Análise de regressão múltipla com estudo relacionado à emissão dos gases de efeito estufa na Amazônia

Abstract

Studies aimed at preserving the environment are becoming more common in the Statistics area and, therefore, it gets even more necessary to understand the reliability of those studies. This work analyzes models of multiple linear regression to predict future values of greenhouse gas emissions in the Amazon forest. It has a brief installation tutorial and a guide to some of the statistical analysis on SAS University Edition. The database used in this study is from the “Instituto Nacional de Pesquisas Espaciais” that includes estimates of the main greenhouse gases emissions in the Amazon: carbon dioxide, methane and nitrous oxide. The regression analysis showed that the deforested area and the absorbed carbon dioxide are the variables that are most related to the greenhouse gas emissions and, therefore, are present as regressors in the prediction models for carbon dioxide and other gases. From the results, we can conclude that, as the deforested area increases, the emissions of polluting gases increase, since the decomposition and burning of organic material in the vegetation releases those gases. On the other hand, as the absorbed carbono dioxide increases, that is, the more preserved areas, the lower the emissions of greenhouse gases.