Table Of ContentsNext Page

A crop simulation model for sorghum

G.L. Hammer and R.C. Muchow

QDPI/CSIRO Agricultural Production Systems Research Unit, PO Box 102, Toowoomba QLD 4350
CSIRO Division of Tropical Crops and Pastures, 306 Carmody Road, St. Lucia QLD 4067

Sorghum, Sorghum bicolor (L.) Moench., is one of the major summer crops grown in the subtropics. The high rainfall variability and limited planting opportunities make crop production in these regions risky. A robust crop simulation model can assist farmer decision-making by quantifying the production risk. Accordingly, we developed a simple, yet mechanistic crop simulation model for sorghum (1) for use in assessing climatic risk to production in water-limited environments.

We required a generally applicable crop model capable of predicting sorghum yield accurately in diverse situations. This was achieved by using the "top-down" approach to model design, in which additional detail is incorporated in the model only if the need can be demonstrated by improved model performance. Whilst this approach avoids unnecessary complexity, biophysical rigour is maintained by considering general phenomena controlling crop growth and development. The model simulates grain yield accumulation from the product of biomass accumulation and harvest index, which increases linearly with time after anthesis. Low temperature and severe water limitation restricted seed set and harvest index increase, respectively. Daily increase in above-ground biomass was calculated as the product of incident radiation, the fraction of that radiation intercepted by the crop, and the efficiency with which the intercepted radiation was used to produce biomass. Biomass increment was restricted when soil water was limiting. Crop leaf area was calculated as the product of leaf area per plant and plant density. Leaf area per plant was determined as the difference between total leaf area produced per plant and that senesced. Both total and senesced leaf area per plant were related to a function of thermal time from emergence and total leaf number. Leaf area production was restricted when soil water was limiting. The timing of each phenological stage was predicted from known responses of rate of development to temperature and/or photoperiod. A soil water balance, which accounted for additions from infiltration and losses from soil evaporation, transpiration, and drainage, was maintained. The model uses a daily time-step and readily available weather and soil information and assumes no nutrient limitation.

The model was tested on data from experiments spanning environments in the semi-arid tropics and subtropics. Potential limitations in the model were identified and examined in the testing procedure by using combinations of predicted and observed data in various modules of the model. The model performed satisfactorily, accounting for 94% and 64% of the variation in total biomass and grain yield, respectively. The difference in outcome for biomass and yield was caused by limitations in predicting harvest index.


1. Hammer, G.L. and Muchow, R.C. 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, UK). pp. 205-232.

Top Of PageNext Page