Field Soil Sampling Density for Variable Rate Fertilization

Variable rate fertilizer application being commercially performed today is most often based on a soil test map. The sampling density used to develop a map is often selected without background information regarding field soil test variability. The objective of this study was to determine how many samples should be taken from a field in order to locate and describe major areas of fertility affecting variable rate fertilizer applications. Two 40 acre fields were sampled in an 82.5 ft grid each fall from 1989 to 1992. Soil pH, Bray PI, and available K levels were determined on each sample and maps were made using inverse distance squared estimates. Data were taken from the samplings to represent a 165 ft and 330 ft grid pattern. Maps were developed from these less dense grids and compared with the 82.5 ft grid values. In 1992, a separate 220 ft grid sampling was taken. The 220 ft grid estimates were more highly correlated with the 82.5 ft grid values than were the 330 ft grid estimates, however, membership of 220 ft and 330 ft grid estimates within soil test categories were similar. Fertilizer P and K applications were made in one field following the 1992 sampling. Spring 1993 sampling showed the success of the 220 ft grid in directing a variable rate application of P and K. Comparisons to theoretical P and K applications directed by a 330 ft grid map showed the superiority of the 220 ft grid compared with the 330 ft grid. Research QuestionVariable rate fertilizer application is being commercially performed today. Fertilizer application is most often based on a soil test map. The sampling density that is used to develop a map is often picked without adequate background information of field soil test variability. There is little information to suggest how many samples would be necessary from a field to represent the major fertility areas. The objective of this study was to determine how many samples should be taken from a field in order to locate and describe major areas of fertility affecting variable-rate fertilizer applications. Two 40 acre fields were sampled in an 82.5 ft grid to describe the major pH, P, and K areas, and then these values were compared in different ways with maps made using less dense grid patterns. Three less dense grids were evaluated to determine which adequately described the original values at the least dense sampling interval. Literature SummarySoil samples have been taken in the past to describe both a field mean and field variability. The greater the sampling density, the more is known of the field variability. The success of a variable-rate application depends on the ability of the sample grid values to describe the field so that underfertilization and overfertilization are minimized at a reasonable sampling cost. Systematic sampling has been recommended in Illinois for many years, however, recommendations have focused on the need for general variation knowledge, but were not intended to direct a variable rate application. A Wisconsin study has shown that a 200 ft sampling grid may be needed to describe field variability, while variable rate fertilizer grid sampling in Washington state also uses about a 220 ft grid for irrigated crops. Study DescriptionTwo 40 acre fields were sampled in an 82.5 ft grid each fall from 1989 to 1992. Soil pH, P1, and available K levels were determined on each sample and maps were made using inverse distance squared estimates. Sample values were taken from the data of each sampling that corresponded to its position in 165 ft and 330 ft grids. Maps were developed from these less dense grids and compared with the 82.5 ft grid values. In 1992, a second sampling was taken in a 220 ft grid. Correlations between the 220 ft and 330 ft grids with the 82.5 ft grid were made in 1992. Soil test categories were defined and the categories of the estimated values for 82.5 ft grid locations taken from the less dense grids were compared for 1992. A fertilizer P and K application was made in 1992 following sampling based on a 220 ft grid map. The field was sampled in March 1993 to evaluate results. A theoretical fertilizer application based on a 330 ft grid map was also calculated and evaluated. Applied QuestionHow well did 165 ft, 220 ft, and 330 ft grid estimates correlate with the 82.5 ft grid sampling? Both the 165 ft and 220 ft grids were generally better correlated with the 82.5 ft arid samdinn than the 330 ft nrid. There was not as much difference, however, between the number of correct soil test category comparisons of the 220 ft and the 330 ft grids. How well did the fertilizer application based on a 220 ft grid P and K soil test map raise P and K levels compared with a theoretical fertilizer application based on a 330 ft grid P and K map? The fertilizer application based on the 220 ft grid map correctly fertilized most areas that were low in P and K. A fertilizer P and K application based on the 330 ft grid mapping would have resulted in larger areas still needing additional fertilizer. RecommendationsGrowers expect the sampling grid to represent major fertility areas so that yields can be maximized and costs for fertilizer minimized in a large portion of a field. Adequate sampling density is needed to show fertility areas which may have a significant impact on yields and costs due to under or overapplication of fertilizers. In this study, a 220 ft grid successfully guided a variable-rate fertilizer application on a highly variable field. A 330 ft grid was not as successful and would have resulted in underapplication of a sizeable area of the field. Growers beginning variable-rate fertilization of pH, P, and K would benefit from a grid of 220 ft, rather than a less dense grid.