INTERNATIONAL JOURNAL OF CLIMATOLOGY Int. J. Climatol. 23: 1211–1218 (2003) Published online 9 July 2003 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/joc.920 USING SOUTHERN OSCILLATION INDEX PHASES TO FORECAST SUGARCANE YIELDS: A CASE STUDY FOR NORTHEASTERN AUSTRALIA Y. L. EVERINGHAM, a,b, * R. C. MUCHOW, c R. C. STONE d and D. H. COOMANS b a CSIRO Sustainable Ecosystems, Davies Laboratory, Townsville, Queensland 4814, Australia b School of Mathematical and Physical Sciences, James Cook University, Queensland 4814, Australia c SRDC, PO Box 12050, Elizabeth Street, Brisbane, Queensland 4002, Australia d Queensland Centre for Climate Applications/Queensland Department of Primary Industries, 203 Tor Street, Toowoomba, Queensland 4350, Australia Received 13 April 2002 Revised 10 March 2003 Accepted 19 March 2003 ABSTRACT Climate is a key driver of sugarcane productivity. Advance knowledge of the likely climate and its impact on production could add value to production, harvest and marketing efforts. A climate forecast system that incorporates five patterns or phases of the southern oscillation index is used to assess whether an early indication of sugarcane yield anomalies in Australia can be produced. Results indicate that, for certain sugarcane growing regions, the climate forecast system offers better estimates of the direction of the anomaly when compared with no climate forecast system. Improved results of the direction of the yield anomaly can be obtained some 7 months prior to the commencement of harvest. This information can then be used by marketers to plan better the customer allocations, shipping schedules and storage requirements for the next season. Advance knowledge of the crop size can also assist industry decision makers in scheduling when the harvest season should commence. Further research is required to determine if other climatic indices, such as sea surface temperatures, can improve yield estimation. Consideration also needs to be given to determining whether the magnitude and the direction of the yield anomaly can be more accurately forecasted with varying lead times. Copyright  2003 Royal Meteorological Society. KEY WORDS: SOI; climate; forecast; sugarcane; yield; Australia; Monte Carlo 1. INTRODUCTION Knowledge of sugarcane crop size impacts across the farming, harvesting, milling and marketing sectors of the sugar industry. Climate is a key driver of sugarcane productivity (Muchow et al., 1997), so, it may be possible, therefore, to forecast crop size using climate forecasting systems. With a sufficient lead time, forecasts of crop size could then be used to guide strategic planning of seasonal operations in all sectors of the industry. The sugarcane harvesting season in Australia commences around the month of June. To start planning for the coming season, however, industry decision makers, such as mill managers and marketers, begin to estimate the size of the crop around November in the year before harvest commencement. In some years, initial yield estimates made at that time correspond quite closely to the total crop size observed after harvest. For example, in 1995 the estimates were within 0.5% of the actual realized crop. In contrast, in 1991, 1992, * Correspondence to: Y. L. Everingham, CSIRO Sustainable Ecosystems, Davies Laboratory, Townsville, Queensland 4814, Australia; e-mail: yvette.everingham@jcu.edu.au Copyright  2003 Royal Meteorological Society