Ciclo de vida domiciliar, ciclo de vida do lote e dinâmica da cobertura da terra em Machadinho, Amazônia Brasileira

Abstract

Deforestation is one of the main problems associated with environmental degradation of the Amazon borders. Much of the high rates of deforestation in the Amazon are the result of human occupation that generates clearing the forest to create fields and pasture. This occupation was the result mainly of a settlement process induced by the federal government during the Brazilian military dictatorship, from the perspective of a specific regional development. The purpose of this dissertation was to use the theoretical frameworks of the border stages, of household life cycle, and of the property life cycle as a backdrop for understand the context in deforestation Machadinho D'Oeste, Rondônia, Brazilian Amazon, because of the actions of stablished family farms in the region in two different stages of the border. To better understand the process of deforestation in the light of these theories have been proposed two types of models: (i) level (dependent variable measured as square meters); (ii) proportion (dependent variable expressed as a proportion of total lot area). In the model (i) were used multiple linear regression (method of OLS, OLS with correction Hubber & White and weighted least squares) and truncated and multiple regression (Tobit model). In model (ii) the Beta regressions were used, the Generalized model, and the correction per fraction and the regressions Beta Zero inflated, the proportion multivariate regression for combined beta-logistic distribution. The results show that the household life cycle markers, property life cycle markers and integration with the markets markers presented reasonable empirical support to explain deforestation in the Machadinho D'Oeste. In addition, the study shows that deforestation patterns increase non-linearly with the household life cycle, increasing its intensity in the consolidation period of the property and reversing this trend as households age. At the same time, it was shown that there is a non-linear effect of residence time in the property on deforestation, suggesting that the early years are predominantly years of land experimentation.