Historically, nitrogen (N) fertilisers have been applied to cotton fields in uniform quantities to remedy nitrogen deficiencies. These applications have been based on fertiliser recommendations derived from the results of field trials conducted using experimental techniques designed to minimise the impact of spatial variation. However, the applicability of these recommendations when up-scaled to a highly variable environment such as a cotton farming system has been questioned, as very few fields exhibit no spatial variability. This study quantifies the amount of soil variability present within 5 cotton fields in eastern Australia and discusses the ramifications this variability has for both soil sampling undertaken to aid in soil management and on the efficacy of current fertiliser applications.

It was evident from this investigation that sampling in different regions of each field could provide vastly contrasting estimates of the field average N and hence, the amount of fertiliser required. Furthermore, it showed how difficult it is with an economically feasible number of soil samples to determine the optimum fertiliser rate given the variability that is present. In view of this, consideration should be given to appropriateness of fertiliser recommendations that do not contain a spatial component and on the methodology used to determine the N status of the soil.