The field of precision agriculture (PA) uses observations of a production system (eg yield maps, remotely sensed imagery), supplemented by explanatory data layers (eg soil maps, elevation models) to identify zones for which differential management may be appropriate. Differential management aims to improve yields and reduce production inputs (e.g. water, fertilizer).
We evaluated the application of PA to the task of planning the layout of vineyard blocks in a project involving the renovation (replacement of old vines with new) of a Clare Valley vineyard.
We collated high resolution spatial information such as yield maps, aerial imagery, EM38 and RTK GPS survey to produce maps of soil conductivity, and a digital elevation model. Rather than traditional soil sampling on a fixed grid pattern, we stratified sampling on the basis of the high resolution datasets.
Broad scale patterns evident in the soil conductivity map correspond to broader scale geological structure of the area. Additionally, the yield maps and imagery showed significant patchiness, possibly due to disease in the existing vines. k-means clustering was performed to derive areas with similar characteristics.
Smoothing the high resolution datasets reduced the influence of the broader scale geological patterns and localised patchiness, and resulted in the most useful cluster solutions. This suggests that the spatial resolution at which these data are useful for design purposes may not be the same as are needed for management purposes. This research shows that spatial analysis and PA data can make a valuable contribution to vineyard re-design.