Kriging Versus Cokriging and Collokated Cokriging for Soil Physical-Hydraulic Attributes and their Influence on Soybean Growth

ABSTRACT

Abstract in Brazil, management agricultural practices not currently consider the soil spatial variability as a result, crop growth can be non-uniform and yields often is low. This research aims to compare Kriging, Cokriging and Collocated cokriging using soil physical and hydraulic properties and their influences on soybean development. We hypothesized that spatial variability of physical and hydraulic properties has influence on soybean development and this variability can be better represented by Collocated Cokriging method. To test these hypotheses, we accessed the soil physical and hydraulic attributes in a field experiment under no-till system, cultivated with soybean. Geostatistical interpolators were applied to generate maps from which spatial dependence of the variables was evaluated. The experiment was conducted on a sandy clay loam Oxisol, on an experimental station located in Ponta Grossa, Paraná, Brazil. Evaluation of the soil attributes was performed bulk density (BD), particle size distribution, saturated soil hydraulic conductivity (K fs ), total porosity (TP), macroporosity and microporosity. The plant was plant height and stand. Data analysis were performed by geostatistical methods; the spatial dependence was established using experimental univariate and cross semivariograms with datasets. Modeling semivariograms led to the generation of attribute maps by Kriging, Cokriging and Collocated cokriging. The estimation by Cokriging and Collocated cokriging was similar from Kriging. From the semivariogram, it was possible to identify that soil and plant attributes were spatially related with each other. The soya growth was mainly changed by slope of the area and little changed by saturated hydraulic conductivity.