Short-Term Impact of Conservation Agriculture on Soil Strength and Saturated Hydraulic Conductivity in the South African Semiarid Areas

: The severe limitation of agricultural land productivity induced by physical soil degradation has become a major concern in semiarid climates, especially in the Eastern Cape Province, South Africa. A randomized complete block design in a split-split-plot arrangement was used to evaluate the short-term (2012–2015) e ﬀ ects of tillage (no-till (NT) and conventional tillage (CT)), rotation (maize-fallow-maize (MFM); maize-fallow-soybean (MFS); maize-wheat-maize (MWM) and maize-wheat-soybean (MWS)) and residue management (residue removal (R − ) and residue retention (R + )) on bulk density (BD), penetration resistance (PR), soil hydraulic conductivity (Ks) and macroporosity hydraulic conductivity. The interaction of tillage × crop rotation × residue management was not signiﬁcant ( p > 0.05) with respect to BD, PR, Ks and macroporosity. The MFM rotation had the highest BD (1.40 g cm − 3 ), followed by MWM rotation (1.36 g cm − 3 ), and the least BD was observed in the MFS rotation (1.29 g cm − 3 ). Penetration resistance was signiﬁcantly higher in CT (2.43 MPa) compared to NT (1.46 MPa). The study concludes that inclusion of MFS and MWS rotations can potentially reduce BD in the short term. Similarly, conversion from CT to NT reduces soil resistance.


Introduction
South African soils are extremely fragile and have low resilience [1]. This is especially true in the Eastern Cape Province, where most soils are low in soil organic matter (SOM) (<10 g C kg −1 ) [2,3], which has contributed to lowering their productivity and continues to create high crop production challenges [4]. The smallholder farmer's unsustainable farming practices such as repeated tillage, maize monoculture and burning/removal of crop residues further reduce SOM levels, worsening the degradation of the fragile lands [3,5,6]. Soil degradation is linked to poverty as it reduces soil productivity and potential economic returns for the smallholder farmers. Identifying and developing suitable land-management strategies that can help smallholder farmers adapt and mitigate these challenges is therefore imperative. Conservation agriculture (CA) based on no-till (NT), viable crop rotations and permanent soil cover [7] aims to produce crops in a sustainable manner without any compromise on the current and future of soil productivity or resource availability. Adoption of the technology in the province is low, however, owing to lack of knowledge on the technology as well as socioeconomic and biophysical factors [8].
The perception amongst most smallholder farmers is that non-tilled soils are compacted, harder to work on and more resistant to root penetration and water infiltration than tilled soils [9]. Soil compaction is recognized as one of the major threats to soil quality [10]. In fact, Moraes et al. [11,12] reported that soil compaction is a major drawback to productivity in the farming sectors because of its ability to reduce porosity, which restricts root distribution and growth as well as inhibiting water and air transmission and circulation. The knowledge of effects of CA on soil compaction and hydraulic properties would help increase adoption of the technology and, potentially, improve both soil and crop productivity amongst the smallholder farmers in the Eastern Cape Province. Bulk density (BD) and soil penetration resistance (PR) are the most common variables used to measure soil compaction.
The ideal BD for plant growth varies for different soil and types of crops. A low BD results in poor soil-root interplay or contact, whereas a high BD decreases aeration and increases soil compaction [9]. Soils with high bulk density and low porosity have a negative influence on saturated hydraulic conductivity (Ks) and infiltration rate [13][14][15]. Mwendera and Feven [16] found that Ks can decrease directly as a result of an increase in soil compaction associated with reductions in soil macroporosity. Soil strength, which is usually measured as PR, is an essential parameter of soil structure. Soil strength is closely associated with increased bulk density and decreased soil porosity [17] as well as soil moisture content. Nevertheless, studies report that PR is highly sensitive to agronomic practices relative to BD [18]. Soil PR is an ideal soil parameter used for the investigation of the influence of different agronomic operations on soil strength and shows the resilience of roots to navigate the soil volume. According to Pias et al. [19], soil penetration resistance determines the level of soil layer restriction to root growth. The interaction among crop type, soil texture, SOM, mass wetness and particle and aggregate sizes influences the response of soil strength to various crop sequences [20]. Most studies on soil strength have been done in other areas [10,11]; a few have been done in the Eastern Cape Province in south Africa [21].
Various studies have reported the benefits of CA on soil compaction and hydraulic properties, but the results are inconsistent. For example, BD was found to be high under NT compared to CT [9,22,23], while no significant differences were observed [24][25][26], and reduced BD under NT compared to CT [27][28][29]. Similar inconsistences were also reported for PR. Ferreras et al. [30] observed a higher PR in NT at the top 10 cm of soil depth relative to CT. Rusu et al. [31] observed a higher PR in the NT at 20 cm of soil compared to the CT. The same authors reported that tillage effects on PR were because the PR determination is hinged on the state of soil settlement and its humidity.
The inconsistent observations on soil compaction are due to the fact that soil responses to management practices are also influenced by soil type, crop rotation, tillage types and climate [32]. Furthermore, studies that focused on the combined effect of conservation agriculture on soil compaction and soil hydraulic properties are still limited in the Eastern Cape Province context. Spatial variability of soil properties is inherent in nature due to geologic and pedologic soil forming factors, but some of the variability may be induced by tillage and other management practices [33]. Investigating the response of a physical properties of a Haplic Cambisol to management practices of smallholder farmers is vital to understand changes that occur at the local scale. According to Gulser et al. [33], all soil physical properties are weakly spatially dependent for the 6-9 cm soil depth, and moderately spatially dependent for the 27-30 cm soil depth. The study therefore focused on the 0-10 cm depth. The study can be used to apply to a wider scale using applications such as geostatistical tools. Therefore, the objective of this study was to examine the impact of different tillage, crop rotation and crop residue management systems on selected soil physical properties, such as soil BD, PR, soil hydraulic conductivity and macroporosity, on a Haplic Cambisol three years after CA inception in the Eastern Cape region of South Africa. The study hypothesized that in the short-term, the use of CA practices will decrease soil strength and increase soil hydraulic conductivity.

Site Selection
The study was carried out at the University of Fort Hare Research Farm (32 • 46 S and 26 • 50 E), located at 535 m above sea level. The climate is warm temperate semiarid. According to the South African classification system, the soil is in the Oakleaf soil form [34] and is a Haplic Cambisol according to International Union of Soil Sciences working group [35]. The field is on flat land. Initial soil tests indicated that the soils had 1.57% soil organic carbon and a BD of 1.53 g cm −3 . The soil type is clay-loam with a texture of 44%, 22% and 34% for sand, silt and clay, respectively. Prior to the establishment of the trial, the field was under lucerne (Medicago sativa).

Trial Layout and Treatments
The experiment was set up in October 2012, and sampling was done after harvesting maize in April 2015, i.e., after a 3-year period. The experiment had a split-split-plot experimental design replicated 3 times ( Table 1). The main plots were tillage (no-till (NT) and conventional tillage (CT)), the subplots were crop rotation (maize-fallow-maize (MFM), maize-fallow-soybean (MFS), maize-wheat-maize (MWM) and maize-wheat-soybean (MWS)), while the sub-subplots were crop residue management (residue removal (R−) and residue retention (R+)). The main plot area was 325 m 2 , while the subplots and sub-subplots measured 70 m 2 and 35 m 2 , respectively.

Trial Management
The agronomic practices were as described by Mtyobile et al. [26] and Gura and Mnkeni [36]. Before the establishment field was ploughed, it was disked to make a fine tilth to create uniform conditions before the initial crop establishment. Tillage in the CT plots was done at the beginning of the winter and summer seasons every year (Table 2). Ploughing was done to a depth of 20 cm of using a disk plough and subsequently harrowed to make a fine tilth at planting. The maize cultivar BG 5785BR and soybean cultivar PAN 5409RG were sown in the summer seasons, both targeting plant populations recommended for dry land crop production in the province, of 25,000 and 250,000 plants ha −1 , respectively [36]. An early maturing, dryland spring wheat cultivar (SST015) was planted in winter (May-August) at a seeding rate of 100 kg ha −1 . An inorganic fertiliser was only applied to the summer maize crop at a rate of 90 kg N, 45 kg P and 60 kg K per ha in all the plots. A compound basal fertilizer (6.7% N; 10% P; 13.3% K + 0.5% Zn) was applied at planting to supply all the K and P and a third of the N. The rest of the N was applied as lime ammonium nitrate (28), at 6 weeks after planting. The soybean crop was inoculated with Rhizobium leguminosarium before sowing. In the winter seasons, an early maturing wheat cultivar (SST 015) was planted at a seed rate of 100 kg ha −1 . At harvest, crop residues were left as surface covers in plots having the residue retention treatment only. All crops were rain-fed.

Measurements
Soil sampling for bulk density, porosity, saturated hydraulic conductivity and water-conducting macroporosity was done at the end of the 2014/15 summer season. Soil bulk density was determined using the core method as described by Okalebo et al. [37]. Undisturbed samples were collected from a depth of 10 cm by driving open-ended plastic core rings made of polyvinyl chloride pipe (11 cm in diameter and 7.7 cm in height) into the soil. Two soil samples were randomly taken in the middle of the net plot in each sub-subplot. The soil sample was dried at 105 • C for 48 h. Bulk density was then calculated as shown in Equation (1): where BD is the bulk density (g cm −3 ), m is the mass of oven-dried soil (g) and V is the volume of the core (m 3 ). The hammer-type dynamic cone penetrometer (DCP) was used to measure penetration resistance as described by Herrick and Jones [38] to assess the impact of the different tillage, crop rotation and crop residue management practices on soil strength. The hammer-type dynamic cone penetrometer was fitted with a 30 • hard steel cone with a 20.3 mm diameter shaft. The shaft had a striking plate attached on it. This striking plate forces the steel cone into the soil, while the shaft serves to guide the 2 kg hammer onto the plate. A predetermined drop height of the hammer to ensure uniformity and repeatability was achieved through an adjustable collar. The penetrometer cone was placed onto the soil surface with the shaft in a vertical position. The cone base was made to be at the same level with the soil surface while in contact with soil. The required number of blows to reach a penetrating depth of 0-0.15 m was recorded. Every strike had an equivalent kinetic energy of 12.74 joules. Five measurements per plot (sub-subplot) were done.
Macropores in this study were considered to be pores which would be emptied at a suction head of 3 cm with an equivalent pore radius (r) larger than 0.5 mm, following Watson and Luxmoore [39]. Briefly, a 100 mm mini-graduated ring was placed on top of the undisturbed core sample. The ring was slightly pressed into the core to allow water conduction, at the same time minimising soil disturbance. The ring was ponded to approximately 50 mm with water. Macroporosity was calculated using the Watson and Luxmoore [39] approach from the difference between the ponded (double ring) infiltration rate and the infiltration rate at a tension of 3 cm. The theory behind this approach is that the capillary rise equation (Equation (2)) can be used to calculate the maximum pore size ((r) in cm)) that is filled with water at a certain suction head ((h) in cm) [40]: where γ is the surface tension of water (M T −2 ), ϑ the contact angle between the water and pore wall (assumed to be 0), ρ the density of water (M L −3 ) and g the gravitational force (L T −2 ). Equation (1) predicts that all pores with a radius smaller than 0.5 mm will remain water filled at a suction (h) of 3 cm and are therefore responsible for water flux at that suction. Under these conditions, pores with a radius larger than 0.5 mm will not be conducting water.
Saturated hydraulic conductivity was measured using a tension infiltrometer under laboratory conditions to assess the effects of the different tillage, crop rotation and crop residue management practices. For this experiment, undisturbed soil samples were collected using core samplers as described for soil bulk density. The core samples were placed on the top of a gauze and were saturated with water. The calculation of saturated hydraulic conductivity (mm h −1 ) was done using the modified falling head equation shown in Equation (3) [41]: where K S is the hydraulic conductivity, L the length of soil column (L); h 0 the height of water above surface at time 0, h 1 the height of water above surface at the end and t the time until the water level changed from h 0 to h 1 .

Data Analysis
A three-way analysis of variance (ANOVA) was performed on all variables using JMP version 14 (SAS Institute, Inc., Cary, NC, USA). Significant differences among treatments of the measured selected soil physical properties that included bulk density, porosity, penetration resistance, saturated hydraulic conductivity, texture and water-conducting macroporosity were identified at p < 0.05.

Soil Bulk Density
Three-way interactions of tillage × crop rotation × residue management were not significant (p > 0.05) with respect to bulk density (Table 3). Similarly, the two-way interactions of tillage × crop rotation, tillage × residue management and crop rotation × residue management had no significant effect (p > 0.05) on bulk density in the study. Only the main effects of crop rotation influenced BD (p < 0.01) in the short-term study. The rotation of MFM (1.40 g cm −3 ) had a significantly higher BD relative to MFS (1.29 g cm −3 ) (Figure 1). Table 3. p-values of three-way ANOVAs evaluating the effect of tillage (T), residue management (R) and crop rotation (CR) on bulk density (BD), soil resistance (PR), hydraulic conductivity (Ks) and macroporosity after 3 years of cropping in Haplic Cambisol in Alice, Eastern Cape. The descriptive statistics of BD, PR, and Ks are shown in Table 4.

Penetration Resistance
The interaction of tillage × crop rotation × residue management had no effect (p > 0.05) on the PR (Table 1). Likewise, the interactions of tillage × crop rotation, tillage × residue management and crop rotation × residue management were not significant (p > 0.05) with respect to PR (Table 1). Tillage main effects were significant (p < 0.001), while crop rotation and crop residue management were not significant (p > 0.05) with respect to PR. Penetration resistance was 40% higher under CT compared to NT (Figure 2).

Soil Saturated Hydraulic Conductivity and Water-Conducting Macroporosity
There were no three-way interaction effects (p > 0.05) on either soil Ks or water-conducting

Penetration Resistance
The interaction of tillage × crop rotation × residue management had no effect (p > 0.05) on the PR (Table 1). Likewise, the interactions of tillage × crop rotation, tillage × residue management and crop rotation × residue management were not significant (p > 0.05) with respect to PR (Table 1). Tillage main effects were significant (p < 0.001), while crop rotation and crop residue management were not significant (p > 0.05) with respect to PR. Penetration resistance was 40% higher under CT compared to NT (Figure 2).

Penetration Resistance
The interaction of tillage × crop rotation × residue management had no effect (p > 0.05) on the PR (Table 1). Likewise, the interactions of tillage × crop rotation, tillage × residue management and crop rotation × residue management were not significant (p > 0.05) with respect to PR (Table 1). Tillage main effects were significant (p < 0.001), while crop rotation and crop residue management were not significant (p > 0.05) with respect to PR. Penetration resistance was 40% higher under CT compared to NT (Figure 2).

Soil Saturated Hydraulic Conductivity and Water-Conducting Macroporosity
There were no three-way interaction effects (p > 0.05) on either soil Ks or water-conducting macroporosity. Two-way interactions of tillage × crop rotation, tillage × residue management and

Soil Saturated Hydraulic Conductivity and Water-Conducting Macroporosity
There were no three-way interaction effects (p > 0.05) on either soil Ks or water-conducting macroporosity. Two-way interactions of tillage × crop rotation, tillage × residue management and crop rotation × residue management had no effect (p > 0.05) on soil hydraulic conductivity and water-conducting macroporosity. Although hydraulic conductivity was higher under NT (58.4 mm h −1 ) compared to CT (32.8 mm h −1 ), the difference between the tillage treatments was not statistically significant ( Figure 3).

Soil Bulk Density
Low BD under MFS relative to MFM rotation could be explained by the rooting system and the differences in decomposition of soybean and maize crop residues. Kavdir et al. [42] reported that the presence of the soybean in a rotation system has the potential to create macropores after decomposition and hence decrease BD. Bulk density is one of the sensitive soil quality indicators which assists to assess the situation with regard to the ease of root penetration, water movement and soil strength [43,44]. A soil BD exceeding 1.3 Mg m −3 in a silty soil could adversely disrupt soil aeration through the reduction of air-filled pore spaces [45]. Lower BD under MFS compared to MWM and MWS can possibility be attributed to greater aggregate stability due to greater residue accumulation from maize and wheat compared to soybean. The results from this study indicated no significant differences (p > 0.05) in BD between NT (1.42 Mg m −3 ) and CT (1.34 Mg m −3 ). The results of the study show that changes in BD in the Eastern Cape Province are visible in the short term (after 3 years), but are likely to occur only under different crop rotations. These findings were similar with other studies which established that BD requires a long period to be impacted by conservation soil management practices. For example, Gura and Mkeni [36] observed no significant difference between NT and CT under similar environmental conditions. Similarly, Ferreras et al. [30] reported no significant difference in BD between tillage systems in a study done on moderately well-drained Chernozemic loam soils in the southern Pampas of Argentina. Soil BD changes take time, but not with a consistent tendency [46].

Penetration Resistance
The measurement of soil PR ensures a quick and easy way of determining soil strength. Soil strength on a larger scale determines the index and momentum of soil productivity [47] and other soil hydrological processes [48]. According to the Cornell Soil Health Manual [49], soil with PR of 2.07 MPa inhibits good root growth. The higher PR in this under CT relative to NT in this study could be attributed to soil crusting. A soil PR range of 2.0-3.5 MPa was found to be critical [50]. The findings

Soil Bulk Density
Low BD under MFS relative to MFM rotation could be explained by the rooting system and the differences in decomposition of soybean and maize crop residues. Kavdir et al. [42] reported that the presence of the soybean in a rotation system has the potential to create macropores after decomposition and hence decrease BD. Bulk density is one of the sensitive soil quality indicators which assists to assess the situation with regard to the ease of root penetration, water movement and soil strength [43,44]. A soil BD exceeding 1.3 Mg m −3 in a silty soil could adversely disrupt soil aeration through the reduction of air-filled pore spaces [45]. Lower BD under MFS compared to MWM and MWS can possibility be attributed to greater aggregate stability due to greater residue accumulation from maize and wheat compared to soybean. The results from this study indicated no significant differences (p > 0.05) in BD between NT (1.42 Mg m −3 ) and CT (1.34 Mg m −3 ). The results of the study show that changes in BD in the Eastern Cape Province are visible in the short term (after 3 years), but are likely to occur only under different crop rotations. These findings were similar with other studies which established that BD requires a long period to be impacted by conservation soil management practices. For example, Gura and Mkeni [36] observed no significant difference between NT and CT under similar environmental conditions. Similarly, Ferreras et al. [30] reported no significant difference in BD between tillage systems in a study done on moderately well-drained Chernozemic loam soils in the southern Pampas of Argentina. Soil BD changes take time, but not with a consistent tendency [46].

Penetration Resistance
The measurement of soil PR ensures a quick and easy way of determining soil strength. Soil strength on a larger scale determines the index and momentum of soil productivity [47] and other soil hydrological processes [48]. According to the Cornell Soil Health Manual [49], soil with PR of 2.07 MPa inhibits good root growth. The higher PR in this under CT relative to NT in this study could be attributed to soil crusting. A soil PR range of 2.0-3.5 MPa was found to be critical [50]. The findings from this short-term study were inconsistent with studies done elsewhere. Findings by Wilkins et al. [51] asserted that in a long-term study on well-drained loess (Walla Walla silt loam) soils, NT has a higher compaction than CT. In a study done on Albic Luvisols developed on loamy sands, Malecka et al. [52] reported significantly higher PR under NT (1.56 MPa) compared to tilling practices (1.19 MPa) at the 0-0.10 m depth. However, Castellini et al. [53] observed that surface soil compaction can actually occur during springtime regardless of the tillage system adopted in the past season.

Soil Saturated Hydraulic Conductivity and Water-Conducting Macroporosity
The lack of differences in soil hydraulic properties could be because of the short duration of the CA experimental study. More so, it has been observed that under a continuous NT system, it takes about five years for soil properties to fully stabilized [54,55]. However, numerous factors such as soil type, climate and machinery as well as crop types have been found to influence major changes in soil physical properties on a short-term basis [56]. Research has also found contradictory trends without a definite pattern in the effects of CA treatments on saturated hydraulic conductivity [56]. According to Kodesova et al. [57] and Kargas et al. [58], the higher average saturated hydraulic conductivity under NT is because of a good soil structure, and higher total porosity increased SOM and plant root influences. Moreno et al. [59] posited that the presence of preferential flow paths due to earthworm activities was the reason for higher saturated hydraulic conductivity under NT compared to CT. The biological actions of earthworms affect soil structure and influence soil properties such as porosity and water content [60]. However, Six et al. [61] reported that lower saturated hydraulic conductivity in CT was due to mechanical breakdown of aggregates during tillage, leading to structural degradation in the CT plots. Fuentes [62] and Hu et al. [56] noted that hydraulic properties are prone to temporal surface changes, which could sometimes be a result of the tillage operations and practices employed. In contrast, in an 8-year-long study, Raczkowski et al. [63] found out that a no-till system resulted in lower macroporosity and higher microporosity as compared to conservation tillage. Drees et al. [64] also observed less macroporosity on a silty clay loam under NT for 17 years than under convention tillage. They were able to observe significant differences among the tillage treatments because of the long duration of their experiments. In another experiment, Zhang et al. [65] found, after 24 years, significantly greater macropores (more than 11%) under no-till with residue retention than under conventional tillage with residue burnt. In the 0-3 cm layer of a silt loam soil, Blanco-Canqui and Lal [66] found that the macropores in the treatments where residues was retained were about twice the volume in the treatments were the residues were removed. This study, therefore, demonstrated that crop rotations have little impact on the macroporosity on the short-term basis (3 years).
The lack of significant interactions (p < 0.05) with respect to PR and bulk density and soil hydraulic properties reiterates that changes in soil physical properties require time. Changes in soil physical properties are not immediate, and when they change, this is not in a consistent tendency. Seasonal changes in soil properties can be due to factors such as volume and intensity of rainfall, drying and wetting of soil, land position and crop type, among others [67]. Furthermore, the interannual and intra-annual changes in properties such as BD are determined by tillage-induced loosening and soil aggregate fragmentation, soil resettling and aggregation under the impact of biological activity including root growth and rainfall [68]. The scope of this study did not cover seasonal variations, only changes after 3 years.

Conclusions
The study of the effects of CA components on BD and PR is very important because of the dynamics of these parameters in influencing farming, crop production and hydrological processes. Lack of significant effects suggests that soil compaction and hydraulic properties in the Eastern Cape Province are not affected in the short term (after 3 years). The study also concludes that inclusion of MFS and MWS rotations can potentially reduce BD in the short term compared to the traditional MFM cropping system. Conversion from CT to NT reduced soil resistance; therefore, no-till can possibly ensure good soil conditions in the Eastern Cape Province.
Author Contributions: G.I.N. designed and set up the experiment, collected data, analyzed data and contributed to writing the manuscript. A.M. contributed to writing and reviewing the manuscript. T.A. contributed to writing and reviewing the manuscript. J.v.T. designed and set up the experiment, collected data, analyzed data and contributed to writing the manuscript. All authors have read and agreed to the published version of this manuscript.