Previous PageTable Of ContentsNext Page

Improving the reliability of sorghum production in the farming system

G.J. Butler1, S. Cawthray2, M. Castor3, S. Yeates4 and T. Christian1

1NSW Agriculture, Tamworth Centre for Crop Improvement, RMB 944, Tamworth, NSW.
2CSIRO/Agricultural Production Systems Research Unit, Toowoomba, QLD.
3Michael Castor & Associates, Goondiwindi, QLD.
4CSIRO, Cotton CRC, PMB 44, Winnellie, NT.

ABSTRACT

Discussions between agribusiness, farmers and researchers identified the potential of altering row configuration to improve the profitability and reliability of sorghum in the lower rainfall zone of the northern grain region. The traditional method of growing sorghum on one-metre rows (SP) was compared with single skip (SS) and double skip (DS) treatments. Replicated on-farm research trials were conducted in the average to above-average rainfall seasons of 1998/9 and 1999/00 in northern NSW and southern QLD. Yields were in the range 2.2-5.5 t/ha and generally SP>SS>DS. Water was conserved till flowering and grain filling in SS and DS treatments. Further research is necessary before APSIM can be used to evaluate skip configurations in sorghum.

KEY WORDS

Sorghum, skip, row configuration, water, maturity.

INTRODUCTION

In an environment with abundant plant-available water, maximum economic yield is the obvious goal for sorghum production. However, pursuit of this single goal increases the risk of crop failure when rainfall is low, with sometimes damaging consequences on farm profitability (e.g., experimental sorghum yields in the 1976-78 study of Thomas et al. (8) in Qld ranged from 0-3.8 t/ha). An alternative approach to maximising yield is to minimise the likelihood of crop failure in almost all years and accept a possible yield penalty in the good seasons. A suggested method of achieving this is to alter row configuration using skipped (unplanted) rows to conserve water during the vegetative stage of crop development for use during flowering and grain filling. This management strategy has been used to enhance productivity of peanuts (7) and to increase gross margins for dryland cotton production (4).

Row-width research in sorghum has generally shown that narrow rows (<0.5 m) outyield wider rows (>0.5 m) except when moisture is severely limited (5, 8). Myers and Foale (6) quantified this relationship by calculating that only when yield was less than 1.1 t/ha would wide rows (> 0.75m) outyield narrower rows. However, Blum and Naveh (1) showed that sorghum grown on 2 m double rows could outyield sorghum on 1 m rows at yields of up to 5 t/ha, provided plant density was maintained.

Most sorghum in northern NSW and southern Qld is currently grown on 1m rows to buffer against variable soil moisture and rainfall. There is now considerable interest in, and some adoption of, skip row configurations. It was clear in meetings attended by growers, agribusiness and researchers in NSW and QLD that the benefits or costs of this management strategy needed to be properly evaluated. Thus, on-farm research was undertaken to compare different row configurations for sorghum production and to provide data for simulation modelling using APSIM for broader-scale (seasonal) evaluations.

MATERIALS AND METHODS

Field experiment

During the two summer-growing seasons of 1998/9 and 1999/00, six experiments were conducted in northern NSW and southern QLD. Site and trial details are summarised in Table 1. All sites contained 3 replicates except M9899 (2 replicates). Main treatments were solid plant (SP) sorghum sown on 1-m rows, single skip (SS), in which every third row is unplanted, and double skip (DS), which has two planted rows adjacent to two unplanted rows. At each site, the SP, SS and DS row configurations were used for either the quick-maturing variety, MR Goldrush (G) and/or the medium maturity variety, MR Buster (B). For the NS9899 and NM9899 experiments, plant densities within rows were standardised, resulting in different area densities; at the other 4 sites, area density was standardised across treatments (different intra-row densities) (Table 1). Note that the NM9900 SP-G population was only half that of the other treatments in that experiment. All trials were sown, sprayed and harvested using the cooperating farmer's machinery.

Table 1. Site and experimental details for sorghum at six sites in northern NSW and southern Qld in the 1998/99 and 1999/2000 seasons.

Location

Year of Trial

Soil type

PAWC (0-1.5 m) (mm)

Treatment

Planting Date

Plot Size (m) †

Plants established (no/ha)

North Star, NSW (NS9899)

1998/9

Red vertosol

140

SP-G
SS-G
DS-G

28/9/98

12 x 50
(9.3 x 50)

40000
30000
20000

North Moree, NSW (NM9899)

1998/9

Grey vertosol

190

SP-G
SS-G
DS-G

27/9/98

16 x 50
(9.3 x 50)

50000
39000
28000

Meandara, Qld (M9899)

1998/9

Red vertosol

135

SP-G
SS-G
DS-G
SP-B
SS-B
DS-B

6/1/99

8 x 100
(8 x 100)

46000
62000
46000
61000
64000
54000

Croppa Creek, NSW (CC9900)

1999/00

Grey vertosol

254

SP-G
SP-B
SS-B
DS-B

20/10/99

16 x 120
(11 x 50)

61000
61000
60000
58000

North Moree, NSW (NM9900)

1999/00

Grey vertosol

250

SP-G
SP-B
SS-B
DS-B

28/9/98

16 x 50
(9.3 x 50)

22000
44000
44000
44000

Meandara, Qld (M9900)

1999/00

Grey vertosol

250

SP-G
SP-B
SS-B
DS-B

29/9/99

16 x 100
(8 x 100)

63000
70000
64000
60000

† Area harvested in parenthesis

Plant and soil measurements

Plant density was estimated from 10 x 5m lengths of row in each plot. Grain yield was determined by harvesting one header width of the length of the plot, then weighing the grain using an electronic weigh bridge. The size of the harvested area for each trial is in Table1.

Soil water to a depth of 1.5 m was estimated gravimetrically. Soil cores were taken at flowering and harvest from between (0.5m from planted row) and within planted sorghum rows in all treatments (SP, SS, and DS). Samples were also taken 1m from the planted row in the middle of the skip for SS and 1m and 1.5 m from the planted row for DS. Plant-available water capacity (PAWC) and plant-available water (PAW) was calculated using soil bulk density, drained upper limit and crop lower limit as outlined in Dalgleish and Foale (3). Plant area at flowering and harvest (PAF and PAH, respectively) were defined as the areas bounded by the planted row and extending to 0.5m from the row. Plant-free area at flowering and harvest (PFAF and PFAH, respectively) were defined as the area more than 0.5m from the planted row (applies only to SS and DS treatments). For SS and DS, the PAW was estimated for a point 0.5m into the skip (average of within-row PAW and the PAW 1m into the skip). In-crop rainfall (Table 2) was estimated as rainfall that fell between planting and 3 weeks prior to harvest. The chance of achieving a better season (4th column, Table 2) was determined using chance-of-rain curves generated using RAINMAN (2).

APSIM was calibrated using CC9900 data. Simulated yields, with no harvest losses, were approximately 0.5t/ha higher than observed machine-harvested yields. Simulations were calibrated for a grey vertosol, with unlimited N, 50,000 plants/ha, with half or full starting soil-water profiles, and run for 100 years at Goondiwindi.

RESULTS AND DISCUSSION

Yield, in-crop rainfall, harvest index and PAW at flowering and harvest for the six trials are presented in Table 2. The 1998/99 and 1999/2000 seasons were average or above-average, with respect to rainfall. DS yields were lower than SS or SP yields at all sites. In the case of M9899, yields of DS-B and SS-B were significantly higher than DS-G and SS-G (P<0.5), respectively. Solid plant, using either Goldrush or Buster, was the highest yielding treatment at all sites. Buster outyielded Goldrush in 4 of the 6 comparisons. Harvest indices of 0.37 to 0.57 suggested that water stress was not severe at flowering or grain filling at any site. Harvest index was relatively constant at each site except NS9899 (0.37-0.51).

Available soil water was higher in the PFAF than in the PAF and there was generally more water in the PFAF in double skip compared to single skip. Apart from NS9899 (where there was late rainfall), soil water in the PAH and the PFAH was less than in the PAF and the PFAF respectively. Available soil water in double skip was higher than single skip in the PAH and the PFAH.

Table 2. Summary of yield, soil water and rainfall data for sorghum at six sites in northern NSW and southern Qld in the 1998/99 and 1999/2000 seasons.

Site and Treatment

Yield†
(t/ha)
(% of SP-G)

In-crop Rain (mm)

Chance of achieving a better season ‡

Harvest Index

PAW PAF (mm)

PAW PFAF (mm)

PAW PAH (mm)

PAW PFAH (mm)

NS9899 SP-G
SS-G
DS-G


3.66a (100)
3.73a (102)
3.22b (88)


226


50


0.37
0.51
0.48


22
34
46



37
61


37
38
50



42
61

NM9899 SP-G
SS-G
DS-G


4.70a (100)
3.92b (83)
3.29c (70)


300


27


0.43
0.40
0.41


53
71
55



79
72


20
49
50



57
71

M9899 SP-G
SS-G
DS-G
SP-B
SS-B
DS-B


4.32a (100)
3.17c (73)
2.22d (51)
3.83b (89)
3.62b (84)
3.06c (71)


260


28


0.53
0.53
0.55
0.53
0.53
0.54


24
47
40
22
51
36



72
74

70
80


0
4
15
10
22
38



12
8

19
39

CC9900 SP-G
SP-B
SS-B
DS-B


5.17a (100)
5.09a (98)
4.28b (83)
3.45c (67)


284


24


0.57
0.57
0.55
0.53


119
143
135
142




151
190


21
26
44
57




51
103

NM9900 SP-G
SP-B
SS-B
DS-B


4.48c (100)
5.48a (122)
4.78b (107)
4.03d (90)


296


28


0.46
0.49
0.48
0.52


130
113
105
104




133
227


62
57
45
47




46
122

M9900 SP-G
SP-B
SS-B
DS-B


4.41b (100)
5.20a (118)
4.45b (101)
3.60c (82)


302


16


0.55
0.54
0.56
0.54


n.d.


n.d.


n.d.


n.d.

† Different letter denotes significance (5% LSD)
‡ Percent chance of obtaining higher total rainfall within the growing season

When water was not limiting at flowering, yield differences were related to the number of heads per unit area (data not shown). However, water stress at flowering and grain fill at NS9899 resulted in SS having more grains/head and larger grain, compensating for lower number of heads per unit area, compared to SP (data not shown).

Fig1. APSIM-simulated double and single skip sorghum yields compared to solid plant for a range of yield levels

The APSIM simulations suggest at yield levels greater than 3.7 t/ha, SP>SS>DS. Trial results agreed with this relationship except when different maturity varieties for SP and SS were included. At yield levels less than 3.7 t/ha, simulations suggested DS>SS, which did not agree with our limited data, although clear separation between DS and SS did not occur until yield was <2.5 t/ha (Fig1). Further research to calibrate APSIM for skip configurations is being undertaken.

Sorghum has traditionally been grown on 1m rows using a quick maturing variety. In better than average seasons using a longer season variety, yields were reduced in DS by 18-33%, compared to losses of between 0-18% in SS. In these seasons, DS and SS were unable to use all the water in the skip area. However there is potential for this unused water to increase both the likelihood and profitability of double crop options. Double crop chickpeas at CC9900 were taller and had more pods in the skip rows (Clark, Pers. Comm).

CONCLUSIONS

Altering row configuration in sorghum while maintaining population density can preserve soil moisture for use at the flowering and grain filling stages of crop development. In better than average seasons this management strategy will result in yield losses, although selection of a longer season variety for use in skip configurations is likely to reduce these losses. The potential for improved yields in the following crop from residual water in the skip area must be considered. APSIM results were in general agreement with trial data. The simulations suggested there were yield benefits using skip rows in below average seasons. Therefore, trials are being conducted to assess the impact of altering row configuration in a low rainfall season.

ACKNOWLEDGMENTS

We gratefully acknowledge the involvement of cooperating farmers Philip and John Coggan, Malcolm Lillyman, Alan Taylor, Allan Hunter and James Clark. This research was supported by GRDC through the Eastern Farming Systems Project (DAN363).

REFERENCES

1. Blum, A. and Naveh, M. 1976. Agron. J., 68, 111-16

2. Clewet, J.F., Smith, P.G., Partridge, I.J., George, D.A. and Peacock, A. 1999. AUSTRALIAN RAINMAN Version 3, Department of Primary Industries Queensland.

3. Dalgliesh, N.P. and Foale, M.A. 1998. 'Soil Matters', APSRU, Toowoomba, Qld.

4. Goyne, P.J. 2000. Final Report CRDC, February 2000, QDPI.

5. Holland, J.F. and McNamara, D.W. 1982. Aust. J. Agric. Anim. Husb., 22, 310-16

6. Myers, R.J.K. and Foale, M.A. 1981. Field Crops Res., 4, 147-54

7. Schubert, A.M., Pohler, C.L. and Smith, D.H. 1983. Texas Agric. Exp. Stn. Prog. Rep. 4058.

8. Thomas, G.A., Myers, R.J.K., Foale, M.A., French, A.V., Hall, B., Ladewig, J.H., Dove, A.A., Taylor, G.K., Lefroy, E., Wylie, P. and Stirling, G.D. 1981. Aust. J. Agric. Anim. Husb. 21, 210-17

Previous PageTop Of PageNext Page

...

Search Site