Agronomic drivers of yield in rain fed wheat production systems of the Northern Grains Region

¹NSW DPI, Tamworth, ²NSW DPI, Trangie

Take home messages

• Time of sowing (TOS) was found to be a key determinant of yield, in all three targeted geographic regions/environments, demonstrating significant declines in yield with delayed sowing vs. early-main season sowings. Higher yielding environments were impacted less by delays in TOS than lower yielding environments, with yield losses of 13% for the Liverpool Plains (LPP) vs. 34% and 46% for Trangie and Nyngan respectively. Variety and/or maturity type did not have a significant effect on yield with timely sowing in the early-main season window indicating that varieties were generally ‘plastic’ in their responses and broadly adapted to environments.
• Variety and maturity type did however influence yield potential with a delayed TOS at Nyngan and Trangie. The faster maturing varieties LRPB Crusader and LRPB Spitfire out yielded the main season variety EGA Gregory, with increasing plant population further enhancing yield potential. This demonstrates the scope for genotype (G) x management (M) options with delayed TOS in lower yielding environments.
• Altering variety and maturity type, and increasing targeted plant populations in response to delays in TOS could not fully compensate for the yield losses associated with delayed sowings.
• Yield responses to Nitrogen (N) and Phosphorus (P) fertiliser application rates, were variable and influenced by starting soil nutrition and seasonal conditions.
• Crown rot (CR) was a significant factor affecting yield potential at both Nyngan and LPP sites, decreasing yields by 25% and 12% respectively. This highlights the potential impact of CR on yield and the underlying need to ensure that yield potential is not constrained by biotic factors such soil-borne pathogens.

Background

It is currently estimated that growers in the Northern Grains Region (NGR) are achieving around 49% of water limited yield potential, translating to a yield gap of approximately 1.9 t/ha (www.yieldgapaustralia.com.au). Water limited yield potential is defined as the potential yield achieved under non-limiting nutrition and biotic stresses using best management practices, but subject to environmental constraints namely plant available water and temperature (Hochman et al., 2012). To put this into perspective, leading growers in Australia using best management practices and available technology are estimated to be achieving around 80% of water-limited yield potential (van Rees et al., 2014), thus indicating that yield is being limited by factors other than available water. Based on these observations, there is an ‘exploitable yield gap’ between actual and attainable yields (i.e. 80% of water limited yield potential) considered the approximation of where grower’s yields plateau within most major cropping systems, as a result of economic constraints and climatic variability (van Ittersum et al., 2013).

Identifying the key drivers of yield in water-limited, rain fed environments is clearly an important strategy for reducing the exploitable yield gap and for increasing dryland wheat production. The aim of this research was to benchmark yield potential across a range of growing environments in the NGR of NSW, over two consecutive seasons, and to quantify the impact of genotype (G), management (M) and environment (E) on yield. Possible yield limiting factors investigated included variety selection (maturity type), time of sowing (TOS), plant population and fertiliser inputs (nitrogen and phosphorus). In addition to these factors, crown rot a major disease of wheat and barley crops in the NGR was also incorporated into this study.

Experimental program

Experiments were conducted at 4 sites in both 2014 and 2015, across three different environments within the NGR of NSW. Sites were located on the Liverpool Plains (LPP), Trangie on the edge of the Western Plains and Nyngan in the Central West. Sites were soil cored to ~ 1.2 m prior to sowing to determine plant available water capacities (PAWC) along with starting soil nutrition and other soil properties.  Site descriptions including site and year, in-crop rainfall (May – October) are shown in Table 1, with soil type and starting soil nitrate nitrogen and Colwell P outlined in Table 2.

Table 1. Time of sowing, growing season rainfall and plant available water holding capacity

Site and year	TOS 1	TOS 2	Growing season rain*(mm)	PAWC (mm)
Nowley 2014	14 May	1 July	174	~120
Mullaley 2015	20 May	8 July	185	~140
Tamarang 2014	9 May	30 June	170	~210
Tamarang 2015	19 May	9 July	252	~150
Nyngan 2014	12 May	5 June	212	~90
Nyngan 2015	8 May	28 May	166	~75
Trangie 2014	15 May	11 June	144	~671
Trangie 2015	7May	25 May	200	~76

*May to October. ¹Estimation based on 25% of Effective rainfall – November to March (Gobbett et al., 2016).

A series of 36 treatment combinations (2 times of sowing x 18 treatments) were examined in a partially factorial split-plot design, with three replicates at all sites (Table 3). Treatments were designed similar to an exclusion experiment, with the high input treatment aimed at providing the perceived optimum combination of factors and a low input treatment comprising a base set of agronomic factors (Table 4), so as to benchmark agronomic or management variables.

Four commercial spring wheat (Triticum aestivum) varieties widely grown and well adapted to targeted growing environments were selected and sown at each location. Varieties were from two different maturity groupings, two main season-moderate maturing varieties EGA Gregory and Sunvale, and two fast-moderate maturing varieties LRPB Crusader and LRPB Spitfire. At each location, varieties were sown at two times of sowing (TOS) an early-main season and a delayed TOS (Table 1). Targeted plant populations were grouped as low, moderate (district practice) or high, with the lower yielding environments of Nyngan and Trangie targeting a lower plant population of 40, 80 or 160 plants/m² compared to 60, 120 and 180 plants/m² for the higher yield potential LPP sites. 

Table 2. Soil type, starting soil N (nitrate N) and Colwell P

Site and year	Soil Type	P (Cowell) (mg/kg) 0-10 cm	Soil N03 (kg N/ha) 0-120 cm
Nowley 2014	Black Vertosol	25	123
Mullaley 2015	Grey Vertosol	46	178
Tamarang 2014	Brown Vertosol	77	167
Tamarang 2015	Brown Vertosol	60	213
Nyngan 2014	Red Chromosol	39	147
Nyngan 2015	Red-Brown Chromosol	39	105
Trangie 2014	Grey Vertosol	76	94
Trangie 2015	Red-Brown Chromosol	23	127

Fertiliser treatments looked at both Nitrogen (N) and Phosphorus (P) with rates, products and application methods outlined in Table 3. Crown rot treatments were applied at sowing using sterilised durum grain colonised by at least five isolates of Fusarium pseudograminearum (Fp) as described by Dodman and Wildermuth (1987), with inoculum rates as per Table 3.

Table 3. Summary of treatments: time of sowing, variety, plant population, nitrogen and phosphorous rates and crown rot inoculum levels

Treatment	Details
Two times of sowing (TOS)	TOS 1 - early/main season TOS 2 - delayed
4 Varieties	EGA Gregory (TOS 1 & 2), LRPB Spitfire (TOS 2), Sunvale (TOS 1), LRPB Crusader (TOS 1 & 2)
3 Plant populations	60, 120 or 180 plants/m2(LPP sites) 40, 80 or 160 plants/m2 (Central West sites)
5 Nitrogen rates	0, 50, 100, 150 or a 50 + 50 kg N/ha split application all applied as urea (46% N). Treatments were side banded at sowing, apart from the split application which was applied at sowing and broadcast at stem elongation (GS31).
4 Phosphorus rates	0, 10, 20 or 30 kg/ha P applied as triple super at sowing
4 Crown rot (CR) inoculum rates	0, 0.5, 1.0 or 2.0 g/m row sterilised durum grain colonised by at least 5 different isolates of Fp +/- added at sowing i.e.; 0, CR+, CR++ or CR+++

Small plot sowing equipment based on 33 cm row spacing’s x 5 rows, with two metre centres was used to sow experimental plots of ~10 metres length. Grain yield was obtained using small-plot harvesters.

Table 4. List of experimental treatments

Treatment	TOS	Variety	Applied Nitrogen (kg/ha)	Population (plant/m2)*	Applied Phosphorus (kg/ha)	Disease
1	Early/Main	EGA Gregory	0	120	20	Nil
2	Early/Main	EGA Gregory	50	120	20	Nil
3	Early/Main	EGA Gregory**	100	120	20	Nil
4	Early/Main	EGA Gregory	50+50	120	20	Nil
5	Early/Main	Sunvale	100	120	20	Nil
6	Early/Main	LRPB Crusader	100	120	20	Nil
7	Early/Main	EGA Gregory	100	60	20	Nil
8	Early/Main	EGA Gregory	100	180	20	Nil
9	Early/Main	EGA Gregory	100	120	0	Nil
10	Early/Main	EGA Gregory	100	120	10	Nil
11	Early/Main	EGA Gregory	100	120	20	CR+
12	Early/Main	EGA Gregory	100	120	20	CR++
13	Early/Main	EGA Gregory	100	120	20	CR+++
14	Early/Main	EGA Gregory	0	120	0	Nil
15	Early/Main	Sunvale	0	120	0	Nil
16	Early/Main	LRPB Crusader	0	120	0	Nil
17	Early/Main	EGA Gregory	150	120	20	Nil
18	Early/Main	EGA Gregory	100	120	30	Nil
19	Delayed	LRPB Spitfire	0	120	20	Nil
20	Delayed	LRPB Spitfire	50	120	20	Nil
21	Delayed	LRPB Spitfire**	100	120	20	Nil
22	Delayed	LRPB Spitfire	50+50	120	20	Nil
23	Delayed	LRPB Crusader	100	120	20	Nil
24	Delayed	EGA Gregory	100	120	20	Nil
25	Delayed	LRPB Spitfire	100	60	20	Nil
26	Delayed	LRPB Spitfire	100	180	20	Nil
27	Delayed	LRPB Spitfire	100	120	0	Nil
28	Delayed	LRPB Spitfire	100	120	10	Nil
29	Delayed	LRPB Spitfire	100	120	20	CR+
30	Delayed	LRPB Spitfire	100	120	20	CR++
31	Delayed	LRPB Spitfire	100	120	20	CR+++
32	Delayed	LRPB Spitfire	0	120	0	Nil
33	Delayed	LRPB Crusader	0	120	0	Nil
34	Delayed	EGA Gregory	0	120	0	Nil
35	Delayed	LRPB Spitfire	150	120	20	Nil
36	Delayed	LRPB Spitfire	100	120	30	Nil

*Nyngan and Trangie targeting 40, 80 or 160 plants/m². **High input combinations EGA Gregory TOS 1 and LRPB Spitfire TOS 2.

Results

Yields varied considerably between sites, TOS and year and ranged from 5.91 t/ha at Tamarang for TOS 1 in 2014, to 0.79 t/ha at Nyngan for TOS 2 in 2015 (Table 5). Experiments were grouped into ‘three environments’ based on Statistical Local Areas (SLA), actual yield and estimated water limited yield potential using the ‘Yield Gap Australia Map’ (www.yieldgapaustralia.com.au). 

Table 5. Mean site yield (t/ha) and corresponding yield range (t/ha) for two times of sowing averaged across varieties

Site and year	TOS 1 Mean	Range	TOS 2 Mean	Range
Nowley 2014	3.80	3.97 – 3.77	2.62	2.83 – 2.44
Mullaley 2015	4.34	4.38 – 4.21	3.56	4.00 – 3.06
Tamarang 2014	5.91	5.94 – 5.50	4.40	4.52 – 4.01
Tamarang 2015	4.23*	4.25 – 4.06	4.67*	4.93 – 4.29
Nyngan 2014	3.61	3.72 – 3.57	2.33	2.72 – 2.08
Nyngan 2015	2.65	2.70 – 2.63	0.79	0.85 – 0.65
Trangie 2014	3.07	3.07 – 2.74	1.82	1.92 – 1.72
Trangie 2015	4.24	4.49 – 3.59	3.04	3.35 – 2.74

* All TOS contrasts significant (P<0.05) except Tamarang 2015.

Liverpool Plains (LPP)

The LPP included the site locations of Tamarang, Mullaley and Nowley, with all experiments conducted on Vertosol soils (Table 2). Yield results varied between site and year, and ranged from 5.91 t/ha at Tamarang in 2014 for TOS 1, to 2.62 t/ha at Nowley TOS 2 in 2014, averaged across treatments (Table 5).

When looking at the across site analysis, timeliness of sowing was found to be a significant driver of yield, delays in TOS reducing yields by 0.60 t/ha or 13.1 % when comparing high input (100 kg N/ha, 120 plants/m² , 20 kg P/ha) EGA Gregorytreatments (Table 6). On an individual site basis, when comparing EGA Gregory for TOS, delays in sowing date resulted in yield declines of 6.0 kg/day up to 28.8 kg/day. The only site not to show a yield response due to an earlier TOS was Tamarang in 2015. This was most likely due to the impact of frost induced sterility, with minimum temperatures of < 0°C occurring during the period from the 28^th August to the 1^st September, coinciding with head emergence/anthesis, with a 14.5% decrease in the yield of EGA Gregory between TOS 1 and TOS 2.

Table 6. Effect of management and crown rot (Fp) on grain yield potential – LPP across site analysis

Variety	Population (plants/m2)	Applied N (kg/ha)	Applied P (kg/ha)	Fp (CR+++)	Yield (t/ha)	Yield Gap (t/ha)
TOS 1
EGA Gregory	120	100	20	0	4.571
EGA Gregory	120	100	20	+++	4.26	-0.31*
TOS 2
EGA Gregory	120	100	20	0	3.971	-0.60*
LRPB Crusader	120	100	20	0	3.65	-0.32*
LRPB Spitfire	120	100	20	0	3.55	-0.42*
LRPB Spitfire	120	100	20	+++	3.13	-0.84*

^*Contrast are significant (P<0.05). ¹EGA Gregory TOS 1 vs. TOS 2 contrast.

Maturity type was not a significant factor in TOS 1, with no difference (P<0.05) in yield between varieties. Variety choice did however, impact yield in TOS 2, with EGA Gregory significantly (P<0.001) out yielding both the quicker maturing varieties LRPB Crusader and LRPB Spitfire by 0.32 and 0.42 t/ha respectively. The yield contrast between EGA Gregory in TOS 1 and LRPB Crusader in TOS 2 equated to 0.92 t/ha or 20%, compared to 0.60 t/ha or 13.1% for EGA Gregory in TOS 2.

Increasing crown rot (CR) disease pressure (± Fp applied at sowing) resulted in a decrease in yield in TOS 1 of 0.31 t/ha equating to a 7% decrease in yield (Table 6). Similarly in TOS 2, yield was also impacted by CR, with LRPB Spitfireexperiencing a 0.42 t/ha or 12% decrease in yield due to CR, when all other variables were held constant. Importantly, when contrasting the combined effects of TOS, CR disease pressure and genotype, potential yield decreased by 1.44 t/ha or ~31.5%.

Varying N and P application rates had a limited effect on yield potential, most likely due to the relatively high starting soil N and Colwell P values (Table 2). There was a small but significant (P<.001) response to P application rates (Nil vs 30 units of P) of 0.17 t/ha (3.44 vs. 3.61 t/ha) in TOS 2 (data not shown). Interestingly, when looking at high input (100 kg N/ha + 20 kg P/ha) versus the low input treatments (Nil N and P), there was a 0.31 t/ha or ~7% difference in yield in TOS 1 (data not shown). The only site to show a significant yield response to N application was Tamarang TOS 1 in 2014, with a 9% increase in yield with the application of 100 kg N compared with the Nil treatment.

Other agronomic factors to adversely impact yield potential included plant population at Tamarang in 2014 with the delayed TOS 2 treatment. The low target population of 60 plants/m² was significantly (P<0.05) lower yielding than the 120 and 180 plants/m² treatments, at 3.67 t/ha vs. 4.01 t/ha and 4.04 t/ha respectively, supporting the accepted principal of increasing targeted plant population with delayed sowing times.

Trangie

Experiments were conducted over two consecutive years at Trangie Agricultural Research Centre. In 2014 the experiment was conducted on a Grey Vertosol and in 2015 on a Red/Brown Chromosol. Yield results varied between site and year, ranging from 4.24 t/ha for TOS 1 in 2015, to 1.82 t/ha for TOS 2 in 2014 (Table 5). Timeliness of sowing was again shown to be a key driver of yield, with delayed sowing resulting in a 1.20 t/ha decrease in yield in the across sites analysis, equating to a 34% decline for EGA Gregory, TOS 1 vs TOS 2 (Table 7). When comparing individual site years, delays in sowing time equated to yield declines of 42.5 kg/day in 2014 and 69.1 kg/day in 2015.

Variety selection was found to be a contributing factor to yield potential in TOS 1 with the quicker maturing variety LRPB Crusader out yielding the longer main season variety Sunvale by 0.61 t/ha, with no significant difference between LRPB Crusader and EGA Gregory(Table 7). TOS 2 results for Trangie in 2015, showed that the faster maturing variety LRPB Crusader was higher yielding than EGA Gregory, 3.35 t/ha vs. 2.74 t/ha respectively (data not shown).

Plant populations (40, 80 or 160 plants/ m²) had no effect on yield in TOS 1 at Trangie. In contrast, population had a significant (P<0.001) effect on yield when sowing time was delayed, with yield improving with increasing plant population (Figure 1). The low population treatment (40 plants/m²) being 0.44 t/ha or 17% lower yielding and the 80 plants/m² approximately 8% or 0.20 t/ha lower yielding than the high population (160 plants/ m²), with significant differences between all treatments (Table 7).

Table 7. Effect of main agronomic factors on grain yield potential – Trangie across site analysis

Variety	Population (plants/m2)	Applied N (kg/ha)	Applied P (kg/ha)	Fp (CR+++)	Yield (t/ha)	Yield Gap (t/ha)
TOS 1
LRPB Crusader	80	100	20	0	3.78
Sunvale	80	100	20	0	3.17	-0.61*
EGA Gregory	80	100	20	0	3.531	ns
TOS 2
EGA Gregory	80	100	20	0	2.331	-1.20*
LRPB Spitfire	160	100	20	0	2.58#
LRPB Spitfire	80	100	20	0	2.38#	-0.20*
LRPB Spitfire	40	100	20	0	2.14#	-0.44*

^*>Contrast are significant (P<0.05).  ns = Not Significant. ¹EGA Gregory TOS 1 vs. TOS 2 contrast. ^#TOS 2 contrasts relate to LRPB Spitfire.

Figure 1. Yield response of wheat to plant population and time of sowing - Trangie across site analysis

Nyngan

(Table 8). When comparing individual site years, delays in sowing time for EGA Gregory, resulted in yield declines of 59.5 kg/day in 2014 and 95 kg/day in 2015.