Título: Um Novo Método para a Predição do Sequestro de Carbono em Áreas de Reflorestamento usando Rede Neural
Autores: Nunes, Clodoaldo; Martinho, Valquíria R. C.; Minussi, Carlos R.
Resumo: Economic and industrial activities have increased the rate of emission of greenhouse gases (GHG) on the planet. In recent decades, there has been a process that’s going to increase the temperature of the globe, depending on the rate of increase of emissions of these greenhouse gases, mainly from emissions of carbon dioxide (CO2). Thus, it becomes essential to adopt measures to reduce carbon emissions, and also, the development of efficient methods for quantifying the flow of carbon into the atmosphere. This research investigates an intelligent system to quantify emissions and carbon sequestration in deforested areas, in the short, medium and long term. The proposed system consists of a combination ART neural network-fuzzy architecture and a multilayer feedforward neural network with training performed through the use of the retropropagation algorithm, making the quantifying and predicting of the carbon absorption more accurate and precise than the original methods. Aiming to test the proposed system, we present an application in an area located in a reforested area in the Amazon region of Mato Grosso-Brazil, on a farm with land area of approximately 10.000 hectares.
Palavras-chave: Estimation of Carbon; Neural Networks; Reforestation Area
Código DOI: 10.21528/CBIC2013-004
Artigo em pdf: bricsccicbic2013_submission_4.pdf
Arquivo BibTex: bricsccicbic2013_submission_4.bib