Geoestatística na avaliação dos atributos físicos em latossolo sob floresta nativa e pastagem na região de Manicoré, Amazonas

Abstract

The use of geostatistical techniques allows detection of the existence of dependence and the spatial distribution of soil properties, thus constituting an important tool in the analysis and detailed description of the behavior of soil physical properties. The aim of the present study was to use geostatistics in assessment of physical properties in a Latossolo (Oxisol) dystrophic under native forest and pasture in the Amazon region of Manicore. Grids with of 70 × 70 m were established in native forest and pasture, and points were marked in these grids spaced at every 10 m, for a total of 64 points. These points were then georeferenced and in each one, soil samples (128) were collected at the depths of 0.00-0.20 and 0.40-0.60 m for a survey of their physical properties. These grids are parallel at a distance of 100 m from one another. The following determinations were made: texture, bulk density and particle density, macroporosity, microporosity, total porosity and aggregate stability in water. After tabulating the data, descriptive statistical analysis and geostatistical analysis were performed. The pasture had a slight variation in its physical properties in relation to native forest, with a high coefficient of variation and weak spatial dependence. The scaled semivariograms were able to satisfactorily reproduce the spatial behavior of the properties in the same pattern as the individual semivariograms, and the use of the parameter range of the semivariogram was efficient for determining the optimal sampling density for the environments under study. The geostatistical results indicate that the removal of native forest for establishing pasture altered the natural variability of the physical properties.