Previous PageTable Of ContentsNext Page

Identifying strategies to minimise climatic risk for grain sorghum, using the SORKAM model

L.J. Wade1, T.J. Gerik2, W.D. Rosenthal2 and R.L. Vanderlip3

1 Queensland Department of Primary Industries, Emerald QLD 4720
2
AES Blackland Research Center, Temple, Texas 76502 USA
3
Department of Agronomy, Kansas State University, Manhattan, Kansas 66506 USA

Summary. The SORKAM model was used to examine the consequences of alternative plant density and row spacing for rainfed grain sorghum in Australia. Simulations were conducted for two crop maturities (early and late) and two levels of maximum capacity of plant available soil water (low and high), for each of 4 plant densities (2.5, 5.0, 10.0 and 20.0 plants/m2) and three row spacings (0.5, 1.0 and 1.5 m), at Katherine, Dalby and Emerald. At Katherine, the best strategy was to seek maximum productivity. Production strategy interacted strongly with season at Dalby and Emerald, but there was a greater scope for manipulating risk at Dalby, because of the higher yields attained there. Simulation provided a powerful basis for evaluating crop management decisions.

Introduction

The SORKAM model (1) was tested against experimental data for rainfed grain sorghum in Australia, and considered suitable for examining the consequences of alternative plant density and row spacing, within its calibration boundaries (3). The model was then used to examine the response of grain sorghum to plant density and row spacing at three locations in Australia. The objectives were to examine how cultural practices may influence variability in grain yield, and to identify management strategies likely to reduce the risk of crop failure.

Methods

Simulations were conducted for three locations: Katherine (tropical, strongly summer dominant rainfall on light soils), Dalby (sub-tropical, more general rainfall on deeper cracking clay soils) and Emerald (semi-arid tropical, generally summer dominant rainfall on shallow cracking clay soils). Default settings of the model chosen for each site were specified elsewhere (3). The model was set to run continuously through 30 years of daily meteorological data for each site, with the following modifications. The earliest and latest feasible dates of planting were specified: 15 December to 30 January at Katherine, 28 September to 1 January at Dalby, and 21 December to 23 February at Emerald. For each season at each site, planting could proceed on the first feasible day in which plant available soil water exceeded 10 cm (6 cm at Katherine), but was less than 90 % of the maximum plant available water capacity (PAWC) in the top 15 cm of soil. Above this capacity in the surface layer, the soil was considered too wet to plant. Losses due to pests, diseases, weeds and poor nutrition were not considered. Harvest was assumed to coincide with physiological maturity, and harvest losses (e.g., lodging, rainfall interference) were ignored.

Simulation treatments and default values were input via the SORKAM parameter menus (1). The treatments chosen included four plant densities (2.5, 5.0, 10.0 and 20.0 plants/m2), and three row spacings (0.5, 1.0 and 1.5 m), which fall within the respective ranges validated in the model and feasible in practice. Crop maturity was simulated by specifying an early or late cultivar in the control menu. The influence of soil type and depth was examined by varying the maximum PAWC: 6.4 and 12.8 cm at Katherine, 11.2 and 22.4 cm at both Dalby and Emerald. Cumulative probabilities were plotted for predicted values of grain yield for each variable at each site. These graphs provided an opportunity to examine, for each location, the proportion of years in which a single crop could be grown, and the likely distribution of grain yield which may result.

Results and discussion

Cumulative distribution functions for PAWC, plant density, and crop maturity by row spacing are plotted in Figures 1 to 3 respectively, for Katherine, Dalby and Emerald. Simulated grain yields ranged from 1.2 to 6.5 t/ha at Katherine, 0 to 9.5 t/ha at Dalby, and 0 to 5.8 t/ha at Emerald. These yield levels are in accord with reported values and grower experience (2,3).

Figure 1. Cumulative distribution functions for sorghum yield based on simulation analysis for two maximum capacities of plant available soil water (- - - low; ---- high) at (a) Katherine; (b) Dalby; and (c) Emerald.

Figure 2.Cumulative distribution functions for sorghum yield based on simulation analysis for four plant densities (- - - 2.5; …. 5.0; - . - 10.0; ----- 20.0 plants/m2) at (a) Katherine; (b) Dalby; and (c) Emerald.

Higher PAWC was advantageous at all sites, with the magnitude of response varying with seasonal favourability (Fig. 1). Greater PAWC was more advantageous in poorer seasons at Katherine, in average seasons at Dalby, and in better seasons at Emerald. This response reflects the increasing importance of water limitation from Katherine to Dalby to Emerald, with direct consequence for the optimisation of production strategy.

Figure 3.Cumulative distribution functions for sorghum yield based on simulation analysis for early and late cultivars in two row spacings, (- - - early in 150 cm; . . . early in 50 cm; - . - late in 150 cm; and ----- late in 50 cm), at (a) Katherine; (b) Dalby; and (c) Emerald.

Productivity at Katherine was favoured by high plant density, late maturity and narrow rows (Figs 2a and 3a). Choice of early maturity in wide rows was of no advantage in the poorest seasons, but resulted in considerably lower yields in most seasons. Even at low PAWC, early maturity and wide rows resulted in reduced yield (Wade, unpublished data). The best strategy was to seek maximum productivity, by growing the late sorghum cultivar in narrow rows at high density.

In contrast, production strategy interacted strongly with season at Dalby and Emerald. At Dalby, low plant density, early maturity and wide rows were favoured in the poorer seasons (Figs a and 3b). The combination of late maturity and narrow rows resulted in higher yields in more favourable seasons, at the cost of crop failure in 30% of years. Choosing early maturity in wide rows would virtually eliminate crop failure, but with a substantial yield discount in many years. Considerable scope exists at Dalby to alter risk exposure by manipulating cultural practice. The favoured strategy should be influenced by the level of debt of the individual farmer. The conservative strategy would maximise the proportion of years in which receipts should cover expenses. If financial commitments permitted the luxury of additional crop failures, choice of late maturity in narrow rows may result in higher average profit. The results provide a quantitative basis for such economic assessments.

At Emerald, manipulation of cultural practice was unable to preclude yield failure in the poorest seasons (Figs 2c and 3c). Low density and early maturity were generally favoured at the lower yield levels attained. Row spacing was of little consequence at low PAWC, but closer rows were favoured at high PAWC (Wade, unpublished data). The advantage of early maturity was reduced in more favourable seasons. Row spacing did not greatly influence yield stability, with wide rows failing to alter the proportion of crop failures for early maturity, and only reducing crop failures by 7% for late maturity. The results provide limited scope for adjusting risk through manipulation of cultural practices at Emerald, because of the low yield levels attained. Current cultural practices appear to be the best compromise (2).

The presented evidence permits a quantitative assessment of risk associated with alternative cultural practices for different locations and plant available water capacities. We believe that crop simulation models provide a powerful basis for evaluating crop management decisions, in order to identify strategies which minimise climatic risk.

References

1. Rosenthal, W.D., Vanderlip, R.L., Jackson, B.S. and Arkin, G.F. 1989. TAES Computer Software Documentation Series MP1669. Texas A&M University, College Station, Texas, USA. 205 pp.

2. Wade, L.J. and Douglas, A.C.L. 1990. Aust. J. Exp. Agric. 30, 257-264.

3. Wade, L.J., Myers, R.J.K. and Foale, M.A. 1991. In: Climatic Risk in Crop Production: Models and Management for the Semiarid Tropics and Subtropics. (Eds R.C. Muchow and J.A. Bellamy) (CAB International: Wallingford, United Kingdom). pp. 263-282.

Previous PageTop Of PageNext Page