Context and objectives
Countries with large agricultural production have a financial interest in production forecasting as production levels have a sensible impact on the product prices at European scale. For Belgium, which is rather a small agricultural producer within EU, this argument won’t stand up. This probably explains why so little was done for agricultural production forecasting up to now. Nevertheless, production-forecasting methods might be useful to solve several other problems concerning Belgian agriculture.(e.g.: assessment of the extent of calamity zones due to exceptional climatic conditions, control of available statistics on crop productions, stock management strategies).
All these politico-economical reasons as well as the scientific aspects of this topic justify that Belgium should develop its own Crop Growth Monitoring System in order to provide reliable, timely and objective estimates of crop yields and calamity monitoring at regional scales (Belgian agricultural regions and agro-statistical circumscriptions). This was the aim of the Belgian Crop Growth Monitoring System project (B-CGMS). The final product of this research was to provide the Ministry of Small Enterprises, Traders and Agriculture and the National Statistical Institute with a renewed and enhanced tool for the prediction and the estimation of crop yield levels.
Expected scientific results
The main result of the research is the establishment of an operational crop growth monitoring system adapted to the Belgian context. Six major annual crops are simulated by the system: winter wheat, winter barley, fodder maize, winter rape seed, potatoes and sugar beet. Final estimation and prediction of yield for agricultural regions and circumscriptions are obtained by combining the technological trend functions with the outputs of the agrometeorological model and the biomass index derived from satellite imagery. For winter crops best predictions or estimations were obtained by the integration of the three components of the system whereas for sugar beet and rape seed only the technological trend still gives best results.
It was shown that images from AVHRR and VGT do contain relevant information on the yield of the main crops and that linear unmixing of the compound signal can be best applied on the VGT-images because of their higher geometric precision. Nevertheless this topic still requires further investigation before it can be integrated in a fully operational yield-forecasting scheme.
The chosen approach also provides more efficient and more precise estimations of the acreage of the different crops during all the campaign. The ability of the System was checked in a near-operational way in may 2000 with the publication of the first agrometeorological bulletin giving yield and production forecasting for the last decade of May 1998 (test on historical data only). An interactive interface (http://b-cgms.cragx.fgov.be/) has also been developed to allow an easier use of the System by potential end-users.
|Project leader(s):||ULg - Département des Sciences et Gestion de l'Environnement|