Geostatistics applied to the study of the spatial variation of soil fertility using Kriging

Abstract

The objective of this work was to evaluate the impact of soil management on some soil fertility properties and their spatial distribution. The study was conducted on 3 lots of a Typic Hapludands soil, dedicated to forest, sugar cane and coffee. Soil sampling was carried out using a regular grid with 38 geo-referenced points 40m from each other. Data were interpolated usin Kriging and analyzed by traditional statistics as well. Maps made from interpolated data showed correspondence between the type of management and the variation of evaluated chemical properties on space. Sugar cane soil had the lowest values on Ca, Mg and K, which was related to a highest absorption and low recuperation by fertilization of those nutrients to the soil. Coffee soil had the highest values of P, Ca, Mg and K due to the intensive fertilization practice, as well as a tendency to higher exchangeable soil acidity due to heavy N fertilization. The forest area had values between sugar cane and coffee. The traditional statistics analysis detected these general differences, but lacked the capacity for showing the changes on the spatial level and the gradient in concentration of the elements as well. It was concluded that spatial distribution of the soil fertility properties was strongly related to that the agronomic management and that the interpolation analysis was reliable and useful for predicting this landscape distribution.