Sampling redesign of soil penetration resistance in spatial t-Student models

Aim of study: To reduce the sample size in an agricultural area 167.35 hectares, cultivated with soybean, to analyze spatial dependence soil penetration resistance (SPR) outliers. Area Cascavel, Brazil Material and methods: The reduction was made by univariate effective ( ) methodology, assuming that t-Student model represents probability distribution SPR. Main results: radius intensity have inverse relationship estimated value . For depths SPR dependence, highest reduced 40%. From new size, sampling redesign performed. accuracy indexes showed differences between thematic maps original designs. However, lowest values standard error parameters structure evidenced design appropriate. Besides, models semivariance function were efficiently estimated, which allowed identifying existence all depth SPR.Research highlights: 40%, allowing for lesser financial investments data collection laboratory analysis samples next mappings area. able influence outliers structure.