The effect of crop rotation on soil health in the north-western Free State region, South Africa

Abstract

Soil health is an important aspect of sustainable agriculture. By maintaining soil health, crop yields can be consistently successful without negatively affecting the environment. The objective of this study was to evaluate the effects that different cropping systems have on soil health in the north-western Free State region. The cropping systems focused on maize production, with soybean and cover crop included as rotational crops and monitored for three consecutive seasons. Soil samples were taken seasonally and their physical, chemical and biological properties, as well as the subsequent soil health score, were measured using the Haney Soil Health Test (HSHT). Results from the HSHT were statistically analysed and showed that there was no statistically significant interaction effect between cropping systems and season on soil properties, nor on the subsequent soil health score. These findings suggest that soybean and cover crop rotations with maize have the potential to naturally maintain soil health in a condition similar to those subjected to an artificially fertilised monoculture of maize. In addition, season variation influenced soil health irrespective of cropping system. Therefore, including soybean and cover crops in cropping systems with maize can be a means of promoting sustainable agriculture in the north-western Free State region, when seasonal variation and inherent soil properties are considered.

Keywords:

• crop rotation
• north-western Free State
• soil health
• sustainable agriculture

The concept of soil health emerged in the 1990s and has seen regained interest with the recent emphasis on sustainable agriculture (Lehmann et al. 2020; Magdoff and Van Es 2021). The definition of “the continued capacity of a soil to function as a vital living ecosystem that sustains plants, animals, and humans” stresses its importance in our life support system (Ribeiro et al. 2018; Lehmann et al. 2020; Norris et al. 2020; Strauss et al. 2021). Soil health is promoted when all three of its components (physical, chemical, and biological) are optimal (Magdoff and Van Es 2021). Soil is physically healthy when sufficient volumetric aggregate stability allows water infiltration, limits surface runoff and provides adequate aeration (Cardoso et al. 2013). The volumetric aggregate stability, which encapsulates organic matter (OM), should be more than 45% (Sundermeier et al. 2017), while there should be at least 3% OM in soil (Magdoff and Van Es 2021). Soil must be able to sufficiently absorb, release and transform essential nutrients including nitrogen (N), phosphorous (P) and potassium (K). In order to obtain ideal nutrient measurements, the pH of the soil should be slightly acidic (5.50–6.50) (Morgan et al. 2005; Hinsinger et al. 2009; Opala et al. 2018). Furthermore, if the soil is healthy, biodiversity is increased and beneficial organisms are active, which stimulates plant growth (Magdoff and Van Es 2021). The greater the microbial respiration (measured as CO₂-C) soils produce, the more abundant and active is the microbial life within them (Magdoff and Van Es 2021).

The north-western Free State region in South Africa is a semi-arid region that is dominated by extremely sandy soils (Beukes et al. 2019; Strauss et al. 2021). Only about 1–2% of the soil is made up of silt, while the clay content is normally less than 10% in the A-horizon and less than 15% in the B-horizon (Nortjè and Laker 2021). These sandy soils are characterised by low nutrient and fertility status and are vulnerable to erosion and compaction (Beukes et al. 2019; Nortjé and Laker 2021; Strauss et al. 2021). Wind erosion is especially problematic during August and September, removing topsoil and nutrients (Strauss et al. 2021). Cultivation leads to compaction and destruction of soil aggregates, which further contributes to low nutrient status and reduced water content for crop production (Magdoff and Van Es 2021; Nortjè and Laker 2021). It is therefore important to use agricultural management practices that conserve or improve soil health.

Maize (Zea mays L.) production is dominant in the area and is normally grown as a monoculture system (Loubser and Nel 2004; Nortjè and Laker 2021). However, the combination of maize with soybean (Glycine max (L.) Merr.), a leguminous crop, has been recognised for its commercial potential and its ability to fix atmospheric N in symbiosis with Rhizobium bacteria (Coskan and Dogan 2011; Acevedo-Siaca and Goldsmith 2020). It has been suggested that including a cover crop mixture as a third crop improves the cropping system’s productivity (Magdoff and Van Es 2021; Smit et al. 2021). Since the success of cropping systems appears to be site-specific (Nortjè and Laker 2021), the key is to find cropping systems ideal for the north-western Free State region. As part of identifying ideal cropping systems, the objective of this study was to monitor soil health and its properties in different cropping systems.

The study was conducted on the farm Christinasrus in the district of Wesselsbron in the north-western Free State region. A trial comparing different cropping systems to monoculture maize (MM) was established. Cropping systems included Maize–Maize–Soybean (MMS), Maize–Soybean (MS) and Maize–Cover crop–Soybean (MCS). The MMS system was split into MMS1 and MMS2 to distinguish between the first (MMS1) and second (MMS2) season of maize. The cover crop mixture was made up of 60% grasses (sorghum and pearl millet) and 40% legumes (Dolichos and cowpeas).

A randomised complete block design with three replicates was used for the trial layout (Table 1). There were 27 plots in total, each being 80.0 m × 24.4 m in size. Cropping systems were assigned to plots and each crop within each system, representing a different stage, was assigned to a plot to be able to distinguish between seasonal and rotational effects. The trial was monitored for three consecutive seasons: 2020/2021, 2021/2022 and 2022/2023.

Table 1: Plot design of trials on the farm Christinasrus for three consecutive seasons

Plot number	Rotational system	Crops planted
		Season 1 (2020/2021)	eason 2 (2021/2022)	Season 3 (2022/2023)
Replicate 1
Plot 1	MMS	Maize	Maize	Soybean
Plot 2	MS	Maize	Soybean	Maize
Plot 3	MM	Maize	Maize	Maize
Plot 4	MCS	Maize	Cover crop	Soybean
Plot 5	MCS	Cover crop	Soybean	Maize
Plot 6	MMS	Soybean	Maize	Maize
Plot 7	MS	Soybean	Maize	Soybean
Plot 8	MCS	Soybean	Cover crop	Maize
Plot 9	MMS	Maize	Soybean	Maize
Replicate 2
Plot 10	MCS	Maize	Cover crop	Soybean
Plot 11	MS	Soybean	Maize	Soybean
Plot 12	MCS	Cover crop	Soybean	Maize
Plot 13	MMS	Maize	Maize	Soybean
Plot 14	MCS	Soybean	Maize	Cover crop
Plot 15	MM	Maize	Maize	Maize
Plot 16	MMS	Maize	Soybean	Maize
Plot 17	MS	Maize	Soybean	Maize
Plot 18	MMS	Soybean	Maize	Maize
Replicate 3
Plot 19	MM	Maize	Maize	Maize
Plot 20	MMS	Soybean	Maize	Maize
Plot 21	MCS	Maize	Cover crop	Soybean
Plot 22	MS	Soybean	Maize	Soybean
Plot 23	MMS	Maize	Maize	Soybean
Plot 24	MMS	Maize	Soybean	Maize
Plot 25	MS	Maize	Soybean	Maize
Plot 26	MCS	Cover crop	Soybean	Maize
Plot 27	MCS	Soybean	Maize	Cover crop

The trial plots were prepared through the practice of rip-on-row (as per the farmer's preference) at a depth of 0.75 m in August or September of each season. Trials were planted in December, after pre-plant fertilisation. Pre-plant fertiliser consisted of 100-140 kg urea ha⁻¹. Maize plots were fertilised again at planting and top-dressed. This resulted in the addition of 116–135 kg ha⁻¹ N, 12–16 kg ha⁻¹ P and 5–9 kg ha⁻¹ K per season. Soybean and cover crop received no additional fertiliser. Maize and soybean plots were sprayed with round-up (glyphosphate) for the control of weeds. All field actions were carried out with commercial equipment. No biomass was removed or utilised by any farm animals.

Rainfall was measured manually using a rain meter at the trial location and the distribution is shown in Figure 1. A favourable rainfall season was experienced during Season 1, with a total of 689 mm measured at the trial location from September 2020 to May 2021. Season 2 (2021/2022) had a very wet start compared to the previous season, with 309 mm of rain measured in December 2021 alone (seasonal total of 922 mm). In Season 3 there was more widespread rainfall with 700 mm measured between September 2022 and May 2023.

Figure 1: Rainfall distribution over Christinasrus, Season 1 (2020/2021, Season 2 (2021/2022) and Season 3 (2022/2023), respectively

Soil samples were collected from all maize plots in the trial during each season. The samples were taken approximately 100 days after planting and collected at a depth of 0–150 mm. Three samples were taken randomly within each maize plot, 200 mm from the plant row, and combined to form a composite sample. Soil properties of 15 composite samples were measured and integrated according to the Haney Soil Health Test (HSHT) to obtain a soil health score for each season (Ward Laboratories Inc. 2019; Regen AgLab 2021) (Table 2).

Table 2: Soil properties measured and integrated to obtain an overall soil health score using HSHT

Soil pH (H2O)
Physical properties	Volumetric stability
	Organic matter (OM)
Chemical properties	Total water extractable N (WEN)		Soil health score
	Inorganic N (NO3-N and NH4-N)
	Organic N (WEON)
	Total P
	Total K
Biological properties	Soil respiration CO2-C
	Water extractable organic C (WEOC)

The HSHT takes a different approach to traditional methods of evaluating soil health by focusing on soil as a dynamic living system, rather than just physical or chemical processes. (Haney et al. 2018). Haney et al. (2018) highlights the importance of including the biological aspect into the soil health score. Therefore, the HSHT not only estimates plant available N, P, and K, but also provides an indication of soil health with respect to nutrient and C cycling (Haney et al. 2018).

The HSHT uses an H3A (Haney, Haney, Hossner and Arnold) extractant to measure plant available nutrients including N (NH₄–N, NO₃–N), P and K (Chu et al. 2019; Singh et al. 2020). This extractant is made up of organic acids (citric acid, oxalic acid and acetic acid), which mimic plant root exudates better than other extractants (Haney et al. 2018; Singh et al. 2020). Since the majority of nutrient cycling is due to a natural drying – rewetting effect (rainfall and irrigation), water extracts are also investigated to better represent the natural environment (Haney et al. 2018). The water extractable organic N (WEON) is the active organic soil N (Haney et al. 2018). The water extractable C (WEOC) reflects the quality of soil organic C as it provides the energy source for soil microbial activity (Haney et al. 2018). Therefore, WEOC estimates the active pool of soil C (Haney et al. 2018). This HSHT approach considers the WEOC : WEON ratio to be a more sensitive indicator of microbial activity than the traditional total C:N ratio (Chu et al. 2019).

An important factor in the HSHT is the microbial respiration, measured as CO₂-C (Chu et al. 2019). The CO₂-C released by soil microbes in 24 hours after the soil has been dried and rewetted is an indicator of soil respiration, and consequently microbial activity (Haney et al. 2018; Singh et al. 2020). The properties measured by the HSHT are integrated using Haney’s formula to obtain an overall soil health score as related to nutrient and C cycling. (Haney et al. 2018; Ward Laboratories Inc. 2019). The resulting score can range between 0 and 50, with a higher score meaning better soil health (Haney et al. 2018). Most soils do not score more than 30, while scores below seven are regarded as low (Singh et al. 2020).

Data from the HSHT was cleaned and prepared for Statistical Package for the Social Sciences (SPSS) version 29 (IBM 2023), where it was further analysed using descriptive and inferential statistics. Descriptive statistics included measures of central tendency and were represented as mean ± standard deviation. Inferential statistics included two-way analysis of variance (ANOVA), which were run to determine if there was a statistically significant interaction effect of cropping system and season on individual soil properties as well as overall soil health. Assumption testing, which included testing for outliers, normal distribution, and homogeneity of variances, was conducted before the ANOVA was run. Statistical significance was accepted at p ≤ 0.05.

Table 3 illustrates the mean measurements and ANOVA results of soil properties and subsequent soil health score for each cropping system in each season as determined by the HSHT. Many soil properties and the subsequent soil health score were affected by season but not cropping system, with no interaction between these factors. Although this is not what was expected, the fact that cropping systems did not affect soil health is not an isolated case. Studies by Bavougian et al. (2019) and Roper et al. (2017) also found that soil health scores were not sensitive to management changes, including crop rotations. However, the fact that the monoculture maize received more fertiliser over time than maize rotated with soybean and cover crops, without significantly affecting soil health, could be an indication that by incorporating these crops, nutrient levels can be maintained naturally. Soybean is able to fix atmospheric N in symbiosis with Rhizobium bacteria (Coskan and Dogan 2011; Acevedo-Siaca and Goldsmith 2020). Similarly, the legumes in the cover crop are also able to act as a source of nitrogen (Fageria et al. 2005).

Table 3: Mean measurements ± standard deviation of three replicates, ANOVA results of soil properties, and subsequent soil health score for each cropping system in each season

Soil properties		Cropping system		Season 1 (2020/2021)	Season 2 (2021/2022)	Season 3 (2022/2023)	Mean	ANOVA results
								Cropping system		Season		Cropping system × Season
								F	p-value	F	p-value	F	p-value
Soil pH (H2O)		MM		6.04 ± 0.10	7.13 ± 0.29	7.43 ± 0.15	6.87 ± 0.66	0.57	0.69	91.96	<0.001	0.80	0.61
		MMS2		6.16 ± 0.10	7.23 ± 0.10	7.83 ± 0.29	7.07 ± 0.78
		MMS1		5.95 ± 0.08	7.43 ± 0.60	7.53 ± 0.29	6.97 ± 0.84
		MS		5.97 ± 0.20	7.03 ± 0.55	7.70 ± 0.36	6.90 ± 0.83
		MCS		6.22 ± 0.27	6.93 ± 0.42	7.60 ± 0.10	6.92 ± 0.65
			Mean	6.07 ± 0.18	7.15 ± 0.43	7.62 ± 0.26
Physical properties	Volumetric aggregate stability:	All systems		2%	2%	2%	2%
	OM (%)	MM		0.50 ± 0.17	0.37 ± 0.12	0.37 ± 0.12	0.41 ± 0.14	0.63	0.65	6.36	0.005	0.67	0.72
		MMS2		0.47 ± 0.06	0.37 ± 0.06	0.40 ± 0.00	0.41 ± 0.06
		MMS1		0.43 ± 0.12	0.40 ± 0.00	0.33 ± 0.06	0.39 ± 0.08
		MS		0.40 ± 0.00	0.30 ± 0.10	0.37 ± 0.12	0.36 ± 0.09
		MCS		0.47 ± 0.06	0.27 ± 0.06	0.40 ± 0.10	0.38 ± 0.11
			Mean	0.45 ± 0.10	0.34 ± 0.08	0.37 ± 0.08
Chemical properties	WEN (mg kg^-1)	MM		11.83 ± 4.21	10.80 ± 0.44	10.53 ± 1.04	11.06 ± 2.26	0.73	0.58	0.75	0.48	0.84	0.58
		MMS2		11.80 ± 0.53	10.37 ± 0.37	13.87 ± 4.27	12.01 ± 2.60
		MMS1		12.70 ± 4.27	11.80 ± 0.87	9.63 ± 1.25	11.38 ± 2.65
		MS		10.23 ± 2.54	9.97 ± 0.87	10.80 ± 1.15	10.33 ± 1.51
		MCS		11.87 ± 3.67	10.67 ± 0.68	13.17 ± 2.18	11.90 ± 2.42
			Mean	11.69 ± 2.95	10.72 ± 0.85	11.60 ± 2.59
	Inorganic N (mg kg^-1)	MM		5.83 ± 1.78	5.26 ± 0.73	3.43 ± 0.25	4.84 ± 1.46	0.57	0.69	1.51	0.24	1.15	0.36
		MMS2		4.63 ± 0.81	5.77 ± 0.44	5.90 ± 4.11	5.43 ± 2.19
		MMS1		5.57 ± 1.63	5.96 ± 0.90	4.10 ± 0.36	5.21 ± 1.27
		MS		4.60 ± 0.79	5.38 ± 1.11	3.37 ± 0.45	4.45 ± 1.14
		MCS		5.47 ± 1.85	4.27 ± 0.80	5.50 ± 1.49	5.08 ± 1.39
			Mean	5.22 ± 1.33	5.33 ± 0.92	4.46 ± 1.99
	WEON (mg kg^-1)	MM		8.23 ± 2.16	5.43 ± 0.15	7.10 ± 0.92	6.92 ± 1.69	0.83	0.52	22.63	< 0.001	0.94	0.50
		MMS2		8.60 ± 0.92	4.57 ± 0.55	8.10 ± 0.44	7.09 ± 2.00
		MMS1		9.70 ± 2.38	5.80 ± 1.65	6.07 ± 1.16	7.19 ± 2.45
		MS		7.37 ± 1.75	4.03 ± 0.61	7.23 ± 1.31	6.21 ± 1.99
		MCS		8.77 ± 2.83	5.50 ± 0.26	7.63 ± 0.81	7.30 ± 2.06
			Mean	8.53 ± 1.95	5.10 ± 0.98	7.23 ± 1.08
Chemical properties (contd)	Total P (mg kg^-1)	MM		32.00 ± 8.54	35.37 ± 7.86	24.00 ± 7.81	30.46 ± 8.89	0.26	0.90	4.60	< 0.001	1.69	0.14
		MMS2		35.33 ± 2.08	33.71 ± 5.86	23.67 ± 6.66	30.90 ± 7.12
		MMS1		42.00 ± 5.57	31.52 ± 1.03	25.00 ± 3.61	32.84 ± 8.15
		MS		31.33 ± 5.69	31.00 ± 7.77	31.00 ± 8.54	31.11 ± 6.44
		MCS		28.33 ± 3.22	29.89 ± 4.34	32.00 ± 7.81	30.07 ± 5.01
			Mean	33.80 ± 6.89	32.30 ± 5.42	27.13 ± 7.07
	Total K (mg kg^-1)	MM		93.00 ± 34.77	80.71 ± 4.82	104.00 ± 29.46	92.57 ± 25.04	1.14	0.36	1.36	0.27	0.88	0.55
		MMS2		82.67 ± 30.75	83.37 ± 17.66	102.67 ± 30.73	89.57 ± 25.44
		MMS1		83.67 ± 25.33	75.24 ± 9.08	71.33 ± 12.86	76.75 ± 15.88
		MS		62.00 ± 7.55	87.69 ± 15.58	81.67 ± 20.82	77.12 ± 17.85
		MCS		85.00 ± 19.08	61.28 ± 9.67	90.33 ± 17.47	78.87 ± 19.24
			Mean	81.27 ± 23.94	77.66 ± 14.03	90.00 ± 25.54
Biological properties	Soil respiration (mg kg^-1 CO2-C)	MM		21.26 ± 1.75	16.67 ± 2.89	11.33 ± 1.53	16.42 ± 4.68	0.45	0.77	77.99	< 0.001	1.30	0.28
		MMS2		19.63 ± 3.63	13.67 ± 1.15	13.00 ± 0.00	15.43 ± 3.69
		MMS1		21.37 ± 3.80	13.33 ± 3.06	10.67 ± 0.58	15.12 ± 5.41
		MS		21.28 ± 0.24	14.33 ± 2.31	11.00 ± 1.00	15.54 ± 4.71
		MCS		22.24 ± 2.91	12.00 ± 1.00	12.33 ± 1.53	15.52 ± 5.32
			Mean	21.16 ± 2.52	14.00 ± 2.48	11.67 ± 1.29
	WEOC (mg kg^-1)	MM		84.33 ± 17.50	56.00 ± 9.85	87.00 ± 17.52	75.78 ± 19.97	0.03	0.99	31.79	< 0.001	1.26	0.30
		MMS2		86.33 ± 9.87	50.00 ± 4.00	89.33 ± 13.58	75.22 ± 20.83
		MMS1		98.33 ± 16.07	58.33 ± 5.51	68.67 ± 15.95	75.11 ± 21.43
		MS		80.33 ± 15.31	56.00 ± 9.64	87.00 ± 14.80	74.44 ± 18.34
		MCS		89.33 ± 16.86	46.33 ± 4.73	86.00 ± 9.64	73.89 ± 23.00
			Mean	87.73 ± 14.42	53.33 ± 7.63	83.60 ± 14.56
Overall soil health		MM		4.64 ± 0.51	3.33 ± 0.49	3.60 ± 0.26	3.86 ± 0.71	0.17	0.95	48.14	< 0.001	1.27	0.30
		MMS2		4.55 ± 0.63	2.83 ± 0.15	3.90 ± 0.26	3.76 ± 0.83
		MMS1		5.07 ± 0.93	3.07 ± 0.45	3.07 ± 0.45	3.74 ± 1.15
		MS		4.47 ± 0.46	2.97 ± 0.06	3.57 ± 0.46	3.67 ± 0.73
		MCS		4.89 ± 0.91	2.70 ± 0.00	3.77 ± 0.31	3.78 ± 1.06
			Mean	4.72 ± 0.65	2.98 ± 0.34	3.58 ± 0.42

The role that seasonal variation (mainly rainfall) plays on soil health was evident throughout the study. Many studies show that there is a seasonal effect on soil health (Mitchell et al. 2017; Babur and Dindaroglu 2020; Sharma et al. 2020; Sherbine et al. 2023). With the prediction of more extreme weather in the future (Chemura et al. 2022) it is important to take note of these effects. Three different significant seasonal trends were identified in this study. The first was an increase in measured values over the duration of the study. This was seen with soil pH (H₂O) and could be due to the application of lime before the onset of the trial, which is one of the most common practices to reduce soil acidity (Li et al. 2019), as well as salts in the topsoil being diluted due to the wet conditions (Rengel 2011).

The second trend identified was a decrease in measured values of total P and CO₂-C over the study period (Figure 2). This could be linked to pH, as a higher pH is associated with constraints in P (Rengel 2011). The variation in P is related to the magnitude of soil respiration and could have been further suppressed by the anaerobic conditions caused by the heavy rainfall of the Season 2 (2021/2022) (Parent et al. 2008; Cusack et al. 2019; Magdoff and Van Es 2021). The kinetic energy of raindrops causes surface sealing and crust formation which reduce soil infiltration (Vaezi et al. 2017).

Figure 2: Seasonal trend for total P and microbial respiration (CO₂-C), measured as Mg kkg⁻¹, over the duration of the study

The third, and most common, trend was a significant decrease in measured values of organic components from Season 1 (2020/2021) to Season 2 (2021/2022), with an increased value in the third season (2022/2023). The OM is generally low in sandy soil but in our study it was further degraded by the anaerobic conditions caused by the rain intensity and duration of Season 2 (2021/202) (Horneck et al. 2011). Sitthaphanit et al. (2009) explains that when rainfall intensity is high, decomposition occurs too quickly, causing OM and nutrients to be lost through leaching, rather than being released slowly into the soil. This results in lower WEON and WEOC levels, which are key determinants of the overall soil health score. The more widespread rainfall in the Season 3 (2022/2023) allowed some physical, chemical and biological properties to recover (Figure 3) due to their ability to withstand short-term fluctuations (Turner et al. 2015) and because the surface sealing and crust formation was less severe (Vaezi et al. 2017). These improvements resulted in a 17% increase in overall soil health from Season 2 (2021/2022) to Season 3 (2022/2023) (Figure 4).

Figure 3: Seasonal trend for WEON and WEOC, measured as mg kg⁻¹, over the duration of the study

Figure 4: Haney soil health scores in each cropping system for Season 1 (2020/2021), Season 2 (2021/2022) and Season 3 (2022/2023). MMS = Maize–Maize–Soybean, MS = Maize–Soybean, MCS = Maize–Cover crop–Soybean. The MMS system was split into MMS1 and MMS2 to distinguish between maize Season 1 (MMS1) and maize Season 2 (MMS2)

In conclusion, this study showed that soybean and the legumes incorporated in cover crops are able to naturally maintain soil health at levels which are similar to those of soils exposed to an artificially fertilised monoculture of maize, therefore minimising the need for fertiliser application. This is a promising sign for the incorporation of crop rotation systems as a means of sustainable agriculture in this area. In addition, with the prediction of more extreme weather in the future, it is important to note that while recovery rates of different soil properties vary after intense, high rainfall, the overall soil health is generally weakened by such events. Seasonal variation (especially rainfall intensity and duration) in combination with inherent soil properties are important determinants of soil health in the region.

Acknowledgements

The authors would like to thank the Department of Higher Education and Training (DHET), South Africa for their financial contribution towards the study. The Sandy Soils Development Committee (SSDC) is also thanked for the access to field trials and data.