ABSTRACT
Influencing points in agricultural spatial analysis may change considerably results on spatial dependence and hence map building. With regards to physico-chemical soil properties and crop yield, such maps should efficiently estimate current field conditions, being important for an agricultural site-specific management, optimizing thus input applications in order to increase yields. This study aimed to analyze hair-plot graphic techniques, with local influence (Ci and |Lmax|) to identify influencing points within a set of georeferenced spatial continuous data. These information were gathered from an experimental area with 167.35 hectares, wherein an agricultural site-specific management has been adopted. As a result, we obtained potentially influencing points and then outlined maps with and without the use of them. By comparing both maps, we could note by metric comparison that it is of major importance to identify those points on a spatial database. Thus, such investigations must be carried out to understand cases of unusual performance, since they considerably modify the generated maps.
geostatistics; hair-plot; local influence; maximum likelihood; chemical soil properties
