The effect of nitrogen and sulphur on the agronomical and water use efficiencies of canola (Brassica napus L.) grown in selected localities of the Western Cape province, South Africa

Abstract

The grain yield per unit of plant-available nitrogen (N) and sulphur (S) nutrients in the soil is usually low in most crops, including canola. Besides the ability of different crops in utilising nutrients, soil and climatic conditions can also influence uptake and use of major plant nutrients. Hence, to determine the effect of soil and climatic differences on canola agronomic efficiency response to N and S application, a factorial split-plot experiment consisting of three S (0, 15 and 30 kg S ha⁻¹) and five N (0, 30, 60, 90 and 120 kg N ha⁻¹) rates was conducted at selected localities (Altona, Elsenburg, Langgewens, Roodebloem and Welgevallen) in the Western Cape province during the 2009 to 2011 growing seasons. Agronomic efficiencies of N applications were improved if 15 kg S ha⁻¹ was applied, but not with applications of 30 kg S ha⁻¹. Sulphur use efficiency also improved when high N (120 kg N ha⁻¹) was applied. The application of both N and S resulted in higher water use efficiencies at some localities.

Keywords:

• agronomic use efficiency
• canola
• water use efficiency

Introduction

Nitrogen (N) is often the most limiting factor in crop production (Blumenthal et al. 2008) with its application usually being the largest production input cost in most grain crops (Sands et al. 2011). Most crop plants may only be able to convert 30–40% of the applied N to grain production (Raun and Johnson 1999; Hirel et al. 2007). Nutrient conversion rates are even worse in canola (Masson and Brennan 1998), with N conversion values of even less than 10% (Malagoli et al. 2005; Sylvester-Bradley and Kindred 2009). The crop is generally a heavy nutrient feeder, particularly with reference to N and S (Oplinger et al. 2000; Hocking and Strapper 2001).

Moll et al. (1982) defined nitrogen use efficiency (NUE) as being the yield of grain per unit of available N in the soil (including the residual N present in the soil and the fertiliser). This NUE can be separated into two processes: uptake efficiency and the utilisation efficiency (the ability to use N to produce grain yield) (Lea and Azevedo 2006). It is difficult to determine how much N is in the soil and how much is taken up, hence N efficiency can be extended to agronomic use efficiency, which is the increase in grain yield obtained when N is applied as fertilisers (Smith et al. 1988). Besides the differences observed in canola genotypes in N remobilisation (Balint et al. 2008; Gan et al. 2008a), efficiency is also affected by agronomic factors that have an influence on plant morphology and physiology (Hirel et al. 2007). Nitrogen use efficiency can also be affected by environmental conditions, such as soil nutrient and moisture levels, temperature and light radiation (Rathke et al. 2006). Increasing N application rates have also been shown to reduce its utilisation efficiency, especially at levels beyond 100 kg N ha⁻¹ (Gan et al. 2008a).

Canola also shows responses to sulphur (S) because the requirements are four times that of wheat or maize (Abdallah et al. 2010). This higher S requirement in canola compared to cereals is ascribed to the higher protein content of these cultivars combined with the higher proportion in brassicas of cysteine and methionine, which contributes to the larger S requirements (Durrani and Khalil 1990). However, the canola grain yield response to applied S depends on N application rate (Janzen and Bettany 1984), as well as soil and climatic conditions (Ngezimana and Agenbag 2014b).

Fertilisation with N and S have been shown to increase yields (Balint et al. 2008; Breennon and Bolland 2008; Gan et al. 2008b; Ngezimana and Agenbag 2013), but absorption (Koenig 2012), translocation (Malhi et al. 2007) and partitioning (Jackson 2000) of these nutrients have been shown to be the highest if there is sufficient available soil moisture (Ahmadi and Bahrani 2009). The utilisation of available moisture can be described as water use efficiency (WUE), which is the grain yield per millimeter of rainfall received (Perry and Hillman 1991). Canola is very prone to water stress (Faraji et al. 2009) and water is best utilised when all growing conditions are favourable (Angadi et al. 2003). Measurement of WUE can thus be of major importance to evaluate growing conditions and identifying constraints to yield, other than lack of water (Cocks et al. 2001).

Canola production and yields per hectare in South Africa are still low compared to the global major producers and efficient fertiliser programs may help to improve both agronomic and water use efficiences, which will result in higher grains yields. For this reason, a study was conducted to determine the effect of soil and climatic differences, as experienced at selected localities in the Western Cape production area, on the agronomic N, S and water use efficiency of canola in response to N and S application rates.

Materials and methods

A factorial split-plot experiment with three S rates (0, 15 and 30 kg S ha⁻¹) as a main factor and five N rates (0, 30, 60, 90 and 120 kg N ha⁻¹) as a subfactor was conducted at Elsenburg (33°51′ S, 18°51′ E; 117 m above sea level [asl]), Langgewens (33°17′ S, 18°40′ E; 91 m asl), Roodebloem (34°22′ S, 19°52′ E; 132 m asl) experimental stations of the Western Cape province of South Africa during the canola seasons of 2009, 2010 and 2011, as well as at Welgevallen (33°52′ S, 18°42′ E; 119 m asl) during 2010 and at Altona (33°42′ S, 18°37′ E; 42 m asl) during 2011. Soil physical and chemical properties at the various locations were as reported by Ngezimana and Agenbag (2014a) with low N and S levels at all localities with the exception of Welgevallen. Nitrogen was applied as limestone ammonium nitrate (LAN) containing 28% N, while S was applied in the form of gypsum (CaSO₄.2H₂O) containing 16% S. All S was applied at planting. Nitrogen application was done in splits of 30 kg ha⁻¹ at planting and thereafter at 30 d intervals depending on the treatment. The 30 N treatments, for example, received all N at planting, whereas the 120 N treatments received 30 kg N ha⁻¹ at planting as well as 30 kg N ha⁻¹ at 30, 60 and 90 days after planting (DAP). Plots receiving no fertiliser were regarded as the control. At all localities, a planting density of 4 kg ha⁻¹ was used. Canola ‘Stubby’ was planted across all localities in 2009, whereas ‘Bravo’ was used in the 2010 and 2011 seasons.

At crop maturity, canola grain yields harvested from 5 m × 3 m plots were converted to kilograms per hectare. Data on grain yield responses to N and S have been reported previously (Ngezimana and Agenbag 2014b). Agronomic use efficiencies of N and S (kilograms seed per kilogram N or S applied) were calculated according to Smith et al. (1988) using the following equations for N and S, respectively:

and

Rainfall was recorded at all localities from 2009 to 2011 during the canola growing season (May to October). In 2009, Elsenburg received the highest total rainfall (May to October) of 568.3 mm, whereas Roodebloem and Langgewens received total rainfall of 339.4 and 285.2 mm, respectively. In 2010, Elsenburg received higher rainfall (442.9 mm) than Welgevallen (426.6 mm), Langgewens (366 mm) and Roodebloem (320.9 mm). During 2011, Elsenburg received a higher total rainfall (325.8 mm) than Langgewens (308.6 mm), Altona (295.9 mm) and Roodebloem (241.5 mm). Total rainfall received was used to calculate water use efficiency (WUE) as grain yield per millimetre of rainfall received using the following equation (Perry and Hillman 1991):

Water use was calculated as follows:

This equation was developed for Western Australian conditions (Perry and Hillman 1991), which are very similar to conditions in the Western Cape, and most of the trials in this study were also planted in early May and harvested towards the end of October or early November. The equation was therefore considered to be relevant for this study.

Data analysis

Data were analysed with analysis of variance (ANOVA) using Statistica 11 (StatSoft, Tulsa, OK, USA). To compare different soil and climatic conditions and measure response to treatments at different localities, locality was regarded as a factor. As three localities (Roodebloem, Elsenburg and Langgewens) were planted in all seasons (2009–2011), whereas Welgevallen and Altona were included in 2010 and 2011, respectively, data for the three seasons were analysed separately. Main and interaction effects were compared using the least significant difference (LSD) test at the 5% level of probability. Any treatment means found to be significantly different were separated using Fisher's protected LSD_0.05 and only statistically significant data are presented and discussed (Table 1).

Table 1: Summary of results from ANOVA on use efficiencies of agronomic nitrogen (NUE) and sulphur (SUE), as well as water use efficiency (WUE) during the 2009, 2010 and 2011 growing seasons

	NUE			SUE			WUE
	2009	2010	2011	2009	2010	2011	2009	2010	2011
Locality	*	**	ns	***	***	*	***	***	***
Nitrogen	*	*	ns	ns	***	***	***	***	***
Sulphur	**	**	ns	ns	*	ns	***	***	***
Locality × Nitrogen	ns	***	ns	ns	***	ns	ns	***	ns
Locality × Sulphur	ns	ns	ns	ns	ns	ns	ns	*	ns
Sulphur × Nitrogen	ns	ns	ns	ns	ns	ns	ns	***	ns
Locality × Nitrogen × Sulphur	ns	ns	ns	ns	ns	ns	ns	**	ns

* P < 0.05

** P < 0.01

*** P < 0.001

ns, non-significant at P = 0.05

Results and discussion

Agronomic use efficiency of applied N

Agronomic use efficiency of applied N (NUE) differed between localities in 2009 with Elsenburg (14.03 kg yield increase per kg N applied) showing significantly higher NUE values compared with Roodebloem (9.53 kg yield increase per kg N applied) and Langgewens (6.50 kg yield increase per kg N applied). Elsenburg received more rainfall, hence higher yield and agronomic efficiency at this locality were observed as canola is sensitive to water stress during the flowering and pod growth phases (Morrison and Stewart 2002; Tesfamariam et al. 2010).

The NUE decreased with the increase in N application rate from, on average, about 14 kg yield increase per kg N applied at a rate of 30 kg N ha⁻¹ to less than 8 at N application rates of more than 60 kg N ha⁻¹ in 2009 (Figure 1). Gan et al. (2008b) also reported low N fertiliser use efficiencies at high N fertiliser application rates. In 2009, no significant differences in NUE were shown in comparisons of treatments where no S (11.37) or 15 kg S (12.71) were applied, but a significant decrease was observed when more than 15 kg S ha⁻¹ was applied. This trend was observed at all localities because the soil showed deficient levels of S (Peverill et al. 1999) and low soil carbon contents of less than 1% at all localities (Ngezimana and Agenbag 2014a) with the exception of Roodebloem.

Figure 1: Agronomic use efficiency of applied nitrogen (NUE) of canola at increasing nitrogen application rates during 2009

In 2010 a significant locality × N interaction was shown with regard to NUE. This was due to contrasting results at different localities (Figure 2). At Elsenburg, and to a lesser extend Roodebloem, NUE tended to increase with the increase in N application rate from 30 to 60 kg N ha⁻¹, followed by a decrease with further increases in N rate. At Langgewens, NUE decreased with increases in N application rate, whereas no significant trend was shown for Welgevallen. However, a decrease in NUE with increasing N application (60–120 kg N ha⁻¹) was shown at all localities. Maman et al. (1999) also found that NUE responds differently to an increase in N fertiliser supply at different localities. During 2011, neither N nor S application rate, nor their interactions, had a significant effect on agronomic efficiency of applied N at any of the localities (Table 1) in spite of low soil S content (Ngezimana and Agenbag 2014a).

Figure 2: Agronomic use efficiency of applied nitrogen (NUE) of canola at increasing nitrogen and application rates at Langgewens, Elsenburg, Roodebloem and Welgevallen during 2010

Agronomic use efficiency of applied S

In 2009, agronomic use efficiency of applied sulphur (SUE) was significantly higher at Elsenburg (22.73 kg ha⁻¹ yield increase per kg S applied) compared to Langgewens (3.15) and Roodebloem (6.46) (Table 2). The high SUE values at Elsenburg in 2009 were attributed to the higher and extended period of high rainfall at this locality, because adequate availability of moisture and a lengthy grain filling phase ultimately lead to higher grain yield per kilogram S applied (Hassan et al. 2007). Neither N nor S application rate, nor their interactions, had a significant effect on SUE values in 2009.

Table 2: Agronomic use efficiency of applied S (SUE) of canola at Elsenburg, Langgewens, Roodebloem, Welgevallen and Altona, during 2009 to 2011 seasons. Means within a season with the same superscript letter are not significantly different (LSD_0.05)

Season	Locality
	Elsenburg	Langgewens	Roodebloem	Welgevallen	Altona
2009	22.73^a	3.15^b	4.46^b	–	–
2010	16.55^a	7.11^b	7.19^b	13.49^a	–
2011	12.05^b	12.07^b	19.48^a	–	14.66^ab

The SUE values of canola at the different localities in 2010 were similar to those of 2009, but results for the different localities were highly dependent on N fertiliser application levels. Although SUE tended to increase with increasing N application rates at most of the localities tested during 2010 (Figure 3), trends were very confusing. At Elsenburg, for example, SUE initially decreased with increasing N rate from 0 to 60 kg N ha⁻¹, but increased with further increase in N rate. At Roodebloem SUE increased with increasing N rate to reach its highest value at 120 kg N ha⁻¹, whereas the highest SUE value at Welgevallen was obtained with an application of 90 kg N ha⁻¹. At Langgewens SUE values were generally low and did not show a significant response to N application rate. Although these results generally confirmed those of previous research (Fismes et al. 2000; Hassan et al. 2007), which showed that N fertilisation can improve SUE in canola, differences in response between localities, in spite of low S values at all localities (Ngezimana and Agenbag 2014a), suggested that other factors such as rainfall and especially rainfall distribution may also be important.

Figure 3: Effect of nitrogen fertiliser rates on agronomic use efficiency of applied sulphur (SUE) of canola grown at different localities during 2010

The SUE in 2010 was also affected by S application rates. Increasing application rates of S from 15 kg S ha⁻¹ to 30 kg S ha⁻¹ resulted in a decrease in SUE from 12.66 to 9.52 kg ha⁻¹ yield increase per kg S applied on average. Although these results suggested that the application of S becomes less effective at rates higher than 15 kg ha⁻¹, application rates of 30 kg S ha⁻¹ may still be economically viable because grain yield still increased with these rates (Ngezimana and Agenbag 2014b).

In 2011 SUE was significantly affected by N application (Figure 4) and differed between localities with Roodebloem (19.48 kg ha⁻¹ yield increase per kg S applied) showing significantly higher SUE values than those of Altona (14.66), Langgewens (12.07) and Elsenburg (12.05) (Table 2). On average, SUE increased with increasing N rate and the highest SUE value was obtained with an N application rate of 120 kg N ha⁻¹ (Figure 4). These results indicated that, although SUE is largely affected by locality (rainfall), S can only be used efficiently if N supply is sufficient.

Figure 4: Agronomic use efficiency of applied sulphur (SUE) of canola at increasing nitrogen application rates during 2011

Water use efficiency

Water use efficiencies of canola differed between localities. In 2009 WUE at Elsenburg (6.36 kg grain mm⁻¹ water) was significantly higher than that at Roodebloem (5.34 kg grain mm water⁻¹) and Langgewens (4.64 kg grain mm⁻¹ water) (Table 3), due to the better distribution of rainfall during the growing season at Elsenburg (Ngezimana and Agenbag 2014b).

Table 3: Water use efficiency (WUE) at Elsenburg, Langgewens, Roodebloem Welgevallen and Altona during 2009 to 2011 seasons. Means within a season with the same superscript letter are not significantly different (LSD_0.05)

Season	Locality
	Elsenburg	Langgewens	Roodebloem	Welgevallen	Altona
2009	6.36^a	4.64^c	5.34^b	–	–
2010	7.53a	4.44c	4.79c	6.20b	–
2011	4.63c	4.07c	12.18a	–	9.65b

In 2009 WUE increased on average from 3.91 kg of grain mm⁻¹ water with no N (0N) to 6.69 kg grain mm⁻¹ water with an application of 120 kg N ha⁻¹ (Figure 5). During 2009 WUE also differed significantly due to S application rates at all localities from 5.04 kg grain mm⁻¹ water when no S was applied to 5.84 kg grain mm⁻¹ water with an application of 30 kg S ha⁻¹ (Figure 5). Soils at almost all localities had low S contents (<6 mg kg⁻¹ soil) and soil carbon contents (Ngezimana and Agenbag 2014a), indicating the importance of S fertiliser aplications on soils low in S and carbon.

Figure 5: Canola water use efficiency in 2009, 2010 and 2011 seasons at different nitrogen (upper) and sulphur (lower) fertilisation rates. Means within a season with the same letter are not significantly different (LSD_0.05)

In 2010 a significant interaction with regard to WUE was shown between locality and both N and S application rates, suggesting that responses differed at different localities. However, from Figure 6 it is clear that, although the degree of response did differ, WUE increased at all localities with an increase in N application rate. Responses with increasing N rates were larger at 15 and 30 kg S application rates compared to where no S was applied. These results confirmed those of a previous study (Ahmadi and Bahrani 2009) and showed the complexity of canola yield responses to N and S at localities differing in rainfall and, to a lesser extent, soil conditions. During 2011, Roodebloem (12.18 kg rain mm⁻¹ water) had a higher WUE than Altona (9.65 kg grain mm⁻¹ water), Elsenburg (4.63 kg grain mm⁻¹ water) and Langgewens (4.08 kg grain mm⁻¹ water) (Table 3). These differences in WUE were most likely due to differences in distribution and total amount of rainfall received during the canola growing season. Rainfall distribution within the growing season is of major importance, because the response of canola to water stress is dependent on growth stage (Ahmadi and Bahrani 2009). Gan et al. (2004) reported that canola stressed during earlier growth stages exhibits recovery, whereas if stressed during pod development, most of the yield components are reduced.

Figure 6: Canola water use efficiency in 2010 at Elsenburg, Langgewens, Roodebloem and Welgevallen localities at different sulphur (0, 15 and 30 kg ha⁻¹) and nitrogen (0, 30, 60, 90 and 120 kg ha⁻¹) fertilisation rates (LSD_0.05 = 0.48)

During 2011 the highest WUE of 8.5 kg mm⁻¹ water was observed when 120 kg N ha⁻¹ was applied (Figure 5). This increase in WUE resulted from the increase in grain yields with increasing N application (Ngezimana and Agenbag 2014b). Nitrogen fertilisation results in increased absorption and translocation of assimilates and stimulates growth (Wright et al. 1988; Koenig 2012), hence increasing yield components (Ahmadi and Bahrani 2009). Fertilising with S also increased the grain yield per millimetre of rainfall received in 2011 with high S application rates, resulting in a WUE of close to 9 kg mm⁻¹ water (Figure 5) compared to 6.5 kg mm⁻¹ water when no S was applied. Besides N fertilisation, addition of S can have profound effects on canola, increasing the grain yields (Breennon and Bolland 2008; Abdallah et al. 2010). As with N, S absorption and translocation increased when water is available increasing the WUE of the canola plant when there is an adequate soil S level.

Conclusions

Agronomic efficiencies of N applications were improved if 15 kg S ha⁻¹ was applied, especially in high rainfall areas. Agronomic use efficiencies of applied S were higher when adequate rainfall was available. Moreover, SUE was improved when rainfall was well distributed throughout the growing season, coupled with high N fertilisation (120 kg N ha⁻¹). Although WUE (kilograms grain yield per millimetre of May-to-October rainfall) differed between years and localities, fertilisation improved WUE.

Acknowledgement

The authors are grateful to the Protein Research Foundation for financial support.