Yield forecasting for marketers

Abstract

Marketers rely on early and accurate yield forecasts to increase industry profitability by improved forward selling strategies. Crop forecasts are required 6-7 months prior to the commencement of harvest. These forecasts need to be updated regularly during the growing season. In this paper, we describe how crop growth models and remote sensing models can be used to provide Queensland Sugar Limited with yield forecast information. An assessment demonstrating how these approaches would have performed in 'forecast mode' using historical yields for the Mackay terminal region is presented. Each method has varied strengths and weaknesses. For this region, the remote sensing model has produced more accurate yield forecasts than the crop growth model. However, the remote sensing model cannot be used until later in the growing season. Conversely, an advantage of the crop growth model is that it can be utilised much earlier in the season when marketers have more flexibility in planning. While these desktop models are still under development, and need to be benchmarked against forecasts that have been provided historically by industry, the crop and remote sensing models offer marketers advance knowledge about the size of the forthcoming crop. This information is invaluable to marketers who manage the forward sales of sugar.