Análise estatística da fertilidade e atributos físicos do solo: áreas de Santarém e Terra Alta-PA.

Abstract

Human intervention in the forest, implies some degree of change in stmcture and ecosystem fimctioning, although some operations are non-destmctive and can be sustained for a reasonably long period. Amazonian soils seem particularly inclined to rapid declines in natural fertility. One reason for this is the mid and low levei of soil organic matter, generally associated with rapid mineralization of carbon and nutrients due to the favorable moisture conditions during most of the year. Agroforestry systems aims to optimize land use, reconciling forest production with food production, conserving soil and reducing the pressure for land use for agricultural production. The soil under agroforestry has lower density, higher porosity, lesser resistance to penetration and higher aggregate stability, compared to the same soil under conventional tillage. In Brazilian agriculture, currently, large areas are considered homogeneous; them the average need of use of inputs promotes the use of equal doses of fertilizers, disregarding the spatial variability and can overwhelm a plot that is fertile and not reach optimal leveis for high productivity in other disabled plots. The spatial variability of soil has always existed and should be considered every time the field sampling is made, it may indicate places that require differential treatment as the management, subject to the representativeness, allowing greater detail the area. The use of methods such as multivariate analysis and geostatistical proved to be a fundamental tool to study the spatial variation of soil properties and produce high quality maps. Soil nutrient Índices have been used to produce maps of spatial soil classification, in order to apply fertilizers in specific and strategic locations, which could save resources and preveni potential environmental problems derived from super or sub-applications. The fundamental theory of geostatistics is the hope that, on average, the next samples in time and space are more similar to each other than those who are distant. The geostatistics has proved very useful in soil science to characterize and study the spatial variation of its properties. This study aimed to provide Information on the variability of soil attributes using alternative methods of statistical analysis and multivariate analysis and geostatistics. Thus the work was divided into two chapters; Chapter 1 was an analysis of the fertility variability of grain production areas in Santarém using multivariate analysis methods, and in Chapter 2, sought to investigate the variability of soil physical attributes in an agro-ecosystem using geostatistics.