Forecasting of Brazilian CO2 emissions from nonlinear models: a review

Abstract

Brazil is a developing country that emits high amounts of CO2 and, in order to comply with the Paris Agreement, it has committed to reducing 43% of its emissions by 2030 in relation to 2005. To achieve this objective, it is necessary to understand which variables contribute with these emissions and develop Energy Policies that reduce them without harming the country's energy generation, which is fundamental to its development. To understand the main factors that affect Brazilian CO2 emissions and predict their trend in the coming years, mathematical modeling has been used as a tool. This article provides a synthesis of the literature, presenting the main methods used to predict Brazilian CO2 emissions, as well as the variables with the greatest influence on them. From this review, it is concluded that non-linear mathematical models, such as Artificial Neural Networks and Gray Model, are more accurate in their predictions compared to linear models. Furthermore, variables such as economic growth and a country's energy consumption have a great influence on CO2 emissions, especially when fossil fuels are the main energy source.