimum bulk density (1.25 g cm-3) was obtained at 0–10 cm depth for one pass in FYM treatments, whereas the maximum bulk density (1.61 g cm-3) was recorded at 10–20 cm depth in the CT treatment with five passes (Fig. 2). The depth of soil compaction in the soil profile is dependent on the axle load, soil moisture content, tyre size, contact pressure, traffic density, and soil organic matter content attributes such as aggregation, aggregate stability, porosity, etc. (Hamza and Anderson, 2005). The greater the axle load and the wetter the soil is, the deeper the soil consolidates in the soil profile. Since in our case these two dependents were constant, the effects of organic matter treatments are rather apparent. The stronger structures induced by a larger amount and decomposition-resistant organic substances scattered the force to a larger area, which minimised the compaction-induced bulk density differences in FYM-treated plots at any given pass number. At any steady state condition, in terms of organic matter addition, each treatment may be regarded as at fixed conditions. However, with a differing number of passes for each treatment, as in our case, the increasing traffic density resulted in increases in the bulk density deeper in the profile (Fig. 2). Parras-Alcantra et al. (2015) similarly reported that organic farming compared to conventional tillage significantly improved soil organic carbon stocks of soil, which resulted in a decrease in soil bulk density in the soil profile as deep as 76.1 cm. Conversely, organic matter loss from conversion of forest soil to agricultural lands has made soil progressively bulkier for vertisols and ultisols in the last 20 years (Bruun et al., 2015), which in turn explains the significance of organic matter in managing bulk density.

The main effects of organic matter incorporation, depth, and number of passes on porosity were significant (Table 3). The effects of treatments, in relation to soil depth and field traffic density, on porosity in descending order were FYM > GM > CT. The main overall effect of depth on total porosity was detrimental at 10–20 cm depth where a significant decrease was observed. The initial average (0.565 cm3 cm-3) of total porosity calculated from Table 2 showed a detrimental decrease down to 0.418 and 0.441 cm3 cm-3 after five and three passes (Table 3). From all plots considered, the maximum and minimum values of total porosity were 0.53 and 0.39 cm3 cm-3 respectively (Fig. 3). The effects of organic matter treatments, depth, and field traffic density on microporosity were statistically significant (Table 3). Microporosity and total porosity parameters similarly responded to organic matter treatments. The volume of micropores was significantly higher in surface layer than the one observed at 10–20 cm (Table 3). Maximum microporosity (0.38 cm3 cm-3) was recorded in the FYM treatment at 0–10 cm depth with three passes, whereas the minimum (0.32 cm3 cm-3) was determined in the CT treatment in the surface layer (0–10 cm) with five passes (Fig. 4). There was an increase in the microporosity of the control plot after one pass compared to its initial porosity, whereas a decrease was recorded for annual FYM- and GM-incorporated plots. After three passes were done, microporosity increased in all treatments including the control. The initial microporosity value and the value after five passes were nearly the same in the CT treatment but they dropped below even the initial porosity for FYM and GM treatments. Although the organic matter incorporation was only implemented for 4 years, the enhancement was recorded for porosity parameters. The enhancement induced by organic matter amendments in long-term studies at various locations was even more astonishing and is similar to our findings. For example, Rasool et al. (2008) and Arthur et al. (2013) reported increases in total porosity and water retention, which are related to pore nature of soil, with an increase in organic matter content depending on agricultural practices or organic matter amendments. FYM and GM, to a relatively smaller extent, promote total and microporosity in the current study. Organic matter application promotes the development of better soil structure by binding the soil particles with polysaccharides and bacterial exudates, which results in decreased bulk density and hence porosity (Bhatia and Shukla, 1982). As the level of soil compaction increased, the amount of water held in high matric potentials decreased, whereas the magnitude of water held at low matric potentials increased (Gupta et al., 1989) due to conversion of some macropores into micropores from compression stress. Similarly, Seker and Isildar (2000) reported an increase in the plant pore volume for holding plant-available water after four passes. The descending order of the treatments was FYM > GM > CT for both of the depths and each treatment, and as pointed out by Celik et al. (2004), microporosity increased with organic matter amendments.

Macroporosity, which is critical for soil aeration and soil water circulation, was changed as a function of soil depth, field traffic density, and organic matter amendments (Table 3). The main effect of organic matter was FYM > GM > CT in descending order. In this study, organic matter amendments significantly improved macroporosity (P < 0.01). However, field traffic at three and five passes (Table 3) reduced the macropore volume below the critical level of 0.1 cm3 cm-3 (Hakansson and Lipiec, 2000) and covered the effects of organic amendments. The maximum macroporosity (0.16 cm3 cm-3) was observed at 0–10 cm depth of the FYM treatment at one pass, whereas the minimum was recorded for the CT treatment at 10–20 cm depth at five passes (Fig. 5).

Enhancement of soil structure traits by reduction of aggregate wettability (Zhang and Hartge, 1992) and enhancement of strength of aggregate stability by incorporation of organic matter partially eliminated the effects of field traffic on macroporosity after 4 consecutive years of FYM and GM application. The organic-matter-bound ambiguity was attributed to type of organic matter, C / N ratio, and the degree of resistance to decomposition. Readily decomposable soil organic matter was reported to be more relevant than total organic matter in mechanical characterisation of the soil (Ball et al., 2000). For example, the less-humified organic matter, such as green manure, was reported to highly efficiently increase aggregate porosity (Zhang, 1994). However, this effect was not found to be durable and resistant to field traffic as compared to the effects of FYM in the current study. Since the overall tendency of GM to increase soil organic matter is much less than the tendency of FYM, which is more stable and resistant to microbial decay, such behaviour is likely in soils that are poor in organic matter.