ABSTRACT
ABSTRACT Satellite observations are addressing agricultural demands at various scales: from the plant level to the national level. Such observations, however, have their specific accuracy, in terms of both their attribute value and their location coordinates. Moreover, their intended use is different. In this chapter, we study the role of satellite images at the two extreme cases. We first address mapping at the field scale, where subpixel resolution mapping is done to identify individual plants. Second, yield mapping is performed at the national level where upscaling is done using geographically weighted regression. The study shows how at these extreme cases, skillful spatial statistical methods collect information that may be useful for decision making, and what the role of spatial data quality is. It concludes that the quality of the spatial information is to be further improved in order to have even better statements than to date.
