Carbon emission forecasting and scenario analysis in Guangdong Province based on optimized Fast Learning Network

As the most economically developed province in China, Guangdong is facing the severe challenge of reducing carbon emissions. The aim of this study is to explore whether Guangdong can achieve the carbon emission peak by 2030 and the carbon neutrality by 2060. We calculate the energy-related carbon emissions, technology carbon emissions from cement production and forest carbon sinks from 1995 to 2019, and construct a Fast Learning Network (FLN) forecasting algorithm improved by Chicken Swarm Optimization (CSO) to predict carbon emissions in 2020-2060. The superiority of CSO-FLN model is confirmed by three error indicators (MAE, MAPE, RMSE). Based on the different change rates of carbon emission influencing factors, nine scenarios are set up to pursue low-carbon development paths. The results show that (1) Guangdong's carbon emissions generally showed an upward trend in 1995-2019; (2) The forecasting effects of CSO-FLN model surpasses Fast Learning Network (FLN) and Extreme Learning Machine (ELM) by comparing the three error indicators; (3) For Guangdong Province, the peak of carbon emissions can be achieved by 2030 or before only under scenario 3, scenario 4, scenario 5 and scenario 7, while only on scenario 3, can carbon neutrality be realized by 2060. According to the research results, some countermeasures and suggestions to reduce carbon emissions are put forward.