1 Introduction

Soil physical, biological, and chemical properties are influenced by soil-forming factors such as climate, parent material, vegetation, topography, and time, developing natural variations (Lin 2011). Agricultural practices, in particular tillage and traffic of machinery, add an extra variation, affecting the movement of water and nutrients, the carbon and nitrogen balance, and the root growth of plants (Strudley et al. 2008). These effects drive the development of alternative tillage systems, like no tillage, with the purpose of decreasing the negative impacts of plowing such as carbon losses or erosion (Haring et al. 2013; Mehdizade et al. 2013; Seitz et al. 2019).

Tillage modifies the hydraulic properties of the soils by changes in their structure and porosity. Conventional tillage creates temporal macroporosity which increments the saturated hydraulic conductivity (Ks) (Coquet et al. 2005), but these values decrease through the season because of particle settlement (Alletto and Coquet 2009). On the other hand, in conservation tillage system, Ks might be lower than under conventional tillage because of the high pressure exerted by seeders (Martínez et al. 2008), but physical behavior is more stable over time (Alletto and Coquet 2009) as well as the infiltration rate in near-saturated conditions could increase with the presence of organic residues (Lampurlanés and Cantero-Martínez 2006). It was proven that any factor that changes the porosity will influence the water movement, with local effects as those promoted by wheel tracks and structure regeneration during the season (Beven and Germann 2013; Moreira et al. 2016).

Studies on the impact of machinery on the soil surface evidence higher compaction in the wheel track compared with positions outside. This leads to changes in pore size distribution, especially at the macropore level, with the subsequent reduction of water mobility (Strudley et al. 2008). Horn et al. (2003) observed a deteriorating effect of machinery traffic on soil physical properties. Schaffer et al. (2007) compared the structure and shapes of macropores after the passage of machinery, evidencing structure deterioration and a greater distance between pores. Etana et al. (2013) studied the compaction of harvesters and its effect on soil functioning, concluding that the effects of compaction had a long-time persistence, especially penetration resistance and water movement. Other effects deriving from tillage and machinery passage include hardpan formations and superficial sealing, preferential water flow, and structure alterations (Hillel 2004; Chan et al. 2006; Pathak et al. 2011; Etana et al. 2013; Sasal et al. 2017).

Hydraulic conductivity (K), both saturated and unsaturated, and their dependence on different tillage systems, had been compared with contradictory results (Martínez et al. 2008; Jonard et al. 2013), because of a high variability, kind of soils, method of measurement, and sampling time during the season, among others (Lampurlanés and Cantero-Martínez 2006; Strudley et al. 2008; Villarreal et al. 2020). Moreover, the temporal variability is often analyzed in the long term, controlling the K value evolution over several years (Fuentes et al. 2004; Pöhlitz et al. 2018; Seitz et al. 2019). Regarding short-term tillage effects, some studies evaluated physical properties depending on wheel track and tillage operations (Barik et al. 2014; Sivarajan et al. 2018); however, few studies consider the K variations during the crop season of a specific plant species. For example, Osunbitan et al. (2005) illustrated the K variability on cultivated soils for an 8-week period after sowing, comparing no-tillage, manual, and conventional tillage systems. Their results evidence a constant decrease of K in all cases, associated with the reordering of particles after sowing. Hu et al. (2009) evaluated the variation of hydraulic conductivity, saturated and unsaturated, in four different crops with different managing systems (Glycine max, Caragana korshinskii plantation, Stipa bungeana pastureland, and Medicago sativa just established) over 4 months, where a general decrease of K in time was observed. However, this study does not relate to representative times like phenology stages or field operations, as it only measures at regular intervals.

On the other hand, the root system affects the soil porosity depending on its architecture. For example, alfalfa increases the macropores in the soil, because of its coarse and deepening roots (Uteau et al. 2013), which could be similar in other species with similar root architecture, like rapeseed (Brassica napus). However, Poaceae has shown only the superficial effects of rooting (Sawchik et al. 2012). This is the reason why crop rotations optimize not only the usage of nutrients and decrease disease incidences but also the root systems over time leading to a better soil structure and a decreasing spatial variability within the soil (Chan et al. 2006).

We hypothesize that the saturated hydraulic conductivity of the soil and its temporal variability is higher in conventional tillage compared with that in no-tillage system at the beginning of the crop season but then decreases to a similar or lower value, with an evident effect of wheel tracks caused by machinery. The general objective of the study was to evaluate the saturated hydraulic conductivity under no tillage and conventional tillage in a wheat-rapeseed rotation, characterizing its variation during wheat crop season as well as the pressure effects of wheel track.

2 Material and Methods

2.1 Site Description

The study was carried out from March 2014 to January 2015 at the Antumapu Experimental Station, in Santiago, Chile, maintained by the Agronomy Faculty of the Universidad de Chile (33° 34′ 17″ S; 70° 38′ 17″ W). According to the Köppen-Geiger classification, the station is located in a warm-temperate area, with a dry period of 7 to 8 months, winter precipitations of 308 mm, and 992 mm of annual ETo; the annual mean temperature is 14 °C, with January being the warmest (mean 28.2 °C) and July being the coldest month (4.4 °C).

The soil has an alluvial origin, belonging to the Santiago soil series, of the coarse loamy over skeletal sand, mixed, and termic of the Entic Haploxerolls family (Soil Survey Staff 2010), resting on deep gravels and stones in a coarse matrix representing 40 to 60% of the volume (CIREN 1996). Organic matter contents range from 4.35 to 5.75% in the no-tillage (NT) system and from 2.55 to 2.57% in the conventional tillage (CT) system.

At the Experimental Station, a tillage systems treatment has been implemented by the Soil-Water-Plant Relationship Laboratory, where diverse studies have been performed, comparing conventional tillage (CT) against no tillage (NT) since 1997 until present (Martínez et al. 2008, 2013). The study location is under rainfed conditions and had a wheat-maize rotation until 2008, and a wheat-rapeseed rotation thereafter. This study was conducted during the wheat period of the wheat-rapeseed rotation.

2.2 Study Design

A comparative study was performed with on one side a no-tillage (NT) system and on the other side a conventional tillage (CT) system during a wheat crop season. In each essay, a block design was imposed, using two treatments and four replicates. Treatments corresponded to machinery passage with an in-the-wheel-track (WT) and outside-the-wheel-track (OWT) measurement and were established on 5-m-wide and 10-m-large plots (eight plots in total). The experimental unit was defined as the area delimited by the width of the track and the length of the plot for WT, and a parallel area with similar size but 0.4 m away from the wheel track for the OWT treatment.

For each tillage system (NT and CT), a complete random block design was set with four replicates. Sampling and measurements were done in 6 different timesteps:

  • Before tillage (13–15 March)

  • After (conventional) tillage (3–9 April)

  • After sowing (20–25 June)

  • First knot (11–16 September)

  • Anthesis (3–7 November)

  • After harvest (26–30 January)

For the first sampling, which was before tillage, the already present wheel tracks from the previous machinery were considered, using the trampled straw still present in both areas as a reference. The sowing started at the end of May and fields were maintained under rainfed conditions during the season with one exception before sowing, where sprinkler irrigation was performed. This irrigation was equivalent to a precipitation of 15 mm and enough to reach the field capacity in the first 30 cm of the soil, according to the previous characterization of water retention curve in the same study site and managements (Martínez et al. 2008). Thus, an appropriate water availability to establish the crops was ensured.

A Ford 4310 tractor was used (4622 kg and 60 HP) and the WT treatment was established under the tracks of the back wheels (14.9″ diameter, 28″ tire ring, and 30 cm width), applying a pressure of 138 kPa on the ground. Before sowing, Anagram Plus® (commercial mix of mancozeb and carbendazim for fungus disease control) was applied using 125 g for each 100 kg of seed as well as Ajax® (metsulfuron metal for weed control) with a 19-g ha−1 dosage. After sowing, two applications of Pirimor® (pirimicarb for aphid control) were performed, using a dosage of 250 g ha−1. Maintenance fertilization was performed with 120 kg ha−1 of P2O5 in the form of triple superphosphate and without mitrogen application. The sampling periods were chosen considering relevant seasonal managements (before and after sowing) and important phenological timesteps for the crop (first knot to anthesis), when the number of grains is determined and there is no significant radical system growing from that period. The two tillage systems with two treatments each, four replicates, and six different periods add up to a total number of 96 different measurements.

2.3 Soil Properties

The hydraulic conductivity (K) was measured at the surface with a mini-disc infiltrometer in a near saturation condition (Perroux and White 1988). Supply tensions of 0.1, 0.2, 0.4, and 0.6 kPa were used on the soil surface. Before measurement, any organic mulch in the NT treatment was removed and plants (crop or weed) were carefully cut out if necessary, accurately avoiding any soil disturbance. Additionally, a fine sand layer was added to ensure the contact of the porous plate to the ground (Reynolds and Zebchuk 1996). After installation of the device, the infiltrated water was registered at 30-s intervals, starting from 0.1 kPa until a constant water infiltration was achieved, approximately 10 min after the beginning of the measurement. This procedure was repeated for each water tension increasingly (0.2, 0.4, and 0.6 kPa).

Data obtained was analyzed after Zhang (1997), based on the fluid equation of Richards, assuming that for a determined tension h, unsaturated hydraulic conductivity (Kns) is:

$$ {K}_{\mathrm{ns}}=\frac{C_1}{A} $$
(1)

With C1 being the curvature of the function determined by the accumulated infiltration I for a specific tension and the function of the square root of time t, the polynomic equation is described as:

$$ \mathrm{I}={C}_1t+{C}_2\sqrt{t} $$
(2)

where C2 is an adjustment parameter for the equation, and t is the time. In Eq. (1), A corresponds to a non-dimensional coefficient, depending on pore geometry in direct relationship with the textural class and water entry pressure. Considering the loamy textural class of the study site, values of 5.72, 6.27, 7.53, and 9.05 were used for 0.1-, 0.2-, 0.4-, and 0.6-kPa tensions, respectively (Decagon Devices 2012). Kns values in function of supply water tensions (h) allowed to perform a linear regression, determining the saturated hydraulic conductivity (Ks) as h zero tension.

During measurements of Kns, undisturbed soil samples were taken about 20 cm away from the mini-disc infiltrometer by the core ring method, to measure soil bulk density (Db) and volumetric water content (θ), both according to Sandoval et al. (2012). Values were obtained to test structural changes through the season using Db and to compare the homogeneity of θ during Kns measurements.

2.4 Statistical Analysis

To compare the treatments and tillage systems effects at different timesteps, diverse analyses of variance (α < 0.05) were done for soil bulk density and water content, after checking the assumptions of normality and homogeneity. In all cases, the assumptions were achieved and no transformations or alternative analysis was necessary. The in-the-wheel-track (WT) and outside-the-wheel-track (OWT) treatments were compared between timesteps and additionally; means for each period were compared as well. When appropriate, a multiple range comparison test (Tukey) was performed to separate the interactions among treatments and tillage systems and to discriminate differences between timestep measurements.

Regressions were performed for both the stabilized accumulated infiltration and the hydraulic conductivity, in the function of time and water suction, respectively. In all cases, adjustments were significant (α < 0.05). A mixed linear model was used were the intercepts differences from the linear regressions were compared at the saturation point (h = 0 kPa) with the four replicates of each treatment, and an analysis of variances was performed for the values on time and treatments. In the same way, the data obtained was compared with a Tukey multiple range test, considering the multiple interactions (treatment × tillage, treatment × time, tillage × time, and treatment × tillage × time). Finally, linear regression was performed for saturated hydraulic conductivity in the function of bulk density (α < 0.05) to explain the structure dynamic during the season depending on wetting and drying processes.

Data analyses were performed with Infostat software (di Rienzo et al. 2017). Normal distribution was checked and the residuals did not show irregularities in normality or homogeneity of variances.

3 Results

3.1 Soil bulk density

The soil bulk density (Table 1) showed values ranging between 1.18 and 1.33 Mg m−3, with little variation through the season and the lowest values at the anthesis measurement.

Table 1 Soil bulk density (Db, Mg m−3) obtained by the core ring method on no tillage (NT) and conventional tillage (CT), in the wheel track (WT) and outside the wheel track (OWT) at six timesteps. Average values (n = 4) with their corresponding standard error
Full size table

There were not any significant differences between no tillage and conventional tillage, nor between the WT and OWT treatments at different timesteps, except for the case before tillage, where an interaction between track position (WT and OWT) and tillage system was detectable.

3.2 Water Content

Table 2 illustrates the volumetric water content (θ) for the treatments in the different timesteps, for the upper 5 cm of the soil.

Table 2 Volumetric water content (θ, %) in no tillage (NT) and conventional tillage (CT), in the wheel track (WT) and outside the wheel track (OWT) at six timesteps. Average values (n = 4) with their corresponding standard error
Full size table

There was an interaction between the treatments and the tillage system for the first-knot sampling, where the highest water content outside the tractor tracks was registered in NT (21.2%), and the lowest value was registered in CT at the same position (OWT), with 11.3%. There are significant differences between the water content through the season, where the highest value is related to the irrigation before the sowing (24%). After that timestep, θ decreases with highest evidence at the anthesis timestep, as there was not any rain for a two-month period before the sampling. Strongest rainfalls were recorded between June and September, but the absence of rainfall made artificial irrigation necessary in April (Fig. 1).

Fig. 1
figure 1

Precipitation (mm) at the study site during the investigation period at Antumapu Experimental Station, Chile. The arrow indicates the initial irrigation and segmented bars the Kns sampling timesteps

Full size image

3.3 Unsaturated Hydraulic Conductivity

In Fig. 2a, the result of an infiltration test is shown as an example to understand the behavior of water infiltration obtained with different water tensions through the mini-disc infiltrometers.

Fig. 2
figure 2

a Accumulated water infiltration in the soil (cm) versus the square root of time in seconds for the wheel tracks (WT) in conventional tillage (CT) evaluated before the sowing, for each tension (kPa). b Unsaturated hydraulic conductivity (Kns) in function of water tension supply for each replicate of WT. Each triangle represents a value for each tension and different replicates. The line in the middle represents the linear regression and the near curves are the confidence intervals (p value < 0.05)

Full size image

The regressions in Fig. 2a follow the adjustment indicated on Eq. (2). If the square root of the time is considered, the C1 values are obtained for each tension (h), equivalent to the curvature coefficient (Decagon Devices 2012). From Eq. (1), Kns is calculated for each replicate of a treatment, as can be observed in Fig. 2b. Finally, the saturated hydraulic conductivity (Ks) can be obtained based on the intercept of the equation with the Y-axis.

Considering the homogeneous textural class at the study site, Kns will depend on the organic matter content and the secondary porosity generated by structuration processes. Table 3 shows the mean results of Kns in the function of water suction for each period.

Table 3 Unsaturated hydraulic conductivity (Kns) at different water tension supplies for conventional tillage (CT) and no tillage (NT), in the wheel track (WT) and outside the track (OWT). Average values (n = 4) with standard error
Full size table

A general tendency can be observed, with higher Kns values on CT than NT, with the exception of the start and the end of the season (before sowing and after harvest). WT and OWT did not show differences at the evaluated tensions, though the NT system tends to present higher Kns values in OWT than WT, except for the first-knot timestep. Besides, with CT, there was a high seasonal variability, related to the farm works and crop development.

3.4 Saturated Hydraulic Conductivity

The obtained Kns values processed with the above-mentioned method (Fig. 2) and the approximation through regression to a tension zero gives the saturated hydraulic conductivity. These values are presented in Table 4, with their respective comparisons per timestep.

Table 4 Saturated hydraulic conductivity (Ks, cm h−1) for no tillage (NT), and conventional tillage (CT), in the wheel track (WT) and outside the wheel track (OWT) at six timesteps. Average values (n = 4) with their corresponding standard error
Full size table

As Table 4 indicates, conductivity values were significantly different in both the time and treatments. Differences are present through every sampling time steps, with the exception of the post-harvest measurement. Highlight the high values of NT before tillage, similar to those obtained at end of the season. The CT system is highly dependent on soil preparation, reaching the highest value at OWT during anthesis. In addition to the results shown in Table 4, evaluated by a linear mixed model and considering the whole dataset analyzed with the multifactorial ANOVA test, interactions with factors can be established; the results are shown in Table 5.

Table 5 Statistical analysis of the saturated hydraulic conductivity with their F-values and p values
Full size table

In Table 5, a statistically significant interaction between tillage and timestep can be seen, with no other significant interactions and no significance for tillage system or wheel tracks. Table 6 details the results of tillage x timestep interaction, with the lowest values after sowing (0.74 cm h−1 in average) and the highest value in anthesis, with 2.82 cm h−1 for CT system.

Table 6 Saturated hydraulic conductivity (Ks, cm h−1) in no tillage (NT) and conventional tillage (CT) system at six timesteps, each value with its corresponding standard error. For the average, the ± SD was included with the respective CV in brackets
Full size table

The hydraulic conductivity at the start of the season had a mean of 1.46 cm h−1, decreasing in time in NT, with the lowest value after sowing; the CT system showed a more variable behavior, but it also reached the lowest value after sowing. After that, Ks values of CT rose again at the first-knot time step, not so with the NT system, which did not change at the first knot and reached a maximum Ks value at the anthesis timestep (2.25 cm h−1). Finally, there is a slight decrease in Ks at the harvest period, but with values similar to the beginning of the season.

As a result of the dynamic changes in soil properties during the season depending on wetting and drying events and structure regeneration, a negative correlation between Ks and bulk density was found (Fig. 3), but no effect of wheel track was observed. When the pre-tillage timestep is excluded (from the previous season), the adjustment improved.

Fig. 3
figure 3

Dependence of saturated hydraulic conductivity (Ks) on soil bulk density for samples belonging to conventional tillage (CT) and no tillage (NT), measured at different moments during the season (BT, before tillage; AT, after tillage; AS, after sowing; FK, first knot; A, anthesis; AH, after harvest). Adjustments significant at *0.05 and **0.01. In continuous line, the measurements before tillage (BT) were excluded

Full size image

4 Discussion

4.1 Soil Bulk Density and Volumetric Water Contents

The bulk density (Db) values are in the range of expected values for soils with a loamy textural class (Hillel 2004), considering that soils contain about 45% of sand particles at the study site (Martínez et al. 2008). The constant behavior of Db in the CT treatment can be explained by the continuous mechanical loosening of the soil profile, leading to a fast settlement after tillage, with an increment of the soil bulk density (Osunbitan et al. 2005). For NT, the constant Db during the season occurs because of the high mechanical stability and the absence of disturbance in this tillage system; nevertheless, all values are slightly high according to Reynolds et al. (2009), who point out that values over 1.25–1.30 Mg m−3 potentially cause yield losses due to inadequate aeration.

The lower value of Db before tillage in the WT treatment within CT, compared with the OWT treatment in NT, results from wetting and drying cycles in the previous season, which acts more efficiently (with higher changes) in compacted soils (Seguel and Horn 2006). Nevertheless, at the end of the season (post-harvest), there are no significant differences, because the effect of wetting and drying cycles depends on the weather conditions during the season (Dörner et al. 2009). Sivarajan et al. (2018) determined the highest values of mechanical strength after harvest and the lowest values after planting in the next season, because of cycles of freezing and thawing during winter time, but without differences between most trafficked and least trafficked rows, as was observed in our study, with differences in Db only before tillage for the wheel track treatment.

Martínez et al. (2008) compared Db under CT and NT in a wheat-maize rotation on the same study site, observing variations from 1.39 to 1.52 Mg m−3 on the first 15 cm of the soil, without differences between tillage systems, but with the highest aggregate stability in NT. This was depending on a higher root density and was also observed by Sawchik et al. (2012) for wheat. In our study, the maximum growth of roots is expected during anthesis (Martínez et al. 2008), which is in coincidence with the lowest Db value in the season. On the other hand, cereals are not sensitive to local compaction, and roots can grow independently of macroporosity levels (Athmann et al. 2019), contributing to renew the porosity system during the season.

In relation to water content, Mulumba and Lal (2008) highlighted the superficial effect of mulch in NT system, which reduces the evaporation from the surface, raising the water availability for the crops. This can be observed in NT, with a tendency to have higher water content compared with that in CT. The 24% of water content after sowing is close to field capacity, determined by the pressure chamber in a previous study of Martínez et al. (2008) in the same study site. In the previous characterization of Martínez et al. (2008), the permanent wilting point (WP) determined for this area was close to 12%, no matter the tillage system, because WP is a soil property which strongly depends on soil texture (Warrick 2002). It was possible to observe values close to field capacity (laboratory results of Martínez et al. 2008) after sowing, which decreased to values even lower than the WP after anthesis.

The low water content after the first knot can severely reduce the yields (Reynolds et al. 2008), a condition expected for regions where the water resource is limited (Kang et al. 2002; Brunel et al. 2013). In Mediterranean climates, where precipitations are concentrated on the initial stage of wheat, the water demand gets higher in spring. At that time, the water availability is decreasing, affecting the final yield of the crop (Fischer et al. 2014). Nevertheless, it should be noted that data were obtained from the first 5 cm of the topsoil layer, which does not represent the water content of the soil profile, because this can also be stored in lower layers and extracted by plants (Fischer et al. 2014). Martínez et al. (2011), studying tillage systems in an oat-wheat rotation in a sandy clay loam soil, found mean values of volumetric water content from 5 to 12% in the first 30 cm of the profile. Nonetheless, on every tillage system, the water content increases at higher depths (Barik et al. 2014), which reflects the storage function of the soil, especially under NT, but without the influence of the wheel track.

Finally, wetting and drying cycles promoted by the rainy winter helped to recover the soil structure (Seguel and Horn 2006) in both tillage systems. Anyway, the effect on the porosity function will be in equilibrium with the aggregate stability to watering (Dörner et al. 2009), affecting the hydraulic conductivity.

4.2 Unsaturated and Saturated Hydraulic Conductivity

It has been observed that fine-textured soils, if they are well structured with a high porosity and macropore continuity, can present high Kns values in near saturation conditions, similar to coarse-textured soils (Ugarte Nano et al. 2015). However, Kns at few to non-disturbed sites can reach low values, because of an abundance of coarse pores, drained at lower tensions (Warrick 2002). Before tillage, CT treatments showed lower values of Kns for all tensions compared with NT. This is explained by NT having little and well-defined soil aggregates compared with CT at the end of the previous season, where after structure alteration in CT, coarse units are regenerated (Dec et al. 2008). This implies more flux channels between aggregates in NT in comparison with that in CT (Hunt et al. 2013) at the end of a crop season. However, as tillage is effectively creating macroporosity, this is explaining the increments of Kns after tillage on CT (Centeno et al. 2020).

After sowing, lower values of Kns compared with the other time steps can be observed for all tensions and both tillage systems, whereas it was more evident with CT for both track positions and with NT for the WT position. This can be explained by the rupture of the structure by tillage and by the raindrop impact of the sprinkler irrigation as well as natural rains before measurements (splash effect, cf. Goebes et al. 2014), which seal the soil surface, increasing the risk of runoff and water erosion (Martínez-Mena et al. 2008; Centeno et al. 2020). Kayser et al. (2013) established that soil compaction in no-tillage systems is generated near the surface, because of the high mass of machinery for sowing transiting on wet soils. Even though superficial mulch prevents damage caused by rainfall (Jordán et al. 2010), this is not enough to prevent the pressure transmission to the ground.

As the crop grows, Kns starts recovering, first in CT which can be easily observed at the first-knot time step and then in NT, which can be seen during anthesis, where the highest values are observed in both cases. This can be explained by the natural structure regeneration through the wetting and drying cycles which restore pore continuity (Dec et al. 2008; Villarreal et al. 2020). Besides, the root activity of the crop increases, using more water from the first-knot timestep to anthesis because of the active growth (Fischer et al. 2014), drying the soil, and favoring the assemblage of solid particles. On the other hand, after harvest, roots die and their regeneration creates new and connected pore space (Beven and Germann 2013; Sivarajan et al. 2018), allowing to find similar values of Kns to those before tillage, at the beginning of the season.

According to Schoeneberger et al. (2012), there was a moderately high Ks at the study site during the whole season. The absence of tillage in the NT system generates a stable porosity, observed as higher values outside (OWT) compared with the cases inside (WT) the wheel track for the first three timesteps. However, initial rains generate a soil swelling for NT-OWT conditions, with a strong decrease of Ks in time (Alletto and Coquet 2009). This is not the main process in NT-WT, and the decrease in time is because of the mechanical stress, similar to the effect observed by Dörner et al. (2009) in an Andisol in southern Chile and Singh et al. (2014) in a wide kind of soils. This study showed that the mechanical stress hampers the swelling of the soil because of increasing contact points between the aggregates. Later in the season, at the first-knot timestep, Ks recovered because of the intensified drying, causing cracks with great continuity (Bhattacharyya et al. 2005).

For the CT system, in a general view, the machinery was effective in creating cracks and porosity (Athmann et al. 2019), with increasing Ks once the tillage was performed, for both, in and outside the wheel track. Values ranged from 1.22 to 1.83 cm h−1 and from 1.25 to 1.59 cm h−1, respectively (WT and OWT), decreasing to the lowest values after sowing because of the applied mechanical stress (Villarreal et al. 2010). The later recovery of Ks is a response to the wetting and drying cycles of the soil and the root activity of the crop, both crucial factors to recover soil porosity (Zúñiga et al. 2015). For much of the study period, the Ks values were superior in CT. However, before tillage and after harvest, the highest values were observed in NT. This is in agreement with other studies, like Fuentes et al. (2004), who found that saturated hydraulic conductivity in the soil was higher in NT during the period without wheat crop and decreasing once it was cultivated. Villarreal et al. (2020) concluded the same, but indicate that in soils with low clay content the differences between tillage systems disappear.

Alletto and Coquet (2009) found a gradual decrease of Ks through time, and in general lower values in the wheel track than outside, depending on variations of coarse porosity and bulk density, as was observed by Soares et al. (2020) at a large-scale approach. Wahl et al. (2004) found a correlation of the infiltration values with the macroporosity depending on the dynamic of CT in time, being the initial mechanical loosening and subsequent soil settlement the main factors during the season. For this study, Ks was negatively correlated with bulk density, as was observed by Pöhlitz et al. (2018) for different tillage systems and shown and interaction between tillage system and timestep, as was stated by Villarreal et al. (2020) as a consequence of climatic fluctuations during the season. Bormann and Klaassen (2008) explained that soil water dynamics depend mainly on the physical properties of the soil, but the increase in Ks would also be conditioned by the water demand of the plant. In our study, the lowest Db values and the highest Ks values were observed in anthesis measurement (Fig. 3), when the maximum root activity is expected. This contrasts with Alletto and Coquet (2009), who indicated that there should not be variations of Ks because intrinsic factors such as soil texture are more relevant than external ones, and they are expected to be constant in the short term. Nevertheless, in a large-scale analysis, Centeno et al. (2020) indicate that the changes in macroporosity are the main factor to explain the variations in Ks.

For NT, Ks was initially higher than CT, with a marked decrease at the second sampling, even though the soil was not disturbed in this period, reaching its minimum after sowing with the lowest values registered (0.49 to 1.01 cm h−1). The latter can be explained by the effect of the compaction made by the seeder machine. This can be confirmed by a previous study about the machinery mass increasing the penetration resistance at soil surface on the same study site (Martínez et al. 2008). At the end of the season, Ks values in NT stayed stable and obtained higher values than CT in post-harvest evaluation, which might be related to higher biological activity in NT systems (Chenu et al. 2019). This is generally favored in NT by higher organic matter content in the soil, as described, e.g., by Sanaullah et al. (2020).

Even though CT presents the highest values of Ks throughout the season (Table 6), the variability in both tillage systems is similar, with CV values close to 40%, which according to Warrick (2002) corresponds to a low level. Nevertheless, the absolute variation in CT is higher than in NT (Δ 2.07 and Δ 1.52 cm h−1, respectively). The interaction of the tillage system with time is definitely the most important factor to explain the dynamics of Ks and it is necessary to continue studying other important timesteps within the phenology of crops, to understand the water dynamics in soil and to optimize its use.

5 Conclusion

In this study, it was possible to show that machinery passage and tillage systems (conventional tillage (CT) and no tillage (NT)) affect the soil physical properties of a loamy soil subjected to a wheat-rapeseed crop rotation in a Mediterranean climate, in central Chile. In particular, the saturated hydraulic conductivity within the topsoil horizon showed to act as a controller of the fluxes into the profile.

Through the season, a low variability of the soil bulk density was observed, with no differences between the positions in and outside the wheel track or between tillage systems. The water content showed a temporal variation, with the highest values before sowing and with differences between tillage systems at first-knot timestep, where CT showed the lowest values. Nevertheless, bulk density and its variations during the season were the key factor to explain the changes in hydraulic conductivity.

Saturated hydraulic conductivity (Ks) did not show differences between wheel track positions for the CT system. In NT instead, the Ks values outside the wheel track were higher compared with the ones within the track until the period after sowing, maintaining similar values at the first-knot timestep and for the rest of the season. On the other hand, wetting and drying cycles and re-structuring processes explain the dynamic of Ks during the season, with the lowest values after sowing and a recovery during the anthesis, with values even higher than CT at the post-harvest timestep. These results highlight the effect of crop development and wetting and drying cycles during the season as key factors for Ks behavior in NT and CT systems.