Eficiência de malhas amostrais utilizadas na caracterização de atributos químicos em latossolos manejados com agricultura de precisão

Abstract

The georeferenced soil sampling using sampling grids is the main strategy of precision agriculture (PA) to recognize the spatial variability of soil attributes, allowing manage them in site-specific. In this context, the objective of the work was to evaluate the efficiency of using different sizes of sampling grids in the characterization of spatial variability of chemical attributes in Oxisols managed with PA in Southern Brazil. Therefore, the work consists of three studies. In studies one and two were used 60 areas (6,046.55 ha), located in 23 counties of North region of Rio Grande do Sul (RS). The areas were sampled using regular grids of 100 x 100 m (20 areas), 142 x 142 m (20 areas) and 173 x 173 m (20 areas), being that 10 areas of each grid were collected from 0.00-0.10 m and 10 areas from 0.00-0.15 m. The spatial variability of acidity attributes (pHwater and base saturation) and bases related (Ca and Mg) (study one) and the levels of P and K (study two) were analyzed by descriptive statistics and geostatistics. In study three, was used an area of 41.96 ha, located in Boa Vista das Missões - RS, where was made the collect of soil in seven sizes of sample grids, 50 x 50 m, 75 m x 75, 100 x 100 m, 125 m x 125, 150 x 150 m, 175 m and 200 x 175 x 200 m, at depth of 0.00-0.10 m. The levels of P and K were analyzed by descriptive statistics and geostatistics, and the similarity of thematic maps was compared by the coefficient of relative deviation (CRD) and by Pearson correlation matrix (p <0.05). The areas of Oxisols managed with PA in RS, show moderate acidity and high fertility, with the exception of subareas that presents limiting P levels. Independently of the depth sampled, the grid sample sizes used in RS state, in general, are not efficient in capturing the different scales of variability in soil chemical properties. According as increase the sample grid the thematic maps of P (CRD: 24.0 to 36.2%) and K (CRD: 11.7 to 19.4%) show more dissimilar to those obtained in smaller grid (50 x 50 m) considered as reference. Thus, the reduction of sampling grid sizes increases the accuracy of the information generated through thematic maps, allowing to do prescriptions of lime and fertilizers on site-specific with greater efficiency, and should be recommended in future sampling plans of soil adopted in PA areas in Southern Brazil.