Uso de indicadores para estimativa da produção de CO2eq empregando rede neural artificial

Abstract

Currently, global warming is becoming increasingly evident. The urgency to take action to address the effects of this phenomenon is crucial to minimizing increasingly severe consequences for biodiversity and societies. June 2024, so far, is the hottest month in the modern history of planet Earth. This work aims to generate estimates of CO2 equivalent emissions based on economic indicators, allowing decision-makers to predict the impact of economic actions on CO2 equivalent generation. Thus, based on the economy, CO2eq emissions can be estimated, enabling a better approach to their environmental impacts. The price of corn, oil, exchange rate, Ibovespa, and GDP were input into a two-layer artificial neural network, with 5 input neurons and 1 output neuron, a bias of 0.12, and a learning rate of 0.18. CO2eq emission estimates for the years 2023 and 2024 were generated, compatible with the Greenhouse Gas Emissions and Removals Estimation System (SEEG). The combination of economic data with artificial neural network resources allows for projecting expected greenhouse gas emissions. This information can be valuable for addressing this global problem, enabling an understanding of the sustainability of economic growth and changes in the way wealth is generated. It is a tool to quantify CO2eq emissions, a reality that needs to be faced.