An evaluation of soil chemical properties variation in northern and southern Indiana
Abstract
Soil spatial variability is a major problem affecting the reliability of soil testing for fertilizer recommendations. The challenge is how to obtain a representative sample of the area. This study examines the spatial variation of some common soil chemical properties and compares the methods of soil sampling by field, sampling by soil type and using the kriging technique in terms of relative precision and efficiency. Seven Indiana fields were selected to provide a range of contrasts in soil variability. Soil samples were taken using grid sampling. Soil properties such as pH in water, SMP buffer pH, available P and exchangeable K, Ca, Mg and CEC by summation were determined. Conventional statistics and geostatistical methods were used to analyze data. Geostatistics was used to construct semivariograms for those properties which exhibited spatial variation. This has helped to develop contour maps of P and K, thus dividing the field in different fertilizer management zones. Isotropic semivariograms of P, K, Ca, Mg and CEC revealed spatial structure in all locations but one. This indicates that spatial variability of soil properties can exist even within a field composed of one mapping unit. The study also indicated that the number of samples required to estimate a soil property with a given level of precision can be reduced if spatial variability is considered. In some landscapes, sampling by soil type can greatly improve precision. Economical comparison of four fertilizer management practices indicated that farming by field, by kriging and by sampling unit can reduce to a certain extent the percent of area under and over fertilized, but the cost of services for the two last technologies is prohibitive.
Degree
Ph.D.
Advisors
Mengel, Purdue University.
Subject Area
Agronomy|Agricultural chemicals
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.