Characterisation of the structure of spatial variability for initial pesticide concentration in a field trial

The spatial variation in the concentration of four herbicides in two trials with maize (metolachlor, terbuthylazine) and winter wheat (pendimethalin, isoproturon) was assessed during the treatment, using classical statistical and geostatistical techniques. Spray trap and soil samples were collected immediately after the treatment to evaluate the herbicide concentration when the solution reaches the soil and the initial herbicide soil concentration, respectively. The results show that the values of the herbicide concentration at the moment of treatment are lower than expected, indicating that under field conditions only a part of the pesticide reaches the soil during the distribution. A decrease of CV in spray traps in winter wheat trials is partially explainable by the greater accuracy used in treatment planning and the better weather conditions. Geostatistical analysis was undertaken on the herbicide concentration in the spray traps, as 80 samples per herbicide were collected, while there were not enough samples for initial soil concentration to perform this kind of analysis. The data showed a spatial correlation only for maize herbicides. The data are discussed in order to establish a more efficient soil sampling strategy.