Abstract
Recent global agreements sponsored by UN to sustainable development and reduction of greenhouse gases emission, undergo by the reduction of deforestation and an increase of productivity of the areas occupied by agricultural production systems. In 2014 the Brazilian Amazon deforestation reached 760.305,5 km2. This area has different land uses with different financial and social returns and different environmental impacts. To know the land use and land cover (LULC) and the predominant production systems in this deforested area is a basic condition for planning actions and public policies for sustainable development. This study developed methodologies for detailed LULC mapping, and to regionalization of agricultural production systems. The mapping of LULC of Brazilian Amazon confirmed previous studies, showing that most of the deforested area is used for livestock activity (about 60%), followed by secondary vegetation (about 20%) and annual crops (about 5%). Regionalization of agricultural production systems was carried out in two stages. The first one was carried out at local administrative boundary, using spatially explicit data of LULC and deforestation, analyzed in conjunction with municipal socioeconomic data spatialized at local level. Based on these results we identified, among five agricultural production systems at the state level. The dimension of territorial configuration showed that each agricultural production system form different types of landscapes. The second stage was conduced with cells of 10 X 10 Km, with LULC spatially explicit data, deforestation and a matrix of distance to infrastructure elements, and using landscape metrics. Was possible to classify the predominant agricultural production system in each cell, and identify the effect of these systems on the landscape. Nine agricultural production systems were found: two in forest domains (Forest Domain, Initial Front), three in agriculture domain (Strict Agriculture, Dominant Agriculture and Coexistence Area) and four in livestock domain (Intensified Beef, Not Intensified Beef, Intensified Beef-Milk and Not Intensified Beef-Milk). The methodologies resulted from this thesis will enable the creation of an operational monitoring system of LULC, continuous and at low cost, and also of the prevailing agricultural production system in each of the deforested territory. With this monitoring system, it will be possible to follow up the effect of public policies in the region, looking for sustainable development.