Previous PageTable Of ContentsNext Page


P.J. Goyne1, D.G. Butler2, and J.M. Hare1

1Department of Primary Industries, Hermitage Research Station, MS 508, Warwick, Qld 4370
Department of Primary Industries, PO Box 102, Toowoomba, Qld 4350

To capture expanding barley markets and fulfil contracts some forward estimate of production is required. A feasibility study has been undertaken to investigate the development of a system to forward estimate production in the northern Australian cereal belt.


As the 1994 drought precluded a real time study, a retrospective analysis was made of the barley receivals at grain depots of the Jondaryan shire on the Darling Downs for the 1988 season.

A grid with points 5 km apart was superimposed on the shire. Each point, recognised from its longitude and latitude, was matched with weather data: daily rainfall, temperature and radiation. A soils map was overlaid and the soil profile water characteristics at each point determined. Barley yield (cv. Grimmett) was simulated, using the barley model QBAR (2), running in the APSIM environment (3) for each grid point identified from a 20 September 1988 Landsat satellite TM image as being cropped. Planting dates for the simulations were determined from the rainfall events, during May and June. Spatial data were processed using the ARC/INFO (1) geographical information system.


Because the type of crop could not be determined, all cropped fields identified on the satellite image were assumed to be barley. Hence the simulated total barley production for the shire was not calculated. In real time forecasting reports from personnel in the field, together with satellite imagery would enable accurate estimates of cropped area to be made. However, the mean simulated yield across the shire was determined (1.12 t/ha) and compared with the mean yield from depot intake (0.94 t/ha). Because the depots could receive grain from both within and outside the Jondaryan shire, comparisons between depot intake and simulated yields are not precise. Nevertheless, the model performed well when compared to these data. In real time forecasting, current, in conjunction with historical weather data, would be used for periodic yield estimates as the crops progress. This study has shown that the necessary technology is now available to test a full scale real time crop forecasting system for barley based on QBAR.


The project was partly funded by the Grains Research and Development Corporation.


1. Environmental Systems Res. Institute, Inc. 1992. ERSI, Redlands, CA.

2. Goyne, P.J., Meinke, H., Milroy, S.P. and Hare, J.M. 1995. (in preparation).

3. Hargreaves, J.N.G., Holzworth, D.P., and Huth, N.I. 1993. Proc. 7th Aust. Agronomy Conf., Adelaide. 385 p.

Previous PageTop Of PageNext Page