Crop yield forecasting at national and regional levels using remote sensing techniques
Crop yield forecasting models are needed to help farmers and decision makers cheaply detect crop condition early enough to assess and mitigate its impacts on grain production. A precise estimate of crop production requires an accurate measure of the total cultivated area and well-established knowledge of crop yield. The first requirement is no longer a problem as is technically solved through various techniques such as area frame sampling. With respect to the second, great efforts have been made to find an accurate definition of the crop yield with respect to the actual factors that shape its growth through out the season. Agrometeorological models have found a wide range of applications in agricultural research and technology and are playing an increasing role in translating information about climate variability into assessments, predictions and recommendations tailored to the needs of agricultural decision makers. However these models have generally been developed and tested for application at the scale of a homogeneous plot. They are criticized for their inability to address large-scale yield estimates at regional or even national levels in addition to their high cost of application. This is because field conditions during the period of crop establishment at the regional scale may be quite variable and poorly represented by standard parameter values of the crop model.