Conservation agriculture effects on plant nutrients and maize grain yield after four years of maize–winter cover crop rotations

Abstract

The effects of maize rotation with oat (Avena sativa cv. Sederberg) and grazing vetch (Vicia dasycarpa cv. Max) winter cover crops on nutrient availability, maize grain yield and maize grain nutrient concentration were investigated. Soil samples were collected from the 0–5 and 5–20 cm depths of experimental plots after four years of continuous maize–winter cover crop rotations. Winter cover crops caused small increases of extractable soil Cu, Mn, P and Zn, but not Ca and K, concentrations. A small dose of fertiliser applied to maize (60, 30, 40 and 1.5 kg ha⁻¹ of N, P, K and Zn, respectively) also caused a significant increase in P and Zn, as well as mineral N, concentrations but only in the vetch–maize rotations. Stratification of Mn, K and Zn in the 0–5 cm soil depth occurred in all treatments. Vetch additionally increased maize grain yield, grain N concentration and soil acidity more than either oat or fallow. Non-fertilisation of maize reduced maize grain yield on oat and fallow–maize rotations more than it did on vetch–maize rotations. A combined application of vetch winter cover crops and small doses of fertiliser could significantly improve sustainability of low input maize-based conservation agriculture systems.

Keywords:

• fertiliser
• grazing vetch
• no-till
• oat
• soil fertility

Introduction

Conventional farming practices that are based on extensive tillage, especially when combined with removal or in situ burning of crop residues, have magnified nutrient losses originating from soil erosion, as well as soil fertility challenges experienced by maize (Zea mays L.) farmers around the world. Therefore, the search for sustainable solutions to soil fertility challenges is increasingly focused on conservation agriculture (CA) (Govaerts et al. 2005, Pretty et al. 2006, Hobbs et al. 2008). The major principles of CA are reduced tillage, crop rotation and permanent soil cover through cover crops. Inclusion of winter cover crops in maize-based cropping systems to replace fallow has implications on nutrient availability as they: (1) capture nutrients that would otherwise be lost to leaching, (2) extract nutrients from deeper layers and convert plant-unavailable mineral forms into organic forms and (3) fix and enhance nitrogen (N) in soils deprived of N (only for legumes). The degree to which a particular cover crop meets these specifications may be dependent on soil, climate, succeeding cash crop, as well as characteristics of the winter cover crop itself. Grazing vetch (Vicia dasycarpa L.) and oat (Avena sativa L.) are examples of fast-growing, winter-hardy cover crops, which have been found to provide dependable biomass in some maize-based CA systems (Murungu et al. 2011).

Maize is the staple food in most African countries and its grain nutrient concentration is important as an indicator of feed value (Graham et al. 2001). Grain nutrient concentration can also provide information related to the necessity for soil nutrient replenishment through fertilisation (Eghball et al. 2003, Heckman et al. 2003). Concerns regarding diminished grain mineral nutritional quality of maize due to genetic selection for yield have recently increased (Vyn et al. 1998, Feil et al. 2005, Ferreira et al. 2012). There are claims that crop nutrient concentration can be increased through CA (Worthington 2001, Graham et. al. 2001).

Long-term experiments are the primary source of information to determine the effects of cropping systems and soil management on soil productivity (Poulton 1995, Körschens 2006). Such an experiment was established in 2007 in the Eastern Cape of South Africa to determine the effects of oat and vetch winter cover crops and fertiliser application to maize on biomass input and weed suppression under no-till and irrigation (Murungu et al. 2011). This paper reports the effects of these winter cover crops on nutrient availability, maize grain yield and grain nutrient concentration after four years of continuous practice in this trial.

Materials and methods

Experimental design and field trial management

The field trial was located at latitude 32°46′ S and longitude 26°50′ E at an altitude of 535 m above sea level in the Eastern Cape, South Africa. The area has a warm temperate climate and an average annual rainfall of 575 mm, which is received mainly during the summer months from October to April. The soil is deep and of alluvial origin, classified as Haplic Cambisol (IUSS Working Group WRB 2006). Mineralogy of the soil is dominated by mica in the clay fraction, with low amounts of quartz and kaolinites (Mandiringana et al. 2005). The soil texture is sandy loam, comprising 64.2% sand, 16.0% silt and 19.8% clay. Prior to establishment of the field trial in 2007, conventional tillage was practiced and maize was the major crop. The field trial was a split plot design and in the main plots, oat (cv. Sederberg) and grazing vetch (cv. Max) were planted. Control plots with no winter cover crops (fallow) were also included. Subsequent to winter cover crop termination, plots were split and maize (cv. PAN 6479) was planted at two fertiliser levels (with and without fertiliser). Fertiliser was not applied to the cover crops and the fallow. There were thus two factors in this experiment, namely type of cover crop and fertiliser level, to give a 3 × 2 split plot design that was replicated three times. The six treatment combinations are presented in Table 1.

Table 1: Summary of treatments used in the study

Type of cover crop (main plots)	Maize (subplots)	Treatment
Oat	Fertiliser applied	1
	No fertiliser applied	2
Vetch	Fertiliser applied	3
	No fertiliser applied	4
Fallow	Fertiliser applied	5
	No fertiliser applied	6

Maize planting was carried out using hand-operated ‘matraca’ planters and no tillage was done. The maize rows were spaced at a distance of 90 cm and the plants at 30 cm to give a planting density of 37 000 plants ha⁻¹. Fertilised maize plots received 60, 30, 40 and 1.5 kg ha⁻¹ of N, phosphorus (P), potassium (K), and zinc (Zn), respectively. One-third of the maize N was applied using the matraca planters as a compound (2:3:4) containing 6.7% N, 10% P, 13.3% K and 0.5% Zn at planting. The remainder was applied as limestone of ammonium nitrate (28% N) at six weeks after planting by banding. These low maize fertiliser rates mimicked smallholder irrigation farmer practice in the Eastern Cape (Fanadzo et al. 2010). Maize was harvested after physiological maturity. Subsequent to harvesting, maize stalks were rolled, glyphosate applied at 5 l ha⁻¹, and oat and grazing vetch winter cover crops were planted. This sequence was repeated over four maize seasons (2007–2011). Supplementary irrigation was applied to the maize and winter cover crops when necessary and the irrigation data has been presented by Dube et al. (2012). Detailed agronomic management of the field trial as well as biomass inputs from the winter cover crops have been reported by Dube et al. (2012).

Determination of maize yield and grain nutrient concentration

Maize grain yield was obtained from a net plot of 6 m × 2 m, representing the two central rows. Maize grain samples from the fourth year of rotations were milled and sieved (<0.1 mm), dry ashed at 450 °C overnight and digested in 1 M HCl for use in the determination of nutrient concentration. Phosphorus (P) was determined using the Murphy and Riley (1962) molybdenum blue procedure. Potassium, calcium (Ca), magnesium (Mg), copper (Cu), manganese (Mn) and Zn were determined by atomic absorption spectrophotometry. Nitrogen was determined before ashing using a near-infrared refractometer (NIR).

Soil sampling and analyses

Soil sampling was done from the inner two-thirds of each plot. Six random soil samples were collected at each of the 0–5 cm (using a small trowel) and 5–20 cm (using a graduated auger, 7 cm diameter) depths in all subplots at the beginning of the fourth year of rotation, immediately before maize planting. The six subsamples of each soil depth were bulked to a composite sample and transferred to the laboratory, where they were air dried, visible organic debris removed, and ground (<2 mm). Soil pH was measured in 1.0 M KCl (1:2.5 soil:water ratio) using a pH meter. Calcium and Mg were extracted from the soil by 1 M KCl. Phosphorus, K, Zn, Cu and Mn were extracted by the Ambic-2 extracting solution (0.25 M NH₄HCO₃ + 0.01 M disodium EDTA + 0.01 M NH₄F + 0.05 g l⁻¹ superfloc N100). The elements in the extracts were measured by atomic absorption spectrophotometry. Soil inorganic N (NH₄⁺ and NO₃⁻) was determined in freshly sampled soil (0–20 cm depth) before planting maize. The analysis was done using a Skalar San Plus System after extraction with 0.5 M K₂SO₄ (1:4 soil:solution ratio). Total soil N was determined using NIR.

Data analysis

Analysis of variance (ANOVA) on the extractable soil nutrients, grain yield and grain nutrient concentration data was carried out using GenStat Release 15 statistical software (Payne et al. 2013). Maize grain yield data were combined across the four maize seasons (2007–2011) and the measurements were subjected to ANOVA to determine yield trends over four maize seasons. This was done after testing for homogeneity of variances across years. Comparisons between treatment means were carried out using the least significant difference at the 5% probability level.

Results and discussion

Effects of cover crop type and fertiliser on soil mineral N and total N

The cover crop type × fertiliser interaction had a significant effect (p < 0.05) on NH₄⁺-N and total mineral N, but its effect on NO₃⁻-N was not significant. Fertiliser application increased NH₄⁺-N and total mineral N on the vetchmaize rotation, but not on the oat or fallow-maize rotations (Figure 1), suggesting that the N in the fertiliser applied was probably immobilised by the oat and fallow systems. Soil in the vetch–maize rotation had the highest NO₃⁻-N (7.49 mg kg⁻¹), whilst that in the oat–maize rotation (5.28 mg kg⁻¹) was not significantly different from that of the fallow–maize rotation (4.51 mg kg⁻¹). All factors and interactions had no significant (p > 0.05) effect on total soil N. Total N content of the soil may be less sensitive to legume cover crops and small doses of N fertiliser than the mineral N in this no-till system.

Figure 1: Winter cover crop type × fertiliser interaction effects on soil mineral nitrogen (N) (mg kg⁻¹)

Effects of cover crop type and fertiliser on extractable macronutrients and micronutrients

Cover crop type × fertiliser interactions were not significant for all the extractable nutrients that were analysed. However, cover crop type effects were significant (p < 0.01) on P, Mg, Cu, Mn and Zn, but not on Ca and K. The soil on winter cover crop–maize rotations had slightly higher amounts of P, Cu and Zn than that under the fallow–maize rotation (Table 2). There are several possible explanations for this observation. The soil type used for this study is weakly weathered, and did not contain large amounts of kaolinite (1:1) clays, or Fe and Al oxides (Mandiringana et al. 2005), and therefore has low to very low P fixation capacity. There is evidence that suggests that organic C derived from cover crops reduces P fixation through its interaction with soil components on P fixation sites, thus increasing P availability (Ohno and Erich 1997). The buildup of organic matter from cover crops is known to favour Zn accumulation, probably by formation of soluble complexes (Rahman et al. 1996). Cover crop residue organic matter itself is a source of Zn, P and Cu, mined from lower soil volumes through extensive rooting systems. It is noted that P and Zn are antagonistic elements in nature, and Zn availability is reduced when available P in the soil is high or P fertiliser is added (Olsen 1972). Phosphorus is thought to precipitate Zn either in the soil or at the root–soil interface. In our study, the simultaneous gain in P and Zn availability in the soil may have been caused by soil organic matter interactions (Ohno and Erich 1997, Rahman et al. 1996). It has been reported that the cover crops improved soil organic matter in this trial (Dube et al. 2012). Thus the possible negative effects of increased soil P on Zn availability were counteracted by organic matter inputs from the winter cover crops.

Table 2: Effects of cover crop type, fertiliser and soil depth on extractable nutrients (mg l⁻¹). Means followed by the same superscript letter within the same column are not significantly different

Treatment	P	Mg	Ca	K	Cu	Mn	Zn
Cover crop type
Oat	30.9^b	389.9^a	1 329	112.4	4.62^a	25.2^b	6.38^a
Grazing vetch	39.9^a	330.4^b	1 347	118.7	4.36^a	31.8^a	7.05^a
Weedy fallow	19.8^c	376.7^a	1 397	102.2	3.96^b	22.6^b	4.40^b
	***	**	ns	ns	**	**	***
Fertiliser
Fertilised	32.5	341.8	1 319	113.3	4.27	27.0	6.53
Non fertilised	27.9	368.8	1 396	108.9	4.35	26.8	5.35
	*	**	ns	ns	ns	ns	*
Soil depth
0–5 cm	32.1	365.1	1 370	139.4	4.28	30.9	6.77
5–20 cm	28.2	345.5	1 345	82.8	4.35	22.9	5.96
	ns	ns	ns	***	ns	**	**

* p < 0.05, ** p < 0.01, *** p < 0.001, ns = not significant

The non-significance of all treatment effects on extractable Ca and K could be due to high inherent total contents of these elements in the relatively young soil used. According to Landon (1984), Ca availability varies enormously from soil to soil and is highly dependent on a number of other factors. Calcium deficiency is mostly limited to highly leached and acidic (pH < 5.5) soils. There were also significantly (p < 0.05) higher amounts of Mn, K and Zn in the 0–5 cm soil layer compared to the 5–20 cm layer (Table 2). These mineral nutrients have limited mobility in the soil and may accumulate in soil surface strata under zero-till regimes, due to surface-applied fertiliser and decomposition of crop residues on the soil surface (Franzluebbers and Hons 1996, Howard et al. 1999). Stratification of nutrients in the upper soil layer under no-till may cause nutritional constraints to productivity when the surface soil becomes dry unless the rest of the soil profile is highly fertile, or receives appropriate levels of fertiliser (Radford et al. 2007).

Effects of cover crop type and fertiliser on soil pH

Interaction effect of cover crop type and fertilizer on soil pH was not significant. However, cover crop type effect on pH was significant (p < 0.01), showing that vetch–maize rotation resulted in slightly lower pH (5.16 ± 0.14) than fallow–maize (5.41), which was similar to maize–oat (5.38). Fertiliser effects on pH were, however, not significant (p > 0.05). The decreasing effect of vetch on pH is most likely due to rapid mineralisation and subsequent nitrification of N in the legume residues as suggested by Liu et al. (1989). Nitrification of 1 mole of NH₄⁺ produces 2 moles of H⁺, thus increasing acidity. Small changes in pH may be associated with large liming requirements for pH correction (Brady and Weil 2008) and, for effective results, the lime is usually incorporated into the soil through tillage. It may therefore be a serious challenge to correct acidification in the subsoil layer under vetch cover crops in the no-till system. There is evidence suggesting that acidity problems are more likely for fields in the early stages of no-till (less than six years) than in longer-term no-till, suggesting that the pH problem could be ameliorated by continued practice of no-till (Crozier et al. 1999). The mechanisms behind this effect, however, are not exactly clear. The extractable Mn content was highest in soil on maize–vetch rotation, but that on oat–maize rotation was similar to that of the maize– fallow rotation (Table 2). These findings are in agreement with Bromfield (1958), who found that the root exudates of vetch had a pH of 5.1 and were more effective in rendering Mn available, compared to root exudates of oat, which had a pH of 7.1. The observed Mn results could therefore be related to pH modification by the vetch cover crops as supported by the soil pH results.

Effects of cover crop type and fertiliser on maize grain yield and grain nutrient concentration

The cover crop type × fertiliser × season interaction effect on maize yield was significant (p < 0.001). Maize grain yield was significantly higher on the vetch–maize rotation for the second season (Figure 2) and lowest on the fallow– maize rotation in the first season. In the fertilised system, the winter cover crop–maize rotations produced higher maize grain yield than the maize–fallow rotations in all seasons. Yield declined from the second season onwards, particularly for the fallow and all non-fertilised treatments. This could be due to the effect of reduced nutrient supply in these rotations, as supported by the significant effects of cover crops and fertiliser on P and Zn (Table 2). In the non-fertilised systems, vetch–maize rotation produced significantly higher maize yield than either oat– or fallow– maize rotations from the second season onwards (Figure 2). It is a well-established phenomenon that the yields of maize crops often increase when they are grown after legumes, mainly due to N fixation by the legume (Tian et al. 2000, Kramberger et al. 2009).

Figure 2: Effects of cover crop type and fertiliser regime on maize grain yield across four seasons (2007–2011)

Cover crop type × fertiliser interaction had no significant effect on concentration of all other nutrients in maize grain yield except for N. While fertilised oat, vetch and weedy fallow plots gave similar grain N concentrations, a lack of fertiliser application reduced maize grain N content on the oat and weedy fallow but not on the vetch winter cover crops (Figure 3), most likely due to the vetch effects on soil N. The total N concentration of grain is a good indicator of the grain protein concentration (Mosse 1990). This implies that vetch winter cover crops could increase protein content of maize grain, hence provide nutritional benefits for resource-poor farmers who cannot afford fertiliser. Bruns and Ebelhar (2006) reported an increase in total N content of grain from increases in soil mineral N content under conditions where soil N was a limiting factor to grain yield. Winter cover crop-type effects on concentration of other nutrients in the grain were not significant (Table 3). Fertiliser application effects were also not significant. Dilution effect arising from substantial increase in grain yield may partly account for the observed lack of response of other grain nutrients to the cover crops and fertiliser. It appears that the winter cover crops and small doses of fertiliser applied had a much greater impact on maize grain yield than on grain nutrient concentration.

Figure 3: Winter cover crop type × fertiliser interaction effects on grain nitrogen (N) concentration

Table 3: Effects of winter cover crop type on grain nutrient concentration

Cover crop	Ca (%)	Mg (%)	P (%)	K (%)	Zn (mg kg^–1)	Cu (mg kg^–1)	Mn (mg kg^–1)
Vetch	0.034	0.106	0.297	0.318	25.0	1.748	8.36
Oat	0.031	0.112	0.307	0.305	27.2	1.313	8.42
p value	ns	ns	ns	ns	ns	ns	ns
CV (%)	25.2	8.4	8.8	10.3	20.9	27.5	15.6

ns = Not significant

Conclusions and recommendations

The practice of winter cover cropping increased the availability of some essential plant nutrients and maize grain yield in low fertiliser input, irrigated maize-based CA systems. The winter cover crops appeared to make a small contribution to the fertility of P and Zn, which are major limiting nutrients in South African soils. Without fertiliser, grazing vetch performed better than oat in terms of improving soil mineral N, maize grain yield and grain N concentration. Grazing vetch, however, also increased acidification of the soil, which could increase the liming requirement. Stratification of Mn, K and Zn in the surface soil also occurred. Future studies need to test the combined application of vetch winter cover crops and small doses of fertiliser under rain-fed conditions in low fertiliser input, smallholder farmer systems, where this technology could make a major contribution.

Acknowledgements

This work was funded by the National Research Foundation (NRF) (grant number 62294) and the Govan Mbeki Research and Development Centre (GMRDC), University of Fort Hare. The views expressed are not necessarily those of the NRF or the GMRDC. The authors also thank the anonymous reviewers, whose comments were valuable in improving this paper.