RESUMO Os atributos do solo podem ser facilmente alterados pelo uso e manejo, comprometendo a capacidade produtiva das áreas. O objetivo da pesquisa foi avaliar a variabilidade e inter-relações dos atributos físicos, químicos e estruturais do solo em agroecossistemas no Semiárido Potiguar Brasileiro, apontando os atributos mais sensíveis na distinção dos ambientes por meio da análise multivariada. A pesquisa foi realizada na comunidade rural de Piracicaba, município de Upanema, Rio Grande do Norte, região Semiárida Brasileira. Foram selecionados quatro agroecossistemas: Área de Mata nativa (AMN), Área de consórcio milho feijão (ACMF), Área de Pastagem (AP) e Área de Cajueiro (AC). Foram coletadas amostras de solo indeformadas e deformadas nas camadas 0,00-0,10, 0,10-0,20 e 0,20-0,30 (m), e realizadas as análises dos atributos físicos, químicos e estruturais. Os dados foram submetidos a técnicas de estatística multivariada, por meio da matriz de correlação, análise de agrupamento, análise fatorial e componentes principais. A análise fatorial revelou os atributos mais sensíveis na distinção dos agroecossistemas, sendo microporosidade, capacidade de campo (CC), ponto de murcha permanente (PMP), água disponível (AD), areia, argila, grau de floculação (GF), Ca2+, Mg2+ e soma de bases (SB) (F1) e densidade do solo (DS), porosidade total (PT), macroporosidade, condutividade elétrica (CE) e carbono orgânico total (COT) (F2). A análise de agrupamento formou três grupos distintos. O COT foi discriminante para agrupar as áreas de mata nativa e de cajueiro, evidenciando a influência do manejo do solo. Os demais ambientes foram agrupados pela diferenciação da textura influenciando na variabilidade dos atributos.

ABSTRACT The soil attributes can be easily changed according to its management and use, compromising the productive capacity of the areas. The objective of this research was to evaluate the variability and interrelationships of soil physical, chemical and structural attributes in agroecosystems in the Brazilian Semiarid region (Rio Grande do Norte State), pointing out the most sensitive attributes in the environment distinction through multivariate analysis. The research was carried out in the rural community of Piracicaba, Upanema municipality, Rio Grande do Norte State, Brazil. Four agroecosystems were selected: Native Forest Area (ANF), Corn Bean Consortium Area (ACBC), Pasture Area (AP) and Cashew Area (AC). Undisturbed and deformed soil samples were collected in the 0.00-0.10, 0.10-0.20 and 0.20-0.30 m layers, and the physical, chemical and structural attributes were analyzed. The data were submitted to multivariate statistical techniques, using the correlation matrix, cluster analysis, factor analysis and principal component analysis. Factorial analysis revealed the most sensitive attributes in distinguishing agroecosystems: microporosity, field capacity (FC), permanent wither point (PWP), available water (AW), sand, clay, flocculation degree (GF), Ca2+, Mg2+ and sum of bases (SB) (F1) and soil density (SD), total porosity (PT), macroporosity, electrical conductivity (EC) and total organic carbon (TOC) (F2). Cluster analysis formed three distinct groups. The TOC was discriminant to group the native forest and cashew tree areas, showing the influence of soil management. The other environments were grouped according to the texture differentiation, influencing the attributes variability.

Resumen La erosión es la principal forma de degradación en el mundo, en consecuencia de la pérdida de suelo, agua y nutrientes, provocando daños directos e indirectos. El objetivo de este trabajo es cuantificar las pérdidas de N, P y K por erosión hídrica en condiciones de lluvia simulada mediante la aplicación de seis pruebas de erosión en cultivo continuo de especies de ciclo estacional, incluyendo cultivos aislados y cultivos intercalados. El estudio se llevó a cabo en el campo en la Estación Experimental de Agronomía en Eldorado do Sul (RS), Brasil, en un Argisol Rojo Distrófico típico. Para el análisis de la pérdida de nutrientes, se aplicaron seis ensayos de erosión con lluvia simulada a los tratamientos entre finales de octubre y mediados de diciembre, utilizando el simulador de lluvia con brazos giratorios de Swanson. Las pérdidas de N y P se vieron influenciadas por las condiciones de la superficie, con la eliminación parcial de los residuos del cultivo y la no movilización del suelo. La excepción fue K, debido a su alta solubilidad en el suelo con baja actividad coloidal, que se perdió fácilmente en la escorrentía superficial del agua del suelo, independientemente de la cobertura total de la superficie del suelo por los residuos de los cultivos.

Abstract Erosion is the main form of degradation in the world, consequently of loss of soil, water and nutrients, provoking direct and indirect damages. The objective of this work is to quantify N, P and K losses by water erosion under simulated rainfall conditions by the application of six erosion tests in continuous cultivation of seasonal cycle species, including isolated and intercropped crops. The study was carried out in the field at the Experimental Station of Agronomy in Eldorado do Sul (RS), Brazil, in a typical Dystrophic Red Argisol. For the analysis of nutrient loss, six erosion tests with simulated rainfall were applied to the treatments between the end of October and mid-December, using the rain simulator with swivel arms of Swanson. The losses of N and P were influenced by surface conditions, with the partial removal of crop residues and the non-mobilization of the soil. The exception was K, due to its high solubility in the soil with a low colloidal activity, which was easily lost in the soil water surface runoff, regardless of total soil surface coverage by crop residues.