Sampling Techniques for Soil Fertility Evaluation: An Exercise in Understanding Variability

Abstract

It is difficult to obtain a representative soil sample from fields with areas of very low to very high soil test values. This field exercise was developed to assist agricultural practitioners with understanding common sampling schemes and their limitations with highly variable soils. A 2.8-ha site with high variability in soil test P and K was selected for use at the University of Minnesota's Extension Field School. Instructors provided background site information including previous crops and yields, factors of soil formation, methodology used in obtaining a composite sample, a representative management zone sample, precision grid sample, and a soil survey map. Before the exercise, 36 soil samples were collected on a 30-m grid and analyzed for P, K, and pH. Each grid point was numbered and flagged. Participants were split into groups, assigned a sampling method, and conducted a site evaluation and determined which grid points to include in their sample(s). Each group was limited to choosing eight grid points to collect their sample. Although P deficiency was evident, approximately half of the groups recommended no fertilizer application due to inclusion of high testing areas in their samples. Overall the grid sampling method resulted in the highest fertilizer recommendation for the site. The management zone method recommended the least fertilizer overall but most of it was applied in low testing areas, and the composite method which often contained nonuniform areas within the sample resulted in fertilizer recommendations unrepresentative of the site.